The Research Of Adaptive Neural Network And Its Application In Wastewater Treatment

The external disturbances and time-delays are widely existed in real engineering,such as the control of the dissolved oxygen concentration in the process of sewage treatment.The existence of external disturbances and time-delays is important factors that lead to system instability.Therefore,we should pay attention to such factors in actual control systems.The paper deeply studies the control for nonlinear systems with time-delay and external disturbances.To solve the defects of the single control method,we combine adaptive control with neural network control method.This paper research content is as follows:Firstly,through simulation experiments for a simple example,we verify the approximation performance of the RBF neural network,while comparing the effects of various parameters for function approximation performance,at the last we compared the approximation performance between gradient descent algorithm and the new hybrid algorithm,we can achieve that the training speed of the hybrid algorithm is faster than gradient descent algorithm.Secondly,for a class of strict-feedback nonlinear systems with time-delays,we proposed a control scheme of adaptive neural network through combining Lyapunov function and backstepping algorithm.In the control scheme,the RBF neural network is used to approximate the unknow nonlinear function,then we design a adaptive neural network controller based on the backstepping algorithm.The simulation experiments demonstrate the feasibility and effectiveness of the controller through the application of MATLAB and by comparing with the simulation result of the original documents,we can obtain that the design method of the controller in this paper improves the tracking speed and accuracy of the system on basis of guarantee the closed-loop control system stability.Then,we proposed a control scheme of adaptive neural network for a class of nonlinear system with input delay and external disturbances,A filter and a virtual observer are constructed to substitute the system state and show the effectiveness of the controller through constructing Lyapunov function.The simulation experiments demonstrate the feasibility and effectiveness of the tracking performance of the system,ensure that the system can reach a steady state in a limited time,it is mean that the design of the adaptive neural network algorithm is reasonable and effective.And by comparing with thesimulation result of previous controlled system,we can achieve that the control algorithm in this paper can work well in reducing the impact of external disturbances.Finally,the research results for the nonlinear system with time-delay and external disturbances is applied to the the experiment platform of wastewater treatment.Compared the results of PID control algorithm with the control algorithm in the paper,we can achieve that the proposed control algorithm in this paper is better than PID control algorithm for the control of the dissolved oxygen concentration in the process of sewage treatment.