Energy Management Research Of Photovoltaic Microgrid Based On Improved Particle Swarm Algorithm

As the main form of renewable energy resources network, microgrid can provide convenience for the access to power grid of renewable energy resources. As the output energy of distributed energy resources is intermittent, the massive distributed energy resources accession to power grid will affect the stability of the power system. As an autonomous system, microgrid can realize the maximum utilization of renewable energy resources by energy management of the system to ensure the power supply reliability of the load and the economical operation of the microgrid.This paper focus on the photovoltaic power plant monitoring s ystem, photovoltaic power generation forecasting and microgrid optimization energy distribution of photovoltaic microgrid. System monitoring of the photovoltaic power plant is the basis of energy management of microgrid. As for the 18 KW photovoltaic system which had been established in photovoltaic labouratory, hardware design is done based on RS485 bus and software design is based on OPC technology and modbus communication protocol. Experiments show that the data acquisition rate of the monitoring system can be up to second grade. The monitoring system basically achieves the function of real-time monitoring of the photovoltaic power plant and accumulating raw data for photovoltaic power generation forecast.As the output energy of photovoltaic power plant is intermittent, the historical power production data of photovoltaic power plant and meteorological data are collected by the monitoring system for photovoltaic power generation forecast. A photovoltaic power generation forecasting model is established based on general regression neural network after analyzing the influence factors of photovoltaic power generation. The proposed model has been proved to be effective by studying a specified case.Energy management model of distributed generations and the energy storage unit in the DC microgrid are established based on the previous research, and three kinds of energy management strategy in different operation modes is proposed. This paper has proposed a microgrid energy management model based on improved particle swarm optimization. The improved algorithm focus on finding feasible solutions of problem as initial position of particle swarm optimization, then use the pattern search to give better solutions. Application of the proposed energy management model on a small DC microgrid system shows the validity of the method. The proposed energy management model can optimize the energy flow in the microgrid to ensure the power supply reliability of the load and the economical operation of the microgrid.