Growing concerns over global warming and increasing fuel costs have pushed the traditional fuel-based centralized electrical grid to the forefront of mounting public pressure. These concerns will only intensify in the future, owing to the growth in electricity demand. Such growths require increased generation of electricity to meet the demand, equating to more carbon footprint from the electrical grid. To meet the growing demand economically by using clean sources of energy, the electrical grid needs significant structural and operational changes to cope with various challenges.

 

Microgrids (μGs) can be an answer to the structural requirement of the electrical grid. μGs integrate renewables and serve local needs within electrical boundaries, thereby, reducing line losses and improving resiliency. However, stochastic nature of electricity harvest from renewables makes its integration into the grid challenging. The time varying and intermittent nature of renewables and consumer demand can be mitigated by using energy storages and dynamic load scheduling. Automated dynamic load scheduling constitutes the operational changes that could enable us to achieve energy efficiency in the grid. Most of the existing work considers non-production line loads (residential or commercial buildings) or flow shop scheduling (industrial manufacturing site).

 

This research aims at developing a framework for dynamic load scheduling in μGs from the perspective of consumers (Demand Side Management) and grid controller (Demand Response). The research consists of building 1) a deterministic optimization model with perfect information sharing and integrating different (types of) electrical loads and (types of) consumers as a benchmark for dynamic models 2) a dynamic load scheduling model from a consumer’s (job shop manufacturing site) perspective 3) a dynamic load scheduling model for different consumers with partial information sharing from a grid controller’s perspective 4) an end-to-end reinforcement learning model for dynamic load scheduling based on guided policy search using optimal solutions built in part 1. While the model developed in this research outperform existing industry standard heuristics, they perform comparatively to the benchmark solutions.