Optimal Control of Electrical Demand within a Sliding Window Billing Structure

On-line computer control of electrical demand in industrial power plants permits savings in energy costs by controlling demand to a target value at the conclusion of each demand period through modulation of turbine power generation. Electrical utilities are with increasing frequency withholding synchronization pulses from their customers, and metering demand on a continuous basis as the W-minute moving average of purchased power (the sliding window). As a result, conventional demand control methods no longer apply, and straightforward power regulation is done instead. However, a theoretical comparison shows that the difference of performances between a minimum variance power regulator and a minimum variance demand regulator is substantial. This encouraging result provides the motivation for a regulator design based on the LQ theory which, as simulations show, achieves in large measure the theoretically derived expectations. The simulation results are then used to derive potential economic benefits in a typical case.