REFRIGERATION COMPRESSOR DRIVER SELECTION AND TECHNOLOGY

QUALIFICATION ENHANCES VALUE FOR THE WHEATSTONE PROJECT

Pankaj Shah, Wheatstone Project, Chevron

Mark Weatherwax, Wheatstone Project, Chevron
Meredith Chapeaux, Wheatstone Project, Chevron
Karl Masani, ConocoPhillips Company
Cyrus Meher-Homji, Bechtel Corporation

ABSTRACT

Chevron Australia, as part of the Wheatstone Project, is constructing a two train liquefied natural gas (LNG)
facility and domestic gas plant at the Ashburton North Strategic Industrial Area, 12 kilometres west of
Onslow on the Pilbara coast of Western Australia. The initial two-train facility design capacity is 8.9 million
tonnes per annum (MTPA). Eventually, the facility could expand to produce up to 25 MTPA of LNG. During
the early stages of project development,
development, a driver selection study was performed based on the ConocoPhillips
®
Optimized Cascade   natural gas liquefaction process. This driver study evaluated a variety of project-
specific parameters and resulted in the selection of a General Electric LM6000 PF aeroderivative gas turbine
based on overall value to the shareholders considering production rates, production efficiency, fuel costs,
greenhouse
greenhous e gas emissions, installed cost, operational and maintena
maintenance
nce cost.

The selection of the LM6000 gas turbine results in the first commercial use of the LM6000 engine in a
mechanical drive application. The final decision to use the LM6000 engine was based on a detailed
technology qualification. A collaborative qualification was performed with representatives from the EPC
contractor, process licensor, end user/operator, and the equipment manufacturer. The technology
qualification followed a systematic approach to identify the risks and novelties within the application. Risk
items identified were analyzed with a combination of detailed studies, computer-based modelling and full-
scale engine testing.

Following the completion of the technology qualification, a detailed risk mitigation plan was developed. The
plan was incorporated into the purchase order of the equipment and subsequently incorporated into the
equipment manufacturer’s Failure Mode Effects Analysis (FMEA) process. Finally, detailed analysis and
testing requirements
requirements were selected to address all risks that were highlighte
highlighted
d in the risk
r isk mitigation plan.

INTRODUCTION

The Wheatstone Project is one of Australia’s largest resource projects – providing both LNG export and
greater security of supply for domestic gas production. The project will provide significant benefits to the
people of Australia, including significant employment opportunities during the construction and operational

phases,
wells andgovernment
associatedrevenue
facilitiesand
as local business
well as opportunities.
the onshore LNG and Thedomestic
project includes offshore
gas plants development
located within the
 Ashburton North Strategic Industrial Area (ANSIA), 12 kilometres west of Onslow in Western Australia’s
Pilbara region.

The Wheatstone Project is a joint venture between Australian subsidiaries of Chevron (64.14%), Apache
(13%), Kuwait Foreign Petroleum Exploration Company (KUFPEC, 7%), Shell (6.4%), and Kyushu Electric
Power Company (1.46%), together with PE Wheatstone Pty. Ltd. (partly owned by TEPCO, 8%).

The Wheatstone Foundation

Foundation Project will process gas from various fields located 145 km offshore in the West
Carnarvon Basin. Eighty percent of the Wheatstone Foundation Project capacity will be fed with natural gas
from the Wheatstone and Iago Field operations, which are operated by Chevron in joint venture with
 Australian subsidiaries
subsidiaries of Shell and Kyushu Electric Power Company,
Company, together with PE Wheatstone
Wheatstone Pty. Ltd.
The remaining 20 percent of gas will be supplied from the Apache and KUFPEC’s Julimar and Brunello
fields. The Foundation Project will consist of two LNG processing trains, each with a capacity of
approximately 4.45 (MTPA), as well as a Domgas plant with a nominal capacity of 190 MMSCFD. The
overall project is expected to consist of at least five LNG trains and additional Domgas facilities with a
nominal LNG production capacity of up to 25 MTPA.

1
 

 A final investment decision to proceed with the Wheatstone Project was made on September 26, 2011, with
construction beginning soon after in December 2011.

LIQUEFACTION TECHNOLOG
TECHNOLOGY
Y

The Wheatstone LNG Plant will utilize the ConocoPhillips Optimized Cascade ®  process. This technology
was first used in the Kenai LNG plant in Alaska in 1969, and since then ConocoPhillips has licensed 22 LNG
liquefaction trains with over 90 MTPA of liquefaction capacity in the world. A simplified process flow diagram
®
of the Optimized Cascade  process for Wheatstone LNG shown in Figure 1.

 ® 
Figure 1: Simplified Process Flow Diagram of the Optimized Cascade  Process

The Optimized Cascade® technology offers the following benefits for the Wheatstone Project:

•   Commercially proven technology

technology that can process natural gas of varying composition which is well suited
to the development of a number of separate gas fields and the Wheatstone Project’s ‘hub’ concept; the
technology has been used around the world, starting with Kenai in1969.
•   Operational flexibility enabling plant throughput to be tuned to market demand and available gas supply.
•   Benchmarks favorably with alternative processes and existing LNG plants in terms of its process
efficiency and reliability.

•
  Uses multiple,
(at reduced parallel
rates) compressor
during circuits
periods of within
planned andeach liquefaction
unplanned gas train thatmaintenance,
turbine allow continued operation
reducing the
number of full plant shutdowns and startups.
•   Use of parallel compressors allows the opportunity to use smaller compressor process drivers for a given
LNG throughput, which complements the use of high-efficiency aeroderivative
aeroderivative gas turbines.

The ConocoPhillips-Bechtel LNG collaboration has provided constant innovation, specifically in the design
and implementation
implementation of LNG driver configuration
configurations.
s. The world’s first application of gas turbines in LNG service
was implemented in 1969 at Kenai. The plant was built with six GE Frame 5 gas turbines and has operated
for 43 years, never missing an LNG shipment. In 2006, ConocoPhillips and Bechtel designed and
constructed the Darwin plant in Australia, which is still the only operating LNG facility to utilize high efficiency
aeroderivative gas turbines (LM2500+G4) in LNG refrigeration service. Since then, eleven LNG liquefaction
trains have implemented the LM2500+ for the primary refrigeration driver. This constant innovation and thirst
for efficiency improvement has led to the development of the LM6000 PF gas turbine in LNG service.

Wheatstone LNG will be the first LNG facility utilizing the LM6000 to drive its refrigeration compressors.
compressors.
DRIVER ALTERNATIVES
2
 

The selection of the driver for the refrigeration compressors within an LNG plant has a key impact on the
overall LNG plant efficiency and capacity of the plant. Natural gas liquefaction plants are generally designed
to the limits of the available refrigeration compressor drivers to maximize train capacities. It is therefore
critical that an adequate amount of effort is put into the evaluation of driver options. Figure 2 shows the
evolution of the LNG process technology since the early years from steam turbine to gas turbines as the
driver of choice for the refrigeration compressors.

Figure 2: Evolution of Refrigeration Compressor Drivers in the LNG Industry

There are three major types of drivers used in the baseload LNG industry as described below.

Steam Turbine Drivers

Most of the earlier LNG baseload plants used steam driven refrigeration compressors. However, their use
has become less common over the past 25 years. The steam turbine driver can be customized to the precise
desired power requirement for the compression system, has high equipment reliability and provides the
ability to vary operating speed over a wide range. However, the use of the steam turbine driver requires an
elaborate steam, water treatment and cooling system, making it relatively complex to operate and high in
initial capital cost. While the equipment reliability of the steam turbine is high, this benefit has to be balanced
against the reliability as well as thermal efficiency of the total system.

Gas Turbine Drivers

The gas turbines come in a variety of discrete sizes with ISO ratings varying from approximately 27 MW (for
a Frame 5 gas turbine) to 130 MW (for a Frame 9E gas turbine). The first use of gas turbines in an LNG
plant was at the ConocoPhillips LNG plant in Kenai, which started up in 1969. This plant, with an LNG
production capacity of 1.5 MTPA, used Frame 5 gas turbines to drive the refrigeration compressors. Initially,
smaller gas turbines were used resulting in plant sizes ranging from 1 to 3 MTPA. As the LNG industry
matured and focused on efficiency and cost reduction, the single train capacities grew to 5 MTPA or more.
The higher capacity trains offered the opportunity to use large-size gas turbines.

Electric Drivers

3
 

In recent years, there has been increasing interest in using electric motors as drivers for the refrigeration
compressors. Some in the industry have proposed that high availability of electric motors can increase the
overall plant production efficiency (availability). While the extent of increase in production efficiency is
debatable, this option typically requires significant additional investment for power plants and systems. The
Snohvit project in Norway is the only LNG plant to use motors as primary refrigeration drivers. To improve
the overall thermal efficiency of electric driver arrangements, combined cycle power generation would be
required with similar pros and cons of the steam systems mentioned previously.
previously.

 AERODERIVATIVE
 AERODERIVATIVE VS. INDUSTRIAL GAS
GAS TURBINES

There are two broad categories of gas turbine drivers used as mechanical drivers in the LNG facilities:

Industrial Gas Turbines

Industrial gas turbines such as Frame 5, 6, 7 and 9 have been used in the LNG liquefaction industry as
mechanical drivers for refrigeration compressors since the late 1960s. The first application was at the Kenai
plant in 1969. Since the mid 1980s, most LNG plants have used industrial gas turbines as mechanical
drivers for refrigeration systems. The thermal efficiencies, at ISO conditions, of these gas turbines range
from 29% to 34% (based on lower heating value). Industrial gas turbines can be either a single shaft or dual
shaft design. Most industrial gas turbines used in power generation application, such as Frame 6, 7 and 9,
are single shaft machines and have a limited range of speed variation. Typical speed ranges for single-shaft
gas turbines are on the order of 95% to 101%. These machines, when used in mechanical drive application,
also require a variable speed startup motor for starting up the compression strings. Unlike the large industrial
gas turbines, the Frame 5 gas turbines have a two-shaft design that allows high startup torques, making it
possible to start the compression strings under refrigeration compressor settle out pressure conditions and
without a large starter motor.

 Aeroderivative
 Aeroderivative Gas Turbines
Turbines

The use of aeroderivative gas turbines for mechanical drive dates back to the 1950s. They are attractive as
mechanical drives because of their small size, fuel efficiency and ease of maintenance. Aeroderivative gas
turbines, such as LM2500 and LM6000, have been developed from aircraft engines. They are relatively
compact, light weight and have thermal efficiencies, at ISO conditions, ranging from 39% to 43% based on
LHV.

 Aeroderivative and industrial gas turbines have inherent technical differences, some
s ome of which may impact
which type is selected for a particular LNG project. Both turbine types are available in a wide range of sizes
that provide flexibility in the selection of the appropriate size for a given application. The sizes are finite and
are not adjusted to fit the application, as are motors, steam turbines and the compressors they drive.
 Aeroderivative units are “multi-shaft” designs, whereas most large industrial units are single-shaft design.
This feature gives the aeroderivatives advantages in part load efficiency and speed variation over a wider
range of operation. The aeroderivatives also have higher full load efficiency due to a higher compression
ratio and firing temperature. Aeroderivative engines are physically smaller than industrials. They are also
modular in design, which allows maintenance and major overhauls to be done “off site.” Industrial units are
typically maintained and overhauled in place. A major item impacting the decision on the use of industrial or
aeroderivative gas turbines is their ability to handle large variations in fuel composition and the rate of
change between these fuels. Given the larger combustion system and longer residence times in the
industrial gas turbines, they are more tolerant of fuel variation compared to aeroderivative gas turbines.
There are a number of additional features that impact the comparison between aeroderivatives and
industrials, including emission characteristics, engine degradation, performance at site conditions (site
rating), and fuel pressure requirements. Both aeroderivatives and industrial gas turbines are well proven
selections for use in LNG refrigeration service.

POWER AUGMENTATION FOR GAS TURBINES

4
 

 All gas turbines have a degradation of power output with increasing ambient temperatu temperature.
re. Power
augmentation can be used to level out the available power from the gas turbine, as well as optimize LNG
production at all ambient conditions. The degradation of power output is normally more pronounced with
higher compression ratio gas turbines, such as aeroderiva
aeroderivative
tive ones. The LM6000 has a compression ratio of
30:1, resulting in a higher power lapse rate with ambient temperature compared to industrial gas turbines.
Compared to other aeroderivative gas turbines, the lapse rate is similar for the higher ambient temperatures
but more significant at the lower ambient ranges. This is shown in Figure 3, which shows the gas turbine
lapse rate between typical average and extreme high ambient conditions. This clearly shows that
aeroderivatives have a much greater lapse rate compared to industrial gas turbines. It also shows that
adding inlet air humidification (IAH), also known as evaporative cooling, to an aeroderivative turbine reduces
the lapse rate from about 25% to 5% when high temperatures occur. This reduction is dependent on site
specific conditions.

Gas Turbine Lapse Rate

(Avg Ambient to Extreme High Ambient)
30%

25%
LM6000PF
20%
LM6000PF w/Evap
w/Evap Cooler
15%
PGT25+G4
10% Frame
Frame66B
B

5% Frame 7EA

0%
Lapse Rate
Rate - Avg Amb. To Ext High Amb.

Figure 3: Gas Turbine Lapse Rate between Average and Extreme Ambient Temperatures

Figure 4 shows the same information over the whole range of ambient conditions. The data is normalized to
the power output at 26 deg C. The purpose of the curves is to show the relative lapse rate given the
available power at the average ambient temperature.

5
 

Gas Turbine Lapse Rate versus Ambient

Temperature (new an
and
d clean)

   C
   6
   2
   t
   a
   r
   e
   w
   o
   P
   o
   t
   e
   v
   i
   t
   a
    l
   e
   R
   r
   e
   w
   o
   P
 
   %
    '

LM6000PF LM6000PF w/Evap Cooler

PGT25+G4 Frame 6B
Ambient Temperature (deg C)
Frame 7EA

Figure 4: Gas Turbine Lapse Rate vs. Ambient Temperature

For gas turbines with IAH, the turbine inlet temperature is a result of evaporative cooling of the ambient air.
The average coincident humidity plus one standard deviation was used in the calculations to determine the
effectiveness of the IAH system. The selection of plus one standard deviation was made to minimize the
occurrences of high humidity conditions limiting the ability to reach the anticipated turbine inlet temperature.
By choosing plus one standard deviation, the humidity will only be higher than the design basis
approximately 16% of the time. As seen in Figure 5, the pinch point in power is only anticipated to occur
between 32 to 39 degrees Celsius. Based on 11 years of coincident hourly data, the dry bulb temperature is
in the 32 to 39 degree C range approximately 10.5% of the year. Therefore, the occurrence of high humidity
coincident with this temperature range is anticipated to be less than 1.6% of the year. This compares to the
use of the average relative humidity which would have resulted in power below the minimum design power
over 8.5% of the time.

6
 

Figure 5: Impact of Coincident Relative Humidity vs. Derated Power with IAH

 As seen in Figure 4, the addition of IAH to the LM6000 for Wheatstone Project site conditions results in a
lower lapse rate at the higher ambient conditions compared to any of the alternative gas turbine selections.
 At the low temperature ranges, the lapse rate is less impacted by the IAH since the humidity at site
conditions under these scenarios is high. At the low ambient conditions, the humidity is close to 100%, giving
very little, if any, benefit from the evaporative cooler. Since the power at the average ambient conditions is
higher for the IAH option, the lapse rate to the lowest ambient conditions is less even though the power
output with IAH and non-augmented LM6000 is equivalent at this condition.

DRIVER SELECTION STUDY

The selection of the type and configuration of the mechanical drivers for a liquefaction refrigeration system is
a key decision during pre-Front End Engineering Design (pre-FEED). For the Wheatstone Foundation
Project, a study was performed to evaluate various refrigeration compressor driver and train configuration
options.

The purpose of the study was to determine the most cost effective refrigeration compressor driver type and
configuration to allow the LNG trains to reliably produce LNG. The selection process for the refrigerant
compressor drivers involved completing a technical and economic comparison of a number of alternative
drivers and driver configurations; this was carried out by study contractor, Bechtel, as part of the pre-FEED
concept selection studies. The information developed by the contractor was then assessed by the owner’s
team, and the capital and operating cost figures were used to prepare a value analysis using the project
economic model.

The study basis that influenced the driver selection was:

•   Feed stream composition and rate

7
 

Based on the upstream resource and offshore pipeline design, the feed gas rate to the LNG facilities was
fixed. Therefore, any excess driver power that might be available was not directly usable for significant
additional annual production of LNG.
Feed stream nitrogen content for the Wheatstone and Iago Field ranged from 5 to 6 mol%. However, the
plant design needed to be flexible to handle feed gas with up to 16 mol% nitrogen from Apache/Kufpec’s
Julimar and Brunello fields. Since excess nitrogen content increases the refrigeration specific power
required, gas turbine power margin improves the ability to process gas with higher nitrogen content and
still achieve design capacity.
•

  Site ambient conditions

The site ambient temperatures that formed the basis of design for the Wheatstone plant
plant are summarized
in Table 1.

Table 1: Site Ambient Temperature

% Exceedance <0.05% 1% 5% 50% 95%

Case Extreme High High-High High Average Low
Temperature (oC) 46 39.6 35.1 25.3 15.2
Design Relative Humidity (%) 15.0 31.6 58.1 81.8 92.1

The key criteria for the selection of the refrigeration compressor driver for Wheatstone Foundation Project
trains included the following:

•   Total installed cost of the refrigeration system

•   Net present value analysis, taking into consideration LNG production and operating cost amongst other
parameters
•   Greenhouse gas emissions and potential greenhouse gas emission taxes
•   Technical, operational,
operational, and schedule risk assessments
•   Thermal efficiency and impact on LNG production
•   Potential for processing feed gas with higher nitrogen content
The type, number, and power output of the LNG refrigeration compressor drivers are key determinants of the
capacity and onstream reliability of the LNG trains. The ConocoPhillips (COP) Optimized Cascade ®  natural
gas liquefaction process comprises three pure component refrigeration loops using propane, ethylene and
methane. The Optimized Cascade technology-based plants built or designed to date have all used simple
cycle gas turbines as their refrigeration compressor drivers, typically in a configuration with at least two gas

turbines driving parallel compressor strings for each of the three refrigeration loops (2 x 2 x 2 configuration).
LNG industry operating experience is an important criteria to owners and financiers when selecting a turbine
for refrigeration service. Both industrial (‘Frame’) and aeroderivative gas turbines have been used in the
Optimized Cascade process to drive the refrigeration compressors. Most plants in operation have used
variants of the GE Oil & Gas Frame 5 gas turbine. However, the Darwin LNG Plant in Northern Australia
uses GE Oil & Gas PGT25+G4 (LM2500+G4 gas generator section with high speed power turbine)
aeroderivative turbine. The Angola LNG plant uses Frame 7 and Frame 6 industrial gas turbines. All trains
designed since the successful startup and operation at Darwin in 2005, have been based on the
LM2500+G4.

Based on Bechtel-COP operating and earlier study experience, a short list of configuration options were
evaluated in parallel with other pre-FEED activities:

i. 6 x LM2500+
LM2500+G4
G4 with mechanical
mechanical refrigeration for inlet
inlet air chilling (IAC)
ii. 6 x LM2500+
LM2500+G4
G4 with Inlet Air Humidification
Humidification (IAH)
iii. 7 x LM2500+G4 with no power augmentation
iv. 6 x LM6000PF with IAH

8
 

v. 7 x LM2500+G4 with IAH

vi. 7 x LM2500+G4 with IAH and waste heat recovery steam generation
generation units, with steam turbines to
generate electric power and steam for process heating
vii. 2 x Frame 7EA and 2 x Frame 5D.

The 6 x LM2500+G4 with IAC (Option i) configuration was the basis for the initial pre-FEED work; this case
was also used for comparing the capital cost for the other alternatives. The cases without steam generation
use gas turbine generator sets for electric power production. Inlet air cooling using evaporative humidifiers or
mechanical chilling was considered in a number of alternatives to augment power output and mitigate severe
seasonal swings in LNG output. Upon further review, the ambient air temperature swings and resulting gas
turbine power fluctuations at the Ashburton North site precluded using aeroderivatives without either IAH or
IAC.

Table 2 summarizes the comparison on some of the parameters for the different alternatives from the study.

Table 2: Summary of pre-FEED driver selection study

st udy alternatives

Driver Description LNG Output, Estimated Total CO2

Alternative Two trains, Power Installed Emissions,
(MTPA) Margin, (%) Cost (TPD)
(US$ MM)
i 6 x LM2500+G4 with IAC Base Base Base Base

ii 6 x LM2500+G4 with IAH Base – 0.06 Base + 0 Base - 63 Base - 100

iii 7 x LM2500+G4 Base + 0.02 Base + 2 Base + 39 Base + 50
iv 6 x LM6000PF with IAH Base + 0.02 Base +12 Base – 57 Base - 40
v 7 x LM2500+G4 with IAH Base + 0.02 Base + 9.5 Base + 119 Base + 70
vi 7 x LM2500+G4 with IAH Base + 0.19 Base + 9.5 Base + 448 Base – 400
and combined cycle
power generation and
steam for process heating
vii 2 x Frame 7EA & 2 x Base - 0.29 Base + 6 Base – 256 Base + 620
Frame 5D

Because the total feed gas supply was fixed for the study, some configurations resulted in excess power
available. The much higher power output of the LM6000 units with IAH provided opportunities for
optimization and cost reduction elsewhere in the process. A list of potential total installed cost (TIC)
reduction opportunities with savings totalling up to ~$200MM were identified if the excess power available
from the LM6000s was used to eliminate equipment and optimize the LNG train design. The TIC for
alternative iv also includes a cost penalty for an extended startup period because LM6000s have not
previously been used in a refrigeration drive application. Cases with lower fuel gas consumption and
adequate refrigeration power could potentially produce more LNG under a feed rate constrained scenario.

While alternative vii has the lowest estimated TIC cost, it also has lower thermal efficiency, with the highest
fuel gas use (resulting in less feed gas available for LNG production) and highest CO 2 emissions. Due to
these factors and the longer maintenance down time required for the industrial (frame) gas turbines,
alternative vii resulted in significantly less annual LNG output compared to the other options.

 Alternative ii has the next lowest TIC, but this configuration has no excess power and is not able to achieve
the target LNG output. Modificatio
Modifications
ns to this design could result in achieving the target output by reducing the
propane condenser approach temperature, adding limited IAC, or other methods, however, any of these
would increase the TIC of this alternative. In addition, the power margin was still insufficient to provide the
flexibility needed to handle higher nitrogen content feed gas, such as that expected from Apache/KUFPEC’s
Julimar and Brunello fields.

9
 

To compare the relative NPV and capital efficiency of the alternative driver configuration cases, the TIC,
LNG production, and operating costs were used to compare each case. The results are shown in the NPV
vs. Present Value Capex plot in Figure 6 below.

1900
Discounted Prof
Prof itability In
Index
dex Y vi
1800
ii iv Discounted Prof itability Index
Index X
  e
  u
   l
1700
  a
   V i iii v
   t
  n
  e
  s
  e
  r
1600
   P  Alternative Descrip tion
   t
  e
   N
i 6xLM2500+ G4 wit h IAC
1500 ii 6xLM2500+ G4 wit h IAH
III 7xLM2500+ G4 wit h no aug ment at io n
iv 6xLM6000P F wwiit h IA H
vii v 7xLM2500+ G4 wit h IAH
1400 vi 7xLM2500+ G4 wit h IAH and St eam Cy cle
vii 2xFrame 7E A

1300
7000 7050 7100 7150 7200 7250 7300 7350 7400 7450 7500
Present Value of Capital Expenditure
Present

Figure 6: Relative Capital Efficiency of Driver Alternatives

The case generating the highest NPV was alternative vi. The increased revenue from higher LNG production
and lower CO2 emission costs clearly offsets the higher capital cost. The NPV and capital efficiency appear
superior to the alternative cases using 7 x LM2500 drivers. However, the investment efficiency parameters,
such as discounted profitability index (DPI) and rate of return (ROR), for alternative vi were approximately
the same as that LM6000PF case (alternative iv), which has a lower TIC of about $500 MM. In addition to
lower TIC and comparable DPI/ROR, the alternative iv has a higher excess power margin, which, as noted
above, can translate into:

•   Higher LNG production if feed gas supply rates can be increased

•   Increased operational
operational flexibility, including the ability to handle higher N 2 content feed gas streams
•   Reduced operating load on the LM6000PF gas turbines, leading to extended maintenance intervals,
reduced downtime, incremental LNG production, and lower operating costs

The combined cycle / cogeneration system in alternative vi provides higher thermal efficiency, but it results in
a more complex operating system. It has higher water demand and additional water treating requirements.
More significantly, it would also be a first-of-a-kind design for the Optimized Cascade process. It was
estimated that the overall FEED timetable would need to be extended by about three to six months to
incorporate the new design aspects resulting from the addition of a steam system, combined cycle power
generation, etc. The LM6000PF drivers, while also not previously used, can be used in the same basic
design configuration used successfully in many operating plants while maintaining the current Wheatstone
schedule.

Overall, the driver configuration using 6 x LM6000PF with IAH provides the optimal combination of relatively
low initial TIC, attractive NPV, DPI and ROR, operating flexibility, and acceptable technical and schedule
risk, and was thus selected for the Wheatstone Project.

HIGH AMBIENT TEMPERATURE OPERATION

The design at various conditions were closely reviewed to ensure that the upstream and LNG plant
operations would not be negatively impacted by the large range of ambient temperatures and relatively high
summer temperatures
temperatures at the plant site.

10
 

The project Basis of Design required that the plant operate at ambient temperatures up to 46oC. At these
conditions, the propane system was approximately 5% short of power and the compressors could not
operate under these conditions. Several options were evaluated to manage the high ambient temperature
operation, as summarized below:

Overfiring of the Gas Turbines

GE
fromallows short-term,
the turbine. periodic overfiring
The overfiring can affectof turbines,
the expectedwhich provides schedule
maintenance a 6% increase
of theinhot
the power Based
section. available
on
the site ambient temperature information, it is expected that overfiring will be required about 25 hours/year.
The overall maintenance schedule of the propane GT requires a major hot section overhaul shutdown at
25,000 hours (or ~3.2 years). Considering that the aeroderivative gas turbine maintenance is an on-condition
based maintenance, it is anticipated that the impact of the overfiring will not be noticeable in the plant
maintenance schedules.

Higher Purity Refrigerant Propane

The current design is based on assuming a 98 mol% pure propane refrigerant. This assumed 1 mol% ethane
and 1 mol% i-butane in the propane. A simple simulation showed that a 99 mol% purity propane (with 1
mol% i-butane) would decrease the overall propane compressor power required by about 2% due to the
ability to condense the propane at a lower pressure when the light components are removed. Based on
information from the propane supplier, higher purity propane should be available. However, no credit for this
enhancement can be assumed until the plant is operating and the imported propane purity is confirmed.

 Add SPRINT
SPRINT to Propane LM6000
LM6000 Gas Turbine
Turbine

Based on GE’s information, the SPRINT system would allow for approximately 3 MW additional power from
the GTs. This is about 9% of the current power requirement and would cover the power shortage. However,
this resulted in about 50% increase in the facility’s demin water system. Furthermore, while this system was
reviewed in detail by GE, there are no operating units utilizing the low pressure injection system for a hot end
drive application. The hot end drive application used for mechanical drives requires the SPRINT nozzles to
be supported on the inlet compressor struts, and this arrangemen
arrangementt has not previously been used.

 Add Inlet Mechani

Mechanical
cal Chillers
Chillers to Gas Turbines
Turbines

The addition of mechanical chillers to the inlet of the GTs was evaluated during the main driver study in the
early phase of the project. Because the possible power increase with the mechanical chillers was estimated
to about 20%, it was determined that it would only be economically viable if it could be used year-round to
produce more LNG. However, under the basis that the facility is feed gas limited, the high capital cost
associated with this option was not warranted to only provide additional power during the extreme ambient
days.

 Add Lithium
Lithium Bromide (LiBr) Chiller
Chiller Package to Gas Turbines
Turbines

High-level research on the use of LiBr chillers showed that their performance tends to drop off quickly in high
ambient temperature environments (>95oF, 35oC) due to crystallization of the absorbing solution. Therefore,
this technology was not deemed feasible to address the issues posed by Wheatstone Project’s ambient
conditions.

11
 

 Add Helper
Helper Motors to the
the Propane Gas
Gas Turbines

The potential of adding a helper motor to supply additional power was evaluated. This too would be a first-of-
a-kind design, because large helper motors have not been used on 2-shaft GTs, as they are able to start
without the assistance from large starter motors. The addition of a helper motor with associated variable
frequency drive is a significant capital cost. Additional power generation may also be required. The high
capital cost associated with this option did not warrant further evaluation.

Utilize Impeller Technology with More Turndown Capability

 As the ambient temperature increases, the pressure ratio in the refrigeration compressor increases and the
power consumed will increase. If the gas turbine is power limited, a reduction of flow through the compressor
will reduce the required compression power. The flow through the compressor can only be decreased until
the flow is at the surge control line, at which time the recycle valves will start opening and essentially lock in
the compression power. Any further increase in the ambient temperature will result in the gas turbine slowing
down and the compressor suction pressures rising. A significant amount of work and qualification was
performed with GE Oil & Gas to utilize their T5.3 high Mach impellers, which offer more turndown capability
than the standard propane compressor impellers.

Many other options were also evaluated for which credit could not be taken during the design phase of the
project but might be available during operations: margins on equipment, quality of the propane refrigerant,
using the GTs with the highest tested power on the propane strings, good operational practices to ensure
that GTs are washed regularly and utilization of high efficiency air filtration, etc.

Based on the overall evaluation, it was recommended that for the limited amount of time where this
additional power is required, the GTs be provided with an overfiring option and the T5.3 impellers
implemented for propane service. Implementation of any other option seemed unwarranted. Future plant
enhancements, once actual operational limitations are known, could be implemented as debottlenecking
options. To limit the over-fire hours, the control system limits the activation of the over-fire mode to high
ambient conditions and times when the compressors are operating with the recycle valves open.
 Additionally, the over-fire controls are only activated on the propane compressor due to the direct impact of
higher temperatures on propane condensing. The selection of the LM6000 gas turbine drivers for
refrigeration compressors resulted in the first use of the LM6000 engine in a mechanical drive application.
The final decision to use the LM6000 engine was based on a detailed technology qualification. A
collaborative qualification was performed with representatives from the EPC contractor, process licensor,
end user/operator, and the equipment manufacturer. The remainder of the paper provides an overview of
LM6000 engine and the technology qualification process.

LM6000 OVERVIEW

The LM6000PC/PD/PF family was derived from the CF6-80C2 aircraft engine that has over 3,731 engines in
service with more that 188.8 million flight hours and 43.8 million flight cycles. The LM6000PF, which is the
specific variant used for the Wheatstone Project, is shown in Figure 7.

12
 

LM6000-PF 
15 ppm

Figure 7: LM6000PF Engine

The LM6000PF engine is a DLE (dry low emissions) configuration of the LM6000 and is the GE current
offering for mechanical drive and power generation applications. The PF was derived from the PD model to
improve operability and fuel flexibility. The following modifications
modifications are made to the LM6000PD engine:

•   Improved combustors and pre-mixers

•    Additional manifold
manifold and
and fuel valve
•   Control software changes
No other changes have been made to the LM6000PD hardware and the rated output of the engines is
identical.

The modifications to the combustors and pre-mixers are shown in Figure 8.

13
 

Design Changes from DLE1 to DLE1.5 technology 

Ori
Ori in
inal
al LM250
LM2500
0 and Current LM2500+G4 and LM6000-

Short wings on inner and outer heat shields 

Wings removed
removed from center heat shields  
Modified pre-mixers to optimize fuel to air ratio in
the combustor 

Figure 8: LM6000PF Combustor and Pre-mixer Details

The experience of the LM6000 engines includes 1076 engines with total operating hours of more than
25,405,471 hours. There are 260 LM6000 PD/PF machines operating with total operating hours of greater
than 5,552,252 hours with the high time engine greater than 95,937 hours. The fleet leader PF engine has
over 54,766 operating hours.

The LM6000 is unique from other 2-shaft gas turbines in that the power turbine (LPT) is connected to the low
pressure axial compressor (LPC) section of the gas turbine (Figure 9). Therefore, the starting and shutdown
characteristics are unique to this engine and were identified as areas requiring detailed review.

14
 

Figure 9: Schematic of LM6000 Engine

The basis of the qualification was a dry low emissions (DLE) unit since all recent mechanical drive gas
turbines currently operating in Western Australia use dry low NOx technology, and environmental permitting
requires that the best available technology to be applied. The LM6000PF engine is capable of achieving a 15
ppm NOx level. The 15 ppm NOx guarantee is based on operation of 70% and higher power levels. The
LM6000PF engines have been guaranteed to operate down to 50% load with a NOx level below 25 ppm.

CHEVRON TECHNOLOGY QUALIFICATION PROCESS

Chevron’s Technology Qualification Process (TQP) is a systematic process for reducing uncertainty
associated with new technology by breaking the system into subsystems and/or components and
determining the risk associated with each. The primary elements of the TQP for the use of the LM6000PF as
a mechanical driver for the Wheatstone Project started in October 2008 and were concluded in December
2009.

During the TQP, emphasis was placed on subsystems and components with the highest degree of
uncertainty because this offers the greatest opportunity for risk reduction. The TQP is one part of an overall
technology deployment strategy that includes leadership behavior, accountability, training, communication,

data capture/metrics, and feedback. The TQP methodology involves the following elements.

Preparation

Preparation involves integration of TQP into the Chevron Project Development and Execution Process
(CPDEP) and mobilization activities, such as TQP team selection, identification of subject matter expert
(SME) requirements, budget and schedule development, and an overview of techniques to be used.

Technology Specification

The Technology Specification defines all requirements for the technology, including target reliability.
Progress of the technology toward deployment is measured by use of initial and target Technology
Developmentt Stage (TDS). The TDS definitions are shown Figure 10.
Developmen

15
 

Figure 10: TDS Designations

Threat Assessment

Threat Assessment identifies threats associated with each subsystem or component. Threats are defined as
failure modes and mechanisms
mechanisms and are identified for the entire lifecycle of the technology.

Threats were converted into ranked risks. The risk rankings were established on a qualitative basis and
considered both the severity and likelihood of actually realizing each identified threat. Risks were ranked as
Low, Medium or High according to the intersection of likelihood and severity.

Qualification Plan

The Qualification Plan identifies all analysis, testing and procedure development needed to reduce the risk of
each threat. Associated schedules and costs are then developed.

Qualification Execution

Performed in accordance with the Qualification Plan, Qualification

Qualification Execution gathers evidence from analysis,
testing and procedure development to assess performance of the technology. Quality, accuracy and
statistical significance of this evidence are critical to the TQP.

Performance Assessment

Performance Assessment evaluates the results from Qualification Execution against requirements of the
Technology Specification. This step determines if the technology can be deployed for a specific set of

16
 

conditions, if additional refinement and another iteration of the TQP are warranted, or if the technology
should be abandoned for that project.

Deployment Preparation

For technology that complies with the Technology Specification, Deployment Preparation documents the
TQP results and recommendations. Thorough documentation and traceability of the TQP will provide an
audit trail for future projects to evaluate any new technology and consider its use.

LM6000PF Mechanical Drive System and Sub-Systems

The LM6000PF mechanical drive system was split into sub-systems to facilitate the qualification process.
The sub-systems identified and evaluated are included in Figure 11. Based on the identified sub-systems,
the following is a summary of the work and major findings.

Figure 11: LM6000PF Mechanical Drive Sub-systems

LM6000 PF Lateral and Torsional Analysis

Lateral analysis, torsional analysis, and unbalance response for the LM6000PF in both a cold end drive and
a hot end drive configuration were provided during the qualification phase. In addition, GE performed a
combined lateral-torsional vibration analysis of the LM6000 engine with the load coupling, gearbox and

compressor on the methane train.

17
 

Temperature & Stress Profiles

GE shared data for the stress levels within the blading from testing during loading as a mechanical drive unit.
GE demonstrated that the stresses developed within the components were all within their standard design
ranges.

Subsequent to the above, GE performed a study including a rapid stop of the LM6000 LP shaft. Based on
this analysis,
standard GEcriteria
design confirmed that
for the the temperature
individual and stress profiles within the LM6000 were within their
components.

 Aero Matching
Matching of Components
Components during Shutdow
Shutdown
n and Startup
Startup

 Aero matching of compone

components nts during startup and shutdown was reviewed. The review was based on GE
simulations that forced the LP rotor to stop in 4.4 seconds simulating the braking load from a centrifugal
compressor. To accomplish this, GE imposed an external load on the engine in excess of the engine rating.
GE utilized this increase in load and fast shutdown to demonstrate a worst case scenario for the engine’s
ability to withstand a fast stop. A load of 59,580 kW was imposed on the gas turbine at the same instant the
fuel valve was closed. The residual fuel in the manifolds continued to provide combustion for approximately
0.75 seconds after the fuel valve was closed. The load was then decreased to 23,012 kW over the next 4.4
seconds. The IGV’s (Inlet Guide Vanes) and VBV’s (Variable Bleed Valves) were stepped fully opened when
the fuel valve was closed. The VSV’s (Variable Stator Vanes) modulate closed based on the speed of the HP
spool. The HP spool decelerated very quickly through the first 4.4 seconds. At 4.4 seconds, the HP spool
was at ~44% of its initial speed.

The simulation had the ability to predict stall within the axial compressor sections. Based on the above
scenario, no stall events occurred within the machine. The simulation also evaluated the component
stresses, and similar to stall, no componen
components
ts exceeded acceptable stress levels.

Figure 12 shows the ramp

r amp down rate for the Bechtel Simulation and the fastest deceleration achieved by GE
using their simulation data without having stall within the gas turbine axial section or modifying their standard
controls.

Figure 12: LM6000 - LP Spool Deceleration Rate

18
 

Based on the simulations performed by Bechtel and GE, there was sufficient margin within the machine to be
able to accommodate a trip of the gas turbine from full load conditions without resulting in stall occurring
within the axial compressor section of the gas turbine.

Continuous Base Load Operation

The mechanical drive applications

applications in an LNG facility required that the gas turbine be run
r un in a base loaded
condition
Figure 13 for extended
includes periods
a subset of time. Aoperating
of engines review was made
under of load
base engines operating
conditions thatunder
were base load conditions.
consideredconditio
during ns.
this evaluation.

Service factor for baseloaded units

1

0.95

0.
0.9
9

0.85

0.
0.8
8

0.75

0.
0.7
7

0.65

0.
0.6
6

0.55

0.
0.5
5
Jan- Feb- Mar-  Apr-
 Apr- May- Jun- Jul-04 Aug- Sep- Oct- Nov- Dec- Jan- Feb- Mar-  A
 Apr-
pr- May- Jun- Jul-05 Au
Aug-
g- Sep- Oct- Nov-
04 04 04 04 04 04 04 04 04 04 04 05 05 05 05 05 05 05 05 05 05

 
 A B C D E F G H I

Figure 13: LM6000 Base Loaded Units

LP Rotor Turning Gear Operation

To perform off-line water washes of a mechanical drive LM6000, a turning gear needed to be provided for
the LP rotor system. This turning gear system was designed based on previous experience with 50 Hz
applications, industrial gas turbine experience, and previous experience with turning gear systems for large
steam turbine driven systems.

For 50 Hz, LM6000 power generation applications, a turning gear is installed on the speed reduction gear
box. For industrial gas turbines, the turning gear is installed on the auxiliary gear box. For the LM6000
mechanical drive application, the turning gear will be installed on the far end of the compressor train in line
with GE experience on previous compressor trains. The actual turning gear components were developed
from proven equipment supplied by Koenig.

The turning gear must overcome the large break-away torque, develop the required speed to operate above
the lift-off speed for the dry gas seals, and at a speed sufficient to perform the off-line water wash. To
perform this function, GE proposed a turning gear with two motors. One motor included an additional
reduction gear to develop the required break-away torque. This motor rotates the LP shaft at approximately

19
 

10 rpm. A second motor will then be used to turn the rotor at approximately 300-350 rpm to allow an off-line
water wash. This configuration was similar to designs previously used by GE on steam turbine applications.

Mark VIe Control Logic

The basis of the evaluation is the use of a dual redundant Mark VIe control system for the LM6000PF. There
were differences in the IGV (inlet guide vane) and VBV (variable bleed valve) schedules between power
generation and mechanical drive application
applications.
s.
Further reviews were performed during the FEED phase of the project for the control system modifications,
but no issues were identified that required further qualification.

20
 

Variable Speed Capability and Performance

The mechanical drive application requires that the gas turbine be capable of variable speed operation. GE
has been working on mechanical drive applications for the LM6000 since 1992. GE has performed testing on
the LM6000 at various times to demonstrate the variable speed capability of the engine from 50% to 105% of
rated operating speed. Variable speed testing of the engine was performed as shown in Table 3.

Table 3: LM6000 Mechanical Drive Testing

Test Date Test Engine Test Type Time on Test
1992 LM6000PA Locked rotor testing to demonstrate break-awa
break-away y
torque capability. This was the second engine to
test (SETT) for the PA model.
Nov-Dec 1996 LM6000PD Variable speed from 1800 to 3780 rpm. 70 hours
May-Jul 1997 LM6000PD Variable speed from 1800 to 3780 rpm including 165 hours
no-load over-speed at 3960 rpm.
LM6000PC Multi-engine vibratio
vibrationn signature validation and 8 hours
testing.
LM6000PC Multi-engine vibratio
vibrationn signature validation and 8 hours
testing.
LM6000PC Multi-engine vibratio
vibrationn signature validation and 8 hours
testing.
LM6000PC Multi-engine vibratio
vibrationn signature validation and 8 hours
testing.
LM6000PC 10,000,000 cycle endurance testing. 60 hours
Sept-Oct 2005 LM6000PC Endurance test for ABS-SVR Certification. 168 hours
Testing from 800 to +3600 rpm.
June 6-7, 2006 LM6000PC Starting torque capability from 700 rpm to 3780 18 hours
rpm. Demonstrated operation at 3400 to 3450
rpm.
January 23-26, LM6000PF Variable speed from 1500 rpm to 3780 rpm in 12 hours
2008 the Evendale test cell using a variable speed
generator.
June 17, 2009 LM6000-PD Locked rotor testing 4 hours
LM6000-PD Low power/low speed testing 16 hours
LM6000-PD No. 1 Bearing Load vs. Variable Speed/Power 24 hours
June 12 – 18, LM6000-PF Variable Speed operation 950 – 3780 rpm 45 hours
2012
July 30 – Aug LM6000-PF Low power / low speed optimizati
optimization
on testing 20 hours
1, 2012
Total test time > 600 hours

Subsequent to placing the order for the Wheatstone Project, all 17 engines ordered (12 mechanical drive, 4
generator drive and 1 spare) have undergone variable speed testing as part of the contract with GE.

Starting Torque Capability

21
 

The starting capability of the engine is a combination of the break-away torque and the engine torque
available across the speed from zero to 105% speed. To support the Wheatstone Project, GE performed a
locked rotor test of the LM6000 engine in a PD configuration on June 17, 2009. This testing confirmed the
break-away torque available from the engine.

Dynamic simulations to predict the starting torque requirements of the Methane compressor string were
conducted during the FEED phase of the project. During the EPC phase, dynamic simulations were also
conducted for the ethylene and propane compressor services. The methane compressor string was selected
during FEED for the simulation because of the use of a gearbox and three compressor
compressor bodies. The results of
the simulations and studies indicate that the LM6000 is capable of starting the compressor strings from full
settle out conditions.

Load Gear Selection

Based on experience, a Lufkin load gear was selected. A design review of the gear was held during the
qualification phase, and analysis was performed to demonstrate the ability to manufacture the gear with a
single piece design.

Couplings

The load coupling proposed for the LM6000 mechanical drive will include a torque meter to aid in monitoring
the performance
supported of the
by the gas gas and
turbine turbine. The torque
will impact lateralmeter will add
dynamics of theadditional weight
gas turbine. to the coupling
In addition, half
the torsional
stiffness of the coupling will impact the overall train torsional. The final coupling selection included
 Ameridrive couplings
couplings following
following previous GE experience for for aeroderivative
aeroderivative applications.
applications.

Train Torsional

 A train torsional analysis was conducted for all compressor

compressor trains. The methane train is considered to be the
most complex from a torsional perspective because it has three compressor bodies and a speed increasing
gear. The analysis was conducted using experience available to GE from previous projects and was
considered preliminary. During detailed engineering, the torsional analysis was finalized and fatigue analysis
was completed to demonstrate that no issues would result in the compressor trains operating range.
range.

Train Alignment

The train alignment study was conducted to determine the amount of axial stretch required on the flexible
element couplings. The study calculated the thermal growth of each piece of equipment as well as the
growth of the coupling spool. The resulting deflections and pre-stretch were used to confirm the axial loading
on the gas turbine.

 Axial Load Limitations

Limitations

Selection of the coupling is critical for the hot-end drive configuration to ensure that allowable axial load limits
are maintained. Based on the alignment study and the growth calculations, the axial loading on the gas
turbine is acceptable and within their standard limits for the machine.

Train Starting

22
 

The LM6000 does not utilize a free power turbine similar to the LM2500 or Frame 5D gas turbines. The
power turbine section of the LM6000 is connected to the LP axial compressor of the engine. Based on this
difference, it was important that the starting capability of the engine be fully analyzed. To address this, GE
performed a locked rotor test of an LM6000PD, and Bechtel performed a dynamic simulation of the LM6000
compressor trains. In addition, testing was conducted on a LM6000PF to demonstrate the engine torque
capability through the complete speed range (0% to 105%). The dynamic simulation was completed and
included the starting torque requirements for a full pressure restart of the compressor train from settle-out
pressure.

Based on all analysis and test data, the LM6000 will be capable of a full pressure restart of the compressor
train.

Fuel Composition

The predicted Wheatstone fuel gas compositions have been reviewed with GE. The primary issue with the
fuel compositions is the rate of change of modified Wobbe index (MWI) impacting the required controls
(Wobbe meter and calorimeter) in addition to the configuration of the fuel system piping for the fuel supply to
the engines. Detailed fuel gas simulations were performed during the EPC phase of the project to
understand the rate of change and to properly design the configuration of the fuel system and the placement
of the Wobbe meters and gas chromatographs. Additionally, the gas turbine controls utilize the Wobbe meter
to actively control the combustion parameters and fuel flow splits of the engine based on the fuel gas
composition.

Compressor Selections

Preliminary compressor selections were made by GE based on the average feed composition operating
conditions to support the LM6000 TQP efforts.

The impellers proposed for the Wheatstone Project using the LM6000 as a mechanical driver fall within
proven experience for GE for the Methane and Ethylene services. The T5.3 impellers used for the propane
service were selected based on a separate technology qualification subsequent to the selection of the
LM6000’s as mechanical drives.

General Arrangement

Based on the
flammable gas project
locatedlayout
under of theairgas
inlet turbines
ducting andgas
to the safety concern
turbine, with of
the basis having the compressors
the qualification was a handling
hot end
drive configuratio
configuration.
n.

The major advantages of the hot end drive are:

•   Gas turbine layout is similar to other gas turbine drives including the LM2500s
•   The torque is not transmitted back through the entire low pressure rotor
•   Lower vibration levels are observed when operating in the hot end drive configuration
•   The inlet ducting is routed away from the process compressors
The hot end drive configuration was reviewed from an operability and maintainability perspective and an
optimized package was designed for the Wheatstone Project. No issues were identified that impacted the
technical feasibility of the LM6000 in a mechani
mechanical
cal drive application
application..

RISK MITIGATION ACTIVITIES

23
 

 A risk mitigation plan was develope

developed d for the LM6000P
LM6000PF F mechanica
mechanicall drive and included as part of the
purchase order with GE. The risk mitigations identified included not only the TQP identified items but lessons
learned from other large gas turbine driven compressor strings. A summary of the major mitigation items are
summarized as follows:

•   Design Reviews
-  Vendor Meetings
a. Pre-award Meetings
b. Coordinati on Meeting
Coordination
c. Design Review Meetings
d. 3D Model Reviews
e. FMEA Analysis
-  Project Meetings
a. HAZID
b. HAZOP/SOA
c. SID Review Meetings
•   Modeling and Simulations
1. Compressor Stage CFD
2. Impeller FEA
3. Compressor Stage Model Testing
4. Lateral Dynamic and Stability Analysis
5. Torsional Analysis
6. Dynamic Simulation
Simulations
s
•   Quality Control and Inspections
-  Pre-inspection Meetings
-  Weld Procedure and Qualificatio
Qualification
n Reviews
-  Verification of Materials
-  In-process Inspection of Manufacturin
Manufacturing
g and Assembly Activities
-  Dimensiona
Dimensionall Checks and Visual Inspections
-  Loop Checks

-  Cleanliness Checks
-   Assembly Com
Compliance
pliance Checks
Checks (per approved drawings)
-   Australian Stan
Standard
dard Compliance
Compliance Checks
-  Non-destructive Examination
-  Final Shop Inspections
-  Preservation and Shipping Inspection
•   Testing at Supplier Facilities
-  Hydrostatic Testing
-  Impeller Overspeed Testing
-  Rotor Balancing
-  Mechanical Run Testing
-  Unbalanced Response Testing
-  Gas Turbine Performance Testing
24
 

-  Compressor Performance Testing

-   Assembly Lea
Leak
k Test
-  FAT Testing
-  Control Functional Testing
-  Control Integration Testing
-  FLFS String Testing
•
  Field Inspections and Testing
-  Receipt Review
-  P&ID Walk-down
-  Preservation Checks and Monitoring
-  Megger Tests
-  Equipment Level and Co-planar Checks
-   Alignment Ch
Checks
ecks
-  Loop Checks
-  Control Function Checks
-  Sub-System Testing
-
  Gas Turbine Solo-run Testing
-  Nitrogen Run Testing
-  Performance Testing
-  Surge Testing
Of specific interest for the qualification
qualification,, special gas turbine testing and full load string testing were specified.

Gas Turbine Performance Testing

 All gas turbines shall undergo a full load performance test. The performance test is used to confirm the
power produced meets or exceeds the predicted power at a heat rate equal to or less than the quoted heat
rate (efficiency equals or is great than quoted). The gas turbine performance testing is the GE standard
production testing and only tests the gas turbine engine.

 All LM6000 engines

engines (GTCs, GTGs
GTGs and spares)
spares) for Wheatstone shall
shall include full power testing at 3060,
3060, 3450,
3600 and 3780 rpm. In addition to the emissions testing taken at full load, emissions testing will be
performed at 50% and 72% load at 3600 rpm.

Full Load Full Speed String (FLFS) Testing

The FLFS testing was required to validate the performance of the LM6000 PF driven compressor trains and
to mitigate operational and schedule risks with the use of LM6000 PFs as mechanical drivers. Based on the
qualification, it was determined that full load string testing shall be performed to confirm the following items:

•   Confirmation of train starting from settle-out conditions

•   Confirmation of train torsional dynamics
•
  Confirmation of compressor stability. Excitation of frequencies with a magnetic bearing and calculating
log decrement values is required for the HP Methane Compressor
Compressor Casing. For all other compressors, the
stability will be demonstrated by meeting the API-617 test criteria for sub-synchronou
sub-synchronous
s vibration levels.

25
 

•   Confirmation of LM6000PF thrust bearing loading

In addition to meeting the requirements from the TQP, the following items are confirmed during execution of
the string testing:

•   Operation of the evaporative cooler system and resulting inlet losses

•   Check-out of complete control system including Mark VIe and B/N System 1
•   Compressor thermal and aero performance under full load conditions
•
  Demonstration of gas turbine stability through rapid changes in gas composition. Nitrogen will be added
to the fuel gas to control the MWI. The tested range of fuel MWI shall cover the entire range specified in
the fuel gas datasheets.

The FLFS test proposed shall be full load, hydrocarbon gas test designed to match the volume reduction
ratio of the individual compressors. This test shall use a gas blend agreed that can simulate compressor
performance at full load.

Because the volume reduction ratios are matched across the stages, mechanical behavior is almost identical
to field operation. This results in similar aerodynamic forces at full load and therefore similar rotordynamic
behavior.

Testing at full load allows for fully loading of the entire train including the LM6000PF and all the drive
couplings. Full load testing also allows for demonstrating the startup and trip behavior of the train at the
suppliers facility.

Strain
and gagesthe
confirm shall be installed
results on the
of the train couplings
torsional to monitor the torsional performance of the compressor train
analysis.

26
 

ACKNOWLEDGEMENT

The authors would like to thank GE Power & Water and GE Oil &Gas for the support provided during the
qualification of the LM6000’s for mechanical drive applications. GE’s support has been instrumental in the
application of the engines for the Wheatstone LNG Project.

27