Calibration of 0-D combustion model applied to dual-fuel engine

Abstract

To acquire a highly accurate 0-D combustion model of the Wiebe function, a genetic algorithm (GA) based on algebraic analysis is developed. By analyzing the in-cylinder combustion process of biodiesel and diesel, a new method is proposed for determining the number of Wiebe functions and the transition angle between premixed and diffusive combustion in dual-fuel engine. First, for different operating conditions, the number of required Wiebe functions and the initial value of Wiebe parameters are determined by algebraic analysis method. Then, genetic algorithm is applied to obtain the further optimized Wiebe parameters and finally compared with Levenberg-Marquardt (LM) algorithm on fitting precision. The algorithm is applied to a dual-fuel engine under the conditions of propeller performance. The results show that, based on the initial value of Wiebe parameters, the fitting process of both LM algorithm and genetic algorithm converges rapidly, the R2 of xb are all at a high level (larger than 0.997), but the RMSE values of genetic algorithm are all in a low range (smaller than 0.013). The fitting effect of genetic algorithm is obviously better than that of LM algorithm. Therefore, genetic algorithm based on algebraic analysis is an incredibly precise algorithm for structuring a 0-D combustion model.