Doctor of Philosophy
Other advisers/committee members
Kishore, Shalinee; Blum, Rick S.; Venkitasubramaniam, Parv; Snyder, Lawrence V.
Future electricity grids will enable greater and more sophisticated demand side participation, which refers to the inclusion of mechanisms that enable dynamic modification of electricity demand into the operations of the electricity market, known as Demand Response (DR). The underlying information-flow infrastructures provided by the emerging smart grid enhance the interactions between customers and the market, by which DR will improve electricity grids in several aspects, e.g., by reducing peak demand and reducing need for expensive peaker plants, or by enabling demand to follow supply such as those from volatile renewable resources, etc. Many types of appliances provide flexibilities in power usage which can be viewed as demand response resources, and how to exploit such flexibilities to achieve the benefits offered by DR is a central challenge. In this dissertation, we design algorithms and architectures to bridge the gap between scheduling appliances and the benefits that DR can bring to electricity grid by utilizing the smart grid's underlying information infrastructure. First, we focus on demand response within the consumer premise, where an energy management controller (EMC) schedules appliance operation on behalf of customers to save energy cost. We propose an optimization-based control scheme for the EMC in the building that integrates both the operational flexible appliances such as clothes washer/dryer, dish washer and plug-in electric vehicles (PEVs), but also the thermostatically controlled appliances such as HVAC (heating, ventilation, and air conditioning) systems together with the thermal mass of the building. Model predictive control is employed to account for uncertainty in electricity prices and weather information. Under time-varying pricing, scheduling appliances smartly using our scheme can incur notable energy cost saving for customers. As an alternative, we also propose a communication-based control approach which is a joint appliance access and scheduling scheme in which the control algorithms are embedded into the communication protocols used by appliances. The control scheme is based on a threshold maximum power consumption set by the EMC; and we discuss how this threshold can be chosen so that it integrates the availability of local distributed renewable energy resources.Then we investigate demand response in the retail market level which involves interactions between customers and utilities. Pricing-based control and direct load control (DLC) are two types of approaches that are used or envisioned for this level. To address pricing based control methods, we propose real-time pricing (RTP) signals that can be designed to work with customer premise EMCs. The interaction between these EMCs and the pricing-setting utilities is modeled as a Stackelberg game. We demonstrate that our proposed RTP scheme reduces peak load and alleviates rebound peaks that are the typical shortcomings in existing pricing approaches. To address DLC methods, we propose a distributed DLC scheme based on a two-layer communication network infrastructure for large-scale, aggregate DR implementations. In the proposed scheme, average consensus algorithms are employed to distributively allocate control tasks amongst EMCs so that local appliance scheduling within each home will eventually achieve the aggregated control task, i.e., to alleviate mismatch between electricity supply and demand.Finally, we study how demand response affects the wholesale electricity market. As is conventional when studying interactions between electricity generators, we employ the Cournot game model to analyze how DR aggregators may impact wholesale energy markets. To do so, we assume that DR aggregators employ a computationally efficient, centralized scheduling mechanism to manage deferrable load over a large aggregate set of consumers. The load reduction from deferrable load can be seen as `generation' in terms of balancing the market and is compensated as such under current regulatory mandates. Thus, the DR aggregator competes with other generators in a Cournot-Nash manner to make a profit in the wholesale market; and electricity prices are consequently reduced. We provide equilibrium analysis of the wholesale market that includes DR aggregators and demonstrate that under certain conditions the equilibrium exists and is unique.
Chen, Chen, "Topics in Demand Response for Energy Management in Smart Grid" (2013). Theses and Dissertations. 1454.