Design Of Marine Alkaline Protease Fermentation Control System Based On Iterative Learning Algorithm

Protease is the most widely sold and widely used enzymes in the market.It is widely distributed in nature.Marine alkaline protease MP,as a protease with new source,has better physicochemical properties than traditional terrestrial protease.It is clean and non-polluting during production and use.It has been widely used in medicine,washing,rubber manufacturing and other industries.In the industrial production of marine alkaline protease MP,it is necessary to control the environmental parameters of its fermentation process to maximize product yield,improve product quality and save cost.However,the growth and reproduction of cells and product metabolism are unpredictable.Influenced by many environmental parameters,the fermentation process presents strong non-linearity and difficult decoupling.Conventional industrial production control methods are difficult to achieve better control effect,which seriously restricts the large-scale production and application of the enzyme.Based on the above problems,this paper studies the fermentation process control of marine alkaline protease MP with the support of National Natural Science Foundation of China and Jiangsu University Advantage Discipline Construction Project.The main work and achievement of this dissertation are as follows.Firstly,Although the fermentation process of marine alkaline protease MP is complex and changeable,it has repeatability between batches.In view of this feature,an iterative learning control strategy for optimizing gain parameters by artificial fish swarm algorithm is proposed to realize dynamic tracking of controller gain parameters.In order to improve the accuracy of the optimization algorithm and reduce the operation time,the artificial fish swarm algorithm itself is included.The optimization work includes dynamic vision,eliminating congestion factor,automatically reducing fish size and exploring fish.The optimized controller structure is designed and its convergence is analyzed.The control effect of the method is preliminarily verified in simulation experiments for linear and nonlinear systems with single input and single output.Secondly,in order to apply the control model to the multi-input and multi-output fermentation control process of marine alkaline protease MP,a brief dynamic model of fermentation process is constructed,and the optimization model of artificial fish flock is established by using the fuzzy neural network.At the same time,an error compensator is designed for the possible errors and disturbances in the algorithm tracking process,which improves the robustness and control accuracy of the system.Finally,a fermentation control device for marine alkaline protease MP based on the S3C44B0 x chip and ARM platform is designed,which can be used in the actual production process.Firstly,the COM component of control algorithm is compiled by Visual C++ in the host computer,which is embedded in the chip and is responsible for the centralized processing of all fermentation tank data,while the data acquisition and control operation of fermentation site are realized by SST89E89516 RD,which corresponds to one-to-one fermentation tank.In order to realize the distributed control of multiple fermentation tanks and ensure the reliable transmission of data,a wireless module CC 1101 driver is compiled.Finally,the feeding and defoaming device in the fermentation process was designed.After practical application and debugging,the effectiveness of the equipment is verified.