Research On Optimal Control For Fermentation Process Based On Improved Interval Algorithm

Interval optimization algorithm with interval variables instead of point variables is a deterministic global method. By using interval optimization algorithm to optimize and control industrial process, the noise and disturbance can be prevented from influencing system.To avoid the lost of optimal interval caused by particle swarm which works as interval accelerated tools, enhance the scope of application and accuracy of interval optimization, the existing interval-particle swarm was modified in this paper. On the basis of it, the modified algorithm was applied to single objective and multi-objective optimization of glutamic acid fermentation. This thesis includes the following aspects mainly:Development process, basic concepts and idea of traditional interval optimization algorithm were introduced. Basic process of traditional interval optimization algorithm was illustrated in detail.Advantages and disadvantages of interval-particle swarm were analyzed.On account of the shortcomings, interval-particle swarm optimization algorithm was improved in three aspects in order to enhance the quality of the optimal interval. Three diversiform functions were used to determine its reliability.The modified interval optimization algorithm was applied to optimize glutamic acid fermentation process. Neural network was used to found an interval model. Rolling optimization idea was used and the acid production rate of fermentation process was taken as the optimization target to optimize single operating variables.At the end of this thesis, research was summarized, and improving direction of interval algorithm and application in control theory ware given. Then the interval optimization curves of acid production rate, conversion rate and the operating variables in each phase were got by optimizing multi-operating variables with acid yield and conversion rate as target.On this basis, improved algorithm was applied to the multi-objective optimization of glutamic acid to optimize multiple operating variables simultaneously with acid yield and conversion rate as the target. According to the actual situation, the whole fermentation process can be divided into two phases of simulation. The simulation results show that the improved algorithm can meet the needs of the two phases of fermentation.