Fault Diagnosis Of The Anode For Aluminum Electrolytic Cell Based On Short-time Fourier Transformation

Aluminum electrolysis is an important method for aluminum treatment in modern industry.Its operation is carried out in an aluminum electrolysis cell.The anode of the aluminum electrolytic cell has an important influence on whether the aluminum electrolytic cell can be normally produced and various economic and technical indicators.If the anode fails,the aluminum electrolytic cell will be disordered,the power consumption will be increased,the current efficiency will be reduced,and the production efficiency will be reduced.The research on the anode fault diagnosis of aluminum electrolytic cell is a great significance to improve the production efficiency and safety of aluminum treatment in modern industry.This article aims to collect signals,study signal characteristics,and diagnose the fault of aluminum electrolysis based on signal characteristics.In order to improve the accuracy of the anode operation fault diagnosis,the subject selected the equal distance drop of the anode guide rod that can reflect the state change in the electrolytic cell as the acquisition signal.Because the operation of the aluminum electrolytic cell is a non-stationary signal that changes in real time and the signal is non-linear,using the short-time Fourier transformation method to analyze the collecting signals.By analyzing the nature advantages and disadvantages of different window functions,comparing the spectrum obtained with three different window lengths and combining the nature of the equal distance drop of the anode guide rod signal,selecting a Hamming window with a window length of 256 for Short-time Fourier transformation analyzes signal data.The time,frequency,and amplitude results of the short-time Fourier transformation analysis are saved to an Excel file.The results of the amplitude obtained by the movement of the window function 4 times are grouped and the top five of each group are extracted.The amplitude and its corresponding five frequency values are used as signal eigenvalues.Three methods,average value,weighted average,correlation coefficient are used to take the fault diagnosis.The method of taking the average value is to statistically obtain the fault feature parameters by averaging the signal eigenvalues,by analyzing the fault characteristic parameters,the critical value for determining whether or not there is a fault can be found,thereby finding all the fault points at the sampling point.The method of weighted average is to compare the results of fault diagnosis under three different weights and take the optimal weight(0.3 0.25 0.2 0.15 0.1)for fault diagnosis.The method for finding the correlation coefficient is to take a sampling point 12800-13312 of the fault point sequence A3 and want to find the fault point sampling sequence,and find the fault point by finding the correlation coefficient of the two sequences.By comparing the three methods,we found that the diagnostic method based on the correlation coefficient can find all the fault points more accurately,so this diagnostic method is applied to the design of the anode fault diagnosis system.In this paper,with the aid of Access database and MATLAB software,a set of fault diagnosis system software was designed and a friendly interface was written.The system implements the functions of MATLAB importing Access data,querying data processing results,drawing curves and diagnostics.