Demand response implementation for improved system efficiency in remote communities

This paper evaluates the performance of a demand response (DR) system, installed in the remote community of Hartley Bay, British Columbia, which is used to reduce fuel consumption during periods of peak loads and poor fuel efficiency. The DR system, installed to shed load during these periods, is capable of shedding up to 15 per cent of maximum demand by adjusting wireless variable thermostats and load controllers on hot water heaters and ventilation systems in commercial buildings. The system was found to be successful in reducing demand by up to 35 kW during the DR event period, but caused a new, time-shifted “rebound” peak of 30 to 50 per cent following each event. A DR “staggering” method is introduced as a tool for reducing and delaying rebound without affecting occupant comfort and safety. In this work, load prediction models based on linear regression and averaging of historical data were also developed for measuring DR shed and rebound, with models based on averaging found to produce more accurate baselines.