What Can We Learn From It? End Use Load Research What Is It? What Can We Learn From It? What Do We Know? What Don’t We Know? Tom Eckman – NPCC Jeff Harris - NEEA

NEEA RBSA metering data provided by Ecotope What’s End-Use Data? NEEA RBSA metering data provided by Ecotope

End-Use Load Research What Is It? Basics Short-interval metering Seconds Minutes Hourly Metering End-use (e.g., refrigerator, space heating system) Sub-end use (e.g., ice-maker, heat pump compressor, electric resistance backup) What Gets Metered Watts, volts & amps (aka-True RMS power) Temperatures (interior, ambient, furnace supply and returns, etc.) Other power characteristics (e.g., harmonics)

What Can We Learn From It? Load forecast, especially for capacity planning Forecast of future load shapes as technology mix and efficiency changes Transmission and distribution system planning Forecast of future load and coincidence factors for system design Impact of changing equipment (e.g. less resistance load, electronics) mix on power factor and other system attributes Demand Response Identification of both “Inc” and “Dec” opportunities Energy Efficiency planning Verification of actual kWh and KW reductions Determining cost-effectiveness and priority for development of efficiency measure based on their system load impacts Rate making Cost-of-service cost allocation based on contribution to peak

The Details Importance of End-use Data Load Forecasting – while many of the regional planners effectively use econometric models, a better understanding of end-use loads would provide the opportunity to isolate and differentiate trends associated with specific technologies, determine impacts of forthcoming codes and standards and assess the effects of changes in energy efficiency policy, not captured by econometric trend forecasting; Capacity Planning and Demand Response – the region is facing a capacity challenge where an understanding of end-use load patterns is necessary for designing effective demand response programs to help manage peaks; Integrated Resource Planning – provides necessary detail on the hourly impacts helping to establish the value proposition for the energy efficiency portfolio, or the measures that make up a cost effective portfolio; Grid Operations and Reliability – used to update critical components of regional transmission operating models with current customer load information helping to operate the grid reliably and economically; Wind Integration – understanding of end-use load patterns is a critical element needed to help integrate intermittent renewable resources into the resource mix; Smart Grid Investments – potentially outdated and inaccurate load shapes are still being used to assess the costs and benefits of demand response including the application of smart appliances from a utility and grid perspective; and Energy Efficiency Planning – provides the basis for determining the benefits and costs associated with measures and programs helping to direct energy efficiency investment decisions; Energy Efficiency Evaluation – assigns time differentiated impacts to energy efficiency measures and portfolios helping to improve program designs to become more cost effective at obtaining energy resources; Rates and Pricing – provides information for use in designing customized rates/riders targeted at specific end-uses and provides insight into the likely impacts of time differentiated rates. An example would be the development of an off-peak charging rate for the promotion of electric vehicles; Customer Service – facilitates partnership between electric service providers and customers through enhanced knowledge and understanding of end-use contribution to typical customer usage patterns fostering consumer education, utility response to high bill complaints, rate impact estimation, energy efficiency target marketing and other customer service touch points.

What Do We Know? The Current State of End Use Load Research The Good News - Bonneville conducted a comprehensive End Use Load Research (EULR) project, the End-Use Load and Consumer Assessment Program (ELCAP)* The Bad News – This year marks the 25th anniversary of when the last ELCAP data was collected There have been no comparable studies in the US or Canada since then There are none planned *The ELCAP was later designated as the Regional End-Use Metering Project (REMP)

What Was Collected in ELCAP? ELCAP began in 1983 and ended in 1991 ELCAP Goal – Collect a comprehensive set of customer characteristics and hourly end-use electricity consumption and weather data from residential and commercial sector consumers ELCAP Product - Created hourly (8760) load profiles at the individual end use and building level along with associated customer characteristics and weather data

ELCAP Sample Residential Commercial Detached Single Family (~340 homes) Three space heating system types Three “vintages” 10 major end uses Commercial 9 building types 15+ major end uses

Where’s The ELCAP Data? We Nearly Lost It All! PNNL collected and archived (paper and electronic) all the project data BPA had some of it, but so far as we could find, doesn’t anymore PNNL had no long term plan for preservation after original ELCAP project staff retired In 2012, the Council’s Regional Technical Forum worked with “retiring” PNNL staff to catalog, retrieve, organize, convert to electronic form and make publicly available on the web all aggregate and site level ELCAP data (http://rtf.nwcouncil.org/ELCAP/)

Common Concerns About ELCAP and Existing EULR Data Sources* They’re Old The last ELCAP data was collected in 1989 They Don’t Reflect Current Building and Equipment Stock Appliance characteristics and usage have changed (e.g., increased efficiency standards); There are new emerging/growing technologies (e.g., computers, plasma TVs, Electric Vehicles, variable speed drives, demand controlled loads, LED’s, etc.); New appliances, lighting and equipment are “electronic” and may have differ Power Factors than existing “resistance” based appliance, lighting and equipment They Don’t Provide Energy Savings Load Shapes Energy Efficiency measure savings, especially controls, do not have the same shapes as their end-use load shapes (e.g., occupancy sensors are designed to turn off lights during the day when no one is present) *Results of a RTF Regional and National Review of the Business Case for End Use Load Research

Are These Concerns Warranted? Let’s Compare To Some More Recent Data The Northwest Energy Efficiency Alliance (NEEA) Research Residential Building Stock Assessment “Test Bed” EULR study of approximately 100 single family homes Preliminary (1st year) results are now available Ductless Heat Pump Pilot Program Detailed 5 minute interval data on approximately 100 ductless heat pumps Heat Pump Water Heater Pilot Program Detailed 5 minute interval data on approximately 50 heat pump water heaters While not as comprehensive as ELCAP these studies do provide some insights into what’s changed and what hasn’t, but only in the residential sector

Four Case Studies Electric water heating- Improvements in efficiency since 1990 have reduced annual energy consumption from 4,700 kWh/yr to 3,000 kWh/yr What was the reduction in coincident peak demand? How will future changes in water heating technology (i.e., the conversion to heat pump water heaters) impact coincident peak demands and opportunities for demand response? Residential Refrigerators – Improvements in efficiency since 1990 have reduced annual energy consumption from 1,500 kWh/yr to slightly less than 500 kWh/yr What impact has this had on the options for demand response? Residential space heating – Supplementing existing zonal electric heating systems (e.g. baseboard, radiant ceiling, wall heaters) with Ductless Heat Pumps can reduce annual space heating use by 40-60% What is the potential reduction in coincident peak demand? What impact will the conversion of resistance heating to compressor-based heating have on Power Factor? What impact will it have on “cold-start” after outages? Residential lighting – Federal efficiency standards are changing the mix of lighting technologies from primarily (90%) incandescent lamps to over 60% fluorescent (CFLs) /solid state (LED) lamps. What is the potential reduction in coincident peak demand from this change? What impact will this have on residential Power Factors?

Testing Conventional Wisdom Despite efficiency improvements, residential water heating and refrigerator load profiles are pretty much the same as they were in 1990. Aren’t they? Heat pumps save energy, but have little or no capacity benefit when it gets really cold. Or do they? Residential consumers use lighting pretty much when they did in 1990. Are we sure? Residential “power factors” are very close to 1.0. Do we know that for sure?

PNW System Daily Load Profile Then (1993) and Now (2012) Winter System Peak: 8 – 9 AM 7 – 8 PM Weather Normalized PNW Loads, without DSIs

ELCAP Residential Water Heating Load Shape System Winter Peak Hours

Residential Water Heating Load Profile Then (1990) and Now (2012) System Winter Peak Hours

Residential Water Heating PNW Hourly Demand Then (1990) and Now (2012) System Winter Peak Hour Savings

Capacity Savings From Residential Water Heating Efficiency Improvements Between 1990 and Today Were Three Times Their Annual Energy Savings   1990 2012 Annual Use (kWh) 4,700 3,000 Savings/Unit (kWh) 1,700 Water heater stock <55g 2,701,101 3,489,689 Water heater stock >55g 300,122 337,815 Water heater stock - Total 3,001,223 3,827,504 Annual Load (aMW) 1,610 1,311 PNW 2012 Savings (aMW) 299 Coincident Peak Load (MW) 2,941 2,034 Coincident Peak Savings (MW) 907

Current Water Heating Load Factors Are Higher

As A Result Using Old (ELCAP) Load Profiles Understate the Capacity Impact of Changes in Water Heater Efficiency   1990 2012 – ELCAP Load Shape 2012 – RBSA Load Shape Annual Use (kWh) 4,700 3,000 Savings/Unit (kWh) 1,700 Water heater stock <55g 2,701,101 3,489,689 Water heater stock >55g 300,122 337,815 Water heater stock - Total 3,001,223 3,827,504 Annual Load (aMW) 1,610 1,311 PNW 2012 Savings (aMW) 299 Coincident Peak Load (MW) 2,941 2,370 2,034 Coincident Peak Savings (MW) 572 907

Federal Standards Will Decrease Water Heating Energy Use Further Altering This End Uses Load Shape Winter System Peak Hours

Capacity Impact of Changes in Water Heater Efficiency Due to New Federal Standards   2012 ER 2012 HPWH Annual Use (kWh) 3,000 2,000 Savings/Unit (kWh) 1,000 Water heater stock >55g 337,815 Annual Load (aMW) 116 77 PNW 2012 Savings (aMW) 39 Coincident Peak Load (MW) 174 108 Coincident Peak Savings (MW) 66 Without More Recent End Use Load Research We Would Not Be Able to Estimate the Capacity Impacts of Heat Pump Water Heaters.

Compared to the Stock in 1990 the Current Stock of Refrigerators Uses One-Third the Energy and Requires One-Half the Capacity Winter System Peak Hours Today’s better insulated refrigerators are less sensitive to ambient air temperature

Since Refrigerator Load Profiles Have Not Changed Significantly ELCAP Data Is Still A Reasonable Representation of Capacity Impacts   1990 2012 – ELCAP Load Shape 2012 – RBSA Load Shape Annual Use (kWh) 1,500 500 Savings/Unit (kWh) 1,000 Water heater stock - Total 4,635,877 7,148,904 Annual Load (aMW) 794 408 PNW 2012 Savings (aMW) 386 Coincident Peak Load (MW) 762 390 391 Coincident Peak Savings (MW) 372 371

On the Other Hand, There Are Subtle “Within Hour” Differences That Might Matter for Demand Response Note difference in use following “defrost” cycle Defrost Cycle

Ductless Heat Pumps Have Both Energy and Capacity Benefits

Ductless Heat Pumps Capacity Benefits Even On Extreme Winter Peak Days

Ductless Heat Pump Provide Air Conditioning – So Summer Loads Will Increase Slightly

Potential System Impacts from Ductless Heat Pumps   Electric Zonal Heat Electric Zonal w/DHP Supplement Electric Zonal Extreme Peak Annual Use (kWh) 9,683 6,359 6.2 (KW) 2.7 (KW) Savings/Unit (kWh) 3,324 3.46 (KW) Existing Baseboard Heated Stock 542,600 Annual Load (aMW) 600 394 N/A Potential Savings (aMW) 206 Winter Coincident Peak Load (MW) 1,726 483 3,343 1,466 Winter Coincident Peak Savings (MW) 1,243 1,877 Summer Coincident Peak Load (MW) 43 112 Summer Coincident Peak Savings (MW) (68)

Lighting Energy Use Scenario LPD (W/ft^2) Annual Energy (kWh/yr) ELCAP 3.54 4500 Current Survey (RBSA) 1.40 1845 Full EISA Compliance w/ EISA targets 1.18 1555 Full EISA Compliance w/ CFLs 0.85 1120 Four scenarios 1990 – Lighting load as measured in ELCAP 2012 – Lighting load as measured in RBSA Full EISA compliance with EISA targets assume all currently non-complying, non-exempt lamps are replaced with their minimum compliance equivalents Full EISA compliance with CFLs assumes all non-complying, non-example lamps replaced with CFL equivalents Annual energy use is for 2,006ft2 house with 1.8 hours per day of on-time (Source: RBSA sample house size and lighting on-time metering)

Residential Lighting Load Profiles Lighting Is Now Contributing Significantly Less to Both Morning and Evening Peaks System Winter Peak Hours

Historical and Forecast System Impacts from Residential Lighting Efficiency Improvements   Lighting Loads (ELCAP) - 1990 Stock & Efficiency Lighting Loads (ELCAP load shape, 2012 stock & efficiency) RBSA Lighting Load Shape - 2012 Stock & Efficiency Post-EISA 2020 Lighting Standards Loads (2012 Stock) Annual Use (kWh) 4,500 1,770 1,080 Savings/Unit (kWh) 2,730 690 Single Family Residential Stock 4,021,706 5,798,217 Total Annual Lighting Loads (aMW) 2,066 1,172 715 PNW 2012 Savings (aMW) 894 457 Coincident Peak Load (MW) - Morning 2,080 1,181 1,227 747 Coincident Peak Load (MW) - Evening 2,884 1,638 1,852 1,127 Coincident Peak Load Savings (MW) - Morning 898 852 480 Coincident Peak Load Savings (MW) - Evening 1,246 1,032 725

What’s All This Mean The region’s development of energy efficiency resources also has capacity impacts Our knowledge of the relationship between energy and capacity savings is based on data collected 25 years ago Let’s see just how much difference using updated data for just two end uses makes

2010 -2012 Utility Efficiency Program Savings

January Day Hourly Savings Change with Updated Data

Annual Energy and Winter Capacity Savings from 2010-2012 Utility Efficiency Programs

Residential Power Factors Are They 1.0? In the next two slides, Serv_A is one leg of the service electric supply to the house while Serv_B is the other Each point is the average power factor over a single day at a single site

Power Factor – Service Leg A

Power Factor – Service Leg B

14 January 2013 14 July 2012 14 April 2012 14 October 2012

Are Residential Power Factors Changing? Traditional residential resistance loads (.e.g., lighting, space and water heating) are being replaced with electronics and compressors (heat pumps) 30% of the lamps in homes are already CFLs At least 60% will be CFLs or LEDs with 7-8 years In addition, there is an increasing saturation of consumer electronics 2.3 TVs/household, 1.5 computers/household, 1.6 set top boxes/household, 1.5 game consoles/household

Summary ELCAP load shapes may still be accurate (e.g., refrigerators) or have errors as large as a simple cycle combustion turbine on peak (e.g., water heating, lighting). Unfortunately, without new load research, we don’t know which is the case for those end uses we didn’t measure. New equipment (DHPs and HPWH) not covered by ELCAP may significantly (>1000 MWp) reduce future capacity needs. Power Factor is becoming significant - Residential averages <0.90; as low as 0.60 Don’t know if PF is leading or lagging; could help or hurt at individual feeder level Future will see even more change: more compressors, less resistance loads, more switching power supplies, smaller inductive loads, Federal standards (HPWHs, lighting, HP dryers.

Summary (2) Significant opportunities for additional peak shifting (HPWH’s, DHP’s ,refrigerator defrost) may exist, but can’t be identified or quantified without further EULR New efficiency opportunities largely involve “controls.” These controls are deliberately designed to change the underlying end- use load shape (e.g., occupancy sensors, day lighting controls, variable speed drives). There are no plans to collect data on the impact of these controls The RBSA meters are now scheduled to be removed at the end of March 2014 (next month) There are no plans to collect data comparable to the RBSA in the commercial, industrial or agricultural sectors

Backup Select RBSA Metering Load Shapes and Power Factor

Annual December Daily System Weather Normalized Loads

14 January 2013 14 July 2012 14 April 2012 14 October 2012

Operating profile of two refrigerators, each in a different house, for two days in the summer. Both are located in kitchens. The refrigerator at site 13400 was manufactured in 1995, and the refrigerator at 14674 was manufactured in 2006. Although both refrigerators show a typical cyclic behavior, a demand response opportunity lies in the defrost control. Both pieces of equipment have automatic defrost, which is indicated by the spike in power to 800 W (in blue) and 425 W (in red). They defrost as needed but do not repeat at any obvious frequency. Post-defrost, both compressors run for a longer period of time, presumably to reduce the temperatures after the defrost. A simple control could be conceived to restrict the times when the defrost occurs to off-peak hours, thereby reducing peak demand.

DHP winter peak day (Jan 11, 2011) compared to winter average day. 90+ sites compared

Residential Water Heating Load Profiles Will Continue to Change 2015 Federal Standard for Electric Water Heaters > 55 gal. capacity

Refrigerator Annual Energy Use by Vintage Manufacture Date Bin Primary Refrigerator Annual kWh by Vintage Mean EB n Pre 1994 699.4 99.8 8 1995-1999 734.8 49.6 19 2000-2004 636.2 49.4 31 2005-2009 496.7 29.6 33 Post 2009 520.2 95.9 Total 604.4 24.8 99