摘要
为实现对电动汽车绝缘栅双极型晶体管(Insulated Gate Bipolar Transistor,IGBT)剩余使用寿命的精确预测,提出了一种基于多策略改进优化的金枪鱼优化算法(Tuna Swarm Optimization on mixed strategies,TSOM)的电动汽车IGBT剩余寿命预测模型。通过改进和优化ELM参数,利用NASA研究中心公开的IGBT老化加速实验数据集,分析并提取了集射极-发射极阻断电压的失效特征参数,以获取阻断电压尖峰信息。针对TSO算法在迭代前期容易陷入局部极值和后期局部开发能力不足的问题,选择Piecewise混沌映射初始化种群,改进其线性权重系数,并引入柯西变异-最优邻域机制来改进解质量增强机制对TSO算法进行多策略改进,增强算法跳出局部最优解的能力。并将该算法用于ELM相关参数的优化,以应用于电动汽车的IGBT剩余寿命的准确预测。通过与其他传统优化算法进行对比分析,结果显示TSOM-ELM方法在电动汽车IGBT剩余寿命预测方面具有高精度的特点。因此,本研究提出的方法可为其他电动汽车IGBT剩余寿命预测方法提供参考依据。
To achieve accurate prediction of the remaining service life of IGBT,a IGBT remaining service life prediction model based on a multi strategy improved optimization tuna optimization algorithm is proposed.By improving and optimizing the ELM parameters and utilizing the IGBT aging acceleration experimental dataset publicly available at NASA Research Center,the failure characteristic parameters of the collector emitter blocking voltage were analyzed and extracted to obtain the peak information of the blocking voltage.In response to the problem of TSO algorithm easily falling into local extremum in the early stage of iteration and insufficient local development ability in the later stage,Piecewise chaotic mapping is chosen to initialize the population,improve its linear weight coefficient,and introduce Cauchy mutation optimal neighborhood mechanism to improve the solution quality enhancement mechanism.The TSO algorithm is improved through multiple strategies to enhance the algorithm's ability to jump out of local optimal solutions.And apply the algorithm to optimize ELM related parameters for accurate prediction of IGBT remaining life.By comparing and analyzing with other traditional optimization algorithms,the results show that the TSOM-ELM method has the characteristic of high accuracy in predicting the remaining life of IGBT.Therefore,the method proposed in this study can provide a reference basis for other IGBT residual life prediction methods.
作者
席小卫
XI Xiao-wei(Lanzhou University of Information Technology,730000)
出处
《移动电源与车辆》
2024年第4期35-42,共8页
Movable Power Station & Vehicle
关键词
寿命预测
金枪鱼优化算法
极限学习机
电动汽车
life prediction
Tuna Swarm Optimization
Extreme Learning Machine
electric vehicles
