Multi-objective Optimal Allocation Of Energy Storage Capacity Based On Wind Power Prediction

At present,wind energy has become one of the world's most important new energy sources as a renewable energy source.Vigorously developing wind power is of great significance for improving climate change and energy scarcity.However,the random variation of wind energy causes the wind turbine output to be difficult to predict and control.In order to facilitate the understanding of wind power power trends and facilitate the formulation of scheduling tasks,this paper takes wind power prediction as one of the research contents.By analyzing the influencing factors of wind speed and wind power variation,the correlation analysis method is used to reduce the dimensionality,the computational complexity of the wind speed prediction model is reduced,and the K-means algorithm and contour coefficients are introduced to cluster the time series data sets.Aiming at the shortcomings of slow convergence and parameter redundancy of neural network wind power prediction model,a wind power prediction model based on multi-scale convolutional neural network is proposed.In order to reduce the model error,it is proposed to add the error as a predictive value to the loss function to further improve the accuracy of the wind power prediction model.On the other hand,the volatility of wind energy will cause the integration of wind power plants to threaten the reliability,safety and economic operation of the power system.Therefore,the introduction of energy storage equipment to stabilize the volatility of wind power generation and achieve peak filling Valley,backup power supply,etc.In order to minimize the economics of the system,the output power volatility and the abandonment rate establish a multi-objective optimal configuration model and constraints for the energy storage capacity.In this paper,Non-dominated sorted genetic algorithm(NSGA-?)is adopted as the model solving method.Aiming at the problem of squeezing distance failure when NSGA-? solves the multi-objective optimization configuration model of energy storage capacity,an improved crowding distance improvement algorithm is proposed to improve the diversity and uniformity of Pareto solution set and utilize explosion.The operator replaces the mutation operator for local search.Finally,in order to reduce the difficulty for decision makers to find a solution in Pareto solution set,fuzzy theory is introduced to achieve unbiased optimal strategy to obtain unbiased optimal solution.