Modeling of 0-D heat release rate of heavy-duty diesel engine based on classification algorithm

Abstract

With the shortage of fossil fuel and the strengthening of emission regulations, the control strategy using in-cylinder pressure signals has caused extensive concern. Modeling the combustion heat release rate is an indispensable work along with the in development of the diesel combustion control hardware-in-loop system. Based on the analysis of experimental data, four combustion modes are proposed in this paper. The Wiebe-based functions developed for each combustion mode are given. According to experimental launch data, the combustion mode is classified automatically using the multi-classification support vector machine algorithm. The LM-Chaos algorithm based on function parameter identification is proposed. The results show that HRR shape is mainly related to the rate of premixed combustion and diffusion combustion under different operating conditions, and the variation of HRR directly affects the structure of Wiebe-based function. Different combustion modes can be simulated by selecting the starting time of combustion, the duration of combustion phases and the co-existence time of combustion phases. This method can simulate the heat release curves of various working conditions and improve the fitting of heat release curves of long combustion time by adding error function.