Research On The Development Mode And Operation Strategy Of Virtual Power Plants Under The New Power System

To build a new type of power system with new energy as the main body,it is urgent to enhance the flexibility of the power system,improve the ability to optimize the allocation of resources and fully consume new energy.At the same time,the decisive role of the market in resource allocation is increasingly prominent,and the power market system has become an important support for the new power system.Power-side and grid-side regulation resources can hardly meet the system’s safe,green and economic operation needs,and it is urgent to guide demand-side resources to participate in the power market to play multiple values,so the virtual power plant that can aggregate and dispatch demand-side resources will become an important way for the new power system to achieve interaction and intelligence on the energy supply and demand side,as well as decentralized resources to participate in the power market.However,internally,virtual power plants are not finely managed according to system demand and resource differentiation,which inhibits the multiple values of resources;externally,the new energy consumption of virtual power plants is mainly through participation in peak-valley regulation to indirectly alleviate wind and light abandonment,and their operation strategies are not coupled with flexible regulation and active new energy consumption,so the multiple values of virtual power plants are not fully exploited.In order to solve the above problems,this paper carries out a study on the development mode and operation strategy of virtual power plants under the new power system.Firstly,by sorting out the operation mechanism and market positioning of virtual power plants,this paper clarifies the trading varieties that virtual power plants can participate in,summarizes the relevant access mechanisms and typical power market trading mechanisms,and then constructs a framework for virtual power plants to participate in intraday power and standby auxiliary service trading;secondly,from"flexible-economic-green Secondly,we extracted the resource features from the three values of "flexible-economic-green",solved the data imbalance problem by adaptive synthetic sampling and self-coding clustering algorithm,reduced the dimensionality of features by using principal component analysis,established a load resource grading model based on improved SVM,and proposed a virtual power plant resource grading and calling model;finally,based on improved long and short-term memory algorithm and demand elasticity theory,we predicted the virtual power plant standby Finally,based on the improved short-term and long-term memory algorithm and demand elasticity theory to predict the standby capacity of virtual power plants,the new energy consumption responsibility weighting mechanism is used to design an assessment strategy for virtual power plants to participate in new energy consumption,and then an operation optimization model is established to couple flexibility,greenness and economy in the context of joint trading of virtual power plants in multiple markets.The improvement of operating revenue under the basic scenario proves the superiority of the proposed model in the economic operation of the new power system,while the reduction of abandoned scenery and regulation deviation under various scenarios proves that the improved strategy is more conducive to the performance of multiple values.This paper innovatively proposes a resource classification model for virtual power plants that takes into account the multiple values of resources and market transaction needs,which solves the problem that the current distributed resources can participate in unclear market transactions and the multiple values cannot be given full play;the paper improves the SVM classification algorithm,solves the data imbalance problem,reduces the feature dimension,improves the accuracy of small sample classification,and provides a feasible way and visual analysis for virtual power plant resource classification.The paper improves the LSTM prediction algorithm to improve the accuracy of small-sample prediction and provide reference for the determination of standby capacity of virtual power plants;the paper establishes a virtual power plant operation optimization model to promote new energy consumption and proposes operation strategies under different scenarios,which solves the current problem of insufficient coupling of multiple values of virtual power plants and helps to bring the multiple values of virtual power plants into full play.