Advanced Models for Fatigue Life Estimation of Combustor Components for Gas Turbine Application
Abstract Modern day combustors operate at very high temperatures which are close to combustor material softening temperatures. At the same time, to meet stringent emission legislations there is a strong drive to improve upon the rich burn combustor technology or shift to advanced lean burn combustor technologies. One of the key driver to improve emission is to save the cooling air budget and use the saved air for primary combustion but this approach would require more advanced and efficient cooling techniques. Fan shaped effusion cooling technology is a very promising technique as it offers high film cooling effectiveness. However, complex cooling features associated with this technology can lead to higher stress concertation and localized triaxial stress state. This stressstrain field in combination with a typical gas turbine engine operating cycle makes such effusion holes highly vulnerable to the thermo-mechanical fatigue failure. Hence to ensure the safety and reliability of combustor liners with such innovative features, it is essential to have thorough understanding of the stress-strain field in the vicinity and accurate prediction of life to first crack. The biggest challenge the designers and engineers face while predicting the initiation life of a structure is selecting the appropriate fatigue damage model for an application. This is due to following reasons: (a) The scatter in fatigue life predicted using different models and experimental values are very huge (b) There is no general universal method which can predict the multiaxial fatigue life accurately for all the materials and loading conditions (c) No general consensus exits among the researchers on which model have to be used for a particular application, material, loading and geometry (d) Application level studies are seldom available on this subject, most of the studies are restricted to laboratory level specimens with very limited implications to industry. Ideally, the fatigue damage model which has to be used for a particular application has to be validated through experiments. To meet this objective, several test specimens featuring novel fan shaped hole geometries were mass-produced using state of the art laser drilling technology. All these specimens were subjected to strain controlled isothermal low cycle fatigue test and the cycles to crack initiation was monitored using potential drop method. Six different multiaxial fatigue damage models (which can be used in low cycle fatigue regime) viz. Walker model, Smith Watson and Topper model (SWT), Fatemi Socie model (FS), Wang and Brown model (WB), Shang model (SW) and Xu model were selected and the life estimated by these models were compared with the experimental values. From the study it is observed that Xu model in which the damage parameter is built using the concept of shear strain energy looks most promising for this application.