Cost-effective energy management of an islanded microgrid

Abstract

The interest in the integration of distributed energy resources in microgrids increased significantly in the last decade. The stochastic nature of some energy sources and the dynamic power demand has brought difficulties regarding the optimal control of microgrids. This current study addresses the energy management challenge in an islanded hybrid energy microgrid that includes three types of renewable energy resources (photovoltaic, geothermal and biomass) and a battery storage system, using intelligent management methods that optimize energy production costs. The proposed intelligent management methods are based on cost function minimization. These ensure optimal energy management operation under uncertain weather conditions, keeping to the minimum the price of the energy produced by the microgrid. Due to the erratic nature of solar irradiance and considering that solar energy is cheaper than its counterparts (geothermal and biomass in this case), forecasting the photovoltaic power production is an essential task for guaranteeing optimal microgrid operation. Day-ahead photovoltaic power forecasts are issued based on solar irradiance forecasts provided by the Global Forecasting System and further integrated into the day-ahead scheduling of the microgrid. This algorithm, together with the cost-optimization algorithms, gives intelligence to the management protocol generating the input parameters (power and operating time for each generating unit) for the microgrid Energy Management Controller. Different cost optimization methods are studied and compared, including mathematical models, such as Mixed Integer Linear Programming, or nature-inspired optimization algorithms, such as Genetic Algorithm, Harmony Search Algorithm, and Particle Swarm Optimization. It is shown that these algorithms provided suitable solutions for the energy balance and scheduling ensuring minimal operation costs for the microgrid. The study also highlights the differences and limitations of these algorithms and shows a method on how these energy management methods can be integrated into a real-world microgrid providing a step forward in the development of intelligent microgrids.