一种基于KPCA-ISSA-SVM的火控计算机电源故障诊断方法
A fault diagnosis method for fire control systems based on KPCA-ISSA-SVM
高锦涛 1李英顺 2郭占男 2佟维妍1
作者信息
- 1. 沈阳工业大学化工过程自动化学院,沈阳 111003
- 2. 大连理工大学控制科学与工程学院,辽宁大连 116200
- 折叠
摘要
传统坦克故障诊断主要靠专家经验,投入人力大、花费时间长.为满足装甲装备的健康管理需求,提出了一种基于核主元分析和改进麻雀算法结合支持向量机的故障诊断方法.针对火控系统信号成分复杂、数据量少的问题,首先利用核主元分析降维提取故障数据的非线性特征,减少其他冗余特征对故障识别的影响,降低数据维度.引入混沌Tent映射和非线性惯性权重因子对麻雀搜索算法进行改进,优化支持向量机核心参数并建立故障诊断模型,同时与粒子群优化和鲸鱼优化的支持向量机模型进行实验对比.实验证明:该方法可以有效对坦克火控系统进行故障诊断,且在准确率和诊断效率方面性能较高.
Abstract
Traditional tank fault diagnosis mainly relies on expert experience,which requires a large amount of manpower and takes a long time.To meet the health management needs of armored equipment,a fault diagnosis method based on kernel principal component analysis and improved sparrow algorithm combined with support vector machine is proposed.In response to the problem of complex signal components and limited data volume in fire control systems,kernel principal component analysis is first used to reduce dimensionality and extract nonlinear features of fault data,reducing the impact of other redundant features on fault recognition and reducing data dimensions.Introducing chaotic Tent mapping and nonlinear inertia weighting factors to improve the sparrow search algorithm,optimizing the core parameters of support vector machine and establishing a fault diagnosis model,while conducting experimental comparisons with support vector machine models of particle swarm optimization and whale optimization.Experimental results have shown that this method can effectively diagnose faults in tank fire control systems,and has high performance in terms of accuracy and diagnostic efficiency.
关键词
故障诊断/火控系统/支持向量机/核主元分析/麻雀搜索算法Key words
fault diagnosis/fire control system/support vector machine/kernel principal component analysis/sparrow search algorithm引用本文复制引用
基金项目
辽宁省科技计划(22JH1/1040007)
出版年
2024
NETL
NSTL
万方数据