State-of-the-art modelling tools for research and development of low-emission gas turbines
Gas turbines need to provide many years of safe and reliable operation while also offering flexibility in their usage. This is true of their output range and cyclical operation and of their fuel composition. In future, hydrogen will play a role in achieving low-emission, carbon-free power generation systems.
Prof. Dr. Kilian Oberleithner from the Institute of Fluid Mechanics and Technical Acoustics (ISTA) at the Technical University of Berlin presented a keynote speech at the FVV Transfer + Networking Event | Spring 2024 outlining a new research approach that will enable the prediction of thermoacoustic instabilities and the flame transfer function. This is important for developing low-emission gas turbines. We visited him after the conference in Berlin.
The key challenges in the development of gas turbines are thermoacoustic instabilities and the prediction of flame transfer functions for swirl and jet flames. Funded by the German Research Foundation (DFG), project 1421 »Dynamic of Swirl and Jet Flames II« has taken a big step towards achieving a comprehensive understanding of thermoacoustic instabilities.
As coordinator of the research project, Siemens Energy worked with other leading gas turbine manufacturers as industry partners and was able to contribute its experience from everyday engineering practice: »We have hundreds of gas turbines in the field and a great deal of practical experience, so we formulated a specific problem and assisted in designing the project and hardware requirements,« reports Dr Lukasz Panek, head of the project user committee. »It is a very complex topic, even for experts. There was no way we could do this fundamental research alone«. It requires a lot of time and personnel, he continues, and the cost is too much for an industrial enterprise that needs to sell products. By contrast, university researchers can conduct in-depth research and understand the fundamental principles. According to Dr Panek, this is necessary for further research and development.
In order to make this complex subject understandable, Professor Kilian Oberleithner offers a tour of the Institute of Fluid Dynamics and Technical Acoustics in Berlin, where he is the scientific head of the research project. Walking through the cellar corridors, past a rotary dial telephone mounted on the wall, we visit various test benches that are currently undergoing conversion. »Depending on the project, we do around two measurement runs a year, in which the test bench operates for a few days. Most of the work is done at our desks, because afterwards we evaluate the data we have gathered, and that can take months,« explains Oberleithner.
Sound influences combustion
The noise level in the small test bench rooms during the tests is extremely high, and this helps to explain what thermal instabilities are: A flame generates noise in the combustion chamber, which is reflected by the combustion chamber walls – and this in turn impedes the flame and causes combustion instabilities. While certain frequencies generate almost no reaction, the flame is heavily influenced by low frequencies in particular.
At the test bench, the researchers are investigating how the flames react at different frequencies and operating points. »The interaction between the flame and the acoustic field in the combustion chamber can cause major pressure pulsations that are very hard to predict,« explains Oberleithner. In addition, flame instability creates areas in which combustion is not completed, which generates higher emissions. The interaction between the flame and the sound waves can be described with a flame transfer function (FTF) and integrated into an acoustic model of the combustion chamber. In turn, this provides a model that makes the stability of the entire gas turbine predictable.
Predicting the flame transfer function
One of the goals of the FVV research project was to be able to make predictions in future at an early development phase as to whether or not a flame is unstable at certain operating points. Until now, FTF has been measured in experiments or calculated through numerical simulation. Now, there is a new approach: Models are being developed that are intended to predict the flame transfer function; they are based on physical conservation equations (momentum, mass, energy), which are linearised around statistical average values. These average values are taken from simulations or experiments such as the predecessor project 1358 »Dynamic of Swirl and Jet Flames I«.
A new development is the multiphysical approach, which combines a range of data from acoustics, thermodynamics and combustion in one holistic model, which can be used to calculate the FTF. The approach is based on linear equations that allow cause and effect to be reversed – this inverse design offers significant potential, as the model shows which details of the combustion chamber design are causing the instability. After all, if engineers can only identify the operating points at which the flame becomes unstable during the testing conducted after a gas turbine is built, this will result in extremely expensive iterations. The flame transfer function needs to be determined for every new burner, before the data is transferred into a model that reproduces all other components of a gas turbine – a huge task.
Prior to this, engineers can modify the geometry of the combustion chamber and adjust the fuel injection; even the thermal conditions can be altered by cooling or heating differently. »It is very difficult to optimise such complex systems. Turbulent flows are very heavily dependent on the starting conditions and surrounding conditions; if engineers make one small change somewhere, it can have a big impact on the whole system,« says Kilian Oberleithner, explaining the challenges. If more hydrogen is to undergo combustion in future, this makes things even more complicated.
Hydrogen changes everything
Efforts to reduce CO2 emissions are making research and development more dynamic. The issue of hydrogen as a replacement for natural gas is also driving innovation, even if gas is considered a sustainable source of energy when produced with green electricity. The problem is as follows: »If we add just 20 per cent by volume of hydrogen to the natural gas, CO2 emissions are not reduced by 20 per cent. You have to burn very high levels of added hydrogen to reduce emissions efficiently,« explains Oberleithner.
However, hydrogen flames behave in a completely different way; the gas burns much faster than methane, the flame speed is higher, and the flame burns at a different point – all of which produces an entirely different flame transfer function. »None of the previous predictions work reliably with hydrogen. If we burn hydrogen in the turbine instead of natural gas, it might start to vibrate, which wasn’t the case before,« says Kilian Oberleithner. Because hydrogen flames are pressure-dependent, he added, it is also extremely difficult to measure the FTF reliably.
And there is another problem: the high reactivity carries a risk of a dangerous flashback. To guarantee safe hydrogen combustion, the researchers must understand the flashback and develop effective countermeasures where necessary. This is to be done in the planned follow-up project, which looks at » Hydrogen Flashback (flasHH, T1723)«.
Oberleithner now leads us into the low-noise room, where work is being done on a burner in the middle of the room. When the burner is in operation, several microphones are arranged around it in a semi-circle to measure the combustion noise. The topic of thermoacoustics is being studied at ISTA in Berlin and in the Thermo-Fluid Dynamics Group at the Technical University of Munich. There is close collaboration with the team led by Professor Wolfgang Polifke at the Technical University of Munich: »Our team in Berlin comes from more of a fluid dynamics background, while Professor Polifke’s group focuses more on thermoacoustics, so we complemented each other perfectly in the project,« explains Oberleithner. While the team in Berlin primarily researched jet flames in the previous project, their colleagues in Munich focused on swirl-stabilised flames.
The project helped us to expand our prediction models and improve the thermoacoustic models. That is worth its weight in gold.
The goal of any FVV project is to transfer the research results into practice. With Siemens Energy, MTU Aero Engines, Rolls-Royce and GE, there were industry partners on board who will benefit from the results. »The project helped us to expand our prediction models and improve the thermoacoustic models. That is worth its weight in gold,« says a delighted Dr Panek. Another project has already been initiated together with Professor Oberleithner – the strong partnership is set to continue.
Nevertheless, Lukasz Panek explains why not every tool developed at a university makes it into industry for the long term: »Tools like this have a ›life‹. They need to be supported and developed, and that is very underestimated. It takes personnel, who are often no longer available when the project ends«. However, he adds that it is a good idea to adopt certain features in existing commercial software, since it is sometimes easier for industry to use commercial software that is continuously supported. //
