Research On Reactive Power Optimization Of Power System With Random Wind Power Generation

In recent years,due to its environmental protection and high reliability,wind power has been greatly developed.The wind power output has the characteristics of randomness.Therefore,wind power grid integration will change the distribution of power flow and bring a lot of uncertainties to the operation of power system.In addition,the fluctuation of load power is a common phenomenon in the actual power system.The traditional research on reactive power optimization is mostly based on the determined system state,without considering the influence of uncertainties such as wind power output and load,which will bring hidden danger to the safe and stable operation of the power system.Therefore,aiming at the uncertain factors in the system,the probabilistic power flow method is introduced to establish a reasonable reactive power optimization model,and the corresponding solving strategy is proposed to establish reactive power allocation scheme suitable for uncertain environment.This paper studies three aspects of optimization algorithm,uncertainty processing and reactive power optimization model.First of all,in this paper,flower pollination algorithm is applied to solving reactive power optimization problem.In order to cope with the local optimum and low optimization precision problems of flower pollination algorithm,a modified flower pollination algorithm is proposed,to which the Sobol sequence sampling technique and adaptive adjustment of switching probability are applied.The effectiveness and accuracy of the improved algorithm on reactive power optimization are verified by an example analysis.Secondly,in order to effectively deal with the impact of uncertainties of wind power output and load on the reactive power optimization of power system,this paper proposes a probabilistic load flow method of power system containing wind farms,and uses this methods to replace the deterministic power flow calculation.In this method,Sobol sequence sampling and Nataf transformation are used to generate wind power output samples with correlation,and the cumulant method based on Cholesky decomposition is used to calculate probabilistic load flow at each sample point.The final probability distributions are obtained by integrating all distributions obtained according to the total probability formula.The example analysis shows that the proposed method has high computational precision unde r the small sampling size,and can obtain the probability distribution of the system state variables accurately.Finally,a reactive power optimization model with the chance constraint of state variables is established by taking the minimum expected value of the active power loss as the objective function.The expected value of the active power loss and the probability distribution of the state variable are solved by the probabilistic load flow,and the modified flower pollination algorithm is used to optim ize the model.The simulation test shows that the reactive power optimization method in this paper can effectively reduce the power loss and the cross-border risk of system state variables in the uncertain environment,thus ensure the safe and economical operation of the system.