Research On Wind Power Grid-Connected Control Strategy Based On Deep Learning Parameter Optimization

This topic to asynchronous wind turbine wind power grid voltage fluctuation on the topic research background,analyzed the Common connection Point(PCC,the Point of Common Coupling)disturbance voltage stability and wind speed,and the relationship between the electrical energy long-distance transmission is put forward based on the optimization of particle swarm optimization algorithm to improve the depth belief network control strategy of Static reactive power Compensator(STATCOM,Static Synchronous Compensator)voltage fluctuation suppression system to suppress the wind field of the PCC voltage fluctuation and flicker,and impact ofThe optimal design of double closed loop control system is studied.The main research content of this paper is as follows:In order to explore the suppression principle of voltage fluctuation in parallel grid of wind farm,the influence relationship between wind speed disturbance and electric power long-distance transmission on voltage fluctuation in parallel grid of wind farm is systematically analyzed and studied,and the problem of voltage fluctuation in parallel grid of wind power system is solved by using STATCOM reactive power compensation device.The voltage regulation principle of STATCOM is further analyzed and studied,and the dynamic mathematical model based on instantaneous reactive power theory is established according to its topology structure and circuit characteristics,and the principle analysis of double closed-loop control strategy is obtained,which provides theoretical and algorithm basis for the improvement of control strategy.In order to improve the limitations of the traditional control system proposed above,the depth confidence network algorithm based on the restricted boltzmann machine and the hybrid multi-objective particle swarm optimization algorithm based on Pareto optimal are studied systematically,and then the particle swarm optimization algorithm is proposed to improve the depth confidence network algorithm.The superiority of the improved method is further verified by an example of wind velocity calculation in the wind field.Under the same condition,the proposed algorithm can reduce the prediction error by 2% compared with the traditional method.In order to improve the response speed of STATCOM and further improve state system performance,the improved depth confidence network algorithm of the optimized particle swarm optimization proposed above is combined with the specific control strategy of STATCOM,hoping to ensure that the controller can obtain relatively ideal control effect under different working conditions.Based on the stability and dynamic characteristics of inner and outer loop of double closed loop system,the PI parameter design of voltage and current loop and its deep learning network control optimization strategy are given.When the proposed control strategy is in local optimal condition,the response time meets the current control requirements of STATCOM that the response time is less than 10 ms.In order to verify the feasibility of the proposed scheme,the STATCOM voltage fluctuation suppression simulation platform is designed and improved by using the existing hardware environment in the laboratory.Wind farm conditions are simulated by switching reactors,starting and stopping induction motors and controlling sudden changes in grid voltage.Through the analysis of the experimental waveform,the rationality and superiority of the new STATCOM control strategy designed in this paper in terms of steady-state voltage accuracy and dynamic response speed were verified.The response time of the experimental system was reduced to less than 10 ms and the average response time was less than 12 ms.