High ratings for gas turbine reliability, availability and maintainability may be ascribed to four key factors, Carlos Koeneke of MHPS tells Tildy Bayar
Efficiency is undoubtedly the Holy Grail for gas turbine OEMs – but reliability, availability and maintainability (RAM) are also crucial aims. Validation of operational turbines is necessary to back up OEMs’ claims.
Carlos Koeneke, vice-president, project engineering at Mitsubishi Hitachi Power Systems Americas (MHPSA), a subsidiary of Mitsubishi Hitachi Power Systems (MHPS), says his firm’s fleet has consistently received top marks from independent third-party validation company Strategic Power Systems, which is used by several major gas turbine OEMs and collects operational data directly from their customers around the world under the auspices of its Operational Reliability Analysis Programme.
Koeneke says MHPS’s high RAM ratings for its advanced class gas turbines are due to several key design and manufacturing decisions as well as a unique method of validation testing.
MHPS’s steam cooled M501G machine features a turbine inlet temperature (TiT) of 1500°C. Koeneke has presented RAM data for 2012–2017 based on 24 M501G turbines in MHPS’s fleet. These data, which have been verified by Strategic Power Systems, show a
99.26 per cent RAM factor for the M501Gs.
Although he offers several factors that he believes explain MHPS’s high RAM ratings, Koeneke cautions that it’s not enough to have a well-designed and correctly installed machine. It’s also necessary to have good operating practices on the part of the turbine owner as well as good maintenance routines.
From an OEM point of view, though, he believes that there are four main factors that have contributed to MHPS’s RAM success.
Using proven materials
MHPS has traditionally targeted a high Turbine Inlet Temperature (TiT) as the most effective way to increase the efficiency of advanced class gas turbines and the corresponding combined cycles. The company says its latest model, the air cooled M501JAC, has demonstrated a TiT
In explaining MHPS’s high RAM numbers for its M501G fleet, Koeneke points first to the use of a Turbine Rotor Cooling Air (TCA) Cooler approach, which effectively reduces the temperature of the air used to cool the gas turbine’s rotating hot gas path components. He says the resulting reduction in cooling air temperature allows for operation at high TiT and high efficiency while using low alloy steel discs and eliminating the requirement for single crystal hot gas path components.
Avoiding the use of sophisticated alloys that are resistant to creep but are also susceptible to failures when exposed to high temperatures and the stresses typically encountered in advanced class gas turbines, he says, is a fundamental premise of MHPS’s design approach.
“We use several external coolers,” he says. “We started with a concept from our Westinghouse partnership time in the late 1960s, all the way up to today.
“You have two choices” for cooling, he explains. “One is to cool the cooling air so that the parts exposed to very high temperatures are cooled with cooled air. Or you start using sophisticated material that can withstand the high stress and high temperature operating environment involved in gas turbine parts, especially the turbine discs.
“If the hot parts are exposed to high temperatures, you have the risk of failures associated with creep. One way to avoid this is by either cooling the air, which is what we do, while designing the plant to recover the heat removed from the cooling air in the bottoming cycle. The other option is the use of sophisticated and more expensive materials. If you go towards sophisticated metals such as high nickel content discs, they do have better creep resistance but are weak in terms of surface imperfections that can lead to cracking. So there is an advantage in terms of sustaining higher temperatures, but there is a disadvantage which can involve disc cracking failures.”
Under one roof
Next, Koeneke describes MHPS’s ‘under one roof’ approach for refurbishing and manufacturing hot gas path components. With this approach, “we can make sure all the parts are manufactured under our quality control system and time delivery requirements,” he says.
“Using sub-suppliers can be more cost-effective, but you end up relying on their quality control. We know of many casting failures that have resulted in significant damage with enormous insurance claims. We have rarely experienced casting issues.”
He notes that purchasing castings from different suppliers, engaging other companies to handle machining and drilling internal cooling holes, and then having other suppliers apply the thermal barrier coatings may achieve certain cost reductions, but there is a downside. “If someone is trying to do the job less expensively, there might be some cut corners, and in some cases the cheaper parts fail”.
“We have the castings done in well-verified Japanese casting companies we have used for years, and then we execute the entire machining process inside our Parts Manufacturing buildings in either Japan or Orlando, Florida. There we do absolutely everything to complete that part, and the part goes out of the door finished.
“Finishing on time and following strict in-house quality criteria can be the difference between being late in your outage or on time, and it directly affects reliability and availability.”
Conservative design approach
According to Koeneke, a conservative design approach is also key to RAM success.
While he notes that this is “somewhat contradictory because increasing turbine inlet temperature involves risks that make this effort the most demanding challenge in gas turbine design”, he says MHPS’s “secret sauce, in our opinion, is to do it gradually”.
“We’ve been doing this for decades, mainly with the incentive of generation in Japan and other Asian countries where fuel is considerably more expensive than in the US. We went from F-class to steam cooled G, then applied conversion to air cooled G, called GAC. Years later, we introduced the steam cooled J and, once again, applied the same successful conversion applied to the G to produce an air cooled J, called JAC.
“We have always avoided focusing solely on flexibility, low loads, emissions and other good flexibility features while neglecting efficiency. This would have led us to the wrong assumption that efficiency is not important in countries with low fuel cost, like the US.
“Despite the low fuel cost, clients here in the US want efficiency. We continued our progressive temperature increase trajectory instead of trying to jack it up at a fast rate with the associated risks I mentioned before.”
‘Our validation is a power plant’
MHPS commissioned its ‘T-point’ validation plant on the site of its Takasago Works in Takasago City, Hyogo Prefecture in 1997. Performance and durability testing of the M501G series gas turbines began at that time.
“For decades in the 1960s and ‘70s,” Koeneke says, “gas turbines were tested at the first client plants. We were doing lab tests for individual components, but no testing of the final assembled product. This product then went to the first plant and obviously details came up at that client’s site, causing headaches, and it took a long time to identify and correct the issues.
“In the ‘80s, companies started to do shop tests. You have a setup in your factory where you install the very first unit, which is a big effort because these units weigh several hundred tonnes. You put it on a stand, connect a lot of sensors/cables/pipes and many other devices needed for operation, then run it for a few hours. You also have to dissipate the energy, which is another considerable effort. We were using hydraulic brakes; other approaches for energy dissipation can be applied.
“The cost associated with large turbines’ fuel consumption is very high. If you just do a shop test you’re not getting any financial benefit other than testing the machine with considerable fuel expenses. Even in the US, to run a large frame machine will burn fuel on the order of $100 million per year – so obviously you cannot sustain this test operation for a long time.
“We wanted to run the turbine long-term because there are many failure mechanisms that don’t occur until a machine has run for a considerable number of hours and starts. You never hear a doctor asking a parent to check the cholesterol of a newborn baby – you’re not going to have high cholesterol until you are older. The same thing happens with gas turbines: low cycle fatigue and other types of failures occur once the machine has accumulated a lot of starts and stops, or has been operated under certain demand conditions.”
MHPS’s solution was to build a power plant inside its facility and then contract with Kansai Electric, the local utility, to sell its power.
“We spent several hundred million dollars building a power station,” Koeneke says, “not with the objective of making money – but having a contract with a local utility allows us to recover all the expenses associated with the investment, fuel, O&M etc.
“We bring in a new frame or or a modified turbine, install it there, and the utility tells us when to start/stop/ramp up/ramp down.
“During periods of low demand we are allowed, depending on dispatch, to look inside the turbine with a borescope, and sometimes we’re allowed to open it. If a part is found to be in bad shape, the designers develop modifications and install the modified part. By the time we sell a similar machine, the tests at T-Point continue while the construction of the new plant takes place. We end up with at least two to three years of operation with the changes that we’ve introduced. This has been found to be great for detecting and correcting problems, and more importantly for testing the countermeasures.”
Because “our validation is a power plant that follows our design criteria, and we’re connecting the gas turbine to a standard generator installed on standard pedestals (see Figure 1), if something is going to go wrong, there is a high probably we will detect it in our validation facility,” Koeneke says.
A unique approach
The reason only MHPS has built such validation facilities involves “a very simple explanation” according to Koeneke.
“Think about spending several hundred million dollars for a company that has strict ROI requirements,” he says. “Under such requirements, it is easy to conclude you are better off spending a fraction of that investment on validation and trying to detect as many potential problems as possible.”
Japanese companies are different, he says. The shareholders are more broad-minded, and the firms’ motivation is also different. For a Japanese company, the reputation of its products is fundamentally important. “There’s an enormous pride associated with this, a way of doing business,” Koeneke says.
The T-point validation plant has enabled MHPS to detect many issues due to its long-term running time, he adds.
“Let’s say we only perform a couple of hundred hours of validation at a test rig facility, and sell the subsequent unit soon after the shop test is finished. The construction of the subsequent project will take two to three years. In the meantime, there will be no further shop test validation done and the new turbine will not be further validated or improved. On the contrary, our T-Point unit continues running while construction of the new plants is ongoing, allowing us to detect, correct and test the corrections through long-term validation.
“For example, one issue we detected in our J class gas turbine involved spallation of the row 1 blades platform. This could have caused catastrophic failures, but we noticed it early. The design team changed the internal cooling scheme – modifying the source of cooling air from a relatively hotter source within the blade to a cooler one – and then we ran the turbine long enough to be convinced that the parts were able to sustain the different environment.
“This approach also helps us define realistic inspection intervals that are based on the condition of the parts after years of validation. We progressively increase the intervals rather than offer long intervals based on engineering assumptions.”
The gas turbines operating in the validation plant run under real-life operating conditions, he notes.
“Kansai Electric decides when we start and shut down, when we run at maximum or minimum load. These are real-world conditions and we have to operate at 3600 rpm because that’s the grid frequency. It’s true that we can’t do load tests at different rpm, but when a client buys a power plant it will be operating at
3600 rpm for 30 years, not under random rpm conditions.
“Small changes in grid frequency can happen during the life of the plant and can induce deviation from the 3600 rpm, but it is difficult to simulate the unit’s response if the test facility is not using a standard generator – not only because of the electric nature of grid frequency disturbances, but because of the generator’s large torsional inertia.”
Carlos Koeneke is a speaker at Electrify Europe. To register, click here.