Design and Implementation of Virtual Power Plant System Based on Equipment-Level Power and Load Forecasting

Abstract

In order to ensure the effective participation of virtual power plants in grid interaction under the novel power system. This paper design and implement a virtual power plant system based on equipment-level power forecasting and load forecasting technology. In order to shield equipment differences, a unified measurement standard for physical equipment of different energy systems is established by constructing virtual machine groups (including equipment attribute information, equipment collection information and equipment evaluation information); using equipment-level neural network power forecasting and multiple load forecasting engines to ensure equipment-level forecast accuracy and flexible aggregation capabilities; building model self-learning capabilities by monitoring equipment historical data to drive model iterative optimization; the cloud edge collaboration solution is used to realize the vertical aggregation of resources, reduce the cloud load and improve the command processing speed. Through deployment and verification, this system can improve prediction accuracy and dynamically aggregate multiple types of resources to participate in grid interaction.