Research Of Set-intervals Optimization Approach For Glutamic Acid Fermentation Process

In the field of control,people expect to get information from the system,so as to effectively model,control and optimize the system.For set-values optimization of complex industrial processes,the conventional way is to set-points optimization.However,in actual process control,due to random disturbance,strong coupling,non-linearity and other characteristics of the process,or lack of sufficient control freedom,it is difficult to achieve precise tracking of the set point at the loop level;even for a class of industrial process,it is not necessary to strictly maintain the control quantity at the set-points,so when optimizing the set-values of complex industrial process,the set-intervals optimization is more appropriate.In this paper,the point-valued process model is established and the optimal solution is used to solve the problem of interval setting value optimization,that is,the interval mathematical model of the process is established based on the process data,and then the interval optimization algorithm is applied to solve the interval optimal solution of each operation variable in the process.A new method based on interval analysis theory is proposed to optimize the set-values of complex industrial processes,which includes establishing the interval model of processes based on interval neural network,and then solving the interval settings based on interval particle swarm optimization algorithm.Taking the glutamic acid fermentation process as an example,the set-values optimization of operation variables in the process was carried out based on the above method,and good simulation results were obtained,which provided a new way for the modeling and optimization of complex industrial processes.The main work of this paper is as follows.(1)Introduce glutamic acid fermentation process,interval analysis theory,random weight neural network structure,weight learning algorithm and particle swarm optimization algorithm.(2)An improved weights learning algorithm for interval random weight neural network is proposed.The penalty factor is introduced into the error function,and the weight of the hidden layer to the output layer is adjusted by the principle of least square method.The simulation results show that the improved weights learning algorithm has excellent convergence and approximation.(3)Based on the basic interval particle swarm optimization algorithm,an interval particle swarm optimization algorithm with fixed particle width is proposed,which replaces the original method of randomly generating interval particles by presupposing the interval width of the population particles in initialization.The numerical simulation results show that the success rate of the algorithm is higher than that of the basic particle swarm optimization algorithm.(4)The prediction model of glutamic acid fermentation process is established by using the interval random weight neural network,and then the simulation experiment is carried out to prove that it has a good prediction effect.Then,on the basis of the interval random weight neural network,using the interval particle swarm optimization algorithm,the product concentration optimization trajectory and the optimization trajectory of each operation variable with the maximum product concentration as the goal are obtained,and compared with the results of the point set value optimization strategy,the necessity and advantages of the set-intervals optimization method are proved.