The Study Of Multivariate Fuzzy Neural Control In The Temperature For Fermentation Process Of Beer

The brewery fermentation jar has inherence characteristics of multivariable, seriouslynonlinear, uncertain, time-variant and large delay. Since it is hard to build the precisemathematical model, the traditional control strategies can't trace the given techniqueflowcharts very well. For it can transform a MIMO system into several independent SISOsystems, Multivariate -decoupling control can make every block run very well. Fuzzy logicputs emphasis on the simulation of human being's thinking and appropriates to expresssomething fuzzy or indefinite. While neural network lays emphasis on the simulation of thestructure of man's brain neural network, and has the capability of concurrent computing,distributed data-storage, tolerance and selfadaptation learning.In this paper, according to the ferment mechanism, I weld these three controlstrategies together and employ segmented control strategy in the process of ferment, that issegmenting the ferment process into ferment prophase and ferment anaphase. In theprophase, control variable is the open degrees of top valve and the middle valve. But in theanaphase, the middle and the top of the jar have the same field. So these two valves havethe same state. The bottom valve and the middle valve form a MIMO system. In everyferment phases, I apply the strategy of fuzzy neural control based on Multivariate–decoupling, and simulate in MATLAB, the results show that this strategy's validity andpracticability in controlling beer ferment temperature.