Development And Application Of LF Furnace Refining Alloy Feeding Model And Temperature Forecast Model

The composition and temperature of molten steel are the process parameters that are mainly controlled during the refining process of Ladle Furnace(LF).At present,most steel plants at home and abroad mainly control the composition of molten steel alloys by using operators to manually calculate the amount of alloy feed,and then send the calculation results to the first-level integrated control system for alloy feeding to adjust the composition of molten steel.For the control of the molten steel temperature in the refining process,due to the limitation of the temperature measuring component,continuous measurement of the molten steel temperature cannot be achieved.The operator needs to take multiple temperature measurements to obtain the molten steel temperature,and then pass the temperature measurement data to the first-level system for molten steel to adjust the temperature.This greatly reduces the production efficiency and production quality,and there are also greater safety risks.Therefore,in order to replace these complicated manual operations to realize the "one-key refining" production process of the LF furnace,it is indispensable and particularly important to establish an accurate alloy feeding model and temperature prediction model.In response to these problems,this thesis using the historical smelting data of Wuhan Iron and Steel Company to develop and apply the LF furnace refining system of the Bilai National Steel Plant in India,and establish the LF furnace alloy feeding model and LF furnace temperature Forecast model.Using the LF furnace refining system designed based on these two models,the calculated amount of alloy feed and the temperature prediction value can be transmitted to the first-level integrated control system in real time to complete the automatic adjustment of the molten steel composition and temperature.The main research contents and conclusions of this thesis are as follows:(1)Taking the LF furnace of the Indian Bilai Steel Plant as the research object,the original production conditions were designed,and an alloy feeding model based on mechanism modeling was established.This model is based on the deoxidized alloy feeding model and the composition alloy feeding model,which greatly improves the calculation accuracy compared with the original calculation method.(2)This paper establish the LF furnace NAS-GA-BP neural network temperature prediction model according to the original temperature measurement sampling method of the steel plant.The model uses the Monte Carlo method to realize the random self-search generation of the optimal structure of the BP neural network.The geneticalgorithm is used to optimize the weights and thresholds of the neural network,which makes the model convergence faster and the prediction accuracy higher.By using this model,continuous prediction of molten steel temperature is realized.(3)Using computer language,database and OPC technology,the design and application of LF furnace refining system based on alloy feeding model and temperature prediction model are realized.The system can realize the automatic adjustment of molten steel composition and molten steel temperature by transmitting the calculated amount of alloy feed and the predicted temperature value to the first-level integrated control system in real time.This system has been put into operation in the Bilai National Steel Plant in India for one year and has been running stably and achieved good results.This is of very important practical significance for improving production efficiency,reducing production costs,and ensuring worker safety.