模式识别在电解铝生产过程故障诊断中的应用

为了提高电解铝生产过程的安全性和可靠性，将模糊理论和神经网络的模式识别方法应用到电解铝生产过程故障诊断中，使用模糊隶属度函数计算并归一化系统状态数据，获得数据隶属异常态的值，再将计算后的值输人至神经网络进行故障诊断，实现电解铝生产的故障诊断与报警，以提高生产过程的智能化水平。最后通过仿真实验，验证了上述诊断方法的有效性和准确性。

Application of Pattern Recognition in Fault Diagnosis of Electrolytic Aluminum Production Process

In order to improve the safety and reliability of electrolytic aluminum pro- duction process, pattern recognition method of fuzzy theory and neural network was applied in fault diagnosis of electrolytic aluminum production. Data on system state were calculated and normalized using fuzzy membership functions to obtain abnormal state values, and then the calculated values were put into a neural network fault diagnosis to realize fault diagnosis and alarming during production, which improved the intelligent level of production process. At last, the validity and accuracy of the diagnosis methods were verified by simulation experiment.