Research On Neural Network Robust Adaptive Control Method Of Power System With FACTS

In recent years,with the rapid development of the power industry and the growing demand for power and energy,the stability of power systems has become increasingly prominent.Power system stability includes power angle stability,voltage stability,and frequency stability.For these stability problems,generator excitation control has proven to be an effective control method.However,due to the adjustment of the power industry structure,the scale and complexity of the power grid are increasing,and the generator excitation control often cannot meet the stability requirements of the power system.Flexible AC Transmission System(FACTS),as an efficient and practical power electronic device,provides a new adjustment method for the stable operation of the power system under the premise of ensuring economic benefits.Considering generator excitation and FACTS devices at the same time,the power system can be made more flexible and controllable.Based on the improved backstepping method and neural network adaptive control,critic adaptive design and adaptive dynamic programming theory,this paper studies the transient stability problem of the power system with thyristor controlled series compensation(TCSC),static var compensator(SVC)and static synchronous compensator(STATCOM)respectively.The main research results are as follows:(1)For single machine infinite power system with TCSC,which is affected by system model uncertainty,nonlinear time-delay and external unknown disturbance,we present a robust adaptive backstepping control scheme based on radial basis function neural network(RBFNN).The RBFNN is introduced to approximate the complex nonlinear function involving uncertainty and external unknown disturbance,and meanwhile a new robust term is constructed to further estimate the system residual error,which removes the requirement of knowing the upper bound of the disturbance and uncertainty term.The stability analysis of the power system is presented based on Lyapunov function,which can guarantee the uniform ultimate boundedness(UUB)of all parameters and states of the whole closed-loop system.(2)An adaptive critic design(ACD)-based robust neural network backstepping control method is presented for the single machine infinite power system containing SVC.First,a robust neural network backstepping control is presented by using ACD approach for the general n-order uncertain nonlinear systems.After that we further develop an optimal robust controller for the SVC system containing model uncertainty and external unknown disturbance.The neural networks(NNs)are used to approximate the unavoidable complex nonlinear items,and meanwhile a robust term is proposed to restrain the disturbance and the corresponding residual error.In particular,through constructing the new primary critic signal and the new form of Lyapunov function,combining with adaptive boundary technology,there is no need for prior knowledge of the disturbance term boundary,the ideal weight boundary and the reconstruction error boundary of the NNs.(3)Aiming at the double machine power system with STATCOM,an adaptive backstepping control scheme based on adaptive dynamic programming(ADP)is proposed under the influence of model uncertainty and external unknown disturbance.The controller is divided into a feedforward controller and a feedback optimal controller for separate design,and a critic neural network is introduced to approximate the cost function.The stability analysis of the double machine power system with STATCOM is carried out by using the Lyapunov function,which ensures the uniform ultimate boundedness of the state quantity of the closed-loop system.