基于IPSO-BP神经网络的钢丝绳断丝损伤识别模型研究

为解决传统钢丝绳断丝损伤识别方法精度低，BP神经网络陷入局部最优等问题，提出改进粒子群算法（IPSO）的BP神经网络识别模型。通过采集钢丝绳断丝损伤信号，提取缺陷信号特征，用峰值、峰峰值、波宽、波形下面积和波动能量5个特征值组成特征向量作为神经网络的输人，断丝数量作为神经网络的输出；利用改进粒子群算法对BP神经网络的初始权值和阈值进行优化；建立基于IPSO-BP算法的神经网络模型，用于钢丝绳断丝的定量识别。结果表明:IPSO-BPS神经网络模型的钢丝绳断丝损伤识别精度、泛化能力均高于传统BP神经网络模型，且改进的粒子群算法迭代寻优速度更快。

Research on identification model of broken wire damage on steel wire rope based on IPSO-BP neural network

In order to solve the problems that the accuracy of the traditional identification method of broken wire damage on the steel wire rope is low and the BP neural network fell into the local optimum,an identification model of BP neural network combined with the improved particle swarm optimization (IPSO) was put forward.The signals of broken wire damage on the steel wire rope were collected,and the characteristics of defect signals were extracted.The characteristic vector,which was composed of 5 characteristic values including the peak value,peak to peak value,wave width,area under waveform and fluctuation energy,was input into the neural network,and the amount of broken wire was taken as the output of neural network.The initial weights and thresholds of BP neural network were optimized by using IPSO,then a neural network model based on IPSO-BP algorithm was established and applied in the quantitative identification of broken wire damage on the steel wire rope.The results showed that the IPSO-BP neural network model had higher identification accuracy and generalization ability for the broken wire damage on the steel wire rope than the traditional BP neural network model,and the IPSO algorithm was faster in the iterative optimization.