Date of Award
Open Access Thesis
Commercial and industrial customers have increasingly adopted fully-automated demand response (auto-DR) as part of their energy management control strategy to mitigate the effects of rising electricity costs. Some of these customers have also contracted with their power provider to permit voluntary curtailment of power consumption to be negotiated by a human operator. Even still, at the customer-side there are needs to maximize cost savings and automate decision making, and at the provider-side there are needs to reduce peaking demand and demand volatility. This thesis describes a centralized multiagent approach that automatically negotiates demand curtailment with a power provider while making the best use of available distributed energy resources. Weather-dependent load and source forecasting methods are implemented to improve decision-making of the agents. Power forecast uncertainty is measured in terms of confidence levels. The effectiveness of the approach was evaluated by simulating the system performance of an actual nursing home that was notionally-augmented with a photovoltaic (PV) power source and battery energy storage system. Actual power consumption data for two representative four-day periods during spring and summer of 2012 were used, together with the simulated performance of the PV and battery systems working together with the agent-based controller. The multiagent system with peak-time load curtailment was able to reduce total electric usage costs by 24%.
Price, T. W.(2012). A Multiagent Energy Management Control System for Commercial & Industrial Applications. (Master's thesis). Retrieved from https://scholarcommons.sc.edu/etd/2202