Demand Response in Smart Grids

Recent changes in the market rules, advancement in control and communication technology, and large-scale deployment of smart meters have made demand response a viable option to improve the reliability and efficiency of the power systems. Demand response is a change in electric usage by end-use customers from their normal consumption patterns in an exchange for a monetary gain. This chapter presents an overview of demand response, including the basic concept, history, control mechanisms, and the current statues of load resources participating in the wholesale market. It also summarizes the lessons learned from those regions which have provided a steady revenue resource potential to demand response participants. Finally, the future challenges for removing market barriers and coordinating a large number of load resources are identified in a future power grid where a large amount of demand response will be incorporated.

In general, end use loads are divided into four categories: residential, commercial, industrial and agricultural. In this chapter, our interest is in modeling residential and commercial building demand response (DR) resources. Unlike generators, end use electricity consumptions are heavily influenced by customer preferences. So, the human factors need to be modeled besides the working principles of DR resources. We will focus on modeling steady-state behaviors of a DR resource so the transient response is ignored.

In this chapter, we will introduce the modeling of thermostatically controlled loads (TCLs), non-TCL controllable loads, and the base load. Electricity consumptions of an ideal DR resource can be curtailed or shifted without causing enduring inconvenience and severe degradation in performance and lifetime. Therefore, TCLs, such as heat ventilation and air conditioning (HVAC) units, water heaters, and refrigeration loads, are generally considered to be the most suitable DR resources. Non-TCL, infrequently used large loads such as washers and dryers, exterior lighting, and pool pumps, are also good DR resources. Must-run loads, such as interior lighting, entertainment, and cooking loads, are usually considered as the base load because their consumptions can be neither delayed nor reduced.

Load management algorithms are widely used to manage and coordinate demand response programs for providing grid services such as peak shaving, load shifting, load following, and frequency response. In general, there are three control approaches: direct load control (DLC), indirect load control (ILC), and autonomous load control (ALC). In DLC, utilities or load services aggregators will directly control the on and off of a controllable load and the owner of the controllable load will not interfere with the control process. In ILC, the owner of the controllable load will respond to a control signal by setting up a triggering threshold or a response curve. In ALC, the controllable load is equipped with an embedded control logic that allows it to autonomously respond to system frequency and voltage deviations without communicating with utilities, load aggregators, or its owner.

Unlike managing generators, loads have limited controllability and observability. Normally, an appliance can either be turned “on” or “off” and its power consumption cannot be continuously varied. In addition, its operation is governed by time-varying external variables such as ambient temperatures, operation modes, and user’s comfort. Furthermore, the cost of controlling loads changes with respect to time-of-the-day, customer comfort preferences, occupancy, etc. Therefore, many well-developed energy management algorithms, such as unit commitment and economic dispatch, are not applicable to manage demand response (DR) programs. It is thus critical to understand a few basic design principles that account for those distinct characteristics of load management to design and develop reliable, robust, and economical DR energy management systems.

The Electric Reliability Council of Texas (ERCOT) is one of the Independent System Operators (ISOs) formed to ensure reliability of the electricity grid in the state of Texas. ERCOT administers both energy and ancillary service market while ensuring a precise balance between load and generation. The market rules at ERCOT have been evolved to allow load resources to participate in these markets and emergency programs. This chapter presents a review of current market design and the statuses of demand response programs at ERCOT, with special focuses on Load Acting as a Resource (LaaR) to provide Responsive Reserve Service, Controllable Load Resources participating in Security Constrained Economic Dispatch (SCED) and 10-min and 30-min Emergency Response Service. The performance requirements and the benefits from deployment of these load resources are illustrated through the real-world events.

Demand response (DR) is a critical and effective measure to stimulate the demand side resources to interact with renewable generation in the power system. However, the conventional scope of DR cannot fully exploit the interaction capabilities of demand side resources, which limits the energy users in the electric power system. With the revolution of the traditional economic and social pattern based on centralized fossil energy consumption, “Energy Internet” is impelling the development of the third industrial revolution, which aims at promoting the incorporation of sustainable energy and internet technology, and facilitating the integration of multi-energy systems (MESs). By integrating electricity, thermal energy, natural gas, and other forms of energy, the smart energy hub (SEH) makes it possible for energy users to flexibly switch the source of consumed energy. With the complementarity of MESs, even the inelastic loads can actively participate in DR programs, which fully exploits the interaction capability of DR resources while maintaining the consumers’ comfort. This novel vision of the DR programs is termed as “Integrated Demand Response (IDR).” In this context, the state of the art of IDR in the MESs is reviewed for the first time. First, the basic concept of IDR and the value analysis are introduced. The research on IDR in the MES is then summarized. The overviews of the engineering projects around the world are introduced. Finally, the key issues and potential research topics on IDR in the MES are proposed. Hopefully, this chapter will provide reference for future research and engineering projects on IDR programs in the MES.

This chapter presents the formulation and critical assessment of a novel type of demand response (DR) program targeting retail customers (such as small/medium size commercial, industrial, and residential customers) who are equipped with smart meters yet still face a flat rate. Enabled by pervasive mobile communication capabilities and smart grid technologies, load serving entities (LSEs) could offer retail customers coupon incentives via near-real-time information networks to induce demand response for a future period of time in anticipation of intermittent generation ramping and/or price spikes. This scheme is referred to as coupon incentive-based demand response (CIDR). In contrast to the real-time pricing or peak load pricing DR programs, CIDR continues to offer a flat rate to retail customers and also provides them with voluntary incentives to induce demand response. Theoretical analysis shows the benefits of the CIDR scheme in terms of social welfare, consumer surplus, LSE profit, the robustness of the retail electricity rate, and readiness for implementation. The pros and cons are discussed in comparison with existing DR programs. Numerical illustration is performed based on realistic supply and demand data obtained from the Electric Reliability Council of Texas (ERCOT).

In this chapter, a real-time demand response (DR) framework and model for a smart distribution grid is formulated. The model is optimized in a distributed manner with the Lagrangian relaxation (LR) method. Consumers adjust their own hourly load level in response to real-time prices (RTP) of electricity to maximize their utility. Because the convergence performance of existing distributed algorithms highly relies on the selection of the iteration step size and search direction, a novel approach termed Lagrangian multiplier optimal selection (LMOS) is proposed to overcome this difficulty. Via sensitivity analysis, the energy demand elasticity of consumers can be effectively estimated. Then the LMOS model can be established to optimize the Lagrangian multipliers in a relatively small linearized neighborhood. The salient feature of LMOS is its capability to optimally determine the Lagrangian multipliers during each iteration, which greatly improves the convergence performance of the distributed algorithm. Case studies based on a distribution grid with the number of consumers ranging from 10 to 100 demonstrate that the proposed method greatly outperforms the prevalent approaches, in terms of both efficiency and robustness.

With integration of more and more renewable energy resources, it is becoming increasingly difficult to maintain adequate primary frequency control (PFC) capability for a future power grid, especially under low system inertia conditions. Load resources (LRs) equipped with under-frequency relays can participate in PFC supplementing to the governor responses from synchronous units. In this chapter, we describe an energy, inertia, and frequency response reserve (FRR) co-optimization formulation in the day-ahead market where both primary frequency reserve (PFR) from synchronous generators and fast frequency response reserve (FFR) from LRs are procured in a cooperative way to meet the desired FRR need tied to the system inertia condition. First, the performance of the governor-provided PFC is evaluated as a benchmark to quantify the effectiveness of the load-provided PFC at each typical system inertia condition, using the actual dynamic network models and operation data of the ERCOT system. Second, since FFR is more effective than PFR in arresting the frequency decline, an optimization approach is proposed to yield different marginal prices for FFR and its PFR counterpart to award the speed of response. The case study shows the effectiveness of the proposed approach and the correctness of the quantities and prices of the cleared reserve.

Demand response control for the residential house load, which represents the largest share of load demand, has been designed with a simple logic in the past, and this could limit the gain and economic benefits which the end-users may obtain. This chapter presents a novel appliance commitment algorithm that schedules thermostatically controlled household loads based on price and consumption forecasts considering users’ comfort settings to meet an optimization objective such as minimum payment or maximum comfort. The formulation of an appliance commitment problem is described using an electrical water heater load as an example. The thermal dynamics of heating and coasting of the water heater load is modeled by physical models; random hot water consumption is modeled with statistical methods. The models are used to predict the appliance operation over the scheduling time horizon. User comfort is transformed to a set of linear constraints. Then, a novel linear-sequential-optimization-enhanced, multi-loop algorithm is used to solve the appliance commitment problem. The simulation results demonstrate that the algorithm is fast, robust, and flexible. The algorithm can be used in home/building energy-management systems to help household owners or building managers to automatically create optimal load operation schedules based on different cost and comfort settings and compare cost/benefits among schedules.