Research On Dynamic Stochastic Optimal Power Flow

In the power system,factors such as load fluctuations and grid faults have certain degree of uncertainty in the scheduling and operation of the power system.However,the determination of the source end can make the above uncertainty negligible.With the continuous change of the power grid operating environment,especially the large-scale wind farm access system,its random fluctuation makes the randomness of node injection power appear more and more obvious,which brings new challenges to the traditional power system power flow analysis.Stochastic optimal power flow is a powerful tool to solve the impact of various uncertainties on the power system.However,the actual power system as a complex dynamic system,the traditional stochastic optimal power flow only considers the optimal target of the power system under a single time section,neglects the mutual influence of each time period,and its calculation result has more limitations in production..In addition,the existing stochastic dynamic optimization models have various forms,and there is not yet a unified description for the application of power system analysis,which is not conducive to the in-depth development of the technology branch.Therefore,this paper introduces the sequential dynamic stochastic optimization model,and generalizes it to the dynamic stochastic optimal power flow calculation of power system.In the optimization process,the randomness of wind power and load is comprehensively considered,and the heuristic adjustment algorithm based on cost strategy is used to solve the opportunity-based approach.A constrained dynamic stochastic optimal power flow model was used to verify the practical validity of the model method through numerical examples.The details are as follows:First,construct a probability density function for predicting wind power;first obtain the short-term forecasting wind power output data through wind generator output model,and then establish a short-term wind power prediction probability density function model based on non-parametric nuclear density estimation;construct a description of non-parametric nuclear density;A multi-objective bandwidth optimization model for estimating accuracy and smoothness,applying the multi-objective Pareto optimization method,solving a sub-target Pareto optimal solution representing the smoothness and accuracy of the nuclear density estimation,and then obtaining the optimal bandwidth and relating the optimization result to The comparison of the bandwidth results obtained by the rule of thumb shows the validity of the method and the conclusion that the method is more applicable than the empirical method.The desired optimal bandwidth value is brought back to the previously established model,and finally the predicted wind power probability density function is obtained.This method overcomes the defect of the presupposition of probability density in the parameter estimation and is suitable for wind power prediction in different regions.It has universal applicability.Secondly,four kinds of solving methods for opportunity constrained stochastic optimal power flow are presented.The stochastic optimal power flow genetic algorithm with opportunistic constraints,particle swarm optimization,backward mapping method and heuristic adjustment algorithm are systematically analyzed.Simulation calculations were performed on a 5-node system with load stochasticity,and the solutions to stochastic optimal power flow using the three algorithms of genetic algorithm,particle swarm optimization,and heuristic adjustment algorithm were compared with solutions of deterministic optimal power flow.The results showed that: The heuristic adjustment algorithm has better adjustment margins,is easier to implement,and is more suitable for practical engineering simulation calculations,providing a theoretical basis for choosing the solution strategy below.Finally,the existing stochastic optimal power flow models are mostly static.They do not consider the insufficiency of the links between various time profiles and the diversity of dynamic stochastic optimization models.They introduce a sequential dynamic stochastic optimization framework and construct a stochastic dynamic based on sequences.Optimized power system dynamic stochastic optimal power flow model;comprehensive consideration of the randomness of wind power and load,using heuristic adjustment algorithm based on cost strategy to solve,and simulation calculation of improved 5-node system and IEEE-14 system example The analysis shows that the results of dynamic stochastic optimal power flow can better adapt to the forecasting error and fluctuation of load and wind power.The proper access of wind farm can effectively reduce the cost of conventional generator power generation and has a good engineering application prospect.