Energy Peak Load Management: Overview
As my secondary stream of research, I have studied electricity demand management during peak load periods, when the supply cost far exceeds the revenue. In the following paper, I study the implementation of Direct Load Control Contracts that allow utility companies to directly control the participants’ electricity consumption:
 Peak Load Energy Management by Direct Load Control Contracts. Operations Research. 3rd major revision was invited. (with Reza Ahmadi and Sriram Dasu)
Peak periods refer to the time intervals when electricity demand exceeds primary generators’ capacity (e.g., renewable, nuclear, coal, and gas) and secondary sources, which are significantly more costly, such as diesel and gasoline-powered generators must be activated. The supply cost far exceeds the revenue that can be generated during peak periods. Therefore, many utility companies focus on managing demand as opposed to increasing the supply. In , we study implementation of Direct Load Control Contracts (DLCCs) that permit utility companies to directly reduce a customer’s energy usage using a remote control device that is installed on site (e.g., see Southern California Edison (www.sce.com). DLCCs help utilities effectively manage demand during peak periods, ensure the reliability of the electric grid, reduce “greenhouse gas emissions and carbon footprint,” and help “customers reduce their energy use and utility bills” (e.g., see Los Angeles Department of Water & Power (www.ladwp.com).
DLCCs are executed as follows. Each day, the firm must decide which groups of customers to interrupt, the starting time, and the duration for each interruption. The objective is to minimize the total power generation cost over the contract horizon subject to constraints on the total number and time of interruptions for each participating customer. The decision problem faced by the utilities can be modeled as a finite-horizon stochastic dynamic program with a high-dimensional state space. We develop an asymptotically optimal approximation scheme for solving it. We apply our solution approach to the data provided by three major utility companies in California. Our analysis indicates that applying our approach can potentially reduce the peak load consumption by approximately 11% on a hot summer day, which corresponds to a 50% cost reduction during peak hours. Overall, our study shows a potential for 5-7% savings in energy generation cost.
I currently study fairness in the assignment of interruptions to customers. Fairness policy dictates that the total times and/or the number of interruptions for different groups must be almost equal during different time intervals, e.g. daily, weekly, or monthly. For example, under a weekly fair policy, it is not permitted to assign 1 call with a duration of 2 hours to one group and 2 calls with durations 5 and 6 hours to another group. This is obviously unfair to the second group and it can result in their dissatisfaction. We aim to design a near optimal implementation policy that ensures fairness.