Research On Fault Diagnosis Of Short Circuit And Modeling Method Current Distribution Of Copper Electrolysis Cell

In the process of copper electrolysis cell, due to the limitation of various process conditions, the short circuit fault, is commonly called burning plate, often appears between the cathode and anode plate in the electrolytic cell. If you can't find and deal with faults effectively, the current will increase dramatically, which will lead to the sharply increase of the temperature. In the same, the increase of short circuit current will lead to the decrease of the electrolytic efficiency, which leads to the decrease of the output of copper and brings the economic loss to the enterprise profit.At present, most of the electrolytic copper enterprises in the country have not very good prevention and detection equipment to the fault. In a domestic factory electrolytic copper plant, for example, in the length of about 400 meters, a width of about 30 meters of the workshop, the number of electrolytic cell have reached 860. And now, for the detection of the fault mainly rely on workers to check it using electromagnetic sensors, and need close to the cell surface circulation investigation, so have many drawbacks, such as, heavy workload, time-consuming, very prone to leakage phenomenon.In view of the above situation, this paper put forward a kind of online monitoring method of the copper electrolysis burning plate of cathode rod, and on this basis, current distribution model was established, reached the purpose of estimating the current values, the glass of rail, in the distance of 11 meters from the cell surface, installed the patrol car. In control room the computer controled the operation of the car. When the car along the orbit, the infrared camera installed at the bottom of it sequentially captured the surface image of electrolytic cell. Image through wireless AP is transmitted to central communication box in the central workshop. Finally, cable to infrared images and the car's running status information transmission to the computer. The GUI interface of the computer calibrated to the malfunction of the cathode bar, and which current values is displayed.Extraction the original image of the infrared camera, due to the existence of the characteristics of the abnormality and weak contrast, we carried out image preprocess, including the geometric correction, contrast enhancement, noise elimination and electrolysis cell edge extraction and segmentation steps. Through the analysis of the geometric features in fault location of the images as the input variables, the algorithm of BP neural network could accurate location and extraction for the fault, and then according to the severity of the fault, it divided into five grades on the classification. On this basis, the main factors affecting the current distribution of the cathode bar were analyzed, and the mathematical model of the current distribution of the cathode bar was established by 3 different methods. On the basis of the analysis of the performance index, the current values of the 3 models is verified by the field test data. Finally, the GUI interface is shown to display the fault image and the fault level of the cathode bar. In addition, the current value of all the cathode rod in the image is displayed in real time by using 3 model data and model average value, which realized the on-line estimation of the current values.