Research On Decision-making Model And Operation Optimization Of Multi-micro Energy Grid Joint System Based On Distribution Networ

In order to fully utilize the existing energy network and reduce or delay the construction of new power grid structures,while ensuring the safety and economy of the distribution network,achieve full consumption of clean energy,enrich user energy types,and improve the operational flexibility of the distribution network.At the same time,the uncertainty of distributed energy and the fluctuation of clean energy consumption rate should be addressed,and collaborative control should be implemented for various types of energy,power grid,user side resources,and energy storage,This article conducts research on the decision-making model and operational optimization of a multi micro energy grid joint system based on distribution networks.This article innovatively studies the optimization operation of the joint system of multiple micro energy grids under the distribution network regulation,better playing the supporting role of the distribution network energy transmission network,coordinating the complementary operation of multiple energy sources,reducing carbon emissions,and improving economic efficiency.Secondly,the integration of multiple heterogeneous energy sources such as electricity,gas,heat,and cold into unified optimization scheduling has expanded the existing comprehensive energy optimization model from multiple aspects such as equipment dimensions and operating modes,and expanded the optimization space of the energy system considering the distribution network,meeting the diverse energy needs of users.At the same time,the combination of deep learning technology and adaptive dynamic programming has improved the real-time performance of the algorithm.Through the Frank Wolfe multi-agent distributed online learning algorithm,the collaborative optimization operation capability of the distribution network and the micro energy grid has been strengthened.Through simulation based on actual grid operation data,in the scenario where wind power and photovoltaic installations account for 50%respectively,the renewable energy consumption capacity has increased from 75%before optimization to 100%,The island operation capacity after branch failure has been increased from 1.3 before optimization to 5.3;In the scenario where wind power accounts for 25% of installed capacity and photovoltaic accounts for 75%,the total network loss of the distribution network on that day decreased from 104.02 MW before optimization to 6.726MW;In the scenario where wind power accounts for 75%of installed capacity and photovoltaic accounts for 25%,the qualification rate of node voltage has increased from 45% before optimization to 82%.In all scenarios,the operational flexibility of the distribution network can achieve good improvement effects.At the same time,through the collaborative optimization operation of the distribution network and multiple micro energy grids,a virtual energy storage capacity of 4.40 MW was provided for the system,meeting the 1.99 MW electricity storage configuration requirements required by the system,saving an investment of0.31362 million yuan.The main research content and achievements of this article are as follows:(1)The flexibility of connecting to the distribution network and regulating resources endow it with schedulability at the distribution network level,while the development of integrated energy systems endows it with the possibility of energy coupling and regional autonomy.This makes the original definition of distribution network resilience unable to fully cover the operational attributes of the distribution network under the new power system.Based on this,this article proposes three types of micro energy grid operation systems comprehensively based on equipment operation characteristics and modes,and it is the foundation of the optimization modeling of micro energy grids.At the same time,the concept and definition of grid resilience were systematically analyzed,the existing distribution network resilience characteristics and evaluation indicators were analyzed,and a new evaluation system for distribution network operation resilience under the new power system was established.(2)The distribution network has the ability to flexibly adjust due to the integration of micro energy grids and their energy conversion,transfer,and storage devices.In order to achieve collaborative control of micro energy grids,it is first necessary to achieve online optimization scheduling of a single micro energy grid.This paper constructs a thermal-electric coupling micro energy grid optimization model and an electric-gas-thermal fully coupling micro energy grid optimization model based on the energy coupling relationship.Secondly,the constrained nonlinear optimization method is combined with deep learning to construct a deep learning strategy for the optimal operating behavior of the micro energy grid;At the same time,the implementation dependency heuristic dynamic programming method was improved,and the implementation network was replaced by a pre-trained deep neural network.The online optimization algorithm and strategy of the micro energy grid were constructed.The correctness of the model and the rapidity of the algorithm were verified through simulation,which ensured the real-time optimal scheduling of various equipment in the micro energy grid,and realized the full consumption of clean energy at the micro energy grid level.(3)In order to achieve online optimization and scheduling of the distribution network-micro energy grid collaborative system,this article first presents a micro energy grid construction plan that combines regional characteristics analysis.Based on the operational needs of distribution network in typical regions,two types of micro energy grid optimization models are constructed,which are suitable for load demand centers and renewable energy centers.A bi-level optimization model of distribution network-multi micro energy grids was constructed,taking into account the cost of auxiliary services provided by the distribution network to the micro energy grid,reflecting the responsibilities and interests of various entities in the actual production environment.Secondly,fully considering the economy and security of the distribution network,the optimal power flow algorithm based on second-order cone programming and the adaptive dynamic optimization method based on pre training were combined to establish an online optimization scheduling strategy for the system of distribution network-multi micro energy grids.The proposed model and optimization strategy were verified and analyzed in a testing system based on the actual network parameters of Guizhou power grid,and branch overload faced by the operation of distribution networks in different regions were eliminated,achieved full consumption of wind and photovoltaic power,and ensured the economic efficiency of the collaborative system.(4)With the large-scale integration of low-voltage distributed photovoltaics,wind power,and electric vehicles,power fluctuations in daily distribution grid operation have become increasingly severe.The distribution network needs to actively optimize the micro energy grid to ensure safety and achieve better economy,and use the distribution network operation resilience evaluation index to objectively evaluate the operation status of collaborative systems.This article first further expands the optimization model for the operation of micro energy grid,introducing an ice storage air conditioning model to achieve the integration of full energy micro energy networks.By utilizing the spatiotemporal complementary characteristics of multiple energy systems,the optimization operation space of the distribution network is further expanded.Secondly,in the optimization of the distribution network layer,a multi-agent distributed online optimization algorithm based on the Frank Wolfe algorithm is combined to achieve online optimization scheduling of the distribution network to the micro energy grids,and to adjust the system operation situation in real time.Finally,a simulation case based on actual system construction was used to verify the operational economy of the joint system,and the effectiveness of improving the operational resilience of the distribution network at this time was evaluated,ensuring the real-time demand of users for multiple types of energy.At the same time,the ability of the distribution network to dynamically respond to the uncertainty of both source and load sides and extreme situations of the power grid was improved.