Research Of End-point Prediction Method To Vacuum Induction Furnace

Vacuum induction furnaces which are mostly used to produce super clean steel and alloy materials have been found numerous applications in aviation, navigation, nuclear power, military and chemical engineering.During the process of steel-smelting in vacuum induction furnaces, end-point parameters such as tapping temperature and carbon content in molten have badly affected the quality of the ingot steels. There is a close relation of the parameters and the performance of vacuum induction furnaces with high efficiently and low consumption. However, it is difficult to measure the temperature and the carbon content during the steel-smelting because it is a high-temperature and strongly corrosive process. As a result, the end-point prediction is very important during the steel-smelting process.Based on the research of a lot of correlative literature, the existing prediction method and the existing state have been proposed firstly. Then an end-point forecast model was built based on the RBF neural network by analyzing the melting technology and the RBF training algorithm.The building of the prediction model for end-point temperature and end-point carbon content has been detailed introduced. A two-round prediction method which can give error correction has proposed based on the neural network prediction model. It can give reliable predictions of end-point time and carbon content of molten steel in the first-round prediction. And the prediction accuracy can be improved by the error correction in the second-round prediction. Total 120 set of data are used for the two model training and validation, respectively. The predicted hitting ratio is high. The results show that the proposed method is effective, indicating the potentials for real-world application.