Intelligent and Autonomous Monitoring of Peak Demand Avoidance in Commercial Refrigeration Systems
The utility service providers calculate the peak demand charges based on the highest level of power consumption that a facility uses in any interval (usually 15 mins) during the billing cycle. The peak demand charges in facilities such as supermarkets could represent nearly up to 40% of the total utility bill. In supermarkets, besides the building, refrigeration systems could potentially play a major role in affecting the peak demands. Traditional threshold-based load shedding rules are not effective in controlling the peak demand of the refrigeration system since tailoring the power demand solely due to the reduction of the peaks above a specific threshold could create new peaks with higher values in the future. To reduce the peak demands effectively, forecasting methods to estimate the power demand based on historical data and weather predictions and operational optimization methods to improve the power demand flexibility are required. The goal of this project is to develop highly accurate forecasting/optimization systems using neural networks and machine learning models to reduce the peak demands based on refrigeration systems and expand the capabilities of existing edge devices for autonomous decision-making.