基于多级模糊神经网络的故障诊断在化工生产过程中的应用

化工生产过程一般都非常复杂，如柠檬酸蒸发。由于控制回路与测控参数很多，生产过程的故障检测与诊断问题非常困难，难以做到实时检查，得到其故障信息。所以本文提出一种基于神经网络的多级故障诊断系统。采用三级递阶模糊神经网络，降解整个系统故障诊断问题的复杂性，同时采用所有子神经网络全局并行的推理方式，具有快速处理能力，适合系统实时在线故障诊断。

Application of mulit - grade fuzzy neural networks in fault diagnosis of chemical processes

In general,chemical processes in industry is extremely complicated such as evaporation of citric acid.It is so difficult to detect and diagnose the fault in industrial processes that we can't obtain the information of on-line fault for there is so many control circuits and parameters of inspection.This paper presents a fault diagnostic system based on multilevel neural networks.By using three level neural networks,the complication of fault diagnosis in real system is simplified and it also has quickly handling ability to match the on-line fault diagnosis by using the rational measurement of juxtaposing sub-neural networks.