Study Of Intelligent Optimization Setting Based On Case-Based Reasoning For Clarifying Process In Sugar Mill

The clarifying process is one of important technology section in the sugar mill, and its result have direct influences to the refined sugar. The sulphitation process, which have the characteristics of short process, simple equipment, less investment, etc, is widely used in the world. But there are a large number of complex physical and chemical reaction in the process, that make the system with the characteristics of nonlinear, time-delaying, time-varying and strong coupling, and led to the some important technologies indexes cannot be directly measured. The operating variable of controller setting can be given based on the delaying test results and the rich experience. That is the main reason why the product indexes instability. Therefore, solving the problem of optimization setting in the sulphitation process to the clarifying section, that have the great significance to improve the quality of refined sugar and increase the economic benefits of cane.This paper detailed analyses the flow of the clarifying process. According to the characteristics of the clarifying process and the operating experience, the operation parameters optimal setting based on case-based reasoning is proposed to clarifying process. Firstly, according to the characteristic of technology process, selecting the system parameters as the input that have a great influence to the process indexes, the echo state network (ESN) prediction model is established. And the model of ESN based on Kalman filter is used to predict process indexes, and enhance the real-time performance of the system. According to the condition evaluation model, it can judge the good or bad to the current process indexes. Secondly, the paper analysis the application method to the case retrieve, case reuse, case revise, case retain and case maintenance, establish the initial database, and compute the weight coefficient of variable based on the neural network.Based on case-based reasoning, a optimization software package is developed, which is work in the Microsoft Visual2008platform, and complete the mixed programming with the other software, the software use the OPC protocol to communicate with KingView software, that complete the connection between the cane monitoring system and the data transmission. That has predicted the indexes and optimized the operation parameters based on the data, and provided certain scientific basis to the decision of the operator.