Control Model Based On Ann Panzhihua Iron And Steel Bof Endpoint

It is an important prerequisite of organizing production reasonably,improving the quality of steel and reducing the cost that the accurate prediction of the ending steel-making in converter and optimizing the steel-making process parameters. On the status of the Iron and Steel Plant of Panzhihua, there is five 120 tons top and bottom blowing converters, three of which built in the early 1970s, equipped comparatively backward and only completed basic automation at present. They have no control-gun or other dynamic conditions and even still depend on artificial experience. so the fitting rate is low. The direction of technological transformation is building the static control model and put it in practice.The topic focused on the status of the Iron and Steel Plant of Panzhihua completed the converter optimal control system. The system includes forecasting function modules, control modules and Web users pages. In the subject, we begin with the establishment of the BP neural network forecasting model. The forecast demonstration are based on the 2000 historical data of which C content are less than or equal to 0.06 and C content of 0.06 to 0.29.For the former, in dual-output model, the fitting rate of C is 72.02 percent while the fitting rate of temperature is 61.87 percent on the conditions of the C content prediction accuracy:±0.01 percent, temperature prediction accuracy:±15℃. For the second, in dual-output model, the fitting rate of C is 77 percent while the fitting rate of temperature is 61.53 percent on the conditions of the C content prediction accuracy:±0.02 percent, temperature prediction accuracy:±15℃. For the last, in dual-output model, the fitting rate of C is 79.21 percent while the fitting rate of temperature is 58.13 percent on the conditions of the C content prediction accuracy:±0.03 percent, temperature prediction accuracy:±15℃.Finally based on the forecasting model, the optimal control function module is completed. Under the conditions of Semi-steel quality,original C content,original S content,original temperature,scrap,high magnesium lime and compound material. For the target of output C and output temperatures on the moment of widening carbon, get the carbon, oxygen and sludge, optimization process parameters by the module. In the optimal control function module, we used alternative model method and the "coordinate the rotation search method" and completed optimization process with one minute .That which is the subject of innovation lies is fully in line with the needs of industrial production. By Offline operation, the result is as the followed: for the kind of output C content which are less than or equal to 0.06, the hit rate of C is 70.2 percent while the hit rate of temperature is 61 in the dual-output model; for the kind of C content which are between0.06 to 0.29, the hit rate of C is 74.43 percent while the hit rate of temperature is 60.7 in the dual-output model; for the last, the hit rate of C is 76.12 percent while the hit rate of temperature is 57.8 in the dual-output model. In the single-output model, the temperature hit rate was 80.02 percent.