摘要
为了避免风量单一特征进行故障位置诊断的不适定性,提出基于风量-风压复合特征的故障位置诊断方法,实现特征信息的多维互补,提高故障位置诊断的准确度。利用蒙特卡洛方法生成大致满足实际故障风阻值分布的故障仿真样本,为了避免不同变量之间不同量纲、不同数量级造成的数据损失,对原始风量、风压数据进行标准化处理,并分别以风量单一特征、风压单一特征、风量-风压复合特征作为支持向量机(SVM)的输入,构建通风系统阻变型故障位置诊断模型。通过故障模拟实验研究表明:风量、风压单一特征进行故障位置诊断的准确度分别为89.80%,90.34%,风量-风压复合特征进行故障位置诊断的准确度为98.23%,说明风量-风压复合特征进行故障诊断可以消除风量、风压单一特征进行故障诊断的不适定性,提高故障诊断的准确度。
In order to avoid the ill-posed characteristic of fault location diagnosis based on the single feature of air volume,a fault location diagnosis method based on the composite features of air volume and air pressure was proposed to realize the multi-dimensional complementarity of feature information and improve the accuracy of fault location diagnosis.The fault simulation samples which could approximately satisfy the actual distribution of fault air resistance were generated by using the Monte Carlo method.In order to avoid the data loss caused by different dimensions and orders of magnitude between different variables,the original air volume and air pressure data were standardized,and the single feature of air volume,single feature of air pressure and composite features of air volume and air pressure were used as the input of support vector machine(SVM)respectively,then a model of resistance variant fault location diagnosis for the ventilation system was built.The results of fault simulation experiments showed that the accuracy of fault location diagnosis based on single feature of air volume and air pressure was 89.80%and 90.34%respectively,while the accuracy of fault location diagnosis based on composite features of air volume and air pressure was 98.23%,the air volume and air pressure composite feature could eliminate the discomfort of air volume and air pressure single feature in fault diagnosis and improved the accuracy of fault diagnosis.
作者
刘剑
刘丽
黄德
邓立军
周启超
LIU Jian;LIU Li;HUANG De;DENG Lijun;ZHOU Qichao(College of Safety Science and Engineering,Liaoning Technical University,Huludao Liaoning 125105,China;Key Laboratory of Mine Thermal Dynamics and Prevention,Ministry of Education,Liaoning Technical University,Huludao Liaoning 125105,China)
出处
《中国安全生产科学技术》
CAS
CSCD
北大核心
2020年第1期85-91,共7页
Journal of Safety Science and Technology
基金
国家自然科学基金项目(51774169)
国家重点研发计划项目(2017YFC0804401)
关键词
风量-风压复合特征
矿井通风系统
阻变型故障
故障诊断
蒙特卡洛
SVM
composite features of air volume and air pressure
mine ventilation system
resistance variant fault
fault diagnosis
Monte Carlo
SVM
