The Research Of Aluminum Electrolyzer Fault Diagnosis System Based On Fuzzy Neural Network

Aluminum electrolysis is a non-steady state, nonlinear, high energy, strong magnetic coupling process. In this paper, to point against the vibration of electrolyzer voltage and temperature, the author design a diagnostic system based on the Fuzzy Neural Network Model, using advanced control technology, for improving the current efficiency, reducing energy consumption and increasing labor productivity. There's such a great significance for promoting the rapid development of enterprises.At first, this paper introduces the model technology of aluminum electrolyzer. Describe the present conditions of aluminum electrolysis and the ordinary fault type and the dynamic model in diagnosis and prediction.Using the wavelet packet transform for the data acquisition. It could easily disintegrate every single vibration between the electrolyzer's voltage and isolate the attribute data from each frequency channel.It's perfect for the fault feature extraction. Using a fuzzy expert system and a new temperature sensing device, make it easy to solve the difficult point of electrolyzer temperature measurement. This article focuses on the establishment of improved diagnostic model Fuzzy Neural Network, using electrolyzer's voltage and temperature coefficient to judge the various status of abnormal groove. Adding subtractive clustering method to the Fuzzy Neural Network can be well coordinated control algorithm accuracy and calculation speed becaused the system greatly reduced the number of fuzzy rules.ANFIS improved diagnostic model inosculate the neural networks and fuzzy inference system, determine the optimal distribution of membership function and the parameter of neural network through the hybrid neural network. Ultimately find out the relationship between input and output, implement the controllable ability of the electrolyzer.Finally, the author selected a reasonable parameter for the model, through the simulation, then transplanted it to the locale control system. The electrolyzer forecasting function can reduce the consumption of energy, improve the engineering efficiency.