The Research And Design Of Converter Steelmaking Endpoint Guiding System

The control of converter end point has a direct effect on the cost of production, steel quality and labor productivity. Currently, most of the country steel plants still use the method of manual experience control which of low hit rate, not only consumes the cost, prolongs the blowing time but also affects the later operation such as continuous casting. Therefore, to improve the converter end point hit rate is of particular importance.This thesis which based on Dalian special steel plant production process makes a systematic analysis on converter steelmaking mechanism and builds a converter static mechanism model. Because the converter steelmaking process is very complex and highly random, having low hit rate when testing the mechanism static model, artificial intelligence control for improving the end point hit rate becomes a very effective way.Through analyzing the converter end point control influencing factors, this thesis firstly builds a converter expert directing system and has proves the feasibility of this control method based on the static mechanism model. Secondly, the thesis builds a LM-BP neural network prediction model. Based on Dalian steel plant production data, a training and test is made to this neural network and the result shows the model prediction has high accuracy. The end point carbon content is within the range of 0.032%-0.125%, liquid steel temperature within the range of 1645℃-1688℃and the end point dual-hit rate up to 65%. According to the feeding liquid steel information, the thesis combines the expert system and the BP neural network and finds a most close target value feeding material information through the model prediction, in this way providing a more precise guide for the operator.