Assessing Demand Response Potential

Determine Key Parameters

To calculate the demand response achievable technical potential, there are several key inputs parameters needed: global inputs (economic and system) and product inputs (impact and cost).[1] The global inputs are in the table below:

DR Model Global Inputs

                       * Labor cost was not used in the 2021 Plan analysis

Of these parameters, only a few are specific to the demand response model: peak definition, number of events, and event length. For the 2021 Plan, the peak definition was not limited, i.e. any hour could be the peak and is determined based on top hours of the regional system load profile within a season.  The top hours are based on event length (four hours) and number of events per season (five) and assumed for all products. The remaining economic and system inputs were consistent with other elements of the 2021 Power Plan, largely the demand forecast. Because the portfolio analysis models can only incorporate a single demand response supply curve, while the demand forecast is stochastic under a wide array of economic and climate assumptions, a single set of assumptions was used for the DR potential - the high economic forecast under the CanESM2 climate model. The high economic scenario was chosen to emulate peak events when DR is likely to be deployed, and the CanESM2 climate exhibited the widest temperature range.

The product specific cost and impact inputs are in the following table:

DR Model Product Inputs

Council staff worked in conjunction with the demand response advisory committee to determine the cost and impact input parameters for each DR product. In addition, the product eligibility was largely based on external data sources such as: the residential building stock assessment, the commercial building stock assessment, and the regional technical forum’s work products. Where appropriate, assumptions are consistent between the DR and energy efficiency supply curves.

Estimate Achievable Technical Potential

The 2021 Plan estimated the DR achievable technical potential using two approaches, referred to as the “bottom-up” and “top-down” approaches. Two approaches were used since, depending on the DR product, the understanding of DR impacts may be better at the end-use equipment level (bottom-up), or instead at the sector or end-use load level (top-down). In general, the residential and small commercial controllable products were modeled using the bottom-up methodology and the larger non-residential and all pricing products were modeled using the top-down methodology.

The figures below (courtesy of Cadmus) illustrate how the key parameters are used to calculate the achievable technical potential using the two approaches. Note, these figures include an intermediate calculation of the “technical potential”, which determines the total amount of potential that could technically be reduced without any market or adoption barriers.

Illustration of Bottom-Up Methodology to Estimate Potential

Illustration of Top-Down Methodology to Estimate Potential

As is illustrated in the figures above, the key difference between the two methodologies is that in the bottom-up methodology, the potential is calculated from per-unit impacts multiplied by the number of available units. In the top-down approach, the calculation is more complex as the impact is estimated from the percent of peak load reduction of the eligible load basis, which is determined by the availability of end-use load during the peak hours, as calculated from the hourly load profile.

This table provides the breakdown on which methodology was used for which product, with applicable sector provided in parentheses.  

Products Listed by Assessment Methodology*

* Res = Residential; Com = Commercial, Ind = Industrial, Ag = Agricultural

Ramp Rates

In addition, a program ramp rate is included to reflect the number of years it may take to reach full participation. For example, for the AC Switch product, the full participation rate is 10 percent, but the assumption is that it will take 5 years to reach that level, linearly increasing over that time – so two percent of the population in year one, four percent in year two, and so on. For all price-based demand response (time-of-use, critical peak pricing, real time pricing), irrigation load control, and bring-your-own thermostat products, the time to full potential was assumed to be three years. Grid-enabled heat pump water heaters have a 10 year ramp, due to the low prevalence of these efficient water heaters currently available in the market. All other products assumed five years ramp. These ramp rates are meant to reflect how quickly a utility could implement a full-scale demand response program, if the economic motivation were present.   

Calculate Net Levelized Cost

In addition to the total impact that could be achieved, by year, for a given product, the supply curves also include an estimate of what it would cost, on a per-kilowatt-year basis, to acquire these impacts. All costs are in 2016 dollars. In order to fairly compare all resources in the strategy analysis, these costs are expressed on a net levelized cost basis, which accounts for both positive and negative costs (e.g. benefits). These are outlined here for all resources evaluated in the 2021 Plan. For DR, the only negative cost included is the value of demand response to defer transmission and distribution investment (T&D deferral). The T&D deferral was estimated to be about $9.80 per kilowatt-year, where $3.10 is from transmission deferral and $6.70 from distribution system losses and is based on input from regional utilities.[2]

For the 2021 Plan, the DR product-specific costs included are further described above and include: setup, operation and maintenance, equipment, marketing, and incentives. These, plus the T&D deferral, are the components of the net costs, which are then levelized to 2016 dollars using a 3.75 percent discount rate. Note, not every DR product may include all costs. For example, a bring-your-own-thermostat program will not have any equipment cost as it is assumed the customer already owns the thermostat.

One important contribution to the costs is the incentive, namely for controllable products (price-based products do not generally have a fixed incentive as the end-use customer is compensated through bill reductions). The 2021 Plan only included a portion of the fixed incentive that offsets the end-use customers loss of utility experienced during a DR event. The portion that goes toward the net levelized cost is varies by sector and was assumed to be as follows:[3]

Portion of Incentive Impacting Levelized Cost

In addition, the 2021 Plan assumes a variable incentive of $150 per megawatt-hour. This represents the cost at which DR would be deployed for economic reasons. In other words, if the price for electricity meets or exceeds this amount, DR would be deployed regardless of adequacy needs.

[1] Cadmus developed the Northwest Demand Response Model used for this analysis; an overview of the model can be found here.

[2] More information can be found here

[3] This approach is largely consistent with the California public utility commission guidance.