Research On System Identification Method Of Adaptive Reduced Order In Process Industry

System identification is a vital research branch in process industry,which is highly concerned by domestic and foreign experts and scholars.The effectiveness of the identification method will directly affect the accuracy of the system model,and then influence the application effect of the model.For the real process industry,identification is usually conducted by the specified order model.Since the system model order does not match the actual order,it becomes more difficult to obtain an accurate mathematical model.Therefore,it is of high research significance and application value to obtain the system model order more accurately.At the same time,with the development of computer and information technology,various intelligent solving algorithms emerge in endlessly.Among them,Particle Swarm Optimization(PSO)is a very practical intelligent optimization algorithm with strong global optimization ability.It has been widely applied in the field of system identification,and has achieved remarkable results.Therefore,this paper makes use of Particle Swarm Optimization(PSO)algorithm and proposes an adaptive order reduction system identification method.The simulation results show that the proposed method is effective under the ideal environment and white noise environment.In this paper,the main research work is as follows:(1)In the study of process industry system identification,for the system,the input and output data of the system can only be obtained.It is difficult to obtain the model order of the transfer function accurately based on these data,which influence the identification of the mathematical model.In order to solve the problem that the order of the model is unable to be determined,an adaptive order reduction system identification method based on PSO is proposed.In this method,the high-order model is designed as the initial model of the system.The PSO algorithm is used to update the velocity and displacement of the particles,iteratively searching for the parameters of the original model.The highorder parameter and fitness function of the model are used to judge whether to perform adaptive order reduction identification.Then,a simulation case is given to verify the effectiveness of the adaptive order reduction system identification method.(2)Generally,the production data of the process industry is a time series with white noise.First,through MATLAB simulation experiment,we simulate the unit step response containing white noise,filter using band pass filter,and then use adaptive order reduction system identification method to identify the unit step response after filtering.Finally,we get the mathematical model of the controlled object.This section verifies the feasibility of the adaptive order reduction system identification method in the case of noise interference.(3)The identification of complex controlled objects and the tuning of PID parameters.Firstly,the identification method proposed in this paper is used to identify complex controlled objects.Then,the identified controlled objects are integrated into a typical PID control system to analyze the error response of the system.Finally,PSO algorithm is used to iteratively search the global optimal solution,and then to get PID parameters.At the same time,compared with the PID parameters obtained by the traditional empirical method,the superiority of PSO-based adaptive order reduction system identification method is demonstrated.