一种基于KPCA-WOA-SVM火控系统故障诊断方法

A Fault Diagnosis Method for Fire Control Systems Based on KPCA-WOA-SVM

针对目前坦克火控系统维修成本高、人力投入大、状态评估涉及指标多等问题,为了提高故障诊断的准确率并降低成本,提出了一种基于核主元分析和鲸鱼算法结合支持向量机的故障诊断方法。首先应用核主元分析对火控计算机的性能参数进行特征提取,减少数据维度,其次应用鲸鱼优化算法参数寻优的支持向量机构建多分类的故障诊断模型,同时与遗传算法和粒子群算法优化的支持向量机进行对比,最后以火控计算机及传感器分系统的电源模块为研究对象进行实验验证。结果表明,该方法可以在较短时间内对火控系统的故障做到准确诊断,提高装备可靠性。

In order to improve the accuracy of fault diagnosis and reduce costs,a new method based on Kernel Principal Component Analysis and the Whale Algorithm combined with support vectors is proposed to solve the problems of high maintenance cost,large manpower investment,and too many indicators involved in state evaluation for tank fire control systems.Firstly,kernel principal component analysis(KPCA)is used to extract the performance parameters of the fire control computer to reduce data dimensions.Secondly,a multi-class fault diagnosis model is constructed by using the support vector machine(SVM)for parameter optimization of the whale optimization algorithm(WOA).Then it is compared with the support vector machine optimized by the genetic algorithm and the particle swarm algorithm.Finally,experimental verification is carried out with the power module of the fire control computer and the sensor subsystem as experimental objects.Results show that the method can accurately diagnose the fault of the fire control system in a relatively short time and improve the reliability of the equipment.