Chapter 9

Liquid Pipeline Operation

The operations directly related to hydraulics are covered in Section 5.1 and batch op-
erations in Section 5.2. This chapter discusses various techniques and tools necessary
to improve safety and efficiency in operations; SCADA, leak detection, DRA injection,
tank farm operation and volume measurement, and power cost control.

9.1 SUPERVISORY CONTROL AND DATA

ACQUISITION (SCADA)

9.1.1 Introduction
Pipeline systems are automated to provide the capabilities of operating pipeline sys-
tems reliably, efficiently and thus economically. Pipeline operation involves moni-
toring and controlling of a pipeline system, and monitoring is required for checking
the pipeline states and controlling facilities such as pump and valve stations. Modern
pipeline system operation is centralized because a centralized operation of the pipe-
line systems benefits the stakeholders including the pipeline company, producer of the
product, and the shipper of the product. A centralized system provides the capability
to monitor and control the complete pipeline system in a safe and efficient manner. It
allows the stakeholders to meet the changing demands for the product being shipped
expediently and to move the product from source to market safely and quickly in the
most economical way possible. A SCADA system provides the pipeline companies
with centralized monitoring and controlling capabilities [1].
SCADA is an acronym for Supervisory Control and Data Acquisition; supervisory
because human operators always issue control commands, not providing a closed-loop
control function. A SCADA system is a computer-based data acquisition system de-
signed to gather operating data from an array of geographically remote field loca-
tions, and to transmit this data via communication links to one or more control center
location(s) for monitoring, controlling, and reporting.
A SCADA system is designed to assist pipeline operators in the operation of the
pipeline system using real-time and historical information. Pipeline operators typically
regulate pipeline pressure and flow, start and stop pumps at stations, and monitor the
status of pumps and valves through the SCADA system. Local equipment control sys-
tems monitor and control the detailed process for the pump and its associated driver.
They may then issue commands of a supervisory nature to the remote locations in
response to the incoming data. Additionally, software programs implemented within
the SCADA host can provide for specific responses to changes in field conditions, by
reporting such changes or automatically sending commands to remote field locations.
Pipeline system control is accomplished by setting a controlling variable at the
desired level and the control system responds to reach the set point. Depending on the
controlling functions, the controlling variable can be pressure, flow, and sometimes
temperature. The controllers monitor and change the controlling variables through the
SCADA system, which transmits the control signals to remote stations such as pump,

551

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 552    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Hardware/Physical
Supervisory
Master Remote
(Host)
and Terminal
Data Acquisition

Software/Protocol

Figure 9-1. Supervisory control and data acquisition system

lifting, delivery or valve station. Figure 9-1 illustrates the relationship between the
master and remote terminal through the computer and communication system.
Traditional SCADA users include the pipeline system dispatchers or controllers, oper-
ation engineers, system engineers, maintenance and measurement staff. System dispatchers
use the SCADA for safe and efficient pipeline system operation while meeting transporta-
tion requirements. Operations engineers analyze pipeline operational problems to increase
operation reliability, efficiency, and throughput as well as troubleshooting, while system
engineers configure and maintain the SCADA system including instruments and remote
terminals. Maintenance staff analyzes equipment performance based on historical data, and
measurement staff validates volume measurements.
Current business environment requires fast access to operational information. As
a result, other groups use the SCADA data to improve the pipeline business. These
groups include accounting personnel who account liquid volumes and issue invoices,
liquid marketers who use estimated batch data to schedule and market liquids move-
ments, and management who make management decisions regarding normal and ab-
normal conditions including emergency situations.
In order to accommodate a rapidly changing business condition or environment,
corporate-wide information access has become critical to the efficient operation and
management of a pipeline system. Not only is it important to provide accurate informa-
tion to operation and management staff, but timely access to this information is of vital
importance to the successful operation of the pipeline company’s business. Companies
that are able to acquire, process, and analyze information more efficiently than their
competitors have a distinct market advantage. Such expansion of the scope, function-
ality, and capabilities is made possible by continuing improvements in computer and
telecommunication technologies.
A properly designed, installed, and operating SCADA system is a keystone in the
operation and management of a pipeline in today’s competitive deregulated pipeline
market. The SCADA system has become the hub for corporate information systems.
Refer to Figure 9-2 for an overview of an integrated corporate and SCADA system.
Looking at the information requirements of a pipeline company and considering both
operational and business aspects, the key requirements can be broadly grouped into the
following categories [2]:

·· Measurement information — Measurement information is used for the

safe and efficient operation of the pipeline system. It includes pipeline data

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    553

Corporate Enterprise Volum e &

Internet/ Sales/
User Level Resource Revenue
Intranet Marketing
Planning Accounting

Corporate
Database
Interface

Non-Real- Historical Real-Tim e

Operation Tim e Database Applictions
User Level Applications

SCADA Real-tim e
Database

Field Level Pum p/

Meter Pipeline &
(PLC, RTU) Com pressor Storages
Stations Valves
Stations

Figure 9-2. Integrated corporate and SCADA system

acquired from field telemetry equipment such as volumes, flows, pressures,

temperatures, product quality, and equipment status. It would also include any
calculated data originating from the SCADA host.
·· Simulation information — Simulation information incorporates measurement
data and simulated data to diagnose current pipeline states and predict future
behavior of the pipeline. The simulation information can be used for system
optimization, line pack and capacity management, storage management, prod-
uct scheduling, and training-related applications on the pipeline system. This
data would originate from a modeling application that may use measurement
information.
·· Business information — Business information combines measurement data
and possibly simulated data along with business and economic data. The infor-
mation is used in business applications related to custody transfer, preventative
maintenance, cost tracking, contracting, marketing, inventory, scheduling and
accounting. This is where SCADA and simulation data are aggregated with
other business data to feed into business processes.
·· Decision support information — Decision support information is a summary
of the key measurement, simulation, and business data required for executive
level decision support. Extracting this key data is generally the function of
a Management Information System (MIS). Such a system has the ability to
gather and aggregate data from numerous corporate and operational databases
to supply key performance data.

It is becoming more and more common for pipeline applications to be tightly

integrated with SCADA systems and to be part of a higher level Management Informa-
tion System (MIS). The advances in computer and communication technologies have
made it much easier to connect SCADA systems to business systems. This allows for
both physical integration of SCADA and business systems as well as business process

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 554    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

integration. Process integration means that SCADA systems are becoming a key part of
business processes. This provides for both proactive business processes as well as the
ability to provide better information and thus better service for customers.

9.1.2 Pipeline System Monitoring and Control

The pipeline system has to be monitored for control. The control of a pipeline system is
achieved mainly at pump stations located in a tank farm and along the pipeline, meter
stations in receipt and delivery locations, and valves along the pipeline and in these
stations. In addition, flow direction and tank levels in a tank farm have to be monitored
and controlled. Therefore, their status and values have to be monitored first and then
controlled if required.

·· Pump station monitoring and control — Since a pump is a primary pres-

sure control device, the pipeline pressure is controlled by starting or stopping
pump units or adjusting the pressure set point at the operating pump station.
Most pump stations are equipped with multiple pump units. Depending on the
pressure or flow requirements, the operating pumps can be arranged in series,
parallel or a combination of series and parallel through open/close operations
of valves. The pump station pressure generated by fixed speed pumps are con-
trolled by a control valve at the station, while the pressure generated by vari-
able speed pumps is controlled by the pump rotation speed. Refer to Chapter 4
for a detailed discussion of pump station control.
·· Pipeline monitoring and control — The main function of a pipeline control
system is to monitor flows, pressures, and sometimes temperatures along the
pipeline. Pipeline companies receive nominations, schedule volumes, and lift
and deliver them by controlling flow rates while maintaining the pipeline pres-
sure within the operating pressure limits and the flow rates within the capacity.
Valves along the pipeline are used to change the flow direction, isolate some
sections of the pipeline, or open other sections. The quality of the products has
to be monitored and batch movements have to be tracked for proper delivery.
·· Meter station monitoring and control — An active pressure control does not
take place at a meter station. Instead, the flow rate and volumes are monitored
to control the flow directions to an appropriate location such as tank or pump.
The most important function of a meter station is custody transfer by providing
metering information.
·· Tank farm monitoring and control — Petroleum products are either lifted at
or delivered to a tank farm. A tank farm is composed of multiple tanks, booster
pumps, meters, valves and piping. The correct product should be lifted from
the correct tank at the injection station and delivered to the correct tank at the
correct tank farm. The product movement, flow direction, and tank level have
to be monitored and controlled by opening or closing various valves along the
flow path to make this happen.

9.1.3 Control Center and SCADA System

Since the SCADA system plays a critical role in the success of the pipeline business, it
must satisfy the following requirements:

·· Capability to operate the pipeline system safely and economically,

·· Provide timely and accurate data for monitoring and controlling the pipeline
system in real-time,

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    555

·· Be reliable with high availability,

·· Provide security, protecting valuable corporate information from inside and
outside intruders.

Most modern SCADA systems can provide the functionality to meet these re-
quirements. However, the combination of the SCADA system together with its control
center should be configured to fulfill them.
There are three basic tiers in a SCADA system as shown in Figure 9-2, namely,
field device, control room, and corporate. The field to SCADA connection is some form
of a telecommunications network, and the connection between SCADA host and the
corporate or enterprise environment is made with a communication network. A backup
control center located at an offsite may be connected to the main control system.
In US, PHMSA incorporates American Petroleum Institute (API) recommended
practices 1165, 1167 and 1168, which are the recommended practices for Pipeline
SCADA Displays, Pipeline SCADA Alarm Management, and Pipeline Control Room
Management, respectively. Each document describes the following:

·· API RP 1165 — Pipeline SCADA Displays [3] focuses on the design and im-
plementation of displays used for the monitoring and control of information
on SCADA.
·· API RP 1167 — Pipeline SCADA Alarm Management [4] provides guidance
on industry practices that include alarm definition and determination, alarm
philosophy, alarm functionality and design, alarm handling, alarm documenta-
tion, alarm audit and performance monitoring, roles and responsibilities, man-
agement of change, etc.
·· API RP 1168 — Pipeline Control Room Management [5] addresses pipeline
control room personnel roles, guidelines for shift turnover, pipeline control
room fatigue management, and pipeline control room management of change.

The operational nerve center of today’s pipelines is the pipeline control center. It
is from this central location that a geographically diverse pipeline is monitored and
operated. It is also the center for gathering information in real time that is used for real-
time operation, for making business decisions and for operational planning. Figure 9-3

Figure 9-3. C
 ontrol Console (Cerda J., 2008, “Oil Pipeline Logistics” Instituto de Desarrollo
Tecnológico para la Industrial, August 11–21, Mar del Plata, Argentina, http://cepac.
cheme.cmu.edu/pasi2008/slides/cerda/library/slides/jcerda-pasi-2008-1page.pdf )

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 556    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

shows a modern control console that the pipeline operator uses for minutes by minutes
system operations. Usually, several large screens are made available to monitor the
entire pipeline system.
Since the control center provides real-time information, it may also include an
emergency situation-room adjacent to the control room. This room may be dedicated
to addressing dispatching issues and particularly to resolving emergency or upset con-
ditions. Several stakeholders, including technical support and management, may be
assembled to address emergencies.
A backup control center may be required in order to operate the pipeline sys-
tem continuously in the event that the main control center is severely disrupted. This
backup is normally in a physically separate location from the main control room. The
backup center is equipped with the same equipment and devices as the main control
center. One option is that the backup system receives the real-time data directly from
the field devices each cycle, so that it is the exact replica of the primary system. The
other option is that the entire backup system is refreshed with the required data re-
ceived from the primary system at a regular interval.
The division of control between a central location and the local pump station var-
ies widely. A large complex pipeline system may be divided into multiple control sec-
tions defined in terms of size of the pipeline network, complexity of the network, or
number of shippers. This division allows the operators, assigned to each section, to
efficiently monitor and safely control the pipeline system.
A control center houses most of the equipment used by the operators on a daily
basis. The equipment required includes the SCADA system computers and terminals,
printers, communication devices, and network equipment used to implement Local
Area Networks (LAN) and/or Wide Area Networks (WAN). In addition, pipeline sys-
tem maps and schematics may be displayed, and operator manuals and other informa-
tion required for performing dispatching functions can be made available.
A SCADA system consists of three main components; host or master, commu-
nication system, and remote terminals. A SCADA host or “master” is a collection of
computer equipment and software located at the control center and used to centrally
monitor and control the activity of the SCADA network, receive and store data from
field devices and send commands to the field. A SCADA system gathers the data from a
variety of field instrumentation, typically connected to remote terminals. See Figure 9-4

Figure 9-4. Typical SCADA system

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    557

for a modern SCADA architecture of both the main control center and backup control
center.
The architecture of SCADA systems can vary from a relatively simple configura-
tion of a computer and modems to a complicated network of equipment. In whatever
form it takes, however, SCADA architecture will incorporate the following key hard-
ware and software capabilities:

·· Ability to interface with field devices and facilities for control and/or monitor-
ing, usually through remote terminals.
·· Provision of a communication network capable of two-way communication be-
tween the remote terminals and the control center. This network might also pro-
vide communication between the control center and a backup control center.
·· Ability to process all incoming data and enable outgoing commands through a
collection of equipment and software called the SCADA host.

Modern SCADA systems provide additional capability:

·· Business applications such as meter ticketing, volume accounting, nomination

management, etc.
·· Application software such as leak detection, inventory management, and training
·· Interface to corporate systems

The network is normally an internal private network. However, there are now
SCADA systems that utilize secure connections to the Internet that replaces the private
network. Web-based SCADA systems are ideal for remote unattended applications, as-
suming that an RTU or flow computer is available. In other words, they are suitable to
pipeline systems or remote locations where centralized computing or control require-
ments are not intense and the primary function is remote data gathering. A web-based
SCADA system offers several benefits. The main advantages are as follows:

·· It provides an economical solution with wireless technology using the Internet

infrastructure.
·· It allows data access from anywhere without extra investment in communica-
tion and software.

Here, it needs to be mentioned that a distributed control system (DCS), instead of a

SCADA, can be used for controlling pipeline systems. The goals of DCS and SCADA
are quite different. A DCS is process oriented. It looks at the controlled process (the gas
processing plant or chemical plant) as the center of the universe, and it presents data to
the operators as part of its job. SCADA is data-gathering oriented; the control center
and operators are the center of its universe and the remote equipment is merely there to
collect the data — though it may also do some very complex process control.
DCS systems were developed to automate process control systems. These systems
are characterized by having many closed loop control elements controlling an analogue
process in real time. The key differences and characteristics of DCS and SCADA are
as follows:

·· A DCS normally does not have remotely (i.e., off-site) located components
and is always connected to its data source. Redundancy is usually handled by
parallel equipment.
·· SCADA needs to have secure data and control over a potentially unreliable
and slow communication medium, and needs to maintain a database of ‘last

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 558    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

known good values’ for prompt operator display. It frequently needs to do event
processing and data quality validation. Redundancy is usually handled in a
distributed manner.
·· A DCS does not poll data but rather needs to be able to process a high number
of transactions at a high speed in order to implement multiple real-time closed
loop control.
·· The majority of operations, such as start/stop commands and alarm detection
of a SCADA system are digital. They also gather/poll analogue readings but
do not implement closed loop control; humans determine if set points need to
be adjusted. A DCS is process control oriented and therefore is designed to be
able to implement many control loops as well as standard operator initiated
start/stop commands.
·· When the DCS operator wants to see information, he usually makes a request
directly to the field I/O and gets a response. Field events can directly interrupt
the system and the operator is advised automatically of their occurrence.

The remote terminals are located where the process values are monitored and
interfaced with the host SCADA. They can be a remote terminal unit (RTU), program-
mable logic controller (PLC), or flow computer. The remote terminals collect data from
the process devices, transmit data to the host SCADA, receive supervisory commands
from the host SCADA, and issue these commands to the process devices. Supervisory
commands may include pump/compressor station or unit start and stop commands,
valve opening and closing commands, and set point settings.
An RTU acquires process values independent of the host by scanning hardware
and software points, and communicates with the host, field I/O points, and other com-
puter systems. It can detect and report alarm conditions, which include I/O error, bad
measurement, high/low limit violations, rate-of-change alarm, and other deviations
from set-points. An RTU provides limited control functions at field devices. The func-
tions range from simple on-off or open-close control to logical control sequences such
as ESD. It supports diagnostic checks with diagnostic software running in the remote
watching for a number of possible problems. Some RTUs provide electronic flow
measurement capability, by performing calculations of AGA, API and other standards,
storing the measurement data, and allowing instant access of the measurement data.
A PLC provides extensive control, communication and operator interface capa-
bilities. PLCs are used as remote terminals on a SCADA system, the heart of station
control for field equipment (pumps, drivers, lube oil systems), communicating with
the host. At a pump station, it can perform all the monitoring and control functions of
pump unit and driver, station valve, station suction and discharge, station electrical
and auxiliary equipment. It may have its own memory for the data to be transferred, or
logic control for the gathering of data and error-checking with the host. PLCs can also
be networked to provide a complete control system for a complex station.
It has to be noted that DCSs are not only economic for large installations but can
be a solution choice for larger pump stations. They would certainly be considered for
installations where there is a station and an associated processing facility or a refinery
that would utilize a DCS for its control. The traditional boundaries between various
control system solution options have become blurred due to the flexibility of today’s
control equipment. For small systems, the control system will generally be imple-
mented using a PLC. As the facility gets larger and more complex, several options are
now available of choosing between installing a control system using networked PLCs
or a DCS system, requiring a careful consideration to ensure the operating require-
ments are met while at the same time the design dovetails with corporate business
information gathering and processing.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    559

SCADA SCADA ARCHIVE ARCHIVE

MASTER MASTER SERVER SERVER
A B A B
(PRIMARY) (BACKUP)

LAN A

LAN B

OPERATOR OPERATOR TERMINAL TERMINAL

WOKSTATION WOKSTATION SERVER SERVER
1 2 A B

CROSSBAR
SWITCH

Figure 9-5. Redundant SCADA system

Reliability and availability requirements particular to individual installations will

determine the configuration of redundant SCADA and database computers and redun-
dant networks. Reliability provides an indication of how frequently a system or device
will fail, while availability is the amount of time a system is fully functional divided
by the sum of the time a system is fully functional plus the time to repair failures.
­Figure 9-5 illustrates a fully redundant SCADA architecture, in which both computer
and communication systems including the associated equipment are duplicated.
SCADA host software architecture is different for every product. However, they
all have the following key components:

·· Operating system such as Unix, Windows or Linux

·· Relational database for historical data management, interfacing with corporate
databases
·· Real-time database for processing real-time data quickly
·· Real-time event manager, which is the core of the SCADA
·· HMI manager for user interfaces

In addition, various utilities and development software are important for system
development, configuration, and maintenance.
The SCADA will manage the polling of data, processing of that data and the pass-
ing of it to the real-time database. It will make data available to the presentation layer
consisting of the HMI Manager and interfaces to other applications, as well as process
control and data requests.

9.1.4 Data Communications

Data communications for a SCADA system require various components; modem, pro-
tocols, network, transmission media, and polling. A modem is an electronic device that
encodes digital data on to an analog carrier signal (a process referred to as modulation),
and also decodes modulated signals (demodulation). This enables computers’ digital
data to be carried over analog networks, such as the conventional telephone network.
In general, modems are used for the connection between an RTU and the SCADA

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 560    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

network or where it is not feasible to have a high-speed network, connection directly

to the RTU.
In the context of data communication, a network protocol is a formal set of rules,
conventions, and data structure that governs how computers and other network devices
exchange information over a network. In other words, a protocol is a standard proce­
dure and format that two data communication devices must understand, accept, and use
to be able to exchange data with each other. A wide variety of network protocols exist,
which are defined by worldwide standards organizations and technology vendors over
years of technology evolution and development. For a retrofit or upgrade project, it is
important to ensure that the SCADA system can support all of the protocols that exist
in the legacy equipment that will be connected to the SCADA system. In some cases
where there may be proprietary protocols, converters may need to be implemented.
On a new SCADA system, there is no need to be concerned about existing equipment
and protocols. However, it is important to ensure that the SCADA system utilizes in-
dustry standard protocols and not just proprietary ones. This will make expansion and
addition of new equipment easier. It will also provide more flexibility in being able to
choose equipment from a wide range of vendors and not be tied to a specific vendor’s
equipment.
A SCADA system will usually incorporate a local area network (LAN) within a
control center and one or more wide-area networks (WANs). The major improvement
in current generation SCADA systems comes from the use of WAN protocols. Not only
does this facilitate the use of standard third party equipment but more importantly it al-
lows for the possibility to distribute SCADA functionality across a WAN and not just a
LAN. In some WAN distributed systems, pipeline controls are not assigned to a single
central location. Instead, control operations can be switched or shared between numer-
ous control centers. Responsibilities can be divided vertically according to a control
hierarchy or horizontally according to geographic criteria. In both cases co-ordination
and integration of control commands issued from various centers are maintained. In
the event of the loss of one or more control centers, the operation can be switched to
another center.
The SCADA network requires some form of communication media to implement
the WAN connection between the SCADA host and remote locations. Ultimately the
choice of which media to use to implement a connection to a remote site will be based
on cost, availability of a particular medium and technical factors such as reliability,
data transfer rate, geography, etc. A second choice to be made is whether the commu­
nication should be leased from a 3rd party or owned and operated by the pipeline
company. This decision needs to be consistent with the corporate IT and operating
guidelines. Commonly adopted communication media include:

·· Metallic line is a hardwired physical connection between the SCADA host and
the remote location. This is a good practical choice in SCADA applications
where the distances between the SCADA host and the remote locations are not
significant and there may be a limited choice of other media. An equivalent is
usually leasing “lines” from a telephone company. The connection will utilize
the internal network of the telephone company and may be any combination
of wire, fibre optic cable, and radio. Another alternative is to utilize mobile
telephone networks which provide good coverage in populated areas.
·· Application of radio transmission on a pipeline SCADA usually takes two
forms. The simple case is where a radio link is used as the last communication
link between the SCADA and a remote site. The main communication back-
bone of the SCADA system is some other media other than simple radio. A long
distance pipeline that may be geographically located in remote areas as well as

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    561

near occupied areas may well incorporate a mix of radio links and fixed links
(leased lines, fibre optic, etc.)
·· A fibre optic cable uses coherent laser light sent along a “cable.” The cables
are not lossless and repeater equipment is required at spacing of up to 100 km.
The growth of the need for data transfer capability for the internet and private
networks has spurred advances in fibre optic equipment. Because a fibre optic
cable uses light and not electricity to transmit data it has the benefit of being
unaffected by electromagnetic interference. On new pipeline projects, some
pipeline companies have installed fibre optic cable in the same right of way
as the pipeline. This can be a cost effective way of providing a transmission
medium to implement the SCADA WAN.
·· A satellite can provide a cost effective communication solution for pipelines
under certain conditions. This solution is usually considered when the RTU is
in a very remote location where the ability to utilize other media is not practical
or very expensive. The capital cost is typically more than alternative techniques
but when operating costs are factored in, this option can be a cost effective
solution. However, poor weather conditions can adversely affect the reliability
of communications.

Polling is the term used to describe the process of the SCADA host communicat-
ing with a number of RTUs connected on a network and exchanging data with each
RTU. The arrangement between the SCADA host and the remote RTU is sometimes
referred to as ‘master-slave’ implying that the SCADA host is in charge of each com-
munication session with an RTU. The types of polling schemes are as follows:

·· Polled Only or multi-drop scheme: The SCADA host will sequentially initiate
communication with each RTU in sequence on a fixed schedule. There will be
a fixed number of attempts to establish communication with an RTU before
reporting that communications with the RTU are faulty. One can imagine that
for a system with a large number of points to be updated at the SCADA host,
this may take some time and therefore there will be some time lag between the
sample time for the first data point and the last.
·· Freeze scheme: One variation of multi-drop scheme is the ability of the master to
issue a freeze command to all RTUs. The RTUs then store their data samples and
the master begins polling and retrieves the data. This results in a database update
at the master where all data was taken more or less at the same time. One way of
mitigating this is to have all the RTUs take and store data samples at the same time.
The major disadvantage of the above two schemes is that the status and value of all
data base points are transmitted every polling cycle, which can be costly.
·· Polled Report by Exception (RBE): In this scheme, a local history of each data
point is saved and the RTU will only send back those points that have changed
since the last poll. In the case of an analogue value, these will have a dead
band that the value must exceed before a new value is sent back to the SCADA
host. This reduces the amount of data traffic on the network. The user must be
careful in choosing dead bands for analogue values for example to ensure that
information is not lost.
·· Unsolicited RBE: In this case, the host does not poll on a regular basis, but each
RTU “pushes” data back to the host when it has updated data to send. This can
reduce data traffic even more than the polled RBE. However, it has the disad-
vantage of the host not knowing if data points have not changed or failed. A
variation can be to have a system that incorporates a guaranteed polling time.
For example, all RTUs may be scanned at least once every 15 minutes.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 562    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

9.1.5 Data Management

Typical data required for the safe and efficient operation of pipeline systems from vari-
ous locations include the following:

·· Pump station — values and quality of suction, casing and discharge pressures.
Sometimes, temperature value and its quality are made available, especially
if a heater or chiller is installed. Status of various valves is also required. For
variable speed pumps, unit speeds are made available for the operator to review
the performance of the unit. Data to allow monitoring of unit operating point is
also useful for determining operating efficiency.
·· Meter station — values and quality of flow, pressure, and temperature. In addi-
tion, a densitometer reading may be available for liquid pipelines, particularly
batch pipelines. The status of various valves is also required.
·· Control or pressure reducing valve station — values and quality of pressure or
flow.
·· Pipeline — values and quality of pressures along the pipeline. Sometimes,
values and quality of temperature are available. These may be retrieved from
automated block valve sites to take advantage of the need for an RTU for valve
control. The incremental cost of pressure and temperature measurements in this
situation is minimal.

Alarm messages are generated to signal the potential or real interruption of normal
operation at any monitored location on the pipeline.
There are four basic data types in a SCADA system, namely, “discrete,” “ana-
logue,” “internal,” and “parameter”:

·· The term “discrete” reflects the fact that these points can only be in one of two
(or more) predefined states. Discrete points are generally binary in nature, i.e.,
they only have distinct states. This can represent open/closed, on/off, normal/
alarm, etc. They are referred to as digital, status or binary points. Some systems
will implement three or four state points, such as a valve status, to indicate that
the valve is “open,” “in transit,” or “closed,” or a pump is running, in start-up,
in shutdown or off.
·· “Analogue” or “Analog” refers to points that have a numeric value rather than
two or more discrete states. Analogue inputs are field data points with a value
that represents a process variable at any given remote location such as pipeline
pressure, oil temperature or pressure set point on a control valve. Analogue
output points can also be sent as commands from the SCADA host, such as set
points for controllers.
·· A third type of data point is determined internally by the SCADA host as op-
posed to being sent by an RTU. The internal data type is also called derived
data. This can range from a simple calculation to change the engineering units
of a field value to more complicated calculations such as the corrected volume
measurement in a tank based on tank level, temperature, and product density of
a flow calculation that uses API corrections.
·· Parameters or factors are generally used to calculate derived values. Exam-
ples include orifice plate sizes, AGA calculation parameters, and performance
curves.

All SCADA systems work in a real-time environment, consequently they have

a real-time database to process real-time data. The real-time database must be able

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    563

to process large amounts of real-time data quickly. A typical corporate relational

database cannot meet such requirements. Conventional database systems are typi-
cally not used in real-time applications due to their poor performance and lack of
predictability.
The SCADA host must be able to meet the requirements of a real-time environ-
ment and easily interface to standard external databases for the purposes of making key
data available to other business processes. One method used is to utilize some form of
a data repository or data historian to store SCADA data for access by other applica-
tions. This reduces the transactions in the real-time database and improves response
performance.
Creating the SCADA database consists of populating the database with each of
the individual data sources in the SCADA network. Each point will require a number
of information fields to be entered to complete a record in the database. This effort
is a time-consuming task and must be done accurately. Typically, the SCADA host
provides a high-level software utility for interactive creation and modification of the
system database. A key feature of a SCADA system is the ability to download RTU
configuration information from the database thus eliminating the need to re-enter data
at each RTU. This also eliminates another source of possible error. Database changes
(e.g., addition, deletion, or modification of points) can generally be performed on-line
and should not require recompiling the system software
All data points will be stored with a time stamp indicating when they were sam-
pled by the RTU. A “quality” flag may also be stored indicating the quality of the
value. Some examples of quality indicators are “Good” meaning that the data has been
scanned recently and is within range, “Stale” indicating that the point has not been re-
freshed for some configurable period, “Bad” meaning that the point’s value cannot be
relied upon, etc. Analogue values are processed by the SCADA host and stored in the
real-time database, usually along with the original or raw value received from the RTU.
Typical processing of analogue points could include conversion to engineering units,
alarm checking against pre-set values for each reading, rate of change alarm, instru-
ment failure alarm, averaging, and totalizing such as volume going into a tank.
SCADA data security and integrity features must be consistent with the corporate
IT standards and should be outlined during the development of the SCADA require-
ments. SCADA manuals should include detailed procedures for generating accurate
and complete copies of records, while the system should allow for each user’s account
to limit the access and function the user can execute. All SCADA historical records
should use secure, computer-generated, time-stamped audit trails to independently
record the date and time of operator entries and actions that create, modify, or delete
electronic records.
A historical database provides for internal analysis and reference as well as meet-
ing the requirements of regulating agencies to review pipeline system operation.
For example, operation engineers use the historical data for operational analysis for
performance enhancement. The regulator may require emergency scan data to track
events leading to and following an emergency condition and eventually to determine
the cause/effect relationship. SCADA historical data includes time-stamped analogue
values and other control-related analogue values. It can also include digital points and
host generated points including alarm and event logs. Operator task logs are also typi-
cally included.
Since a large amount of data can be accumulated, the historical data needs to be ar-
chived periodically. Archived data refers to data that has been stored on archival media
(CD, digital tape, etc.) and is stored in a separate location from the SCADA host sys-
tem as required by corporate policy. The period of time after which data should be ar-
chived is determined by corporate policy. The data archive should include all ­analogue

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 564    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

and digital data, alarms, events and operator actions. The SCADA system must be able
to retrieve archived data without interrupting ongoing process operations.
To facilitate analyzing system upsets and events, the SCADA system can have
a feature known as “playback.” This functions much like a rewind on a VCR and al-
lows a user to replay historical data through an off-line operator terminal in order to
more easily analyze and determine the root cause of an upset. It can also be used to
do “post-mortems” with operators to provide feedback on actions that were taken and
to determine if remedial action taken was done correctly and in a timely fashion. The
SCADA data base manager needs to store and time tag all operator actions (alarm ac-
knowledgment, commands, etc.) as well as all incoming and outgoing data to get the
most benefit from this feature.

9.1.6 Alarms
For safe pipeline operation, potential alarm situations should be addressed by annunci-
ating alarm messages. High-priority alarms may require audible alarming. Alarm condi-
tions are expected during the course of pipeline system operation. The alarm processing
function can help to identify potential alarm conditions before actual alarm conditions
occur. Examples of potential alarms include high-pressure violation, high-temperature
violation at a compressor discharge, leak detection, etc.
The alarm processing function should be able to limit the number of alarms to
those that are important. If the number of alarms is too large, the operator’s attention is
consumed reviewing and acknowledging alarms instead of monitoring and controlling
the pipeline system. An overabundance of alarms also desensitizes the operators and
can result in them ignoring critical alarm conditions.
In general, alarms are prioritized according to their critical nature in order to give
the operator an indication of which alarms need to be attended to first. Emergency
alarms require the operators to take immediate action to correct the condition, while
communication alarms may require them to contact support staff immediately. Warn-
ing alarms are not usually critical, requiring preventive measure without immediate
action. The severity of alarms should be configured to be one of multiple levels of
severity (for example, high, medium, or low) for all alarm generating points. Alarms
are usually color coded, requiring a different color for each level of alarm. In addition,
an audible signal should be generated for high-level alarms.
Analogue alarms are generated when a current value for an analogue point reaches a
limit pre-defined in the data base attribute for that point. Figure 9-6 is a typical alarm sum-
mary display, which in this case shows the conditions both “in alarm” and “not in alarm”
as well as both the “unacknowledged” and “acknowledged” statuses. The first two alarm
messages are in an alarm condition because the tank is in “Low-Low” level.
Alarm levels will typically include the following:

·· High-High (or Alarm) means that the point has reached its maximum allowable
value. This will generally mean that it is close to or has reached a point where
local automatic protection systems may be initiating action.
·· High (or High Warning) means that the point has reached a warning level. If
remedial action is not taken, the point may reach High-High. The trending sys-
tem will allow an operator to display such a point to see how long it has taken
the point to get to the warning level.
·· Low-Low (or Alarm) similar to High-High but for a lower limit
·· Low (or Low Warning) similar to High but for a lower limit
·· Rate of Change: The slope of a trend line has exceeded a pre-defined limit. This
means the process value is changing more rapidly than would be expected.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    565

Figure 9-6. Alarm summary (courtesy of Telvent)

Discrete alarms are generated upon a change of state of the data base point. These
can represent:

·· Change from normal to abnormal such as a high-temperature alarm on a com-

pressor station outlet.
·· A change of status that was not the result of an operator control action. For
example, a valve closes or a pump shuts down with no initiation from the
operator.

All such alarms will be reported and logged, as will any change of status of a point.
This will provide not only a record of all abnormal events but will also show when
equipment was acted upon by an operator.
A basic alarm management scheme consists of detecting the alarm and reporting
the alarm to the operator. An alarm management system will also log and provide an audit
trail of each alarm. This will include the time that the alarm was reported, when it was ac-
knowledged by the operator and when the alarming point returned to normal. This informa-
tion along with the database log will provide key information for post-event analysis.
In any system upset, there will be an initiating event followed by secondary indica-
tions or alarms. For example, a control valve may fail causing pressure to rise, which may
then cause pressure relief valves to operate and flow rates to exceed expected values. Some
SCADA systems may incorporate some form of artificial intelligence to process alarms
automatically to advise the operator of what the potential root cause may be.
The SCADA database will have the ability to assign various levels of alarm sever-
ity to individual points to provide an easy means of reporting high-priority alarms to
an operator. In an emergency condition, it is important to not overload an operator and
allow concentration on priority items.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 566    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

The alarm message includes the date and time of the alarm, the point that caused
the alarm, the severity of the alarm denoted by color and an audible signal, and the state
of the point. The message is displayed in the alarm window and in the tabular summary
of alarms. The alarm window lists all unacknowledged alarms, which should be made
available on the screen at all times. Alarms are always logged in an event summary,
including not only all the information in the alarm message but also the time when the
alarm was acknowledged and by whom.
The operators should be able to easily monitor alarm messages and quickly re-
spond to the messages. Therefore, messages should be made readily available to the
operator. The current alarm summary is mainly used for monitoring and acknowledg-
ing the messages, while the alarm history summary is mainly used for reviewing the
alarm status and pipeline system operation.

9.1.7 Human Machine Interface (HMI) and Reporting

Key features of displays and reports are discussed in this section. Typical data x/y in-
cluded in displays and reports are as follows:

·· Telemetered data, including analogue, digital, and derived values and quality
·· Parameter data, such as orifice plate size
·· Schematic information, including station yard piping, facility locations on the
pipeline system, and other pertinent information

The displays need to be designed to meet the needs of individual operators, be-
cause they are the prime users of SCADA displays. Displays need to:

·· provide a fixed area on the screen for alarm and emergency annunciation
·· refresh the displays dynamically and within a short time (at most a few sec-
onds) after a command is issued
·· allow the operators to be able to navigate the displays easily and quickly
·· maintain a consistent “look and feel” and use intuitive design industry-­accepted
display design methodologies and standards.

All SCADA vendors will have a comprehensive HMI system, which will include
tools for creating and modifying displays and reports. In fact, the capabilities of most
systems can be bewildering and intimidating. Since a typical SCADA host will have
a large real-time database, the challenge is to design an HMI that presents relevant
information to the operator in an easy to understand set of displays.
It is suggested that a fixed area be reserved on the screen for alarm and event mes-
sages, system performance monitoring, and annunciation of emergency scan. In other
words, this information remains on the screen always until it has been acknowledged.
It is important to develop some guiding principles for each system before the displays
are created. These guidelines should include some variation of the following:

1. Have a hierarchical approach: Top-level displays will show key summary

­information but also have the ability to “zoom” in quickly for more detail.
­Typically, the top level display is a pipeline system overview or a pipeline sys-
tem schematic. The system overview display allows the operator not only to
view the current pipeline states including set points and alarms of the ­system but
also to ­access a particular station for viewing control points and/or ­modifying
their ­values. It not only displays all pump/compressor stations and current alarm
­messages but also flow, pressure and temperature including set points. In ­addition,

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    567

this display may show the link to pump/compressor, meter, or valve station con-
trol panels through which the operator can send a control command.
2. Screen navigation should follow the current expected features found in most
window-type navigation software to reduce operator-learning time and to make
the system as intuitive as possible.
3. Ensure a consistent “look and feel” of displays to minimize training and the
chance of operator error. These will include the use of color and a consistent
and logical approach to the use of buttons, menus and toolbars. A judicious
choice of colors is important as certain types of color blindness can result in
some colors appearing the same to some people.
4. Keep screens as uncluttered as possible while still supplying the required in-
formation. The possibility of confusion should be minimized and care taken to
reduce the possibility of information being lost or “buried” on the screen.

The following displays are considered to be key display requirements for effective
operation of the pipeline system:

·· Display or schematics of an entire pipeline system

·· Pump station overview including measurements on piping, unit and driver
·· Meter station information including flow rate and accumulated volume, total
station flows, etc.
·· Pipeline elevation and pressure profiles with MAOP
·· Batch tracking information along the pipeline system
·· Tank and storage information such as tank inventory
·· Alarm and event annunciation and summary
·· Communication summary
·· Measurement and equipment status summary
·· Security-related information including system status and police contacts

The displays are either in tabular or graphical format. In some cases, it may be
useful to have both tabular and graphical formats for displaying data. The selection of
format depends on how the data is used. For example, it is more useful to display pres-
sure drop along the pipeline in graphical format. Most modern SCADA systems use
several display mechanisms, which include textual and graphical images augmented
with real-time information. Color and shape can be used to relate discrete information
in an intuitive manner, plot and trend display types can be used to display graphs of
analog data in an x/y format or a horizontal or vertical bar graph.
There are other display types such as pushbutton for selecting a button to perform
a specific function, meter/gauge for showing a meter/gauge device with values, and
region for marking a location on a display. Some SCADA display systems support
display format control. The format control functions include popup and pan/zoom.
For example, the functions such as set-point control and communication control can
be supported by pop-ups. A large display area can be easily navigated by means of a
panning/zooming feature of the display system.
Figure 9-7 is a display of a pipeline system. It shows the operating statuses and
parameters of the entire pipeline system; pump stations and the operating pump units,
current station pressures and flow rate, density, and list of alarm messages at the bottom
of the display. From this display, a desired pump station or alarm message is selected
to review the detailed data for the station.
Figure 9-8 shows a typical pump station diagram. To monitor or control a pump
station, it can be directly selected from the display of the pipeline system. Then, the
operator can monitor the measured variables and controlled parameters of the selected

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 568    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Figure 9-7. Screen display of a pipeline system (courtesy of Telvent)

Figure 9-8. Pump station diagram (courtesy of Telvent)

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    569

station as well as its station operation statuses including pump unit and valve statuses.
From the pump station display, the pump units and valves can be controlled and the
control point can be set. Normally, the suction or discharge pressure is controlled from
this screen.
Figure 9-9 displays a meter station in a tank farm, installed with a meter prover.
The operator can view the current meter station statuses and meter data including flow
meter and valve positions as well as control the meter station from this display. The
operator can acknowledge any alarm messages related to the meter station operation.
These alarms are listed in the alarm summary at the bottom of the screen. In addition,
the flow meter can be proved by means of a meter prover.
Figure 9-10 displays the elevation profile, and pressure profile with MAOP. The
pressures can be presented in terms of head so that all three units are the same. This
allows the operator to visually detect trouble spots such as slack flow conditions along
the pipeline.
Data trending capability is one of the most important functions of any SCADA
system because it helps the dispatchers and operations staff to identify potential prob-
lems before they arise and to diagnose alarm conditions. Data trending is used to dis-
play any analogue values which are stored in the historical database over time at a
specific location or locations. Data trending displays are in graphical format due to the

Figure 9-9. Meter station with a prover (courtesy of Telvent)

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 570    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Figure 9-10. MAOP, pressure and elevation profiles (courtesy of Telvent)

large amount of data. Figure 9-11 shows a typical trend plot of flow rate, pressure, tem-
perature, and API gravity at any specific point of a pipeline. Any data can be trended,
and the trended data can be analyzed to detect any anomaly at the point.
All SCADA systems have some type of reporting capability. This will typically
consist of both standard reports generated automatically by the system and user-defined
reports. These reports are generated from the SCADA databases containing real-time,
historical and calculated data. The standard reports are of a predefined structure, while
the user-defined reports meet the user’s specific needs. Examples of standard reports
include operating summary reports and billing reports, and those of user-defined re-
ports include such things as command/alarm log sorted by station.

Figure 9-11. Data trending

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    571

The types of reports usually found on a pipeline SCADA system would include
some of the following:

·· Operating reports
·· Shift or daily operating summary reports
·· Product movement report
·· Alarm summary report
·· System availability, communication and reliability report
·· Emergency scan report, containing operating data during emergency conditions

Government regulators may require pipeline companies to submit regulatory re-

ports. Normal operation reports may need to be submitted regularly, but emergency
reports are mandatory in the event of emergency conditions.
The SCADA system provides system administration tools to configure and main-
tain the system, and allows the SCADA users to access various logs.

·· Command log, containing a record of all commands issued by the operator

·· Alarm log, containing all generated and acknowledged alarm messages for
tracking operational problems
·· Database maintenance log for recording commands used to change any SCADA
database values
·· System log for recording the SCADA system performance including error data
such as the start/stop time, abnormal running time, etc.
·· Communication log for recording the statistics of the communications with the
RTUs such as the number of attempts, the number and types of error, etc.

The number, content, and style of reports will vary widely depending on the pipe-
line type, the business requirements, and the regulatory environment. It is important
that the SCADA system provides an easy to use, flexible reporting package that does
not require programming changes to create and implement reports.

9.1.8 Security
A SCADA system will provide for user password access and the ability to configure
specific levels of access for each user. For example, there may be users who may ac-
cess the SCADA system but are allowed only the ability to read some pre-configured
reports. For example, only those who are directly responsible for the database are al-
lowed to make changes to the database and this is done with password protection.
SCADA systems have long been thought to operate in a secure environment
because of their closed networks, which are not exposed to external entities. In ad-
dition, the communication protocols employed were primarily proprietary and not
commonly published. Recent advances, such as Web-based reporting and remote
operator access, have driven the requirement to interface with the Internet. This
opens up physical access over the public network and subjects SCADA systems
to the same potential malicious threats as those that corporate networks face on a
regular basis.
Typically, compliance with industry standards and technologies is regarded as
a good thing. However, in the case of newer SCADA systems, recent adoption of
commonly used operating systems and standards makes for a more vulnerable target.
Newer SCADA systems have begun to use operating systems such as Windows that
are commonplace in corporate networks. While this move offers benefits, it also makes
SCADA systems susceptible to numerous attacks related to these operating systems.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 572    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

RTU to host protocols are now utilizing industry standard protocols, which may com-
promise their security.
Due to cyber terrorism, the security associated with the SCADA network needs to
be designed and assessed by the same policies utilized in other areas of the company.
If there are no such clear network security policies in place, then they need to be estab-
lished before taking specific actions on the SCADA network. For detailed information
on SCADA security, refer to API Standard 1164 — Pipeline SCADA Security [6]. This
standard provides guidance to the operators of pipeline systems for managing SCADA
system integrity and security.

9.2 OVERVIEW OF PIPELINE LEAK DETECTION SYSTEM

9.2.1 Introduction
This section discusses various aspects of pipeline leak detection without an emphasis
on any particular techniques. Anyone, who is interested in the detailed discussion of
the leak detection techniques and their implementation considerations, are referred
to other volumes [1]. This section introduces the selection criteria of a leak detection
system and various leak detection techniques.
Pipeline leak detection is only one aspect of a pipeline leak management pro-
gram; it encompasses leak prevention, detection and mitigation procedures. In order
to minimize the consequences of a leak, pipeline companies require a comprehensive
leak management program. A leak detection system by itself does not improve on a
pipeline’s integrity nor reduce potential failures of a pipeline system. However, such a
program will not only help prevent and monitor the degradation of a pipeline that may
eventually lead to failure, but will also minimize the consequences of pipeline leaks
if they occur.
Pipeline companies minimize leaks through a leak prevention program. The main
causes of leaks are: outside or third party damage such as excavation equipment hitting
the pipeline, geophysical forces such as floods and landslides, improper control of the
pipeline system, and pipe corrosion. Figure 9-12 shows leak statistics in US, Canada
and Europe.

Figure 9-12. Leak statistics

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    573

Even though the statistics are about ten years old, they can be relevant to ad-
dress key issues on leaks. Incidents resulting from damage by a third party are signifi-
cantly higher in Europe than those in Canada, mainly because the population density
in Europe is much higher. Proper control of third-party damage is achieved through:
marking of the right of way; education of employees, contractors, and the public; and
effective use of systems such as “One-Call.” Geophysical forces cannot be controlled
but can be monitored and their effects can be mitigated. Pipeline integrity management
is a significant subject by itself and discussed in separate volumes [7].
Leak mitigation is the attempt to reduce the consequences of a leak when it occurs.
If a leak can be detected quickly and isolated quickly, the spillage can be minimized.
This requires that the leak alarm and its associated information are reliable and accu-
rate. Having effective procedures in place and the proper resources and tools to enact
them are critical in addressing the leak mitigation issues efficiently. The leak confirma-
tion and isolation issues should be part of leak detection. The scope of leak detection
does not normally include spillage management issues such as cleanup procedures and
manpower mobilization.
Historical data indicates that leaks were predominantly detected by local opera-
tion staff and third parties. Successful detection by means of a single leak detection
system was random. This was because no single leak detection system could detect
leaks quickly and accurately or provide reliable leak detection continuously and cost-
effectively. Therefore, more systematic approaches to leak detection are required, such
as a combination of line patrol, sensing devices and/or SCADA-based systems with
automated leak detection capability.
Since SCADA systems have become an integral part of pipeline operations, a
particular consideration has to be given to leak detection methods that can be easily
implemented on the SCADA system. API Publication 1130 [8] addresses various Com-
putational Pipeline Monitoring (CPM) methodologies, integrated with a host SCADA
system. In association with the CPM, API Publication 1149 [9] and API Publication
1155 [10] are briefly discussed with respect to how they are used for specifying and
evaluating leak detection performance.
Pipeline Leaks
This chapter uses the definition of leaks as defined in “Petroleum Pipeline Leak De-
tection Study [11].” There are two types of leaks: an incipient leak and an actual leak.
“Incipient leaks” are defined as those that are just about to occur. Certain incipient
leaks can be discovered by inspecting the pipeline and dealt with before they become
actual leaks. Here, an actual leak is called a pipeline leak when fluid is leaking out of
a pipeline system.
All pipeline leaks are associated with certain external and internal phenomena.
External phenomena include the following:

·· Spilled product around the pipeline

·· Noise generated from leakage at the hole in the pipe
·· Temperature changes around the hole

Internal phenomena include:

·· Pressure drops and flow changes

·· Noise around the hole
·· Temperature drop for gas pipeline

All leak detection systems take advantage of the presence of one or more leak
phenomena.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 574    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Standards on Leak Detection

In North America, a leak detection system is normally required on new liquid pipelines,
but not on existing pipelines unless mandated otherwise by the appropriate regulatory
agency. In general, there is no leak detection requirement on gas pipelines other than a
few new gas pipelines. The same is true of multi-phase gathering pipelines.
Pipeline companies are using various leak detection methods with varying degrees
of success. At present, no single method truly stands out as an ideal system able to de-
tect the wide ranges of leaks with accuracy and reliability, and having low installation
and operating costs. Some are accurate and reliable but too expensive, and some are
economical but unreliable.
Different countries have developed different leak detection regulations and prac-
tices. A few references and standards are introduced below. However, in general, the
codes and standards on pipeline leak detection are not well defined.
American Petroleum Institute (API) has published several standards on pipeline
leak detection. They are listed below:

·· API 1130 “Computational Pipeline Monitoring” addresses the design, imple-

mentation, testing and operation of Computational Pipeline Monitoring (CPM)
systems. It is intended as a reference for pipeline operating companies and
other service companies. The publication is used as a standard by regulatory
agencies in many parts of the world.
·· API 1149 “Pipeline Variable Uncertainties and Their Effects on Leak Detecta-
bility” discusses the effects of variable uncertainties and leak detectability eval-
uation procedures for a computational pipeline monitoring methodology. This
publication describes a method of analyzing detectable leak sizes theoretically
using physical parameters of the target pipeline. It can be used for assessing
leak detectability for new and existing pipelines.
·· API 1155 “Evaluation Methodology for Software Based Leak Detection Sys-
tems” describes the procedures for determining CPM’s leak detection perform-
ance. Unlike API 1149, this publication addresses the performance evaluation
procedures based on physical pipeline characteristics and actual operating data
collected from pipeline operations.

The Canadian standards applicable to oil and gas pipelines are specified in Z662,
“Oil and Gas Pipeline Systems.” Section 10.2.6 of Z662 specifies leak detection for
liquid hydrocarbon pipeline systems. The specifications in Section 10.2.6 for liquid
pipeline systems states:

“Operating companies shall make periodic line balance measurements for sys-
tem integrity. Operating companies shall periodically review their leak detection
methods to confirm their adequacy and effectiveness. Installed devices or oper-
ating practices, or both, shall be capable of early detection of leaks. Measuring
equipment shall be calibrated regularly to facilitate proper measurement.”

The title of Annex E is “Recommended Practice for Liquid Hydrocarbon Pipeline

System Leak Detection.” The annex describes a practice for leak detection based on
computational methods, particularly material balance techniques. It does not exclude
other leak detection methods that are equally effective. The annex emphasizes that op-
erating companies shall comply as thoroughly as practicable with the record retention,
maintenance, auditing, testing, and training requirements, regardless of the method of
leak detection used.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    575

Leak Detection System Selection Criteria

It is essential that the objectives and requirements for employing the leak detection
system are defined. The objectives of the leak detection system are to assist the pipeline
operators with:

·· Reducing spillage of product and thus reducing the consequences of leaks,

·· Reducing operator’s burden by detecting leaks quickly and consistently with-
out relying heavily on operator experience,
·· Satisfying regulatory requirements.

A leak would be initially detected and located by the leak detection system and
then confirmed by some means such as visual inspection. After, or even before the
leak is confirmed (depending on the company’s leak response procedures), the leak
must be isolated by closing block valves adjacent to the leak. After the leak is iso-
lated, a significant volume of product can be lost depending on the leak location
and terrain of the leaked pipeline section. The spillage during the detection phase
is often relatively small compared to potential total spillage. Therefore, the impor-
tance of rapid detection time as a valuable feature of a detection system cannot be
over-emphasized.
It is important to define a set of selection criteria for use in assessing the perfor­
mance and selection of various leak detection systems. Typical performance criteria
are listed in Table 9-1 [12, 13]:

Table 9-1. Leak detection system performance criteria

Criteria Description
Detectability Detectability of leaks is measured in terms of leak detection time and range of
leak size.
Sensitivity Sensitivity is defined as the minimum leak size that the leak detection system
can detect.
Reliability Reliability of a leak detection system is defined in terms of false alarm rate. If
the frequency of false alarms is high, the operators may not trust the leak detec-
tion system, increasing the confirmation time and thus spillage volume.
Robustness Robustness is defined as a measure of the leak detection system’s ability to
continue to operate and provide useful information in all pipeline operating
­conditions.
Operability The leak detection system needs to operate not only continuously but also in all
operating conditions (shut-in, steady state and transient state). In addition, the
system should not interfere with normal operations.
Accuracy Accuracy is defined as a measure of the leak detection system’s ability to esti-
mate how close the estimated leak location and size is to the actual leak location
and size.
Cost The cost includes the installation and operating costs of a leak detection system,
including instrumentation or sensing devices.

An effective leak detection system helps pipeline operators mitigate the risks and
consequences of any leak. It can shorten leak detection time, increase reliability (not
miss actual leaks and at the same time not produce false alarms), and reduce leak con-
firmation and isolation time with accurate leak location estimates. Simply put, overall
cost of a leak can be reduced using an effective leak detection system. However, there
are costs to implement and operate a leak detection system.
Therefore, the decision-making process of implementing and operating a leak de-
tection system can be made by balancing the risk and consequences of possible leaks
against the cost of a leak detection system and mitigation program. The following

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 576    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

p­ rocess may help in analyzing potential risks of leaks in terms of cost and the cost sav-
ings resulting from the implementation of a leak detection system:

·· Estimate likely probabilities of various leaks and thus the potential number of
leaks.
·· Estimate the direct and indirect costs of leaks over a period of time without
a leak detection system by using historical data for the consequences of the
leaks.
·· Assess attainable leak detection performances of various leak detection sys-
tems by applying the above criteria.
·· Determine the costs of implementing and operating these leak detection sys-
tems over the same period of time used in the cost calculation above.
·· Estimate potential cost savings from the use of a leak detection system.

9.2.2 Overview of Leak Detection Techniques

Broadly, there are three different types of leak detection methods: Inspection Methods,
Sensing Devices, and Computational Pipeline Monitoring Methods.
9.2.2.1 Inspection Methods
In general, inspection methods provide very accurate, sensitive and reliable results.
Particularly, ultrasonic and magnetic inspection techniques can detect both actual and
incipient leaks by determining the pipe wall thickness. However, internal inspection
methods are very expensive requiring specialized tools and expertise, and a pipeline
cannot be inspected continuously. Due to the nature of intermittent operation, only
leaks that occurred prior to the inspection will be detected and any occurring after
will remain undetected until the next inspection. Inspection techniques include visual
inspection, magnetic flux technique, ultrasonic technique, hydrostatic test, and others:
Visual Inspection — Current visual inspection methods rely on detecting hy-
drocarbons along the pipeline right of way either visually or by using an instrument.
Spillage evidence includes spilled hydrocarbons, vegetation changes caused by hydro-
carbons, odor released from the pipeline, or noise generated by product escaping from
a pipeline hole. For inspecting transmission lines, pipeline companies often use an
inspection airplane equipped with hydrocarbon detection sensors and cameras.
Magnetic Flux Technique — Strong magnets are mounted on a magnetic inspec-
tion pig. When a strong magnetic field is applied to steel pipe, magnetic flux is formed
in the pipe. If the pipe is uniform, so is the resulting magnetic flux. If the magnetic flux
is distorted, the magnetized pipe may contain defects. Since changes in magnetic flux
induce electric current, transducers measure the induced current. A magnetic inspec-
tion pig can detect pipe defects reliably and locate them accurately. It can run without
interrupting normal pipeline operations. In general, it can produce a wealth of informa-
tion for detailed defect assessment. However, a magnetic inspection pig tends to miss
longitudinal defects and cracks, and is expensive to purchase or operate.
Ultrasonic Technique — Pigs mounted with high-frequency ultrasonic equip-
ment are used to inspect internal and external defects and pipe welds on manufactured
pipes and operational pipelines. An ultrasonic inspection tool can detect small defects
accurately under clean conditions assuming that it is well coupled with the pipe sur-
face. This technique does not interfere with normal pipeline operations nor adversely
affect the pipeline system safety. However, it is sometimes difficult to maintain good
coupling between the transducer and pipe wall. Recent advancements with these tools
have resulted in the ability to detect stress corrosion cracking (SCC).
Hydrostatic Test — Hydrostatic testing must be performed on new pipelines,
as specified in ASME B31.4 and other standards, prior to in-service use. The main

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    577

a­ dvantage is that it can detect not only incipient failure but also existing small pin
hole size leaks. However, non-critical cracks may not be found and crack growth can
­accelerate due to pressurizing at the time of testing.
Hydrostatic testing was also used on operating pipelines to assess their structural
integrity. It is another method of identifying SCC problems. When an operating pipe-
line is tested at a pressure above normal operating pressure using the fluid normally
transported in the pipeline, it is called a dynamic pressure test. The purpose of this test
was not to accommodate the increase in operating pressure level, but to confirm the
pressure capability of the pipeline system. However, a pressure test can be destructive
if a line break occurs during the test. As a result, pressure testing of operating pipelines
with hydrocarbon liquids is not allowed by code and is not practiced anymore. Proce-
dures for hydrostatic testing and ILI are outlined in API Standard 1160 — Managing
System Integrity for Hazardous Liquid Pipelines.
9.2.2.2 Sensor Methods
Sensing Devices can be used to continuously sense particular characteristics of leaks
such as sudden pressure drop, noise, electrical impedance, or other signals caused by a
leak or interference around a pipe. Some sensing devices can detect not only leaks but
also third party interference around the pipeline system. Traditionally, these techniques
have been relatively unreliable and impractical. There are a few emerging technologies
in sensing devices such as fiber optics that are showing increasing promise. Certain
techniques such as specialized fiber optic cables can be expensive for existing pipe-
lines, as the pipeline has to be retrofitted with the cable or sensing devices.
Acoustic Sensing Device — The figure above shows the schematics of this tech-
nique. The principle of this technique relies on the fact that when a fluid passes through
a hole under high pressure, the resulting turbulence creates acoustic pressure waves
that travel through the fluid and pipe. Acoustic sensors are placed on the pipe, regularly
spaced along the pipeline, to detect these acoustic waves. An acoustic leak detection
system continuously monitors the pipeline for the sound characteristic of a leak. The
signals, after the background noise including operation characteristics are filtered, are
compared to the appropriate thresholds to confirm or reject a leak. The acoustic leak
detection system can also determine the leak location by correlating the sensor spacing,
velocity of sound, and propagation time difference.
The advantages, if it is installed properly, include the detectability of small leaks
in a short time, accurate leak location and continuous operation. However, it tends to
generate frequent false alarms particularly for small leaks in the presence of large back-
ground noise in the pipeline and can be expensive for a long transmission line, because
of the need for many acoustic sensors (Figure 9-13).
Optical Fiber Sensor System — This is an emerging technology that uses an
optical fiber sensor to detect leaks and/or impending pipe damage [14]. It requires the
installation of an optical fiber cable along the entire length of the pipeline. It operates
by detecting optical properties, temperature change, and/or micro bends of the pipe.
The latter capability allows the detection of activities outside the pipeline which can
be picked up by micro-strain sensors.
The advantages of this system, if it is installed properly, include the capability of
detecting and locating outside third party damage and fluid theft as well as continuous
operation. However, the installation cost on an existing pipeline can be high. Even
though several successful trials were reported [15], its performance has not yet been
fully proven for long pipelines.
Vapor Monitoring System — A vapor monitoring leak detection system [16]
detects leaks by placing a sensor tube next to the pipeline. In the event of a leak, the
hydrocarbon vapors will diffuse into the sensor tube. Its operation is shown in the figure
below (Figure 9-14).

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 578    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Leak Detection - Communication processor

Computer - Monitor

Communications Link

Signal Signal Signal Signal

Processor Processor Processor Processor

Acoustic Acoustic Acoustic Acoustic

Sensor 1 Sensor 2 Sensor 3 Sensor 4

Figure 9-13. Schematics of acoustic pressure wave technique

The system consists of a suction pump, gas detector and a plastic cable or tube that
is installed adjacent to the pipeline. When a leak occurs, some hydrocarbon molecules
diffuse into the tube as a result of the hydrocarbon concentration difference between
the inside and outside of the tube around the leaking section. In due course, the affected
area of the tube will have a higher hydrocarbon concentration than the rest of the tube.
When the pump pulls the air, the affected air is also pulled toward the detection unit,
which analyzes the hydrocarbon concentration. Because the air is pulled at a constant
speed, the system can determine the leak location. Leak size can be estimated from the
concentration of hydrocarbons.

Monitored pipe

Permeable
sensor tube
Clean dry air
Pump Sensor Electrolysis
cell
Gas Test peak
concentration Leak signal (hydrogen)

Arrival time
of leak signal
Arrival time
of test peak

Figure 9-14. Vapor monitoring method

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    579

This method of leak detection and location can detect a very small leak and locate it
accurately. It can be used for both onshore and offshore pipelines as well as multiphase leak
detection. In addition, this methodology can be used to detect many different substances. A
system based on this technology has been used in an Arctic pipeline (Northstar Develop-
ment). However, this method may be too slow to react to large leaks, and the installation
and operation costs can be very high. This system should be used in conjunction with other
leak detection systems in environmentally sensitive areas.
9.2.2.3 Computational Pipeline Monitoring (CPM) Methods
The CPM methods are based on mathematical or statistical computations of certain
quantities using commonly available measured values such as flows and pressures ob-
tained through the host SCADA system. Each scan, a CPM system receives an updated
set of SCADA data and sends a set of the modelled results back to SCADA through the
SCADA interface software. In general, the cost is relatively reasonable but the sensitiv-
ity is lower than other methods.
Any pipeline monitoring system that continuously checks for leaks can be con-
sidered a real-time leak detection system. All CPM methodologies are classified as
real-time leak monitoring systems. Real-time leak detection as discussed in this section
includes only the methods based on leak detection software operating in conjunction
with a host SCADA system.
Any CPM system consists of the following components:

·· Field instrumentation and RTU which sends the field data to the host SCADA
·· SCADA system, which collects the field data, sends them to the real-time leak
detection system, and annunciates event and alarm messages. The SCADA sys-
tem requirements for leak detection are discussed in ref. [1].
·· Hardware and software interfaces which integrate the functions of the host
SCADA and real-time leak detection system
·· Real-time leak detection computer and software

The key advantage of the CPM methods is that they seldom need additional in-
struments and equipment to those that already exist for normal pipeline operations. As
a result, the implementation and operating costs are typically lower than the costs for
inspection and sensor methods.
API Publication 1130 defines the following eight CPM methodologies:

·· Line balance technique

·· Volume balance technique
·· Modified volume balance technique
·· Compensated mass balance technique
·· Real-Time transient model (RTTM) method
·· Flow/pressure monitoring method
·· Acoustic/Negative pressure wave method
·· Statistical techniques

The first five methodologies are based on mass balance principle and will be dis-
cussed in that context. The mass balance principle applied to a pipeline means that the
difference between the amount of fluid that enters and leaves the pipe over a given time
must be the same as the change in fluid inside the pipe over the same period of time.
This principle is expressed mathematically as follows:

Imb = Vin - Vout - DLP (9 – 1)

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 580    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

where
     Vin = corrected volume entering the pipeline over a fixed time interval
  Vout = corrected volume leaving the pipeline over the same time interval
DLP = change in line pack over the same time interval
Imb = imbalance

The non-zero imbalance can be attributed to a number of factors including mea­

surement errors and line pack calculation errors. It is interpreted as a leak if the imbal-
ance is positive beyond a predefined limit, and as an unaccounted flow if it is negative,
assuming that the measured flows are accurate. The flows that go into and out of the
pipe are measured quantities and line pack changes are calculated quantities. Depend-
ing on how line pack changes are treated, the mass balance method has several forms.
Line Balance — Line balance methodology uses only the measured flows into
and out of the pipeline system, ignoring line pack changes. It assumes that a leak may
have occurred if more fluid enters the pipeline than leaves the pipeline over a certain
time period. Since the inlet and outlet flow rates are measured at different pressures and
temperatures, they need to be corrected to base conditions and then compared. If pres-
sure and temperature measurements are not available, leak detection is less accurate.
The line balance methodology is very simple to implement and the computational
requirement is not extensive. It is a suitable method for pipeline systems with very
small pipe size and short pipe length, assuming that reliable flow measurements are
available at both injection and delivery ends. This technique is less popular now, be-
cause it does not offer much advantage over the other techniques.
Volume Balance — This technique uses both the flow difference and line pack
change terms in the mass balance equation, roughly compensating for the difference
between the volumes into and out of the pipeline with line pack changes over a cer-
tain time period. Imbalance beyond a set limit is interpreted as a potential leak. Line
pack change depends on fluid properties such as compressibility and thermal expan-
sion, pressure, and temperature, and pipe data such as size and length. Since line pack
change is a dynamic quantity, the fluid compressibility and pressures are the most
important parameters in calculating line pack change. Even though pressure changes
in a transient condition, it is assumed that the pressures are in a steady state at the time
of the pressure measurements. Because of the steady state assumption, the calculated
line packs and its changes under transient operations are not very accurate.
At a minimum, this method requires flow and pressure measurements at all fluid in-
jection and delivery points. Short-pressure measurement spacing helps to reduce error in
line pack calculation, but no additional instrumentation is required. The volume balance
technique is better suited to heavier hydrocarbon liquid pipelines with relatively small line
pack and short-pressure measurement spacing, because the line pack calculation error can
be minimized. Its implementation is simple and its operation is reliable. However, it has
limited leak detection capability for long pipelines with highly compressible fluid in which
the line pack changes significantly. Figure 9-15 illustrates a trend of volume balance, which
is composed of the volume in and out difference and the estimated line pack change.
Modified Volume Balance — The modified volume balance method is a modi-
fied version of the volume balance method; it differs in that it uses a more accurate
accounting of product properties. The method tracks batches along the pipeline and
calculates the average bulk modulus dynamically in each pipe segment. The segment
bulk modulus is applied to calculate the segment line pack and its changes, and the
whole line pack and its changes and then all changes are added together.
Most comments made for the volume balance method are valid for this method,
except that its implementation is not so simple but it generally calculates the line pack
change more accurately than the volume balance particularly for batch pipelines.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    581

Figure 9-15. Trend of volume balance

Compensated Volume Balance — The compensated volume balance method is

an enhanced version of the modified volume balance. It calculates temperature profiles
along the pipeline by solving an energy equation with the temperatures at the injection
as a boundary condition. The method takes into account the fluids movements includ-
ing batch, fluid blending, and product characteristics. To reduce line pack calculation
error during transient operations, a filtering technique is applied to line pack changes.
Most comments made for the volume balance method are valid for this method,
except that its implementation is more complex whereas it generally calculates the line
pack change more accurately than the volume balance particularly for batch pipelines
and light hydrocarbon liquids such as propane and ethane. Also, this method may be
able to estimate a leak location if the pipeline state after the leak reaches a steady state
condition.
RTTM — The Real-Time Transient Model (RTTM)-based leak detection meth-
odology performs the functions of determining the pipeline state in terms of flow,
pressure, temperature and density profiles based on real-time data and then detecting
anomalies of pipeline state including leaks. API Publication 1130 defines the Real-
Time Transient Model-based leak detection methodology as follows:

“The fundamental difference that a RTTM provides over the CMB method is that
it compares the model directly against measured data, i.e., primarily pressure and
flow) rather than use the calculated values as inputs to volume balance. Extensive
configuration of physical pipeline parameters (length, diameter, thickness, pipe
composition, route topology, internal roughness, pumps, valves, equipment location,
etc.), commodity characteristics (accurate bulk modulus value, viscosity, etc.), and
local station logic (e.g., pressure/flow controllers) are required to design a pipeline
specific RTTM. The application software generates a real-time transient hydraulic
model by this configuration with field inputs from meters, pressures, temperatures,
densities at strategic receipt and delivery locations, referred to as software boundary
conditions. Fluid dynamic characteristic values will be modelled throughout the
pipeline, even during system transients. The RTTM software compares the mea­
sured data for a segment of pipeline with its corresponding modelled conditions.”

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 582    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Theoretically, the RTTM approach of real-time modelling and leak detection

can provide the most accurate modelling and leak detection sensitivity results. Most
RTTMs can provide a wealth of information on the pipeline state. In practice, however,
real-time data quality and availability are often not sufficient for reliable operation
of this leak detection approach, and certain values such as viscosity are not readily
measurable on-line. In addition, modeling in transient conditions sometimes increases
uncertainty when data quality is questionable. The main disadvantages include longer
installation time, requirements for a high level of expertise to operate, and strong de-
pendency of measurement quality. As a result, many companies have attempted to
make this methodology work in actual operations with limited success.
Pressure/Flow Monitoring Technique — This technique is used on liquid pipe-
lines to indicate unusual conditions and potential rupture conditions. This monitor-
ing methodology monitors rapid or unexpected changes in pressure and/or flow rate,
depending on their availability. There are four types of pressure/flow monitoring
techniques used on liquid pipelines to indicated unusual conditions and potential leak
conditions:

·· Pressure/Flow Limit Monitoring — ensures that measurements stay within pre-

defined operating conditions and emergency limits.
·· Pressure/Flow Deviation Monitoring — ensures that measurements stay within
a predefined tolerance of an expected operating value. Often, separate deviation
limits are established for active and inactive conditions and for positive and
negative deviations.
·· Pressure/Flow rate of change (ROC) Monitoring — ensures that any rapid
measurement change, above a predefined value per defined time period, is
made known. Often, separate ROC limits are established for the positive and
negative directions.
·· Pressure/Flow ROC deviation — modified version of the Pressure/Flow ROC
Monitoring, that projects expected ROC values during transient conditions.
Often, separate ROC deviation limits are established for positive and negative
directions.

Mathematically, a projected value is expressed in terms of a linear regression to

predict the next pressure or flow rate using a specified number of pressures or flow
rates collected over a specified period. In principle, if the current measurement drops
outside a predefined threshold from the predicted value, an alarm condition is satisfied.
Normally, a second violation check is performed with the next value in order to avoid
generating frequent alarms. If a second consecutive violation is detected, pressure and/
or flow rate violation alarms are generated.
This method is simple and easily implemented on the host SCADA system. The
main difficulties with this method are as follows:

·· Normal operations can produce rapid changes in pressure and flow rate that do
not necessarily indicate a leak.
·· Pipeline pressure increases can mask a leak.

This method may be useful for detecting unusual events or ruptures. For leak
detection purposes, it is normally used in conjunction with other leak detection
methods.
Acoustic/Negative Pressure Wave Technique — This technique works similarly
to the acoustic sensing technique, except that pressure sensors are used instead of
acoustic sensors. API Publication 1130 defines this method as follows:

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    583

“The acoustic/negative pressure wave technique takes advantage of the rarefaction

waves produced when the commodity breaches the pipe wall. The leak produces
a sudden drop in pressure in the pipe at the leak site which generates two negative
pressure or rarefaction waves, travelling upstream and downstream. High response
rate/moderate accuracy pressure transmitters at select locations on the pipeline con-
tinuously measure the fluctuation of the line pressure. A rapid pressure drop and
recovery will be reported to the central facility. At the central facility, the data from
all monitored sites will be used to determine whether to initiate a CPM alarm.”

This technique provides accurate leak location and rapid detection of relatively
large leaks, assuming that the leaks occur rapidly, the sensor spacing is close and oper-
ating pressure is high. However, it tends to generate frequent false alarms.
Statistical Technique — This technique is not a leak detection technique in itself,
but a decision tool based on the data generated by some of the above methods. Two
statistical techniques specifically applied to leak detection are described below. Even
though it is time-consuming to determine all possible “no leak” conditions for all pos-
sible operating scenarios, several successful implementations have been reported.
Sequential Probability Ratio Test (SPRT) Technique — A statistical leak de-
tection method applied to this problem is a sequential probability ratio test (SPRT)
technique to determine an alarm status. It provides a means of making a leak alarm de-
cision by analyzing time series data statistically. For pipeline leak detection, the SPRT
is applied to the time series data of the volume imbalances or flow differences.
To achieve reliable and sensitive leak detection performance, test values such as
imbalance data, should be reliable and the statistical parameters properly set during
a tuning process. A sufficient amount of normal operational data must be analyzed
in order to obtain the correct statistical tuning parameters. The tuning parameters in-
clude the number of time series data points, probabilities that determine the thresh-
olds, leak sizes to be detected with minimum standard deviation, and mean value
correction.
The SPRT offers good fault detection capability including pipeline leak detection.
The sequential probability ratio test expression includes the standard deviation and
mean value terms that indicate variability of the incoming data and inherent measure-
ment bias. Therefore, the equation automatically takes into account the pipeline opera-
tions in terms of changes in test values and bias correction. This technique responds
to changes quickly, and if properly tuned, it can provide sensitive and reliable leak
detection capability.
However, successful operation of the SPRT technique requires that the smooth
time series data to be tested be reliable. Since it relies on other calculation methods for
its test values such as volume imbalance, the selection of a proper imbalance calcula-
tion method is an important factor in achieving good leak detection performance. In
general, the SPRT tends to use a lot of test data for proper trending analysis, and thus
it may respond too slowly to respond to pipeline ruptures that require immediate leak
detection and confirmation.
Bayesian Inference Technique — Another statistical approach to leak detection
uses a Bayesian inference technique in order to make a leak/no leak alarm decision.
In other words, assuming known prior probabilities of no leak for a set of no leak pat-
terns, the Bayesian inference technique applies the Bayes’ rule to determine the prob-
ability of a no leak alarm condition. The same Bayes’ rule is applied to a leak condition
to determine the probability of a leak occurring. This technique has been success-
fully implemented in a simple pipeline, because prior probabilities can be determined.
However, it is not simple to apply it to complex pipelines with several pump stations
because it is much more complex to determine prior probabilities.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 584    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Initially, a large number of operation scenarios including leaks are simulated or

alternatively past operational data is used off-line to obtain the leak and non-leak pat-
terns. This data becomes the basis for patterns in the pipeline system. These patterns
are then refined with actual operation data obtained while the system is operating in
real-time. As a result, reliability and detection sensitivity can be improved as more
operating data is accumulated over time and used to refine the pipeline state and
p­robabilities.
A Bayesian inference method has been successfully applied for fault detection,
and this pattern recognition technique, at least in theory, can be applied to any pipeline
operation. For pipelines with a simple configuration, with no more than one inter-
mediate pump station, the probability of a leak and no leak condition may be easily
determined. For complex pipeline systems, however, it is time-consuming to determine
prior probabilities and establish pipeline operation patterns. Since the Bayesian infer-
ence technique needs to build an accurate probability database for almost all possible
operations, extensive field and maintenance tuning efforts are required for reliable
operation; this may take a long time to acquire for a complex pipeline system.

9.2.3 Implementation and Operation

Whatever the leak detection method or methods used, the implementation and opera-
tion issues are critical to get the full benefits of the leak detection system. For example,
installing adequate instrumentation or sensing devices is critical to the implementation
of an effective CPM or sensor-based system. For optimum performance, it is important
that the installed instrumentation or sensing devices be consistent with leak detection
requirements.
The real-time leak detection systems such as CPM and sensing methods are closely
integrated with the host SCADA. Therefore, CPM and sensing methods require an in-
terface with a host SCADA system. The interface allows all field data used and data
generated by these systems to be exchanged with the SCADA system, so that the pipe-
line operators can respond to an emergency expediently according to the company’s
emergency response procedure.
As part of the implementation phase, the commissioning and tuning tasks have to
be performed after these systems are installed on site. The following tasks are usually
performed (the tasks listed below are not necessarily required for all these systems):

·· Check the SCADA functions and interface

·· Check the instrumentation or sensing devices for their availability, accuracy
and other behaviors
·· Check and tune the performance of the installed system or systems during nor-
mal pipeline system operations
·· Check if other operation-related problems exist and correct them before the
system is put into service
·· Perform several levels of system acceptance tests to check if the system satis-
fies the leak detection requirements

The pipeline operator identifies and analyzes pipeline operation problems via the
user interface. The user interface should be easy to use and provide appropriate in-
formation in order to make correct decisions. It is critical to have accurate and timely
information in an easy-to-interpret format. The decision may include leak confirmation
and location, so that the operator can shut down the pipeline in accordance with the
pipeline’s operating and alarm conditions. A proper response must be made quickly
when an emergency such as a leak occurs. In addition, the user interface should be

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    585

Figure 9-16. Example display of leak alarm messages

consistent with the SCADA system, because the operator gets the information through
the SCADA. Alarm messages are the most critical information that the operators must
pay attention to, and a pipeline map is required to efficiently execute emergency re-
sponse procedures. The map gives detailed information about the names and contact
numbers of responsible parties, pipeline route and terrain, population close to the pipe-
line route, responsible officials including police, critical environmental concerns, etc.
An example display of leak alarm messages is shown in Figure 9-16. It shows the
leak alarm status, estimated leak location and size, and other information that helps the
operator to quickly identify the potential problem.
Alarm messages are critical information that the operations staff must pay atten-
tion to. It is strongly recommended to display alarm messages including leak detection
alarms on the SCADA alarm display screens. The following features and qualities
should be part of the alarm displays:

·· Consistent with SCADA system alarms and have an appropriate priority.

·· Have different colors for each category of alarm.
·· Acknowledged and unacknowledged alarms should be accessible to the pipe-
line operator in one step. Acknowledged alarms still in the alarm state should
remain readily available to the pipeline operator.
·· Have a time stamp as part of the displayed alarm.
·· Should have both audible and visual cues. Each alarm should have a unique
audible tone. Visual cues for any given alarm should persist for a long enough
period of time so as not to be overwritten irrevocably by newer alarms.
·· Not easily defeated, or inhibited without just cause. The use of screen savers or
any other screen blanking is strongly discouraged.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 586    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Trending measured and calculated values of the SCADA and CPM system help
determine what caused an alarm. Trending may be in graphical and tabular forms:
Graphical presentation makes it easier to identify anomalies. The tabular form is use-
ful for analyzing data in detail. API 1130 suggests that a trend cover a long enough du-
ration to see values before a CPM alarm occurred and continue right through to when
the alarm ends, or the current time. The following values need to be trended:

·· Measured pressures and temperatures

·· Measured densities, particularly for batch pipelines
·· Measured flow rates and their differences between inlet and outlet flows
·· Calculated line pack changes if they are made available
·· Imbalances for the CPM methodologies using mass balance principle

Effective operation of a leak detection system requires a thorough understanding

of the system installation, operation, capabilities and maintenance. The pipeline opera-
tions staff must have extensive training including practical on-line operating experi-
ence. Emphasis is placed on how to operate the system effectively and how to analyze
the results accurately. The pipeline operators and other operation engineers should
learn the skills necessary to effectively monitor the system and diagnose anomalies and
to effectively manage and maintain the system.
A leak detection system manual should be readily available for reference by those
employees responsible for leak detection on the pipeline. The manual may contain the
following information:

·· A system map, profile, and detailed physical description of each pipeline segment
·· A summary of the characteristics of each product transported
·· A tabulation of the measurement devices used in the leak detection procedure
for each pipeline segment and a description of how the data is gathered
·· A list of special considerations or step-by-step procedures to be used in evaluat-
ing leak detection results
·· Details of the expected performance of the leak detection system under normal
and line upset conditions
·· The effects of system degradation on the leak detection results

API 1130 recommends on-going testing — establishing the policy of periodic test-
ing, test frequency, and test methods. Testing methods may include removal of fluid
from the pipeline, if permitted. The main purposes of the testing are to check if the
installed system is effectively operating and to test whether operators follow the com-
pany’s emergency response procedure. Effective emergency response is one of the key
tasks for mitigating the consequences of the leak when a leak is detected. Emergency
response procedures must not only be clearly written but understood and practiced by
pipeline operating staff.
API 1130 recommends keeping design records, software changes and test records, and
specifies the record retention length. Records of tests should include the following:

·· Date, time and duration of the test

·· Method, location, and description of the commodity withdrawal
·· Operating conditions at the time of the test
·· Analysis of the performance of the CPM system and, for tests, the effectiveness
of the response by operating personnel
·· Documentation of corrective measures taken or mitigated as a result of the test
·· SCADA data generated during the test

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    587

It also recommends that records detailing the initial or retest results should be
retained until the next test.
To maximize performance of the implemented CPM system, operating companies
need to establish a procedure and schedule for maintaining all instruments, communication
tools, and hardware and software that affect the leak detection system. Refer to API 1130
which describes several aspects of a system maintenance and support program.

9.2.4 Leakage Response

A pipeline leak, particularly a large leak, is treated as an emergency. At the event of a
leak, line pressure should be reduced as fast as possible in the leaking section to mini-
mize spillage caused by the leak. The leak detection system should be able to identify
at least the leak section. To reduce the line pressure, the originating station should be
taken off quickly, and then other stations should be dropped.
If possible, the operator should shut down the closest station upstream of the leak
section and keep the downstream station on as long as possible to pull pressure down at
the leak site. Practically, the operator is able to shut down the station downstream of the
leak after reducing the suction set point as low as possible. This allows the operator to
monitor line pressure between the stations. When the line pressure drops significantly,
the leaking section should be isolated by closing the upstream sectionalizing or isolat-
ing valves first and then the downstream valves.
Whenever a line is shut down due to either a confirmed or suspected leak, clear-
ance must be obtained prior to starting the line back up. All relevant data must be
retained along with any other information that would help analyze the causes of the
leak and facilitate the leak report preparation. This report is also required for the man-
agement and for the next shift operator in case line start-up is delayed.

9.2.5 Summary
A leak detection system is a tool for mitigating the consequences associated with a leak
by fast but reliable detection and accurate location. The operator should be well trained
in using the leak detection system so that any emergency due to a leak can be effec-
tively managed. If a leak detection system is effective, it can be good insurance for
reducing risks. An appropriate leak detection system should help pipeline companies
operate their pipeline systems safely by reducing the consequences related to a leak.
A SCADA system is an integral part of daily pipeline operations. The CPM and
sensing methods of leak detection take advantage of real-time capability and the
effectiveness of the SCADA system as a monitoring and controlling tool. As the his-
torical data indicates, the current CPM technologies are far from satisfactory in their
performance. They need further improvement in their reliability and leak detection
sensitivity. Also, a single CPM system may not satisfy all the criteria of an effective
leak detection system. Combining a few CPM and sensing methodologies, however,
may be able to satisfy not only most regulatory requirements but also effectiveness
criteria.

9.3 DRAG REDUCING AGENT (DRA)

9.3.1 Introduction
A drag reducing agent (DRA) is primarily used to increase the capacity of petroleum
liquid pipelines. It is also known as drag reducer, flow improvers, or just as DRA.
It is reported [17] that the Trans-Alaska Pipe Line System (TAPS) started to use a

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 588    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

commercially available DRA since 1979, after its effectiveness in increasing through-
put was proven. Nowadays, many North American pipeline companies use DRA to
increase throughput well beyond their design capacities. In addition to oil transmission
pipelines, DRA has been applied to district heating and cooling systems.
Different types of DRA are used in these systems. Most DRA products consist
of a long chain and high molecular weight polymer which is injected into the petro-
leum liquid in small amount (in the range of 10 to 30 ppm of the petroleum liquid)
for reducing the frictional pressure drop. It is injected into the petroleum liquid
flow stream to raise line throughput or lower line pressure to maintain the line
within operating parameters while increasing flow for a given power input. There-
fore, it can save power costs and also relieve the pipeline company of ­capacity
restriction.
Main classes of additives and chemicals used for transport are summarized in
Table 9-2 [18].
Table 9-2. Main classes of additives and chemicals
Additive Remark
Drag reducer For both liquid and gas transport
Corrosion inhibitors Both chemical inhibitors and biocides
Paraffin inhibitors Prevent paraffin depositions
Pour point depressants Lower pour point of waxy crudes
Gas hydrate inhibitors For gas and multi-phase transport to prevent hydrate
deposits
Surfactants For multi-phase transport of heavy crudes
Odorizing additives For safety

9.3.1.1 Drag Reduction Mechanism

Drag reduction is a phenomenon in which the friction of a liquid flowing in a pipeline
in turbulent flow is decreased by using a small amount of an additive. The DRA is
believed to damp flow turbulence of petroleum products near the pipeline wall, so
that friction and thus frictional pressure drop can be reduced. This dampening effect
reduces frictional pressure drop, reducing energy consumption or increasing flow rate.
Its concentration in the liquid affects the turbulent characteristics of the liquid in the
pipeline. DRAs do not coat the pipe wall or change fluid properties such as density and
compressibility.

Figure 9-17. Mechanism of drag reduction

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    589

In most petroleum pipelines, the liquid flows through the pipeline in a turbulent
flow regime. In this flow regime, the molecules move in a random motion. As shown in
Figure 9-17 [19], a turbulent flow regime has three flow zones; laminar sub layer, buffer
region, and turbulent core. Working with the molecules in the second and third zones,
a DRA can reduce the energy waste caused by the random motion. In other words,
drag reduction takes place through an interaction between DRA and the turbulence of
flowing fluid.
The DRA is effective only for reducing friction and thus friction pressure loss.
Since the total pressure drop is caused by both friction and elevation gain, it is benefi-
cial only in the sections where the frictional pressure loss is significant but not useful
where the primary pressure drop is caused by elevation gain.
Figure 9-18 [20] shows that lowering these internal fluid pressure losses increases
the bulk throughput of the pipeline for a given pumping energy, hence operating costs
are reduced.

Figure 9-18. Effect of chemical drag reducers on pipeline pressure and flow

9.3.1.2 Benefits of Using a DRA

There are several options to increase the pipeline capacity; installing additional pump
stations, adding parallel loops, and increasing the pipe diameter. However, these op-
tions are capital intensive and time-consuming. The key benefits of using a DRA are
operating cost saving and flexibility in dealing with throughput with relatively small
investment in facility. Specifically, the following benefits can be realized:

·· The construction of new pump stations is not required to satisfy throughput

increase, especially if the increased throughput requirements over the design
capacity do not occur frequently.
·· Significant increase in pipeline throughput can be handled with relatively small
investment in the construction of DRA facilities. In some instances, it is re-
ported that the throughput increase is expected to be more than 30% [21].
·· Significant throughput can be maintained even for a de-rated line during main
pump maintenance.
·· Energy cost savings, particularly if the power cost is high.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 590    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

·· The DRA facility construction time is shorter and the cost is significantly lower
compared to the time and cost for the construction of pump stations.

However, the cost of DRA is relatively high, so a break-even analysis has to be

performed to justify the DRA facility installation and operation.

9.3.2 DRA Characteristics and Performance

To be effective, the DRA should have the following characteristics [22]:

·· Solubility: The DRA must be soluble within the fluid being pumped.
·· Shear stability: The DRA must be stable enough not to break down in turbulent
pipe line flow.
·· Downstream effects: The DRA should have no downstream effects on refinery
equipment when used in crude pipe lines and, on product systems, no effects
on engine performance.

A DRA can be degraded due to shearing while flowing through the pipeline and
f­ acilities such as pump and valve stations. The degradation through pipes is roughly
­proportional to travelling distance. The degradation of DRA through the facilities may
be caused by the operating pumps as well as the station piping and fittings. The deg-
radation due to the station piping and fittings may be small if the station is bypassed,
but it can be larger if the DRA flows through the pump station. The degradation due
to the pumps may be different for each pump, and the degradation for each pump in
series is larger than that for a single operating unit. Field data indicates that the DRA
passing through a running pump degrades roughly by one third of its value. Therefore,
the DRA degrades roughly by two thirds if a pump station is operated with two pumps
in series.
Several properties of the fluid being pumped can also affect DRA performance as
follows:

·· Turbulence: The pipeline must be operated at turbulent flow conditions for the
DRA to be effective. Most DRAs are not effective for heavy crude transporta-
tion flowing in laminar flow conditions [23].
·· Viscosity: Decreasing the viscosity of the fluid increases the effectiveness of
the DRA.
·· Temperature: Increase in the temperature of the fluid decreases the fluid viscos-
ity and increases the solubility of the DRA, thus improving performance.
·· Wax or water contents: It is indicated that high wax or water contents in the
fluid reduces the effectiveness of the DRA.

9.3.3 DRA Operations

9.3.3.1 DRA Facilities
Figure 9-19 below shows an example of a test installation for application of the new
ExtremePower™ DRA (warm climate). Some additional equipment would be required
for cold climate installation.
A DRA injection facility is constructed of injection and feed pumps, flow meter, stor-
age tank, pressure sensors, control equipment, and equipment for safety such as pressure
relief valves. Figure 9-20 shows a DRA injection system. Multiple tanks are normally used
for DRA storage, and gear pumps driven by variable speed motors for injecting it. The DRA
injection rate is measured by a positive displacement flow meter.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    591

Figure 9-19. Example of field test facilities (courtesy of ConocoPhillips Specialty Products,
www.liquidpower.com)

9.3.3.2 DRA Injection

The effectiveness of DRA is measured in terms of the reduction in frictional losses
in the pipeline. It varies with the DRA concentration, viscosity of the solvent fluid,
pipeline temperature, fluid velocity, and pipeline diameter. Since a DRA is composed
of long polymer strands, the DRA can be sheared when it passes through the pipe-
line and equipment such as pumps and control valves. This results in degradation of
its effectiveness. The DRA effectiveness depends on the length of the pipeline and
the amount of shearing. Therefore, DRA must be injected downstream of all pumps,
meters and valves to prevent shear degradation, requiring the DRA pump pressure be
higher than the pump station discharge pressure.

Figure 9-20. DRA injection system

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 592    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Startup of DRA Injection: The DRA injection system is usually automated and
controlled remotely by an operator from the control center. Refer to Figure 9-20, which
illustrates a DRA injection system at a pump station. In anticipation that the remote
control system may not work, the facility needs to provide local manual control capa-
bility. Described below is a normal DRA injection starting procedure:

·· Select stations where DRA is to be injected, while checking to ensure that in-
compatible product (such as jet fuel) will not be affected by wrongly injecting
a DRA into the passing batches.
·· Determine the DRA flow rate set point based on the target flow rate in order to
obtain an optimum drag reduction. The required DRA flow rate is calculated
automatically if the line flow rate is known. Normally, DRA injection initially
begins at a high flow rate, and then lowers to the required flow rate.
·· Select and start the DRA pump if there are several DRA pump units.
·· Check if the DRA flow rate agrees with the DRA flow rate set point.

DRA injection operation is shown in Figure 9-21. The DRA flow rate is control-
led to reach the DRA set point. The pump station is equipped with two variable speed
pumps (two VFD drivers) in series, and the DRA injection system is installed down-
stream of the station. A booster pump is installed to boost the suction pressure of the
mainline pump.
In order to make sure that any jet fuel batch is not contaminated with DRA, the
following steps for the startup and shut-down of DRA injection are taken:

·· DRA injection should not start about one hour after a jet fuel batch has passed
the DRA injection station.
·· DRA injection should be shut-down about an hour before the jet fuel batch ar-
rives at the injection station.
·· Where a batch tracking application is employed, DRA lockout can be triggered
by the approach of a jet fuel batch to the DRA injection station with a status
returned to SCADA that can be used to lockout the DRA pump.

Figure 9-21. DRA injection at a pump station (courtesy of Telvent)

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    593

Shut-down of DRA Injection: When a pipeline operates near the pressure op-
erating limit, shut-down of DRA injection could cause line operating pressure to be
exceeded if the flow rate remains constant. Described below is a normal DRA injection
shut-down procedure:

·· Select the station where DRA injection is to be shut down.

·· Check if DRA is no longer required to obtain line flow rate for pending batches
or to maintain the line pressure within the operating limits.
·· Select and stop the DRA pump at the station where DRA injection is shut
down.
·· Check if the DRA flow rate decreases to zero.

The above figure shows a DRA injection and shut-down control through the
SCADA screen of the pump station, where the DRA injection takes place.
9.3.3.3 DRA Concentration Tracking
The DRA concentration is measured in parts-per-million in the flowing product. The
DRA concentration is tracked as it moves down the pipeline, and the concentration in
the subsequent section includes the degradation due to moving along the pipeline and
passing through running pumps. The inclusion of the DRA will create a new ‘batch’
blended with the DRA when the DRA is being injected.
Both sheared and non-sheared DRA concentrations need to be tracked to prop-
erly operate DRA injection. A DRA injection rate is used with measured or calculated
product flow rate to calculate the DRA concentration. When a DRA passes through a
pump, it is sheared and no longer active. The DRA tracking function tracks the sheared
and active DRA concentrations and checks the concentration against the maximum
DRA concentration allowable in the product. For example, DRA is not allowed in jet
fuel and thus its concentration should be checked against zero concentration level. A
graphic view of the DRA contents within a pipeline can show active, sheared and total
concentration of DRA in the product as well as the positions relative to DRA injectors
or pump stations.
9.3.3.4 DRA Limitations on Operation and Design
If the throughput is restricted by the pipeline capacity, it is generally cost-effective
to install DRA facilities at pump stations. However, if the desired flow rate is higher
than the pump capacity, the pumping capacity must be increased to accommodate the
increased throughput requirements.
Figure 9-22 illustrates the pump operating point change due to capacity increase.
Since the throughput increases in the presence of a DRA, the existing pumps may not
be able to accommodate the flow rate increase without modifying the pump charac-
teristics. Note that the pump does not operate at the best efficiency point (BEP) when
DRA is injected into the fluid.
DRA can be used in the transportation of crude oil and refined petroleum products
(except jet fuel) in order to increase pipeline throughput. The DRA can accumulate
on turbine blades and may damage the turbine. Therefore, it cannot be used for jet
fuel transportation, not because of its effectiveness but because of its potential safety
concern. The original DRA did not work with heavy crude, but ConocoPhillips has
recently developed a DRA [23] that has proven effective for heavy crude.
Since much higher flow rate can be achieved with DRA, the flow velocity can be
fast. It has to be noted that the higher velocity can also increase the surge pressure.
Therefore, a check must be made to see if the existing pipeline can meet the new tran-
sient pressure requirements.
As the DRA is injected into the pipeline section, the throughput increase takes
place slowly because the increased rate is linearly dependent on the flow velocity.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 594    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Figure 9-22. DRA injection and changes in pump operating point

The desired throughput can be achieved only after all the liquid in the pipeline section
contains the required DRA concentration. For example, it would take about 10 hours
to reach the intended throughput in a 100 km section if the flow velocity is in the order
of 10 km an hour.
As discussed in the previous section, DRA is sheared as it moves through pipes
and particularly pumps. In other words, it has to be injected at every operating pump
station. Therefore, it can be very costly for a long pipeline with short pump station
spacing.

9.3.4 DRA Correlations

When undertaking pipeline hydraulics simulations, generally, information required is
the type of DRA used plus the injection rate in ppm (parts per million). The injection
rate depends on the following:

·· Type of DRA,
·· Supplier,
·· Pressure,
·· Temperature,
·· Pipeline product flow rate,
·· Distance travelled.

The DRA injection rate needs to be determined to achieve the desired pressure or
flow rate. DRA manufacturers provide their own DRA correlations. When studying a
DRA operation, one of the following correlations can be used: AESOP [24], Burger
[25], Conoco or Simplified Conoco correlations [26]. They relate the effective friction
factor with the DRA concentration. The AESOP correlation has been recently devel-
oped by a joint consortium of academic and industrial bodies under European Union
funding.
Aesop: This is a correlation that requires data which is specific to the fluid in a
line.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    595

Burger: This is a widely-used correlation that depends on the DRA concentration,

product viscosity, fluid viscosity, and the pipe diameter.
Conoco and Simplified Conoco: This option is a correlation that requires data
which is specific to the fluid in a line.
The drag reduction factor is a direct multiplier to the calculated friction factor.

f = fm ´ (1 – F)

where
f = effective friction factor
F = drag reduction factor
fm = model-calculated Moody friction factor

Aesop drag reduction factor

The Aesop correlation is of the form

æ ACe ö
F = Bç
è 1 + ACe ÷ø
(
1 + C ×10 -5 NRe )
Ce = e -ld ppm

where A, B, and C are AESOP coefficients

F = drag reduction factor (%)
Ce = Effective concentration of additive (ppm)
NRe = Reynolds number
d = Distance travelled (km)
gλ = Degradation coefficient

Burger drag reduction factor

æ æ ppm ö ö
0.5
ç v * çè C ÷ø ÷
F = k1 * ln ç s ÷ +k
ç d 0.2 ÷ 2
ç ÷
è ø
where
k1 and k2 = Burger equation constants
v = local fluid velocity, ft/sec
ppm = DRA concentration, parts per million
Cs = fluid viscosity, centistokes
d = pipe diameter, ft

Conoco drag reduction factor

The Conoco (CDR) drag reduction factor is of the form

ppm
F=
(a * ppm + b)
where a and b are product-specific constants and ppm is DRA concentration in parts
per million. For full details of the Conoco CDR Correlation contact Conoco Speciality
Products, Inc. This DRA correlation is only active for velocities in excess of 0.6m/s
and Reynolds Number in excess of 7500.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 596    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Figure 9-23. D
 RA concentrations and their effects (courtesy of Energy Solutions International
Inc.)

Simplified Conoco drag reduction factor

The Simplified Conoco drag reduction factor is of the form

ppm
F=
( A * ppm + B)
where A and B are product-specific constants. The equation above is used directly with
the product-specific constants entered by the user.
Figure 9-23 shows a set of typical DRA effect curves for the Conoco, Burger, and
AESOP correlations. These curves can only be regarded as typical as the drag reduc-
tion effects depend on the pipeline dimensions, fluid properties and velocity as well as
the DRA properties themselves. Note that the DRA takes effect even when the DRA
concentration level is very low, less than 10 ppm, and that drag is not reduced signifi-
cantly even when the concentration keeps increasing beyond 30 ppm.

9.4 TANK FARM OPERATION AND VOLUME MEASUREMENT

Crude oil and petroleum products, including light hydrocarbons, are likely to deliver
their products to/from tankage. These products are often stored in tanks in various
locations such as producing areas, refineries, petrochemical plants, and/or distribution
centers. Petroleum liquids are stored underground or in aboveground storage tanks.
Storage allows for flexible pipeline transportation and efficient transportation manage-
ment through the existing pipeline system and minimizes supply/delivery disruptions.
The stored liquids need to be measured and accounted for accurately in order to keep
track of all volume movements including custody transfer when required. Refer to
Chapter 8 for the detailed tank and tank farm design.
Oil and petroleum product pipelines lift their products from and deliver to tank
farms. A tank farm refers to a collection of tanks located at a refinery, shipping terminal
or pipeline terminal. A tank farm at a refining operation is used to store feedstock and
various products produced by the refinery and to hold them until they are scheduled for

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    597

Figure 9-24. Simple tank farm (courtesy of Telvent)

injection into a pipeline for transportation. Similarly, tank farms at shipping terminals
hold products until a shipping route is scheduled. The shipping route may be via tanker
ship, truck, railcar, or another pipeline. Tank farm operation covers tank control, vol-
ume measurement and inventory.
A schematic for a typical tank farm is shown in Figure 9-24. On the left hand side,
distillate and gasoline tanks are separately connected to each set of valves, which in
turn are connected to a booster pump. The distillate or gasoline is lifted through the
dedicated booster pump into the mainline pumps. The lifted fluid passes through the
flow meter and in this case the meter prover before it reaches the mainline pumps.
Any alarm or warning messages related to tank or tank farm operation are listed at
the bottom of the SCADA screen, to which the operator can respond remotely. The
operator responses may include simply acknowledging the message or taking correc-
tive action.

9.4.1 Tank Farm Operation

Tank farms are located at the flow lifting or receiving and delivery terminals as well
as side-stream injection and side-stream delivery points in the pipeline system. Full
stream injection or delivery takes place at the lifting and delivery terminals. Either par-
tial or full stream injection can take place at the side-stream injection points, while strip
or full stream delivery at the side-stream delivery point. All flows in and out of the tank
farms in the pipeline system must be measured and recorded for custody transfer.
9.4.1.1 Normal Batch Lifting Sequence at a Product Lifting Tank Farm
The first step of lifting a product is to select the tank or header from which the batch
will be supplied to the first main line pump station, and then the appropriate booster
pumps to be used for the batch.
The next step involves the valve manipulation in sequence:

·· Open the valve on the tank or header from which the batch is to be supplied.
·· Open the suction valves on the selected booster pumps.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 598    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

·· Close the valve on the booster pumps bypass lines.

·· Close the valve on the pump discharge lines.
·· Close the valve downstream of the minimum flow bypass control valves.

Once the conditions to start are satisfied, the start sequence is initiated by opening
the bypass valve on the selected booster pump(s). After the flow through the bypass
line has stabilized, the valve in the discharge line is opened to direct the flow to the
main line pump suction through the meter designated for the batch.

·· The flow rate for the batch is metered and accumulated volumes are recorded.
·· Pumping of the leading batch would have been just completed prior to the in-
troduction of the new batch.
·· Close the booster pump discharge header valve to stop lifting the leading
batch.
·· Stop the booster pumps that were being used for pumping the leading batch.
Close the booster pump suction valves and tank valves.
·· Set the flow accumulator to zero at the start of the new batch.

9.4.1.2 Operation at the Delivery Terminal

The pressure levels of tanks are low unless the tanks are for high vapor pressure liq-
uids, while the pressure level of the incoming fluid may be high. Therefore, a pressure
control valve is installed at the entrance of the delivery terminal to keep the pressure
below the maximum tank pressure.
The combination of estimated time of arrival of batches, the interface detection by
the densitometer located upstream of the delivery terminal, and the detection of batch
arrival by the densitometer is used to identify the arrival of various batches. Depend-
ing on the tank management procedures established for the delivery terminal, the batch
will be directed into designated tanks by opening and closing of tank valves.
One or more densitometers or dye detectors are provided near the delivery termi-
nal for batch interface detection. One detector will be installed a few kilometers from
the terminal in order to provide lead time for taking actions to direct batches into ap-
propriate tanks, and the other should be installed near the tank manifold.
The batch interfaces or transmix may be directed into a designated tank at the
delivery terminal. The start and finish of batch interfaces will be indicated by the den-
sitometer. It is normal practice to collect samples of the batch at frequent intervals to
confirm the start and completion of batch interface. This will ensure the purity of the
arriving batch. Any off-spec interfaces must be sent to a slop tank for separate treat-
ment such as re-processing.
9.4.1.3 Side-Stream Injection
The side-stream injection facility, including densitometers, is normally installed on the
upstream or suction side of the intermediate pump station. The upstream densitometer
provides lead time to initiate or terminate the incoming batch injection. The injection of
the batch starts or stops when batch interface arrival at the intermediate station is con-
firmed by a change in densitometer reading on the suction side of the pump station.
For partial injection, the injected product is the same as that of the incoming batch,
resulting in the reduction of the upstream flow rate while maintaining the previous flow
rate downstream of the injection point. For full stream injection, the line upstream of
the injection point is shut in, but the downstream flow rate can be maintained unless
the delivery rate is changed.
For partial injection, it is a normal practice to start injection by opening the in-
jection valve after the batch interface has passed in order to avoid any unnecessary
mixing. The flow rate at the lifting point has to be cut to compensate for the injection,

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    599

because the flow rate upstream of the injection point has to be reduced if the down-
stream flow is maintained. The pressures in the upstream section have to be adjusted to
maintain the selected injection rate. It may be necessary to take off some pump units
and/or stations right after the partial injection starts. For a full stream injection, the
mainline valve just upstream of the injection point is closed and all pump stations are
shut down in the sections upstream of the injection point.
Figure 9-25 illustrates a side stream injection into the mainline at a pump station.
A booster pump takes a batch or fluid from a tank and pumps the fluid into the mainline
pump. The flow rate is measured before it is injected into the pump.

Figure 9-25. Side stream injection (courtesy of Telvent)

9.4.1.4 Side-Stream Delivery
Normally, two densitometers are installed at the delivery location to detect batch in-
terface during batch operation. The first densitometer is installed a few kilometers up-
stream of the take-off location on the main line and the second densitometer installed
at the delivery station. The flow into the take-off is shutdown by closing the appropri-
ate valve on the take-off when the head of trailing batch arrives near the take-off as
indicated by the upstream densitometer. Downstream of the take-off point, a pressure
control valve is installed to maintain the required delivery pressure.
Normally, the take-off valve is opened when part of the stream is passing. Care
must be taken to maintain the pre-determined downstream rates to prevent stretch-
ing out interfaces between batches. Therefore, the operator has to calculate the down-
stream flow rate before the strip delivery starts. If the delivery rate is large relative to
the mainline flow, it is essential to take off some pump units and/or stations right after
the strip delivery starts.
For a full stream delivery, the operator has to open the take-off valve and close
the mainline block valve immediately downstream of the take-off connection. If the
delivery is to be full stream, the downstream section should shut down with reasonable
pressures in the section.
9.4.1.5 Break-Out Operation
The figure below shows a simple tank farm that allows a break-out operation
­(Figure 9-26).
At a certain location, a fluid breaks out full stream into tankage and other fluid is
simultaneously pumped out of another tank, called a break-out operation. The break-out

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 600    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Figure 9-26. Simple break-out tank farm

operation tankage is usually located in the middle of a pipeline, not at the lifting or delivery
terminal unless the terminal is connected to another pipeline system. The batch volume has
to be determined by providing flow accumulators on the flow meters.
In the above tank farm diagram, the incoming fluid can bypass the tank farm or
be stored in a tank. For example, a batch received from the mainline is stored in Tank
2 and the other batch in Tank 1 is injected into the mainline on the downstream side of
the tank farm.
9.4.1.6 Sump System
A sump system collects drainage from various sources such as pumps and pig traps,
and pumps them to other facilities when a sump tank is full. The system may include
a tank that collects any slop or off-spec liquids such as used lubricating oil of a pump.
Sump tanks are installed at pump stations and receiving/delivery points. When a sump
tank is getting full, the liquid in the tank is transported to a refinery or if acceptable can
be blended with other liquids being transported at the location.

9.4.2 Tank Control

A tank farm operation includes two levels; tank farm operation and tank operation.
From the perspective of the control center, tank and tank farm operations are fully
automated and controlled by a tank farm control system.
The purpose of a tank farm control system is to assist the operator in moving
products and maintaining the inventory of the products. Terminals that handle multiple
products (i.e., a batched pipeline) with a large number of tanks and interconnecting
pipelines can have quite a complicated routing within the terminal. There will be a sig-
nificant number of motor-operated valve controls and tank level monitoring systems.
The system generates and stores product delivery and shipment feed information in
business applications such as inventory tracking, billing for product receipts and deliv-
eries, as well as feeding the same information into a pipeline scheduling system.
A tank farm control system can assist the operator by verifying that proposed valve
line-ups represent a valid path before he initiates the sequencing and starts the pumps

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    601

to move the product. This ability helps avoid an error of lifting a product from or de-
livering to a wrong tank. This is important, as an error such as the injection of crude oil
into a refined product tank would be costly.
The tank control requires the tank information and several control variables. The
tank information for a tank farm includes product name, tank type (net or gross), tank
volume correction factor, maximum and minimum tank volumes, maximum and mini-
mum working level, and others. The tank volumes are normally measured in tank level,
which is converted into tank volume (see Chapter 7). The tank control variables are as
follows:

·· Flow rate: Flow rate into and out of a tank is calculated by dividing the net/
gross volume change by the elapsed time.
·· Volume-to-fill: The net/gross volume in the tank is subtracted from the maxi-
mum safe tank volume to calculate the volume-to-fill.
·· Volume-to-pump: The tank bottom volume is subtracted from the net/gross
volume to calculate the volume of liquid that can be pumped from the tank.
·· Time-to-fill/empty: The time to fill or empty the tank is calculated from the
current flow rate into or out of the tank, and the volume-to-fill or volume-to-
pump.

A tank control requires alarms and events to be generated in response to vari-

ous conditions. These may include alarms when maximum or minimum tank levels
have been violated and alarms for abnormal rates of change. Typically, the information
contained in a tank report includes such data as tank level and water level, measured
and corrected gravity, temperature, gross and net wet volume, S&W volume, net dry
volume, and flow rate (Figure 9-27).

Figure 9-27. Tank control (courtesy of Telvent)

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 602    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

The above figure displays the information on the tank status, product in the
tank, tank volume and level, and other tank control data. The tank level should be
between the maximum and minimum levels, and the level gauge is used to control the
flow moved into or out of a tank. Tank level conversion to tank volume is discussed
below.

9.4.3 Tank Volume Measurement

One way to measure the volume of a stored liquid is to determine the level of the liq-
uid in the tank and then calculate the volume from a capacity or strapping table that
relates the level to the corresponding gross volume of liquid in the tank. The strapping
table is established during the tank proving process, using a tank prover which has
thermometers mounted in the measuring section to accurately measure temperature.
The API Standard 2550, Measurement and Calibration of Upright Cylindrical Tanks,
describes the strapping procedures, and API MPMS describes the strapping procedures
for cylindrical as well as other types of tanks. See Chapter 7 for detailed description of
the calibration procedures for tankage.
Tank level to volume conversion requires that the parameters and strapping table
or equation associated with the tank be defined. In addition to the level measurement,
the gravity and suspended sediment and water (S&W) content and the temperature of
the liquid and ambient temperature near the tank need to be measured to determine the
net volume and liquid head stress caused by high hydrostatic pressure on a large tank.
The accurate calculation of the volume in the tank requires parameters such as tank
roof types (fixed or floating) and the level of free water. The volume conversion can be
performed by a field automation system device such as a PLC.
Once the tank level has been measured, whether manually or automatically, the
level data is converted to a gross volume using a volume conversion process. The
process uses either the strapping table data for each individual tank or an incremental
table that defines incremental volumes per number of level increments for the tank. The
conversion equation associated with the tank can be used for the volume conversion.
The gross volume should be corrected for tanks with a floating roof by taking into ac-
count the weight of the roof and any snow load.
The level of free water is also required to determine the gross volume of the pe-
troleum product in the tank. This value is converted to its equivalent volume using the
volume conversion table and then subtracted from the gross volume to determine the
gross volume of the product by assuming that the water is on the bottom of the tank.
A gross volume is converted to a net volume using the density and temperature of
the fluid in the tank. The density or API gravity is used to calculate the temperature cor-
rection factor, which is detailed in the API Standard 2550. Once the temperature cor-
rection factor is determined, it is multiplied by the gross volume to obtain its equivalent
net wet volume. If the sediment and water (S&W) contents are present, their values are
used to determine the net dry volume.

9.4.4 Tank Inventory

The tank inventory functions include tank calculations such as flow rates and volume
conversions, volume validation and correction, floating roof adjustments and tank tick-
ets, limit violation alarming, and inventory data collection and storage.

·· API 2550 standard is used to measure and calibrate tank volumes.

·· The tank inventory and ticket data are used for daily scheduling, volume bal-
ancing, and gain/loss analysis.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    603

·· The tank data includes the tank ID, inventory date and time, shipper, product
name and gravity, temperature, tank gauge level, and roof loading value.
·· The host collects the tank data from each tank and stores them in the tank in-
ventory database.

The tank inventory volume is determined from the measured gauge level through
a multi-step process (API 2550 procedure):

·· The tank level gauge is converted to a gross volume using a volume conversion
table, which may be built by means of either increment or strapping table.
·· An increment table defines incremental volumes per number of level incre-
ments, while a strapping table defines levels with corresponding tank volume.
Linear interpolation is performed if the level is between two defined increments
or strapping table entries.
·· The volume of a tank with a floating roof has to be corrected by applying a cor-
rection factor to compensate for the effect of the floating roof weight.
·· Assuming that sediment and water (S&W) are on the bottom of the tank for
strapping purposes, a free water level is subtracted from the measured tank
level to obtain a true gross volume of product.
·· Gross volumes are converted to a net volume using the temperature and density
of the liquid in the tank.
·· First, the current API gravity is measured and corrected to base condition of
15°C.
·· Second, the temperature correction factor is calculated using the API equation
or API tables.
·· The temperature correction factor is multiplied by the gross volume to obtain
a net volume.

In addition to individual tank inventory, a tank farm inventory needs to be taken.

Tank farm inventory is a balancing process, typically performed on a regular hourly
or daily time period. All transactions at a tank farm are analyzed, receipts to tanks,
tank transfer to pipeline, pipeline transfer to tanks, tank to tank transfers, etc. to ensure
that the accumulated transaction volumes from all inputs and outputs match the actual
inventory in the tanks.
As a result of individual tank inventory, leak detection can be performed on each
tank by monitoring “dead” tanks for changes in level. A tank is determined to be dead
if all valves to the tank have been closed for a certain period of time measured in min-
utes. The time delay allows the tank level to stabilize. The tank readings are captured
in regular intervals over a specified period. If the tank readings show a downward trend
over the period, a leakage from the tank can be suspected and should be investigated
for its confirmation.

9.5 POWER COST CONTROL

In general, the energy and payroll costs are the two highest in pipeline system opera-
tion, depending on the locations of pipelines. As the energy cost increases, the energy
cost control is even more important and pipeline companies should address the cost
issue. The purpose of the energy cost control is to minimize the energy cost and thus
pipeline system operating costs. Assuming that pumps are driven by electrical drivers,
power is mostly consumed by pumps and thus this section discusses the power cost
control in pump station operation.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 604    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Power cost can be reduced by controlling power demand at each pump station
run by electrical drivers. Power demand can be controlled by scheduling products or
batches properly and by minimizing the usage of power or energy. A proper schedule
includes such factors as product or batch lifting and delivery schedule, throughputs,
pumping order, etc.
The power cost can be minimized in four different ways; controlling power demand
based on the power contracts, monitoring and collecting pump unit operating statistics, op-
erating pump station and units at or near the best efficient point, and selecting an optimum
set of pump stations and their control pressure in the entire pipeline system.

9.5.1 Power Demand Control

Power demand charge is the first area that has to be addressed to control power cost
[27]. Power demand is the maximum rate of electrical energy used for a given period
of time. Normally, power is measured in kilowatts (kW).
Typically, the power company is responsible to determine the quantity of output
that will be supplied to each station, and to provide real-time data at the request of the
pipeline company. The power contract specifies electricity rates, time of the day when
and the locations where power is delivered. Also, the contract includes penalty clauses
that are applied to either side when a clause is breached.
Power contracts have power demand charges with on-peak and off-peak hour
charges, penalty clauses for high-power demands, and charges for unused power. On-
peak hours are the power contract time period when power usage charges are at the
highest rate, while off-peak hours are the time period when power usage charges are
at a reduced rate.
The demand charge is most likely based on the maximum power demand during
the month, but the billing is based on the power company’s peak power used at one
point in time during the billing period. Therefore, if the power usage is limited to a
contracted level or even zero during the peak periods, the penalties can be reduced to
zero. In order to reduce the power cost, power companies may encourage the custom-
ers to use power during off-peak hours when power is readily available and charges are
low. Since the on-peak and off-peak hours of the day are defined in the contracts, the
operator can control the power usage during the on-peak periods.
For a long pipeline, it is likely that several power companies provide power to pump
stations and the time zones of the stations may be different. This implies that the on/off-peak
hours can be different. Therefore, a summary table of power contracts can be made avail-
able to the operator in order to check the different on/off-peak hours of all the pump stations
in the pipeline system and select the least cost power company to each pump station.

9.5.2 Pump Unit Operating Statistics

The unit operating statistics are supplied through the station PLC. The pump unit op-
erating statistics are useful to run pumps efficiently and safely and to determine the
pump and driver maintenance schedule. The unit operating statistics may include the
following data:

·· A count of limit violations such as power.

·· A count of unit starts and unit total operating hours to check against the al-
located number of annual starts for a pump unit. The count of unit starts is
segregated into the number of attempted and successful starts.
·· Measured input power, calculated output power and station efficiency
·· Accumulated driver operating hours for maintenance purposes

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    605

These statistics are made available to the operator through the SCADA system.
They are determined at regular intervals and displayed on the SCADA screen along
with the efficiency at all pump stations. As a minimum, the following data may be
required for each pump unit:

·· On-peak run time, products and volume moved

·· Off-peak run time, products and volume moved
·· Total run time
·· Total number of on-peak starts
·· Total number of starts
·· Date and time the unit was last running and started
·· Limit violations and their counts
·· Measured input power
·· Calculated output power and station efficiency

Station efficiency for each station is calculated by dividing the calculated output
power by the measured input power. The output power can be obtained by multiplying
the flow rate with the differential pressure, which is the difference between the case
pressure and suction pressure.
When a pump unit start is initiated, the operator has to check if starting the unit
will violate the following constraints:

·· Exceed the on-peak or off-peak maximum station power specified in the con-
tract. This check is intended to reduce power cost.
·· Exceed the maximum number of times the unit is allowed to be started and the
minimum time required between starts. This check is intended to protect the
unit from overuse.

The ultimate purpose of collecting unit operating statistics and validation is to

control the overall operating cost.

9.5.3 Pump Station Monitoring

The pump station monitoring functions monitor and display the pump unit and driver
performances in order to operate pumps efficiently. Operation efficiency of a pump
station can be improved by monitoring unit performance and taking corrective ac-
tion if required. A pump performance monitoring function calculates pump station
performance, monitors the trends of each unit’s performance, and displays the per-
formance of pump units including alarms for deviation in performance. Based on the
monitoring result, the operator tries to operate pumps at or near the best efficient points
(BEP).
The pump station monitoring function is concerned mostly with the pump unit
efficiency at the operating point on the pump performance curve. If the driver is a vari-
able speed, then the pump performance curves are bounded by the minimum and maxi-
mum speeds, with the efficiency related to the flow, head, and speed. On the other hand,
the fixed speed pump has a single performance curve with the efficiency related to the
flow and the head controlled by a control valve. The operating efficiency is determined
from the flow and pressure at the operating points, control logic, fixed or variable speed
pump performance curves, and different combinations of pump units.
The operating point of a pump is plotted on the pump performance curves. Plots
of the current and historical operating points are superimposed onto static perfor­
mance curves including the minimum and maximum operating ranges. The operators

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 606    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

Figure 9-28. Operating point trajectory on performance curve

use these plots to operate pumps close to the BEP of the pumps. In addition, the ef-
ficiency can be trended to identify improper throttling operations or degradation of
pump unit efficiency. Such information can be used to determine the operator training
and equipment maintenance requirements and to re-rate the pump curves.
Figure 9-28 illustrates current operating point and historical trajectory of operat-
ing points superimposed on performance curves of a variable speed pump. It shows
how efficiently this pump is and has been operating for varying flow rates. Based on
this operating data, more efficient pump operating strategies can be developed. A simi-
lar trajectory can be plotted for fixed speed pumps to exhibit how efficiently throttling
actions have been taken.

9.5.4 Power Optimization

Power optimization refers to short-term power minimization for current pipeline op-
erations and off-line optimization for future operation planning. It is mainly concerned
with system-wide optimum operations of facilities such as pump stations and pressure
reducing stations. The results of a short-term power optimization are typically treated
as recommendations and are not used for a closed-loop control.
For a large pipeline system, a mathematical model is used to obtain system-wide
optimum solutions. The power optimization model deals with power consumption and
DRA usage for liquid pipelines. It determines an optimum pump station selection and
unit line-up as well as pressure set points at the on-line stations so as to minimize
power/DRA cost. It may compare the DRA cost against the power cost for the given
flow rates. The model may adjust flow rates to take advantage of lower energy costs
during off-peak hours.
An optimization model can provide information regarding the following:

·· Pump stations and units to be brought on-line

·· Optimum pump station suction or discharge pressure set points, pump unit on/
off switching schedules, and minimum power cost for a specified time period.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    607

·· Pump unit line-up and operating point, considering that a station may consist of
different pump units and that the units can be combined in various modes. The
operating points, overlaid on the pump performance curve, can be displayed on
the host SCADA screen.
·· Calculation of the overall pumping costs. When drag reducing agent (DRA)
is injected for a liquid pipeline operation, the cost without DRA is compared
against the cost with DRA.

In addition, some optimization systems may provide the following information for
analysis to help improve pipeline operation efficiency performed by operation staff:

·· Key optimization results and historical records

·· Flow rate vs. suction/discharge pressure trends with set point change records
·· Flow rate vs. number of pump units brought on-line and power consumption
·· Cumulative pump operating records
·· Pump efficiency trends

The model employs the following data in addition to the pipeline configuration
and facility data:

·· Pipeline hydraulics and equipment such as pumps

·· Pipeline and facility availability data
·· Power contract data
·· DRA cost, if the DRA injection systems are installed
·· Line fill and batch schedule data and injection and delivery flow rates for
batched liquid pipelines

The primary criterion for an optimization model is to minimize power costs. A

secondary criterion is to balance pump unit operating hours, avoiding frequent unit
start-ups and shut-downs. The solution from the optimization model should not vio-
late any pipeline and facility constraints. These constraints may include maximum
and minimum pressures and flows at certain points in the pipeline network such as
minimum delivery flow, maximum and minimum pump flows and compression ratio,
maximum power, etc.
Optimization models are difficult to apply on complex network configuration and
pump stations. However, it was reported that certain mathematical techniques were
successfully implemented for liquid pipeline energy optimization [28, 29].
A power optimization system can be implemented as a part of the host SCADA
system, and connected via an interface with the SCADA system. Through the interface,
the SCADA system sends the current pipeline states required for an optimization run,
controls its execution with the data, and receives the optimization results along with
alarm and event messages such as new batch lifting and station startup or shutdown.
The current states may include the following data:

·· Lifting and delivery flow rates

·· Pump stations and units which are on-line and off-line
·· Batch and DRA tracking data for liquid pipelines
·· Pipe roughness or efficiency to improve hydraulic calculation accuracy
·· Unit utilization data and maintenance schedule

If it is implemented on the SCADA system, the accuracy of the batch/DRA track-

ing data and friction factor need to be improved in order to calculate the hydraulic

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 608    n    Hydrocarbon Liquid Transmission Pipeline and Storage Systems

profiles accurately. In order to calculate the pipeline hydraulics accurately, accurate

pipe roughness or pipe efficiency along the pipeline may be required. A real-time batch
tracking capability can provide a more accurate hydraulic calculation. Some optimiza-
tion models can re-rate pump performance curves by analyzing recent data automati-
cally to determine actual pump performance and efficiency.
A power optimization system is typically configured to run at regular intervals as
well as on demand by the operator. Running the system at regular intervals ensures that
the system will notify the operator of any system changes required due to changes in
the pipeline line fill (e.g., batched operation, etc.). When there is a need for flow rate
change, the operator will enter the new parameters and obtain new system changes.

REFERENCES
[1] Yoon, M., Warren, B., and Adam, S., 2007, Pipeline System Automation and Control, ASME
Press, New York, N.Y.
[2] Chudiak, G. J., and Yoon, M., 1996, “Charting a course in the 90s — From field measurement to
management information systems,” Proc. of International Pipeline Conference.
[3] American Petroleum Institute, 2007, API RP 1165 — Pipeline SCADA Displays, API Publication,
1st Edition.
[4] American Petroleum Institute, 2010, API RP 1167 — Pipeline SCADA Alarm Management, API
Publication, 1st Edition.
[5] American Petroleum Institute, 2008, API RP 1168 — Pipeline Control Room Management, API
Publication, 1st Edition.
[6] American Petroleum Institute, 2009, API Standard 1164 — Pipeline SCADA Security, API Publi-
cation, 2nd Edition.
[7] Mohitpour, M., Murray, A., McManus, M., and Colquhoun, I., 2010, Pipeline Integrity Assurance —
A Practical Approach, ASME Press, New York, N.Y.
[8] “Computational Pipeline Monitoring,” API Publication 1130, 3rd Edition, American Petroleum
Institute, 2007.
[9] American Petroleum Institute, 1993, Pipeline Variable Uncertainties and Their Effects on Leak
Detectability, API Publication 1149, 1st Edition.
[10] American Petroleum Institute, 1995, Evaluation Methodology for Software Based Leak Detection
Systems, API Publication 1155, 1st Edition.
[11] Yoon, M. S., and Yurcevich, J., 1985, “A Study of the Pipeline Leak Detection Technology,” Gov-
ernment of Canada, Contract No. 05583-00106.
[12] Yoon, M. S., Mensik, M. and Luk, W. Y., 1988, “Spillage Minimization Through Real-Time Leak
Detection,” Proceedings of OMAE Conference, ASME.
[13] Yoon, M. S., Jacobs, G. B., and Young, B. E., 1991, “Leak Detection Performance Specification,”
Proceedings of ETCE Conference, ASME.
[14] Jeffrey, D., et al., 2002, “An effective and Proven Technique for Continuous Detection and Loca-
tion of Third Party Interference Along Pipelines” Proceedings of IPC, ASME.
[15] Strong, A., et al., 2008, “A Comprehensive Distributed Pipeline Condition Monitoring and its
Field Trial,” Proceedings of IPC, ASME.
[16] “Northstar Development Project Buried Leak Detection System,” Intec Project No. H-0660.03,
1999.
[17] Small, S. R., 1983, Increase Throughput With Drag Reducing Additives, Pipe Line Industry, June.
[18] Fink, J., 1998, “Additives and Chemicals Used for Transport,” ULS, NWP 0905.
[19] Johnston, R., Lauzon, P., and Pierce, P., 2008. “New Heavy Crude Oil DRA Enhances Dilution for
Flow Increase,” Hydrocarbon Engineering, March.
[20] Seto, S. P., 2005, “Investigation of Pipeline Drag Reducers in Aviation Turbine Fuels” CRC
Report#642, Project CA-68-97. http://www.crcao.com/reports/recentstudies2005/CRC%20642.
pdf.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

 Liquid Pipeline Operation    n    609

[21] http://hercules.us.es/aesop/aesop_presentation.pdf for the Assessment of Energy Saving in Oil

Pipelines (AESOP) project.
[22] Dreher, W. R., et al, 2008, “New Heavy Crude Oil Flow Improver Increases Production —
Application Scenarios,” Proceedings of IPC, ASME.
[23] ConnocoPhillips’ ExtremePower Flow Improvers website for heavy oil applications, www.­
extremepowerflowimprovers.com.
[24] ESI uses these correlations to model DRA and then try to use original references for the correla-
tions themselves. The AESOP correlation was developed under EU Project ENK6-CT2000-00096
and provided by ESI.
[25] Burger, E. D., Munk, W. R., and Wahl, H. A., 1982, “Flow Increase in the Trans-Alaska ­Pipeline
through use of a Polymeric Drag reducing Additive,” J. Petroleum Technology, 34(2), pp. 377–386.
[26] These and other DRA correlations and parameters are available from ConocoPhillips and other
DRA vendors.
[27] Neaderhouser, D. L., and Wray, B. C., 2000, Monitoring Electric Pump Costs in Real Time,
PSIG.
[28] Short, M., and Meller, S., 1996, “Elements of Comprehensive Pipeline Optimization,” Proceed-
ings of International Pipeline Conference, ASME, New York, N.Y.
[29] Jefferson, J. T., 1998, “Procedure allows calculation of ideal DRA levels in products line,” OGJ.

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms

Downloaded From: http://ebooks.asmedigitalcollection.asme.org/ on 12/30/2014 Terms of Use: http://asme.org/terms