Design Of Physical Model Control System On The Experimental Platform And Modeling For Temperature Prediction Of LF Furnace

The thesis is exposed a sub-project in "985 Project", named’experimental research platform design of the control system integration and process comprehensive optimization of steelmaking-refining-casting process’. The modeling and control system of LF furnace experimental platform has been completed. This experimental platform can simulate the industrial site production process and the actual control process in a laboratory environment, which lays the foundation of the research for production process optimization control and production scheduling control.The main research contents are as follows:(1) The realization of movement control of electric furnace and LF furnace physical modle part. It has been completed to control the electric furnace and LF furnace automatically on the experimental platform, including the movement of buggy ladle, the lifting of furnace cover, the lifting of electrode, the tilting of Furnace, the opening of cover and the flashing of alarm lamp. At the same time, the visual monitoring and control of overall process of steel-making have been realized by the building of WinCC monitoring interface configuration on PC and the communication between WinCC and PLC.(2) The establishment of the mixed prediction model of molten steel temperature of LF furnace. Firstly, establish the mechanism model of molten steel temperature in the process of refining, according to the law of conservation of energy. Secondly, identify the mechanism model parameter using the standard particle swarm optimization algorithm. Then, establish the mixed prediction model. Finally, confirm the model accuracy after the production data preprocessing.(3) The improvement of the temperature prediction model. The improved adaptive inertia weight mutation particle swarm optimization is used to improve the accuracy and stability of the mixed prediction model, because of the precocious phenomenon in the standard particle swarm optimization algorithm. With the running of refining process, the accuracy of the mixed temperature prediction model descends, which is caused by the model mismatch, time-varying and interference. It can ensure the accuracy of model to utilize reference furnace optimization strategy.The thesis has finished the control system design for the physical movement of electric furnace and LF furnace, and the automatic control of electric furnace and refining operations on the experiment platform through the design of the hardware and software in PLC.Besides, establish the mixed molten steel temperature prediction model of LF furnace using a combination algorithm of the mechanism and the particle swarm optimization, then improve the model considering the defects of the mixed model.