The Application Of Intelligent Control In Stored Grain Aeration Process

Grain is the essential material to maintain people’s normal life. Achieving grain storage safety plays a very important role, not only in maintaining people’s daily lives and ensuring market and social stability, but also in preparing against natural disasters. Aeration is an effective method to ensure grain safe storage. In this thesis, some intelligence control methods were studied and be used for stored grain aeration simulation. By comparing and analyzing the aeration simulation results, the better control method which could be used in grain aeration process control was obtained.The main contents of this thesis wre show as follows:1. PID control algorithm was attempted to be used to control the stored grain aeration process. An optimization objective function was built. Particle swarm optimization (PSO) was used to optimize the control parameters of stored grain aeration temperature PID controller and humidity PID controller. A simulation experiment for stored grain aeration process which was controlled by PID controller was conducted using Matlab. The parameters of temperature PID controller and humidity PID controller were those obtained from using PSO to optimize simulation aeration process. Moreover, a simulation experiment for stored grain aeration process which was controlled by PSO algorithm was conducted. By comparing and analyzing the two aeration simulation results, a conclusion that PID algorithm could control the grain temperature and moisture content to objective value faster than PSO optimization was drew.2. In this thesis, support vector machine technology was attempted to be applied to stored grain aeration process control by taking advantage of its powerful data processing capacity. By mining the aeration law implied in grain situation data which were acquired from aeration experiment in laboratory, a support vector machine stored grain aeration model was obtained. By using cross-validation method, the kernel function parameters and penalty factor which make the model with highest prediction accuracy rate were obtained. Mean square equation was used to calculate the prediction error of the obtained model. In order to verify effectiveness of the obtained model, a simulation experiment for stored grain aeration process which was controlled by the obtained model was conducted using Matlab.3. The simulation results of stored grain aeration process controlled separately by PID algorithm, PSO algorithm and SVM-based stored grain aeration model were compared and analyzed.