Coordinated integration of distributed energy resources in unit commitment

Distributed energy resources, such as electric vehicles and roof-top solar panels have become increasingly popular and essential components of power distribution systems. However, their growth poses significant challenges to system operation, such as bidirectional power flow, intermittent power generation, increased peak demand, and unexpected frequency/voltage fluctuation. To tackle these challenges, we developed a vehicle-to -grid model, which incorporated dynamic electric vehicles usages, such as driving times, driving distance, and charging/discharging locations. The machine learning method of principal component analysis and XGBoost were used to develop the solar energy prediction model. We also developed a novel unit commitment model to coordinate and aggregate the distributed energy resources in power generation economic dispatch. Electric vehicle charging was used as elastic demand and electric vehicle discharging was used as power generation sources. The solar energy was considered as a negative load. Simulation results showed that uncontrolled electric vehicle charging and solar energy could negatively impact the power systems. For example, the load ramping rate was increased by 384%. Simulation results also showed that the proposed models could mitigate the negative impact and improve energy efficiency. For example, the peak demand, load ramping rate and average generation cost were reduced by 21%, 79%, and 27%.