The Research Of Intelligent Control Method Based On Particle Filter About A Class Of Complex Industrial Process

With the development of the society,the intelligent control method plays an important role in a class of complex industrial process,such as petroleum,chemical,pharmacy,metallurgy,and electric power and so on.However,this class of complex industrial process often has many characteristics,including non-line,non-Gaussian,large lagging,and uncertainty.Because the traditional intelligent control methods,such as the fuzzy control,neural network control and expert system and so on,always need to rely on a lot of expert information and human experience to complete the modeling and optimization of the controlled system,which results the debugging process is much longer than other ways.Besides,once the control variable is changed,the control method also needs to be redesigned,which makes those control methods restricted in universal applicability of a class of complex industrial process,and cannot meet the requirement of industrial process in the robustness and control precision.The process of synthetic ammonia decarburization is a typical complex system,which has many characteristics such as time-varying,big lag,non-line,and non-Gaussian,so this paper chooses the complex industrial process of the synthetic ammonia decarburization as the research object.In order to overcome the shortcomings such as the slow convergence speed and low control precision caused by the traditional intelligent method in the non-line and non-Gaussian system,the particle filter algorithm which has the special advantage of deal with the problem of the estimation and prediction in the non-line and non-Gaussian system is put forward to control the process of the synthetic ammonia decarburization.And then aiming at the problems which are the particle degeneracy and need a lot of particles when the initial state is unknown in particle filter algorithm,the particle swarm optimization is used to optimize the control,that it can implement a bounded control for a controlled system effectively.This paper takes the crystal size of the soda in the synthetic ammonia decarburization as the research background,and uses the improved RBF neural network to build the reference model of the controlled object.According to the rule of minimum variance,the control quality assessment based on minimum variance is provided to make evaluation for the controlled process,so that it can supervise the quality index of the whole controlled process and also can find the problem in time,and then this paper uses the particle filter to estimate the reason and also provides a theoretical guidance to solve the problem.At last,according to the change of the control quality index,this paper using the proposed improved method of adaptive particle number,which can make the control method improved and perfected,and the intelligent method based on the particle filter can provide an effective research approach for a class of complex industrial process.Meanwhile,the yield and quality of products and the economic benefit and market competitiveness are all improved.