TECHNICAL PAPER

Title: Matching Steam Turbines with the New Generation of Gas Turbines

Authors: Ram Narula, James Olson, Justin Zachary—Bechtel Corporation

Date: May 2004

Publication/Venue: PowerGen Europe 2004

Matching Steam Turbines with the New Generation of Gas Turbines

By
Ram Narula
Principal Vice President
and Manager of Technology

Justin Zachary
Principal Engineer

James Olson
Assistant Mechanical Chief Engineer

Bechtel Power Corporation

5275 Westview Drive
Frederick, Maryland U.S.A.

Presented at PowerGen Europe 2004

Barcelona, Spain
May 2004

1.0 INTRODUCTION

To meet the requirements of the highly competitive combined cycle (CC) power plant market,
steam turbines (STs) have had to follow the same evolutionary path as heavy-duty gas turbines
(GTs). These two types of equipment have become even more interdependent with the
introduction of the new “G” and “H” GT technology classes. These new classes have created an
inseparable thermodynamic and physical link between the primary and secondary power
generation systems by using steam, instead of air, in a closed loop to perform most if not all of
the GT cooling duty. The previous generation of GTs, the “F” class, has also undergone
extensive modifications and upgrades. In addition to producing higher output, the “F” class
machines are producing substantially more exhaust energy for the bottoming steam cycle.

1
This paper describes Bechtel’s experience with modern STs in all phases of a CC project: from
the initial selection through construction, startup, and testing. The STs include all the major
manufacturers, such as Alstom, GE, Mitsubishi, Siemens Westinghouse, and Toshiba, in several
CC configurations.

The paper presents several CC projects that have been completed or are in advanced stages of
construction. These projects employ the new generation of GTs and STs.

2.0 CYCLE SELECTION AND OPTIMIZATION

In the last 10 years, STs in CC applications have had to evolve significantly, from small 80 MW
dual admission nonreheat configurations to multiple-pressure-admission reheat turbines with
outputs reaching the 350 MW range. Because of this rapid evolution, many original small turbine
designs have had to be modified or adjusted to compete in performance and capacity.

Significant basic differences exist between STs designed for CCs and conventional Rankine
cycle (RC) applications. First, feedwater heaters are not normally used in the thermal design of
the bottoming cycle for a CC; therefore, exhaust flows are much higher. Secondly, the steam is
added to the steam path flow at several pressures from the heat recovery steam generator
(HRSG) to allow the maximum amount of heat to be extracted from the GT exhaust energy. The
CC low pressure (LP) exhaust steam flow, when compared with RC LP steam flow, can be up to
35 percent greater than the main steam flow. Thirdly, CC designs may use supplemental or duct
firing in the HRSGs to compensate for the GT’s reduced output at high ambient temperatures,
which coincides with maximum summer demand for power. In the U.S., it has become quite
common to almost double the ST output by using massive amounts of supplementary firing to
capitalize on peak summer demand. Finally, competitive market pressures have pushed the ST
suppliers to offer a compact plant layout with axial steam exhaust using only two standard
cylinders, thus reducing the cost of manufacturing and installation as well as length of the
erection schedule (Ref 1).

2
Although a modular building block system with standardized components and turbine parts is
typically employed by all ST manufacturers, it must be remembered that the steam blade path is
individually designed for each application to achieve high efficiency. Even small blade path
improvements can translate into large operational savings.

Figure 1 shows the typical thermal efficiency values for three modules (high pressure [HP],
intermediate pressure [IP], and LP) recorded in several recent projects for both unfired and fired
cases. As the figure indicates, turbine cylinder efficiency in the range of 94 to 96 percent is not
uncommon. In this analysis, the power output for the fired case, with main steam pressure at 131
bara (1,900 psia), was 55 to 65 percent higher than the unfired case, with the main steam
pressure at 69 bara (1,000 psia). It should be noted that with sliding pressure operation, HP and
IP module efficiency does not change significantly between the fired and unfired cases,
indicating that ST operation at part load does not adversely affect the power plant heat rate. The
LP module efficiency, however, does change significantly, since it is affected by the specific
selection and sizing of the module and heat sink, which determine the ST operating
backpressure.

96
95.1
95
94.1
94
93.3
93.0 93.0
Efficiency

93
92.2 HP
92.2 92.3 92.7 92.1
92 91.9 92.3 IP
91 LP
89.9 90.0 90.1
90 89.7

88.8
89
88.2
88
fired unfired fired unfired fired unfired

A A B B C C

manufacturer and state of duct firing

Figure 1. Steam Turbine Module Efficiencies for Duct-fired and Unfired Cases

3
3.0 PLANT CONFIGURATION SELECTION

One of the earliest tasks in the development phase is selection of plant configuration. Bechtel
maintains a portfolio of standardized plant designs for each of the following configurations:
ƒ 1 x 1 (one GT, one HRSG, and one ST)
ƒ 2 x 1 (two GTs, two HRSGs, and one ST)
ƒ 3 x 1 (three GTs, three HRSGs, and one ST)

This approach allows the EPC contractor to build a plant that best suits the owner’s operating
goals. A single ST for multiple GTs is less costly than a separate ST for each GT. However,
multiple 1 x 1 configuration trains offer some significant advantages, as outlined below.

ƒ Phased construction flexibility. Owners can add units as dictated by market conditions.
ƒ Speed to market. As the construction duration of a 1 x 1 configuration is shorter than a 3
x 1 configuration, the first train can be brought on line more quickly and start generating
revenues, while the remaining units are still undergoing construction.
ƒ Closer matching with dispatch demand. A site with multiple 1 x 1 trains using
supplemental duct firing can better and more efficiently match dispatch requirements.
ƒ Greater plant redundancy. Each unit is completely redundant.
ƒ Optimized spare parts inventory. Identical components are used for all trains.

4.0 EQUIPMENT SELECTION

Before selecting equipment from different suppliers, a thorough investigation is necessary to

ensure that the owner’s pro forma objectives are met for power output, heat rate, startup times,
reliability and availability, etc.

The process includes an independent technology assessment of the equipment’s operating history
and quality control for the engineering and manufacturing processes. In addition, the
performance offered by the original equipment manufacturers (OEMs) for a specific project is
normalized and reconciled with past performance of the same equipment in a similar
configuration, as achieved on other projects. Bechtel maintains a performance database of all
past projects that is routinely updated with information from field tests.

4
Special consideration is given to cogeneration plants where selection of cycle pressures and
locations of the steam extractions is limited by the commercial availability of STs that can meet
the full range of specified conditions. Another decision needed on most cogeneration projects
involves the ability to cold start the ST while the plant continues to supply full process steam
requirements to the host. At the other extreme, it is also critical to determine whether to design
the ST and associated steam cycle equipment for the maximum steam case when no steam is
required for the process. In this scenario, a larger LP section and an increased capacity for
downstream electrical equipment (transformer, isophase bus, etc.) are required. In making the
selection, careful consideration should be given to economic benefits to be realized from the
availability of the extra power and operational constraints due to possibly infrequent occurrences
of such conditions (Ref 2). Equipment selection requires the unique expertise and value-added
service that only an experienced EPC contractor can provide to the customer.

5.0 CORPORATE EXPERIENCE

Over the years, Bechtel has executed more than 96 projects using 170 GTs and more than 96 STs
in CC applications. In the last 5 years alone, Bechtel has executed 25 projects, installing and
starting up 58 GTs and 38 STs.

Table 1 lists STs from major suppliers that have been used in Bechtel CC projects around the
globe. The following discussion is based on the experience gained in the design and
commissioning of these projects.

Table 1. Steam Turbines from Major Suppliers

Nominal
Power CC
Supplier No. of Units (MW) Characteristics Configuration
Alstom 7 250 2x1
Alstom 12 85/155 Unfired/duct-fired 1x1
Alstom 1 300 3x1
GE 1 300 2x1
GE 2 250 3x1
GE 1 150 1x1
Desalinization
GE 1 120 noncondensing unit 1x1

5
 Nominal
Power CC
Supplier No. of Units (MW) Characteristics Configuration
GE 3 200 2x1
Cogen with large
MHI 2 100 process steam 2x1
SWPC/MHI 1 120 1x1
SWPC 4 150 1x1
SWPC 1 295 3x1
Toshiba 1 300 3x1
Westinghouse 1 230 2x1
Total 38

6.0 STEAM PATH OPTIMIZATION AND SECTIONS

While ST design is a mature technology, manufacturers continue to improve the output and
efficiency of their equipment through improved blade designs and superior materials to
accommodate higher main and reheat steam temperatures. Computational fluid dynamics (CFD)
techniques coupled with rig and field tests have been extensively applied to develop three-
dimensional blades, vanes, and passages. Competitive pressures to maintain a small footprint
have also pushed new designs to focus on fundamental blade path parameters and offer increased
stage count, reduced blade root diameter, and optimized reaction values to minimize leakages.
Comparison of individual loss mechanisms can identify and correct the major contributors to
losses such as valves, and nozzle and blade profiles (Ref 3).

As design performance has improved, margins have decreased, making ST performance more
sensitive to detailed design issues such as surface finish and seal design.

7.0 EVALUATION OF THE EXHAUST SYSTEM: AXIAL, DOWNWARD, OR

LATERAL

With the high LP flows associated with CC STs, greater emphasis is placed on the last-stage
blade (LSB) dimensions and material. One of the major loss mechanisms in the ST is the kinetic
energy of the steam as it leaves the LSBs—the lower the kinetic energy that can be achieved, the
higher the resulting ST efficiency. The amount of loss is proportional to the ratio of the
volumetric steam flow rate through the LSBs and the annular area of the turbine exit. To

6
decrease this loss, a larger turbine exit annulus is required (Ref 4). Numerous challenges face
designers, who must marry complex 3D aerodynamic designs with robust mechanical features in
this erosion-prone environment. One typical example deals with aeroelastic instability, which is
one of the more challenging design problems for very large LSBs. Aeroelastic instability is a
vibration induced by flow, which occurs at off-design conditions in the regions of low axial
steam flow and/or at high condenser pressures. The phenomenon can lead to flow separation
from the airfoil, blade stall flutter, and buffeting of the blades. Some manufacturers have opted
for titanium as the LSB material because of its high strength/weight ratio and superior corrosion
resistance. However, titanium has reduced damping properties; therefore, it is generally
necessary to change blade construction from freestanding to an interlocked design for increased
mechanical stiffness. Despite numerous rig tests, theoretical analysis, and empirical data
verification for new blades, designs with proven operational records at severe off-design
conditions should be preferred.

The axial exhaust design is preferred because of its superior thermodynamic performance
through lower pressure losses, greater pressure recovery in the diffuser, and simpler and less
expensive construction. Because of the compact design and lower elevation, the capital cost of
the ST and the associated turbine building can be reduced by 40 percent over using a double flow
LP module with a bottom exhaust arrangement (Ref 1).

8.0 CONSTRUCTION

With the pursuit of compact, more efficient designs by ST manufacturers, construction tolerances
and startup requirements have become more critical. As the capacity of CC STs has continued to
grow, economic factors have pushed manufacturers to try to maintain the same cylinder sizes.
These designs have resulted in reduced design margins and equipment that is much more
sensitive to construction variables.

Reducing the number of bearings used to support the rotor in the turbine generator set allows for
shorter overall length and lower costs. Use of fewer bearings results in higher bearing loadings
and increased sensitivity to vibration, which in turn leads to additional startup delays due to the
added balancing time.

7
With the axial configuration, the location of the outboard LP bearing is an additional concern.
Normally, bearing housings are easily accessible from the outside, and the housing is maintained
at atmospheric pressure or at a slightly negative pressure by the vapor extractors on the lube oil
reservoir. With an axial design, this bearing is located within the exhaust region and is subject to
condenser backpressure conditions. In some cases, a small amount of leakage across the housing
flanges or inspection openings introduces oil into the cycle. Because this leakage is difficult to
detect, the resulting cleanup can be extensive and time consuming.

The axial exhaust design has an additional feature that is sensitive to design tolerances—the
large axial loads on the turbine during startup that are transmitted to the turbine foundations. The
foundation is expected to handle this loading with minimal deflection, and the slide plates under
the turbine are designed to freely allow the cylinder to grow. This large axial load, due to both
differential and thermal pressure, results in varying amounts of deflections, and the variation in
slide plate friction results in further variation. Further temperature stratification of the HP and IP
cylinder due to differential warming or cooling during startup or shutdown has resulted in rotor
binding if not adjusted properly. The sensitivity of the ST to these conditions results in further
delays in project startup.

The conversion of standard 50 Hz designs to 60 Hz conditions has also resulted in unforeseen

problems. Steam turbine sets that have had many reliable operating hours in their original 50 Hz
configuration have caused a variety of startup delays. The most significant difficulty is the
introduction of high vibrations that are transmitted through the STs and generators into the
piping or foundations.

9.0 STARTUP

The ST startup flexibility and commissioning time play a significant role in startup of the entire
CC plant. Heavy penalties for failing to bring the unit on line as scheduled mandate that
predicted unit startup times be achieved. In addition, due to higher fuel costs and increased
electrical reserve margins, CC plants are being dispatched as intermediate-duty units rather than
base load, as originally envisioned. Achieving the goal of a fast and reliable startup requires

8
careful design and integration of the ST and BOP requirements. On one hand, the ST supplier
should provide more flexible ST startup parameters (such as greater steam temperature mismatch
and more relaxed steam purity) while maintaining reasonable constant life consumption (Ref 5),
controlling low cycle fatigue by monitoring the maximum wall temperature differences and
permissible ramp rates. On the other hand, an EPC contractor can employ the entire arsenal of
auxiliary equipment available to assist in providing the narrow preoperational and operating
conditions that the modern ST requires (Ref 6) . Improving heat retention after shutdown, using
advanced water treatment systems to achieve steam purity more quickly, providing additional
lines for warm-up, and using an auxiliary steam boiler to reach desired condenser vacuum more
rapidly are only a few examples of the measures that can be taken. However, many of these steps
can significantly increase the overall cost of the project.

10.0 THERMAL PERFORMANCE

Contractually, the EPC contractor must conduct a performance test for the entire CC power plant
to demonstrate to the Owner that thermal performance guarantees are met. However, on many
occasions, component performance testing for GT, HRSG, and ST is also performed
concurrently with the CC plant test. It should be noted that a new ASME Performance Test
Code, PTC 6.2, is in the final stages of approval. This new code was conceived especially for
STs in CC applications, wherein the power output is the only guarantee. The new code demands
specific correction curves to account for the interaction between the HRSG and ST and sets forth
stringent requirements for instrumentation accuracy.

The outcome of ST tests conducted by Bechtel for 13 units, manufactured by 5 different

suppliers, is presented in Figure 2. As can be seen, many of the units did not meet their
guarantees, and additional work was required to improve their performance. It is not hard to
understand that some OEMs succumb to competitive pressures and offer aggressive performance
guarantees.

9
 Performance test results for steam turbines

6.0
4.48

4.0
% from guarantee

2.0 1.44
0.94
0.54
0.15
0.0

1 2 3 4 5 6 7 8 9 10 11 12 13
-2.0 -1.58
-2.07 -0.90 -2.10
-2.55 -2.00
-2.88
-4.0

-5.22
-6.0

Figure 2. Steam Turbine Output Guarantee Test Results

11.0 SUMMARY AND CONCLUSIONS

Steam turbine manufacturers are investing heavily to improve turbine efficiency by reducing
internal losses and by using improved construction materials. This process is ongoing, with high
expectation of further progress. The equipment manufacturers have identified and corrected
many deficiencies associated with introduction of new technologies. Regarding the failure of
many STs to achieve their performance goals, Bechtel, as an experienced EPC contractor, has
learned from this experience and has made provisions at the design stage as well as during
execution to ensure that the total power plant performance guarantees are met. The integration of
other major equipment (GT, HRSG, heat sink) was done to allow better-than-guaranteed
performance. In a domain under its sole responsibility, Bechtel has implemented other design
measures, resulting in additional design margins to compensate for the higher performance
sensitivity of the ST’s high performance requirements.

Even though performance test codes governing plant testing are quite detailed, demonstrating
performance shortcomings is not an easy task. Many issues are still subject to interpretation by

10
the parties involved in the testing, in particular measurement uncertainty. Because of the
significant monetary consequences, contract documents concerning test execution and the
methodology to determine liquidated damages (or bonuses) must be very explicit, and the actual
testing must be conducted “by the book,” with the full participation of the equipment supplier.

Having executed a large number of projects with STs from all the leading manufacturers, Bechtel
has accumulated a large database of actual performance. The lessons learned and best practices
have evolved into a due diligence process that is used at the equipment selection stage of all new
projects. The process is refined and improved as more data becomes available from newly
completed projects.

Continuous effort and collaboration among Bechtel, owners, and equipment suppliers have been
required to arrive at this juncture in process improvement. For all parties to accomplish their
goals under these conditions, design and construction approaches had to align with the business
objectives.

The relationships and modes of interaction among all the participants in building a power plant
from conceptual design to final acceptance play a significant role in successful project outcome.
Technical openness, free exchange of ideas, input based on personal experience, and most
important, mutual trust among the parties provide the right ingredients for a successful project
conclusion.

12.0 REFERENCES

1. Ernette, D. Steam Turbine Technology: The Way Forward. Power

Engineering International, Dec. 2000 issue.
2. Ugolini, D. J. Design Considerations for Combined Cycle Cogeneration
Plants. Paper presented at PowerGen International 2003, Las Vegas, NV, Dec.
10-13, 2004.
3. Maughan, J. R., L. D. Willey, et al. Development of the Dense Pack Steam
Turbine: A New Design Methodology for Increased Efficiency, IGPGC 2000-

11
 15008. Proceedings of 2000 International Joint Power Conference, Miami
Beach, FL, July 23-26, 2000.
4. Mujezinovic, A. Bigger Blades Cut Costs. Modern Power Systems, Feb. 2003
issue, pp. 25- 27.
5. Ulbrich, A., E. Gobrecht, et al. High Steam Turbine Operating Flexibility
Coupled with Service Interval Optimization, IJPGC2003-40072. Proceedings
of 2003 International Joint Power Generation Conference, Atlanta, GA, June
16-19, 2003.
6. Chrusciel, A., J. Zachary, et al. Challenges in the Design of High Load Cycling
Operation for Combined Cycles. Paper presented at PowerGen International
2001, Las Vegas, NV, Dec. 10-13, 2001.

12