Gas Turbine Performance Modeling and Test Data Analysis

The following sections present methods to model the individual effects occurring when water is encountered. These techniques are suitable both for predictions and test data analysis. Figure 10–27 presents a flowchart of how they interact; clearly many effects are common to several situations.

Formulae

Formula 10.1. Specific Heats Factor for Moist Air = fn(Water Air Ratio)

$\text{CPfac}=\left[\text{WAR}\text{.molar}\times \text{CPW}+\left(1-\text{WAR}\text{.molar}\right)\times \text{CPA}\right]/\text{CPA}$

$\text{WAR}\text{.molar}=\text{WAR}\times 28.96/18.015$

(i) CPfac is the ratio of CP for moist air to that for dry air.

(ii) WAR.molar is the ratio of water to dry air, by number of moles.

(iii) CPW, CPA are the specific heats of water and dry air.

(iv) WAR is the ratio of water to dry air, by mass.

(v) 18.015, 28.96 are the molecular weights of water and dry air respectively.

Formula 10.2. Gas Constant Factor for Moist Air = fn(Water Air Ratio)

$\text{Rfac}=\text{Ro/}\left(\text{MW}\times \text{RA}\right)$

$\text{MW}=1/\left\{\left(\text{WAR/}18\text{.}015\right)+\left[\left(1-\text{WAR}\right)/28.96\right]\right\}$

(i) Rfac is the ratio of R for moist air to that for dry air.

(ii) Ro is the universal gas constant, 8.31 kJ/mole K.

(iii) MW is the molecular weight of moist air.

(iv) RA is the gas constant of dry air, 0.28705 kJ/kg K.

(v) WAR is the ratio of water to dry air, by mass.

(vi) 18.015, 28.96 are the molecular weights of water and dry air respectively.

Formula 10.3. Gamma Factor for Moist Air = fn(Water Air Ratio)

$\text{GAMMAfac}=\left(\text{WAR}\text{.molar}\times \text{GAMMA}\text{W}\right)+\left(1-\text{WAR}\text{.molar}\right)\times \left(\text{GAMMAA}\right)/\text{GAMMAA}$

$\text{WAR}\text{.molar}=\text{WAR}\times \mathrm{28.96.18.015}$

(i) GAMMAfac is the ratio of GAMMA for moist air to that for dry air.

(ii) WAR.molar is the ratio of water to dry air, by number of moles.

(iii) GAMMAW, GAMMAA are the ratio of specific heats for water and dry air.

(iv) WAR is the ratio of water to dry air, by mass.

(v) 18.015, 28.96 are the molecular weights of water and dry air respectively.

Formula 10.4. Enthalpy of Saturated Steam (kJ/kg) = fn[Temperature (°C)]

$\text{H}=-{7.352}^{-06}\times {\text{T}}^3-{2.333}^{-03}\times {\text{T}}^2+2.437\times \text{T}+24926949/\left(\text{T}-387.5\right)$

Accuracy is within 0.6%, range is 0–370°C.

Formula 10.5. Temperature Level of Evaporation (°C) = fn[Pressure(bar)]

For P = 1 to 25 bar:

$\text{T}=-{18.11}^{-05}\times {\text{P}}^6+0.0014006\times {\text{P}}^5-0.043\times {\text{P}}^4+0.67482\times {\text{P}}^3-5.9135\times {\text{P}}^2+33.2486\times \text{P}+72.1585$

For P = 25 to 210.5 bar:

$\text{T}=-{1.11726}^{-11}\times {\text{P}}^6+{8.97543}^{-09}\times {\text{P}}^5-{2.9476}^{-06}\times {\text{P}}^4+0.00051476\times {\text{P}}^3-0.05329436\times {\text{P}}^2+3.933136\times \text{P}+152.0676$

Accuracy is within 0.5°C.

Formula 10.6. Partial Pressure of Water Vapor in Moist Air (Bar) = fn[Specific Humidity, Pressure (Bar)]

$\text{Pw}=\text{P/}\left[\left(\text{0}\text{.}622\text{/SH}\right)+1\right]$

(i) Pw is water vapor partial pressure.

(ii) SH is specific humidity, kg water vapor per kg dry air.

(iii) 0.622 is the ratio of molecular weights of water and dry air.

Formula 10.7. Enthalpy of Liquid Water (kJ/kg) = fn[Temperature (°C)]

$\text{H}={3.1566}^{-12}\times {\text{T}}^6-{2.9348}^{-09}\times {\text{T}}^5+{1.0407}^{-06}\times {\text{T}}^4-{0.16703}^{-03}\times {\text{T}}^3+0.0120915\times {\text{T}}^2+3.87675\times \text{T}+\text{0}\text{.}74591$

Formula 10.8. Latent Heat of Melting Ice (kJ/kg)

$\text{DHif}=333.5$

Formula 10.9. Work on Liquid Water in Compressor (W) = fn[Water Flow (kg/s), Mean Blade Speed (m/s)]

$\text{DPW}=0.5\times \text{Wwater}\times {\text{U}}^2$

For an axial compressor this work is done in each stage.

Formula 10.10. Enthalpy of Dry Steam (kJ/kg) = fn[Temperature (°C), Pressure (Bar)]

Superheated:

$\text{H}={2.98}^{-04}\times {\text{T}}^2+1.83\times \text{T}+2500-{5.14207}^{08}-\text{P}{\left(\text{T}+276\right)}^3-\left({1.03342}^{37}\times {\text{P}}^3-{6.42613}^{31}\times {\text{P}}^5\right)/{\left(\text{T}+276\right)}^{17.787}$

Supercritical:

$\text{H}=\text{P}\times \left[{\text{T}}^3\times \left(1+0.001634\times \text{P}\right)+{\text{T}}^2\times \left(1094.941-1.663087\times \text{P}\right)-8169907\right]/{5.37}^7+0.004505\times \text{T}\times \left(738.0074+\mathrm{31,929.78}/\text{P}-0.30077\times \text{T}\right)+611.736+39.63429\times \left(9.551098-0.002642\times \text{P}\right)\times \text{arctan}\left[\left(\text{T}+0.005184\times \text{P}\right)\right]$

Formula 10.11. Specific Heat of Steam (kJ/kg K) = fn[Temperature (°C), Enthalpy (kJ/kg)]

$\text{CP}=\text{dH/dT}=\left[\text{H}\left(\text{T}+\text{DT}\right)-\text{H}\left(\text{T}-\text{DT}\right)\right]/\left(2\times \text{DT}\right)$

Formula 10.12. Theoretical Maximum Rate of Condensation = fn(Temperature, Pressure, Mach Number, Humidity)

The following outlines an approximate method:

The main approximation is to neglect the loss in total pressure that occurs due to the momentum change when heat release causes an increase in flow velocity.

ATS Advanced 3D Compressor

The W501G incorporates the first 16 stages of the 27:1 pressure ratio, 19-stage, ATS compressor with slight modification of the last 3 stages, and with vanes 1 and 2 fixed instead of variable. As a result, the W501G operates the ATS mass flow of 558 kg/sec (1230 lbs./sec), but at a pressure ratio of 19:1—optimized for the G cycle performance.

The design is based on three-dimensional inviscid flow analyses and on custom-designed, controlled-diffusion airfoil shapes. Controlled-diffusion airfoil design technology has been successfully applied in the aircraft industry for many years. The mechanical integrity of each stationary and rotating airfoil was verified by finite element analyses to satisfy steady stress and endurance strength criteria. Each airfoil was tuned to avoid potentially harmful resonant frequencies.

To verify the aerodynamic performance and mechanical integrity of the new high-pressure ratio design, the full-scale W501ATS compressor was manufactured and tested in 1997 at a specially-designed facility at the US Navy Base in Philadelphia. To reduce the required power to the 25 MW available at the test facility, the compressor test was carried out at subatmospheric inlet conditions.

The ATS-developed compressor technology has also been retrofitted into the W501F product line. Using the analytical techniques developed and proven in the ATS program, the W501F compressor was upgraded in the latest improvement to this successful frame. This advanced compressor is utilized on new W501F units. In addition, the redesigned compressor can be retrofitted to any 42 W501Fs that were built with the original W501F compressor. Applying this ATS technology to the W501F expands the benefit of the ATS program since the W501F compromises more than 70% of future units that are sold or on order at Siemens Westinghouse.

Brush Seals and Abradable Coatings

To minimize air leakage, as well as hot gas ingestion into turbine disc cavities, brush seals were incorporated into the W501ATS engine design at several locations: under the compressor diaphragms, at the turbine disc front, under turbine rims, and at the turbine interstages. Tests were carried out on test rigs for the different brush seal locations to develop effective, rugged, reliable, and long life brush seal systems. At the Philadelphia US Navy Base, full-scale brush seals were tested as part of the ATS compressor test, which verified the brush seal low leakage and wear characteristics.

To date, ATS-developed brush seals have been successfully incorporated and operated in W501G and later W501F product lines. Pre- and post-upgrade tests have demonstrated performance improvement in retrofit applications.

Considerable performance benefits can be obtained by reducing compressor and turbine blade tip clearances. Abradable coatings permit tip clearances to be minimized without fear of damaging hardware, and they provide more uniform tip clearances circumferentially. Abradable coatings, identified for compressor and turbine applications, were tested to determine abradability, tip-to-seal wear rate, and erosion characteristics. These ATS-developed abradable coatings have been incorporated into the W501F and W501G compressor and front turbine stages (1 and 2). The later turbine stages (3 and 4) employ shrouded blades with honeycomb seals.

Closed-Loop Steam Cooling

Using closed-loop steam cooling on transitions and turbine stationary components has two advantages. First, more compressor delivery air is available for premixing with the fuel gas in the combustor hot end. This allows very lean premixed combustion and makes possible the restriction of NO_x emissions to single digits. Second, closed-loop steam cooling significantly improves cycle efficiency by reducing the amount of chargeable air used for cooling and sealing.

The ATS transitions, which duct the hot combustor exit gases to the turbine inlet, are closed-loop steam cooled with air as an alternate coolant at part load. Steam enters the engine through four external connections and is routed to each transition supply manifold through internal piping. The supply manifold feeds the steam to an internal wall cooling circuit. After cooling the transition walls, the steam is collected in an exhaust manifold and ducted out of the engine. The W501G employs the ATS transition.

High-Temperature Thermal Barrier Coatings

Thermal barrier coatings (TBC) are an integral part of the W501ATS engine design. A development program is in progress to develop an advanced bond coat/TBC system with a projected service life of more than 24,000 hours. Different bond coats and ceramic materials were evaluated under accelerated oxidation test conditions and down selected. An advanced bond coat/TBC system mechanical integrity and durability was demonstrated in more than 24,000 hours of cyclic testing at 1010°C (1850°F). This advanced bond/coat TBC system has been incorporated on the W501G Row 1 and 2 blades. This coating will improve both the life and durability of these parts, and it can potentially improve future engine performance by reducing the amount of cooling air required.

Steam-Cooled Vane

Development activities are focused on extending the W501G frame to ATS efficiencies through the introduction of additional technology advancements. The next major step will add a thin-walled, closed-loop, steam-cooled Row 1 turbine vane to the W501G. The steam-cooled vane will extend the benefits of the steam-cooled transition by eliminating most cooling air from the Row 1 vane. The result will be a combination of increased rotor inlet temperature and decreased burner outlet temperature. The benefit will be improved efficiency and reduced NO_x. Single-crystal casting trials have been successfully completed at PCC Airfoils, Inc. in Mentor, Ohio.

The ATS steam-cooled vane will be first tested in an engine sector rig. The test rig consists of a full-scale combustor basket and transition and a ¹⁄₁₆th-sector vessel, which will operate up to full ATS pressures and temperatures. The rig will be located at Arnold Engineering Development Center at the Arnold Air Force Base in Tennessee. The vane will be instrumented to verify analytical predictions of metal temperatures, heat transfer coefficients, and stress.

After validation in the ¹⁄₁₆th-sector rig, the vane will be retrofitted into a W501G. A comprehensive test program will verify vane performance and improved plant performance. Test parameters will include vane metal temperature, stress, and steam temperatures.

Coatings

Thermal barrier coatings (TBC) are an integral part of the W501ATS engine design. A development program is in progress to develop an advanced bond coat/TBC system with a projected service life of more than 24,000 hours. Different bond coats and ceramic materials were evaluated under accelerated oxidation test conditions and down selected. The mechanical integrity and durability of an advanced bond coat/TBC system was demonstrated in more than 24,000 hours of cyclic testing at 1010°C (1850°F).

Under a related program, DOE-Oak Ridge National Laboratory (Contract DE-ACO5-95OR22242), new ceramic compositions and TBC concepts were identified which have a sintering resistance and phase stability superiorted to that of 8YSZ TBC. These compositions and concepts are being further optimized and will be transferred to components for an 8000 hr engine demonstration under a DOE-Chicago contract DE-FCO2-000H11048.

FEA and CFD

Finite element analysis and computational fluid dynamics are very useful modeling techniques mechanical engineers employ to design hot section components.

FEA is used to predict the operating stresses and temperature that a component will experience in service. Heat transfer and fluid dynamic boundary conditions determined by CFD are applied to the FEA model. The resulting stresses and temperatures are then entered into a material model to predict the component “design life.”

From an O&M perspective the overall system model can be used to evaluate remaining life. Even more useful, these models can be used to predict remaining life based on operating condition. For example, if the user were to operate their components at a different load than design, the life impact of this can be directly estimated with the model.

Long-Term Material Steady Behavior

Such behavior can be characterized using Larson-Miller curves and creep rate equations combined with safety factors to account for the unknowns in the model. The Larson-Miller curve is a generally accepted method used to determine the creep life of a component given time, temperature and stress and is represented in Figure 10–36.

Material creep strain rates are used in modeling the nonlinear plasticity experienced in hot section components. For example, the long-term load shedding of a turbine root form can be estimated to predict a more representative design stress. Material creep strain rates can be represented by the classical power law for creep known as the Bailey-Norton law. The expression for uniaxial creep strain in terms of the uniaxial stress and time is represented in the following equation:

$\text{εc}={\text{C0}}^{\ast }\sigma \left(\text{C}1\right)\text{t}\left(\text{C}2\right)\text{e}\left(-\text{CT/T}\right)$

where

Long-Term Material Cyclic Behavior

This behavior can be classified as either high cycle or low cycle fatigue. The classification of fatigue behavior into two types characterizes the different material model damage mechanisms. Low cycle fatigue or LCF deals primarily with 1000s of cycles.

From a practical perspective, LCF is tied with engine starts or load variations which are linking with thermal and stress cycles. High cycle fatigue or HCF results from cycles much greater than 1000s and more likely in the 10⁷ range. Practically HCF is associated with cyclic stress loading tied to the engine operating speed. For example, blades rotating past a number of vane wakes will accumulate a number of stress cycles every time the blade passes by during one revolution. Both HCF and LCF are dealt with using different material models.

HCF can be modeled using a Goodman diagram as illustrated in Figure 10–37. A mechanical designer will design the HCF loading of the part to fall in a region on the Goodman diagram where safe operating stresses occur that ensure near infinite operating cycles.

LCF material properties are primarily deduced from laboratory testing. Most of the laboratory testing conducted to characterize the low cycle fatigue resistance of materials is carried out at constant temperature using simple waveforms with steady strain rates. By contrast a startup or trip cycle on a gas turbine hot gas path can generate quite a complex stress cycle with transient thermal loading.

To predict the fatigue life of a component exposed to the more complex cycles based on the lab test results, some simplifications must be made. The first simplification addresses the isothermal nature of most lab testing. The assumption is made that the damage accumulated during the thermal-mechanical fatigue cycles is equivalent to damage generated in isothermal testing at the peak temperature of the thermal-mechanical fatigue cycle.

Secondly, the differences between the complex strain cycle experienced by parts and the simpler cycles used in lab testing must be addressed. Under low cycle fatigue conditions, fatigue life is a function of the strain range (the maximum strain in the cycle less the minimum strain in the cycle) and test results are most commonly presented as graphs of cyclic life to failure versus strain range.

The complicated strain cycle a component sees in service is thus characterized by the strain range, to predict low cycle fatigue life from simple lab cycles. The predicted life can then be read from the LCF curve for the material at that strain range (see Figure 10–38).

Mechanical approach pros:

Mechanical approach cons:

Systems Engineering Approach

As demonstrated above metallurgical, mechanical and performance engineering takes a very different approach to O&M. Although different, each discipline can be tied together in one central computing server and provided as a web service to facilitate O&M.

A component tracking and management system or CTMS has been developed to track component data at the serial number level. Its features are listed as follows:

Two potential CTMS application examples follow.

Example 1: Trailing Edge Thinning of Rotating Frame 7EA Row 2 Turbine Bucket. Metallurgical analysis revealed an engine axial inter-granular crack along the span of the trailing edge up to 7/8″ long. Metallographic investigation concluded that a creep related mechanism was responsible for the trailing edge cracking. Additionally oxidation and wear had reduced the thickness of the trailing edge to 76% of the new condition.

Mechanical analysis showed that the strain range experienced during a shutdown and a full load trip was between 0.003 and 0.004. The boundary conditions that resulted in this strain range were determined from a transient performance analysis which provided the gas path transient temperatures, pressures and flows. The LCF life was predicted to decrease significantly with reduced trailing edge thickness. So much that a 25% thickness reduction limit was set for the bucket’s trailing edge before retirement from service.

As one can see, each analysis on its own could not establish a minimum wall thickness criterion. Metallurgical analysis revealed the type of failure. Mechanical evaluation predicted strain ranges using the result of performance engineering transient gas path analysis. By combining the results of the three disciplines, a retirement criterion has been established. To implement this in practice a systems health monitoring solution would track each bucket by serial number in a database. This database would contain a remaining life quantity for each serial number. GTAP would be added to the health monitor to track gas path conditions to calculate remaining life in the part based on the mechanical engineering models described above. This remaining life number would be updated periodically in the database and when it reaches a certain number, alarms and notifications would be given by the health monitoring system to the O&M engineer.

Example 2: Over-/Under-Firing a Frame 3 Gas Turbine. In some cases when the market conditions are right an operator may elect to over-fire their gas turbine to get additional capacity out of the unit during peak operation. The economic conditions in certain peak hours can be so advantageous that over-firing is very attractive and the consequential impact on maintenance needs to be estimated. Conversely there are cases where a reduced firing rate would substantially offset the maintenance cost to make it economically viable. These decisions can be made with a greater degree of confidence if a systems approach to O&M were implemented.

By archiving and tracking component remaining life in a database, the life impact of over-/under-firing can be managed and optimized. Metallurgical and performance history will dictate the life consumed thus far and forward looking mechanical analysis can predict the impact of the load changes on the remaining life. Figure 10–39 shows the impact of over-firing on the Row 2 bucket as a function of the original life and % over fire. The integral of this curve could be taken to estimate the remaining life in the bucket and then be used to optimize load vs. economics.

The next case is another example of industry working in concert with and benefiting the practical education within academia. (See Chapter 17, Training and Education.)

Improving the Permitting Process

Since the air permit is usually secured as quickly as possible, “ideal case” assumptions are sometimes made. The concern is that key factors—project design, fuel selection, type of operation, back-end equipment selection, and performance—may be determined upon submission of the air application even if all pertinent information and inputs are not available. While this tactic is useful, it is invaluable for projects having advanced GTs as prime movers.

The process definitely requires a team effort. The environmental consultant provides valuable input to the permitting strategy, giving the owner and EPC contractor broader project perspective. Both owner and contractor will ultimately need to consider issues such as constructability, initial compliance testing, long-term emission compliance, and operability and maintenance. Consequently, the process should be structured and comprehensive, encompassing the following:

Examples of key site arrangement issues include:

Equipment Selection

Before selecting the equipment from different suppliers, a thorough investigation should be conducted to ensure that the owner’s pro forma objectives for power output, heat rate/efficiency, exhaust energy, emissions, reliability, and availability are met.

The information in the database is constantly updated with results from field tests. Figure 10–40 represents a selection of 37 GT units recently tested.

When establishing the HRSG and ST design conditions, it is a good engineering practice to allow for some variability of GT exhaust flow and temperature values. This way, the design could accommodate either shortfalls or better-than-guaranteed exhaust energy (flow and temperature). Thus, more realistic and competitive values can be predicted for the CC performance.

Our experience indicates that owners/operators are well advised to engage an experienced and bankable EPC contractor in purchase or reservation agreements for GTs or power islands. As mentioned above, such collaboration becomes crucial for projects involving advanced GTs. The EPC contractor can verify that terms and conditions essential to managing project design, construction, and commissioning are adequately covered in these agreements. An EPC contractor, with direct experience with this type of GT, can also work with the equipment supplier and the customer to ensure that the scope is complete and all interfaces are well defined. Areas where the EPC contractor can add value to an owner’s reservation agreement include:

Gas Turbine Startup and Commissioning Issues

EPC contractors are responsible for starting up a CC plant and incur significant penalties if successful, timely operation is not achieved. Therefore, EPC contractors must evaluate the cost-effectiveness of approaches that facilitate rapid startup. The problem becomes even more critical for merchant power plants, where cycling and part-load operation are often required. Depending on specific completion requirements and previous experience with a particular type of advanced GT, the startup schedule must allow time for unforeseen events. Some examples of these include:

However, this activity impacts the EPC contractor. The execution schedule is extended to perform a water wash of the compressor before the tuning process and to install and remove temporary instrumentation for full-blown performance testing of the GT. Because emissions limits must be met at all ambient conditions, adjustments made in the field might modify the performance correction curves for ambient temperature.

Additional Gas Turbine Lessons Learned

Advanced GTs with DLN combustion technology may require the ability to both heat and cool fuel gas during startup. As GTs proceed to higher outputs and efficiencies, the fuel gas supply pressure requirements typically exceed 30 barg and can sometimes exceed 35 barg for can-type combustion systems.

These supply pressures often require supplemental fuel gas compression. During startup at low fuel flow rates, operation of supplemental compressors can result in significant increases in fuel gas temperature due to compression heat. However, DLN combustors typically require “cold” fuel for initial startup. As a result, a startup or pilot cooler may be required downstream of the supplemental fuel gas compressor. Later on, as the GT ramps up in load, heated fuel would be required for full pre-mix operation and the cooler would need to be isolated or bypassed.

On projects with widely varying fuel gas supply pressures, supplemental compressors can often be bypassed, even during startup. Therefore, both a dewpoint heater and a startup/pilot cooler would be required for the fuel gas supply system to meet startup fuel gas requirements under all scenarios.

Exhaust Loss Curves

One of the important features of STs used to integrate other components and, in particular, optimize the heat sink is the exhaust loss curve. The ST thermal kit provided by the manufacturers to plant designers contains information on LP exhaust losses in the traditional form of an “exhaust losses curve.” This curve gives the specific enthalpy loss for an exhaust average steam velocity. The steam velocity value depends on the back pressure and/or steam flow. The use of this curve requires corrections for moisture content. Since the average steam exhaust velocity is a function of the volumetric flow for a given geometry, its value depends on both back pressure and steam mass flow. Figure 10–41 provides a number of exhaust loss curves for various exhaust areas. The increased exhaust loss at lower exhaust velocities (less than 550 feet per second [fps]), occurring at low part-load conditions, is due to the formation of a reverse vortex at the blade root.

Development of new generation and often non-conventional profiles, with large variations along the blade height, has produced considerable benefits in terms of the stage’s efficiency and reduced exhaust losses. Figure 10–42 presents the exhaust curve loss for blades from two manufacturers that have almost identical exhaust and blade lengths. It can be concluded that a more advanced 3D profile design (Curve A) leads to a lower exhaust loss. Since the total exhaust loss curve also includes the contribution of the exhaust hood, its design should be aerodynamically effective.

Steam Turbine Startup and Commissioning Issues

ST startup flexibility and commissioning time play a significant role in startup of the entire CC plant. Due to higher fuel costs and increased electrical reserve margins, CC plants are dispatched as intermediate-duty units rather than base load units, as originally envisioned. Achieving the goal of a fast and reliable startup requires careful design and integration of the ST, GT, 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, 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 the process. Examples of such measures include means to improve the heat retention after shutdown, design of advanced water treatment systems capable of achieving steam purity more quickly, provisions for additional warmup lines, and use of an auxiliary steam boiler to reach desired condenser vacuum more rapidly.

Additional Steam Turbine Lessons Learned

Some GT/HRSG/ST combinations are better suited for rapid cycling than others. It is not possible to accurately predict CC power plant startup times from the typical startup curves for each piece of equipment. Each major equipment vendor will make assumptions about startup conditions that are most favorable for that vendor’s individual piece of equipment. These assumptions are usually different, often conflicting, and sometimes incompatible. The overall startup integration must be considered early in the project, and consistent requirements need to be provided at the bid phase for each vendor. The drive to increase ST efficiencies results in tighter clearances to reduce leakages, but these usually require longer startup times to avoid rubbing and reduce thermal stresses. Stringent steam temperature matching requirements may also be requested. These requirements vary between ST vendors and cannot always be met by every GT/HRSG combination. The GT operating at low loads has a limited capability to maintain high exhaust gas temperature. This capability varies significantly between manufacturers.

Combining a GT with limited ability to control exhaust gas temperatures at low loads with an ST that has very strict temperature requirements at startup can lead to unacceptably long startup times (greater than 8 hours) for cycling units. The HRSG may not be capable of accommodating the GT and ST requirements simultaneously without significant bypass capability. Although this scenario might seem unlikely to occur, Bechtel has witnessed this circumstance more than once, when owners procured each piece of major equipment separately, considering only performance and price with no attention to integrated startup requirements.

Cycling Considerations

The merchant plant concept implies that electric power must be supplied to the grid only when it is commercially justifiable. Such operation requirements must typically be met on very short notice. Therefore, these plants need to be started up quickly and must have flexible operating ranges. Merchant plants are normally in cycling service, which can be considered part-load operation or daily on/off duty. Heat sink considerations for a cycling plant include the following:

Balance of Plant Equipment Selection

BOP equipment for a CCPP includes boiler feedwater pumps, condensate pumps, circulating water pumps, closed cooling water pumps, air compressors, fuel metering equipment, etc.

Proper system integration of BOP equipment also requires that interconnecting piping/support systems be optimized to the power island (GT, ST, HRSG, heat sink). Based on the arrangement of the plant and amount of BOP equipment, the engineering evaluation will start by determining the amount of circulating water, steam, air, and fuel required to operate the plant at the expected power output. The optimization of sizes, material, layout, and cost of BOP equipment and critical piping is no less important to the performance and cost of the entire plant than the power island major equipment. One of the important lessons learned was to review and continuously validate the specifications for each of the major BOP systems (main steam, feedwater, condensate, and circulating water) to rapidly respond to new requirements and incorporate the experience gained from previous projects.

Auxiliary Loads

A critical step in each project is to assess, with a high level of accuracy, the plant’s total auxiliary load at given guaranteed conditions. Auxiliary loads, representing the amount of power consumed by each piece of equipment during normal plant operation, can vary significantly due to the type of equipment and its configuration. The theoretical estimation at the beginning of the plant design must also be confirmed during performance testing. This process requires a careful engineering evaluation of the equipment in operation and the utilization factor to be applied to each piece of equipment.

A good indication of a successful design is to achieve a close match between the auxiliary load calculated and the actual values measured during the test. By conducting upfront detailed engineering to obtain actual pump performance curves and using a database of past projects’ auxiliary loads (both calculated and measured), Bechtel was able to predict with a very high degree of accuracy the auxiliary loads for the projects associated with the turbines presented in Tables 10–8 and 10–9.

Material Selection

Tailoring material selection to the specific operating conditions of the power plant reduces the cost and ensures proper plant operation. Over the years, Bechtel has developed a material selector guide that assists the design engineer in selecting the proper material and in matching its properties to the required design parameters.

One interesting lesson learned relates to the selection of a high temperature pipe material. In the initial stage, the selected material was adequate and met the operational requirements. However, a better material, which exceeded the requirements, was found to be less expensive because the pipe diameter could be smaller and a standard schedule size could be used instead of minimum wall pipe. In another example of tailored material selection, metal pipe was replaced with high-density plastic in a low-pressure and temperature application, resulting in substantial savings.

Water Chemistry

Controlling the amount of dissolved oxygen in the condensate is critical, especially during startup. The most critical operational backstop to prevent dissolved oxygen (DO) from entering the feedwater is to avoid breaking condenser vacuum overnight (or ever). Maintaining condenser vacuum and preventing the hotwell from subcooling dramatically decreases the morning startup time and almost eliminates the huge swings in chemistry associated with cold startup. Following are lessons learned on how to minimize corrosion potential due to DO:

If the HRSG is taken off line for maintenance under cold conditions, apply and monitor a nitrogen blanket above all water spaces and follow the manufacturer’s recommendations for chemical dosing of the drums. As the steam space pressure decreases, align the nitrogen system to push and replace the steam, which will minimize air in-leakage.

Plant Configuration

One of the earliest tasks in the development phase is selection of plant configuration. Large contractors tend to maintain standardized plant designs to accommodate the OEM preferred packages. (Note that OEMs are frequently their own contractor and owner/operator.) For this contractor, they include:

A single ST for multiple GTs is less costly than a separate ST for each GT. However, multiple 1 × 1 configuration trains offer some significant advantages, as follows:

OEM Selection

Before selecting equipment from different suppliers, a thorough investigation is necessary to ensure that the plant owner’s requirements for power output, heat rate, startup times, reliability and availability, etc. are met.

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.

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. Most cogeneration projects require 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. For this instance, a larger LP section and an increased capacity for downstream electrical equipment (transformer, isophase bus, etc.) are required. Economic benefits to be realized from the availability of the extra power and operational constraints due to possibly infrequent occurrences of such conditions must be considered. Equipment selection requires the unique expertise and value-added service that only an experienced EPC contractor can provide to the customer.

Experience History

Table 10–12 lists STs from major suppliers that have been used in this contractor’s projects (96 projects using 170 GTs and more than 96 STs in CC applications).

Figure 10–64 provides a comparison of the relative efficiencies of all three configurations (1 × 1 × 1, 2 × 2 × 1, 3 × 3 × 1) as a function of the amount of process steam duty. Relative efficiency RCOG_SEP is defined as the ratio between the thermal efficiency of each configuration and a reference efficiency of a conventional cycle consisting of a combined cycle producing electricity and a boiler producing process steam. In the example, for 300 MW process steam duty, a 3 × 3 × 1 configuration is slightly more efficient (RCOG_SEP = 1.1323) than a 3 × (1 × 1 × 1) configuration, each exporting 100 MW (RCOG_SEP = 1.1304). However, for cases where the steam duty varies, three trains, each in a 1 × 1 × 1 configuration, provide substantially greater efficiency than a 3 × 3 × 1 configuration due to the increased operating flexibility.

Steam Path Optimization

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. Pressures to maintain a small footprint have improved blade path parameters and offer increased stage count, reduced blade root diameter, and optimized reaction values to minimize leakages.

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 major loss 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 decrease this loss, a larger turbine exit annulus is required. Challenges face designers, for instance 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 use 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.

The axial exhaust design has 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 turbine building can be reduced by 40% over using a double flow LP module with a bottom exhaust arrangement.

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.

Fewer 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.

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 in 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.

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 must 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 experience a variety of startup delays. The most significant is high vibrations that are transmitted through the STs and generators into the piping or foundations.

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.

Also, 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 careful integration of the ST and BOP requirements. 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, controlling low cycle fatigue by monitoring the maximum wall temperature differences and permissible ramp rates. The 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.

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.

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. ASME Performance Test Code, PTC 6.2, is 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 this contractor (13 units, 5 OEMs) is in Figure 10–65. As can be seen, many of the units did not meet their guarantees, and additional work was required to improve their performance. OEMs may under competitive pressures offer aggressive performance guarantees.

Cogeneration

Factors to be considered include:

Steam Turbine

The steam turbine (ST) is started with power output in mind, which means that it is rapidly loaded with reduced and nearly constant steam temperatures until the bypass valves are closed with the GTs at base load. This start method eliminates steam vents to atmosphere at the HRSG, due to the incorporation of an auxiliary boiler. Since there are no HRSG- or ST-induced holds of the GT startup process, special care must be taken in handling the generated steam in order to meet the optimal steam temperature for fast loading of the ST.

The steam turbine is designed for full-arc admission without the use of a control stage. This design greatly reduces blade stresses during the startup process. The combination of full-arc admission with variable-pressure operation (which has become a standard for HRSGs in the industry) provides high efficiency, maximum operating flexibility, and minimized maintenance.

The application of high-capacity steam bypass systems for HP, IP, and LP steam is key to the operating flexibility and reliability/availability of any plant. The fast-start plants incorporate special provisions to route the HP bypassed steam into the cold-reheat of the HRSG to minimize thermal stressing of the HRSG reheater section during startup.

In order to provide further flexibility, the authors’ company’s patented Turbine Stress Controller (TSC) is supplied. The TSC consists of a stress evaluation system based on an ST startup program with an on-line life cycle counter. It calculates and controls stresses in all critical ST “thick wall” components (including the valves, HP casing and rotor body, and the IP rotor body) depending on three different ramp rates, i.e., planned, accelerated, and fast. The function of the TSC is to control the startup ramp rates in order to minimize material fatigue within the constraints of operating requirements and at the same time to calculate the cumulative fatigue of the monitored turbine parts. The TSC assigns an appropriate number of equivalent starts for each planned, accelerated, and fast start. With this knowledge of the effect of each type of start on the maintenance schedule of the steam turbine, the owner can make prudent business decisions regarding whether or not the asset should be used to satisfy transient dispatch requirements.

Thermal Cycling

Thermal cycling is the premier offender in HRSG system failures. The cumulative fatigue damage from cycling cannot be reversed. Eliminate cyclic fatigue damage, and the HRSG’s life cycle will increase substantially.

From vendor data (about combating cyclic fatigue):

Thermal Stress

The need to understand and positively control and guide thermal expansion within the boiler is paramount. The HRSG vendors offer these design features to minimize thermal stress:

SCR and CO Contribution to Extended Startup Times

The presence of a CO catalyst does not contribute significantly to the startup time required for the HRSG. Since the CO catalyst is located forward in the HRSG, the CT exhaust gases typically heat up the blocks within 30 minutes.

Conversely, the selective catalytic reduction (SCR) system requires significant startup time, especially if the unit is cold. The ammonia injection grid feed piping contains a tremendous amount of metal mass and requires significant heat to obtain operating temperature. If the HRSG uses an internal gas return system (hot gas feed from downstream of the catalyst) to vaporize the ammonia, then an additional heat source is required to enable fast HRSG startup by quickly getting the SCR catalyst and piping in the required operating temperature range of 600–800°F. Several methods exist to preheat the catalyst blocks and piping:

Water Chemistry

Controlling the amount of dissolved oxygen in the condensate is critical, especially during startup. The most critical operational backstop to keep dissolved oxygen (DO) out of the feedwater is not to break condenser vacuum overnight (or ever). Maintaining condenser vacuum and preventing the hotwell from subcooling dramatically decreases the morning startup time and almost eliminates the huge swings in chemistry associated with cold startup. Successful strategies recommended to minimize corrosion potential due to DO include:

Water Treatment System Design Features

Equipment design approaches that should be evaluated to minimize water treatment problems inherent in cycling and part-load operations include: feedwater, drum water, and steam purity requirements. Ramp-up rates may be limited by the need to avoid exceeding ASME and equipment suppliers’ limits for transient drum water and steam purity.

Figure 10–66 shows a typical fluctuation of drum water quality as an 1800 psig steam generator is cycling.

As shown in Figure 10–67, the startup data graphs for an actual combined-cycle plant, a typical warm startup should take about 2½ hours, while a cold startup takes about 3½ hours. If this time is extended due to water treatment problems, peak rate revenue opportunities may be jeopardized.

Table 10–18 shows some of the inherent causes and effects that can be caused by an inability to consistently achieve and maintain the target feedwater, drum water, and steam purity requirements.

During these transient conditions, drum water characteristics, especially pH, are difficult to control and stabilize.

As a result of these problems, passivating films can be removed; corrosion products can be displaced and transported; feedwater internal treatment chemicals can be overfed or underfed; and steam purity can be compromised.

Also, when extended high-rate blowdown and corresponding makeup rates are required to achieve and maintain the drum water and steam purity requirements, costs are incurred to provide demineralized water production and storage capacity. Energy costs can also become significant.

Erosion, corrosion, deposition, hot spots, and all other associated problems are well understood. The most immediate risk to owner-operators and contractors with startup responsibility is exceeding equipment manufacturer’s limits.

Longer term, these phenomena affect equipment serviceability and facility service life. Long-term effects can be monitored to provide “early warnings” with test heat exchangers and a comprehensive corrosion coupon program.

Condensate System

Feedwater Internal Chemical Treatment

Air-Cooled Condenser Heat Rejection