By Yoshie Usune, Masao Terazaki, Yasuoki Tomita, Mitsubishi Heavy Industries, and Jun-Hee Lee, Mitsubishi Power Systems Americas
Editor’s note: This article is excerpted from a paper given at POWER-GEN International in Orlando in December 2010 that was selected by the conference planning committee for recognition as one of the best papers of the show. The complete paper is available at the Power Engineering magazine web site www.power-eng.com.
Gas turbine combined cycle (GTCC) plants are designed and built with emphasis on high availability, reliability and performance. In 2009, the M501F gas turbine fleet (including auxiliary equipment, generator and exciter) attained an availability rating of 95 percent. This high availability is possible through the application of the following four activities:
- Maximizing the reliability of gas turbines through design verification
- Minimizing scheduled maintenance duration
- Providing quick and close operational support to clients
- Gas turbine upgrades
Forced outage rate can be decreased by continuously monitoring GT operation and reacting quickly to GT trip/runbacks. The analysis of disruptive events and the lessons learned can be used to improve the durability of hot parts, other gas turbine components and sensors. Abnormality diagnosis technique is also a great contributor for preventing extended forced outages.
Another important factor is reducing the duration of scheduled outages. In this regard, field experience is used to design parts that can sustain longer intervals, specifically 12,000 hours between inspections. Efficient field work schemes and improved tools are also developed to reduce outage duration.
Finally, effective countermeasures to recover performance degradation and combustion system upgrades targeting for reduced emissions with improved combustion dynamic margin can help increasing availability of the power plant. The remote monitoring center is a powerful tool to define appropriate compressor water wash intervals through compressor efficiency monitoring.
This article describes some of the processes and technologies which are applied to achieve these goals.
In the U.S. power generation market, combined cycle power plants have to achieve high reliability and availability to meet the dispatch requirements of electric grids at a short notice. Therefore, measuring the reliability, availability and maintainability (RAM) performance of the power plants fairly without any bias becomes important because the RAM data is a reflection of both product and service quality for original equipment manufacturer (OEM).
Since January 2009, Strategic Power Systems Inc. (SPS) has been collecting and processing the RAM data of Mitsubishi’s M501F Fleet and advanced M501G Fleet as an unbiased third-party. SPS specializes in measuring reliability and availability for the power plant owner/operators and the OEM using Operational Reliability Analysis Program (ORAP), based on industry measurement standards.
For Mitsubishi’s advanced M501G Fleet, 16 units are currently being monitored. An additional 56 unit-year’s worth of data collected by Mitsubishi since 2004 was audited by SPS in accordance with IEEE 762 and ISO 3977 standards. Table 1 shows the average statistics from 2004 to 2009. Table 1 also shows the SPS-reported RAM statistics of the M501F fleet during year 2009 compared with the SPS-monitored “F” Class fleet data including non-MHI.
The challenge is to improve the availability and reliability statistics further. The first thing to look at is gas turbine hot gas path parts (HGPP) that are exposed to the most extreme conditions and therefore most likely to experience fall-out. HGPPs can be noticeably enhanced through a feedback approach that incorporates fleet wide field experience gathered from actual GTs into the original part design process to give more robust part design.
Second, quality in field service work is an area that has a critical impact on the availability and reliability of gas turbines. The duration of scheduled outages can be reduced by effective project management such as pre-planning hard and soft assets, executing with knowledgeable personnel and so on. Accumulated technical knowledge via lessons-learned from previous outages is an important asset for the field service personnel.
Third, the use of remote monitoring systems has proven to be an effective tool for early detection of potential problems. Early detection requires software algorithms to isolate the signature of a symptom from other data. Mitsubishi has developed and implemented the Mahalanobis Taguchi method (MT method) for automatic early detection. It identified that minor changes in the blade path temperature (BPT) deviation was caused by incipient damage to the combustor baskets and subsequent bore scope inspections prevented worsening of this damage. In addition to the MT method, trip analysis was conducted to reduce the number of trips by offering prioritized countermeasures to customers according to the statistics.
Fourth, gas turbine upgrades provide valuable solutions for power generation users to counter the negative effects of thermal performance degradation. In this respect, Mitsubishi has been focusing on the development of gas turbine hot part component improvements and system upgrades. For example, Mitsubishi introduced row 1 blade improvement, advanced DLN combustors and an advanced combustion pressure fluctuation monitoring system (A-CPFM).
In this article, a four-pronged approach for achieving higher availability and reliability is outlined: (1) robust design utilizing the field data gathered during scheduled outages, (2) efficient scheduled outage management with emphasis on quality, (3) proactive intervention with remote monitoring technology and (4) improved design and upgrades for longer parts life.
From Field Data to Design
This section explains the design process of HGPPs and how input from the field experiences improves GT reliability. Increasing the turbine inlet gas temperature (TIGT) is the most effective way of increasing the GTCC efficiency together with a higher compression ratio. Because of increased TIGT, HGPPs are exposed to severe conditions.
To provide upgraded parts, various steps are required from design to manufacturing and verification. It is essential to ensure performance and reliability by long-term verification after laboratory tests have been conducted. In-house verification of components in a full sized GT is essential part of the product development cycle for Mitsubishi. Knowledge gained from actual field experience is fed back into the design stage. This distribution of information leads to constant improvements in GT performance and reliability.
The process of improving HGPP design consists of analysis of thermal barrier coating (TBC) degradation, careful application of life assessment techniques on substrate metallurgy and analysis of operational data.
Design & Development
Design specifications to meet a GT performance requirement are first determined as allocation of cooling air and then detailed HGPP design is begun. It is important to apply optimized cooling scheme, materials and TBC together with state-of-the-art aerodynamics, heat transfer and mechanical design technologies. Computational fluid dynamics is used to determine the flow field, and finite element methods are used to determine the details of GT component design.
Furthermore, operational data and individual technology verification results play an important role in HGPP reliability design. A number of dedicated test facilities at Takasago Machinery Works are utilized such as a high-pressure combustion test rig, a scale model turbine section to study the effects of turbine blade and vane cooling mechanisms and a scale model compressor to test modified compressor blade and diaphragm designs.
The original combustor basket design for F class engines has a plate fin cooling wall structure. Field experience revealed that partial oxidation at the plate fin ring edge was identified as the main damage mode especially for the duel fuel units because of localized flame hold at the steps between the rings. An upgraded structure called MT-Fin (Mitsubishi Takasago Fin) with an acoustic liner was developed for cooling the basket wall. This resulted in improved durability of the baskets because the inside wall of the new combustor basket has a smooth surface and the cooling effectiveness is enhanced.
Combustion dynamics are the most difficult issue with DLN (dry low NOX) combustors. This can destroy combustion hardware if large peak amplitudes are allowed to continue for long periods of time. The benefit of the acoustic liner is to dissipate acoustic energy generated by combustion pressure fluctuations to protect the hardware. This technology has been applied to 43 units so far with the lead unit accumulating 13,819 actual operating hours and no severe combustion dynamics have been observed since its application.
Transition pieces require improved cooling and thermal barrier coatings because combustion gas from the combustor basket exit is directed toward the first stage of the stationary airfoils via transition pieces. Originally, masking tape was applied to prevent TBC entering cooling holes as) and consequently the area around the cooling holes was exposed to high temperatures. The steep thermal gradient due to lack of TBC around the cooling holes caused low cycle fatigue (LCF) after many GT start-up and shut-down cycles. TBC coating process with no-masking was developed and verified. TBC coating is applied around all cooling holes so that thermal stress is reduced and LCF life is improved. Consequently, the number of parts with damage indications was significantly reduced, while the fall-out rate became almost zero.
The application of integrally cast multiple vanes at the forward turbine stages is known to cause cracks at the airfoil root that are driven primarily by its thermal expansion differentials. An alternative design utilizes field experience to overcome the above issue. Turbine row 2 vane utilizes a mechanically bolted configuration and turbine row 1 vane is installed a single vane configuration. The design intent is to allow for the thermal expansion of the shroud for a segmented vane. An added benefit of this row 2 vane bolted configuration is better TBC process control because the vane surface is more accessible for robotically deposited TBC coating.
Original F class turbine row 3 vanes made of cobalt based alloy X-45 showed axial creep deformation toward the downstream-side after long-term operation. The material was upgraded to nickel based alloy MGA2400CC, which is Mitsubishi patented super alloy with higher creep strength. A deformation measurement tool was also developed. This improvement resulted in a reduction in axial creep deformation by 70 percent.
Minimizing Scheduled Outage Duration
In general, power plant availability can be increased not only by preventing unexpected interruption of power generation, but by shortening scheduled outage duration or extending a scheduled maintenance interval. On the other hand, extension of scheduled outage intervals requires robust design and field-proven upgraded parts that should sustain a longer maintenance interval with repairable damage modes.
The scheduled maintenance duration can be reduced by detailed planning of outage works, necessary hardware and manpower. Accumulated technical knowledge from lessons-learnt meetings is an important asset for the field service personnel and helps to prevent the unnecessary extension of outage works.
Quality and process control is one of Mitsubishi’s continuous improvement goals. Ongoing quality programs at the Field Service include technical documentation, field quality audits, clearly defined goals and expectations, and training. To enhance the quality of Field Service, a regular technical review of how to improve the specific tasks and a field service lessons learned program were introduced.
One of the “Lessons Learned Notices” in May 2009 recorded five sensors detecting signals for the combustion pressure fluctuation monitoring (CPFM) and identified as having “noise” issues prior to the start-up of the unit. Due to a high probability of having issues during the tuning of the unit, Field Service notified the customer that the CPFM sensors would need to be changed out in order to tune the unit successfully.
As corrective action, the unit was shut down and cooled for several hours, allowing the exchange of the relevant CPFM sensors. As a result, Mitsubishi implemented a human performance program, which includes a mechanism for communication among customer, project manager and site workers. It is included in the project manager’s turnover document to ensure that all necessary corrective actions are implemented prior to startup. Also prior to startup, a meeting will be conducted between the customer and Mitsubishi personnel to ensure all open items are closed.
Outage duration has to be managed by effective project management covering pre-planning of hard and soft assets, executing with knowledgeable personnel and controlling work with quality documentation. During the pre-planning stage, work procedures specific to the type of GT are prepared and all inspection and quality records are reviewed. Pre-outage meetings are held with the customers for discussing outage scope and any concerns that the customer has. Evaluation of parts list and tooling inventory is made to prevent any delay due to lack of any GT parts, tools, or consumable parts. The outage schedule is developed showing daily activities and key milestones in a visual format.
During an outage, the project manager will meet with the customer daily to discuss outage status, a review of critical path and planned activities. When the outage scope is extensive covering gas turbine, steam turbine, generator and HRSG (heat recovery steam generator), a daily meeting with the customer becomes important to review the critical paths.
Communication between the outage site and the Gas Turbine Engineering Service Department was vastly improved by using an internet system. Any deviated conditions of service-run parts from the drawing requirements are reported to Engineering personnel via e-mail. Disposition requests are made and detailed inspection findings are shared in the photos and report via internet. Disposition by the Engineering Department will be made in less than 24 hours.
Gas Turbine Upgrade
While addressing component durability issues, an effort was made to extend inspection intervals and life since this also improves unit availability. Extending inspection and repair intervals by implementing durability enhancements means that longer operating intervals will not sacrifice the repair ratio of HGPPs. Mitsubishi has evaluated extensive field data and developed combustor and turbine front stage modifications to allow for extended intervals.
The extension of scheduled outage intervals requires robust design and field-proven upgraded parts that should sustain a longer maintenance interval with reparable damage modes. For example, a service interval of 8,000 EOH (equivalent operating hours) requires annual scheduled maintenance, but the extended 12,000 EOH interval requires a maintenance once every 1.5 years. Any increase in the fall-out rate of the parts or any severe damage to the parts due to extended service interval will not be acceptable to the owner or the operator of the power plant because of the much higher costs associated with new replacement parts or extensive repairs.
Turbine Row 1 Blade Upgrade
The rotating blade platform is the most difficult area for design engineers due to the end-wall flow. Initially, cracks were observed on seal pins or cooling air holes. The key to reducing the low cycle fatigue (LCF) distress is the uniform cooling to minimize the thermal gradient between the airfoil and platform-edge. The platform cooling was enhanced to address the issue and at the same time the cooling air balance for the design modifications was maintained. This helped optimizing the back-flow margin at the airfoil surface as well as having uniform metal temperature distribution. The latest modifications were implemented and successfully operated for 23,203 actual operating hours without any indications of damage.
A 15 ppm dry low NOX combustor technology for G-class gas turbines, called Mk8-4, was developed and tested at T-point. Several key design features including improved inlet aerodynamics, fuel/air mixing, combustor aerodynamics and acoustic dampers are used in the Mk8-4 in order to reduce NOX and improve performance.
Validation of the Mk8-4 combustor was performed in a M501G gas turbine at T-point verification plant. More than 1,500 special measurements were collected during rigorous tests which were conducted for verification of the new design. The test results confirmed the following features:
- 1. Less than 15ppm NOX and 10ppm CO emission with turn down to 60 percent load
- 2. Stable combustor dynamics at full load range
- 3. High start-up reliability (combustor ignition)
- 4. Suitable for daily start and stop (DSS) operation, and
- 5. Confirmed reliability and durability of hot parts.
Long term verification of the new combustion system was also conducted at T-Point under grid demand conditions for several months.
Advanced Combustor Pressure Fluctuation Monitoring (A-CPFM)
Combustor instabilities due to sudden changes in fuel calorific value or fuel quality, large ambient temperature swings, or sudden changes in operating load conditions can trigger high-pressure fluctuations. If the resulting combustion dynamics are not suppressed in a relatively short time, they can cause considerable damage to the gas turbine hardware.
In order to reduce runbacks or trips initiated by the protection system due to combustion instabilities and to minimize the frequency of combustion tuning to make adjustments to compensate for seasonal or fuel composition changes, Mitsubishi developed a self-tuning system called advanced CPFM (A-CPFM). This on-line monitoring and protection system automatically adjusts the air bypass valve, main and pilot fuel flows to maintain an appropriate fuel/air ratio depending on the combustion chamber flame stability.
A-CPFM system can be retrofitted to existing units that were commissioned before the advanced modules were commercially available. A combined cycle power plant in Texas experienced combustion dynamic instabilities due to fuel gas composition changes. After retrofitting the ACPFM system, the units have successfully avoided load runbacks and trips during transient fuel gas changes that would have previously disrupted plant operation.
A four-pronged approach for achieving higher availability and reliability is outlined: (1) robust design utilizing the field data gathered during the scheduled outages, (2) efficient scheduled outage management with lessons learnt, (3) proactive intervention with remote monitoring technology and (4) improved design and upgrade for extended parts life.
Properly scheduled outage works also reduce the scheduled maintenance duration by detailed pre-planning of hardware and skilled man-power. Accumulated technical know-how of the field service crews via lessons-learned meetings played an important role in reducing the down-time and enhancing the quality of the outage works. That prevented the unnecessary extension of outage works.
Remote monitoring improved the speed and the accessibility of quality operational data for engineering review, data trending, and deviation analysis. It made early detection of potential problems possible by using sophisticated software algorithms to isolate the signatures of certain problems from the other data.
Finally, effective countermeasures to recover performance degradation and combustion system upgrades targeting reduced emissions and improved combustion dynamic margins helped increase availability ratings.
1. S.A. Della Villa and C. Koeneke, “Monitoring and Measuring the Availability and Reliability Performance (RAM) of one OEM’s Advanced Gas Turbine Class Fleet: Computation Method Based on Continuous Data Collection,” PowerGen Europe, Amsterdam, Netherlands, June 2010.
2. S.A. Della Villa and C. Koeneke, “A Historical and Current Perspective of the Availability and Reliability Performance of Heavy Duty Gas Turbines: Benchmarking and Expectations,” GT2010-23182, ASME Turbo Expo, Glasgow, UK, June 2010.
3. Y. Usune, M. Onda, Y. Tomita, and B. Jayaraj, “OEM’s Approach to Maintain High Reliability of Hot Gas Path Parts,” GT2010-23241, ASME Turbo Expo, Glasgow, UK, June 2010.
4. N. Mikami, S. Kumano, I. Moriwaki, and E. Funabiki, “Remote Monitoring for Power Generation Equipment,” ACGT 2009-TS49, Proceedings of the Asian Congress on Gas Turbines, Tokyo, Japan, August 2009.
5. Y. Tomita, D. Hamajima, and N. Champ, “Gas Turbine Performance Upgrade Technology and Verification,” ACGT 2009-TS60, Proceedings of the Asian Congress on Gas Turbines, Tokyo, Japan, August 2009.
6. H. Arimura, T. Komori, C. Koeneke, T. Ai, and T. Kishine, “Evolution on Large Frame Industrial Gas Turbine for Environmentally Friendly Power Generation,” PowerGen Asia, Kuala Lumpur, Malaysia, October 2008.
7. C. Koeneke, H. Haruta, and T. Kawakami, “Update on MHI Self-Tuning Combustion Dynamics System ACPFM,” PowerGen International, Orlando, Florida, December, 2006.
Power Engineerng Issue Archives
View Power Generation Articles on PennEnergy.com