基于人工智能的矿用带式输送机故障诊断模型研究  

Research on Fault Diagnosis Model of Mining Belt Conveyor Based on Artificial Intelligence

作　　者：李洪亮 王庆华 LI Hongliang;WANG Qinghua(School of Mining and Coal,Inner Mongolia University of Science and Technology,Baotou 014010,China;Inner Mongolia Key Laboratory of Mining Engineering,Baotou 014010,China;Inner Mongolia Research Center for Coal Safety Mining and Utilization Engineering and Technology,Baotou 014010,China;Inner Mongolia Cooperative Innovation Center for Coal Green Mining and Green Utilization,Baotou 014010,China)

机构地区：[1]内蒙古科技大学矿业与煤炭学院,内蒙古包头014010 [2]内蒙古自治区矿业工程重点实验室,内蒙古包头014010 [3]内蒙古自治区煤炭安全开采与利用工程技术研究中心,内蒙古包头014010 [4]内蒙古煤炭绿色开采与绿色利用协同创新中心,内蒙古包头014010

出　　处：《金属矿山》2025年第2期152-160,共9页Metal Mine

基　　金：内蒙古自治区高等学校科学研究重点项目(编号:NJZZ22444);内蒙古自治区直属高校基本科研业务费项目(编号:2023QNJS102);内蒙古自然科学基金面上项目(编号:2024LHMS05007)。

摘　　要：为了提高矿用带式输送机故障诊断的准确率,针对传统的BP神经网络故障诊断模型诊断精度较低、泛化能力差、对初始权值和阈值敏感、容易产生过拟合等问题,提出了一种新的智能故障诊断模型。首先,采用主成分分析法(PCA)对数据样本进行处理,以降低数据的噪声和维度;然后利用隐藏层经验公式确定BP神经网络隐藏层神经元个数选取范围,在此基础上采用穷举算法和粒子群算法(PSO)组成嵌套粒子群算法(NPSO)对BP神经网络隐含层神经元个数、权值和阈值进行全局寻优,最终构建了基于PCA-NPSO-BP的智能故障诊断模型。基于实例,通过MAT-LAB软件仿真测试的结果显示,PCA-NPSO-BP故障诊断模型的诊断精度高于基于灰狼算法优化BP神经网络和基于遗传算法优化BP神经网络的诊断模型。In order to improve the accuracy of fault diagnosis of mine belt conveyor,a new intelligent fault diagnosis model is proposed to solve the problems of low diagnostic accuracy,poor generalization ability,sensitivity to initial weights and thresholds,and easy over-fitting of traditional BP neural network fault diagnosis model.Firstly,principal component analysis(PCA)is used to process the data samples to reduce the noise and dimension of the data.Then,the empirical formula of hidden layer is used to determine the selection range of the number of hidden layer neurons in BP neural network.On this basis,the exhaustive algorithm and particle swarm optimization(PSO)are used to form a nested particle swarm optimization(NPSO)to optimize the number,weight and threshold of hidden layer neurons in BP neural network.Finally,an intelligent fault diagnosis model based on PCA-NPSO-BP is constructed.Based on the example,the results of MATLAB software simulation test show that the diagnostic accuracy of PCA-NPSO-BP fault diagnosis model is higher than that of BP neural network optimized by grey wolf algorithm and BP neural network optimized by genetic algorithm.

关 键 词：带式输送机 主成分分析法 嵌套粒子群算法 BP 神经网络 故障诊断 

分 类 号：TD528[矿业工程—矿山机电]