Domestic electric water heater control for peak shaving & frequency control

Thumbnail Image



Journal Title

Journal ISSN

Volume Title


University of New Brunswick


The traditional operation paradigm in power systems is to dispatch controllable generators to follow a variable electricity demand. In such a way, the objective of real-time generation/load balance is achieved. The power system infrastructure is designed to meet the highest level of demand; consequently, it is underutilized during off-peak periods. Hence, peak shaving helps to reduce capital investment, operating, and maintenance costs. With the increasing penetration of renewable energy resources (RERs), the requirement for traditional generators to provide electricity is decreasing, then on-line generators' reserves are decreasing. Hence, RERs require additional reserves to defend the fluctuating generation. Therefore, additional ancillary service resources are required, such as frequency control services. Conceptually, to control some loads to follow power production is possible, which was proposed in the 1980s, to help to provide peak shaving and ancillary services. The advantages of a domestic electric water heater (DEWH) make it a primary candidate for direct load control (DLC). The operation of a conventional DEWH is dependent on its internal thermostat, which reacts to water temperature directly and is uncontrollable. An external relay can be installed to exercise control on conventional DEWHs, but the internal end-user comfort may not be guaranteed under the relay control without temperature measurements. In the thesis, an individual DEWH state estimation method is introduced to provide a reference for selecting control actions. Based on hot-water consumption patterns and characteristics of hot-water consumption activities, the worst cases are considered to maintain the end-user comfort without temperature measurements. A customer satisfaction prediction index (CSPI) is proposed to gauge the comfort level of DEWH users, without temperature measurements, and is generated by a time-varying weight matrix. Subsequent hot-water consumption is considered in the CSPI. A peak shaving algorithm based on the CSPI is proposed in the thesis. As DEWHs are resistance heating devices, they can be used to provide frequency control services for utilities. DEWHs can be aggregated as a virtual frequency control provider (VFCP) to provide frequency control under normal and contingency conditions, and the available capacity is estimated and sent to the system operator (SO) for its controls. Two control algorithms are proposed to provide frequency control services by the VFCP. The results of this thesis show that: (1) The end-user comfort can be maintained without temperature measurements with the proposed individual DEWH state estimation method. (2) The proposed peak shaving method reduces morning peak by 24.0% and evening peak by 31.3%. (3) The VFCP can be an effective resource to provide frequency control in power systems under different situations.