An analytical user-friendly methodology to transform compressor and turbine supplier characterization maps dedicated to 1D engine simulation: Modelling of turbocharger heat transfer and friction losses

Abstract

Available turbocharger characterization maps are usually obtained on a gas stand where the turbine intake temperature is high, resulting in heat transfer affecting both the compressor and turbine. A correction is then mandatory to take into account heat transfer and calculate the real amount of work produced or consumed.An experimental campaign is conducted to characterize heat transfer and friction losses on two different turbochargers. Then, semi-empirical correlations are developed minimizing the sum of square errors of a multiple regression to modelize these two phenomena. Compared to existing methodologies, these correlations minimize the requested input data: only the supplier maps are required whereas no additional test or calibration is needed.The methodology to transform efficiencies and model heat transfer is presented. A gain of 10 % is observed on the confidence interval +/−3 °C on the compressor outlet temperature calculation (passing from 42.2 to 53 %), beside the turbine outlet temperature +/−10 °C interval is improved by around 20 %. These improvements are observed on two engine 1D models using two different turbochargers. The second engine's turbocharger is not involved in the establishment of the correlations and then demonstrates the good predictability of the proposed methodology.