| The method of cane sugar with carbonate is very important in the Southern of China, while clarification process is the key part of the cane sugar, in which complex physical and chemical reactions was contained, with the Larger lag, many operating variables, while the strong coupling characteristics is very hard to detect. Currently, the setting of the operating parameters of the clarification process is basically relying on the on-site personnel to determine. But, the operators’experience is different, which makes the clarification process unstable. So the only way to improve the quality of the refined sugar is to solve the problem.This paper analyzes the process of clarification, and case-based reasoning methods are used to optimize the operating parameters settings. Also, summarizes the characteristics of parameters in the process of clarification to design the monitoring system including expert stations, operator stations, and field devices. On the basis of mechanism analysis, some important parameters are selected as the input of the prediction model based on T-SRFNN to predict the value of calcium salt and pigment. The model of clarification conditions is created to distinguish the good condition form the bad. If the condition is not good enough, CBR must be used to change related parameters until the condition get better. This paper analysis the application of case-based reasoning technology, what is more, a simple example is given. Unsupervised clustering algorithm is used to classify data and the similarity between two cases is calculated by the Euclidean distance.This article uses the Visual C++6.0to development CBR software which supports the OPC specification. With CBR system, parameters of the process of clarification can be monitored at any time. Users can entry Offline data by the software to get the sample of the neural network training. The test proves that it runs well in field and provides users the best way to solve problem. |
PDF Full Text Request