柴油机故障诊断中的变权重多证据体融合应用研究  

Application Research on Weight-varying Multi-evidence Fusion in Fault Diagnosis of Diesel Engine

作　　者：张玲玲[1,2] 廖红云[2] 乔龙[2] 任金成[2] 张晓倩[3] 

机构地区：[1]军械工程学院火炮工程系,河北石家庄050003 [2]军事交通学院汽车工程系,天津300161 [3]太原卫星发射中心技术部,山西太原030027

出　　处：《车用发动机》2012年第2期84-88,92,共6页Vehicle Engine

摘　　要：为了能够确定各证据的权重,对证据进行权重调整,应用了基于遗传算法的加权平衡证据调整方法。首先,利用遗传算法确定最优化的或者近似最优化的权重系数,然后依据加权平均值和优先权的证据分布形态对证据进行加权调整,且保证调整后的证据加权平均值不变以及优先权证据分布形态稳定。柴油机故障诊断实例表明,采用基于遗传算法的加权平衡证据调整方法能提高故障诊断的信任度和准确率,明显优于普通D-S等权重证据组合方法。To determine the weight of each evidence,the method of weighted evidence balance based on the genetic algorithm was used to adjust the weight of evidence.The optimal or nearly optimal weight coefficients were determined by the genetic algorithm and then the evidences were weighted and adjusted according to the weighted average and distribution of preferred evidence that should be constant before and after the balance.The experimental results show the weighted evidence balance method can improve the confidence and accuracy of fault diagnosis,which outgoes the basic evidence theory method.

关 键 词：柴油机 故障诊断 信息融合 遗传算法 D-S证据理论 

分 类 号：TK428[动力工程及工程热物理—动力机械及工程]