Customer modeling and pricing-mechanisms for demand response in smart electric distribution grids

We describe and contrast different market mechanisms to incentivize residential electricity customers to perform demand response (DR) via load shifting of schedulable assets. A customer-incentive pricing (CIP) mechanism from our past research is discussed, and compared to flat-rate, time-of-use (TOU), and real-time pricing (RTP). The comparison is made using a for-profit aggregator-based residential DR approach to solve the “Smart Grid resource allocation” (SGRA) problem. The aggregator uses a heuristic framework to schedule customer assets and to determine the customer-incentive price to maximize profit. Different customer response models are proposed to emulate customer behavior in the aggregator DR program. A large-scale system consisting of 5,555 residential customer households and 56,588 schedulable assets using real pricing data over a period of 24 h is simulated and controlled using the aggregator. We show that the aggregator enacts a beneficial change on the load profile of the overall power system by reducing peak demand. Additionally, the customers who are more flexible with their loads, represented as a parameter in the proposed customer α-model, have a greater reduction on their electricity bill.1 ∗Electrical Engineering and Computer Science Department, South Dakota State University, Brookings, SD, USA ∗∗IRTES, Université de Technologie de Belfort-Montbéliard, Université Bourgogne Franche-Comté, Belfort, France †Department of Electrical and Computer Engineering, Colorado State University, Fort Collins, CO, USA ‡Department of Computer Science, Colorado State University, Fort Collins, CO, USA 1This chapter is an expanded version of a journal article [1]. Project codes and data sets have been made available for use through the open source BSD 3-Clause license at https://github.com/IPEMS 136 Cyber-physical-social systems and constructs in electric power engineering 6.1 Customer modeling introduction The US Energy Information Administration predicts a 24% increase in residential electricity use from a 2013 reference case to the year 2040. Additionally, since 1982 the growth in peak electricity usage has exceeded the growth in transmission capacity by almost 25% each year [2]. Given the combination of these trends, it is expected that peak energy demands will approach, and eventually exceed, the available transmission system capability (the remaining power that could be transferred from generation to consumption). A side consequence of these trends, in addition to the possibility of system outage, is more congestion on transmission lines, leading to increased locational marginal prices (LMP) – different marginal prices (price of providing the next MW of power) at different buses in the transmission network. Studies show that small and targeted reductions in peak demand can have large impacts on wholesale electricity prices [3]. As shown in Reference 2, it is unlikely that additional spending will be allocated for increasing transmission capability, leading to research in the areas of distributed generation (DG) and, in the case of this chapter, shifting or curtailing load during peak hours. Given that residential customers can account for over half of the system peak demand in summertime, such as in markets like the Electric Reliability Council of Texas (ERCOT) [4], residential demand response (DR) programs are attractive solutions for relieving the stress on the system and market. In this work, we define the term DR to be the reduction in peak demand by shifting or shedding loads in response to system or economic conditions to alleviate stress on the electric power system. As presented in Reference 5, incentives can influence customer behavioral changes. Dynamic pricing programs are one method for accomplishing DR. These utility-offered programs, such as time-of-use (TOU) and real-time pricing (RTP), fluctuate the price of electricity throughout the day in accordance with system load levels to elicit a change in the consumption of electricity [6]. In this chapter, we introduced an additional pricing method, customer-incentive pricing (CIP), which provides the residential end-user an additional competitive pricing scheme for participating in a targeted DR program through an aggregator. The aggregator in this chapter is a forprofit entity in a deregulated market structure that interfaces a DR exchange market (DRX) and a set of customers. The aggregator uses the combination of many customer schedulable assets of the participating customers to perform large-scale load shifting. In many, if not all energy markets, there is a minimum power rating required to bid into the market (e.g., 0.1 MW in the PJM market [7]). The aggregator entity is able to enact a change on the electric power system load profile by bidding the aggregate load of customer assets into the bulk market through the DRX. By voluntarily opting into the aggregator DR program, the customer is provided the opportunity to participate in the bulk power market. Residential customers can change electricity use to take advantage of the time-varying rates provided by the utility and aggregator to reduce their electricity bill. The challenges of an effective residential DR program are (a) the uncertainty in the time-varying price of electricity and (b) that as a customer, the benefit received from changing energy usage must exceed the inconvenience caused. To overcome these challenges and to Customer modeling and pricing-mechanisms for demand response 137 IS O Market, control (cyber) Power system (physical) End-user (social) SG RA tility CPSS Figure 6.1 The proposed SGRA CPSS lies at the intersection between the power system network, market and controls, and end-user maximize the benefit of dynamic pricing, we introduce the aggregator-based residential DR program, denoted Smart Grid Resource Allocation (SGRA), where given a set of participating customers with schedulable assets, subject to customer constraints (i.e., availability of customer assets and customer incentive requirements), the aggregator sets the CIP and schedule of assets to maximize profit of the aggregator while not inconveniencing the participating customer. The SGRA is formulated as a cyber-physical-social system (CPSS), as illustrated in Figure 6.1. The proposed SGRA CPSS lies at the intersection of the physical electric power network, cyber market and control layer, and social residential end-users. The independent system operator (ISO) interfaces the bulk power market and physical equipment. Local utilities traditionally deliver power to end-user customers through the distribution network. The aggregator entity provides the residential end-user a path to participate in the market. The SGRA problem is solved using resource allocation methods analogous to those used in the computing discipline, where compute tasks must be allocated to machines. In general, such problems are NP-complete [8–10], so heuristic optimization is used to find near-optimal solutions. In this chapter, we use heuristic optimization techniques to find near-optimal solutions to the SGRA problem in a reasonable time frame to be used as a day-ahead scheduler of a large number of residential end-user assets. According to the California Energy Commission (CEC), residential loads are not easily controlled and need to be composed of a large portfolio of assets to provide a strategic DR product [11]. The work in this chapter directly addresses the CEC DR strategies (direct DR participation with the ISO, new market and auction mechanisms, e.g., our proposed DRX, improving customer willingness to participate, and the introduction of time-variant pricing) by offering direct DR participation through 138 Cyber-physical-social systems and constructs in electric power engineering