Identification Of Nonlinear System Driven By Non-uniform Sampling Data

In practical industrial applications,there are a wide range of nonlinear systems where the input refresh and or output sampling of a class of systems exhibit unequal time intervals.Aiming at such system identification problems,traditional identification methods are no longer applicable to such systems.To solve the problem of identification of non-linear systems driven by non-uniformly sampled data,this paper proposes a chaotic natural selection particle swarm algorithm based on the standard particle swarm algorithm.The chaotic natural selection particle swarm optimization algorithm applied to natural selection different non-uniform sampling data driven nonlinear systems,respectively discusses the non-uniform sampling data with known base driven Hammerstein model,Wiener model and Hammerstein-Wiener model,with a dead zone characteristics of non-uniform sampling data driven Hammerstein model,Wiener model identification and chaotic natural selection particle swarm algorithm in the application of the electric arc furnace electrode adjustment system.The specific work of this article is as follows:(1)Aiming at the shortcomings of the standard particle swarm algorithm,such as easy to fall into the local optimum,and the diversity of the population in the late convergence period,a chaotic natural selection particle swarm algorithm is proposed.By introducing chaotic sequences in the setting of the initial population position,the initial coverage of the particle position is increased,there by increasing the accuracy of finding the optimal solution.In the control algorithm,the ability of algorithm exploration and development also plays an important role in the performance of the algorithm.In the control algorithm,the exploration and development ability of the algorithm also plays an important role in the performance of the algorithm.By introducing the nonlinear decreasing inertia weight adjustment mode and the asynchronous learning factor adjustment mode,the development and exploration ability of the algorithm is increased.By introducing natural selection,the particle population is updated to improve the algorithm's population diversity.Through the improvement of the above three aspects,the convergence accuracy and convergence speed of the algorithm are improved.(2)Aiming at the identification of a class of Hammerstein system,Wiener system and Hammerstein-wiener system driven by non-uniform sample data with known basis functions,the chaotic natural selection particle swarm algorithm proposed in this paper is used to identify the system.The proposed algorithm is further extended to the identification of non-uniformly sampled data-driven Hammerstein systems and Wiener systems with dead band characteristics,and simulation examples are used to verify the effectiveness of the proposed algorithm for non-uniformly sampled data-driven nonlinear systems.(3)Discuss the identification of arc furnace electrode adjustment system.This article introduces the arc furnace steel making process equipment,and further studies the core part of the arc furnace adjustment system: the arc furnace electrode adjustment system.First,the electrode adjustment of the electric arc furnace is modeled.The system can be represented as a Hammerstein-wiener type non-linear block structure composed of a series of dead-band nonlinear links,hydraulic third-order linear links,and arc nonlinear links.The proposed chaotic natural selection particle swarm algorithm identifies the system.