The Research On The Application Of Artificial Intelligence And Chaos Theory In The Realtime Simulation And Optimization And Decision Of Copper-Matte Converting Furnace

Copper matte converting system is a complicated process which has the characteristics of multivariable, nonlinearity, strong coupling, large inertia, time varying and uncertainty and very difficult to carry out real-time on-line control. The purpose of the optimization, decision-making and control of copper matte converting process is to improve the productivity and decrease energy consumption. The automation procedure of copper matte converting process is more and more important as copper matte converting reactors are widely applied in our country and better converting indices are required. Therefore, improving automation control level of copper matte converting process has already become an important direction of the development of converting process.The basic principle of copper matte converting process is briefly introduced in this paper. The basic characteristics, control targets and affecting factors are analysed. Spontaneous heating process, temperature field in lining and gas jetting phenomenon are studied. The optimization of slag making system is also studied. Underthe guidance of the method of thinking and technical line, "mathematical simulation------holographic emulation?-integral optimization'', aimed at economized energy consumption, the optimization strategies of the temperature around tuyere, flux adding system, cold burden adding system and blasting system are put forward. Original samples are self-standardized, noise samples are filtrated and variables used for modeling are selected. Pattern recognition techniques such as Primary Components Analysis(PCA), Optimal Decision Plane(ODP) and Partial Least Squares(PLS) are applied in the historical sample data. Wavelet neural network forecasting model is formed to predict the slag weight and components. A new forecasting model is formed by combining the auto-regressive (AR) method with the triple exponential smoothing method to dynamically predict the copper matte grade. A Chaos Genetic Algorithm(CGA) is come up with in this paper based on the ergodiciry of Chaos and the inversion of Genetic Algorithm. The optimization efficiency of CGA is evaluated quantitatively. The result compared with other algorithms, shows that the optimization efficiency of CGA is higher than others. Cold burden smelting dynamical model is developed responding to the characteristics of many different cold burdens and its corresponding components. Based on the mechanism analysis and AI technique, an optimization and decision-making model about operating parameters and an on-line simulation and detection model in the copper matte blowing process are built. A model to forecast the copper-making end point is also built by using smoke temperature in a factory.On the research mentioned above, the Intelligent Decision Support System of the Operation-Optimization is developed. After it is applied in a copper smelting converter, every production quota has been obviously improved. Blister copper productivity is increased by 6.0 percent, cold burden input is increased by 7.8 percent and average converter life-span is improved from 213 to 235 by 10.3 percent.To improve the Intelligent Decision Support System, an information integration system is developed including management, monitoring and decision-making. It is applied in a factory. The computerized decision-making, status monitoring and management are realized. Therefore, better economic benefit is achieved.