Full-load, full-scale testing is key to reliability20 April 2001
Extensive testing and design validation have been crucial elements in Mitsubishi Heavy Industry’s approach to advanced gas turbine development and commercialisation.
The evolution of MHI gas turbines over the past two decades is shown in the diagram below. Through the use of new materials and advanced blade cooling concepts, the firing temperature of the F class was raised by some 200°C to around 1300-1400°C, resulting in a combined cycle efficiency of 55-57 per cent (LHV). In 1997, MHI introduced the G class, with a firing temperature of 1500°C. The G machines have steam cooling of the combustor baskets and achieve a combined cycle efficiency of 58 per cent (LHV). Subsequently, in the H class, steam cooling has been further extended to include other hot gas path components, with closed circuit steam cooling of combustor baskets, row 1 and 2 turbine vanes, blades, and turbine rings. The H, currently under verification testing, is aiming at a thermal efficiency of 60 per cent or more.
Contributors to reliability
The use of large advanced turbines in combined cycle applications in a deregulated environment is imposing new challenges on gas turbine technology, with more frequent startups/shutdowns, load variations and other factors driven by fuel costs. Widely publicised problems with the introduction of some technologies, and their financial implications, have pushed RAMD (reliability, availability, maintainability and durability) to the top of the agenda.
The following features are major contributors to reliability:
• Use of proven design structural features
The rotor configurations of the Mitsubishi D, F, G and H gas turbines are strikingly similar. Except for the spool type compressor in the M501D, the compressor and turbine sections in the F, G and H gas turbines even maintain exactly same bolt count - 12.
The basic structure of the H series gas turbine is modelled after the F and G, inheriting their proven technologies, as follows:
• Two-bearing support for the rotor, with the compressor and turbine portions joined by a centre coupling.
• Connection with the generator at the (cold) compressor end, where thermal effects, such as expansion, are small, and a flexible coupling is not required.
• Axial flow exhaust structure, which is optimum for the layout of a combined cycle plant.
• Compressor side bearing supported by 8 radial struts and turbine side bearing supported by tangential struts, able to easily absorb differential thermal expansion by keeping the shaft at the centre.
• For the connecting structure, discs with torque pins are connected with bolts on the compressor rotor side, while discs with curvic couplings are connected with bolts on the turbine rotor side, both able to deliver torque without fail.
• Maintaining same scaling procedures
The scaling procedure from 60 Hz designs to 50 Hz designs has been steadfastly maintained throughout the MHI product lines. For example, the first two rows of blades and vanes of the turbine are common across the 60 Hz and 50 Hz versions. Likewise, the combustors are common between the two frame sizes. The field experience and design calibration with this scaling approach is retained across MHI’s D, F, G class of gas turbines. By contrast others have deviated significantly in scaling method, rotor structure, materials and coatings.
• Use of proven materials
The same materials are used in the F, G, and H class machines. For example, MGA 1400 for turbine blade rows 1 to 3 and MGA 2400 for turbine vane rows 1 to 4. The commonality in materials simplifies transfer of field repair experience and practices. Also, the use of fewer standardised materials makes it possible to concentrate on long-term material property characterisation. By retaining materials and field proven structural design features, MHI’s design focus has been on improving stage efficiencies in the compressor and turbine through state-of-the-art 3D aero computational techniques. It has proved possible to achieve a 25:1 pressure ratio in the H, as compared with 20:1 in the G machine, while at the same time using two fewer stages. Thus, similar rotor spans are maintained for the G and H, optimising design operating experience and reliability.
• Extensive verification testing
It has been MHI’s policy to make new technologies commercially available only after extensive full scale full load testing. The need for testing is now more important than ever before.
The G test plan included extensive factory testing and extended field evaluation. After several development tests, including a 2D cascade test, a first stage scale model test was conducted on the MF-221 gas turbine, which has a half-scale compressor. A full load shop test was carried out with the MF-221. In addition a full-scale rig was built to test the front three stages of the compressor.
Trial operation of the 60 Hz version of the G, the 254 MWe M501G, started in February 1997 at the company’s in-house full-scale combined cycle power plant at Takasago (known as T-Point), with 1800 instrumentation probes measuring flows, temperatures, pressures, strains, sound, emissions, etc.
The facility usually undergoes daily start-stops and power is transmitted to a local utility. Thus reliability and availability have to be up to commercial contractual levels. Thorough inspections following the peak dispatch season enable continuous validation to be carried out, with subsequent modifications, as needed. For example, minor design adjustments made during early stages of machine verification and operation resulted in enhancements to cooling circuits, resulting in superior thermal distributions and improved material/coating performance and component durability prior to commercial introduction of the G machines, which was later in 1997.
Trial operation of the 50 Hz version of the G, the M701G, was started at Tohoku’s Higashi Niigata plant in October 1998, with commercial operation in July 1999. Prior to site installation, combustion tests were conducted on a full-scale combustor and fuel nozzle, while full load shop testing was done in 1997 to verify performance and reliability of the M701G.
Trial operation of the 60 Hz version of the H (the M501H) started at T-Point in February 1999.
Evolving towards the H
AS already mentioned the G series incorporated a number of advanced features compared with the F. To achieve a higher Mach number compressor the G uses multiple circular arc (MCA) airfoils and controlled diffusion airfoils (CDA) in the front stages of rotating blades, as well as in the front stages of stationary vanes, in the higher Mach number region.
The G combustor is based on that used in the F. However to achieve the same NOx level as the F, the flame temperature must be kept around 1500°C in the combustion zone by the circulation of cooling steam in the combustor wall. By eliminating the combustor cooling air from the steam cooled wall, all the combustion air is introduced into the primary combustion zone to maintain the flame temperature at the same level as the F series - even with higher turbine inlet gas temperatures.
In the G turbine section, effective cooling is needed to achieve the same durability at the 1500°C turbine inlet temperature as was obtained with previous generation machines. The G therefore incorporates: full-coverage film cooling (FCFC) of the blades and vanes; thermal barrier coating (TBC); and directionally solidified (DS) blade technologies.
The H machine builds on the G technology features. Compared with the G, the H has 20 per cent greater output with efficiency 2 percentage points higher – but keeping the same 1500°C turbine inlet temperature.
The main innovation of the H, as already mentioned, is that the row 1 and 2 blades and vanes are cooled with steam, instead of high pressure air from the compressor, as used in the previous generation designs. Turbine row 3 is air-cooled and row 4 non-cooled.
Thanks to the steam cooling, the amount of cooling air is reduced to about half that used in the previous designs and the mixing loss associated with the cooling air is also reduced, with consequent improvement in overall plant thermal efficiency. In addition, the combustion gas flow increases by an amount corresponding to the reduction in cooling air, and the efficiency of the turbine section is increased.
Also, for a given pressure ratio, the exhaust gas temperature increases due to the reduced amount of low temperature cooling air mixing with the combustion gas. The exhaust gas temperature is kept at the same level as the previous machines due to the increased pressure ratio. The H compressor achieves a pressure ratio of 1:25, which is higher than previous designs, but with a reduced number of stages, namely 15. This is achieved through the use of the most advanced blade and vane technology available. The compressor design has undergone thorough verification using a 0.29 scale model.
Development of the H series gas turbine was started in 1996, and trial operation began in February 1999. Such a short development time was supported by reliable verification through various component tests conducted in parallel. Any improvements required could be incorporated rapidly. Some examples are discussed below.
For the connections to and from the supply and recovery passages of the steam cooled blades and vanes, a special seal that had not been used for the existing gas turbines was adopted. As already noted the H series uses steam for cooling. However, if the same amount of leakage as for the previous cooling air was allowed, there would be significant reduction in total plant thermal efficiency, requiring increased makeup water. Therefore, the steam sealing is of critical importance. In order to address these issues in advance, the sealing ability of the selected seal technology was confirmed by means of various component tests.
First of all, a component test was conducted to check sealing ability under stationary conditions to select optimum structure and dimensions. Second, the components selected for the rotating system were tested using the model rotor, which could simulate actual rotating conditions.
Steam cooled blades and vanes
Several of the turbine row 1 vanes of the M701F gas turbine installed at MHI’s Yokohama Machinery Works (K Point) were converted to steam cooled vanes. These modified vanes were cooled with steam under actual operating conditions and data on cooling characteristics obtained.
Then, prior to trial operation of the H series turbine, a high temperature cascade test simulating the actual engine was conducted, with steam cooling of the row 1 blades.
As already mentioned, a 0.29 scale compressor was manufactured as part of the H series development programme. This compressor was driven by a two shaft gas turbine (M252) and equipped with a control valve for increasing pressure at the outlet. Tests were conducted simulating various operating conditions.
High pressure combustion test
A high pressure combustion test was conducted using the 0.29 scale compressor, which was equipped with a high pressure combustion test rig on the downstream side.
High temperature rotation test
A high temperature rotation verification test rig was built. This was equipped with a combustor to which high pressure air was supplied from the 0.29 scale compressor, along with natural gas. 1500°C combustion gas from the combustor was introduced to a 0.6-scale single stage turbine. The rotating load is controlled by absorbing it with a dynamometer to keep the rotating speed constant at 6000 rpm. In developing the H series turbine, six blades of a G series turbine were modified to accommodate the steam cooling configuration. A steam passage was introduced into the bladed rotor enabling steam to be supplied and recovered from the shaft end, as in the H series turbine. The simulated H gas turbine blades were operated under conditions where steam was supplied to the rotor from the shaft end for cooling and then recovered from the shaft end through the rotor. During operation, data for the cooling characteristics of the blades, temperatures and vibration levels at various points were obtained, and confirmed to be in line with design values.
Trial operation of the M501H gas turbine was conducted using the T-point in-house combined cycle plant at MHI’s Takasago works. Generally, this in-house verification plant has been used for verifying the long term reliability of the M501G gas turbine. This was temporarily replaced with an M501H machine for the H trials. High pressure steam generated at the exhaust heat recovery steam generator was introduced to the high pressure steam turbine, whose outlet, in turn, was introduced to the steam cooled blades and vanes for cooling. Following cooling, the steam was recovered to the inlet of the medium pressure steam turbine.
Trial operation was started on 8 February 1999, and 220 MW (GT output 160 MW, ST output 60MW) was achieved. On the basis of this trial operation, verification of the high pressure compressor characteristics and the steam system (cooling characteristics of steam cooled blades) was successfully completed.
From the measurements it was confirmed that cooling steam flowed through the turbine rotor, blades, blade rings and vanes, and it was shown that cooling effectiveness was as designed. Also, equipment condition after the various tests was good.
In the trial operation, special measurements were carried out at 1500 points. These showed the following:
• Confirmation of stable operation of steam cooled rotor
Steam was supplied from the rotor shaft end to the steam cooled blades through the inside of the rotor, and recovered to the rotor shaft end back through the inside of the rotor. Stable vibration characteristics were obtained in the trial operation. This trial was the first run for the H series gas turbine, so the operation was generally carried out manually, to allow timely corrective action to be taken if necessary. In the future, automated valves will be used.
• Confirmation of cooling characteristics of steam cooled blades and vanes
Results from trials show that expected performance is obtained. Output and turbine inlet temperature were up to 220 MW and 1200°C respectively in the trial. However, it was verified that the metal temperature was within the allowable limit by extrapolating the measurement data theoretically up to an output of 385 MW and turbine inlet temperature of 1500°C.
• Confirmation of performance of compressor and turbine
For the compressor, the data for start up characteristics and performance were obtained in advance from a performance test using the 0.29 scale compressor, as already noted. The results from the trial operation suggest characteristics very similar to those obtained from the performance test. Start up characteristics and compressor adiabatic efficiency were confirmed to be as designed. Turbine aerodynamic efficiency was confirmed to have achieved the design efficiency estimated by calculation, with correction for steam leakage. The plant total thermal efficiency was confirmed to be very close to the design value.
Second trial and verification
In March 2001 the second full-scale verification trial of the combined cycle M501H was successfully completed at the T-Point facility. This included full speed full load testing with closed loop steam cooling in the combined cycle system. Following a full-speed rotating test of the steam cooled rotor and pre-verification, a full load test was conducted with more than 2000 measurement points to evaluate the closed circuit steam cooling system, including the dynamic response of gas turbine and HRSG during changeover between air and steam, cooling effectiveness, component efficiency etc. Finally, steady state data were recorded at full load and transient data during load changes.
At the time of writing, extensive verification of the test results, data and hardware are proceeding, while the T-Point power plant is being scheduled for the switch over to the M501G gas turbine for summer peak load dispatch operation to Kansai Electric. On the M501G itself further reliability and durability verification continues on upgraded design features that will be introduced in future commercial applications.
TablesMain features of successive generations of MHI turbines