基于BP网络和专家系统的铝电解槽分层故障诊断

Aluminum Cell Hierarchical Fault Diagnosis Method Based on BP Network and Expert System[HS)]

摘要:

针对铝电解槽故障种类繁多和不易诊断的问题,设计了基于BP网络和专家系统的分层故障诊断系统,包括前层分类和后层预报；通过对槽电阻信号的频谱分析,提取了故障特征信息,并对故障进行分类；建立了基于BP网络的前层分类器,用于诊断特征显著的故障；制定了故障诊断和控制规则,完善了专家系统的知识库,根据前层分类结果对余下故障进行诊断；通过制定规则,将前层分类和后层预报相结合,实现了故障诊断系统的整体设计；仿真结果及理论分析表明,该系统可有效预报单一及复合故障,提高故障诊断的准确率,保证铝电解槽工作状况的稳定。

Abstract:

In order to solve the problem of great variety and diagnostic difficulty of the aluminum cell’ fault,a layered fault diagnosis expert system based on BP neural network and expert system was designed,which included former layer classification and after layer prediction. Based on the analysis for the frequency spectrum of the cell resistance signal,the fault feature information was extracted to classify the fault,A former layer classifier based on BP network was established to diagnose the obvious features of faults. Furthermore,with fault diagnosis rules and control rules as well as the consummate knowledge of expert system,the remaining fault was diagnosed according to the former layer classification results. The overall design of fault diagnosis system was achieved by setting the rules and combining the former layer classification and after layer prediction. The simulation results and theoretical analysis show that the system can effectively forecast the single and compound fault and then improve the diagnostic accuracy,ensure the working conditions of the aluminum cell.