A SEMI-EMPIRICAL MODEL TO PREDICT AIRCRAFT SOOT EMISSION IN RICH ZONE OF RQL COMBUSTOR

The reduction of particulate matter emissions is becoming a requirement for aircraft turbine engine combustor design. This requirement leads to the need to estimate particulate emissions during the conceptual design phase. Current prediction methods are based on detailed numerical simulation techniques such as CFD, which are unsuitable for conceptual design due to high computational cost. This paper introduces a new approach employing a semi empirical model for prediction of the soot emission indices of RQL combustors. The proposed approach dramatically improves computational efficiency by avoiding complex numerical calculations. The model is based on the response surface developed from experimental data for turbulent non-premixed flames. The data has been extracted from the literature, employing statistical methods such as machine learning techniques and polynomial regressions to apply the turbulent flame data to the actual operating conditions in the primary zone of aircraft engine combustors. The model is developed by first representing the combustor primary zone by chemical reactor networks constructed in CHEMKIN based on a statistical PDF approach to simulate the non -uniform distribution of time -evolving local mixture fraction with a beta distribution. The reactor networks are used to estimate the concentrations of soot precursor species in the rich zone. The empirical equations are then used with the predicted concentrations to predict the soot foimation rate. Finally, these results are used along with the turbulent non premixed flame data to develop the final model through a model calibration process.