基于PSO+SOM神经网络的无人机装备故障智能诊断研究  

Research on Intelligent Diagnosis of UA V Equipment Failure Based on PSO+SOM Neural Network

作　　者：沈延安 陈强 杨克泉 SHEN Yanan;CHEN Qiang;YANG Kequan(Army Artillery and Air Defense Academy,Hefei 230031,China)

机构地区：[1]陆军炮兵防空兵学院,合肥230031

出　　处：《火力与指挥控制》2025年第1期152-159,168,共9页Fire Control & Command Control

摘　　要：针对当前无人机装备故障人工诊断效率低、智能诊断方法少、故障识别正确率低以及SOM神经网络收敛速度慢等问题,提出一种基于PSO+SOM神经网络的故障智能诊断方法。通过改进PSO算法优化SOM神经网络和对比PSO、GA、ACO对SOM神经网络的改进效果,以及比较LVQ、BP、传统SOM、PSO+SOM神经网络的故障诊断效果,结果表明PSO+SOM神经网络的故障诊断模型具有适度值小、判别时间短、迭代次数少、准确率高、收敛速度快的优点,为实现无人机装备故障智能诊断提供一种高效的方法。In view of the current problems of low manual fault diagnosis efficiency,less intelligent diagnosis methods,low fault identification accuracy rate of UAV equipment and slow convergence speed of SOM neural network,a fault intelligent diagnosis method based on PSO+SOM neural network is proposed.By improving PSO algorithm to optimize SOM neural network and comparing the improvement effects of SOM neural network with PSO,GA and ACO and comparing the fault diagnosis effects with LVQ,BP,traditional SOM and PSO+SOM neural network,the results show that the fault diagnosis model of PSO+SOM neural network has small moderate value,short discrimination time,less iterations,high accuracy and fast convergence speed,an efficient method for implementing the intelligent fault diagnosis of UAV equipment is provided.

关 键 词：无人机 SOM神经网络 PSO算法 智能化 故障诊断 

分 类 号：TB302.3[一般工业技术—材料科学与工程]