20

Guide on the Best

Practice for Tanker
Performance Monitoring
 Guide on the Best Practice for
Tanker Performance Monitoring

All rights reserved. No part of this publication may be reproduced in any material form (including photocopying
or storing it in any medium by electronic means and whether or not transiently or incidentally to some other
use of this publication) without the written permission of INTERTANKO. Applications for INTERTANKO’s written
permission to reproduce any part of this publication should be addressed to the publisher.

© INTERTANKO 2020

While every effort has been made to ensure that the information contained in this publication is correct,
neither the authors nor INTERTANKO can accept any responsibility for any errors or omissions or any
consequences resulting therefrom.

No reliance should be placed on the information or advice contained in this publication

without independent verification.

Cover image courtesy of LATSCO Marine Management Inc. Vers.1 – Nov 2020
About this document and Notice of Terms of Use

This Guidance document has been produced with help from the INTERTANKO Working Group on Performance
Monitoring.

While every effort has been made to ensure the accuracy of the information provided, none of the contributing
individuals, organisations or associations nor any of their members past, present or future warrants its accuracy
or will, regardless of its or their negligence, assume liability for any foreseeable or unforeseeable use made
of this publication. Consequently, such use is at the recipient’s own risk on the basis that any use constitutes
agreement to the terms of this disclaimer. The information contained in this publication does not purport to
constitute professional advice from the various content contributors, and none of the contributing individuals,
organisations corporations or associations nor their members accept any responsibility whatsoever resulting
directly or indirectly from applying or relying upon recommendations contained in the Guidance, even if caused
by a failure to exercise reasonable care.

This document may provide guidance supplemental to the requirements of international and local regulations.
However, nothing herein is intended to replace, amend, supersede or otherwise depart from such requirements.
In the event of any conflict or contradiction between the Guidance contained in this document and regulatory
requirements, applicable laws shall prevail.

Any enquiries regarding this document should be sent to us via [email protected] and
[email protected].
Contents

1. Scope  7
2. Performance Monitoring in the tanker industry 8
2.1 Background  8
2.2 Future scope of work 10

3. Regulatory framework  11
3.1 IMO CO2 reduction framework  11
3.2 IMO Data Collection System (DCS)  11
3.3 IMO Green House Gas Initial Strategy  12

4. Industry standards and commercial requirements  13

4.1 ISO 15019 and ITTC recommended procedure  13
4.2 ISO 19030 14
4.3 Charter Party and TMSA  14
4.4 Energy Efficiency Management Plan 16
4.5 Examples of ship energy efficiency measures and Energy Efficiency  16
Technology (EET)
4.6 Performance indices 16
4.7 Data intellectual property  18

5. Trend analysis  19
5.1 Questionnaire  19
5.2 Results of the online poll  19

6. Understanding of uncertainty  21
6.1 Basic definition 21
6.2 Sources of error  22
6.3 Introduction to uncertainty in vessel performance 22
6.4 The process of understanding uncertainty in vessel performance  23
6.5 Baseline uncertainty 23
6.6 Proxy errors 24
6.7 Uncertainty in parameters used in analysis 25
6.8 Uncertainty in sensor measurements 26
6.9 A case study in applying different methods recommended in ISO 19030 26

7. Data source  28
7.1 Auto-logging sensors onboard the ship 28
7.2 Performance reporting 31
7.3 Quantification of uncertainties in manual reporting and auto-logging 31
7.4 Uncertainties in manual reporting 32
7.5 Manual reporting systems 35
8 Data analysis  36

9 Data driven decision  40

9.1 New building data and sea trial 40
9.2 Fuel management 40
9.3 Hull cleaning and propeller polishing 40

10 Summary  42
Takeaways 43
References  44
Contributors 45

Appendices46
Appendix 1. Basic principles of ISO 19030 (Section 4.2) 46
Appendix 2. Background of the speed loss concept in ISO 19030  51
(Section 4.2)
Appendix 3-1 Examples of operational measures for energy efficiency  54
(Section 4.6)
Appendix 3-2 Classification of Energy Efficiency Technologies (EET, Section 4.6) 57
Appendix 4. Non-Disclosure Model Agreement (Section 4.8) 60
Appendix 5. Results of the online poll (Section 5) 63
Appendix 6. Theoretical background of uncertainty analysis (Section 6.2) 76
Appendix 7. Uncertainty in sensor measurements (Section 6.5)  82
Appendix 8. Uncertainty, today and ISO 19030 (Section 6.9 and 7.4)  90
Appendix 9. How to measure shaft power (Section 7.1) 95
Appendix 10. Energy Efficiency Management Plan – A case study of a  97
shipping company (Section 9.1)
Appendix 11. Information to be agreed at technical specification stage  104
(Section 9.1)
Appendix 12. How to manage fuel oil consumptions using flow meter  106
(Section 9.2)
Appendix 13. Explanatory notes for Specific Fuel Oil Consumption (Section 9.2) 109
Appendix 14. Hull-propeller performance assessment methods 112
(Section 9.3.2)
Appendix 15. A case study into decision-making processes about propeller  114
polishing and hull fouling (Section 9.3.1 and 9.3.2)
List of abbreviations

AIS Automatic Identification System

ASP Analogue Signal Processing

CFD Computational Fluid Dynamics

CSI Clean Shipping Index

DCS IMO (Fuel Consumption) Data Collection System

EEDI Energy Efficiency Design Index

EEOI Energy Efficiency Operational Indicator

ESI Environmental Ship Index

EVDI Existing Vessel Design Index

GHG Green House Gas

GPS Global Positioning System

HFO Heavy Fuel Oil

HullPIC Hull Performance & Insight Conference

IMO International Maritime Organization

IGS Inert Gas System

ISO International Standard Organization

ISTEC INTERTANKO Safety & Technical Committee

ITTC International Tank Towing Conference

LNG Liquefied Natural Gas

MCR Maximum Continuous Revolution

MEPC Maritime Environment Protection Committee

MRV EU Monitoring, Reporting and Verification of CO2 emissions

OCIMF Oil Companies International Marine Forum

PLC Programming Logic Circuits

RPM Revolutions Per Minute

SCDA Supervisory Control and Data Acquisition

SEEMP Ship Energy Efficiency Management Plan

SFOC Specific Fuel Oil Consumption

SOG Speed Over Ground

SOW Speed Over Water

SPI Shipping Performance Index

WMO World Meteorological Organization

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  5

INTERTANKO WG/Performance Monitoring Chairman’s notes

INTERTANKO WG/Performance Monitoring

Chairman’s notes
The INTERTANKO Working Group on Performance Monitoring (“WG”) was established in 2017. During the
course of its activities, it has involved not only ship owners but also other industry stakeholders including ship
managers, paint manufacturers, service suppliers, sensor suppliers and Class Societies. The WG Members’
multiple disciplines put the WG in a unique position to share knowledge and experiences and to focus on
deliverables. The nature of this project involved multiple brainstorming sessions, which was a recognition of
the complexity of what the WG was expected to achieve.

This Guide contains a valuable set of principles that set out the approaches adopted in ISO standards, regulatory
framework reshaping the industry and the best practices relating to ship performance monitoring. It will assist
INTERTANKO Members in understanding the complexities of the subject, formulating their respective policies
and implementing associated industry standards, rules and regulations that are suitable to their economic,
social and environmental circumstances.

The WG finalised the development of the Guide on 28 August 2020 and it was approved by the INTERTANKO
Safety & Technical Committee (ISTEC) on 21 September 2020. It will be later expanded and updated as
necessary.

As Chairman of the WG, I would like to encourage all Members of INTERTANKO to promote better and wider
understanding of vessel performance monitoring because the subject requires a degree of understanding in
theoretical background of ship design, operation and maintenance, effectiveness of various energy saving
devices and their limitations. The WG attempted to bridge many gaps between design and maintenance,
and theory and practice, bringing the border between industry standards and the International Maritime
Organization’s (IMO’s) greenhouse gas (GHG) policy at close-in range.

INTERTANKO Members wish to express sincere thanks to all participants and particularly to the tireless work
of the Secretary of the WG, Gil-Yong Han, without whose project leadership this guideline would never have
been published.

Francesco Bellusci
Chairman of the INTERTANKO WG/Performance Monitoring
(Managing Director of Scorpio Group)

6 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 1. Scope

1. Scope
Evaluating the efficiency of a ship is not an easy task. Ships are different to other means of transportation,
moving on the boundary of two different fluids, air and water. The ship’s interactions with these fluids results
in a unique energy dispersion phenomena, the influences of which makes it difficult to calculate and forecast
the combined effect.

Despite rapid progress of Computational Fluid Dynamics (CFD) techniques, the ship’s contractual performance is
still measured with sea trials and propulsion is pre-evaluated by lengthy and costly model tests using correlation
factors that have been empirically derived.

While a model test is a laboratory test at a certain scale, the sea trial is a full-scale replica. This is a verification
procedure of the model test results with experts applying their knowledge and experience together with
movable data loggers, with the aim of eliminating the interference of external factors and establishing the
relationship between power and speed as accurately as possible. However, it is difficult to carry out such sea
trial tests periodically during the ship’s in-service period.

The difficulties with assessing ship performance in simple algorithms have led to the development of vague
contractual clauses to cover all possible unexpected variations. In the tanker industry, the Charter Party
describes a ship’s ability to meet a few parameters, for example “in laden condition about 13 knots on about
30 tons of fuel a day up to Beaufort 5” with the Master’s obligation to report such a performance based on
their assumptions.

For years, the difficulties of monitoring the ship’s performance, the ambiguous performance requirements of
the Charter Party and obligations arising from the Bill of Lading signed by the Master are some of the reasons
why interest in performance monitoring has been low in the past.

Things are now changing. owners, Charterers and institutional investors are looking for tools to understand
their assets’ performance and carbon exposures.

The emerging interest on monitoring performance can be seen in times when the freight market was very
poor and fuel costs were high or, as is the case now, when new and challenging regulations make monitoring
of the ships’ performance mandatory. There is an ever-growing body of new regulations on climate change,
ship’s energy efficiency, fuel consumption reporting and GHG reduction targets. There is also a movement in
financial institutions pushing for more detailed carbon footprints-related financial disclosure.

With this background, the INTERTANKO Safety and Technical Committee (ISTEC) established a Working Group
with the aim of developing an industry best practice for a ship’s performance monitoring. Initially, the plan
was to collect information on experiences gained from the use of the ISO 19030 standard and, if appropriate,
to provide a set of suggestions to assist its Members in applying their fleet performance monitoring in an
easy and efficient manner. This was in response to the needs of stakeholders who were looking for tools to
understand a set of data generated from ship performance monitoring and its meaning.

In addition, the Working Group considered additional steps to enable the owners to monitor the trend of their
ships’ performance and to take decisions to intervene by applying physical means to improve ship performance
or improving the accuracy of the methods applied.

This Best Practice Guide examines the usefulness of various energy efficiency devices and the methods applied.
It will help the ship operators to benchmark performance of their sister ships and even determine how well
their different types of ships are performing.

It should be noted, however, that the scope of this Best Practice is not to invent a new means of monitoring
performance nor to pursue intellectual property rights by presenting a new standard or amendments to the
ISO 19030 standard.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  7

2. Performance Monitoring in the tanker industry

2. Performance Monitoring in the tanker industry

2.1 Background
In the 1980s, ship performance monitoring was given a low priority despite its “relative” importance. Noon
reports were still on telex, GPS data was not properly evaluated and radars were kept switched off to avoid
costly repair and complaints from companies.

Once fax replaced telex, the initial idea involved building up a comprehensive noon report with machinery and
weather data trying to evaluate hull and propeller fouling conditions and machinery performance.

With the advent of an email communication system, noon reports became an excel file and data was ready to
be analysed despite inherent difficulties. A very similar format of the noon report is still in use 20 years later.

Next came a focus on commercial performance measured on a single speed and a single consumption up to
Beaufort 5, and there were plenty of ways to avoid underperformance. The freight market, although cyclical,
was stabilised at a good average rate, the cost of fuel was low and there was no focus on performance
monitoring.

With the introduction of a GPS tracking systems, shipping companies started to look at voyage performance,
comparing the Master’s declared weather data with the external provider’s weather data, the Master’s estimated
position with the GPS based one, and the Master’s reported ship speed against the speed over ground (SOG).
Building more efficient ships was on everyone’s mind. Suddenly, the 2002-2008 period saw huge capital
invested into the shipping industry. The global trade was steadily on the increase and the freight market was
so high that the impetus for increasing ship performance and energy efficiency was reduced to almost zero,
with the focus again placed on delivering high block coefficient vessels with maximum cargo-carrying capacity
as quickly as possible with standard (better than saying ‘substandard’) efficiency.

Things have changed since 2008. The freight market collapsed, fuel costs increased, and emission control
measures slowly began to dominate regulatory framework. Wise ship owners will continue to adapt to this
changing regulatory environment and lead the industry with innovative designs and operation.

Efficiency expressed in terms of fuel consumption (or equally emission) became a nightmare (or an opportunity)
for shipping owners, designers and regulators. Everyone at all levels in shipping started to sell or advertise ‘eco’
designs, fuel efficiency devices and outperforming paint systems. However, do they provide the efficiency gains
that they promised to deliver and how do we best measure those efficiency gains?

No efficiency devices can be managed if the efficiency is not measurable according to solid and recognised
standards. Existing standards were mainly aimed at defining contractual speed/power points for sea trials of
new building ships while monitoring ship performance during operation was left to the shipping company
alone. Again the primary focus was on commercial performance, of which its vague definition had not been
changed in tandem with the degree of understanding gained within the industry.

Noon report data was the only tool at the time to check the effect of an eco-design or other energy efficiency
device.

A new standard for sea trial was introduced (ISO 15016) which was again a compromise between technical and
economical interests and with a number of undefined indicators about current effects and using overcorrecting
algorithms for weather factors.

8 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 2.1 Background

The Energy Efficiency Design Index (EEDI) regulations, introduced post 2008, took a leading role as a GHG-
regulating effort in new building and is based on the ISO 15016 standard.

It was after 2008 that for the first time we began to see ship designers no longer looking at an efficient new
building ship balance sheet: new hull forms with lower Cb and improved wake factors were introduced,
engines with high torque plus low revolution were introduced by engine manufacturers, bigger propeller
diameters were re-introduced and various fuel saving devices were invented or re-invented.

Despite these developments, the question still remains - how do we best monitor the ship’s efficiency
continuously after delivery including the effects of fuel saving devices?

Data sensors and data loggers were introduced to capture performance-related parameters, which are listed
in ISO 15060, such as torsionmeters and flow meters.

Producers of fuel-saving devices - and not only ship owners, began needing more precise monitoring methods
and were all looking at new opportunities.

A two-tier market for eco and non-eco ships was naturally created: huge interest evolved, experience grew
and regulatory pressures on emissions reduction triggered a group of experts to develop the first performance
monitoring standards, ISO 19030, albeit with a limited scope. It was intended to monitor a trend in changes
of ship performance, not to attempt to calculate absolute values.

In parallel, a structural consolidation of the collapsed freight market with reduced (or negative) margins and
the stabilising fuel oil cost at higher levels began to act as a catalyst to make fuel saving a business strategy,
not merely fulfilling regulatory requirements.

Calls for ship owners to reduce GHG emissions from their ships present opportunities for development and
exploitation of new technologies aimed at providing meaningful data on the commercial and technical
performance of ships.

In case of bulkers and tankers, the difference of draft between a ballast voyage and laden voyage is huge.
Experience suggests that it is difficult to find a commonly agreeable efficiency indicator due largely to this
so-called ship utilisation factor. EU MRV and IMO DCS are based on unfiltered and un-normalised rough
commercial data.

For the same reasons, the Energy Efficiency Operational Indicator (EEOI) would not represent the ship’s energy
efficiency accurately. Many factors contributing to the EEOI are dependent upon the subject ship’s operational
profile rather than the technical one.

To overcome these challenges, it would be natural to look into ISO 19030, which addresses high frequency
data, as a starting point. Based on the experience gained from the application of ISO 19030, efforts can be
made to reduce the inaccuracy of data sources, data filtering and data normalisation processes. We can leave
this task to a group of industry experts who developed the first version of ISO 19030.

The whole industry is now experiencing a transition towards a data-driven, decision-making culture using big
data and digitalisation. It presents the best environment for performance-monitoring practitioners to look
beyond ISO 19030.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  9

2. Performance Monitoring in the tanker industry

2.2 Future scope of work

Once the background of performance monitoring methods is well understood, we can start to look at
practicalities beyond cleaning the hull and propeller over time.

Potential areas include:

• better understanding of the sea trial standard,

• calculation of EEDI,
• de-structuring various components and liabilities associated with so called commercial performance,

• efficiency of transportation assigning proportionate responsibilities to the right parties.

With emerging market-based instruments and defragmentation of shipping responsibilities amongst

complicated stakeholders, it is important to clarify how a commercial and technical decision should be taken,
and how such decisions are impacting the environment. The United Nations-termed “sustainable development”
shall be a shared responsibility rather than it being given to the last link in the chain of command, which is
normally a technical manager of a company.

We can start to challenge such time-honoured concepts in shipping as utmost despatch, no deviation, ordered
speed, etc., where the legal weight is so heavy that the development of more sophisticated technology and
new modes of operation is hitherto hindered. The new modes of operation include virtual time of arrival,
weather routing and speed optimisation.

Most importantly, it will be our task to examine various ways of complying with the IMO GHG reduction
targets.

10 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 3.2 IMO Data Collection System (DCS)

3. Regulatory framework
3.1 IMO CO2 reduction framework
International maritime transport is still the most energy-efficient mode of mass transport and only a modest
contributor to global CO2 emissions (2.7% in 2007) while carrying 90% of world trade. However, a global
effort for further improvements in energy efficiency and emissions reduction is still needed as sea transport
is predicted to grow significantly in line with world trade. The IMO’s work on enhanced energy efficiency
and GHG emissions control has three distinctive building blocks: technical measures, operational reduction
measures and market-based measures.

The most important technical measure is the Energy Efficiency Design Index for new ships (EEDI) that requires a
minimum energy efficiency level per capacity mile (e.g. tonne mile) for different ship types and size segments.
With the level being tightened incrementally every five years, the EEDI will stimulate continued technical
development of all the components influencing the fuel efficiency of a ship. After the EEDI regulation entered
into force in 2013, IMO, after re-evaluation of the incremental reduction factors phrased as “Phase 1, Phase 2
and Phase 3”, will adopt a set of amendments to the EEDI regulations which will advance implementation of
the EEDI regulations earlier than originally planned.

On the operational side, a mandatory management tool for energy-efficient ship operation, the Ship Energy
Efficiency Management Plan (SEEMP), was developed to assist the international shipping industry in achieving
cost-effective efficiency improvements in their operations using the Energy Efficiency Operational Indicator
(EEOI) as a monitoring tool and benchmark.

Development of the technical and operational measures was a remarkable step in assisting the shipping
industry to continuously reduce its GHG emissions. It was recognised that these measures alone would not
be sufficient to satisfactorily reduce the amount of GHG emissions from international shipping in view of the
growth projections of world trade. Therefore, market-based mechanisms had been considered at IMO over the
years under the GHG work plan. A market-based mechanism would serve two main purposes:

• off-setting in other sectors of growing ship emissions (out of sector reduction); and

• providing an economic incentive for the maritime industry to invest in more fuel-efficient ships and
technologies and to operate ships in a more energy-efficient manner (in sector reductions).

It remains to be seen how these market-based mechanisms will be shaped when the IMO GHG initial strategy
is fully enacted.

3.2 IMO Data Collection System (DCS)

IMO adopted mandatory MARPOL Annex VI requirements for ships to record and report their fuel oil
consumption in 2016. Under the amendments, ships of 5,000 gross tonnage and above are required to collect
consumption data for each type of fuel oil they use, as well as other, additional, specified data including proxies
for transport work. The aggregated data is reported to the Flag State after the end of each calendar year and
the Flag State, having determined that the data has been reported in accordance with the requirements, issues
a Statement of Compliance (SoC) to the ship. Flag States are required to subsequently transfer this data to
an IMO Ship Fuel Oil Consumption Database. IMO will be required to produce an annual report to MEPC,
summarising the data collected.

The first annual aggregated fuel oil consumption data on a ship-by-ship basis will be collected for the year of
2019 and published in 2021.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  11

3. Regulatory Framework

Figure 1: IMO DCS diagram, 2016 INTERTANKO

Prior to the IMO’s DCS, the EU enforced its own Monitoring, Reporting and Verification of CO2 emissions
from ships from 1 January 2018 (“EU MRV”). Both EU MRV and IMO DCS have a clear objective to drive the
shipping industry to reduce GHG emissions from ships. They are mandatory and are intended to be the first
step to provide accurate CO2 emissions data for the IMO’s future debates about the IMO’s long term GHG
strategy.

3.3 IMO Green House Gas Initial Strategy

In April 2018, IMO adopted an initial strategy on the reduction of GHG emissions from ships, setting out a
vision to reduce GHG emissions from international shipping and phase them out, as soon as possible in this
century. Under the identified “levels of ambition”, the initial strategy envisages for the first time a reduction in
total GHG emissions from international shipping which, it says, should peak as soon as possible and to reduce
the total annual GHG emissions by at least 50% by 2050 compared to 2008, while, at the same time, pursuing
efforts towards phasing them out entirely. The strategy includes a specific reference to “a pathway of CO2
emissions reduction consistent with the Paris Agreement temperature goals”.

Figure 2: IMO’s Initial Strategy on the reduction of GHG emissions from ships,
Resolution MEPC.304(72), adopted on 13 April 2018

12 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 4.1 ISO 15019 and ITTC recommended procedure

During the course of the development of this Best Practice, it was recognised that the ship operators shall
examine their ships’ CO2 emissions while keeping their reporting obligations for their ships as well as shore
systems and routines. By doing this, improvements in areas of energy efficiency can be identified and addressed
under the regulatory framework.

4. Industry standards and commercial

considerations
4.1 ISO 15019 and ITTC recommended procedure
Amongst many variables, the ship’s speed is the most important one to determine the ship’s contractual
performance at the time of sea trial and the ship’s operational performance during service.

For contractual performance, the purpose of speed and power trials is to determine the ship’s performance
in terms of ship’s speed, power and propeller-shaft speed under prescribed ship’s conditions and thereby
verify the satisfactory attainment of the ship’s speed stipulated by the Energy Efficiency Design Index (EEDI)
regulations and/or contract. The conditions stipulated by contract and/or the EEDI regulations are usually a
combination of smooth hull and propeller surfaces, no wind, no waves, no current and deep water of 15°C.
It is an impossible task to conduct a sea trial under no wind, no waves and no current conditions. Therefore,
“speed trial analysis procedures” for evaluation and correction of speed trial data are needed to compensate
or remove the effect of environmental conditions encountered during sea trials.

ISO 15019 and ITTC recommended procedures address the methods of analysing the results obtained from
speed trials. The ISO 15016 speed trial analysis guidelines are known to be a detailed method that takes
into account various effects, such as waves, wind, steering, drifting, water temperature, salt content, vessel
condition, and shallow water.

ISO 15019 reads “thereby enabling owners and others to have confidence in the validity of the final results”
with the desired target accuracy within 2 % in shaft power and 0.1 kn in speed.

Also, once the ship is built, the speed Vref to be entered in the EEDI equation has to be verified in a sea
trial. MARPOL Annex VI specifies two alternative standards for conducting and evaluating such sea trials:
ITTC Recommended Procedure 7.5-04-01-01 (2014) and ISO 15016:2015 (from 1 June 2015) and their latest
revisions.

The IMO’s 2014 Guidelines on survey and certification of the EEDI (MEPC.1/Circ.885/Rev.1, 2015) states as
follows:

4.3.5 Sea conditions should be measured in accordance with ITTC Recommended Procedure 7.5-04/01-
01.1 Speed and Power Trials Part 1;[2014 or the latest edition] or ISO 15016:[2015 or the latest
edition].

4.3.6 Ship speed should be measured in accordance with ITTC Recommended Procedure 7.5-04-01-01.1
Speed and Power Trials Part 1 or ISO 15016.

Both standards were recently amended to overcome previously identified drawbacks. The ITTC Recommended
Procedure contained black boxes leading to inaccurate results and the ISO standards were too complicated
leading to ambiguous results.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  13

4. Industry standards and commercial considerations

4.2 ISO 19030

The establishment of ISO 19030 in November 2016 was a breakthrough in the marine industry. Although
limited in scope, the aim of the standard is to prescribe practical methods for measuring changes in ship-specific
hull and propeller performance during the ship’s service period and to define a set of relevant performance
indicators for hull and propeller maintenance, repair and retrofit activities.

Some of the procedures described in the standard address both performance monitoring and digitalisation.
Operators and other stakeholders quickly understood that data could be used for “smarter” analytical-driven
decisions, significantly improving the efficiency of their fleet and at the same time reducing the whole shipping
industry’s environmental footprint to a greater level.

According to INTERTANKO’s online poll results, many in the industry are not well-aware of the technical
contents of ISO 19030 (roughly 30% of respondents said “I am aware of ISO 19030”). It would be worthwhile
exploring the technical background of ISO 19030 for a deeper understanding.

Appendix 1 describes the basic principles of the ISO 19030 standard, highlighting best practices and its
limitations. ISO 19030 uses the speed loss concept. It relies on the previously mentioned conclusions that relative
changes in total resistance reflect changes in frictional resistance when factors that affect total resistance are
similar. Appendix 2 provides more details about the background of the speed loss concept.

4.3 Charter Party and TMSA

4.3.1 Charter Party
The Charter Party standard clause on vessel performance (Ref.: BIMCO Model Clause 18, Performance of
Vessel – speed and consumption):

• Unless otherwise ordered by Charterers, the Vessel shall perform all voyages at the service speed stated
in the Questionnaire.

• Owners warrant that the Vessel is and shall remain capable of maintaining, throughout the Charter
Period, the speeds and bunker consumptions for propulsion described in the Questionnaire under
normal working conditions and in moderate weather (which for the purpose of this Clause shall exclude
any periods of winds exceeding Force 5 on the Beaufort Scale).

4.3.2 TMSA, A Best Practice Guidance (3rd edition, 2017, OCIMF)

The Oil Companies International Marine Forum (OCIMF) Tanker Management and Self Assessment (TMSA)
programme was introduced in 2004 as a tool to help companies assess, measure and improve their management
systems. The programme encourages companies to assess their safety management system (SMS) against key
performance indicators (KPIs) and provides a minimum expectation (level 1) plus three levels of increasing
best practice guidance. In the context of the TMSA, KPIs are measures against which a company can track its
effectiveness on meeting its aims and objectives.

TMSA Element 10 “Environmental and energy management” is relevant to our study. A summary is provided
as follows.

Main objective
Companies establish a proactive approach to environmental and energy management that includes the
identification of sources of marine and atmospheric emissions and implementation measures to avoid or
reduce potential impacts. That includes:

• effective fuel management,

• optimising energy efficiency,

14 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Main objective

• identifying and implementing energy saving opportunities,

• effective use of current and emerging technology for existing ships and new builds,

• internal and external benchmarking of environmental performance.

Element 10.2.1 specifies that energy management may include monitoring and reporting requirements for the
following:

• daily fuel consumption including ME, boilers, IGS and auxiliaries,

• vessel’s speed and distance travelled,

• vessel’s condition (laden or ballast),

• vessel’s trim,

• weather, sea state and wind direction.

Such data is recorded on a voyage-by-voyage basis, for individual vessels and on an overall fleet basis. Time
spent alongside and at anchor is included.

Element 10.2.2 specifies that:

Actions may include:

• establishing baseline criteria and targets to be achieved,

• operational measures to improve environmental performance such as engine room waste management,
garbage management, slop management, VOC management,

• regular performance reviews including the calculation of specific fuel consumption trends, monitoring
of hull condition and propeller fouling, the performance of ME, boilers, IGS and auxiliaries and the
generation of waste.

Element 10.2.3 specifies that:

Measures may include:

• optimisation of vessel trim,

• speed optimisation where practical,

• weather routing,

• optimising onboard power management such as the use of generators and boilers,

• propeller polishing / cleaning,

• hull cleaning,

• most efficient method of ballast water exchange / treatment.

Element 10.4.1 specifies that: Available technology is used to enhance energy efficiency:

This may include:

• emerging coating technologies,

• real time performance monitoring and comparative analysis of vessels,

• condition monitoring of environmentally critical equipment,

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  15

4. Industry standards and commercial considerations

• engine auto-tuning,

• ME de-rating,

• alternative energy efficient fuels,

• fitting of appendages to the hull to aid efficiency.

Element 10.4.3 specifies that performance is measured within the company and benchmarked across the
industry periodically.

4.4 Energy Efficiency Management Plan

See Appendix 10. It discusses how a shipping company views and implements the energy efficiency management
plan. The approach addresses recommendations contained in IMO’s SEEMP and ISO 50001 with a case study
of Aframax tankers.

4.5 Examples of ship energy efficiency measures and Energy Efficiency Technology (EET)
Reference is made to TMSA 10.2.3 (see P15).

Once the best practices on ship performance monitoring are established, the ship operator would want to
transition the knowledge towards the next phase of challenges facing the industry. IMO has charted the clear
course to decarbonisation of shipping with an initial ambition of GHG emissions reduction by at least 50% of
2008 levels by 2050 (IMO MEPC, 2018)1. It is essential for owners and operators to consider their positions
now in order to take informed decisions for the immediate future.

There are operational measures to achieve these challenges:

• speed optimisation,

• virtual time of arrival,

• weather routeing,

• cargo optimisation.

Appendix 3-1 discuss these measures in detail.

Apart from operational efficiency measures, there are energy efficiency technologies (EET) that can be used to
maximise the efficiency. Appendix 3-2 classifies each EET with respect to readiness of deployment, initial costs,
and appropriateness for retro-fitting.

4.6 Performance indices

Performance indices in the industry have a wide range of purposes, however, the primary purpose of any metric
is to establish a level playing field and to allow comparison within a peer group in a quantifiable manner. While
companies can have internal KPIs to ascertain performance of their vessels within their companies, independent
indices are designed to provide a more easily comparable and transparent view of vessel performance. This can
be used for various purposes:

• Regulation: port specific, national, regional or international regulations need indices in order to assess
compliance or drive a goal-based approach to achieve targets such as emissions reduction, safety or
environmental impact levels.

• Rewards: ports, Flag States or organisations can implement incentives in the form of discounts. The
Green Award and Environmental Ship Index (ESI) are good examples.

1
IMO, 2018 Initial IMO Strategy on Reduction of GHG Emissions from Ships. Resolution MEPC 304(72), 2018

16 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 4.6 Performance indices

• Commercial decision making: The Charterers want to consider a performance index in order to shortlist
the vessels they wish to hire. This may be compared to findings from the TMSA vetting process but with
more focus on a performance-based approach. Shippers may also want to choose more efficient vessels
to transport their cargo.

• Financing: Financiers have already begun to acknowledge the risk of financing projects that may become
stranded under more stringent regulatory regimes. Thus, well designed performance indices can provide
a way of evaluating risks associated with their financial investments.

• Newbuild design and second-hand market: when commissioning a newbuild or buying a second-hand
vessel, a performance index can give a better idea of what is to be expected from the vessel under
consideration.

Performance indices can be of two types:

• Design index: this index is based on the design parameters of the vessel, including the hull form, installed
power, design speed, design draught and theoretical operating condition. This can be calculated at the
design stage and remains with the vessel for its lifetime providing that no significant changes to the
vessel are undertaken. An example is the Energy Efficiency Design Index (EEDI). Note that EEDI is not a
measure of how the ship is being operated or to be operated.

• Operational index: this index is based on operational parameters taken from in-service measurements in
combination with design data. This aims to describe the actual performance of the vessel rather than its
expected performance. An example is the Energy Efficiency Operational Index (EEOI).

Table 1 presents a non-exhaustive list of current vessel performance indices.

Table 1: List of indices

Governing
Acronym Name Unit Description
body
Based on design data, ratio of CO2 emissions and
Energy
possible transport work based on DWT and design
EEDI Efficiency IMO gCO2/tnm
speed. Mandatory for vessels built from 01/01/2013
Design Index
onwards.
Energy
Based on operational data, ratio of CO2 emissions
Efficiency
EEOI IMO gCO2/tnm and possible transport work based on DWT and
Operational
actual speed over a voyage. Voluntary.
Index
Considers GHG emissions (SOx, NOx, CO2), PM,
Clean
CSI CSI n/a chemicals, waste and water. Scale of 0 – 150.
Shipping Index
Independent.
Same methodology as EEDI but applied to all vessels
GHG
Existing Vessel based on design data from vessel database. Vessels
Rating / Rightship gCO2/tnm
Design Index compared to each other in peer groups and allocated
EVDI
an A to G rating within group. Independent.
Environmental Considers GHG emissions (SOx, NOx, CO2) and
ESI WPSP n/a
Ship Index onshore power use. Independent.
Holistic index considering environmental, health
Shipping
and safety, HR management, navigational safety,
SPI Performance BIMCO n/a
operational, security, technical and port state control
Index
performance. Scale of 0 – 100. Independent.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  17

4. Industry standards and commercial considerations

As previous studies have shown, the application of performance indices has to be very carefully considered
(INTERTANKO/UCL 2016, Faber & ‘t Hoen 2017). Work is currently in progress at the IMO to strengthen the
EEDI regulations and introduce goal-based performance indices in the form of carbon intensity as candidate
short-term measures for the initial GHG reduction strategy. This signals the prospect that performance indices
will shape the near future of shipping and emphasises the importance of establishing salient, legitimate and
credible indices.

Methodologies for indices:

• EEDI: MEPC.245(66)
• EEOI: MEPC.1/Circ.684

• EVDI/GHG Rating: https://site.rightship.com/wp-content/uploads/2018/05/GHG-Rating-Methodology_

Vessel-Verification.pdf

• CSI: https://www.cleanshippingindex.com/wp-content/uploads/2019/04/2018-10-26-Methodology-
and-Reporting-Guidelines.pdf

• ESI: http://www.environmentalshipindex.org/Public/Home/ESIFormulas

• SPI: https://www.shipping-kpi.org/book/pages/SPI

4.7 Data intellectual property (Appendix 4)

Many shipping companies are now using external service providers for the purpose of ship performance
monitoring. This trend explains how the shipping industry has seen a huge increase in demand for more
accurate ship performance data analysis. Appendix 4 provides a model agreement to be signed between the
ship owner and an external service provider protecting data and data compilations as intellectual property
assets. It addresses key considerations in data owernship, data transfer, the scope of data usage, confidentiality
and control.

18 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 5.2 Results of the online poll

5. Trend analysis
5.1 Questionnaire

The INTERTANKO WG developed an online Questionnaire seeking Members’ input on the following:

• To what extent are Members applying vessel performance monitoring systems?

• To what extent are industry standards, ISO 19030, or equivalent, applied in practice?

• Which areas are the most challenging in day-to-day performance monitoring work?

Appendix 5 contains the Questionnaire and answers.

5.2 Results of the online poll

The online survey of tanker operators carried out across the INTERTANKO Membership from 15 April to 31 May
2019 asked 25 questions on respondents’ perspectives and attitudes towards performance monitoring. This
section provides an overview and summary of key analytical points of the survey.

In summary, Members may not know a great deal about how ISO 19030 is formulated and works, but the
majority have a sense that ISO 19030 or similar methods can provide a useful tool, and that they also do not
entirely rely on ISO 19030 or similar to ensure the correctness of performance monitoring. This brief does not
translate into widespread application of, say, a dedicated in-house monitoring team, a third-party provider or
a combination of both. How they actually and commonly measure and compare the performance is still an
ongoing issue.

1.  Hull and propeller performance monitoring is viewed as the most important aspect of vessel
performance management. Commercial performance and machinery performance are seen as an
important part to a lesser extent.

2.  Most Members possess awareness of ISO 19030, however, the majority are not necessarily conversant
with it. It is not widely adopted in the industry.

This means that the large majority of Members connect their performance monitoring with ISO 19030,
however, their understanding of ISO 19030 and its connection to performance monitoring is superficial.
The survey asked multiple questions about auto-log systems, manual reporting, data frequency, sensors
fitted and data quality. Looking at individual responses, we find that all gave accurate answers pertaining
to their monitoring management. This means, on the other hand, that performance monitoring is a
truly essential part of their tanker operation.

o It should, however, be noted that the focus here is on hull and propeller performance (as covered
by ISO 19030), and less on machinery performance (which is probably much better established
already).

o In the context of performance monitoring, the main application now is performance-based

cleaning (of hull and propellers).

3.  The survey asked a question (Q6) about how they do performance monitoring within their companies.
More than 30% of answers chose a hybrid system consisting of an in-house team and a third-party
service provider. The same number of responses chose manual performance monitoring with email
and excel reporting. Almost 20% of responses chose a third-party service provider alone.

The survey results show that all of them are implementing a performance monitoring system, manual,
automatic, a third-party service or a combination of these methods.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  19

5. Trend Analysis

4.  The survey asked three questions (Q8~Q10) about auto-log systems. More than 60% of Members
responded that they have auto-log systems fitted on their ships. Surprisingly, about 40% of Members
who responded do not have auto-log systems on their ships. The smaller their fleet is, the more likely
they are to fit auto-log systems. It is noteworthy that about 90% of Members fit torque meters on their
fleet, followed by RPM, SOG, wind, STW, heading guages and so on. For this reason, this document will
place prominence on torque meters. See Appendix 9 (How to measure shaft power).

Members do, however, place very high corporate importance on manual reporting of performance
related data (Q14). More than 90% of Members are using manual reporting.
o Every operator uses noon reports. Not all operators count this as part of their reporting (because
“of course” everybody does). Hence, “hybrid” and “auto-log” reporting may mean the same thing
in the answers.
o There seems to be a conceptual mixing between “data acquisition” and “performance assessment”
(of whatever data one has).

5.  The survey asked two questions (Q11 and Q12) about data gathering frequency and averaging. 70% of
Members are opting for high-frequency data gathering, e.g. between 1 second and 1 hour. The same
number of Members are averaging the so-collected data over time.

6.  A very high level of confidence in sensor data (Q3) was noted. About 50% of Members answer that for
more than 90% of time, their sensor data was reliable.

7.  Data of interest (Q17). Whereas torque meters are fitted on 90% of Members’ ships, data
of interest for record-keeping purposes is wind data (95% of Members). This may be due to
the fact that wind data is the most frequently observed and recorded data on board ships.
o If ships use auto-logging, it is likely that they also have torque meters. The difference between Q10
and Q17 seems to be due to the difference between auto-logging and manual data acquisition.

8.  The survey asked how they measure fuel oil consumption (Q19). About 70% of Members are using
either mass or volume flow meters. Members’ confidence in the accuracy of flow meters is neither high
nor low (Q20).

9.  The survey also looked at key messages that could assist in building the case for common methods
of performance monitoring. The most convincing statement can be found in Q24. About 70% of
Members state that the characters accept their ships’ performance data.

10.  Members’ narrative comments suggest that implementing a performance monitoring system is
considered a fundamental must-do action by shipping companies even though it is not made mandatory
regulators and commercial contracts. The large majority of Members say that they are willing to support
actions to further develop the best performance monitoring methods even if the trade-offs of high
costs, a lack of best practice guidance, or more diverse choices are presented to them. The next step for
the shipping companies is to make this issue urgent and increase the saliency of the subject.

o Performance monitoring will gain in importance (A1, A3, A4, A9, A10, A11, A13, A15, A21 and
A22), with a market demand for user-friendly reporting (management report by exception and in
standardised form) (A6, A14, A19, A20 and A28).

o The perceived large variation in performance monitoring (more accurately performance assessment
/ correction) methods undermines trust, with some vendors / approaches seen as black sheep (A2
and A5).

o In-water cleaning is a related issue that needs to be addressed (A12).

20 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 6.1 Basic definitions

6. Understanding sources of uncertainty

The path to zero uncertainty is infinitely long and very expensive. Be sure to only travel as far as you need.

This section will not include mathematics behind the uncertainty but this knowledge is critical in understanding
the impact of a particular uncertainty. For example, the cubic relationship between speed and power can
magnify the impact of uncertainty inherent in the ship’s speed. Nor does this section attempt to provide a
comprehensive list of all sources of uncertainty. Rather, it seeks to help the reader understand how to identify
and deal with sources of uncertainty through relevant examples.

6.1 Basic definitions

The quality of a measurement must be assessed with its associated uncertainty. Uncertainty defines a range
within which the true value lies. It is quantification of a doubt. It is thus different from the error which
represents the difference between the true and measured value. Evidently the lower the uncertainty, i.e.
narrowly spreading measured values, the higher the quality of measurement. Uncertainty consists of type A
components evaluated by statistical means (standard deviations of various degrees of freedom in a series of
observations) and type B components not derived from observations. Type A uncertainty is obtained from a
probability density function derived from an observed frequency distribution in a set of measurements, while a
type B uncertainty is obtained from an assumed probability density function called subjective probability based
on a priori knowledge such as manufacturer specifications or engineering judgement.

Combined uncertainty from various sources is governed by composition of all sources of uncertainties.
Expanded uncertainty is a combined uncertainty multiplied by a factor depending on a desired confidence
level. The higher the required confidence level (typically 95%) the higher is the factor to multiply the combined
uncertainty. This factor is also known as the coverage factor. A confidence level of 95% means that 95% of the
measured values will lie within twice the standard deviation of the measurement (for a Gaussian distribution).

Drift is a change in output over a period of time that is unrelated to input. This can be due to aging, temperature
effects, sensor contamination, etc.

Bias is a systematic discrepancy between an indicated, assumed or declared value of a quantity and the
quantity’s true value. It is the difference between the average of measurements with the reference/true value.
It constitutes KPI for the systematic error. Systematic error is a constant error and it characterises accuracy or
lack of it. It is desirable to have low bias and low standard deviation (accurate and precise respectively).

Precision is the proximity/variance among repeated measurements.

Error is due to random and systematic effects. Randomness on errors is the result of temporal and spatial
variations affecting the repeatability of the observations. Systematic error is due to a recognised effect and it
can be quantified and accounted for.

Accuracy: This means closeness of agreement between a measured value and the true value of a quantity. If a
measurement is accurate, the average of the measurement results is close to the “true” value (which may be
the nominal value of a certified standard material). If a measurement is not accurate, this can sometimes be
due to a systematic error. Often this can be overcome by calibrating and adjustment of instruments.

Precision: This describes the closeness of results of measurements of the same measured quantity under the
same conditions, i.e. the same thing is measured several times. It is often quantified as the standard deviation
of the values around the average. It reflects the fact that all measurements include a random error, which can

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  21

6. Understanding sources of uncertainty

be reduced, but not completely eliminated. Measurements can be accurate, but imprecise, or vice versa. The
ideal situation is precise and accurate.

6.2 Sources of error

Uncertainty analysis might be more time-intense than the actual measurements due to its complexity.
Nevertheless, it is necessary to include uncertainty range when reporting measurements. Sources of error in a
measurement are:

• scale effects,
• incorrect modelling (geometry, weight etc.),

• incorrect environmental modelling,

• instrumentation & measurements error,

• error in analysis and interpretation of tests results.

Precision errors are easier to identify but harder to generalise, since they usually depend on types of transducers
and details of the experimental facilities and test set-up.

For uncertainty in measurements of data, it is necessary to analyse each sensor depending on its manufacture
and type. The manufacturer of the sensor provides information about the expected accuracy of measurements
by the sensor and this information can be used as an indicator when evaluating the data.

We will discuss uncertainty associated with sensor measurement data in Section 7, Section 8 and Appendix 7.

Theoretical background of uncertainty analysis is provided in Appendix 6.

To examine uncertainty in the context of vessel performance monitoring, we need to understand sources of
uncertainty and how they might lead to uncertainty in our understanding of vessel performance. There are a
few different definitions of “performance”. For example, we may have a question of how using calm water
performance measurements might introduce uncertainty into predictions of real world fuel consumption. As
this examples shows, it is critical to follow the uncertainty from source to business decision.

6.3 Introduction to uncertainty in vessel performance

Our goal is to produce a series of useful indicators defining vessel performance and different indicators will be
used in different decision-making processes. Different indicators might have types of uncertainty that make
them more suited to one application than another.

Every time we take a measurement, use data or theoretical formulae, we introduce a level of approximation
into our analysis. This has the potential to make some indicators misleading and result in poor or erroneous
decision-making. However, good decision-making will only come from good performance knowledge.

Performance indicators are critical to good decision-making but failing to understand the uncertainty within
them can lead to their inappropriate use resulting in misguided and poor decisions.

The breadth of possible areas of uncertainty outlined here might lead to some dismay in terms of the possibility
of creating valuable performance indicators. This is not the case and there are many valuable performance
indicators that can be developed, but each has a level of uncertainty that needs to be understood if they are
to be applied wisely.

22 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 6.5 Baseline uncertainty

Uncertainty is the sum of all the possible ways in which an indicated value might differ from reality. In any
indication it is essential to understand the sources of uncertainty in order to understand their impact on the
conclusions that are being drawn from the indicator. Understanding uncertainty might lead to a conclusion
that certain sources of uncertainty need to be reduced but it might equally lead to the conclusion that the
decision, use of the indicator, does not depend on a particular source of uncertainty. Removing uncertainty can
be costly and needs to be evaluated against the benefit to the received indicator within its application.

To reinforce this last point, consider the simple example of comparing a current vessel’s performance with:

1. the same vessel’s performance three months earlier, to determine whether the hull is fouling,
2. a different vessel’s performance, to see which vessel is most efficient.

In the first case we rely on consistent measurements, a flow meter does not have to be accurate in its absolute
measurement, rather consistent over time in representing the fuel flow (with the same errors). In the second
case, a 5% difference in the accuracy of the speed log between the two vessels might translate into a significant
uncertainty in the comparison.

We must focus on removing the uncertainty required to make a performance indicator valid for its intended
application. We must not use an indicator in an application where the uncertainty makes it invalid.

6.4 The process of understanding uncertainty in vessel performance

In order to ensure that we tackle performance indicator uncertainty appropriately, i.e. with a view to extracting
best values at least cost, we should think through the evaluation process:

1. An indicator has a purpose; ensure that we fully understand the context in which the performance
indicator will be used and what decisions are made. Understanding uncertainty depends on this.

2. An indicator has a reference against which to compare; ensure that we fully understand the uncertainty
in the reference or base line; this is part of the uncertainty evaluation.

3. An indicator will be based on certain assumptions from its application; we must understand how
these assumptions place a limit on the level of uncertainty. We want to avoid costly and unproductive
improvements.

4. An indicator will have input values taken from sensor measurement, each with its own uncertainty.

5. In case an indicator is dependent on human interaction, manual data entry has its associated errors.

6. Interpret the result. Note how easy it is to create a performance indicator then misuse it.

6.5 Baseline uncertainty

This discussion raises an important concept of baselining. We must determine a “reference” against which
we compare our performance indicator. As an indicator is simply a number, it has no meaning unless it has a
reference for us to be able to determine that something is better or worse. Frequently these are presented as
“baselines” on graphs or nominal performance (100%) against which trends are presented. A reference might
be something generated during a sea trial on vessel delivery, or historic performance, industry standard and a
sister ship operation.

It should also be recognised that baselines have uncertainty too. This is a challenge in our industry.

As an example, consider that someone wishes to determine the performance benefits of a new propeller
design through a trial on a single vessel prior to installation across a fleet. It sounds simple:

• determine the performance (the baseline),

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  23

6. Understanding sources of uncertainty

• replace the propeller,

• determine the new performance (the indicator), compare the two.

However, considering the cost of dry docking to replace the propeller, it would be wise to clean the hull at the
same time. Now the process has changed:

• determine the performance,

• correct that performance to allow for a hull clean (the baseline),

• replace the propeller,

• determine the new performance (the indicator).

The uncertainty analysis must be carried out prior to the activity. There may be cases where the baseline
uncertainty is so significant that the whole exercise becomes pointless. This can be determined in advance
(there is no excuse for carrying out such an expensive exercise only to discover the baseline is inadequate).

The route to determining performance indicators from a set of input parameters will introduce uncertainty
into the indicator. We also need to examine sensor measurement uncertainty in this context (see section 6.8).
A baseline example for hull and propeller performance is provided in section 8.

6.6 Proxy errors

This section focuses on measurement uncertainty but it is essential to distinguish uncertainty in measurement
from uncertainty in value. Let us try to understand exactly what is being measured. In many cases, a measurement
is a proxy for the value required. This introduces an additional level of uncertainty and becomes the reason why
certain measurements are rejected in the decision-making process.

It has long been understood that to take a measurement, the subject being measured needs to be disrupted.
However, this is not possible in shipping. The presence of the ship itself can have such a massive impact on a
measurement as to make it of little to no value.

Some examples are outlined below to demonstrate the breadth of the challenge:

1. The measurement of the speed through water at the hull as a proxy for the speed through water at a
distance from the ship.
The presence of the hull significantly impacts the value required even if the speed through water at the
hull is being measured accurately. See Figure 3.

Figure 3: measurement of ship speed

24 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 6.7 Uncertainty in parameters used in analysis

2. The measurement of the wind at the vessel as a proxy for the real value of the apparent wind.
The presence of the vessel has significant influence on the flow of the wind and the output from
an anemometer might be accurately representing the wind as it sees. However, the anemometer
measurements might be significantly different from the real value of the apparent wind (which would
be desired in any performance calculation).

Figure 4: Measurement of wind

3. The measurement of the volume of flowing fluid.

A volume flow meter can measure the volume of fluid passing through a pipe. However, fluid is a
combination of gas and liquid. The output of a volume flow meter might accurately represent the flow
of fluid whilst being a poor proxy for the flow of fuel because poor positioning could lead to significant
amounts of air still being kept in the flow.

4. The use of mass of fuel consumed as a proxy for the energy consumed.
In performance monitoring, we are sometimes interested in the energy that is being used to propel
the ship. Mass of fuel consumed by the main engine is taken as a proxy for this where we use the
calorific value of the fuel to determine energy extracted. Errors in the calorific value because of poor
measurement or fuel mixing will introduce variations in the perceived energy used. Laboratory testing
of the fuel can reduce these uncertainties.

6.7 Uncertainty in parameters used in analysis

It is easy to have a false confidence in parameters that are widely used or are buried in the depths of a calculation,
but these too can be sources of uncertainty. Many performance indicators are developed from calculations that
have some dependence on parameters with uncertainty. Block coefficient is a useful approximation of a hull
form used in many calculations. Formulae, apparent constants and assumptions all need to be challenged as
each might have an impact on the uncertainty in any performance indicator.

Two interesting examples are presented overleaf.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  25

6. Understanding sources of uncertainty

Shallow water corrections

A recent study concluded that longstanding assumptions about corrections for shallow water are in fact
significantly in error, see the graph on Figure 5 below (speed correction at depth draught ratio).

Figure 5: Speed correction at depth draught ratio)

The significant point is that using theoretical knowledge (any) in order to create a performance indicator can
be a source of uncertainty.

Shaft material for torque meters

Torque meters determine the “torque” using shaft dimensional changes, but these are dependent on the shaft
material and its consistency at the point of measurement. These are entered as part of the set up but again
they are subject to error and represent another value that has uncertainty. The challenge with these variables
is that they frequently disappear in the depths of a calculation and are not visible at the point of analysis, but
they can be significant in some uses of performance indicators.

6.8 Uncertainty in sensor measurements

Typically, many measured parameters are fed into our performance calculations. Any uncertainty in these will
result in uncertainty in the final indicator. Each needs to follow through the performance indicator creation
process to determine its impact on the result, although many can be quickly shown to be insignificant in
comparison to others and therefore not critical in the analysis.

We discussed uncertainty inherent in a few sensor measurements. In Appendix 7, various types of measurement
uncertainty are outlined in greater detail with simple examples. The aim is not to create a comprehensive list of
sources of uncertainty, rather enable the reader to know how to explore areas of uncertainty.

6.9 A case study in applying different methods recommended in ISO 19030

There are various types of sensors, instruments and control systems that are processing signals from many
sensors in all vessels. They are providing an operational environment for the ship’s crew to perform their duties
and are defined as follows.

Signals from sensors are either analogue or digital. Normally, a sensor is combined with a single or a multi-step
processing unit, with a visual and/or digital output, that transforms the signal into a meaningful format to an
instrument (i.e. torque meter, fuel oil meter) and/or a Programming Logic Circuit (PLC) (i.e. cargo control unit)
that a person can read. It is common signals to be used by PLCs primarily for control/safety purposes.

26 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 6.9 A case study in applying different methods recommended in ISO 19030

Instrument is a device that transforms signals into measurements for a predefined purpose (i.e. GPS, weather
station, echo sounder, FOC etc). The instrument may have an electronic output (i.e. RS232, NMEA, Ethernet
etc) providing a digital output in a sampling rate and format as programmed by the instrument vendor.

PLCs/SCADA systems typically integrate signals from many sensors (thousands in a modern ship) in order
to perform predefined tasks in a fully controlled manner. The sampling rate per sensor varies from a few
nanoseconds up to event triggered sensors (i.e. pulse, level or proximity sensors). Those systems are typically
equipped with a digital output (i.e. USB, RS232, RS485, NMEA, Modbus, Canbus Ethernet). This digital output
can be programmed in order to provide the predefined measurements from the PLC programmer output in
a defined sampling rate and relative to the processing ability of this subsystem due to processor/memory/
bandwidth restrictions.

Sampling rate of an analogue sensor (i.e. 4-20mA, PT-100, 0-10mV) means a signal relative to physical
phenomenon measurement per a given period of time (i.e. 1 ms).

Sampling rate of an analogue signal processing (ASP) unit means a new signal for a period of time that
contains x number of analogue signals based on a given formula (i.e. median, average, start value, end value,
multi averages etc) and a period of time i.e. 500msec.

Analogue to digital converter is a processing unit that transforms signals into a digital signal of series of bits
(voltage or not - “1” or “0”) or similar in a formatted manner.

Sampling rate of a digital processing device means a new signal for a period of time that contains y number of
processed signals of an ASP unit based on a given formula and a period of time i.e. 15 sec.

Converter is an additional processing unit that transforms a signal of a given protocol to another protocol (i.e.
4-20 mA to Serial) based on a formula and predefined by the vendor conditions.

Data protocol is a given format for a device to communicate with other devices.

Measurement is the transformation of a signal into a value using the appropriate linear or non-linear equation,
given reference points and a metric system (i.e. Temperature in Celsius or Fahrenheit or Kelvin) and presented
using a predefined user interface (i.e. LCD Display, Monitor) with a predefined sensitivity, accuracy and duration.

A gateway is a device that reads digital signals from predefined protocols and generates data packets in a
format that software running on a computer can read.

Data processing software is an application that reads data packets and transforms them into data stored in a
computer.

Automated data collection system is the ability to gather measurements in defined format stored as tables in
a computer for further use.

ISO 19030 defines a set of rules and conditions that measurements received generate a comparable set of
various measurements used to define the actual performance of the hull, minimising the uncertainty generated
from the various operational envelopes within which the ship operates.

Appendix 8 discusses the application of ISO 19030. Appendices 6 and 7 give further details on the theory
behind uncertainty analysis and its application to ship performance and sensors.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  27

7. Data source

7. Data source
7.1 Auto-logging sensors onboard the ship
The data from several sensors installed on board the vessel is used to feed the performance system, either by
readings done by the crew or by an automatic feed to the system.

Figure 6: Performance measures and influencing factors, Søren Vinther Hansen (2015), PhD thesis,
Performance Monitoring of Ships, pp 51-72, Technical University of Denmark

The most important sensors are listed and described below.

Speed log
The speed log is a crucial source for data in the performance calculation. The accuracy of the sensor is often
questioned and in many cases the speed log readings are omitted as the speed indicator for performance
analyses.

A well-functioning speed log is the best source for speed since it measures speed through water and therefore
is not affected by current set and drift. With careful maintenance, proper settings and calibration, it can be the
most valuable source for speed through water.

With regards to accuracy, various manufactures of speed logs give varying information. Once calibrated, the
speed log measurements can have an accuracy of 0.1%, but since it is sensitive to a number of factors, a
general approach is that average speed logs have an accuracy of 0.5 knots.

A number of settings and environmental factors influencing the speed log measurements are:

Water clarity The measurement of the speed through water depends on acoustic reflection from solid
particles in the water such as microorganisms or suspended dirt. In extremely clear water the
quantity of scatters may be insufficient for adequate signal return.

Aeration Aerated water under the transducer may reflect sound energy which could erroneously be
interpreted as sea bottom returns. Sailing in heavy weather can be the source of this effect
and so could create non-laminar flow around the transducer. By placing the transducer near
the bow, the effect of non-laminar flow is reduced considerably.

Ships trim Changes in the trim (affects fore/aft speed) and list (affects transverse speed) of the ship will
and list affect the measured speed. (Example: 5° trim change gives 0.4% speed change).

28 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 RPM and torque meter

Current profile Speed through water is measured relative to a water layer beneath the ship (> 3m). Sailing in
strong tides and current, the direction and magnitude of the surface current can be different
from the measured layer and this can give an error in the measured speed.

Eddy currents Sailing in eddies in boundaries of ocean current where the flow can be opposite or normal to
the direction of the primary current and will affect the speed measurement.

Sea state Following seas result in a rhythmic change in the vessels speed. This produces a fluctuation in
the measured speed. The speed measured could be inaccurate due to the lag introduced by
the speed filtering in the log.

Speed over Speed logs of the Doppler type can measure speed over ground as well as speed through water
ground It is important to ensure that when setting the sensor up it is measuring the speed through
water.

GPS receiver
The GPS receiver delivers information about the vessel’s position, the speed and the course over ground. The
sensor makes use of GPS satellite signals to determine the vessel’s longitudinal and transverse ground speeds
and the motion vector represents the magnitude and direction of its motion over ground.

In cases where the ship is not subject to any set and drift caused by current, the speed is equal to the speed
through water.

In cases where the GPS speed is used in performance analyses, one will often see a higher scatter in the results,
since the variations in current during a voyage will influence the performance result.

Wind anemometer
The wind anemometer can be a helicoids propeller type with a vane for direction measurement or it can be of
the transducer type with no moving parts. It can be mounted in the mast head on top of the navigating bridge
or it can be mounted in the forward mast. The general rule for the positioning of the wind anemometer is
that it should be placed as high and far ahead as possible in the vessel, to be free of distortion of the airflow
to the sensor.

Measuring accuracy is in the order of +/- 0.3 m/s or 1% of the wind speed and +/- 3° of the wind direction.
The accuracy is given to the range in which the anemometer is calibrated.

Vessel’s draft
Draught sensors used onboard are normally of the pressure transmitting type suitable to measure with adequate
accuracy the hydrostatic draught forward, aft and mid ship starboard and port, while the vessel is static. While
vessel is en route these type of sensors can introduce an error. It is suggested that dynamic draft measurements
systems providing more accurate measurements on draft, trim and list are used.

RPM and torque meter

Two types of meters are in common use. One measuring RPM and torque by light/laser and the other measuring
torque by strings and RPM by light/laser. Both types have shaft rings mounted and they are placed as close
to the main engine as possible. The measuring accuracy is in the order of 0.5 % and the update period
of the measurements can be varied. The RPM torque meter is calibrated at installation and depending on
manufacture and type, various calibration intervals and methods are suggested.

The torque and RPM measurements are used to find the shaft power.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  29

7. Data source

Thrust meter
The thrust meter measures the RPM, torque and thrust on the shaft. The measuring accuracy is in the range of
0.1 % and the update period of measurements can be varied.

The thrust and the torque measurements can in combination be used to identify the efficiency of the propeller.
Appendix 9 further discusses how to measure shaft power.

Fuel flow meter

The fuel flow meter measures, when placed in line with the liquid flow, the linear, nonlinear, volumetric and
mass flow rate of the liquid so that the quantity of the fluid flowing can be determined. The number of flow
meters depends on the installation in the vessel and to which extent the fuel consumption will be separated
between the different consumers.

Rudder indicator
The rudder indicator measures the rudder angle continuously and the measuring accuracy is in the range of
+/- 0.5° at angles near midships and +/- 1.5° at hard over rudder.

Several factors can influence the rudder angle measurements:

• An index error in the rudder angle measurement device.

• A permanent rudder angle to counteract the turning effect of the right-hand propeller.

• A permanent yawing effect induced by the ship’s hull during sailing.

The echo sounder

The sensor measures the water depth. The frequency ranges for the sensor are in the interval from 28 to 210
kHz and the measuring accuracy is in the order of 2.5% of the measured depth. The sensor is to be used in
confined waters for navigational purposes. The sensor frequency is normally set for 50 kHz and the detection
level for this frequency is around 90 – 150 m depending on sea water salinity and temperature. Anything
above this level will therefore not be detected. Environmental factors which can influence the measurements
are shown below:

Sea state Violent pitching in bad weather.

Sea water temperature Rising cold water in several sea areas. Hot water discharges from power
plants.

Noise From bow thrusters, main engine vibrations and propeller running reverse.

Air temperature & air pressure

Air temperature is measured by thermometer and air pressure is measured by barometer, both placed outside
on the aft part of the navigation bridge. Temperature changes during the day and night, during seasonal
change and during passage of areas with different climates. Air pressure changes with temperature and with
various weather conditions.

Gyro compass
The gyro compass measures the ship’s heading with reference to true North. The heading is used to calculate
true wind speed and directions based on relative measurements from the wind anemometer in the foremast.

The compass gets input from the speed log (for correction of speed error) and the GPS (for correction of
latitude error).

30 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 7.3 Quantification of uncertainties in manual reporting and auto-logging

Sea water temperature

The sea water temperature is usually measured at the sea chest suction. The water temperature at this level is
more or less following the air temperature, except on occasions at high latitudes in the winter period where
cold winds have not affected the sea.

Wave radar
Traditionally the wave estimation is done visually by the crew, by estimates from the wind speed or from
hindcast weather information. Since all of these have their different uncertainties, a wave radar can be used
to measure the actual sea state at any given time. The installation of a wave radar requires typically a X-band
radar and a processing unit linked to speed, wind and heading sensors.

Motion sensors
The bow and aft motions can be measured by accelerometers. They measure the vertical acceleration / vertical
velocity and the heave motion is found by integration. They can be used to estimate sea state conditions and
in combination with draught sensors to monitor the dynamic trim of the vessel.

7.2 Performance reporting

The use of telemetry is evolving, vessels are becoming smarter and the possibilities look endless. Nevertheless,
the vessels that use manual reporting, namely noon reports, for any kind of performance monitoring (hull and
propeller, boiler, auxiliaries, etc.) are the vast majority of the world fleet. INTERTANKO’s online survey analysis
indicates that this figure approaches 90%.

This figure is constantly dropping, mainly due to sophisticated new builds that come into market and retrofitting
programmes, but the rate is such that it ensures that manual reporting data will be used for many years to
come.

The typical manual reporting system from a vessel to shore can serve for regulatory compliance (SOLAS, EU
MRV & IMO DCS), and also can be used for evaluating commercial and technical performance, post fixture
financial use, maintenance of inventories and enhancing ISM systems.

Data that comes from manual reporting can even be used as proxies or to supplement information collected
from automatic data acquisition systems.

The majority of the industry use noon reports which comprise of information (speed, distance, consumption,
etc.) collected over the last 24 hours (+/- 1 hour due to crossing of time zones) while the vessel was at sea.

A different approach to manual reporting is event-based reporting, where the vessel provides information for
each event (e.g. at sea, anchorage, drifting, canal crossing, discharging, etc.). This is the preferred method
since it allows for a comprehensive allocation of time and consumptions per mode and ensures a continuity
of data. The latter is also almost a prerequisite for producing accurate and automated EU MRV and IMO DCS
emission reports with the use of the same data, manually reported, without having the vessel producing
additional dedicated emission reports per voyage, which is actually a duplication and requires significant effort.

For any of the above-mentioned methods, the diversity of the recipients add to the uncertainty of the data
reported from vessels, since the absence of common definitions amongst stakeholders is often noticed.

7.3 Quantification of uncertainties in manual reporting and auto-logging

Regardless of the method used to evaluate hull and propeller performance of a vessel, operators should
consider sampling frequency and uncertainty. As such, the use of data collected through noon reports (low
frequency) will lead to outputs of higher uncertainty. The uncertainty in the output of an evaluation of a

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  31

7. Data source

vessel’s performance, with the use of ISO 19030 methods, using datasets from 180 days with the use of noon
reports approaches that of datasets of 90 days of auto-logging, (ISO 19030 Part I, Annex A).

High precision sensors will affect equally both approaches in a positive manner while introducing human
error, which if not accounted for in the determination of uncertainties above, will further increase significantly
uncertainty to any attempt of quantification of performance with noon data.

Human error is introduced mainly through the determination of the environmental conditions that the
vessel encounters within the report duration (usually 24 hours) and collection of information in the absence
of appropriate measuring equipment, which results in subjective estimations. The absence of appropriate
guidance to crewmembers onboard involved with collection and reporting is also noticed and has an adverse
effect in the quality of reporting. If the influence of human error is minimised, then uncertainties of 5% can
be achieved, Aldous (2015), and noon reports can be used to derive meaningful results, Schmode et al (2018)
and Montazeri et al (2018). The latter requires the existence of high-precision sensors.

A different approach to reduce uncertainty in a vessel’s performance analysis is to reduce the reporting period
to cover steady state periods of limited duration, provided that appropriate care has been paid to the collection
of the required data. This can provide a good insight on a vessel’s performance but this is a time-consuming
task for the crew, since the requirement for reporting the rest of the information required from shore is not
waived, and cannot be repeated on a daily basis.

7.4 Uncertainties in manual reporting

An attempt to provide some insight on the sources of uncertainty in parameters used for evaluating hull and
propeller performance of vessels through noon reports follows.

Wind
Wind speed and direction is a significant source of uncertainty, since relative position of wind and vessel can
change constantly within the day, while wind speed is also fluctuating. In addition, there is a tendency from
crewmembers to overestimate its impact, especially in low-wind conditions. Another source of uncertainty is
partially linked to the absence of a common reference height of reporting / estimating wind force. The World
Meteorological Organization2 defines that the reference height for wind speed associated with determination
of the Beaufort scale in open sea is 10m from sea level while the anemometer of large vessels, especially while
these are sailing in ballast could be more than 40m, resulting in deviations of more than 1 Beaufort scale.

The following graphs depict a comparison between wind data collected from vessels’ anemometers (corrected
to reference height of 10m as per ISO 15016:2015, Annex C), with data provided by a weather provider and
data reported from vessels while they were sailing for a period of over a year.

2
Chapter 5.9.3, “Anemometers at sea”, CIMO Guide 2014 (The WMO Guide to Meteorological Instruments and Methods
of observation). On fixed platforms and ships, it is of the utmost importance that wind sensors be exposed sufficiently
high above the platform and its superstructure to avoid the often extensive influence of the platform on the local
wind structure. In general, it is never safe to assume that a wind sensor is unaffected by the platform structure, even
if it is exposed at least 10m above the height of the tallest obstruction on the platform, unless the platform is relatively
small. WMO (1981) concludes that, at sea, good exposure should have higher priority in obtaining accurate and useful
measurements than standardisation of the measurements at 10m (WMO, 1989).

32 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 7.4 Uncertainties in manual reporting

Figure 7-1: Comparison of wind date collected from vessels and data from a weather provider, Vessel A)

Figure 7-2: Comparison of wind date collected from vessels and data from a weather provider, Vessel B

The agreement between weather provider and vessels’ anemometer recordings is evidenced. The average
difference experienced was in the order of 1% with the highest difference being 4% for vessel A at Beaufort
Scale 4. On the other hand, differences between noon and anemometer was deviating on average by 5% with
the highest deviation being in the order of 9% for both vessels at lower wind speeds.

The above depicts that if auto-logging is not available, a reliable weather provider can be used as a proxy and
reduce the uncertainty in evaluating hull and propeller performance of a vessel. On the other hand, for vessels
which utilise auto-logging, it is also recommended that they use weather providers as a backup method in case
of anemometer failures or even identify issues that might be affecting their accuracy.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  33

7. Data source

Sea condition (swell and wave)

For determination of sea state, the absence of appropriate equipment onboard results in relying on subjective
observations of the officers on watch. Uncertainty is further increased due to the fact that more than one
personnel are involved in the process, since watches change periodically and due to the absence of daylight
for a significant part of the report duration.

Existing technology in that area is rather too expensive (wave radars) or immature (determination of wave
height with the use of X-Band radars).

Sea currents
Currents’ speed and direction fluctuate within the reporting period, having a significant impact on a vessel’s
performance. A quick fix for this could be adding the requirement that the vessel reports the distance sailed
through the water over the requested period and as such, the overall impact of the current is indirectly
estimated. The error is reduced.

Use of current data from weather providers to derive a vessel’s speed through the water just by adding vectors
of speed over ground and current often results in increasing the uncertainty, in the contrary with the impact of
weather providers to the rest of the parameters listed above.

Depth of sea and sea water temperature

Two underestimated parameters with regards to their impact to hull and propeller performance are depth and
sea water temperature. Their combined adverse impact can be up to 5% in power demand, depending on the
vessel’s size and trade.

Sailing in shallow waters will result in speed loss as a result of increased resistance and reduced propulsion
efficiency. The impact of shallow water will be greater at low ship speeds.

With regard to water temperature and salt content, they affect the density of the sea water and so the ship’s
resistance. The lower the water temperature, the higher the ship’s resistance becomes.

These parameters should be part of noon reports or otherwise collected from a third party.

Power measurements
In order to separate hull and main engine performance, the installation of a torque meter is a prerequisite. If
not, any deviation in the vessel’s fuel performance will not be able to be allocated to its source, especially for
vessels fitted with electronically-controlled main engines.

For the power measurements, the same principles with recording of wind and current apply. Dates of their
calibration should also be made available.

Fuel measurements
Fuel measurements can either be based on mass or volume flowmeters and tank readings. For fuel mass
measurements the requirement is straightforward since it involves recordings of flowmeter values at the
beginning and at end of each reporting period. When volume flowmeters are used, collected values need
to be converted to mass with the use of density and temperature of fuel at the flowmeters, as such, the fuel
density needs to be available, which is a potential source of error.

Accuracy of determination of fuel consumption is also highly dependent on the number and location within
the fuel piping system of the flowmeters available. Regardless of whether the method of collection is manual
or automated, we need to ensure that recorded values reflect actual consumptions of the intended equipment
and do not include leakages, returns and recirculation.

34 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 7.5 Manual reporting systems

For example, for a significant number of vessels, a fuel measuring system for the main engine and auxiliary
engines are common, making it difficult to determine the allocation of consumption of each equipment in case
of malfunction of any of the system flowmeters. This is something to be taken into account and addressed
through relevant checks.

Regardless of the setup of the fuel measuring and reporting system, it is strongly recommended that flowmeter-
derived figures are validated through sounding measurements of fuel tanks while the vessel is stationery in
frequent intervals. Fuel tanks should be calibrated and measuring tapes and bobs used should be of the
required accuracy.

Draft measurements
It is also common that reported draft is not updated to reflect the actual sailing condition of the vessel and that
is usually observed in long voyages. Depending on the vessel’s type, size and sailing condition trim can have a
significant impact on the resulting consumption and as such, any deviation should be reported. The distinction
between dynamic and static drafts should be made.

7.5 Manual reporting systems

Operators that wish to use noon reports for evaluating hull and propeller performance of their vessels should
ensure that the reporting system used will minimise uncertainties associated with the parameters addressed
above. In this respect, the following should be considered:

• The reporting system used shall be event based and allow for identification of all performance
influencing factors such as navigating in ice, slowing downs, etc. and their corresponding duration and
consumption.

• The reporting system used should be able to provide for declaration of consumption of each equipment
at each event.

• Involved crewmembers should receive training on the use of the system used for manual reporting,
which should include voyage-specific examples and detailed instructions on the requirements of each
parameter to be reported.

• Strict validations for key parameters should be introduced to the reporting system based on logical
functions and ship-specific particulars.

• In order to minimise the workload onboard, information required to be collected from vessels should be
scrutinised and reduced to the minimum practical extent to produce each report, for example; use of
distance and time to produce speed.

• The use of drop-down lists should be maximised in order to minimise clerical errors and typos.

• Supplement/enrich submitted information through technology (for example; use of GPS to derive
vessel’s speed, distance and position, and AIS).

• Use of weather providers in order to collect hindcast environmental data to supplement the information
collected from the vessel.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  35

8. Data analysis

8. Data analysis
Once data is collected, filtered and validated, it is analysed in a performance system. The data is normalised
towards a baseline and analysed with reference to a model of the vessel. The data is normalised to a standard
condition of the vessel i.e. corrected for performance-influencing factors like weather, environmental factors
and operational conditions. The model can be constructed from theoretical and physical data e.g. from vessel
design information and trial data, from collected data and machine learning techniques or from a combination
of both.

The baseline can be constructed for different areas of performance management e.g. hull/propeller, main
engine, auxiliaries or boiler performance. See section 6.5 for baseline uncertainty.

A baseline example for hull and propeller performance is shown in the following Figure 7. The baseline model
refers to speed/power relations for different draughts. In the following example, two draughts are shown.
One data point is measured on board and normalised towards model conditions. In this case, it is done with
a weather correction.

Figure 8: Principle of baselines and normalisation of values with speed/power relationship

The normalised data is then compared to the model value and the added power 3 is, in this case, an expression
for the deterioration of performance. For hull and propeller performance, the added power can be converted
to a reference value or KPI, where the added resistance due to fouling can be meaningful as KPI. See Figure 8.

3
The added power is defined as the additional power (delivered by the main engine) the vessel has to use to sail at a certain
speed when the hull/propeller is fouled i.e. the model value (or the baseline value) + the added power. So, the added
power is an expression for the performance degradation of the vessel which means that added power equals increased
fuel oil consumption to maintain the speed.

36 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 8. Data analysis

Figure 9: Normalised value compared to baseline for speed/power relationship

There are many advantages to using normalisation and reference values.

On the vessel level, baseline models enable the user to define an accurate reference performance for a specific
vessel under actual conditions, e.g. for actual weather and operational conditions. The performance can be
broken down to components like excess consumption by onboard consumers, route selection by a simulated
speed profile and vessel utilisation. See Figure 10.

Figure 10: Breakdown of fuel oil consumption into components related to

various performance influencing factors

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  37

8. Data analysis

By normalising vessel performance to calm sea conditions and a standard speed, one can evaluate and compare
the technical performance of the ship and the fleet. Grouping of vessels or equipment by their characteristics
(such as type, size, hull form/coating, energy-saving devices) adds context and insight to the analysis. The
technical efficiency insight can be applied for chartering decisions and assist in fleet maintenance prioritisation,
new-building and technology impact evaluation. On an equipment level, an equivalent approach enables
accurate evaluation of equipment deterioration or investment payback.

The reference curves can apply not only to multiple draughts but also to multiple trim and then reference curves
become reference surfaces (three-dimensional curves). Since trim is an important performance parameter for
many vessels, the accuracy of the performance analysis is considered to be better when including this.

The performance KPI can be calculated frequently. The frequency depends on the frequency of the data for
the performance analysis which can vary from a data point per day (noon data) down to frequencies around
15 seconds (according to the ISO 19030).

The performance KPI can then be plotted according to the frequency of the analysis. A sample KPI plot is
shown on Figure 11.

(Figure 11: The KPI development over time, in this case Added Resistance
increasing due to fouling of hull and propeller)

The plot of the performance KPI produces a scattered image. The scatter normally depends on the quality
of the data and the quality of the model work. A trend line is then used to express the increase in added
resistance over time and the trend line result is then an expression of the performance of the vessel. The trend
line can be constructed based on different principles. In general, the principle of linear regression is used.

38 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 8. Data analysis

Figure 12: Linear regression principle in scatter plots

In Figure 11, the red dots symbolise a set of independent variables x which relate to the dependant variables
y with an underlying relationship, shown by the blue line. The blue line can be derived by different fitting
techniques, where the least squares method is a common one to use. To establish a more robust trend and
with less sensitivity to outliers, different estimators can be applied to the linear trend.

The scatter in the analysis affects the result. Too much scatter indicates that there are issues with the data
quality. Outliers can periodically appear. These are values that are at an abnormal distance from other values
in a random sample from a population. The exact definition can vary depending on the method used in the
analysis. See Figure 12.

The outliers can affect the trending used in the performance analysis and might be causing errors or lead to
erroneous decisions.
Figure 13: Example outliers (blue circles) in scatter plots

Figure source: Vessel Performance Solutions

There are different ways to detect outliers. As an example, ISO 19030 mentions the Chauvenets criterion by
which the mean and standard deviation is used over a time bloc to establish an outlier. ISO 19030 Part 2 Annex
1 (outlier detection) provides a detailed overview of this method.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  39

9. Data driven decision

9. Data driven decision

9.1 New building data and sea trial
With emerging digital technologies, every company hires data scientists, installs data management software,
and trains their employees. Shipping companies are no exception. However, this trend would not in itself lead
to innovative designs and cost-effective ship operation. We need to understand what exactly data-driven
decisions in shipping are. This concept is intricately intertwined with the need to understand limitations and
applicability of various data generated from the ship. Without a degree of understanding of how such data is
generated, analysed and used, we would not achieve our objectives: environmentally-friendly and safe design
and operation. This exercise should start from the building stage and throughout the ship’s lifetime.

Appendix 10 presents an example case of a shipping company with this concept in mind. Appendix 11 discusses
information to be provided to the owner at the time of new building contract, construction and sea trial.

9.2 Fuel management

Fuel oil is the energy which drives our shipping commercial venture. Traditionally, from the moment fuel
is bunkered on board until the time it is consumed, management of quality, quantity, processing and
consumption are important factors in the viability of commercial venture. Today, fuel emissions have come
under environment regulatory purview with financial deterrents, requiring very close monitoring of the fuel oil
management onboard. The main factors in fuel oil management are enumerated as:

• quality,

• quantity,

• processing.

A flow meter, if installed, plays an important role in both monitoring the mass of fuel oil newly received and
the mass of fuel consumed every day. For the purpose of the ship performance monitoring, it is important to
regularly check the accuracy of the flow meter.

Appendix 12 discusses fuel management and monitoring in greater detail. Appendix 13 provides a background
of how the Specific Fuel Oil Consumption (SFOC) is calculated.

9.3 Hull cleaning and propeller polishing

9.3.1 Hull cleaning
Practical experience shows that for oil tankers, the potential for fuel savings through efficient hull and propeller
management lies at the same level with the cumulative effect of the rest of the consumers on board. It makes
sense that all efforts for enhancing the energy efficiency of a company start from hull and propeller.

With reference to the hull, efficient management translates into optimising hull strategy at drydock and later
on if so required to proceed with corrective actions such as underwater hull maintenance to minimise the fuel
penalty that the vessel will suffer because of biofouling between consecutive drydocks. A drydock strategy
includes the selection of appropriate antifouling for the type of vessel, its expected trade profile, coating
application, surface preparation (extend and quality) and ensures that the ships are operating in compliance
with both international and local regulations.

It is recommended that every operator deploys the use of an effective hull and propeller performance
monitoring system to determine whether the expected performance targets are met. They should also set up
documented drydock procedures with quantified quality control metrics. INTERTANKO’s “Guide to Modern
Antifouling Systems and Biofouling Management (2020)” publication extensively covers these topics and the
part on antifouling selection is worth examining.

40 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 9.3.2 Propeller polishing

9.3.2 Propeller polishing

Depending on the relation and operation conditions of the ship, the propeller surface will be more or less,
affected by biofouling. The surface roughness of the propeller can be increased by marine biofouling (algae,
slime, mussels, barnacles, tubeworms, etc.), calcium fouling or erosion (cavitation, corrosion because of
electrolytic incidents), poor maintenance or even contact damages. Whereas the upper area of the propeller’s
blade is often relatively clean because of the higher flow velocities, the middle area and blade root can be more
affected by biofouling.

For more detailed guidance, refer to the INTERTANKO Guide to Modern Antifouling Systems and Biofouling
Management (2020). Appendix 14 provides hull and propeller performance assessment methods with reference
to ISO 19030.

Appendix 15 presents a study examining the decision-making process for propeller polishing and hull cleaning
with various different cases.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  41

10. Summary

10. Summary
INTERTANKO aims to become a leading association in moving the shipping industry towards zero carbon
emissions, reducing the impact of our activities on the marine environment. We face other increasing
operational challenges, too, such as incentives from financial institutions for more energy-efficient ships and
ever-rising levels of expectations from Charterers.

Nonetheless, our industry contributed to the development of a design index to reduce the amount of CO2
emissions at new building design stage (EEDI), and also to the development of an indicator which ship operators
could use to monitor over time the efficiency in transportation of an individual ship (EEOI).

At present, various stakeholders have started to use operational indicators in many different ways, beyond
their defined scopes or even to promote their own indicators for rating operational performance of a ship
(difficult with simplistic indexing) or to address only a certain liability of the fragmented chain of responsibility
in shipping (an irrelevant indexing).

We have the IMO Data Collection System (DCS) which requires ships to report their voyage data so that
shipping transport efficiency can be assessed globally, based on facts. Despite the limits that the IMO DCS has,
such a global effort to assess ships’ total annual fuel oil consumption (CO2 emissions) promotes the drive to
further reduce environmental impact from ship operations.

Pending further regulations as well as social expectations that each ship should assess its performance and
continue to improve its efficiency, INTERTANKO Members felt compelled to seek a means of such performance
assessments and increased awareness. Several Members were developing their own monitoring systems or
buying turn-key solutions from the market. Comparison or benchmarking such performance assessment
records is very difficult, and near impossible to benchmark without common understanding about what we
really measure and analyse.

Huge interest in energy efficiency of ships and complexity of analysis of a ship’s energy performance have
led to the development of this Guide. At the same time, more concrete proposals to bring international
shipping in compliance with IMO’s GHG reduction targets are materialising. There are some proposals which
provide incentives to reduce GHG emissions from ships while keeping the flexibility needed to reward the more
efficient ship operations.

While decarbonisation will be achieved only by introduction of disruptive technologies/innovative fuels, the
monitoring of performance will remain high on the agenda of all industry stakeholders including regulators.

Financing serious research and development to achieve shipping decarbonisation is already an integral part of
the IMO Initial Strategy for GHG emissions reduction.

Poseidon Principles is another new reality that the shipping industry faces. It simply requires detailed reporting
on ships’ and companies’ performance and efficiency. There is no doubt that such reporting schemes are not
problem-free at their initial stages but they will evolve and hopefully include obligations for other participants
to the efficiency of transportation at sea like charterers, traders, receivers and port authorities whose
responsibilities in GHG emissions from ships is just as important as those of ship owners.

All of this indicates that it will take a period of time for the process of monitoring and assessment of the energy
efficiency to be properly understood by industry. The vessel performance monitoring itself evolves and changes
its focus.

42 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Takeaways

Therefore, it is suggested that INTERTANKO continues to place its focus on vessel performance monitoring,
including the following:

• Monitoring the development of the various energy efficiency practices and learning from the feedback
received.

• Monitoring developments at IMO which could influence the current practices of monitoring ships’
performance and considering the impacts of these.

• Monitoring the data released from the IMO DSC and assessing its significance where relevant.
• In this context, performance monitoring should be seen as an integral part of a larger ‘big data’
framework but in which data-driven decisions can be taken based on a meaningful quality control of
data.

• Considering other important KPIs contributing to improving performance such as vessel traffic, port
logistics, other bottlenecks in the logistic industry, more dynamic and less rigid contractual terms to
allow choices of use of weather routing, virtual time of arrival, terminal arrangements to accommodate
more efficient ships and a few more elements which address possible areas for leveraging towards
better performance.

• Further considering how ships’ performance monitoring has become an important part of the company’s
Corporate Social Responsibilities plan.

Takeaways
The complexity of ship performance monitoring and management of propulsion efficiency is certainly not
entirely addressed and explained by this Guide. However, readers could learn from the experiences gained by
the members contributing to this Guideline and suggestions on how to consider and address the majority of
problems encountered in the application of performance monitoring models for their own ships.

The Guidelines have purposely avoided entering into the analytics of unproven normalisation methods, mainly
concentrating on propulsion efficiency (hull, propeller and main engine) and leaving aside the entire ship
energy efficiency approach which, in a tanker, should have taken into account other important energy use,
such as electricity production, water production, cargo and accommodation heating, air conditioning, inert gas
production, cargo discharging and other machinery efficiency.

Therefore, the main scope is to monitor the hull and the propeller efficiency, engine efficiency, avoidance
of commercial underperformance, monitoring of emissions, fuel saving, paint selection and energy-saving
device selection. This Guide suggests looking at the desired achievement before deciding which system to use,
making reference to ISO 19030.

The amount and frequency of the data will determine the accuracy of the result. ISO 19030 recommends a set
of data and desired frequency. Different sets of data and different time intervals are considered valid.

The quality of data is important. The adequacy of the output is proportional to the quality of the input.
High frequency data, auto-logged data, manual data, noon report and AIS data are useful in building up a
propulsion efficiency study, although with varying accuracy.

The filtering and normalisation of data can be done by a white box approach, black box approach, and hybrid
approach. This Guide adopts mainly the white box approach without dismissing the benefit of other methods,
which in certain cases are the only viable solution. INTERTANKO Working Group members believe the best
approach is based on assessment through the laws of physics and consider statistics as a powerful tool in case
of a lack of proven analytics.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  43

10. Summary

Normalisation against weather represents the most difficult part of propulsion efficiency monitoring. This Guide
suggests that analytics within ISO 15016 and 19030 are yet to solve the problem with necessary precision.

GHG reporting policy is mainly based on so-called “noon report manual data” which represents the base
for regulatory data acquisition regimes (EU MRV and IMO DCS). The rate of adoption of high frequency
auto-logged data across the whole industry is still low. Better penetration is evident in more sophisticated
sectors (tankers for instance). Further study of more sophisticated approaches will be beneficial for the exact
comprehension of the vessel performance, and for the critical review of GHG policies under discussion.

References

– Francesco Belluci and Gil-Yong Han “Tanker Operators’ perspective, Vessel Performance Monitoring”,
HullPIC 2018

– Intertanko/UCL “A Case Study of Fuel Monitoring and Efficiency Indicators for INTERTANKO” (2016),

– Bazari “Ship energy performance benchmarking/rating: methodology and application” (2007),

– SO 19030 (2016), Measurement of changes in hull and propeller performance, ISO, Geneva

– Faber, J. and ’t Hoen, M. (2017) Estimated Index Values of Ships 2009-2016.

– Analysis of the Design Efficiency of Ships that have Entered the Fleet since 2009. Delft.

– Søren Vinther Hansen (2015), PhD thesis, Performance Monitoring of Ships, pp 51-72, Technical
University of Denmark

– Aldous, L. (2015), Ship Operational Efficiency: Performance Models and Uncertainty Analysis, PhD
Thesis, University College London

– Aldous, L; Smith,T; Bucknall, R; Thompson, P (2015), Uncertainty analysis in ship performance monitoring

– Schmode, D; Hympendahl, O; Bellusci, F (2019), Influence of Data Sources on Hull Performance

Prediction

– Montazeri, N; Petersen, JB; Hansen, SV (2019), Autolog Data Processing for Vessel Performance
Application

44 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Contributors

Contributors
Francesco Belluci Scorpio Group

Gil-Yong Han INTERTANKO

Michael vom Baur MvB euroconsult

Volker Bertram DNVGL

Jean-Marc Bonello University College London

Jose Gonzalez Celis Stolt Tankers

Eleftherios Dedes EURONAV

Christos Giordamlis Prisma Electronics (UK) Ltd.

Lefteris Karaminas ABS

John Kokarakis BV

Kostas Korfiatis Tsakos Shipmanagement

Panos Koutsourakis ABS

Luca Letizia Herbert Engineering

Manolis Levants JOTUN AS

Oliver Masnyk Red Rock Power Ltd

Ivana Melillo d’Amico Group

Stavros Papageorgiou LATSCO Marine Management Inc

Dragos Rauta INTERTANKO

Rajat Saxena Epic Gas

Daniel Schmode StormGeo

Ian Sellwood GreenSteam

Michael Servos MINERVA

Jan Otto de Kat ABS

Edwin Pang Royal Institution of Naval Architects (RINA)

Ian Paul V Ships

Lars Penmann Stena Bulk AB

Soren Vinther Hansen Vessel Performance Solutions (VPS)

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  45

Appendices - Appendix 1. Basic Principles of ISO 19030

Appendices
Appendix 1. Basic Principles of ISO 19030

The basic principles of the ISO 19030 standard, highlighting best practices and its limitations, are explained as
follows. It consists of three parts.

Figure 1: three parts of ISO 19030

Part 1 addresses parameters needed to be logged, definitions of four performance indicators, discussions on
sensors, and analysis on methodology in measuring accuracy of the performance indicators. Part 2 defines
the default method for measuring changes in hull and propeller performance and how to calculate four
performance indicators. The default method is mainly characterised by automated high frequency logging
equipment and is generally considered to be a more accurate method. Part 3 outlines alternatives to the
default method and is mainly applicable to lower data recording frequency or manual reporting systems (e.g.
noon data). Most of the alternative methods will result in lower overall accuracy, but increase the applicability
of the standard.

Performance indicators
ISO 19030 defines a set of performance indicators (PIs) in order to determine the effectiveness of hull and
propeller maintenance, repair and retrofit activities. All PIs are based on the concept of speed loss which
is described in Appendix 2. All PIs use a reference (R) and evaluation (E) period. The value of the PIs is the
difference between reference and evaluation periods. All PIs are described briefly below.

46 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 PI-2: In-Service performance

PI-1: Dry-docking performance

Figure 2: Drydocking Performance (PI-1)

Dry-docking performance is the change in hull and propeller performance following the present out-docking
as compared with the average from previous out-dockings. The main purpose of this PI is to determine the
effectiveness of current dry-docking. This PI is useful in evaluating the effect of surface pre-treatment and
initial coating performance and effect of any retrofit activities. However, it should be noted that it is almost
impossible to distinguish the effect between these two factors.

PI-2: In-Service performance

Figure 3: In-Service Performance (PI-2)

In-service performance is a PI suitable to evaluate and quantify hull and propeller performance over the period
following out-docking to the end of the dry-docking interval. This also includes any maintenance activities that
have occurred over the course of the full dry-docking interval (propeller cleanings, hull cleanings, etc.).

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  47

Appendices - Appendix 1. Basic Principles of ISO 19030

PI-3: Maintenance trigger

Figure 4: Maintenance Trigger (PI-3)

Maintenance trigger is the most interesting performance indicator from an operator’s point of view. When the
value of this PI exceeds a certain threshold (defined by the operator based on their operational needs) then a
maintenance activity should be initiated (ex. propeller, hull maintenance or even a dry-dock).

PI-4: Maintenance effect

Figure 5: Maintenance Effect (PI-4)

Maintenance effect is a PI that allows operators to determine and quantify the effectiveness of specific
maintenance events by measuring the difference in performance before the event and after.

48 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Practical experience with ISO 19030

Practical experience with ISO 19030

ISO 19030 is in general a good methodology in measuring and quantifying changes of hull and propeller
performance. Results are meaningful in most cases, even with manual reporting data as long as its quality is
relatively good and the crew onboard is consistently reporting correct values. However, with respect to the scope
of the standard, there is a common misunderstanding that the standard can be used for general performance
monitoring, overall efficiency estimations and fuel consumption metrics. Some of the procedures described
in the standard can be a good starting point for further evaluation of a vessel’s efficiency in general, but one
should not forget the actual scope of the standard that is only to measure hull and propeller performance in
the most practical way as much as possible.

One of the criticisms of the standard is the applicability of the default method (ISO 19030 Part 2) which requires
an auto-logging high frequency unit. A commonly shared argument is that installation and maintenance of
such equipment is a costly and complicated solution. Practical experience has shown that this argument is
valid. At least in the early stages after installation, these units were found to be unreliable. Recent experience
has shown that these units are becoming much more sophisticated, adopting state-of-the-art fault tolerance
and recovery algorithms. They become easier to install and the associated costs are justifiable due to increasing
market competition. However, post-installation maintenance for this type of equipment still poses a challenge.

Parameter filtering, suggested by the standard, is also a topic of common disputes. The argument is that
if all recommended filters are applied, then a statistically significant dataset cannot be realised. This is of
course subject to logging frequency. Higher logging frequency will result in bigger datasets and lower logging
frequency will result in a smaller number of datasets.

Even with a high logging frequency, most of the collected data could be filtered out:

• if the ship sailed with an intermediate draft (between fully laden and ballast voyage),

and/or

• if the actual operational profile (speed and power) was not within the range of the available speed-
power reference curves (e.g. due to slow steaming).

According to the default method (ISO 19030-2), actual displacement should not deviate more than 5% from
the speed-power reference displacement and speed and power range should be within the range of the speed-
power reference data. Hence the need to address these type of problems in the real world.

Practical ways to deal with these two identified problems are to follow the recommendation of Part 3 of the
standard (ISO 19030-3):

• Regarding draft variations, linear interpolation between the speed-power curves is the recommended
way to solve this problem.

• Regarding operational parameters that are not within the available range of speed-power curves, ISO
19030-3 describes two methods1 by which speed-power curves can be constructed for the operational
parameters in question.

This will have an impact on accuracy, but still yields meaningful results. In general, the most practical
scenario for operators is to try as much as possible to adopt the recommendations of the default
method which will provide them with the most accurate results. However, deviation from the default
method can still give meaningful results for operational decisions.

1
Refer to permanent trial trips and passively monitoring, ISO 19030-3, 5.3.1.2.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  49

Appendices - Appendix 1. Basic Principles of ISO 19030

Finally, the standard attracts some criticisms due to its dependency on speed and by this, results are affected
by operational parameters. This challenge was extensively discussed by the ISO working group. Consensus
on a transparent method to tackle this issue was difficult to achieve. There are methods in the industry to
measure hull and propeller performance which are claimed to be accurate enough and not dependent on
speed. However, most of them are relying on the practitioners who use their own approaches not transparent
to the users2. The true strength of the standard is that it recommends a practical and transparent method,
easily applicable by everyone, leading to analytically-driven decisions and making it easier for all stakeholders
to accept the results.

Source:
Input from Manolis Levantis, JOTUN AS, to INTERTANKO Working Group discussions

Reference:
Svend Søyland and Geir Axel Oftedahl, ISO 19030 – Motivation, Scope and Development, HullPIC 2016
CSC (2011), a transparent and reliable hull and propeller performance standard, IMO MEPC 63/4/8

2
These are so-called “grey” or “black” box approaches.

50 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Total resistance

Appendix 2. Background of the speed loss concept in ISO 19030

The goal of any hull performance measurement method is to measure the impact of underwater hull surface
characteristics on the total resistance of a ship.

Total resistance of a ship is influenced by many factors, but components related to hull, wave resistance and
viscous resistance are generally the most important. Contribution to total resistance by the superstructure
is usually less significant than these factors, but should not be neglected (especially for certain ship types
and sizes). Viscous resistance is dominated by frictional resistance of the underwater hull surface. Frictional
resistance is influenced by:

• the hull’s physical roughness (ex. welding seams),

• the surface characteristics of the applied coating and how these change over time, and

• the effect of mechanical damages and marine growth (both experienced through the ship’s lifetime).

It is obvious from the above that frictional resistance can change significantly over time. The isolation and
measurement of frictional resistance should therefore be the goal of any hull performance measurement
method. The question is how best to measure frictional resistance of a ship in service.
Using model tests, such as measuring the resistance of flat plates with given surface characteristics at relevant
Reynolds numbers, can be very misleading, especially when one attempts to simulate full scale ships in service.
This is mainly because the surface of a small flat plate cannot be representative of the surface of a ship in full
scale and the characteristics of the full scale ship’s hull surface change over time. For a full-scale ship, it is
preferable to measure the total resistance and deduct the contribution of frictional resistance.

Total resistance
Total resistance of a ship is a ratio between thrust delivered to the propeller and the achieved speed through
water. There are many factors that influence this ratio:
• Design
o Hull shape

• Operational conditions
o Draught
o Trim
o List
o Ship speed
o Rudder angle

• External conditions
o Wind strength and direction
o Wave height and direction
o Swell
o Water depth
o Water salinity
o Water temperature

It is almost impossible to isolate and quantify the effect of all these factors that influence a ship’s total
resistance. Complex procedures must be followed, and many assumptions are to be made. ISO 15016 and
ITTC guidelines for speed trials is a practical example of existing complexities even at very well controlled
conditions. However, even if one was able to measure or quantify the effects of the above-mentioned factors

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  51

Appendices - Appendix 2. Background of the speed loss concept in ISO 19030

under ideal or controlled conditions, he/she would only be able to measure the ship’s total resistance. The
question remains: how do we quantify the effect of frictional resistance?

Relative versus absolute effects

The idea of how to isolate frictional resistance out of the total resistance is that instead of trying to measure
the frictional resistance directly, one could focus on the change of the total resistance over time. By measuring
the changes of total resistance over time, the focus now is on relative instead of the absolute total resistance.
With this approach, the effect of factors that remain unchanged over time can be neglected.

As mentioned earlier, the main contributors of the total resistance are wave resistance, viscous resistance
and superstructure resistance. Wave resistance stays the same over time, given a certain hull shape and the
same applies to the superstructure’s resistance. Therefore, viscous resistance is the main factor that changes
over time, because it depends on the actual characteristics of the underwater hull surface. Eventually, hull
coating will degrade and hull fouling will develop, thereby significantly changing the underwater hull surface
characteristics.

Given that viscous resistance (predominantly frictional resistance) is the only component that changes over
time under normal operating conditions, monitoring the change in total resistance over a given period of time 3
would enable us to monitor the change in frictional resistance over the same period of time due to hull surface
degradation. The task of measuring frictional resistance is now simplified to track changes of total resistance at
normalised conditions. As such, by focusing only on relevant changes over time, one does not need to correct
all possible effects influencing total resistance, but rather isolate similar (but not necessarily ideal) conditions.

Speed loss concept

The speed loss concept relies on the previously mentioned conclusions that relative changes in total resistance
reflect changes in frictional resistance when factors that affect total resistance are similar.

The method requires measurements of propeller shaft torque, propeller speed (rpm), speed through water,
draught aft and fore, wind speed and wind direction. In addition, GPS speed and direction, water depth,
fuel consumption, rudder angle, propeller thrust, wave height and wave direction are also measurements of
interest. These can be used to validate the data from the required measurements.

One important practical challenge that needs to be highlighted is the lack of thrust measurements. Because
one has to rely on shaft power delivered to the propeller, rather than thrust delivered to the propeller, this
method measures the combined effect of ship hull resistance and propeller efficiency.

The metric for evaluation of hull and propeller performance is the relative speed loss in percentage called
“speed deviation”. Speed deviation is calculated by comparing the ship’s speed observed at any given time
with the speed that the ship is expected to achieve at a measured shaft power (Figure 1).

3
It should be a reasonably long period of time, e.g. longer than 12 months.

52 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Figure 1: Illustration of speed loss as the deviation from an expected speed-power relation

Figure 1: Illustration of speed loss as the deviation from an expected speed-power relation

Likewise, ISO 19030 defines the ship’s performance value (PV) as the percentage speed loss compared with a
reference speed-power relation. Here, the percentage speed loss, Vd, is calculated as follows:

Vm: the measured ship speed through water

Ve: an expected speed through water

In order to calculate the speed expected, speed power curves from model tests or speed trials are used. This
means that speed deviation values are normalised for draught variations. Other factors need to be filtered out
(e.g. a wind speed higher than 16 knots or power and speed values which are not within the range of the
available model). Operational conditions such as manoeuvring drifting, acceleration, deceleration and steering
should also be eliminated. Additional filtering can be applied if other measurements are available (ex. wave
height). As previously mentioned, the remaining dataset is validated with all the other available information
such as fuel consumption and corresponding power values, speed through water with speed over ground
and rpm. The aim of the filtering is to keep typical and comparable conditions, so that the change in speed
deviation can be associated with change in frictional resistance and propeller efficiency over time.

However, one should be very careful on how much filtering should be applied. It is obvious that over-filtering
may result in less variability in the dataset, but on the other hand, the remaining dataset might not be statistically
significant. Filtering should be dependent on the data collection frequency, variability of the dataset and the
time period subject to analysis. There will always be a trade-off between accuracy and data availability. One
possible way to address this is to increase the logging frequency. This will ensure that enough data remains
after filtering procedures and temporary fluctuations can be tracked and eliminated easily.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  53

Appendix 3-1 Examples of operational measures for energy efficiency

Appendix 3-1 Examples of operational measures for energy efficiency

1. Speed optimisation
The selection of ship speed in operation is affected by several factors and is usually a balance between
contractual obligation of voyage time or warranted speed, maximum cargo capacity and minimum fuel
consumption while ensuring safe operation. Optimisation can refer to commercial or technical aspects which
often have conflicting implications on speed decisions. For the main engine, which is the main consumer of
fuel oils, the theoretical technical optimum operation for any ship can be obtained at an engine load where
the minimum Specific Fuel Oil Consumption (SFOC) for the propulsion system on board (the combination of
the propeller and engine) is applicable.

It is rarely the case that this load allows an appropriate commercial operating speed. Thus, ships can determine
an optimal economical speed that is a compromise between factors such as:

• Commercial pressures.

• Regulatory and safety requirements.

• Contractual commitments.

• Daily operational factors.

It many cases, the optimal compromised speed is different from the design speed of the ship. Interventions
such as engine derating and propeller matching and changes of hull geometry may be considered to improve
efficiency. Operational initiatives such as virtual arrival can also help identify an optimal speed for the voyage
considering port activity and avoiding unnecessary higher speeds and excessive waiting time.

At present, speed reduction is still arguably one of the most efficient measures under consideration at IMO. It
is kept as one of the candidates of short-term measures whose potential impact is being evaluated. This would
create another factor that would have to be considered.

Source:
- IMO Train the Trainer Module 3 – From management to operation

2. Virtual time of arrival

Virtual arrival involves ship speed optimisation by adjusting the ship’s port or terminal arrival time when
there is a delay or change in available berth slots. This usually implies sailing at a lower speed with less fuel
consumption and emissions while also avoiding unnecessary waiting time at berth. This operational method
also supports recommendations contained in the SEEMP and TMSA, thus fulfilling initiatives designed to
improve the efficiency and economics of the industry.

54 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Figure 1: Illustration of speed loss as the deviation from an expected speed-power relation

Figure 1: Visual representation of Virtual Arrival (INTERTANKO/OCIMF (2011)

Besides the fuel and emission savings directly gained from reduced speed, virtual arrival sets up an open
channel of communication between owners, charterers and ports and optimises the outcome of voyages
for all parties. It can also maximise utilisation of port operations planning and ensure port safety by reducing
congestion. Crews can benefit from less uncertainty when coming into port and can better plan their
activities on board.

Planning virtual arrival commences with a pre-agreement between the charterer and the owner setting
out the obligations of each party contractually in order to avoid disputes. These include setting up a clear
decision-making process, appropriate clause in the charter party establishing the terms of implementation
and an agreed method of calculating performance and how benefits will be assigned to both parties. Once
virtual arrival is enacted, a process of verification may be required to ensure that all parties agree to the
actions taken and the resulting benefits.

For a practical guide to the implementation, benefits and issues related to virtual arrival, refer to Virtual
Arrival, INTERTANKO/OCIMF (2011).

The IMO Global Industry Alliance is also working to create a Just-In-Time operational guide to be submitted
to the IMO in 2020. Update information can be found at: https://glomeep.imo.org/global-industry-alliance/
gia-resources/.

Sources:
- INTERTANKO/OCIMF “Virtual Arrival Optimising Voyage Management and Reducing Ship Emissions - an
Emissions Management Framework” (2011), 2 Ed.
- IMO Train the Trainer Module 3 – From management to operation.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  55

Appendix 3-1 Examples of operational measures for energy efficiency

3. Weather routeing
While weather routeing has always been part of the responsibilities of the Master on board to ensure safe
passage, this operational measure is now used to allow detailed voyage planning to optimise performance
depending on the sea route weather conditions. Sailing in rough seas can cause delays, damages to ships
or cargo, increased fuel consumption, and may be unsafe for crew. It is in the interest of both owners and
charterers to include weather routeing in their voyage planning. It is understood that if weather routing is
being used on a charter, the ship may deviate from the warranted performance in the charter party and
possibly take a longer route that avoids heavy weather.

Weather service providers use historic databases, long range forecasts, weather satellite information and
weather buoys to create a model that allows for a comprehensive picture of metocean conditions at a high
temporal and geographical resolution. The service can be used on board the ship for the Master to set and
adapt the voyage plan accordingly or a shore-based service can provide the Master with regular advice as to
what route would be optimal. Route optimisation, similar to the speed decision, is a compromise between fuel
consumption, commercial pressures, time constraints and safety.

Final decisions on navigation are always left to the ship’s Masters and weather routeing is a tool that is to be
used at their discretion.

Source:
- IMO Train the Trainer Module 3 – From management to operation, IMO Resolution A.893(21) Guidelines
For Voyage Planning (1999)

4. Cargo optimisation
Cargo optimisation is not a novel concept and is relatively intuitive as the efficiency of a ship in terms of fuel
consumption per unit transport work increases with utilisation. Safety is always a priority. Therefore, trim and
stability calculations should always be carried out and verified to ensure safe passage. Currently, the voluntary
EEOI is in place allowing the company to compare ships’ performance based on their overall efficiency related
to the amount of transport work conducted over a voyage.

As part of the short-term candidate measures under the theme of IMO GHG initial strategy, an Annual
Efficiency Ratio (AER) metric is proposed to establish a baseline of performance and then to regulate gradual
improvements. AER compares the amount of CO2 emitted to the deadweight of the ship and distance sailed
over a year. If AER or similar metrics comes into being as a mandatory measure, cargo optimisation would
become a high priority in this context.

Source:
- IMO Train the Trainer Module 3 – From management to operation

56 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Table 1: Operational Measures

Appendix 3-2 Classification of Energy Efficiency Technologies (EET)

Appendix 3-2 contains discussions of cost-savings and ROI (return on investment). Reader’s discretion is advised:

- The ranges and savings presented are subject to the conditions of each case and are not lineally to be
added.

- The figures and sums should be taken as indicative only.

The purpose of this section is to classify EETs with respect to readiness of deployment, initial costs and
appropriateness of retro-fitting.

The classification of EETs and practices in terms of cost is considered as:

• ‘Low’ when the annual corresponding amount is between 1 $ and 50,000 $

• ‘Medium’ when the annual corresponding amount is over 50,000 $ and less than 100,000$

• ‘High’ when the annual corresponding amount is over 100,000 $

For the estimation of Return on Investment (ROI), the bunker cost was considered equal to 480USD/MT.

Table 1: Operational Measures

Claimed Implementation
Measure Cost ROI
savings up to: feasibility
Propeller polishing 2% Low Easy < 6 months
Hull cleaning 10% Low Moderate < 1 months
Slow steaming Max. 36% Low Easy 0
Virtual port arrival 6% Zero Hard 0
Propulsion efficiency monitoring 5% (over 5
medium Moderate < 24 months
(including sensor installation) years interval)
Weather routeing/ software 5% Low Moderate < 24 months
Port turn-around time 1% Zero Hard 0
Optimization of ballast & trim 1% - 4% Low Moderate < 24 months
Speed optimisation 5% Low Moderate < 12 months
Autopilot adjustment 1% Low Easy < 18 months
Optimised voyage planning 5% Low Hard < 12 months
Optimum use of fans and pumps 0.5% Zero Easy 0
Optimum use of bow-thruster 0.5% Zero Moderate 0
Efficiency control of HVAC system 2% Zero Moderate 0
Variable Frequency Drive (VFD) for
pumps, fans and other electrical 8% Low Easy < 12 months
equipment
Cargo heating and temperature 10% of
control optimisation cargo heating Low Moderate < 12 months
consumption
Minimisation of subcooling effect on
2% Zero Easy 0
vacuum condenser

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  57

Appendix 3-2 Classification of Energy Efficiency Technologies (EET)

Claimed Implementation
Measure Cost ROI
savings up to: feasibility
Usage of fuel oil additives 1% Low Easy < 6 months
Main engine combustion optimisation 4% Low Moderate < 12 months
Optimum settings of fuel oil purifiers 0.2% Zero Easy 0
Improved machinery maintenance note 1
4% Medium Moderate <24 months
Energy management 1% Medium Hard <36 months
Note 1
Minimisation of air system leakages, proper insulation of steam distribution network, electrical insulation of
electric network and overhauling of M/E and A/E as specified by the manufacturer intervals.

Table 2: Energy Efficiency Technology (EET))

Claimed
Savings Implementation Suitable for
Measure Cost ROI
up to (not Feasibility Retrofitting
equal to):
Rudder surf bulb 5% High Moderate < 36 months Yes
Rudder surf fins 1% High Moderate < 132 months Yes
Propeller boss cap fins 2% Medium Moderate < 14 months Yes
Contra rotating propellers 3% High Hard ~ 132 months No
Mewis duct 3%-7% High Moderate ~14 months Yes
Propeller duct 3% High Moderate < 24 months Yes
Wake equalising ducts 2% High Hard < 18 months Yes
Pre swirl fins 2% Medium Hard < 30 months Yes
Anti-fouling paints 6% High Moderate ~ 9 months Yes
Air lubrication 4% High Hard < 60 months Yes
De-rated M/E < 6% High Easy < 60 months Yes
Part load optimisation 3% High Hard 0 Yes
Turbocharger isolation 3.5% No
Hard 0 Yes
Waste heat recovery
2%-5% High Hard < 72 months Yes
(Rankine Cycle) generator
Fuel oil homogenisers 0.5% Medium Moderate < 36 months Yes
Installation of LED lighting
and automatic motion 1%-2% Low Easy < 60 months Yes
sensors
Solar panels for auxiliary 5% of aux.
High Easy < 60 months Yes
loads fuel
Kites and sails 10% High Hard < 60 months Yes
CLT or Kappel propellers 6% High Moderate < 12 months Yes
Grim vane wheel 3% High Moderate < 60 months Yes

Note: Numbers in this table are presented for illustration purposes only.

However, not every EET can be retrofitted to an existing vessel, nor applied in combination with another EET.
The following table presents the compatibility of each hydrodynamic device with each other for the potential
combination of these technologies.

58 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Table 3: Compatibility of EETs

Table 3: Compatibility of EETs

References:
- ABS Energy Efficiency Measures, 2013
- J.S Carlton, Marine Propeller and Propulsion, 2nd edition, Butterworth & Heinemann, Oxford, 2007.
- PhD Thesis “Investigation of Hybrid Systems for Diesel Powered Ships”, E. K. Dedes, 2013.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  59

Appendix 4. Non-Disclosure Agreement (NDA), Ship Performance Data (Section 4.7)

Appendix 4. Non-Disclosure Agreement (NDA), Ship Performance Data (Section 4.7)

[Shipping Company Name]

Legal entity: [Shipping Company Name “AAA”]

Business address:

[Third Party Service Provider]

Legal entity: [Company Name “BBB”]

Business address:

Definitions
AAA and BBB are individually referred to as a “Party” and collectively as the “Parties”. “Disclosing Party”
means the party who discloses Confidential Information. “Receiving Party” means the party who receives
Confidential Information.

Preamble
Whereas the Parties have expressed their mutual interest in having an open dialogue where each of the Parties
may be prepared to disclose in written or electronic or other form certain data concerning technical, chemical,
financial and marketing information of its relevant companies and for this purpose the Parties agree that said
information disclosed hereunder will be of confidential nature.

Therefore, AAA and BBB intend to enter into discussions in relation to hull & propeller performance and
the needed data for its assessment (hereinafter referred to as “Purpose“), which will involve the disclosure
or exchange of certain non-public information. The Parties agree to the following provisions:

1. Confidential Information

a) “Confidential Information” shall mean any kind of information or material in connection with the Purpose
which, by denotation or reasonable circumstances, is considered confidential (including, but not limited
to, technical, scientific, environmental, commercial, legal and financial information, including, but not
limited to, bids, prices and pricing models, templates, reports, drafts, methodologies, protocols, check-
lists, management systems, plans, drawings, software, algorithms, databases, prototypes or process
technologies, trade secrets, know-how and other materials, electronically transmitted documents
including emails, intellectual property rights and proprietary information) and which is disclosed on or
after the effective date of this confidentiality agreement (hereinafter “Agreement”).

b) The term “Confidential Information” also includes all analyses, compilations, studies, notes or other
documents whether prepared by any of the receiving Parties or their directors, officers, employees, agents,
advisors and/or representatives (hereinafter collectively “Representatives”) which are based or contain or
otherwise reflect the Confidential Information.

60 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 4. Duration of Confidentiality Obligations

c) Information shall not be deemed Confidential Information insofar as it

• is or becomes known to the receiving Party (BBB) from the other Party (AAA) or a third Party without
any confidentiality obligation to the disclosing Party;

• is or becomes generally available in the public domain through no act or failure to act on the part of
the receiving Party;

• has demonstrably been developed by the receiving Party (BBB) independently from this Agreement.

2. Confidentiality Obligations

a) The receiving Party undertakes to the disclosing Party that it shall keep the Confidential Information secret
and confidential and

• use it on a need-to-know basis exclusively for the Purpose, unless stated otherwise in this Agreement
or another contract between the Parties;

• neither disclose, disseminate, or publish any Confidential Information, nor use it for any competitive
purpose, provided however that each Party may share such information with its officers, employees,
affiliates, subsidiaries, subcontractors, suppliers or professional advisors who are subject to
confidentiality obligations reflecting the principles herein;

• immediately notify the disclosing Party in writing in the actual or assumed case of loss or a non-
approved disclosure of Confidential Information.

3. Disclosure

a) The receiving Party may disclose Confidential Information insofar as

• the Confidential Information is requested to be disclosed by any competent court, governmental

agency, flag state administration, EU commission or other relevant public authority in accordance with
applicable law, court order or other public regulation;

• The Confidential Information must be disclosed to the relevant accreditation or notifying body in
accordance with any applicable accreditation or certification scheme;

• The Confidential Information must be disclosed in accordance with requirements derived from
international conventions and organisations (including but not limited to, towards the International
Association of Classification Societies (IACS)).

b) The Parties shall grant access to sites and installations to inspectors and representatives of regulatory,
administrative, accreditation and notifying bodies and authorities to the extent such access is required
under any regulatory framework that is applicable to the Purpose, including flag state requirements.
If applicable, access to vessels shall in particular be given to representatives of the EU Commission as
required under EU Regulation 391/2009.

4. Duration of Confidentiality Obligations

The Agreement shall become effective as of the last date of the Parties’ signature.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  61

Appendix 4. Non-Disclosure Agreement (NDA), Ship Performance Data (Section 4.7)

5. Return of Confidential Information

Upon termination of this Agreement, each of the Parties will promptly return to the respective other
Party all documents or other matter furnished by the Disclosing Party that is part of the Confidential
Information together with all copies or transcripts thereof.

6. Reservation of Rights

Each Party reserves all rights in its Confidential Information. The disclosure of Confidential Information by
one Party does not give the other Party or any other person any licence or other right in respect of any
Confidential Information beyond the rights expressly set out in this Agreement.

7. Final Provisions

a) The failure or delay in exercising any right or remedy under this Agreement by a Party shall not constitute
a waiver of the right or remedy of the other Party.

b) Each Party acknowledges and agrees that this Agreement does not contain any commitment to enter into
any agreement in relation to the Purpose or any part of it.

c) Without affecting any other rights or remedies that any Party may have, each Party acknowledges and
agrees that money damages alone might not be an adequate remedy for any breach of this Agreement
or the obligations and undertakings hereunder. The Parties shall be entitled to equitable relief, including
injunction and specific performance for any threatened or actual breach of this Agreement and the
undertakings under it. Such remedies shall not be deemed to be the exclusive remedies for a breach of
said provisions but shall be in addition to all other remedies available at law or equity.

e) Without prior consent from the other Party, which may not be unreasonably withheld, no Party shall
transfer or assign – entirely or1 in part – this Agreement or rights or obligations resulting from it, unless
the recipient of the transfer or the assignment is a legal successor of the transferring or assigning Party
due to fusion, transfer, merger, acquisition, legal reorganisation, sale of property assets or purchase of
essentially all tangible assets the subject matter of this Agreement refers to.

However, AAA shall be entitled to assign the Agreement to affiliated companies within AAA.

f) This Agreement can be amended or modified only by an amendments in writing signed by both Parties.

g) This Agreement shall be governed by and construed in accordance with English law and the Parties submit
to the exclusive jurisdiction of the courts of England Wales.

Signed by

Place: Place:

Date: Date:

for and on behalf of AAA for and on behalf of BBB

62 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Q3. How many vessels are in your fleet?

Appendix 5. Online survey results

Q1. Participating companies

A total of 64 entries were recorded. There were six cases where two entries were made by the same company.
Hence, 61 companies contributed to the survey. For the purpose of this survey, the company names are
anonymised. Assuming that these three pairs of double entries from the same companies would not skew the
overall trend analysis, and for the sake of simplicity, this analysis will take all entries into account as submitted.

Q2. What business do you conduct?

Q3. How many vessels are in your fleet?

Number of
0~30 31~60 61~90 91~120 120~150 151~180 181~210 211~240 241~270 750 755
Tankers
Number of
27 15 5 3 2 1 1 2 0 1 1
companies

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  63

Appendix 5. Online survey results

Q4. What do you consider the most important aspects of performance management?

Q5. Are you aware of ISO 19030?

64 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Q7. Who is responsible for performance management in your company?

Q6. How do you do performance monitoring in your shipping company?

Q7. Who is responsible for performance management in your company?

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  65

Appendix 5. Online survey results

Q8. Do you have vessels equipped with an auto logging system for performance related sensors
(torque, STW, draft, etc.)?

Q9. How many of your vessels are equipped with an auto logging data logger?

Answered:31, Skipped: 33

Number of auto-logged Tankers 0~30 31~60 61~90 91~120 120~150 151~180 181~210
Number of Respondents 18 8 1 0 0 1 1

Respondent A said: 20% of their entire fleet.

Respondent B said: All of their fleet (100%).
Respondent C said: None of their fleet (zero).
Respondent D and E said: 180 and 200 ships respectively.

Q10. Which sensor signals are integrated into the auto logging system onboard?

Answered:32, Skipped: 32

Each bar of the graph opposite represents a percentage of the total 32 respondents who indicated that they
had installed the particular auto log system.

In the graph:

“More” means: More navigational data such as positions and waypoints

“Others”: 7 Respondents reported the following equipment:

M/E temperatures, Manifold pressure, Aux Boiler Steam Pressure, D/G load, In Test Phase, A complete
alarm and monitoring system, Cargo control console and bridge.

66 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Q11. At which interval of time data are collected?

Q10 graph

Q11. At which interval of time data are collected?

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  67

Appendix 5. Online survey results

Q12. Within the data collection period, the data are:

Q13. How reliable is the system? Data for all relevant sensors is available (Reliability as – Consistent
reporting / logging of the data at the required frequency and of sufficient quality for required use).

68 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Q16. Performance data is captured in...

Q14. Do you have vessels that do manual reporting of performance related data?

Q15. How many vessels do manual reporting?

Number of
0~30 31~60 61~90 91~120 120~150 151~180 181~210 211~240 241~270 755
Tankers
Number of
23 11 3 2 1 0 1 2 0 1
Companies

Q16. Performance data is captured in...

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  69

Appendix 5. Online survey results

Q17. Which data is recorded?

In the graph:
“Others”: 7 Respondents reported the following manual data reporting:
Positions, speed, cargo parameters, slip, electrical load, ballast ROB, M.E RPM running hours,
fuel/LO consumptions, current, all typical Noon Reporting data.

Q18. Manual performance data is available...

70 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Q20. What is your experience with the data quality coming from fuel meters?

Q19. How do you measure fuel consumption?

Q20. What is your experience with the data quality coming from fuel meters?

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  71

Appendix 5. Online survey results

Q21. What is your experience with the data quality coming from your torque meter?

Q22. Sensors are maintained....

72 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Q24. In your experience, did the Charterers accept your ships’ performance data?

Q23. Which decisions do you base on findings from performance monitoring?

Q24. In your experience, did the Charterers accept your ships’ performance data?

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  73

Appendix 5. Online survey results

Q25. Your opinions about the future of hull/propeller performance monitoring. Please write your
opinions in a free-form (narrative – 31 answers)

No. Narratives
SCADA systems will be utilised. Dedicated departments will be created at the shipping companies
A1 and there will be a Performance Monitoring Centre, which will also co-operate with Operations
Department.
Too many suppliers in the market that claim they can do a lot. The reality is the software analysis
are limited to the quality of inputs received. Someone still has to sit and physically evaluate the
A2 results. Also, any person can give some sort of results with a set of data. The question is if those
results are actually correct. The industry should be careful with too many people out there that
they claim they are experts in performance analysis.
I think that the monitoring requirements will increase as the focus of fuel consumption and the
A3 resulting emissions increases. The big problem will be what to do - locations that allow hull cleaning
whilst in port are reducing. Dockings are expensive for simply hull cleaning.
This is a critical component of vessel overall performance. Close monitoring is required to
A4 identify any loss of performance due to fouling or paint damage and rectify same, to reduce fuel
consumption and avoid underperformance claims.
The descriptive method that is used (decomposition of total resistance through empirical formulas/
models) could be good for ship design purposes but when handling actual operation data the
A5
produced results can be untrustworthy unless the applied coefficients can be estimated with high
accuracy.
The future is a fully automated data gathering act with a predetermined algorithm which rings the
bell when something is out of the ordinary. We should achieve to reach the 21st century in sense
of data technology so no additional burden is imposed on the crew. Most important is that we stay
A6
pragmatic and realistic. The basic in- and outputs related to performance monitoring are generally
simple and logical, so it should stay like that so also all involved parties (going from charterers to
crew) fully understand what's going on.
This is very important as the AFs today are not having as high performance as they once did.
We will soon start using a system based on reading auto logged data that is based on a neural
A7
networking feature, we have tested and found it to be well in line with e.g. Propulsion Dynamics
Casper reports (that we also use).
Hull/propeller performance is one of the biggest contributors to fuel savings. With more data we
A8 can now make models without towing tank experiments. Experts should manage performance as
it comprises complex analytics.
A9 To become more data dependent (quality of data and use of highly analytical tools).
A10 It will become even more critical...
It will be more important to monitor hull/propeller performance from 2020 due to higher bunker
A11
prices and increased focus on CO2 emissions.
After tin-free anti fouling we see generally poor performance of the alternative coatings, which
means far more hull polishing. This is a challenge since there are restrictions of what ports allow or
A12
available equipment. We don't see a short term solution to this. There are a number of new types
flaunting that claim long term performance, however, the solid proof of this is scarce.
A13 A valuable tool to monitor efficiency. This will become more and more important in the future.
A14 More automated and accurate prediction software would be required.
Monitoring systems can provide great value for all involved stakeholders. Data driven decision
A15 making is the way forward on a transparent and efficient way. Reducing costs and creating a
competitive advantage for all the companies leading the way at performance monitoring.

74 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Q25. Your opinions about the future of hull/propeller performance monitoring. Please write your
opinions in a free-form

A16 It is important and needs detailed data and AI to eliminate variables that affect the results.
A17 Already included in the CP.
Hull /propeller performance monitoring is very important due to the potential bearing and impact
A18 on the vessel fuel consumption. If this is managed well, the fuel savings will be enormous which
aids to lower operating cost and CO2 emission.
A19 It will be much better if the performance report can be presented in way of Graphical and Chart.
A20 Should be standardised and accepted as best practice within the industry.
Performance monitoring in ships is commercially very important. It directly impacts earnings and
is also important for the sustainability profile of vessels and the company. I believe that in future
the focus on vessel performance will increase, but before making it extremely high tech and
A21
complicated, there is a lot that can be achieved by doing the basics right and training the crew to
operate the systems and machinery correctly. There is nothing more energy efficient than an aware
and motivated crew operating a vessel.
A22 Absolutely necessary (for the forthcoming years).
Increased support or incentive to clean hull from legislation side rather than based on the existing
A23
fuel price.
Hull and performance monitoring is necessary. Some challenges to this are the following:
Communication from ship to shore (reliable connections); Communication between various vendors
of onboard equipment. For example AMS manufacturers, and other equipment manufacturers
do not want to open code. Weather routing companies shore-side may not have access to all
information shipboard; It’s great to have a robust program and goals in place, but often it needs a
lot of dedicated personnel to accurately manage. This includes the marine expertise in operating
A24
the ship and also the data management personnel. Often the marine person does not have a robust
understanding of IT systems, communication, and management of data when systems fail or are
not performing as intended.; Improvements that have taken place; Better providers and available
technology/support for systems; Awareness with companies and the cost management aboard
has elevated the importance of bunker management; Companies seeing and demonstrating real
savings.
Accurate digital sensors (flow meters, tank gauging, live weather readings) combined with AI
A25
software.
A26 Annual hull cleaning and propeller polishing is a must for good performance.
A27 Subjectivity is included because manual analysis is the main.
A28 More accurate and consistent data is required and also easy to use analysis.
Better the logger is connected by IT methods to the shore. However, third-party support is required
A29
to set up.
Necessary to clarify and make distinction between regulatory, commercial and technical
A30
performance.
A31 Will come more into focus with the increase in fuel costs associated with 2020.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  75

Appendix 6. Theoretical background of uncertainty analysis

Appendix 6. Theoretical background of uncertainty analysis

1. Evaluation of uncertainty
Uncertainty of a measurement is evaluated using a mathematical model of the measurement and the law
of propagation of uncertainty. It is thus inherently assumed that the measure and Y in question, when not
measured directly, can be modelled mathematically through a functional relationship:

where, X1, X2, …, XN are the parameters affecting the measureand Y.

These may be represented by probability distributions. If the probability distributions of Xis are known, the
probability distribution of the measure can be determined following determination of its mean value and
standard deviation.

The uncertainty related measure of interest will then be determined as the combined standard uncertainty via
error propagation.

For input quantities it is necessary to distinguish between:

Uncorrelated (independent) input quantities, and
Correlated (interdependent) input quantities.
If the operator uses different measuring instruments/sensors to determine the inputs Xi, for uncorrelated
uncertainties, the combined uncertainty is:

Where:
Uy : uncertainty (absolute value) of the measurand Y, and
Uxi : uncertainty (absolute value) of the input quantity Xi.

Sensitivity coefficient:

This equation, called the law of propagation of uncertainty, is based on a 1st order Taylor expansion.

This is essentially a linearisation for error propagation.

If fully correleted (covariance = 1), input quantities X1 ,…, XN are being used to calculate the measurand Y=
Y(X1, X2, …, XN). The uncertainty of Y can be determined by:

76 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 2. Monte Carlo Methods (MCM)

Where:
UY: uncertainty (absolute value) of the measurand Y
UXi: uncertainty (absolute value) of the input quantity Xi

If a variable is a function of multiple other quantities, its uncertainty is termed combined standard uncertainty,
evaluated considering the weight and correlation of all those quantities through the law of propagation of
uncertainty, also known as the GUM method or analytical method.

The uncertainty evaluated as a systematic error 𝛣 is inherent to every measurement, in that it retains the
uncertainty features of the measurement system and normalisation procedure. The uncertainty derived by
direct analysis of the data scatter could instead represent the repeatability error P. The total (combined)
uncertainty uc will be:

A propagation of the uncertainty approach is not appropriate for the following cases:
• Uncertainty is large (uncertainty is considered large when the standard deviation divided by the mean
is greater than 0.3);
• Distributions are not normal;
• Equations are complex;
• Data is correlated, and
• Different uncertainties in different inventory years.

The GUM method is exact or close to exact in cases of linear relationship among the variables, or where
nonlinearities aren’t significant. Where strong nonlinearities or correlations exist among the variables, the
application of the law of propagation of uncertainty outlined above becomes cumbersome and often violates
the boundary of its applicability. In such cases, Monte Carlo Methods (MCM) can be utilised to propagate
the uncertainties tackling the limitations of the GUM method. These methods propagate the uncertainty
by random sampling from probability distributions assigned to the input variables for a specified number of
iterations.

2. Monte Carlo Methods (MCM)

The Monte Carlo Method may be described in a step-wise manner as follows:

Let 𝑥𝑖 be the estimation of the 𝑋𝑖 input variable, 𝑦 the estimation of 𝑌 or the resulting variable with 𝑦 = (𝑥1,
… 𝑥𝑖 … 𝑥𝑛). Then the procedure is:
1) define the Probability Density Functions, PDF, for every 𝑥𝑖,
2) select the number 𝑀 of Monte Carlo iterations/trials,
3) sample a random value for every 𝑥𝑖 from its corresponding PDF,
4) calculate 𝑦 = 𝑓 (𝑥1, … 𝑥𝑖 … 𝑥𝑛),
5) iterate 3) and 4) 𝑀 times to obtain 𝑀 estimations of 𝑦,
6) evaluate (𝑦) as the standard deviation of its resulting (discrete) distribution,
7) evaluate (𝑦) (expanded uncertainty) by choosing a suitable coverage interval.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  77

Appendix 6. Theoretical background of uncertainty analysis

The steps of a Monte Carlo simulation are depicted in the graph below:

Figure 1: Steps to carry out uncertainty analyses using Monte Carlo

The Expanded Uncertainty for 𝑀 Monte Carlo simulations expected to encompass 95% of the cases is taken,
is given by:

U(Xi ) = X — x 0.025M
where X̅ is the mean value of the X𝑖 measurements and 𝑥0.025𝑀 is the 2.5th percentile of the 𝑀 measurements
of X𝑖.

The Monte Carlo simulations are run using algorithms which generate stochastic (i.e., random) values based
on the probability density function (PDF) of the data. The objective of these repeated simulations is to produce
distributions that represent the likelihood of different estimates. Once the simulations have been run, they
are applied to the model, which could be complex or be a simple equation, developed to calculate the final
estimate. To calculate the uncertainty, the confidence interval can then be identified for the final distributions.
Before running Monte Carlo simulations, it is necessary to identify the PDFs that have a good fit with each of
the data sources with key uncertainty sources identified. Once the data has been adjusted to account for biases
and outliers, a variety of goodness-of-fit tests can be applied to identify the PDF best fit to the data.

The choice of the correct number of Monte Carlo iterations 𝑀 to obtain stable results is open to study. They
6
can be as high as 10 iterations. A certain criterion is the desired convergence of the results.

It may be necessary to truncate the fitted PDF, i.e., specify minimum and/or maximum values for the Monte
Carlo simulation (effectively removing data points that lie beyond the acceptable range). For example, it may
be necessary to truncate certain PDFs with very long tails, such as lognormal and gamma distributions, to
prevent the simulation of unrealistically small or large values.

It is recognised that in ship performance evaluation there will be large uncertainty in the input data, distributions
will often be non-normal, equations can be complex and nonlinear, with correlations between datasets and
temporal variations due to the degradation of the hull from fouling. Monte Carlo is the correct approach,
therefore, the use of Monte Carlo uncertainty analyses needs to be more widely adopted in ship performance
assessment.

78 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 3. Probability distribution functions

3. Probability distribution functions

It is known that if a measurement is repeated infinitely many times, the measured values will follow a Gaussian
distribution around a mean. The Gaussian distribution is called the parent distribution. The Gaussian distribution
is given as:

where μ is the mean and σ is the standard deviation. For N samples, the mean X (=μ) is given as:

And the standard deviation Sx (=σ) is given as:

It should be noted that the mean is itself normally distributed according to the Central Limit Theorem. The
standard deviation of the means depends on the number of samples according to the following relation:

Where γ is the confidence interval, typically γ=0.95, t=2 for a normal distribution.

The probability distribution for a type of measurement error is a mathematical description that relates the
frequency of occurrence of values to the values themselves. Error distributions include, but are not limited to
normal, lognormal, uniform (rectangular), triangular, quadratic, cosine, exponential, U-shaped and trapezoidal.
Each distribution is characterised by a set of statistics. The statistics most often used in uncertainty analysis are
the mean or mode and the standard deviation.

Figure 2: Examples of probability distribution functions

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  79

Appendix 6. Theoretical background of uncertainty analysis

1.  The normal distribution should be applied as the default distribution, unless information to the contrary
is available. The normal distribution is most appropriate when the range of uncertainty is small, and
symmetric relative to the mean. The normal distribution arises in situations where many individual inputs
contribute to an overall uncertainty, and in which none of the individual uncertainties dominates the
total uncertainty.

2.  Apply the lognormal distribution if it is suspected that the distribution of the value of interest is skewed
(i.e., non-symmetric) and bounded on one side. If many uncertain variables are multiplied, the product
asymptotically approaches log normality.

3.  Apply the cosine distribution if the value of interest has been subjected to random usage or handling
stress, and is assumed to possess a central tendency.

4.  Apply the quadratic distribution if it is suspected that values are more evenly distributed.

5.  The triangular distribution may be applicable, under certain circumstances, when dealing with parameters
following testing or calibration. It is also the distribution of the sum of two uniformly distributed errors
with equal means and bounding limits. The triangular distribution is appropriate where upper and lower
limits and a preferred value are provided by experts but there is no other information about the PDF. The
triangular distribution can be asymmetrical.

6.  The U-shaped distribution is applicable to quantities controlled by feedback from sensed values, such as
automated environmental control systems.

7.  Apply the uniform distribution if the value of interest is the resolution uncertainty of a digital readout.
The uniform distribution is a special case of the Beta distribution.

4 Uncertainty analysis
Uncertainty is calculated to support decisions based on measurements. Therefore, uncertainty estimates
should realistically reflect the measurement process. In this regard, the person tasked with conducting an
uncertainty analysis must be knowledgeable about the measurement process under investigation.

The general uncertainty analysis procedure consists of the following steps:

1. Define the measurement process.

2. Identify the error sources and distributions.

3. Estimate uncertainties.

4. Combine uncertainties.

5. Report the analysis results.

The first step in any uncertainty analysis is to identify the physical quantity whose value is estimated via
measurement. This quantity, sometimes referred to as the “measure” may be a directly measured value or
indirectly determined through the measurement of other variables. It is also important to describe the test
setup, environmental conditions, technical information about the instruments, reference standards, or other
equipment used and the procedure for obtaining the measurement(s).

This measurement process information is used to identify potential sources of error.

80 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 5 Reporting uncertainty

Measurement process errors are the basic elements of uncertainty analysis. Once these fundamental error
sources have been identified, then the appropriate distributions are selected to characterise the statistical
nature of the measurement errors.

With a basic understanding of error distributions and their statistics, we can estimate uncertainties. The
spread in an error distribution is quantified by the distribution’s standard deviation, which is the square
root of the distribution variance. Measurement uncertainty is equal to the standard deviation of the error
distribution. There are two approaches to estimating measurement uncertainty:

• Type A estimates involve data sampling and analysis.

• Type B estimates use technical knowledge or recollected experience of measurement processes.

Because uncertainty is equal to the square root of the distribution variance, uncertainties from different error
sources can be combined by applying the “variance addition rule.” Variance addition provides a method for
correctly combining uncertainties that accounts for correlations between error sources. When uncertainties
are combined, it is also important to estimate the degrees of freedom for the combined uncertainty.
Generally speaking, degrees of freedom signify the amount of information or knowledge that went into an
uncertainty estimate.

5 Reporting uncertainty
When reporting the results of an uncertainty analysis, the following information should be included:

1. The estimated value of the quantity of interest and its combined uncertainty and degrees of freedom.

2. The mathematical relationship between the quantity of interest and the measured components
(applies to multivariate measurements).

3. The value of each measurement component and its combined uncertainty and degrees of freedom.

4. A list of the measurement process uncertainties and associated degrees of freedom for each
component, along with a description of how they were estimated.

5. A list of applicable correlation coefficients, including any cross-correlations between component

uncertainties.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  81

Appendix 7. Uncertainty in sensor measurements

Appendix 7. Uncertainty in sensor measurements

1. Erroneous sensor readings

Unfortunately, although sensor technology is ever improving, there are frequent examples of sensors outputting
values that are not linked to the parameter being measured.

• Complete sensor failure requiring intervention.

• Sporadic erroneous readings that are not representative of the measured value.

• Long-term sensor disruption that requires intervention.

• Interference with the sensor from external sources.

• Failure in communications between sensor and data collection.

Sensors with fixed outputs, outputs that are constantly zero and outputs that are significantly different from
another proxy are typical examples. In many cases these can be recognised by an intelligent data collection
system.

Consider a simple example of a log speed indicator:

• It might fail completely resulting in a fixed output where the speed of the vessel was apparently fixed
for a period. These indicators typically have some noise present so this could even be detected over a
period as short as a few minutes.

• It might experience sporadic failures showing zero jump readings, for example, that would be suspicious.
However, other random jumps can be difficult to distinguish from true behaviour.

• It might become damaged such that it suddenly begins to show under reading. This should be detectable
by comparison with another proxy.

As much as the operator can, and as appropriate for the application, all of this erroneous data should be
removed from the analysis. The data is not representative of the parameter being measured and will introduce
errors in the performance indicator.

Comparison with a proxy

We have considered proxy inputs into performance indicators. Frequently a measurement will also have a proxy
against which it can be compared in order to identify erroneous readings.

Take again the log speed indicator. GPS vessel speed measurements are more reliable but represent the
speed over ground. However the separation between the two represents the current and this can be given a
reasonable upper bound.

As an example, we might implement a scheme that removes all speed log measurements (assume them faulty)
that differ from GPS speed by more than 4 knots. This value being defined because we have decided that there
is only a small chance that currents that large will be observed on the vessel’s route.

One word of caution, an astute Captain will typically try to follow the current, meaning that more often
than not, the speed over ground will exceed the speed through water. Our removal tool for poor data is
now inherently biased towards removing lower log speed outputs. The risk is that we introduce bias in our
measurements. This sort of error could be corrected using external current measurement sources if that was
felt necessary.

82 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Disturbance in the measurement technology

Sensor modelling
As an alternative to a proxy, a model, or digital twin, of the sensor can be created. Modern techniques now
enable this to be automated and this can be a useful technique.

2. Noise
Noise in our application is typically identified as high-frequency disturbances seen over a measurement.
However, it is critical to understand the source of the noise so that we know how to remove its influence on
the final value.

With today’s measurement technologies, measurement frequencies are typically much higher than we need to
create vessel performance indicators. We are more interested in identifying longer term impacts. Our challenge
is therefore to address high frequency influences. To achieve this, we then need to separate the noise in the
measured variables from the noise in the measurement system.

Disturbances in the measured parameters

In this section, some indications of the potential problems will be outlined.

Log Speed indicator

Let’s consider as an example the log speed indicator. Regardless of technologies, there will be natural
disturbances occurring in the flow of water around the vessel. These might be high frequency disturbances
caused by turbulence in the flow or perhaps lower-frequency disturbances caused by slamming into waves.
There are a set of disturbances that are genuine variations in the flow of water. These are not erroneous
measurements, yet we want to remove them from our performance indicator. Averaging is common in this
situation. Note, however, that clipping (removing) larger variations might introduce a bias into the resulting
value.

Anemometer
At very low wind speeds, the anemometer is likely to return a scatter of results representing the wind angle.
These are genuine in the sense that they are the output from the anemometer as it spins helplessly with the
vessel motion. They add no value to determining the apparent wind angle. However, knowing that the wind
speed is very low would allow us to remove them from our analysis. This will take the form of ignoring the
value, a form of clipping. Noting that averaging the wind angle is completely misleading, this needs to be
treated as a vector with the wind speed.

Fuel flow meter

It is typical to observe spikes in fuel flow due to filter flushing. The spike actually represents fuel that has been
used, in the sense that it is no longer available. However, it does not represent fuel that was converted into
energy to propel the vessel. In this case, it might be important to include the spike in some performance
indicators whilst removing it from others.

Disturbance in the measurement technology

The process of measurement introduces errors, some of which will appear as noise in the measured value.

Inherent noise
At some level, all measurement systems have noise present. Even the best electronic circuits of measurement
systems will introduce noise. However, in our applications this is rarely a significant problem. This can usually
be identified by isolating the measurement system from the value being measured.

Resolution
Resolution is the granularity with which the measurement system can measure the parameter. This is not
typically an issue with modern technology in our application but can be seen as random low-level noise in a
measured parameter. It can be removed by averaging (low-pass filters).

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  83

Appendix 7. Uncertainty in sensor measurements

Sampling speeds
Modern measurement techniques will yield measurement results periodically with samples of the actual
measurement. This can be exacerbated by the communications systems onboard the ship, especially when
data is lost between the samples. Today, this is not a major issue in shipping where sampling speeds are much
higher than we need for our application.

Outliers
Outliers are results that are clearly not valid. It is possible to consider a sensor-related problem and determine
whether a value could possibly represent the parameter that is being measured. These values are frequently
observed on graphs of collected data where single points stand out from the crowd.

However, these points need to be handled with care. For example, when we know that the ship is not capable
of exceeding 20 knots through the water, we can choose to clip outputs from the log speed indicator above
20 knots. However, we must remember that the log speed indicator has noise present potentially resulting
in readings higher than 20 knots. This is not due to a sudden change in the ship’s speed and it is rather due
to turbulent flow. Clipping of only the high side of the noise would result in the introduction of an offset in
the measurement. This kind of noise should therefore be averaged out. Some analysis is required to justify
removing outliers or to ensure that the impact of removing outliers does not introduce bias. For example, we
recommend a comparison with a proxy where the average error in the proxy can reasonably be assumed to
be zero.

3. Calibration of sensors in vessel performance measurements

This section will focus on the aspects relating to ship performance measurement processes through a number
of examples, avoiding calibration in general. This is to enable the right questions to be asked of the supplier or
installer to ensure appropriate attention is paid to the management of uncertainty.

Sensors can “slip”, as their errors keep changing during prolonged operation. All sensors should be
maintained and re-calibrated according to the manufacturer’s guidelines to ensure their continued operation
to specification. This is a process of comparing a sensor reading with a known good source and adjusting the
sensor to give the same output.

One of the first problems in sensor calibration, and particularly in shipping, is to find a standard or certified
piece of equipment against which to calibrate.

Calibration typically deals with zero offset errors, slope errors and linearity errors (although manufacturers
make every attempt to minimise these).

84 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 3. Calibration of sensors in vessel performance measurements

If the traditional calibration processes fails, we need to find other ways to confirm the accuracy of a sensor:
1. Comparison with a proxy.
2. Installation of a temporary second measurement instrument.
3. Recording at known operational points.

Sensors installed where factory calibration is relied upon

Some sensors are very difficult to calibrate during their use (in-situ). For example, fuel mass flow sensors.

When they are installed, they were already calibrated in the factory with water. These sensors are understood
to be quite robust, but the impact of the very low fuel flows, vibration and mounting location are not well
researched and an in-situ calibration where possible would be preferable.

Sensors installed where it is typically only possible to carry out a single point calibration after installation

Some sensors can be calibrated during the installation process, however, this typically takes place when the
ship is stationary, and only the “zero” can be calibrated. Torque meters are a good example. It is understood
that a well installed torque meter is reasonably accurate. However, the installation is complex and a poor
installation can have a significant impact on the accuracy, given that there is rarely a zero torque on a shaft.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  85

Appendix 7. Uncertainty in sensor measurements

Sensors that are compared and require calibration in pairs

On many of our vessels we have circulating fuel, and systems where there is fuel flow without fuel consumption.
To measure consumption we put fuel flow meters on either side of the consumer and look for the difference.
Frequently we have high flow and low consumption meaning that even minor calibration errors will have a
significant impact on the difference. In this scenario, very few applications will need further calibration (e.g.
with the zero consumption situation, adjust one flow meter to match the other, setting the difference to zero).

4. Sensors designed for other applications

Consider the mass flow meter. Commercial shipping flow meter installations frequently use mass flow meters.
They are designed by manufacturers for a range of applications. When used in main flow measurement, the
requirement to keep the pressure drop across the sensor to a minimum drives the manufacturer to recommend
a larger flow meter. However, when we compare the manufacturer’s specifications, we find that the flow
meters were operated with increasing inaccuracy.

As an example, a typical mass flow meter for our applications might have an offset error of 0.02 tonnes per
day and slope error as small as 0.2% of the measured value. This slope error only applies to flows that are more
than approximately 20% of the maximum flow rate. Below this rate, the error increases exponentially. To keep
the pressure drop low, vessels frequently use flow meters with a high maximum flow, perhaps 150 tonnes per
day. This means we are frequently operating the flow meter with poor sensor performance.

5. Accumulation of errors
It is important to understand whether we are measuring a parameter or the rate of change of that parameter
and hence deducing the parameter itself. We may make different measurements of the same parameter more
applicable in different circumstances.

For example, fuel flow meters measure the rate of change of the fuel remaining on board, not the value itself.
Take the example of total fuel consumed. If we consider the total volume consumed, it is to be for one day,
then we have two ways to determine this; measuring the fuel remaining on board at the start and at the end

86 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Due to the external conditions

of the day and calculating the difference or summing the fuel flow measured by the flow meters. Both would
seem reasonable and their uncertainty can be determined.

However, if we do this for one year (as might be required for MRV), we can take many FROB measurements to
determine the fuel used between measurements for the year. Note that the measurements are unbiased and
errors should not accumulate. However, if we use fuel flow meters for the same calculation, any error in the
measured rate of change of FROB will be consistent and accumulate through the year and into the final result.

Below is an example of FROB measurements (red) against successive FROB calculations using fuel flow readings
(blue, green, yellow and pink, gaps at point of FROB measurements for re-alignment).

850,000

800,000

750,000

700,000

650,000

600,000

550,000

500,000

450,000

400,000

350,000

300,000

250,000

200,000

150,000

100,000

50,000

7-Feb 14-Feb 21-Feb 28-Feb 7-Mar 14-Mar 21-Mar

Date

Due to the external conditions

Some sensors are effected by parameters other than the measured parameter and ones completely outside
of our control. For example, draught sensors. The draught marked on the side of a ship is directly measuring
the draught, but a draught sensor is typically a pressure sensor. It is not directly measuring the draught and
it measures pressure at the bottom of the ship, as a proxy of the draught. The draught is then determined
taking into account the external air pressure and the salinity of the water, two other parameters each with
uncertainty.

Some performance indicators might survive using an approximation. They assume a fixed value for the air
pressure and salinity, or we could reduce the uncertainty by measuring them directly.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  87

Appendix 7. Uncertainty in sensor measurements

Relevance of the point of measurement

We have seen how the point of measurement can impact the result. We carefully mount an anemometer to
avoid turbulence caused by the vessel. But some parameters will be different at different points of measurements
even without challenges of installation. For example, wind speed varies with height above the water. Even in
the absence of turbulence from the ship, anemometers mounted at different heights will produce different
measurements. To reduce the uncertainty, we need to measure wind speed at the right height.

6. Manual entry by crew or shore staff

Today, a complete performance monitoring system still requires that some data is entered by people. This
involves manual work by people taking readings from sensors during which human factors are added to the
uncertainty outlined already, or people adding their own opinions (of the weather for example). We are all
human and make mistakes. We can be driven by motivations that distract us from precisely entering correct
data. There are many commonly observed causes of uncertainty:

• Typing errors.

• Mistakes like time zone errors.

• Misunderstanding like unit errors.

• Errors due to the time difference of taking related readings.

• Misguided, like the desire to be seen in a positive light.

• Potentially even deliberate attempts to mislead.

Much of the earlier discussion applies to correcting this data also; use a proxy, remove outliers etc. Modelling
can be a powerful tool in detecting deliberate manipulation.

A good practice in order to minimise human error is immediate feedback. We do not want to remove
responsibility from the operator in entering accurate data but we do want to provide them with the best tools
by which they can correct their entries at the time the entries are being made.

Opposite is an analysis of a number of manual entries that differed from the fuel flow meter measurements.
In the graph, all entries except the zero errors are mistaken ones in some form. This is not statistical rather a
sample-based observation.

88 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 7. Interpretation of the indicator

Fuel calibration error vs sensor (KG/s)

7. Interpretation of the indicator

Interpretation and use of indicators are not within the scope of this document. Uncertainty analysis is critical
in understanding the limitations of any indicator. We may develop indicators that are valid for use in some
circumstances. These indicators may not be applicable for other circumstances due to uncertainty. It is essential
that appropriate applications are communicated with relevant performance indicators. A frequent scenario
where this might be a problem would be indicators developed for comparison of performance of one ship at
two points in time, which would then mistakenly be used to compare two other different ships.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  89

Appendix 8. Uncertainty, today and ISO 19030 (Section 7.4)

Appendix 8. Uncertainty, today and ISO 19030 (Section 7.4)

This Appendix discusses how uncertainty is translated to measurements and then to decision-making. Various
methods recommended in ISO 19030 will be explained.

We begin with basic principles about data processing units:

• Analogue signals are sensitive to electrical corruption.

• Processing units are sensitive to noise.

• Digital signals are sensitive to conditions for formulas used.

• Measurements are sensitive to filters and calculation methods used.

• Signals received are feeding up different processing units.

• Processing units are feeding different outputs.

Figure 1: Example of digital output datasets from 3 analogue signal inputs

Figure 1 visualises potential digital output datasets from three analogue signal inputs. Without taking into
consideration various stages of processing the signals and the produced output, the expected outcome is
subject only to statistical processing and generates uncertainty in calculations.

Filtering can reduce uncertainty, but if not properly designed, then it hides the actual behaviour of the ship
creating more uncertainty. In addition, measurements may vary relative to operation conditions (i.e. power
measurements when the ship’s speed is increasing).

ISO 19030 is an effort to develop a common and comparable approach on analysing data and calculating
parameters based on the same operational conditions. The outcome is a cluster of measurements that meets
the criteria defined by the standard and filter out uncertainty of measurements about the ship’s condition
data. Quantifying the uncertainty in ship performance studies, related with the measurement aspects, is a
key aspect as it establishes the level of confidence in a decision-making process and is directly associated
with the appropriateness or risk of the decision.

Within the framework of ISO 19030, appropriate use of performance indicators is subject to awareness of
the extent uncertainty influences the accuracy of each performance indicator. Uncertainties inherent in the
sensors used for measuring parameters (either primary or secondary) as well as the measurement procedures
are the key functions that affect the level of uncertainty of the performance value.

90 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Appendix 8. Uncertainty, today and ISO 19030 (Section 7.4)

For the estimation of uncertainty, Part 1 of ISO 19030 describes a method that incorporates the sensor measurement
accuracies and the specification of measurement procedures. It deploys a simulation of a reference ship’s operation
in combination with a simulation of the data measurement, sampling, and analysis processes. The theoretical
background of the method to model the error propagation is described in the uncertainty analysis section of this Best
Practice.

In this sub-section, key results of the uncertainty analysis that associate with uncertainty of measurement sensors and
procedures with accuracy of the performance indicator will be presented.

For calculation of the measurement parameters, ISO 19030 Part 2 defines two variants of the basic method and
Part 3 defines 4 alternative methods. (Table synopsises the sensor type, its associated accuracies when applicable
and the measurement procedures for all methods. Method 2 is described in Part 2, where each variant differs on the
estimation of delivered power. Its Annex B and C refer.

On the other hand, Methods 3-1 to 3-4 are defined in Part 3 as proxies for the measurement of parameters when
the methods of Part 2 are not available.

The methodology for estimation of the accuracy of performance indicators is also based on assumptions of parameters,
related to data measurement, sampling, and analysis process, that could affect the uncertainty of the performance
indicator. For example, sample size and the sensor’s precision affect the precision of the performance indicator, while
the bias and the drift of the sensor affect both precision and bias of the values of the performance indicator.

The effect of some parameters like the speed variability per day, the operational profile and those related with the bias
of the measurements, are not included in the study as it is constant between the reference and evaluation periods
and is therefore cancelled. This is supported also by the fact that the assumptions are principally common between
the periods and the performance indicator targets rather than absolute performance. More details can be found in
Aldous et al (2015).

Table 1: Measurement parameters and the associated sensor accuracies (to 1 σ) per each method
included in ISO 19030. The following contents are derived from ISO 19030

Method Method
Method 3-1 Method 3-2 Method 3-3 Method 3-4
2-Annex B 2-Annex C
Speed Speed log Speed log Speed log 1
Proxy using Speed log 1
Proxy using
(3%) (3%) (3%) SOG (5%) (3%) SOG (5%)
Delivered Torque meter 2
Fuel 3
Fuel Torque meter Torque meter 3Fuel
Power and rpm meter Consumption Consumption and rpm and rpm Consumption
(1.1% to 5.0%) (5.6% to 7%) proxy (10%) meter (3.6%) meter (3.6%) proxy (10%)
Frequency of Every 15s Every 15s Every 15s Every 15s Daily Daily
measurement
Trim/draft Draft gauges Draft gauges Draft gauges 4
Draft mark 4
Draft mark 4
Draft mark
(+/-0.1 m) (+/-0.1 m) (+/-0.1 m) reading (from reading (from reading (from
last port) and last port) and last port) and
tank soundings tank soundings tank soundings
(+/-0.1 m) (+/-0.1 m) (+/-0.1 m)
Water depth Echo sounder Echo sounder Echo sounder Echo sounder Echo sounder Echo sounder
Rudder angle Respective Respective Respective Respective None None
Indicator Indicator Indicator Indicator
Wind speed Anemometer Anemometer Anemometer Anemometer Anemometer Anemometer
and direction
1
The method approximates speed through water from speed over ground measurements with corresponding accuracy of 5% (to
1σ).
2
Method uses the continuous measurements of fuel flow (mass or volumetric), engine-specific SFOC reference curve and bunker
analysis data (calorific value, density and density change rate).
3
Refers to manual methods for estimation of fuel consumption such as physical sounding of tanks or by taking manual meter
readings. Provided the fuel consumed, delivered power is estimated accordingly to procedure of Method 2-Annex C
4
Loading computer can be also used.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  91

Appendix 8. Uncertainty, today and ISO 19030 (Section 7.4)

Under the assumptions mentioned above, the uncertainty analysis is applied for all methods of 1. Due to the
range in the accuracy of some parameters in Methods 2 (Annex B and C), there is a respective range in the
uncertainty values, and therefore the average values of them are derived respectively. The results are expressed
as the uncertainty to a 95 % confidence interval, which means under this standard variation, the ‘true’ value is
expected to lie between the values with a probability of 95 %. Using as performance value the percentage of
speed loss, Figure 2 presents the uncertainty values for the performance indicator for each method, taking into
account several time intervals. It has been derived according to the next formula, as provided by ISO 19030:

For the results of Figure 1, the same duration of the evaluation and reference period has been considered.
It appears that the best results (lowest uncertainty) are achieved by the Method of Part 2 (Annex B) and
followed by those of the Method of Part 2 (Annex C). Methods 3-1 and 3-2 present a small increase of the
uncertainty, while Methods 3-3 and 3-4 provide much higher uncertainty values.
Figure 1: Performance indicator uncertainty to a 2 sigma (95% confidence interval))

In order to visualise the importance of considering the uncertainty value, the following next cases are presented:

• Case 1 has a performance indicator of -2.4 which means a decrease in speed of 2.4 % per annum. The
reference and evaluation period is 12 months each one.

• Case 2 has a performance indicator of -0.8, corresponding to 3-monthly reference and evaluation
period respectively.

In both cases at t=0, the performance value is equal to 0. Figure 3 presents the respective results per each case
and for Methods 2-Annex B, Method 3-2, Method 3-3 and Method 3-4, too. It appears that for the shorter
period, Method 2-Annex B marginally satisfies the required confidence level, while the other methods and

92 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Appendix 8. Uncertainty, today and ISO 19030 (Section 7.4)

especially 3-3 and 3-4 could not provide safe conclusions as the lower and upper bounds are overlapped. For
the longer period (performance indicator is equal to -2.4), the first two examined methods could be considered
satisfactorily, whereas the remaining two would fail.

Therefore for short periods of monitoring and when low deviations of performance are expected, only the
high precision and frequency sampling methods could be an option. For the longer monitoring periods, the
validity of Methods 3-3 and 3-4 depends on the level of performance deviation. To this end, Figure 4 shows
the critical value of the performance indicator for validity of each method. Specifically, for Method 3-3, if the
performance indicator is expected to be above 5 (in absolute value), then it may be used, while the respective
value for Method 3-4 is 9.25.

The first two methods are quite close and would need at least a value of 0.75.

Figure 3: Blue pairs corresponds to annum performance (per reference and evaluation period),
while the orange to an overall 6-month. 95% confidence ranges are shown. Points have been
located on the mean time of each period per each case. Each arrow shows the total time duration
(reference and evaluation period)

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  93

Appendix 8. Uncertainty, today and ISO 19030 (Section 7.4)

Figure 4: Performance indicator values versus validity for acceptable uncertainty (95% confidence
interval) presented with positive difference of bounds. The difference is defined between the
lower and upper bound value of the reference and evaluation period respectively.

Therefore, in general terms, the absolute uncertainty is reduced through any of the following:

• increasing the accuracy of sensors and measurements of the performance indicator calculation’s
inputs;

• increasing the frequency of measurements;

• increasing the time period of the reference and evaluation period.

Finally, as ISO 19030 mentions, it is advised that when using these quantifications of uncertainty, attention is
paid to ensure the applicability of key assumptions and that these results are treated as indicative values only.

References:
Aldous L., Smith T., Bucknall R., Thompson P. Uncertainty Analysis in Ship Performance Modelling Ocean
Engineering. 2015

ISO 19030 Ships and marine technology — Measurement of changes in hull and propeller performance, Parts
1, 2 and 3.

94 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 Measuring principles

Appendix 9. How to measure shaft power

Performance monitoring normally relies on measuring developed power. It is possible to detect variations and
trends using fuel consumption as a proxy for developed power, but in order to distinguish between variations
outside the engine room and engine-related deterioration, fuel consumption figures should be supported by
knowledge of developed power.

The primary value of developed power is the indicated power, Pi, released by the combustion in the cylinder.
With modern sensor techniques, the indicated power can be calculated with good accuracy. The indicated
power can be used directly for calculations of specific fuel oil consumptions etc. bearing in mind that the basis
for calculation is not the commonly used brake power, Pb. For a ship without a torque meter, the indicated
power is an important measured value.

The brake power, Pb, can be calculated from the indicated power by multiplying with the mechanical efficiency,
ηmech which can normally be found in the shop trial record of the engine.

For reliable performance monitoring, it is common to have a fixed mounted torque meter installed at the
output flange or shaft of the engine where the engine brake power, Pb, will be measured. Brake power is
measured at the shop trial with a water brake as load or with electrical generator as load in case of a generator
engine.

For a two-stroke slow speed engine built with thrust bearing, the mechanical efficiency will be marginally
reduced for the engine in operation compared to the shop trial.

For engines with reduction gear, the loss in the gear box will reduce the shaft power Ps with the mechanical
efficiency of the gear box and losses in the thrust bearing.

For a two-stroke engine, the initial project figures of specific fuel oil consumption will use the brake power
measured at the flywheel and in most new building contracts it will be this figure measured at the engine
builder’s shop trial.

The delivered power to the propeller Pd can then be calculated by use of the assumed shaft efficiency. For most
installations, the shaft efficiency is close to 99% of the delivered power.

Whichever way is chosen, to include or disregard the gearbox and shaft efficiencies, all calculations and
comparisons should follow the same assumptions and procedures.

Measuring principles
Torque meters for marine use are based on one of following principles:

Phase difference: Light or magnetic flux change is measured as an indication of the shaft section
rotation. One way is to mount two rings on the shaft and detect the shift in position
when the shaft is twisted.

Displacement: Rotation displacement between sections of the shaft is measured and converted to
rotational angle. Two rings can be physically in contact with each other and the
change in position will be a measure of the torque.

Strain gauge: The strain on the surface of the shaft is measured with the aid of an on-glued strain
gauge which is converted to a shaft rotational angle. This method is common but it
introduces additional error due to ageing of the glue.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  95

Appendix 9. How to measure shaft power

Magnetostrictive: Torsional rotation is changing the magnetic flux in the shaft which can be
converted to a measure of the twist angle.

None of the above meters are directly measuring the power, but instead they calculate the torque and measure
shaft rotational speed and multiply them, P = M*n.

Rotational speed and torque measurements have limits to their accuracy. ISO 15550:2002 requires an engine
test installation to comply with following limits of inaccuracy:

Engine Brake Torque Ttq Permissible deviation: +/-2 %

Engine speed n Permissible deviation : +/- 2%
The product of Ttq * n = Pb, Brake power have Permissible deviation of : +/- 3%

All measuring principles of torque are based in one way or the other on the material shaft properties and
the shaft deformation under torque. The shaft behaviour under external torque is described by the torsional
module, G.

Unfortunately it is difficult to find reliable values on G-modules for intermediate/propeller shafts. Some reports
indicate a wide spread in G module between shafts, even from same maker and time period.

The International Towing Tank Committee (ITTC) is recommending using the value of 82,4 GPa for the torsional
module, G. The International Standard Organization (ISO) has suggested using 80 GPa.

Material tests of propeller/intermediate shafts have returned values between 77 and 82 GPa, with test pieces
from the same maker and by the same laboratory, indicating that the ITTC value might be in the higher region.
Assuming that 80 GPa is a reasonable estimate for the average G module, the deviation could be estimated to
+/- 2 GPa which would indicate an uncertainty of about 2,5% (k=2).

For sea trials in Asia, they often use the value of G=80 GPa.

It can be of interest to calculate the uncertainty for a fictive power measurement:

Assuming deviation of the rpm meter is 2%
Assuming deviation of the torque meter is 2%
Assuming deviation of the G module is 2.5%

Uncertainty = √(〖(2〗^2 ) + 2^2 + 〖2.5〗^2) .. 3,8

Here, the unknown value of the shaft torsional module G for simplicity is treated as an uncertainty.

96 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 2. Energy Management Systems

Appendix 10. Energy Efficiency Management Plan - A case study of a shipping company
1. Energy management
Energy is one of the most imperative aspects of the shipping industry. Key drivers for improving energy
efficiency of shipping operations, through effective energy management are:

• Reducing the environmental footprint,

• regulatory compliance, and

• improving profitability and supporting sustainable development.

Effective energy management provides a competitive advantage to businesses through the reduction of
operational costs and exposure to fuel costs.

Effective energy management can also strengthen a company's corporate social responsibility. Global
consumers are well aware of environmental issues and, as a result, good corporate citizenship behaviour on
energy management is expected by stakeholders, including banks and shareholders.

Work to improve on energy management is complex, and not an immediate fit to some organisations’
structures and business models. Building competence and achieving success in energy management requires
significant effort and resources and can take several years.

2. Energy Management Systems

According to ISO Standard 50001 on Energy Management Systems, an Energy Management System (EnMS)
“…is a set of interrelated or interacting elements to establish an energy policy and energy objectives, and
processes and procedures to achieve those objectives”. A structured EnMS enables an organisation to achieve
its policy commitments and take action as needed to improve its energy performance.

Given the fact that a significant number of companies have already put in place an EnMS, some of them
certified under ISO Standard 140001 (Environmental Management Systems – requirements with guidance
for use), there must be a clear distinction between setting targets for achieving regulatory requirements for
emissions reduction (e.g. EEDI, EU MRV, GHG) and improving energy efficiency in real terms, which are not
always aligned.

Compared to ISO Standard 50001, ISO 14001 has a broader scope covering all environmental aspects while the
former elaborates specifically on energy use, addressing procurement and design parameters for equipment,
systems and processes. Personnel that are working for or on behalf of the organisation and might have an
impact on its energy performance are addressed in both standards, whilst ISO 14001 adopts a loose approach.
Adoption of either of these Standards, or at least their basic principles, and integration of those into one
management system is recommended.

The EnMS approach adopted in this document is based on the principle of monitor, manage and optimise.

2.1 Monitor
Typically, a monitoring system comprises a combination of manual reporting (includes noon data, logs,
dedicated forms that crew need to maintain, etc.), online performance monitoring systems, third-party sources
(AIS, weather, etc.), Planned Maintenance System, external and internal audits, and observations of technical
and marine superintendents.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  97

Appendix 10. Energy Efficiency Management Plan - A case study of a shipping company

Increasing the energy efficiency of a ship operator is the intended result of increasing the efficiency of the
monitoring systems used to collect the required information, which supports the management system
and ensures that its effectiveness is maintained. As such, components of the monitoring system should
be compatible, providing data integrity and maximum data flow, while their integration should focus on
transparency, elimination of duplication and reduction of crew workload and potential human errors.

Transparency is becoming an industry norm, since charterers, cargo owners, banks, investors and insurance
companies are increasingly demanding proof of environmental and operational efficiency commitments prior
to making contract decisions, as do the regulators (EU, IMO, EPA, Flag Administrations).

The monitoring step can be considered as complete when the energy baselines of all of the company’s vessels
in terms of fuel consumption under all conditions are determined.

2.2 Manage
As part of their EnMS, ship operators need to implement a process of identification and evaluation of areas
of significant energy use onboard their vessels, with the aim to establish performance baselines, determine
current status of performance and identify opportunities for improvement. Initially this process should focus
on energy use under the ship operator’s control and be dynamically updated, assuming business-as-usual
conditions. Stakeholders that should be brought to this process are charterers, cargo operators, Classification
Societies, subcontractors, shipyards and manufacturers of equipment used onboard.

The distribution of fuel consumption per each type of ships, onboard process, operational state, consumers
and their anticipated impact on total energy consumption should be used to determine the onboard process
of significant energy use onboard. See Figure 1.

Depending on its energy impact, each onboard process should be categorised as an onboard process of
significant energy use or not. Each onboard process of significant energy use should be drilled down to
its consumers to further determine their significance and components (equipment, systems, processes and
personnel working for, or on behalf of, the company and service contractors) and further investigated for
performance improvements opportunities.

Figure 1 summarises the output of a number of energy-based risk assessments and depicts typically significant
energy users onboard, per each onboard process and types of ship. It serves as a guide on the onboard process
that ship operators should use to improve their efficiency onboard.

98 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring

 2.2 Manage

Figure 1: An example of identification of onboard processes of significant energy use per ship type

Positioning
HandyMax

PanaMax

SuezMax
AfraMax

Dynamic
Operational Onboard

VLCC
Consumer
State Process

Hull & Propeller Efficiency x x x x x x

Voyage Voyage Planning x x x x
Execution/
Propulsion Main Engine Efficiency x x x x x x
Trim & Ballast
Sea AE Management x x x x x x
Passage Electricity
Production AE Efficiency x x x x x x
AB Management x
Steam
Production AB Efficiency x
Incineration Incinerator
Hull & Propeller Efficiency x
Propulsion
Main Engine Efficiency x

Port Electricity AE Management x

Operations Production AE Efficiency x
AB Management x x x x x
Steam
Production AB Efficiency x x x x x

Electricity AE Management x
Production AE Efficiency x
Anchorage
Steam AB Management x
Production AB Efficiency x

Figure 1 is derived from fuel-driven risk assessments, where the primary input is the distribution of fuel onboard
per operational state and onboard process and the efficiency that each consumer/onboard process is being
operated with, at the time of the assessment.

Figure 2, overleaf, provides such an example for a group of Aframax vessels. To assign a consumer as significant
or not depends on how much energy consumption increases at the end of an evaluation period compared to
baseline performance.

In this particular example, the threshold for acceptable increase of consumption per consumer per year is set
to 5%, but this can change with a more conservative or aggressive approach.

For those consumers identified as significant, their components (equipment, systems, processes and personnel
working for, or on behalf of, the ship operator and service contractors) should be further investigated for
performance improvements.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  99

Appendix 10. Energy Efficiency Management Plan - A case study of a shipping company

Figure 2: An example of identification of onboard processes of significant

energy use for aframax tankers

Such changes as trading pattern, adoption of technical measures to increase efficiency and crew training will
have an impact on performance and subsequently allocation of fuel per operational state and onboard process.
As such, the identification of the onboard processes of significant energy use for each vessel is dynamic and
needs to be repeated at least annually within the scope of the energy review of the company.

This process serves as a tool to prioritise onboard processes requiring allocation of resources. However, the final
selection of measures shall be based on corporate strategic goals and input received from all of the company’s
functions at all levels.

This is a multi-objective optimisation task. An optimal decision needs to be taken between two or more
conflicting objectives.

Each adopted measure should come with a dedicated action plan. Action plans should contain specific KPI,
target, verification method and quantification of the expected benefit in terms of fuel and emissions reduction.
It is suggested that these are made part of the company’s SEEMP.

Once the action plans and respective KPIs are adopted, every effort should be made that real time monitoring
of these KPIs is available to interested stakeholders. Input to those KPIs can be either high-resolution data from
an online propulsion performance monitoring system or low resolution data that comes from PMS and other
company’s systems. The intention is that guidance and recommendations are provided to involved personnel
ashore and onboard on how to optimise the way the equipment, system, or vessel is operated and managed,
in a proactive manner, through real-time analysis and identification of trends.

100 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 Figure 2: An example of identification of onboard processes of significant
energy use for aframax tankers

Such an analysis could include assessment of voyage performance, comparison of individual vessels’
performance with best performing sister vessels over predetermined periods and with industry sources when
available in the future.

With respect to operational/technical measures that do not require design modifications, the more significant
are:

• Optimisation of drydock strategy (hull treatment and antifouling systems selection).

• Online performance monitoring and onshore support.

• Condition-based maintenance of hull and propeller.

• Onboard optimisation of electric loads management.

• Monitoring of performance of main and auxiliary engines.

• Speed optimisation & weather routing.

• Optimising discharge operations especially for vessels with steam cargo pumps.

• Optimising heating of cargo.

• Monitor auxiliary boiler efficiency.

• Trim and ballast optimisation.

• Optimisation of hotel loads.

The management process can be considered as complete when the energy baselines of all company’s vessels are
determined and short-term predictions can be done for specific trading patterns within predefined uncertainty
limits and confidence levels.

Most of measures listed above are influencing more than one discipline at the same time within a shipping
company and as such, a clear designation of duties and responsibilities across the company should be made.

Effective communication is critical, both to inform and motivate personnel ashore and onboard. Part of the
above tasks should be achieved through campaigns and training programs.

2.3 Optimise
Optimisation includes periodic evaluation of the implementation of the planned measures and their
effectiveness. Ship operators need to deepen their understanding on the overall characteristics of the ship's
operation, such as what types of measures do/don’t function effectively, and what the resulting trends are.

Identified opportunities for improvement and lessons learnt through that process should be used as an input
to the next cycle of implementation and to the new buildings process, if it is about a design improvement,
including the onboard data acquisition infrastructure supporting the energy management analysis. As such,
feedback loops needs to be established with owners, shipyards and vendors.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  101
Appendix 10. Energy Efficiency Management Plan - A case study of a shipping company

Design
Foundations of energy efficiency must be established as a ground-up approach and as such, every effort
should be made that ships delivered are as energy efficient as possible for the intended trade, taking also into
account the total cost of each design improvement. A non-exhaustive list of design improvements follows:

• Electronically controlled engines and appropriate tuning.

• Propulsion improvement devices (ducts, fins, bulbs, etc.).

• Improved propeller designs.

• Improved rudder designs.

• Optimisation of hull forms.

• Advanced heading control systems.

• Recovery of waste heat.

• High-efficiency converters at motors.

• Frequency converters at motors.

• Composite boilers.

• Independent IG generators.

• Use of renewable sources of energy, etc.

While there are many options available at the new-building phase or as a retrofit solution, their effect is
not always cumulative and may require support of a number of different stakeholders, if they are to be
utilised most effectively. It is essential that each of the stakeholders has the incentives and flexibility to join the
improvement effort.

Concerning retrofitting solutions, the efficiency improvement depends on the trade and sailing area of the
ship, as well as on its remaining service life and the cost of fuel. As a result, it is likely that the pathway to the
most efficient combination of measures will be unique to each ship.

In particular, so-called “smart ships” should be designed with due regard to appropriate data infrastructure,
robust software integrity and cybersecurity. These ships should be able to facilitate the use of aggregated data
gathered from sensors, data communications, data treatment and data synthesis for reporting and decision-
making actions.

Such data infrastructure could be commercially critical if it supports the energy management decisions. For
this reason, data infrastructure designs should be capable of providing guarantees for constant availability
(i.e. minimum data gaps over the physical time continuum of the ship’s expected lifetime) and a high degree
of flexibility so that they can be easily reconfigured to accommodate evolving regulatory and operational
requirements on energy efficiency measures.

Fleet utilisation
Once the baseline and projected energy performance of a ship is known, choosing the right ship for right the
shipper, the charterer and trading routes would become more secure since uncertainty in the Time Charter
equivalent calculations is minimised. Note that shipping trade routes are not static as they are dependent on
oil flows.

102 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 3. Energy management team

3. Energy management team

In order to support an established EnMS, it is suggested that management nominate person(s) responsible
(full or part time) for its effective implementation with clear management authority. The size and nature of
the company, and available resources, will determine the size of the team. For larger organisations, a cross-
functional team can provide a more effective mechanism to engage different parts of the organisation in the
planning and implementation of the EnMS.

Typical responsibilities of the energy management team are as follows:

• Review the results of the energy objectives, targets, energy measures and energy performance and
propose corrective action if the targets are not achieved;

• Evaluate the programme and/or energy measures results to determine whether to:

o continue with the programme/energy measure, or

o propose corrective action if the targets are not being met;

• Discontinue the programme/energy measure;

• Evaluate regulatory, customer or supplier feedback;

• Evaluate impacts due to a change in the:

o activities, products or services provided;

o legal and other requirements;

• Proposal of changes to the system to management as necessary.

References
• IMO Study on the optimization of energy consumption as part of implementation of a Ship Energy
Efficiency Management Plan (SEEMP) – 2016

• ISO Standard 50001 on Energy Management Systems

• Resolution MEPC.282 (70) 2016 Guidelines for the Development of a Ship Energy Efficiency Management
Plan (SEEMP)

• ABS Guidance Notes on Smart Function Implementation

• Tanker Management Self-Assessment 3

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  103
Appendix 11. Information to be agreed at technical specification stage

Appendix 11. Information to be agreed at technical specification stage

The ship owner will need to make data-driven decisions during the ship’s operation to reduce fuel oil
consumption and improve its energy efficiency. The need to ensure accurate data-driving decisions should
be considered at the stage of discussing the ship’s technical specifications during the new building contract.
Section 9.1 of this Guide also suggests that this exercise should start from the building stage. This is indeed
the best timing and opportunity as modification of the ship’s technical specifications after delivery would be
hugely difficult and expensive.

This Appendix discusses a range of information that will fulfil the ship’s future need for operational energy
efficiency.

Optimisation of hull-propeller-engine for the intended operational profile.

Fuel oil system configuration to be based to the various grades of fuel intended to burn.

Owners need ship-specific information for a variety of objectives:

• Estimation of Speed-Power-Consumption Reference Values in variable conditions (draft, weather, etc.);

• Minimum Propulsion Power;

• Barred Speed Range Power Margin as % of Bollard Pull;

• Econometer (load diagram SFOC mesh);

• Hull-Propeller degradation assessment;

• Engine-Fuel degradation assessment;

• Optimum Voyage Planning;

• Other.

In this respect, the following information should be provided, based on model tests carried out with the final
wake adapted and as-built propeller, for at least 3 drafts and not less than 6 speeds, preferably 10 speeds, with
the lowest possible speeds (see note) that the towing tank can deliver accurate results:

Full model test report with towing resistance, self-propulsion and propeller open water characteristics,
including the load variation factors KSIp (ξp), KSIn (ξn), KSIv (ξv). Minimum data to include: VS, RTS,
THDF, wTS, ETAR, ETAO, ETAD, PD, RPM

Note: For speeds lower than above, present that above on the basis of CFD.

Propeller cavitation tests to be carried out at laden and ballast drafts.

The testing facility to follow up a methodology according to ITTC recommended procedures.

Additional data to be provided to Owners:

• Wind Resistance Coefficients to be based on wind tunnel testing at various drafts;

• Full Hydrostatic Tables including block coefficient, length waterline, wetted surface area, projected
transverse areas above and below waterline, waterplane area, pitch radius of gyration to length ratio
(kyy), fore and aft waterline entrance angles.

104 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 Appendix 11. Information to be agreed at technical specification stage

Information to be retrieved from vessel’s manuals, plans, drawings:

• SMCR Power & RPM;

• RPM at higher and lower barred speed range;

• Height of anemometer from keel;

• Propeller diameter;

• Propeller pitch ratio;

• Height of shaft above baseline.

Added weather resistance and power calculations is to be carried out by the builder and the following
information is to be provided to Owners for different drafts, speeds (as per model tests) and different weather
conditions, i.e. significant wave heights and peak wave periods:

• Response Amplitude Operators (RAO);

• Added wave resistance;

• Added wind resistance;

• Power and RPM in weather.

Hull roughness measurements to be provided prior to launching.

Calibration of fuel oil tanks for the as-built configuration, including any changes carried out during the
construction stage. The ullage tables to be available for various pitch and roll conditions.

Joint sampling quality of the fuel used during the main engine shop test and during sea trials.

Contractual Speed and EEDI to be derived with the same method. ISO 15016 latest version to be applied and
the input data and analysis to be provided to Owners. Measurement sensors and instruments to be calibrated
prior to sea trials and the calibration records to be provided to Owners.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  105
Appendix 12. How to manage fuel oil consumptions using flow meters

Appendix 12. How to manage fuel oil consumptions using flow meters

In Chapter 9.2, this Best Practice discusses the importance of fuel oil management. We will further examine
the following three main factors in fuel management. In addition, we will discuss two different types of flow
meters.

Quality
The importance of a proper sampling & sealing procedure while bunkering using the drip sampling method
is paramount. A company’s commercial and financial exposure when bunkering non-compliant quality fuel is
quite high. Adequate seafarer training in the company’s sampling procedures is highly recommended.

Avoiding mixing different grades of fuel is recommended. If commercial reasons dictate fuel mixing, internally
transferring fuel to an empty a tank is suggested, failing which a compatibility test before bunkering would be
a good proactive measure.

Quantity
Onboard any vessel, regardless of what the flowmeter reads, the actual fuel remaining in the tank is what really
matters. To avoid any bunker consumption misreporting, it is imperative that what is in the tank should be in
the log books and be reported in the performance management system. The vessel manager must develop an
extensive fuel oil management plan which covers procedures related to bunker quantity discrepancy during
bunkering operations.

In case of any short supply of bunker, the Bunker Delivery Note (BDN) quantity must match the actual bunkered
quantity onboard. In case the BDN quantity is higher than the actual quantity onboard, the vessel must issue a
Letter of Protest (LOP) for short supply. This LOP must be filed with the BDN and mentioned in the Oil Record
Book using Code I followed by the actual quantity bunkered in Code H. For example, if the actual quantity
received on board is 957 MT and the BDN is for 965 MT, then in the Oil Record Book:

Code I entry must be made stating “Fuel oil Quantity bunkered as per BDN 965MT, Actual quantity received
onboard 957 MT, Letter of Protest issued for short supply of 8 MT”.

Now, when the Code H entry is made, the quantity bunkered should be declared as actual quantity received
i.e. 957MT and not the BDN quantity of 965MT, because the LOP was issued, and its declaration has been
made in the Oil Record Book before the Code H entry.

In case there is any significant loss of fuel quantity due to density difference between the BDN and lab analysis
result, a declaration about that can also be made in the Oil Record Book (using Code I) and logbook. Proof
related to this (lab analysis result and BDN) must be retained to support this.

It is important to remember to never let the temperature of bunkered oil fall below the pour point while it is
stored in bunker tanks. It is not a very good idea to allow fuel to be stored in tanks without consumption for
a very long period of time. If this is unavoidable, churning this oil using the transfer pump can be considered.

Processing
When changing over to a different grade of fuel, it is always good practice to let the settling and service tank
lower down to the lowest possible safe level before transferring a new grade of fuel into the settling tank. This
will minimise intermixing of different grades.

Good purification is important for good combustion, which is in turn related to engine efficiency. As a primary
step, settling and service tanks must be drained of settled sludge and water regularly. It is advisable to use

106 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 Measure

clarifier and purifier in series for good performance. The correct size of gravity disk must be selected depending
on the density of fuel to be processed using the nomogram. It is important to maintain correct temperature of
purification and flow rate, depending on the grade of fuel being passed through the purifier.

At times it is also advisable to use purifiers in parallel for better purification. The de-sludging interval should
also be set according to the quality of the bunkered fuel. Purification guidance can also be sought from the
fuel oil test lab as they will be able to advise on such matters depending on the properties of fuel. Normally oil
test results also advise on the temperature to be maintained to get the correct viscosity for injection. It is very
important to maintain this temperature for proper injection.

A good fine-mesh backwash filter before entry of fuel into the engine is the last precaution against entry of
unwanted particles into the engine which can negatively impact the engine.

In an auto log system, which is not fitted with mass flow meters, it is possible to mis-report density of oil in use
and temperature to alter fuel consumption, hence it is very important to cross-check total consumption against
stock reduction. Ideally, this must be done at every port arrival / port departure. If the reported total consumption
using flow meter readings (or estimated consumption in case of no flowmeters) is compared against reduction
in actual stock, it is then possible to get a fair idea of the accuracy of reported total consumption. This period
can be selected between “Last Line” and “All Fast” of any two subsequent ports.

If the difference between total reported consumption and stock reduction is zero, then the reported daily
consumption is inaccurate. If the difference between the two is rather large, even then the reported daily
consumptions are inaccurate. Ideally, the sum of reported consumption should be approximately 0.5-1% less
than the stock reduction value. This difference will account for the sludge removed during the purification and
back flushing / tank draining process. The limit of 0.5-1% is not mandatory but just a guideline, the actual
difference will depend on the percentage of water in the bunkered oil, removable solids in the bunkered fuel,
frequency of purifier desludging, frequency of auto-back flushing and leaks etc.

Measure
Performance monitoring will normally require the consumed fuel mass to be measured. If the engine load is
constant, the measuring time should preferably be as long as possible, one or several hours. In this case, a
service tank level with installed thermometer can be used for measuring the consumed fuel mass. The accuracy
of using a tank will depend on the geometry of the tank. A high and slim measuring tank inserted in a corner of
a service tank is a good complement to the fuel flow meter and can be used to check the flow meter accuracy.

When the level of a service or measuring tank is not stable enough due to weather or the time is too short
for an accurate level difference, a fuel flow meter can be installed.

The arrangement of the fuel supply system and fuel flow meters are of importance for the measuring result.

Some vessels install supply and return flow meters in the inner fuel booster loop. These meters must be sized
for the flow in the inner fuel loop which is typically 3–6 times higher than the aggregated fuel consumption
of all engines in that loop.

A flow meter with a higher rating will normally have a higher inaccuracy.

If the consumption is calculated by subtracting the return flow from the supply flow, two inaccurate results are
combined. Sometimes the results would appear as if the engine was producing fuel, which is most unlikely.

When the arrangement of the fuel system and the location of the flow meter(s) have been decided it is
necessary to choose which type of flow meter to use. Traditionally volume flow meters have been installed for

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  107
Appendix 12. How to manage fuel oil consumptions using flow meters

engines and boilers. With increased focus on fuel consumption, Coriolis mass flow meters are now becoming
increasingly popular.

Flow meters
Volumetric flow meters have the disadvantage that the fuel mass must be calculated based on the temperature
at the flow meter and a known density at a specified temperature for the fuel in use. Both the inaccuracy in
the temperature measurement and the given density and the formula used for calculating the new density will
introduce errors to the final result.

The Coriolis mass flow meter(s) will present the mass flow directly on its display. Temperature and density of
the fuel at the meter are also displayed. The Coriolis flow meter does also have errors in the presented figures
but they are assumed to be more accurate than calculated figures.

To achieve completeness of the calculation of consumed fuel mass, the leak oil from the engine should be
measured. Normally the leak oil amount is fairly constant for different engine loads. For automated calculations,
a fixed representative mass can be used. Changes of oil viscosity and fuel pump wear will change the amount
of leak oil.

Below is a comparison between Volumetric and Coriolis flow meters.

Flow meter type Advantage Disadvantage

Direct measuring of mass flow; Sensitive to vibration; Mechanical wear of
No mechanical wear; moving parts;
High precision; Sensitive to pipe tension;
Coriolis mass Temperature output; Density output; Flow pulsation might influence accuracy;
High initial cost;
Should be calibrated at the intended
operation temperature;

Proven technology; Mass flow must be calculated based on

Insensitive to vibration and pipe measured temperature and assumed
tension ( within limits); density;
Volume flow Independent of flow direction; Sensitive to changing fuel density;
Measurement of pressure pulses;
Low initial cost;

It is important to regularly check the accuracy of the flow meter. This can be done by comparing service tank
fuel reduction against flow meter readings for a short duration of time (2-4 hours). If done accurately and in
calm weather and regularly, this can give an estimation of the average % deviation in the flowmeter. As in an
auto-logged system, the probability of crew not finding out about such deviations is high, hence it is advisable
to check this using reported data, and by regular accuracy cross-checking as mentioned above.

Flowmeters must also be serviced / zero checked diligently as per the maker’s advice or company PMS,
whichever is applicable.

108 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 Appendix 13. Explanatory notes on Specific Fuel Oil Consumption (SFOC)

Appendix 13. Explanatory notes on Specific Fuel Oil Consumption (SFOC)

Most commercial vessels are powered by either two-stroke or four-stroke diesel engines. Four-stroke diesel
engines are also the most common prime mover for electrical generators onboard.

At the new building stage, project engineers and designers are evaluating different engine configurations
based on weight, space, vibrations and fuel consumption, amongst others.

The estimated fuel consumption figures from design stage will later be confirmed during the engine shop trial
and the fuel consumption records will remain with the vessel as a reference value for the new engine.

The operation of a diesel engine is influenced by the ambient operating conditions and the fuel used. To
compare engine fuel consumption under various ambient conditions and with different fuels, the measured
fuel consumption must be recalculated back to the same reference conditions.

Such reference conditions are recommended in ISO 15550:2002 and ISO 3046-1:2010 as below:
Ambient air temperature [tr]: 25 °C
(Tr = 298K)
Water inlet temp. to scavenging air cooler [tcr]: 25 °C (Tcr = 298K)
Ambient air pressure [pr]: 100 kPa
Ambient air relative humidity [Θr]: 30%
Reference fuel Lower Heating Value [LHV]: 42,700 kJ/kg

ISO 3046-1:2002 is presenting formulas for recalculation of power and fuel consumption back to the reference
for four-stroke diesel engines.

For two-stroke slow speed, rpm<130, turbo charged diesel engines with charge air cooler, ISO 3046-1:2002 is
referring to formulas for recalculation issued by the engine builder / designer. One expression for recalculation
of two-stroke slow speed diesel engines is presented below:

Specific Fuel Oil Consumption (SFOC):

In the above expressions, guidance values for the influence of any deviation of ambient conditions are given
as follows:
Ambient air temperature [tr]: increase 1°C → increased SFOC 0.02%
Water temp to scavenging air cooler [tcr]: increase 1°C → increased SFOC 0.06%
Ambient air pressure [pr]: increase 1kPa → reduced SFOC 0.02%
Fuel Lower Heating Value [LHV]: increase 1% → increased SFOC 1%

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  109
Appendix 13. Explanatory notes on Specific Fuel Oil Consumption (SFOC)

To evaluate the performance of a diesel engine, whether driving a generator or a propeller, the specific fuel
oil consumption (SFOC) should be calculated using the developed power. For an engine driving an electrical
generator, the power is taken from the instruments on the Main Switchboard (MSB), taking into account the
generator efficiency.

For an engine driving a propeller, the influence of exterior factors like weather, hull or propeller fouling etc.
should be separated from factors originating in the engine room. This way, the power developed by the engine
must be measured.

As can be seen in the equation above, both power and fuel LHV are directly influencing the SFOC calculation
and both are under normal conditions associated with uncertainties. Further, the calculation of the developed
brake/shaft power is based on knowledge of the intermediate or propeller shaft material properties under
twist, namely the torsional module G. Power is measured from torque which is proportional to G. The lack
of knowledge of the shaft material torsional properties will act as a permanent bias/offset to the power/
consumption curve.

The Lower Heating Value (LHV) of the fuel in use presents an uncertainty due to the practise to use the LHV
from the Bunker Delivery Note (BDN) or from a fuel analysis report issued by an independent laboratory. Either
way, LHV figures are normally calculated or estimated with LHV relying on density and sulphur content of the
fuel sample. The LHV can be measured in a laboratory by a bomb calorimeter but this is seldom done as this is
a separate test with additional cost.

Sample test records comparing the calculated/estimated LHV with actual measured LHV are hard to find.
Some results show good correspondence between the two values whereas other test sequences indicate
considerable discrepancies. This variation is assumed to be random and will influence the SFOC after every
bunkering and use of new fuel. A vessel engaged in worldwide operation will bunker at different ports and
from different sources. A vessel bunkering at the same port from the same supplier is likely to experience a
smaller variation in calculated SFOC due to the estimation of LHV.

Below is the result from 10 samples of high sulphur HFO 380 cSt:

Average Measured LHV, μ = 40,1 MJ/kg σ = 0.4 MJ/kg k=2 → ± 2% (400 kJ/kg/100/k)
Average Calculated LHV, μ = 40,4 MJ/kg σ = 0.2 MJ/kg k= 2 → ± 1% (200 kJ/kg/100/k)
The error in estimated/calculated LHV for Low Sulphur fuel and Bio fuels are not known.

An estimate of the uncertainty, u, in the SFOC can be calculated for a fictive vessel, assuming that all variables
are independent and normal distributed. The following definitions are provided for clarification:

Y: The measurand, which may be determined by direct measurements or indirectly by measurements of other
parameters and subsequent calculations. Y=F (X1, X2, … Xi) where Xi input parameter.

xi: The estimate of input parameter Xi which may be the result of a measurement, specification or calculation.

U: The expanded measurement uncertainty U of the output estimate y defines an interval Y=y±U having
the sought level of confidence, equal to the expansion factor k times the combined standard uncertainty uc(y)
of y. U=k uc(y)

uc(y): Combined standard uncertainty of output estimate y indicates the measurement uncertainty uc(y) of
the measurand Y taking into account the measurement uncertainties of the input parameters xi.

110 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 Appendix 13. Explanatory notes on Specific Fuel Oil Consumption (SFOC)

The combined standard uncertainty is given by:

where the derivative represents the sensitivity coefficient and u(xi) is the component of combined standard
uncertainty uc(y) of output estimate y generated by the standard uncertainty of input estimate xi.

k: The expansion or coverage factor k is used to calculate the expanded uncertainty U of output estimate y
from its combined standard uncertainty uc(y) in order to achieve the sought level of confidence. For a normal
distribution and 95% confidence level k=2. If k=1 confidence level is 68% and if k=3 it is 99.7. It is thus
equivalent to the standard deviation of the input parameter Xi.

In the calculation table Y→sfoc, input parameters X1, X2, X3, X4, X5, X6→ FLM, Powmet, LHV, tr, tcr, pr and
u→uc(y). The sensitivity coefficients have been incorporated in the last rate column.

Source Estimation error rate

Flow meter FLM ISO max 2% ; k=2 0.01
Power Powmet ISO max 3% ; k=2 0.015
LHV BDN: 40 100 kJ/kg 400 kJ/kg 400/40 100= 0.01
tr ISO max 2 degrees; k=2 0.0002
tcr ISO max 2 degrees; k=2 0.0006
pr ISO max 1 kPa 0.0002

For a confidence level of 95% (k=2) the expanded uncertainty U = 4 %;

Thus a measurement of Specific Fuel Oil Consumption on this fictive vessel in operation with assumed
uncertainties will most likely vary randomly within a band of +/-4%. In addition to the above, the power/
consumption curve will be shifted in a band of approximately ± 2.5 % due to the unknown exact value of the
torsional module, G. Based on the results above, two remarks could be made:

• I n order to keep the uncertainty band as small as possible, maintenance and record-keeping of all
sensors and their calibration as well as cross-checking of developed power to determine an improved
estimation of the G module is essential.

• F uel saving measures due to the uncertainty band of the power/consumption curve will be virtually
impossible to confirm.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  111
Appendix 14. Hull-propeller performance assessment methods

Appendix 14. Hull-propeller performance assessment methods

Methods for assessing hull-propeller degradation, whilst being compatible with the principles of ISO 19030,
are:

• Load Diagram method;

• Resistance Decomposition method.

Both methods can consider external environmental factors, such as fouling, current, wind waves and swell
waves.

For reference:

a) Wind waves are the direct result of wind input and are considered young waves. Wind waves have short
wave periods and multiple wave heights. The direction of wind waves generally follows the direction of
the wind.

b) Swell is the result of a storm or active wind area at a large distance from the observer. Characteristics of
swell are long wave periods and constant wave height. The direction of the swell wave does not need
to agree with the wind direction.

The Load Diagram method

This method is a useful rough way for operators to assess vessel performance in terms of “Delta” on Brake
Power due to an overall increase of hull-propeller roughness. This is achieved by:

• collection of Power-RPM-Speed-Weather high-frequency measurements for specific draft(s), i.e. large

data sets with difference in months and between cleaning periods;

• filter in open seas only (i.e. no manoeuvring canal transit, idle, etc.);

• filter in a specified range of true wind speed and angle (for the calculation of true wind speed see ISO
15016:2015);

• for the more advanced users, filter out acceleration events.

Subsequently, apply a “power” curve through the remaining scatter. Then compare the resulting power curves
for a specific draft at different periods and derive the “Delta” on Brake Power.

In this method, the external factors such as weather, current and fouling are all accounted for in the power-
rpm shift inside the load diagram.

If a company could measure the wave height and period concurrently with power/rpm/speed at the same high
frequency and subsequently apply post-processing to identify the segments associated with long periods and
almost constant wave height, then the swell waves could be filtered out.

112 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 The Resistance Decomposition method

The Resistance Decomposition method

In this method, the external factors are decomposed in terms of their individual resistance.

A challenging resistance component will always be the resistance due to waves. Added resistance due to
waves is induced by (a) reflected waves as a result of incoming waves reflecting against the ships' hull and (b)
radiated waves as a result of ship motions:

a) Short waves do not influence the motion of the vessel, as the wave height of short waves is limited.
Wind waves are generally short waves and give the largest contribution to wave reflection.

b) Long waves or swell waves provide the largest contribution to wave radiation. For wave radiation
calculations it is important to consider the loading condition as this determines the motion response.

The Resistance Decomposition method has been invariably taught at universities, whilst in the industry ISO
15016 remains as a tool.

The challenges for operators are plenty. For instance, are operators in a position to measure (a) and (b) separately
and in terms of “significant wave height”, “period” and “direction”? And at the same high frequency are
power/rpm/speed recorded? Even on the theoretical front there are different schools: those that will estimate
the added wave resistance of each (a) and (b) from each of their significant wave height, and those that
will first calculate a combined significant wave height and then derive the added wave resistance. When
attempting to include the different directions from bow of each (a) and (b), things get even more complicated.

Some companies that are using noon reports for vessel performance purposes, ask for a daily average record
of (a) and (b) separately and in terms of observed wave height and direction from bow. Such data is useful to
have but cannot be consistently integrated with high frequency measurements (ISO 19030).

Owners need to keep in mind that a useful Resistance Decomposition requires information to be provided at
the time of the ship’s delivery so that the ship-specific mathematical model is developed and calibrated after
a few voyages. Such information should be systematically agreed between Owners and Builders during the
technical specification stage of a new construction project, to be made available prior to the ship’s maiden
voyage. Otherwise, such a method would be difficult to implement consistently for the company’s fleet. The
minimum required information is also available in an xls format upon request1.

1
Contact the INTERTANKO Secretariat.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  113
App 15. A case study into decision-making processes about propeller polishing and hull fouling

Appendix 15. A case study into decision-making processes about propeller polishing and
hull fouling
Biofouling is considered to be a large contributor to efficiency loss during service, thus the monitoring and
control is of paramount importance.

Biofouling can contribute to the transfer of invasive aquatic species, which may pose threats to human, animal
and plant life, economic and cultural activities, and the aquatic environment. As such, an increasing number
of IMO member states are stepping up measures to regulate underwater hull treatment either by imposing
recollection of removed biofouling material or totally banning such operations. This regulatory pressure might
create operational restrictions and additional costs. If a hull underwater treatment goes wrong, damage to
antifouling systems or exposing the bare metal can have a significant impact on the performance of a vessel,
resulting in loss of time and increased costs for the company.

Given the above, it is imperative that any action towards hull and propeller cleaning should be carefully planned
and supported by performance analysis. This task involves understanding on how each uncertainty discussed
in the previous sections affects the end results, thus increasing the complexity of answers to the following:

1.  Is the vessel fouled or not?

2.  How much fuel oil is consumed per knot with its present fouling?

3.  How are the ship’s speed and consumption tables affected?

4.  Optimal speed and consumption tables for vessel market placement.

5.  Decision-making to polish the propeller and/ or clean the hull (include ROI and route scenario analysis).

6.  Paint deterioration and re-fouling rate after hull cleaning.

7.  Optimal decision for paint in dry-dock.

This section discusses a realistic approach to the points above.

It will clarify what the expected benefits of propeller polishing are, and using a practical example, how the
results can drive the decision towards a full hull cleaning on a VLCC trade route. Finally, it will discuss what is
useful to analyse after a full cleaning and underline important aspects that should be considered for selection
of a paint scheme.

There are two main methodologies available in the market to derive performance results as discussed in
Appendix 14. They are marketed as digital twin ships, machine learning data analytics or simply the ISO19030
approach to performance.

There is a third approach which combines these two main methodologies. This method utilises models, with a
large amount of data, not only for normalisation but to extend the applicability of methods like ISO19030 off
bounds, and train the regression curves not following a neural network approach. Models are again used for
further calculations which will be demonstrated while providing answers to the above question set.

Question No 1: Is the vessel fouled or not?

This can be answered using a statistical approach and monitoring the trend of power increase versus time or
by identifying the total resistance and by reducing the weather effect using machine learning techniques. The
following two methods have been identified in the market.

114 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 Question No 1: Is the vessel fouled or not?

Method 1:
The first method is a simple regression of data that has been adjusted using mathematical models in order to
bring all values to a near “calm condition”.

How reliable this procedure needs to be is considered as follows:

1. In addition to the uncertainty contributing factors of Appendix 14 in terms of waves, there is no wave
correction in degrees more than 45 from the bow (ISO 15016 STA methods) even if their effect is
evaluated as minimal.

2. Faltsinen and Mauro methods require a lot of data and are not simple formulae to program. This
poses a significant practical challenge.

3. Mean models e.g. Kwon or Aertssen methods are based on a sample of hull forms and cannot
properly describe a particular hull form under investigation.

4. Models using machine learning are very prone to individual programmer’s perception about naval
architecture and connection with each variable, rendering their application questionable to a large
extent.

Thus, any “correction” of data points to the equivalent calm condition has induced error which is dependent
on the selected method. However, it is globally assumed that the error is constant when applied to the same
vessel over time. Nevertheless, this assumption can be rejected because:

• In each draft, the ship shows different behaviour.

• Each model does not behave properly in all directions of wind and waves.

• Machine learning relies on the training periods and amount of train data (a practical question is how
much of the data is used and in which conditions?).

In order to continue our discussion, this section assumes that the error per vessel and per model is the same for
all conditions, otherwise any further discussion is not possible as the accuracy issues become chaotic (accuracy
reduction point A).

Using statistical data, a difference in terms of power (or speed) from a baseline is established. The baseline can
be defined either by analysis of after dry-dock performance, or by setting up a custom benchmark (e.g. model
tests, sea trials CFD, machine learning baseline etc.). The proper selection of methodology is vital in the implied
uncertainty of reference line (accuracy reduction point C).

Direct comparison of reference power, and power as measured by torque meter during the performance
evaluation period, eliminates the uncertainty compared to the estimation of added resistance and then
transformation to power. In further detail, the transformation of resistance to power is made using the
methodologies described in ISO 19030 (methodology A) using an average transmission efficiency factor ηT , or
ISO 15016 (methodology B) using propeller non dimensional polynomials (KT and KQ).

Methodology B is far more accurate than A, however, requires significant calculations per measuring point,
rendering the process more complex (accuracy reduction point B).

Assuming that the models have been correctly used, programmed and results are obtained in a relatively rapid
manner, a linear trend line is established which is then compared to the established reference line.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  115
App 15. A case study into decision-making processes about propeller polishing and hull fouling

The difference of reference line and the in-service performance line (produced by linear regression of scatter
dots) gives the hull efficiency drop, measured as speed loss percentage or power increase percentage overtime
(Figure 1).

Figure 1: ISO19030 Speed loss % before and after the reference

period (end April 2019), a modern VLCC

Method 2:
Another method to determine fouling increase at any sea state (answer to question 1) is to use a “ghost ship”
model that follows the ship in real conditions. The advantages of this approach are:

1) Use of many more samples (constraints imposed only by the applicability of used models).

2) Direct estimation of added power due to fouling as this is a pure comparison of measured Power and
Ideal Power at current speed.

The disadvantages of this method are:

1) Unreliability of models in every ship condition.

2) Wrongly calibrated models can hide the effect of fouling on the effect of added resistance.

3) No knowledge about the ship’s hydrodynamic behaviour in slow speeds (however, this can be partially
solved by CFD methods and test in sea trials (“only as reference” yard statement) .

To counteract the disadvantages of the method, the following should be considered:

1) Performance analysis should be made in low sea states where the effect of added resistance models is
minimised.

2) Slow speed behaviour can be solved with CFD or test in sea trials at new building stage (even if the
results are stated “only as reference”).

3) After dry-dock, trials for the owner at ballast draft condition following the ISO 15016 approach and
extrapolating the results to other draft conditions.

116 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 Question No 2: How much is the effect on ship performance?

Question No 2: How much is the effect on ship performance?

The “digital twin” method directly estimates the added power on a specific speed range.

Nonetheless, for a statistical method that averages the added power overtime, further calculations should be
made, and more complexities should tackled.

The problem relies on the determination of an average power increase over time which does not involve the
observation of various corresponding speeds, which may lead to overestimation of power in lower speed
points.

Due to the scattering nature of added power data, the linear regression can be considered valid (already
considered accuracy reduction point B and C). Nevertheless, it has been observed in variations of the ISO19030
methodology, that for data reduction purposes only one point in every 24 h time block can used, but this
introduces new uncertainty as the average is not weighted for the duration of each validated block (as per ISO
19030 methodology) (accuracy reduction point D).

Assuming again that the methodology is agreed and accuracy reductions have been taken into consideration,
the following two approaches to determine the new speed power curve, thus estimating the ΔP and eventually
the ΔSFOC and ΔFOC, have been encountered:

A. Shifting the model test or benchmark power curve to fit the scatter data, in the same manner as is for
sea trial.

B. Calculate the average ΔP increase and then multiply the model test or benchmark power curve.

The above approaches A and B are depicted on the below figure. It is evident that data scatter around specific
speed points introduce uncertainty at the lower and upper speed range.

Figure 2: Regression of speed-power curves of data points following Approach A and B

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  117
App 15. A case study into decision-making processes about propeller polishing and hull fouling

Hence, the following complexities are summarised:

1) Both approaches can overestimate the effect of friction resistance increase at slow speeds, hence
power increase, due to ship inertia and absence of viscous and flow separation phenomena there is
no ΔP.

2) Very high projection of added power on higher speeds (high absolute values multiplied by constant
percentage).

3) Approach A overestimates most probably the effect of added resistance at low speeds, whilst it tends
to match with Approach B at higher speed ranges.

It is recommended that in the absence of data at high and low points, the above approach is followed at a speed
range equal to ISO 19030 limitations and as least programming effort following Approach A is suggested.

Using the method of “Digital Twin”, the simulation can give a precise value of added power in each point,
reducing the overestimation effect.

Question No 3: How are the ship’s speed and consumption tables affected?
As soon as the added power per each knot is determined, the construction of speed and consumption tables
is a straightforward procedure. However, to proceed from ΔP to ΔFOC, the SFOC of Main Engine(s) should be
inserted per each load point.

Assuming that a two-stroke marine diesel engine has a constant SFOC curve near the propeller curve (shop test
curve) and the effects of ΔRPM per each load do not alter the SFOC, FOC can be determined by the following
equation:

For a very accurate approximation, there is a ΔSFOC for any power-RPM curve that differs from the one of
shop trial (or heavy propeller curve). This can be considered as accuracy reduction point E, however its effect
can be considered small, shown in Figure 3. However, this is not measurable in a non-laboratory environment.

Figure 3: Theoretical effect of ΔRPM on Electronically Controller Two-Stroke Propulsion Engine SFOC

The speed and consumption table can now be modified by including the ΔFOC value.

118 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 Question No 4: Optimal speed and consumption tables for vessel market placement

Question No 4: Optimal speed and consumption tables for vessel market placement
The previously mentioned accuracy reduction points need to be considered. ΔFOC has a range of potential
correct solutions as the following have been affected:

1) ΔP (by construction of a new curve or the applied model or the behaviour of the ship in various other
speed points than the analysed ones.

2) SFOC as is ΔP dependent as SFOC is load depended. Refer to Figure 3 assuming a constant propeller
margin SFOC – Load.

3) SFOC as is power-RPM dependent due to engine thermal efficiency change as shown in Figure 3
(reduction of propeller margin).

As a result, an error bar per each speed point may be constructed. This error bar has a practical meaning
and can be decisive in vessel chartering. Depending on the performance team and the observations of past
performance, model accuracy etc. this error bar can be:

• Low, yielding to a known vessel performance and correct chartering placement.

• Medium, yielding to some cases of speed or warranties underperformance.

• High, yielding to severe doubts of vessel performance description, thus associated claims.

In all cases, the Chartering department must know the limits of the ship and decide how they plan to place
the vessel on the market. Of course, the speed & consumption curves based on the low error bar figure gives
correct descriptions and relies on the strategy of Chartering, the rest (medium and high error bars) though
imply a different approach in vessel market placement, dictating the strategy as there are many unknown
factors that need to be taken into consideration. Therefore, this strategy can be aggressive (high risk as
performance may be lower than expected) or very defensive (to minimise the risk of underperformance claims).

To tackle the above problem, it is suggested getting feedback for every voyage from every vessel and comparing
the actual achieved speed & consumption points to the ones that were used to place the vessel to the market.
This approach will allow a review of the methodology and the selected models (if applied) in order to reduce
the accuracy reduction points.

Question No 5: Decision-making to polish the propeller and clean the hull (include ROI and route
scenario analysis)
Before decision-making, it is important to estimate the contribution of hull fouling and of propeller blade
fouling over the total measured added power. In the literature, the split of propeller and hull effect is described,
however the practical implementation of measurements in full scale is still questionable.

Consequently, the maximum contribution of propeller fouling has been quantified by various studies as
described below.

The loss of efficiency at the propeller depends on nature and extent of deposits. The benefits resulting from
the polishing of a propeller and restoring its condition to Rupert A scale condition can typically range from a
1-6% power increase, being a conservative estimate.

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  119
App 15. A case study into decision-making processes about propeller polishing and hull fouling

Rupert Ra* Rt** Power

Scale (Micron) (Micron) Increase
A 0.65 5 0%
B 1.92 12 0.5%
C 4.7 32 1.5%
D 8.24 51 3.0%
E 16.6 97 5.0%
F 29.9 154 6.0%

*Ra= centreline average (This is the average

deviation of the profile about the mean line)

**Rt= Maximum peak to valley amplitude

Townsin reported (1983, 1985) on the benefits of propeller polishing as did 0. Fagerjord, Det Norske Veritas,
IPaper Series No 80 P022 October 1980) and TE. Svensen, (l.Mar.E. 1982) and others.

Figure 3: Effect of Blade Roughness to Added Propulsion Power by various studies

The above curves are reasonable estimates. However, the actual effect of roughness will depend on ship speed,
RPM, blade area and other ship and propeller characteristics as well as the peakiness (kurtosis coefficient) and
texture of the roughness measured or assessed as being of a particular Ra value.

Polishing of a propeller should be a scheduled task ideally with a periodicity of 4-12 months and conducted
by experienced and specialised personnel. Whenever a propeller-polishing operation takes place, the scope of
works should include a thorough hull inspection about the extent of biofouling, painting condition (existence
of blisters, peeling off or any other sign of deterioration) and hull structural condition including appendages
and Propulsion Improvement Devices, if fitted.

This periodicity ensures that the effect of propeller fouling is kept minimised, while with the visual observation
of hull, the association of performance methodology results with actual hull condition are verified, enabling
much better accuracy in the calculations.

The maximum contribution of propeller fouling can be considered known and is verified also by in-service data
including the time when the hull was free of fouling. Thus, it can be accepted that the maximum added power
contribution due to Rupert scale E propeller blade condition can reach 9%.

120 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 Full hull cleaning – Decision-making and ROI

Full hull cleaning – Decision-making and ROI

The following example concerns a VLCC that is identified as fouled (hull and propeller), as the total added
power is estimated from 7 - 13%. This range greatly affects the decision-making and ROI.

The VLCC is sailing laden from US gulf to China. After China, the vessel is fully fixed to load cargo in Fujairah.
The Performance department is asked whether it is beneficial to:

1) stop the vessel in laden in Singapore (off-hire costs involved) prior to arrival at China for full cleaning
or partial cleaning, or

2) undertake the cleaning in Singapore in ballast after departure from China where other activities are
planned too, or

3) deviate to Fujairah and perform the cleaning prior to arrival at Arabian gulf for loading as the cost and
time windows are advantageous to the vessel at Fujairah.

Example variables (numbers, days and costs mentioned in this section are for illustration purpose only):

For a VLCC, the route from the US Gulf up to China can be 35 days sea passage on 12.5 knots average
GPS speed, covering approx. 10500 nautical miles. The distance from Singapore to Chinese ports that can
accommodate a VLCC for berthing is on average 2,100 nautical miles or equal to 7 days’ sea passage. This
example will use the above distance figures and the involved fees as per Table 1.

Table 1: Estimated Fees

Reference ID Call A Call B Call C

Stoppage at Stoppage at Stoppage at

Action
Singapore (laden) Singapore (ballast) Fujairah (ballast)

Duration of Stay [h] 24 24 18

Cost of Off-hire (24h) [USD] 60,000 0 0

Cost of full hull cleaning and propeller

polishing (incl. mob. / de-mob costs) 26,500 26,500 18,500
[USD]

Deep Draft surcharge diving Fees [USD] 4,000 0 0

Agent Fees [USD] 1,500 1,500 1,800

Vessel Speed [knots] 12.5

Table 2, overleaf, has three scenarios of Added Power labelled as 107%, 110% and 113% respectively. Each
of the scenarios has uncertainty in the amount of power recovery after full cleaning (stated as high recovery
meaning a slight percentage of added power remains, medium recovery and low recovery which is a marginal
improvement action).

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  121
 App 15. A case study into decision-making processes about propeller polishing and hull fouling

Table 2: ROI calculations for stoppage at Singapore (call A) in laden voyage

Power
Added power Daily
difference Daily FOC FOC penalty ROI days
Scenario percentage savings in
after savings remaining after Call A
after cleaning USD
cleaning

Upper ΔP bound (113%) 13% 7.1

Recovery percentage L 8% 5% 2.7 4.4 1,090.1 80.7

Recovery percentage M 5% 8% 4.4 2.7 1,745.9 50.4

Recovery percentage H 3% 10% 5.5 1.6 2,183.9 40.3

Average ΔP bound (110%) 10% 5.5

Recovery percentage L 5% 5% 2.7 2.7 1,092.2 80.6

Recovery percentage M 3% 7% 3.8 1.6 1,530.2 57.5

Recovery percentage H 1% 9% 4.9 0.5 1,968.8 44.7

Lowest ΔP bound (107%) 7% 3.8

Recovery percentage L 3% 4% 2.2 1.6 875.3 100.5

Recovery percentage M 1% 6% 3.3 0.5 1,314.0 67.0

Recovery percentage H 0% 7% 3.8 0.0 1,533.6 57.4

In this case, the Duration of the round-trip between Singapore and China is about 13 days at 12.5 knots.
According to Table 2, the ROI varies from 26 days up to 100 days depending on the fouling scenario, quality
of service, off hire cost and diving operation cost. For all calculations below, let us define the “equivalent cost”
as the cost of unsaved fuel due to postponement of hull cleaning activity along with the diving related costs
(incl. agency appointment). The hull cleaning ROI is calculated using the equivalent cost, while the breakeven
day is defined as shown in Figure 4:

Figure 4: Decision making based on ROI analysis for 3 stoppage scenarios

To determine whether it is optimal to undertake the cleaning, fuel penalty (if no action is taken in call A) is
estimated.

122 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 Full hull cleaning – Decision-making and ROI

Table 3: ROI calculations and equivalent cost estimation for stoppage in Singapore (Call B) in ballast)

Singapore to Theoretical
HFO Equivalent
China China fuel Savings Break even
cost (for cost (fuel ROI days
Scenario to Singapore if cleaning day after
remaining penalty + after Call B
(fuel penalty performed Call B
excess) Table 1 fees)
TOTAL) in Call A
Upper ΔP bound (113%) 92.4 36948

Recovery percentage L 56.9 22776 14172 42172 39 68

Recovery percentage M 35.6 14251 22697 50697 29 37

Recovery percentage H 21.4 8558 28390 56390 26 27

Average ΔP bound (110%) 71.1 28450

Recovery percentage L 35.6 14251 14198 42198 39 68

Recovery percentage M 21.4 8558 19892 47892 31 45

Recovery percentage H 7.1 2855 25595 53595 27 32

Lowest ΔP bound (107%) 49.8 19937

Recovery percentage L 21.4 8558 11379 39379 45 88

Recovery percentage M 7.1 2855 17082 45082 34 54

Recovery percentage H 0.0 0 19937 47937 31 44

It appears that taking into consideration the equivalent fuel cost without hull cleaning in Call A, the breakeven ROI is higher than
the ROI after cleaning in Call B. This means that the additional fuel penalty needed to undertake hull cleaning at Call B is not
sufficient to cover the off-hire cost and additional diving operation fees due to deep draft.

In case the vessel goes to the Arabian Gulf calling at Fujairah (Call C), Table 3 is transformed to Table 4 (overleaf):

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  123
 App 15. A case study into decision-making processes about propeller polishing and hull fouling

Table 4: ROI calculations and equivalent cost estimation for stoppage at Fujairah (call C) in ballast

Theoretical Equivalent Break Break

HFO ROI
China to fuel savings cost (fuel even day even day
cost (for days
Scenario AG via if cleaning penalty Call after Call after Call
remaining after
Singapore performed A and Call B C (cleaning C (cleaning
excess) Call C
in Call B + Table Fees) Call B) Call A)
Upper ΔP bound (113%) 135.0 5,4001

Recovery percentage L 16.4 6,540.9 47,460 81,932 75 32 61

Recovery percentage M 26.2 10,475.4 43,526 86,522 50 22 30

Recovery percentage H 32.8 13,103.2 40,898 89,588 41 19 20

Average ΔP bound (110%) 32.8 13,130.7

Recovery percentage L 16.4 6,553.1 6,578 41,076 38 32 61

Recovery percentage M 23.0 9,180.9 3,950 44,142 29 24 38

Recovery percentage H 29.5 11,813.0 1,318 47,213 24 20 25

Lowest ΔP bound (107%) 23.0 9,201.5

Recovery percentage L 13.1 5,251.7 3,950 35,629 41 38 81

Recovery percentage M 19.7 7,883.8 1,318 38,699 29 27 47

Recovery percentage H 23.0 9,201.5 0 40,237 26 24 37

From Figure 4, it appears that if no action is taken in Call B, the vessel fuel penalty incurred before reaching
Fujairah exceeds the loss of hire, only if the hull was excessively fouled. The other cases render a stoppage
in Call A more costly than the actual benefit of the ship, hence the potential full hull cleaning should be
postponed until the ship reaches the next convenient location.

The results of the examples are summarised below:

• In the scenario of high fouling (113%), stoppage in Singapore in ballast is recommended.

• Especially in the case of excellent underwater service where most of the added power effect is
reduced (recovery percentage H), stoppage in laden voyage can be considered.

• In the scenario of medium fouling (110%), stoppage in Singapore in ballast is recommended.

• In the scenario of low fouling (107%), stoppage in Singapore in ballast together with a marginal
potential for stoppage in Fujairah is recommended?

• Assuming that a partial cleaning is performed in Singapore, the costs are lower and off-hire time
smaller. However, based on the recovery percentage scenarios, this option is considered not beneficial
in terms of ROI.

To conclude, it was demonstrated that the optimal decision is greatly affected by the proper estimation (as
affected by reduction of methodology accuracy) of the hull and propeller fouling. In addition, the knowledge
of power recovery after cleaning influences the decision-making process, too.

124 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
 Question No 7: Optimal decision for paint in dry-dock

Question No 6: Paint deterioration and re-fouling rate after hull cleaning

Regardless of the adopted methodologies about added power or speed loss, the slope of each deterioration
curve after every underwater activity should be calculated and projected over time. In addition, the time
duration between the cleaning event and the date point where the same added power or speed loss prior to
the event is observed, should be recorded, and transformed to days up to the same deterioration point.

Figure 4: Added Power effect between underwater activities interval and refouling rate regression

Adopting this methodology, primarily, the actual benefit of cleaning is determined, but more importantly, the
paint deterioration rate is evaluated. In other words, assuming that the hull was cleaned properly with the
correct brushes or other methodologies, the paint quality and the applied layer thickness can be evaluated as
it gives us the fouling free period after every activity. The higher the duration is, the better the paint serves its
purpose.

Question No 7: Optimal decision for paint in dry-dock

After evaluating vessel performance throughout the five-year period between two dockings, assuming that
the data is reliable and the selected methodologies functional, the paint deterioration model and the re-
fouling rates estimated, the operator is ready to select the optimum painting scheme for the next five-year
period.

The evaluation of paint consists of comparison of end performance and commence of performance between
two dry-docks as per Figure 5 (Paint Benefit quantification).

INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring  125
App 15. A case study into decision-making processes about propeller polishing and hull fouling

Figure 5: Evaluation of painting scheme between consecutive Drydocks (five-year period)

Regardless of this procedure, the slop of paint deterioration should be evaluated. As per the example
previously presented, the added power effect significantly affects vessel placement in the market, thus
freight income.

Finally, what is also recommended for evaluation is the comparison of refouling rates (marked in thick black
lines in Figure 5). The slope comparison can show the diving operations quality, or antifouling deterioration
rate due to the applied paints. Thus, the refouling rate comparison is very delicate in choosing a final paint
product. However, the slope comparison is critical and prone to errors as it is affected by many parameters
and a significant amount of data related to underwater activities, hull condition, cleaning technology and
estimation of areas left with no antifouling after a cleaning event, should be considered and correlated.

References:
G. Aertssen. Service Performance and Trials at Sea. In 12th International Towing Tank Conference,1969.

R. L. Townsin and YJ Kwon. Approximate formulae for the speed loss due to added resistance in wind and
waves. Tran. RINA, 125:199–207, 1983.

Y. J. Kwon. Speed loss due to added resistance in wind and waves. The Naval Architect, March: 14–16,
2008.

Faltinsen, O.M., Minsaas, K.J., Liapis, N., Skjørdal, S.O., 1980. Prediction of resistance and propulsion of a
ship in a seaway. In: Proceedings of the 13th Symposium on Naval Hydrodynamics, Tokyo, pp. 505–529

Maruo, H., 1957. The excess resistance of a ship in rough seas. International Shipbuilding Progress 4 (35)

Maruo, H., 1960. The drift of a body floating on waves. Journal of Ship Research 4 (3), 1–10

Maruo, H., 1963. Resistance in waves, 60th anniversary Series. The Society of Naval Architects of Japan 8,
67–102

126 INTERTANKO Guide on the Best Practice for Tanker Performance Monitoring
INTERTANKO London
St Clare House
30-33 Minories
London EC3N 1DD
United Kingdom
Tel: +44 20 7977 7010
Fax:+44 20 7977 7011
[email protected]

INTERTANKO Oslo
Nedre Vollgate 4
5th floor
PO Box 761 Sentrum
N-0106 Oslo
Norway
Tel: +47 22 12 26 40
[email protected]

INTERTANKO Asia
70 Shenton Way
#20-04 Eon Shenton
079118
Singapore                                                          
Tel: +65 6333 4007
Fax: +65 6333 5004
[email protected]

INTERTANKO North America

801 North Quincy Street – Suite 500
Arlington, VA 22203
USA
Tel: +1 703 373 2269
Fax:+1 703 841 0389
[email protected]

INTERTANKO Athens
Karagiorgi Servias 2
Syntagma
Athens 10 562
Greece        
Tel: +30 210 373 1772/1775
[email protected]

www.intertanko.com