基于GA-BP-Adaboost模型的掘进机截割部故障诊断研究  

Research on Fault Diagnosis of Cutting Part in Roadheader Based on GA-BP-Adaboost Model

作　　者：刘伟健 刘洋[1] Liu Weijian;Liu Yang(Sany Heavy Industry Co.,Ltd.,Shenyang 110020,China)

机构地区：[1]三一重型装备有限公司,沈阳110020

出　　处：《煤矿机械》2023年第3期173-176,共4页Coal Mine Machinery

摘　　要：针对掘进机截割部故障预测问题,选取某煤矿井下掘进机截割部采集数据作为实例进行研究。先对BP神经网络和GA-BP神经网络进行预测对比分析。接着以GA-BP神经网络为基础,用Adaboost方法训练若干个弱分类器和弱预测器,将多个GA-BP分类器通过Adaboost算法组合成强分类器和强预测器。对采集到的掘进机截割部运行数据进行预处理,消除信号中干扰和漂移,提取多个时域特征参数。用MATLAB对GA-BP-Adaboost算法中的重要参数进行分析,得出结果:基于GA-BP-Adaboost算法的模型预测精度较高,说明GA-BP-Adaboost算法具有良好的预测性能和泛化性能,能够准确对掘进机截割部故障进行诊断。Aiming at the problem of fault prediction of the cutting section of the roadheader,the data collected by the cutting section of the roadheader in a coal mine was selected as an example to research.Firstly,BP neural network and GA-BP neural network were used for prediction and comparative analysis.Then based on GA-BP neural network,several weak classifiers and weak predictors were trained by Adaboost method,and multiple GA-BP classifiers were combined into strong classifiers and strong predictors by Adaboost algorithm.Preprocessed the collected operation data of the cutting section of the roadheader,eliminated the interference and drift in the signal,and extract multiple time-domain characteristic parameters.The important parameters of GA-BP-Adaboost algorithm were analyzed with MATLAB.It can be concluded thatthe model based on GA-BP-Adaboost algorithm has higher prediction accuracy.It shows that GA-BP-Adaboost algorithm theory has good prediction performance and generalization performance,and can accurately diagnose the fault of the cutting part of the roadheader.

关 键 词：截割部故障诊断 BP神经网络 遗传算法 ADABOOST算法 

分 类 号：TD421.5[矿业工程—矿山机电]