电动汽车电机故障时间的粒子群优化灰色预测

电动汽车电机故障因素多,可靠性分析需要大样本数据,为准确预测电机的故障时间,建立了故障率较高元件的故障树模型,给出了其可靠性计算式,并将基于小样本数据的灰色算法引入到电机可靠性计算中,利用传统和改进灰色模型进行仿真分析。为了进一步提高预测精度,以两种灰色模型为基础,利用粒子群算法的全局寻优能力,提出了以均方差最小为目标函数的优化模型,对电机故障时间进行预测,并利用两组实测数据进行了验证。结果表明,优化算法的相对平均误差分别为3.36%和5.05%,相对误差最大值分别为5.62%和8.41%。该结果验证了所提算法的有效性,为电动汽车电机的故障预测提供了理论依据。

Grey Prediction Model of Electric Vehicle Motor Based on Particle Swarm Optimization

Due to various motor faults in electric vehicles,a large amount of data are usually required in reliability analysis.In order to accurately predict the malfunction time,we established a fault tree model of components with high failure rate,and proposed an analytic formula.Grey algorithm based on small sample data was introduced into the reliability calculation of motor,and the traditional model and improved model were simulated and analyzed.To further improve the prediction accuracy,particle swarm optimization(PSO) which has the abilities of seeking the global optimum was utilized to fit the two grey models aiming at least mean squared errors,and to predict the malfunction time.At last,the optimization modelwas validated by two sets of measured data.The analysis results reveal that,the average relative errors of optimization algorithm are 3.36% and 5.05%,repectively,and the maximum relative errors are 5.62% and 8.41%,repectively.The results verify the effectiveness of the proposed algorithm,which provides fundamental basis for faults prediction of motors used in electric vehicle.