ZINC 2020 - Particle Swarm Optimization - Model Predictive Control for Microgrid Energy Management

Particle Swarm Optimization - Model Predictive Control for Microgrid Energy Management

Quyen Van Ngo (ETS, Canada); Kamal Al-Haddad (Ecole de technologie supérieure, Canada); Kim Khoa Nguyen (University of Quebec, Canada)

Microgrid is becoming the most attractive solution for integrating distributed renewable sources into the utility grid. Such a system combines renewable generations with conventional distributed generations, storage systems, and loads in one entity operating in both isolated and grid-connected modes. However, it also associates with a high level of uncertainty and volatility following climatic conditions. Therefore, energy management strategies in operating MGs plays a crucial role in term of economic and reliability. This paper investigates a method applying constrained multi-swarm particle swarm optimization without velocity-based model predictive control to optimize the operation cost in small scale PV-MGs. The results are compared with the linear programming algorithm. The results show the effective modified particle swarm optimization embedded in the model predictive control algorithm was perform well. The simulations are run over 24 hours ahead based on the forecast data of PV generation, load demands, and energy price.