Research On Reactive Power Optimization Of Power System With Multi-wind Farm Based On Gaussian Mixture Model Of Uncertainty

In recent years,the large-scale integration of wind power into the power system not only improves the cleanliness and sustainability of energy but also brings strong randomness to the power system.Correctly modelling the randomness of wind farms and doing a good job in reactive power regulation and control of power systems connected to random wind farms are the key areas for researchers to deal with the development of new wind power and efficient grid connection.To solve this problem,this paper proposes a multi-wind farm modelling method based on an improved Gaussian mixture model,and based on an innovative two-layer probabilistic reactive power optimization algorithm,a power system reactive power regulation connected to multi-wind farms is proposed.Strategy.The specific work of this paper is mainly divided into the following parts:（1）Based on the output data of multiple wind farms,the improved Gaussian mixture model is used to accurately describe the correlation of multiple wind farms.This paper first summarizes the relationship between wind speed and output power of wind farms based on time series and proposes the probability density screening process before the parameters of the Gaussian mixture model are solved.The truncated Gaussian mixture model corresponds to the domain.In addition,this paper also proposes a parameter solution method based on the K-means clustering analysis.The example analysis proves that when describing the output of multiple wind farms,the improved Gaussian mixture model has a higher fitting accuracy than the traditional algorithm,and the parameter solution method reduces the complexity of the Gaussian mixture model parameter calculation.（2）Two sampling methods for probabilistic power flow calculation are proposed in this paper.The correlation coefficient matrix that is difficult to calculate in Nataf transform is obtained through the dichotomy and polynomial fitting algorithm,and finally,the wind farm power scene analysis method of point estimation method based on Nataf transformation and the Hammersley sequence sampling of wind farm output based on Nataf transformation are formed.The example analysis proves that,compared with the Monte Carlo sampling method,the Hammersley sequence sampling method based on the Nataf transform improves the sampling accuracy.Compared with other scene analysis methods,the point estimation method based on the Nataf transform has obvious accuracy advantages and high practical application value.（3）A new reactive power optimization algorithm is proposed.In this paper,a multiobjective power system optimization model is established,and the objective function also considers the operation economy and safety of the random multi-wind farm power system.The spider monkey algorithm and the second-order cone optimization algorithm are combined to form a two-layer optimization model.The upper layer adopts the spider monkey algorithm to consider the operation safety,and the lower layer calculates the economy of the power system based on the second-order cone optimization algorithm.Finally,an example analysis is carried out based on the improvement IEEE33 node model,and the optimization strategy verifies the application value of the two-layer reactive power optimization algorithm from two aspects of the operation safety and economic.