A virtual power plant optimal scheduling method based on a master-slave game strategy

Abstract

The invention discloses a virtual power plant optimization scheduling method based on a master-slave game strategy. The method comprises the following steps: 1, analyzing and constructing a virtual power plant model containing multiple load types; 2, establishing a virtual power plant transaction model based on an existing power market transaction mechanism; 3, establishing a risk cost model considering uncertain factors; 4, carrying out optimal scheduling on the virtual power plant by adopting a master-slave game strategy; And 5, solving the optimized scheduling model through a reinforcementlearning algorithm. According to the method, the volatility of distributed energy output in the virtual power plant and the uncertainty of load prediction are considered, an economic dispatching modelof the virtual power plant in the power market is constructed, and a master-slave game strategy and a reinforcement learning algorithm are adopted for optimization solution, so that the purposes of improving the production benefits of the virtual power plant and reducing the load power purchase cost are achieved.

Description

technical field [0001] The invention relates to the field of optimized operation of distributed power sources, and more specifically, the invention relates to an optimal scheduling method of a virtual power plant based on a master-slave game strategy. Background technique [0002] In recent years, clean energy sources such as photovoltaic power plants and wind power plants have developed rapidly in China. Due to the uncertainty of distributed energy output, the connection of large-scale distributed energy to the distribution network will aggravate the peak-to-valley load difference of the power grid, and its safe operation caused a certain impact. On the other hand, with the rapid growth of electric vehicles, the load type of the power grid is no longer a single traditional load, but also a controllable load with temporal and spatial uncertainties. As a new type of energy management system, virtual power plant technology is more and more applied to the coordinated managemen...

AI Technical Summary

Problems solved by technology

[0003] There are many kinds of optimal dispatching methods for virtual power plants: stochastic programming method considering distributed energy output uncertainty and backup tuning method, which have achieved good results in dealing with the randomness and volatility of distributed energy output , but in order to ensure the reliability of power supply, blindly increasing energy reserves has caused unnecessary waste; the two-stage optimal scheduling method is mainly to collect the internal power generation and sales information of the virtual power plant and external market incentive policies, and then build an economic scheduling model The disadvantage of this method is that it needs accurate probability distribution information of large-scale uncertain factors as the premise, and the accuracy of the optimized scheduling method is lacking.

Images