摘要
当前机电故障数据检测算法,不能有效提取高阶特征量,造成电故障数据集离群点检测准确率低、实时性和自适应性较差,据此,提出基于数据挖掘的机电故障数据集离群点检测算法。构建机电故障数据传感信息采集节点分布模型,采用多传感器融合采样方法进行机电故障数据采样,提取机电故障数据的统计特征量,采用高阶特征量提取方法进行机电故障数据的样本信息分析和回归检测,挖掘机电故障数据的离群点特征量,分析机电故障数据集离群点差异分布特性,通过挖掘机电故障数据集离群点的属性特征,实现机电故障数据集离群点的检测优化,根据检测结果实现机电故障。实验结果表明,采用该方法进行机电故障数据集离群点检测的准确性较高,对故障检测的实时性和自适应性较好。
The current electromechanical fault data detection algorithms cannot effectively extract high-order feature quantities,resulting in low accuracy of outlier detection,poor real-time and adaptiveness of the electrical fault data set.Based on this,a data mining-based electromechanical fault data set is proposed.Group point detection algorithm.Construct the distribution model of the mechatronic fault data sensing information collection node,use the multi-sensor fusion sampling method to sample the mechatronic fault data,extract the statistical feature quantities of the mechatronic fault data,and use the high-order feature quantity extraction method for the sample information analysis and regression of the mechatronic fault data Detection,mining of outlier feature quantities of electromechanical fault data,analysis of the distribution characteristics of outliers of electromechanical fault data sets,and detection of outliers of electromechanical fault data sets,realizing the detection of outliers of electromechanical fault data sets Optimized to achieve electromechanical failures based on detection results.The experimental results show that the accuracy of outlier detection of the electromechanical fault data set using this method is higher,and the real-time and adaptive performance of fault detection is better.
作者
范英铭
FAN Ying-ming(College of Rail-Transit,Nanjing Vocational Institute of Transport Technology,Nanjing 211188,China)
出处
《新一代信息技术》
2019年第22期53-59,共7页
New Generation of Information Technology
基金
全国交通运输职业教育教学指导委员会2019年交通运输职业教育科研项目(项目编号:2019B03)。
关键词
数据挖掘
机电故障
数据集
离群点
检测
Data mining
Electromechanical failure
Dataset
Outlier
Detection