The application of granular computing in electric vehicle fault diagnosis

Abstract

In this paper, granular computing algorithm is studied and a granular computing application scheme for electric vehicle fault diagnosis is developed. It is shown that the granular computing has significant advantages on depicting the importance of different attributes for a knowledge representation system. A granular computing-back propagation algorithm can then be designed based on the information of the acceleration failure fault of a parallel-series hybrid electric vehicle including accelerator pedal position, vehicle speed, and so on, in the sense that not only the amount of fault data is effectively reduced, but also better accuracy and efficiency for the electric vehicle fault diagnosis can be obtained. Simulation results are presented to demonstrate the validity and effectiveness of the proposed application scheme.