摘要
现有电炉电能计量装置异常诊断系统的数据处理方式比较有限,只能根据预设规则进行简单的故障判断,未利用机器学习等先进技术对大量数据进行分析和挖掘,难以发现潜在的异常模式和规律,导致诊断精度较低。为此,设计基于关联规则挖掘的电炉电能计量装置异常诊断系统。首先,设计系统功能模块,其中包括主控模块、远程通信模块和数据管理模块;然后,建立系统知识库,并在知识库内通过关联规则挖掘技术挖掘电炉电能计量装置异常数据;最后,将上述挖掘到的异常数据输入到卷积神经网络模型中,通过学习和训练完成电炉电能计量装置异常诊断,并输出异常诊断结果。实验结果表明,所设计系统的电炉电能计量装置异常诊断准确率较高,具有一定的技术水平与实用性。
The data processing methods of the existing abnormal diagnosis system for electric furnace energy metering devices are relatively limited,and can only make simple fault judgments based on preset rules.Advanced technologies such as machine learning are not used to analyze and mine a large amount of data,making it dfficult to discover potential abnormal patterns and patterns,resulting in low diagnostic accuracy.To this end,a fault diagnosis system for electric furnace energy metering devices based on association rule mining is designed.Firstly,design system functional modules,including the main control module,remote communication module,and data management module;Then,establish a system knowledge base and use association rule mining technology to mine abnormal data of electric furnace energy metering devices within the knowledge base;Finally,input the abnormal data mined above into the convolutional neural network model,complete the abnormal diagnosis of the electric furnace energy metering device through learning and training,and output the abnormal diagnosis results.The experimental results show that the designed system has a high accuracy rate for abnormal diagnosis of the electric furnace energy metering device,and has a certain level of technical level and practicality.
作者
杨子成
卢建生
王超
郭海旭
YANG Zicheng;LU Jiansheng;WANG Chao;GUO Haixu(State Grid Shanxi Electric Power Company,Taiyuan 030000,China)
出处
《工业加热》
CAS
2023年第11期38-42,47,共6页
Industrial Heating
基金
山西省企业科技攻关项目(16438246132)。
关键词
电炉电能计量装置
关联规则挖掘
异常诊断
知识库
卷积神经网络
electric energy metering device of electric furnace
association rule mining
abnormal diagnosis
knowledge base
convolution neural network
