Modelling Challenges for Lean Premixed Flames in Industrial Gas Turbine Combustors
13:30 - 14:00
Level 0 lecture hall between Al-Jazri and Al-Kindi (buildings 4 and 5)
The developments of lean premixed combustion systems for heavy-duty gas turbines has been significantly relying in the past on incremental improvements often identified via a trial and error exercise and expensive experimental tests at high pressure conditions. Testing at atmospheric conditions allows to reduce development costs but the transferability of the results to high pressure is sometimes questionable and, in order not to jeopardize the full process, must be carefully evaluated. Numerical simulations allow in principle to reduce development costs but, considering the complexity of the physico-chemical phenomena involved, expectations are sometimes too high. Turbulent premixed flames at high Reynolds and high Damköhler numbers as found in industrial systems are in fact characterized by scales with maximum to minimum size ratio still prohibitively large for Direct Numerical Simulations, even using state-of-the-art supercomputing technology. Large Eddy Simulation (LES) is becoming popular also in industrial developments and has surely introduced a progress over more traditional Reynolds Averaged modelling of the Navier-Stokes equations. At high Damköhler numbers however, heat is released over a length scale which is typically smaller than the LES resolved scales, leading to a highly wrinkled unresolved flame front structure which consumes reactants at a bulk rate whose sound and accurate modelling is still an open challenge. In addition to the traditional turbulence-chemistry interaction problem, where research has focused for many years, other complex phenomena like vortex breakdown in swirl flows, combustion dynamics and emissions play a fundamental role in the behavior of this kind of flames.
While accurate simulations of lean premixed combustion in heavy duty gas turbines remain a challenge, the use of modelling can be still very useful to understand the underlying physics and provide guidance to experimental testing and validation. The scope of the talk is therefore to highlight aspects of lean premixed combustion that are beyond the pure turbulence-chemistry interaction modelling problem. The first part is dedicated to the transferability of emissions and flame stabilization properties from atmospheric to high pressure conditions. The second part focuses on the important problem of combustion dynamics. These can be mitigated using passive measure like acoustic resonators that do not require detailed understanding on how the combustion process interacts with the acoustic field. On the other hand, a better understanding of the frequency response of the flame to the acoustic field (so called Flame Transfer Function, FTF) can also be of help during the burner and combustor design phase. This topic is discussed for the cases of swirl stabilized technically premixed flames with a special focus on the role played by acoustically driven velocity and equivalence ratio perturbations. Experimental and CFD results in case of industrial burners gas will be used in support of this type of analysis.