基于自适应变异的粒子群优化BP神经网络的液压缸故障诊断方法  

Fault Diagnosis Method for Hydraulic Cylinder Based on Adaptive Mutation Particle Swarm Optimized BP Neural Network

作　　者：赵庆浩 周军 荆丰伟[1] ZHAO Qing-hao;ZHOU Jun;JING Feng-wei(National Engineering Research Center for Advanced Rolling and Intelligent Manufacturing,University of Science and Technology Beijing,Beijing100083,China;Nanjing Iron and Steel Co.,Ltd.,Nanjing 210031,China)

机构地区：[1]北京科技大学高效轧制与智能制造国家工程中心,北京100083 [2]南京钢铁股份有限公司,江苏南京210031

出　　处：《机电产品开发与创新》2024年第4期103-107,共5页Development & Innovation of Machinery & Electrical Products

摘　　要：本文创新性地探讨了一种液压缸故障诊断方法,该方法应用BP神经网络算法和自适应变异的粒子群优化方法,实现液压缸故障推理和判断。与传统的PSO-BP神经网络模型对比,该模型借鉴了遗传算法的思路,应用变异理论,使得粒子能够跳脱出先前搜索到的最优位置,再次进行更广泛地搜索。这种搜索方式使得算法搜索空间有较大的提升,使得算法寻优能力大大提高,有效提升了BP神经网络液压缸故障诊断模型的效率。This paper innovatively explores a hydraulic cylinder fault diagnosis method that integrates adaptive mutation particle swarm optimization and BP neural network.In contrast to the conventional PSO-BP neural network model,we introduce a mutation operation,drawing inspiration from the genetic algorithm's concept,allowing particles to break away from previously discovered optimal positions and conduct more extensive searches.The introduction of this mutation operation expands the search space,enhances the algorithm's potential to find better solutions,and effectively improves the efficiency of the BP neural network hydraulic cylinder fault diagnosis model.

关 键 词：自适应变异粒子群 BP神经网络 液压缸 故障诊断 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]