Optimized Control For Aluminum Reduction Cell Based On Knowledge Driven

With the rapid development of national economy, People growing demand for aluminum.At the same time,Our National policies is low-carbon and efficient because of Raw material and the sources of energy is getting strained day by day. Therefore the direction of technology research for Aluminum industry is becoming how to improve the efficiency of aluminum, extend the cell life cycle and save the energy.For the aluminum production control situation at present,The paper takes the background of control system transformation of 160KA prebaked in Chalco Guizhou Branch electrolysis plant,And proposes an optimal control program of the cell which based on knowledge-driven from three different hands.First of all ,This paper gives the description and analysis for the process and technical parameters of aluminum production detail.Describe the skills for knowledge representation,knowledge collection and knowledge manager based on the historical data and real-time data,Utilize the knowledge-driven to solve the model of complex data ,Carrys out the SOM cluster analysis for the raw data with advanced knowledge-driven approach,Then uses BP neural network which has been optimized by Genetic Algorithm to train the classified data and establishes a identify model of cell slots status.Secondly,For the problem that concentration of Al2O3 is difficult to track and control,This paper use the fuzzy decoupling approach to treat the cell's voltage data and get a new data sample.Combine the new data sample with expert knowledge to creat a determine model of alumina concentration.Identify feature control model of cell by using least-square recursion algorithm.Finally,For the blanking clearance (NB) of Al2O3 is influenced by setting resistance,the control system uses an algorithm combine human-simulated intelligent control with fuzzy control.It can draw up the control strategy,change the setting resistance and blanking clearance of Al2O3 based on the feature control model.according to the cell status and alumina concentration which is reasoned by above-mentioned two models.And then cell control will be faster, more accurate, more efficient in this way. We simulate the control system using matlab depending on the specific algorithm,And the results of simulation show that the optimal control based on knowledge-driven control algorithm is superior to the traditional, The control mode which combine human-simulated intelligent control with fuzzy control can make the control program of the cell have a better robustness,be better than traditional control mode in the control effect.This control system can reduce the times of anode effect, improve current efficiency, maintain the balance of material and energy, realize the purpose of energy saving and life-cycle prolonging.The optimal control system based on knowledge-driven has been applied to Chalco Guizhou Branch electrolysis plant successfully.After long time running and analysis,we find the optimized effect is obvious.The accuracy rate of cell slots status'identification and alumina concentration's detection can be more than 90%.The controlling process is rapid and accurate,implement energy-saving 160 degrees every ton of aluminum.