Chen, Yifan (Beijing Institute of Technology), Yang, Chao (Beijing institute of technology), Wang, Weida (Beijing institute of technology), Gao, Pu (Beijing Institute of Technology)

Energy Management for Hybrid Electric Vehicles Using an Intelligent Recognition Method of Engine Characteristics

Scheduled for presentation during the Regular session "Energy Management" (FrB2), Friday, November 1, 2024, 10:50−11:10, Room T2

7th IFAC Conference on Engine and Powertrain Control, Simulation and Modeling, Oct 30 - Nov 1, 2024, Dalian, China

This information is tentative and subject to change. Compiled on January 2, 2025

Abstract

Under extreme conditions such as rapid acceleration, the engine speed of hybrid electric vehicle (HEV) is forced to decrease, making it difficult for the vehicle to meet power demands. However, the inability to obtain the degree of engine speed reduction in different environments in advance makes the energy management strategy (EMS) of HEV less adaptive. In this paper, a simple outcome-oriented engine characteristic recognition model is presented. Firstly, the autoencoder is combined with the traditional clustering algorithm, and the characteristics of the engine are defined artificially. Secondly, the long short-term memory (LSTM) neural network is introduced to recognize engine characteristics online, and different thresholds of EMS under different characteristics are given. In the face of unknown operating conditions and unknown engine states, the recognition model can provide recognition results online and assist EMS in power distribution. Simulation results demonstrate that this strategy effectively controls the SOC trajectory. Compared with the benchmark methods, the fuel economy of the proposed strategy is improved by 6.95%, which is close to the global optimal strategy. The effectiveness of this strategy is validated.