一种露天矿卡车故障的NLP技术挖掘与分析方法

受制于国内露天矿设备的智能化故障诊断水平,露天矿卡车故障信息以人工采集方式为主,故障信息的标准化和故障管理的程序化程度不够,难以进行深入的故障分析工作。为此,提出了基于自然语言处理技术的故障分析方法。该方法首先对故障文本信息进行预处理,利用向量空间模型对处理结果进行统计,从而获得初始特征向量;然后利用主成分分析算法对初始特征向量进行降维处理,建立故障特征空间模型;最后利用这些特征向量训练故障挖掘模型,对露天矿卡车故障信息进行挖掘与分析。通过矿山实际数据验证了该方法的合理性,为露天矿卡车故障信息分析与管理提供了新方法。

A Mining and Analyzing Method of Truck Fault in Open-pit Mine Based on NLP Technology

The fault information of trucks in open-pit mines is obtained mainly by manual collection method,due to the poor ability of intelligent fault diagnosis. With low levels of standardization of fault information and procedure of fault manage-ment,the fault analysis work cannot be conducted deeply. Considering above-mentioned problem,a new fault analysis method based on natural language processing technology is proposed. The initial feature vectors are obtained firstly by pretreating fault text information,the vector space model is used to gather statistics on pretreatment result,and the initial feature vectors are ob-tained;then,the principal component analysis algorithm is adopted to descend dimension of initial feature vectors to establish the fault feature spatial model;finally,the fault mining model is trained by using these feature vectors,so as to analyze and min-ing the fault information of the truck in open-pit mine. The rationality of the mining and analyzing method proposed in this pa-per is verified based on the practical data of mines,and it is further indicated that it can provide a new and effective approach for analyzing and managing the fault information of trucks in open-pit mines.