Study On The Control Of Statcom Based On BP Neural Network

Along with the increasing of modern power system's scale and complexity ,Reactive power'distribution in electric network becoming more and more complex.The requirement to reactive power compensation is more and more strict .More advanced reactive load compensation equipments and modes are need.STATCOM is a new kind of reactive-load compensation equipment with the kemel of converter ,which is developed in 70s late 20 centuries .It has superiority in fast reaction, high pricision, and the character to realize multifunction , multiobject control.So it has get comprehensively attention in the field of power system .Recently,several control modes like optimum control, adaptive control , differential geometry control and etc are available for the STATCOM control ,but each has its shortcoming.This paper based on the deduced dynamic mathematic model,bring forward a kind of control mode that based on back propagation neural network.Based on the coordinate transformation theory, a set of state equations for STATCOM is derived, and it is linearized at the point of δ₀, According to the optimal control theory, a expression which facilitates to exert a control over AS by use of ∫Δu'dt,Δi_q', kΔu_dc' is found. Different K which corresponding to different i_q0' and δ₀ are educed by riccati equation,and apply it to neural network simulation.In order to decrease system harmonic components, the means of creating PWM wave based on harmonic suppression principle is introduced for the STATCOM drive-pulse: attach fillister to certain angle of the square wave to depress useless harmonic and control fundamental component magnitude, then use BP neural network to calculate each angle of the fillister.In this paper, the SIMULINK simulation software is used to study various compensation effects of STATCOM under the control of BP neural network. It is...