Intelligent Optimization Control System Research And Software Development For Sintering Process

In the past decades, along with the development of sintering process control technologies as well as parameter detection technologies, the automation level of steel companies has been enhanced drastically. However, the energy consumption, the degree of coordination operating parameters and the optimization of technical indicators still remain face challenges for the sintering process whose raw material quality fluctuates wildly. Therefore it is of great importance to develop the corresponding intelligent optimization control system.This subject is a part of national key project implemented in Hangzhou Iron & Steel Group. In this project, we divided the sintering process into three segments as proportion, mixing and thermal status. Main tasks are as follows:1. We propose and realize a network structure by which the real-time acquisition and sharing is achieved so that the optimization system and the control system can be integrated smoothly. Hybrid programming with C# and Matlab was adopted to design the modules of object modeling, control strategy and optimization algorithm.2. Radial Basis Function Neural Networks has been used as main modeling method in this project. Peak density function and center clustering algorithm were proposed to design the RBFNN structure, which is the model of mixture moisture soft sensing and thermal status of sintering. Because of the relation of sinter is too complex to mechanism modeling, Principle Component Analysis (PCA) was used to design the sinter quality prediction model.4. The parameters table of ingredients and regulating rules of dosage was proposed to calculate the mount and the proportional. Then, Gradient descendant algorithm was adopted to design the multi channel adaptive controller. Based on that adaptive controller, the closed loop control system for mixture moisture has been implemented in field. Finally, we imported the differential operator and control weights into generalized predictive control, designing the GPC of sintering thermal status.Based on the above techniques, the intelligent optimization control system for sintering process has implemented successfully on No.1 sintering bed of the iron works. Finally by the analysis of the horizontal section influencing factors, it is proposed that dividing the sintering trolley horizontally into multiple channels. Using particles swarm optimization algorithms to improve multi-channel coordination of optimal control technology, the simulation obtained good research results.