Research On The Construction And Operation Strategy Of Virtual Power Plant With Regional Cooling System For Peak Shaving Demand Of Power Grid

In recent years,increasing urban power loads and a large number of new energy sources have been incorporated into the power grid,and the peaking demand of the power grid has also increased.Since the district cooling system itself has many controllable variables,it has gradually become a typical flexible load on the demand side.In addition,virtual power plant provides an effective way for distributed resources to participate in power grid operation,which is a research hotspot in the field of power system in recent years.According to the current research results,it can be concluded that most scholars focus their research on the flexible load inside the virtual power plant on resources such as central air conditioning.However,compared with the former,the regional cooling system has greater adjustment potential due to its large scale and more controllable variables.When most scholars build virtual power plants,they mainly consider the regulatory characteristics of the resources themselves and the geographical location of the resources,but the risk economic problems they face during operation cannot be ignored.In this thesis,aiming at the problems of low resource utilization rate of regional cooling system and increasing peak shaving demand of power grid,an improved clustering method based on parameter adaptive density peak is proposed to aggregate the internal resources of virtual power plant.Based on the risk conditional value theory,a dynamic capacity model of virtual power plant considering operation uncertainty is constructed,and the improved particle swarm optimization algorithm is used to decompose the peak shaving task to realize the purpose of regional cooling system participating in peak shaving of power grid.Firstly,a modeling method of virtual energy storage model for district cooling system is proposed,which is similar to conventional battery energy storage.The energy consumption model of chiller,chilled water pump and terminal fan of district cooling system and the thermal effect model of indoor users are established.The energy storage characteristics are analyzed.Compared with the ordinary battery energy storage model,the proposed virtual energy storage model modeling method is used to encapsulate the district cooling system into a model with similar expression to the ordinary battery model,which provides a theoretical basis for the virtual power plant to aggregate and call the district cooling system.Secondly,a virtual power plant including a regional cooling system is constructed.Based on the improved clustering method of parameter adaptive density peak,the decentralized district cooling system is aggregated to construct callable energy storage components.Considering the factors such as outdoor temperature changes and users ’ willingness to regulate and control,a dynamic capacity model of virtual power plant is established based on risk identification theory.According to different risk preferences,the dynamic adjustment of capacity parameters of virtual power plant is realized to avoid operational risks and meet operational economy.Complete the construction of virtual power plant.The results show that the improved clustering algorithm is trained by the data set,and the internal resources of the virtual power plant are aggregated accordingly.According to different risk strategies,the capacity of the virtual power plant can dynamically follow the operation risk.Finally,decompose the peak shaving task.The output potential of the virtual energy storage components of the district cooling system is evaluated and sorted from the perspectives of response capacity,duration and response time.Based on the improved particle swarm optimization algorithm,the peak shaving task of virtual power plant is decomposed with the objective function of minimizing output deviation.The operation strategy of virtual power plant participating in power grid peak regulation is formulated.The results show that under the premise of completing the prefabricated cooling,the improved particle swarm optimization algorithm can more accurately decompose the peak shaving task and accurately transform it into the parameters of each part of the regional cooling system,so as to participate in the peak shaving task of the power grid.