Data-driven optimization for microgrid control under distributed energy resource variability

Area of Science:

• Electrical Engineering
• Renewable Energy Systems
• Optimization Techniques

Background:

• Smart grids benefit from renewable energy integration for resilience and clean electricity.
• The intermittent nature of renewable sources necessitates advanced scheduling strategies for Microgrids.
• Optimizing hybrid systems with photovoltaic (PV), wind, and controllable distributed generators (CDGs) is crucial for both grid-connected and stand-alone operations.

Purpose of the Study:

• To develop a priority-based cost optimization function for hybrid Microgrids.
• To incorporate uncertainties of intermittent parameters into the scheduling methodology.
• To minimize the overall operational cost of Microgrids through advanced optimization.

Main Methods:

• A priority-based cost optimization function was developed, including operating, emission, battery, grid energy exchange, and load shedding costs.
• Monte Carlo simulations generated multiple scenarios for uncertain parameters, which were then reduced using k-means clustering.
• The Grey Wolf Optimization (GWO) algorithm was employed to minimize the developed cost function, validated on the CIGRE test network.

Main Results:

• The GWO algorithm demonstrated effectiveness in minimizing the Microgrid's cost function under various scenarios.
• Different priorities assigned to sub-objectives yielded distinct operational outcomes, showcasing the flexibility of the proposed method.
• The GWO approach proved superior to Jaya and Particle Swarm Optimization (PSO) algorithms in optimizing the Microgrid's cost.

Conclusions:

• The proposed priority-based cost optimization strategy effectively manages uncertainties in hybrid Microgrids.
• The Grey Wolf Optimization algorithm provides a robust and efficient method for optimizing Microgrid operations.
• The methodology ensures reliable and cost-effective energy supply in both grid-connected and stand-alone Microgrid configurations.