Jason Wayne Black

Emergency demand response for distribution system contingencies

The use of demand side resources to respond to contingencies on electric power systems is typically limited to interruptible service contracts for large commercial or industrial customers and bulk load shedding schemes. There are a few programs that allow bidding of emergency response capability at the transmission level, but are primarily for response to reserve margin shortages. This paper proposes a demand response program that enables the use of residential load resources to respond to contingencies down to the distribution level. This contingency response program enables targeted, selective demand response that expands the range of contingencies that can be mitigated by demand and substitutes the shedding of non-critical loads for bulk load shedding programs. The program is targeted, in that it allows for implementation at any node in the system, and selective in that it chooses the least amount and lowest impact/cost loads to shed to alleviate a given contingency.

Potential problems with large scale differential pricing programs

Flat rate prices for residential customers have historically enhanced the ability of system operators to predict demand by providing a smooth, certain price signal, thus reducing risk in meeting the need to instantaneously balance supply and demand in electricity systems/markets. The desire to reduce peak loads, however, has lead to exploration of dynamic pricing, including time of use and critical peak pricing programs. These programs are currently in the pilot stage throughout the US, with low overall participation by residential load. Large-scale participation in dynamic pricing programs can cause unwanted consequences that will not be observed in small-scale programs. This paper investigates several potential negative consequences from large-scale participation in existing dynamic pricing programs. These include: the rebound effect, coincident load shifting/shedding, and limitations of fixed, uniform pricing periods.

Optimal scheduling of demand response events using options valuation methods

Several demand response (DR) programs, such as critical peak pricing (CPP), stipulate that electric utilities may schedule events only a limited number of times per year. Utilities must therefore use these events judiciously in order to maximize their benefits from the DR program. Traditionally, they have used rather simple decision rules to schedule such events (e.g., when temperature exceeds a certain threshold). In this paper, we present an option value approach for determining when to invoke these events. This approach calculates a dynamic threshold value, which represents the option value of deferring the use of one the events for use at any future point in the planning horizon (typically a year or cooling season). This threshold enables utilities to make the decision on whether or not to call an event based on the expected current benefits versus potential future benefits. This paper presents an example based on an actual DR program, which uses a simple, temperature based threshold, and compares it with our option based approach. Our approach can also be applied to any other criteria besides temperature such as reserve margins and generation cost.

Scheduling demand response events with constraints on total number of events per year

Several demand response (DR) programs, such as critical peak pricing (CPP), stipulate that electric utilities may schedule events only a limited number of times per year. Utilities must therefore use these events judiciously in order to maximize their benefits from the DR program. Traditionally, they have used rather simple decision rules to schedule such events (e.g., when temperature exceeds a certain threshold). In this paper, we present an option value approach for determining when to invoke these events. This approach calculates a dynamic threshold value, which represents the option value of deferring the use of one the events for use at any future point in the planning horizon (typically a year or cooling season). This threshold enables utilities to make the decision on whether or not to call an event based on the expected current benefits versus potential future benefits. This paper presents an example based on an actual DR program, which uses a simple, temperature based threshold, and compares it with our option based approach. Our approach can also be applied to any other criteria besides temperature such as reserve margins and generation cost. Our approach may also be classified as an optimal stopping rule or optimal secretary problem in the operations research literature.

Transforming Electrical Load from an Operational Constraint to a Controllable Resource

Electric utilities have historically treated power demand as an uncontrollable input, requiring generation and transmission resources to maintain the supply-demand balance. In recent years, demand response (DR) has emerged as a means to manage customer loads to balance the grid. This paper presents analytic solutions to enable utilities to optimize DR programs to serve as operational resources for the grid. We developed two sets of analytics. First, we developed a clustering-based method to accurately estimate the load curtailments expected from customers during DR events. Then, we used an option value-based optimal DR event scheduling method to compute a dynamic threshold value that the utility can use to make daily decisions for triggering DR events. In extensive tests, the proposed methods show superior performance over existing approaches. We implemented these analytics in the General Electric (GE) PowerOn™ Precision Demand Response Management System, which GE offered from 2011 to 2015.