Research On The Application Of Soft-sensing Technique In The Distilling And Fermentation Process

The main research of this paper is how to apply Soft-sensing Technique in the process of distilling and fermentation to solve the problem of an effluent quality parameter of the amylum rate which is usually difficult to measure. If we apply the Soft-sensing Technique in this process and build its soft-sensing model ,we can solve the problem there is only one way of chemistry analysis in laboratory to achieve its value. However.this way has many disadvantages such as long time needed, being easily affected by human-factors and difficult to be forecasted and controlled by computer. Therefore.the process can't be operated effectively.This paper mainly includes the following two parts:(1) The review on the Soft-sensing Technique:The first part of this paper introduces the basic theory of Soft-sensing Technique and its engineering design steps, then elaborates the basic theories of MLR,BP neural network and RBF neural network.( 2 ) The research on application of Soft-sensing Technique in the distilling and fermentation processThis part is the core part of the paper. Firstly.it introduces the technology of the alco-distilling and its flow process.Secondly, with the help of MATLAB 6.5 tool, Soft-sensing models about the amylum rate are built separately with three methods of MLR statistical regression.BP neural network and RBF neural network .From the simulations, the results show that: MLR is not so effective as BP neural network and RBF neural network both in the approximation and the prediction of the parameter; BP neural network can basically approximate to the anticipated requirement while it has bad abilities of predication and has other disadvantages such as the existence of partial minimum and the low network malleable pace;RBF neural network has higher precision and higher astringent pace, also it is the simplest way of the three ways. So, RBF neural network can more adapt to the real-time identification of quality parameters in complicated processes such as distilling and fermentation process and will provide a base for the realization of automation in the distilling and fermentation process.