ENERGY CONSERVATION THROUGH DEMAND-SIDE MANAGEMENT (DSM): A METHODOLOGY TO CHARACTERIZE ENERGY USE AMONG COMMERCIAL MARKET SEGMENTS

Transcription

1 ENERGY CONSERVATION THROUGH DEMAND-SIDE MANAGEMENT (DSM): A METHODOLOGY TO CHARACTERIZE ENERGY USE AMONG COMMERCIAL MARKET SEGMENTS K.R. Grosskopf, Ph.D., CEM School of Building Construction University of Florida Gainesville, Florida David Barclay Strategic Planning Gainesville Regional Utilities Gainesville, Florida Paul Oppenheim, Ph.D., P.E. School of Building Construction University of Florida Gainesville, Florida ABSTRACT Managing energy demand can be beneficial for both the energy consumer and the energy supplier. By reducing energy use, the consumer reduces operating costs and improves production efficiency and competitiveness. Similarly, the supplier may reduce the need for costly capacity expansion and wholesale power purchasing, especially if energy reductions occur during peak loading conditions. Energy reductions may also lessen global climate change and reduce many other consequences of fossil-fuel energy use. The following research highlights a methodology to characterize energy use and optimize a DSM program for different types of commercial buildings. Utilizing publicly available records, such as utility billing data and property tax records, the diverse commercial building market was characterized. The commercial building types were matched to relevant submarkets of the North American Industry Classification System (NAICS). These sources were combined to prioritize building type submarket energy use intensity (kwh/sf/yr), load factor and many other energy use characteristics for each market segment. From this information, lower tier performers in each NAICS submarket can be identified and appropriate DSM alternatives selected specific to each. INTRODUCTION Billions of dollars are spent each year in the U.S. on energy DSM programs. DSM programs are intended to reduce operations costs for consumers, defer new generation capacity, and reduce environmental impacts. Of 262 utilities selling 75% of all power produced in the U.S., 55% have DSM programs. Combined, these utilities reduced net energy consumption an average of 1.1% and peak demand 5.3% in 2000 alone (DSM, 2000). With average DSM program costs of USD 0.04/kWh, twenty U.S. utilities reported an avoided supply to program cost ratio of for commercial lighting programs (Sonnenblick, 1994). As conservation programs expanded beyond lighting retrofits however, the absence of appropriate program designs and a rationale for supplier incentives resulted in the stagnation of DSM programs during the 1990 s. Following years of growth from the late 1970s, DSM expenditures fell from USD 2.74 billion in 1993 to USD 2.5 billion in In the aftermath of 9/11 however, the U.S. has begun to reassess its dependence on foreign energy supplies and its contribution to global climate change. As a result, energy suppliers are beginning to revisit marketbased DSM programs. This time, utilities are taking a closer look at the energy use characteristics of individual markets in order to develop customized DSM options that maximize energy use reductions and minimize program costs. Through investment in current technologies, it is believed that a 30% reduction in energy use can be cost-effective for the user, while aggressive demand-side management strategies could reduce energy use up to 75%. The following research presents a methodology for characterizing the energy use of a utility s commercial market in order to identify DSM options that will most cost-effectively reduce energy use intensity. The objectives of this effort were i) to stratify the C&I market of a medium-sized (600MW) utility into major NAICS market segments, ii) to merge utility energy use records from with county property appraiser building data, and iii) to compare the energy use intensity (kwh/sf/yr) among NAICS market segments. The purpose of this research is to present a methodology to qualify and quantify the unique energy use characteristics of major NAICS market segments using available utility and property appraiser data in order to develop DSM options that maximize energy use reductions while minimizing program costs. BACKGROUND Population growth, economic development and other environmental issues increasingly compromise the availability of energy resources worldwide. In 2004, U.S. energy use exceeded 29.37x10 9 MWh (100.25x10 15 Btu or quads ), or approximately 22.6% of the world s fossil fuel consumption (DOE, 2004). The U.S. energy efficiency index provided by the Department of Energy (DOE) indicates that the built environment consumes 36% of all U.S. energy resources, with the commercial sector amounting to more than 40% of this demand.

2 6% 5% Warehouse 14% Public Assembly 5% Other Vacant 4% 17% 14% Food Sales 16% Food Health Care 4% 7% Lighting, 23% Other, 7% Ventilation, 3% Refrigeration, 3% Equipment, 6% Cooking, 4% Water Heating, 15% Heating, 3 Cooling, 7% Figure 1. Distribution of U.S. commercial building use Figure 2. Distribution of U.S. commercial building by floor area. (EIA, 2006). energy use. (EIA, 2006). Commercial buildings consist of a variety of building types including offices, retail, service, education, food service and sales, healthcare, government and public assembly, and many other non-residential uses (Figure 1). Each of these markets has unique energy needs, but as a whole, more than half of all energy used by commercial buildings is dedicated for space conditioning and lighting (Figure 2). As a result, lighting and HVAC systems are often targeted first by DSM programs as cost-effective, low hanging fruit. Commercial office space represents nearly a quarter (2) of all non-residential energy use in the U.S. (Figure 3). Weighted by total useable floor area in each market, the average U.S. commercial building consumed 53.7 kwh/sf/yr in Food service and sales are among the most energy intensive, consuming between kWh/sf/yr, followed by healthcare at kwh/sf/yr. Remaining commercial buildings such as office, retail, lodging and institutional spaces consumed between kWh/sf/yr (Figure 4). Energy use intensity in the southeast U.S. however, was roughly half the U.S. average for all markets (EIA, 2006). Healthcare Warehousing Food Public Assembly Food Sales All Other Vacant 1% 9% 8% 8% 7% 6% 6% 6% 5% 5% 13% 2 0% 5% 10% 15% 20% 25% Percentage of total U.S. non-residential building energy use. Figure 3. Percentage of U.S. non-residential energy use by market sector (EIA, 2006).

3 Food Sales Food Healthcare All Other Public Assembly Warehousing Vacant Energy use intensity (kwh/sf/yr) Figure 4. U.S. energy use intensity of non-residential buildings by market sector (EIA, 2006). In addition to building use, it is reasonable to assume that energy use intensity, like many other building attributes, may be affected by building size and age. Economies of scale generally suggest that larger buildings are more cost effective than smaller buildings of similar use. Larger buildings typically have greater useable floor area in relation to envelope area, making them more cost efficient. Likewise, it would be reasonable to assume that newer buildings constructed to more modern building codes would also be more efficient. However, national data shows that large buildings (>100,000sf) are generally more energy use intensive than smaller buildings within the same use designation (Figure 5). Data also shows little or no correlation between age and energy use intensity among non-residential markets (Figure 6) Energy use intensity (kwh/sf/yr) ,001-10,000sf 10, ,000sf Over 100,000sf Health Care Warehouse Figure 5. Influence of building size on energy use intensity by select U.S. markets (EIA, 2006).

4 120.0 Energy use intensity (kwh/sf/yr) or before Food Health Care Public Warehouse Figure 6. Influence of building age on energy use intensity by select U.S. markets (EIA, 2006). METHODOLOGY To develop and test a method for characterizing the energy use intensity of non-residential markets, a total of 13,827 commercial account records, active as of 31 December 2006, were obtained from a mediumsized (611MW) utility located in Gainesville, Florida. Each record included a unique account identification number (PREMISID) associated with a specific location or metering point as well as a customer name and address for customers having one or more electric, natural gas, water, wastewater and/or telecommunication services provided by the utility. NAICS identification numbers and descriptions were matched to each record using utility account information as well as available telephone records and D&B Data Universal Numbering System (DUNS) numbers. The NAICS assignment achieved an 88.1% match rate on a total of 12,188 records. A confidence code ranging from 1 (low) to 10 (high) was assigned to each NAICS match. Accounts that did not achieve NAICS assignments with a confidence code of 6 or greater were eliminated. A total of 9,650 (69.8%) commercial accounts were assigned a NAICS identification number with confidence code of 6 or greater. Of those records successfully matched with a NAICS identification number, inactive accounts that did not have electric use history after 1994 were eliminated, resulting in a total of 7,048 (51.0%) remaining accounts. Records without electric history were also eliminated as such records could not be used to analyze energy use intensity alone since most, if not all buildings within this study were either electric-only or a mix of predominately electric and some gas. Of 7,048 active electric accounts successfully assigned a NAICS identification number, 4,917 (35.6%) were matched by customer name, service address and other account information to a parcel that had a valid county property appraiser tax identification number. Of those, 4,162 (30.1%) accounts matched parcels having summarized total floor area for that parcel. The remaining 4,162 records were distributed across 1,844 unique parcels. Of these, 1,197 (28.8%) were single accounts located on single parcels. The remaining 2,965 (71.) accounts were located on 647 parcels having between 2 and 115 accounts each. Since floor area could not be readily unbundled and allocated to multiple accounts, each having different use, NAICS designation and energy use history, a mixed-use NAICS designation was created to represent the combined floor area and energy use history for these largely strip mall businesses aggregated on a single parcel. A total of 4,162 commercial accounts were sorted by NAICS identification number and grouped according to major (2-3 digit) NAICS market segments (Table 1). The number of account records remaining in each NAICS market segment was compared to the number of pre-filter account records to determine if any NAICS market segment experienced a disproportionately high rate of attrition.

5 Table 1. Major EIA commercial markets (n=12) and equivalent NAICS submarkets (n=37). Next, monthly energy use records dating from January 1994 to December 2006 for each remaining account were summarized by average and maximum bi-annual electrical energy demand (kw), electric energy consumption (kwh), load factor and natural gas energy consumption (therm). Inactive billing periods resulting from building vacancy or transition of ownership (e.g. customer account changes) were identified and excluded from the analysis. Summary energy use data for each commercial account from each bi-annual period from was merged with NAICS data and building floor area data. Since energy use intensity was measured by the total energy use of the building in relation to its floor area, natural gas consumption (if any) was converted from therms to equivalent units of electrical energy (kwh).

6 The converted average annual gas consumption was then added to the average annual electric consumption and the sum divided by total parcel floor area to determine the average annual energy use intensity (kwh/sf/yr) for each account in each NAICS market segment. The average and maximum annual energy use intensity for each NAICS market segment was identified. The standard deviation or scatter among the energy use intensity data within each NAICS market segment was also calculated. In addition, energy use characteristics (kw, kwh, load factor, etc.) and size (sf) of businesses in the upper an lower quartile of average energy use intensity were identified for each NAICS market segment. Finally, the percentage of total energy demand and energy consumption for each NAICS market segment was calculated. NAICS markets having high energy use intensity, significant spread between upper and lower quartile, and those representing an appreciable load for the utility are considered priority markets for demand-side management (DSM). For the purposes of this study, it was assumed that the same (or similar) type of business occupied each building space from 1994 to present, and that no significant changes to building area or composition had been made during this time. CASE STUDY The C&I market of the greater Gainesville area was selected for study under a research grant provided by Gainesville Regional Utilities (GRU) to the University of Florida (UF). Florida currently ranks third nationally in both population (18 million) and energy consumption. Over the next 5-10 years, the state s energy consumption is expected to increase by 30%. Just under half of the state s energy (47%) is consumed by buildings, with a near even split between residential (55%) and commercial (45%) markets. Electricity powers over 90% of the commercial energy demand in Florida, with space cooling comprising 30%-45% of the total energy load. Anchored by the nation s third largest university with a student enrollment just under 60,000, Gainesville is largely comprised of technology, institutional, office, retail, healthcare and service industries. GRU is a multi-service utility generating roughly USD 77.3M in annual revenue primarily from the sale of 2,110.8GWh of electricity and 22.0 million therms of natural gas (Table 2). In spite of its limited size, GRU has been nationally recognized for its leadership, innovation and investment in energy conservation. Table 2. Gainesville Regional Utilities (GRU) 2006 demographics. Utilities Electric, gas, water, wastewater and telecommunications Total net revenue USD 77.3M Electric customers 88,663 (9,538 commercial) Electric sales 2,110.8GWh (955.9GWh commercial) Electric generation capacity 611MW Electric service area 124sq mi, 1,399 circuit miles Gas customers 32,520 Gas sales 22.0 million therms Gas service area 117sq mi, 705 main distribution miles RESULTS Energy use intensities for buildings in the south U.S. are significantly lower than the national average. Southern regions generally have fewer combined heating and cooling degree days. A degree day is defined as a 24-hour period where the ambient (outside) temperature is one degree above a cooling set-point temperature, or, below a heating set-point temperature. Furthermore, cooling loads which dominate in the south are more efficiently met with direct expansion air conditioning technologies that are more than three times more efficient than either electric or gas heating. The energy use intensities of non-residential buildings within the GRU case study region were somewhat consistent with similar market sectors in the south (Figure 7) having an energy use intensity on average of 28.0kWh/sf/yr, approximately 16% lower than the aggregate south average of 33.5kWh/sf/yr. A notable exemption was healthcare. At the time of this study, information on inpatient hospitals within the GRU case study region was not available. It must also be noted that the energy use intensity data for the GRU case study factors energy use from , whereas the EIA data is from 2003 only. In addition, the GRU case study may be affected by the relative absence of an energy intensive industrial sector and a disproportionately high concentration of food sales, service and retail industries supporting the University of Florida.

7 180.0 Energy use intensity (kwh/sf/yr) GRU South US 0.0 Food Sales Health Care* Public Assembly Warehouse Food *non inpatient (e.g. hospitals) Figure 7. Comparison of energy use intensity between U.S., south and GRU case study region. Within each major EIA market, significant energy use patterns were identified for many NAICS sectors or submarkets. Identifying submarkets with disproportionately high energy use intensity is important for an effective DSM program so that conservation efforts can be prioritized. For example, the average energy use intensity for the retail market in the GRU case study is 17.2kWh/sf/yr. However, the energy use intensity for gasoline stations, a subset of the retail market, was 41.6kWh/sf/yr. Part of this disproportionate increase can be accounted for by the fact that gas stations have significant energy loads outside of the building footprint such as lighting, signage and pumps. In addition, gasoline stations are often illuminated well beyond Illumination Engineering Society (IES) standards, often exceeding 100 foot candles (fc) or more in an effort to create a perception of added safety and security. Furthermore, many gasoline stations and convenience stores have significantly greater hours of operation than many other types of retail. For utilities, the loss of electric sales from conservation can in some cases be offset by a cost savings in needed electric generation, transmission and distribution infrastructure. As a result, energy demand intensity is generally of greater importance to energy suppliers than energy use intensity. An important energy demand metric tracked closely by utilities is load factor, which represents the average load or energy demand in relation to the peak demand. A commercial facility having an average load of 200kW and a peak load of 500kW for example, has a load factor of 40%. As a result, the utility must designate 500kW of wires and generation capacity to meet the facility s peak load, even though on average, only 40% of this infrastructure would be utilized at any one time. Building into this equation safety reserves and line losses, as much as 65%-70% of the utility s invested infrastructure dedicated to this facility may sit idle at any given time. As a result, utilities and energy suppliers are generally more inclined to promote DSM options that improve load factor among low performing markets. As shown from the GRU case study, lodging, food service and retail have a reasonably good load factor whereas government, office and educational facilities have a comparatively low load factor (Figure 8). Again, the disparity in load factor could be largely attributed to greater hours of operation among lodging, food service and retail sectors as opposed to 9 to 5 government, office and educational facilities. such as 24- hour health and personal care stores were found to have exceptionally good load factors, many exceeding 70%.

8 Food Health Care* Other Load factor (kwh/kw/720) * non inpatient (e.g. hospitals) Figure 8. Comparison of average market load factors, GRU case study region. Since roughly half of all energy consumed in commercial buildings is used for thermal comfort and lighting, most DSM options implemented in the U.S. are related to heating, ventilation and air conditioning (HVAC) systems, lighting systems and the building envelope (Figure 9). The potential for demand side management (DSM) within any given market is a function of several factors, including energy use intensity and the proportion of energy consumed by the market. In addition, the relative market penetration of DSM within the market must also be considered. The food sales market for example, is among the most energy intensive of C&I markets, averaging 156.9kWh/sf/yr nationally, 55.0kWh/sf/yr in the south, and 72.2 kwh/sf/yr within the GRU case study region. However, food sales represent only1.8% of the C&I building stock by floor area. HVAC prevent maint Elec ballast lighting Insulated glass Tint, reflect or LoE glass Spec reflect lighting Exterior shading HVAC economizer VAV system HVAC EM&C system Daylighting 79.8% 73.4% % % 31.7% 30.4% 26.6% 20.9% Lighting EM&C system Daylight sensors 7.6% 12. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% Percentage of adoption by total floor area, all sectors Figure 9. Adoption rate of commercial DSM alternatives in the U.S. (EIA, 2006).

9 Table 3. Percentage of market by floor area adopting various HVAC, lighting and building envelope conservation features (EIA, 2006). Conservation Features Food Sales Food Health Care Public Warehousing Average HVAC* Variable air volume system 41.0% n/a % 28.6% 10.4% 48.9% 27.5% 49.3% % 13.8% 30.4% Economizer Cycle 43.4% n/a 20.3% 55.6% 35.6% 24.3% 49.3% 35.3% 41.5% 11.5% 17.3% 14.6% 31.7% Preventative maintenance 90.1% % % 72.4% % 90.6% 73.1% 60.5% 67.6% 79.8% Energy mgmt control system 51.4% n/a n/a 32.9% 12.6% 17.1% 39.3% 25.3% n/a n/a n/a 7.9% 26.6% Fenestration and lighting* Insulated Glazing 65.8% 58.9% % 76.4% % 62.6% 71.0% 66.4% 48.1% 43.0% 62. Tinted or LoE Glazing % 43.7% 72.1% 51.8% 54.3% 80.5% 41.7% % 48.3% 50.4% 52.3% Exterior shading 26.6% 37.1% 51.3% 35.3% 38.0% 44.4% 21.1% 27.0% n/a 26.3% 25.6% 20.8% 32.1% Skylights or atriums 24.9% n/a n/a 35.0% % 20.6% 16.5% n/a 11.0% 14.3% 17.0% 20.9% Daylight sensors 4.8% n/a n/a 9.6% 10.8% n/a 5.1% n/a n/a n/a n/a n/a 7.6% Specular reflectors 44.4% 23.3% % 44.4% 28.6% 55.5% 32.3% 38.9% 29.4% 34.0% 35.0% 37. Electronic ballasts 84.6% 74.0% 66.5% 87.5% 81.4% 69.1% 82.9% 65.0% 82.8% 57.9% 66.1% % Energy mgmt control system 11.6% n/a n/a n/a n/a 9.4% 15.7% n/a n/a n/a n/a n/a 12. Average 45.6% 48.5% 46.0% 55.7% 45.4% 36.3% 47.7% 41.6% 60.3% 38.1% 36.8% 33.3% 44.6% Avg. number of options (1-12) * more than one may apply

10 By factoring the food sales average energy use intensity by building stock floor area, food sales represent roughly 6.0% of the C&I market energy use. Of the total C&I building stock, a weighted average energy use intensity of 1.3kWh/sf/yr is dedicated to the food sales market. Nationally, the food sales market ranks third behind government and healthcare among major markets in the average number of DSM alternatives implemented beyond minimum building energy code requirements with 5.8 (48.5%) options implemented of 12 options surveyed (Table 3). By factoring the relative energy intensity, building stock and subsequent energy consumption of the food sales market (1.3kWh/sf/yr) by the potential for further DSM adoption (48.5% -1 ), a dimensionless DSM prioritization score of 0.7 is calculated for the food sales market (Figure 10). In contrast to the food sales market, the retail market has a comparatively low energy use intensity of 50.6kWh/sf/yr nationally, 25.5kWh/sf/yr in the south, and 17.2kWh/sf/yr within the GRU case study region. However, retail represents 20.1% of the C&I building stock by floor area, and subsequently, 15.8% of the C&I market energy use. Of the total C&I building stock, a weighted average energy use intensity of 3.4kWh/sf/yr is dedicated to the retail market. Nationally, the retail market ranks next to last among major markets in the average number of DSM alternatives implemented beyond minimum building energy code requirements with 4.4 (36.7%) options implemented of 12 options surveyed (Table 3). By factoring the relative energy intensity, building stock and subsequent energy consumption of the retail market (3.4kWh/sf/yr) by the potential for further DSM adoption (36.7% -1 ), a dimensionless DSM prioritization score of 2.2 is calculated for the retail market (Figure 10). Although the food sales market is considerably more energy intensive, retail comprises a much greater share of the total C&I energy market. In addition, DSM market penetration within the retail market is comparatively less than that of the food sales market nationally. Considering only these energy use characteristics, the potential for DSM adoption within the retail market is likely greater than that in the food sales market. Food Health Care Food Sales Public Assembly Warehouse DSM potential Figure 10. DSM market prioritization, GRU case study region. CONCLUSIONS The following study shows how existing databases maintained by utilities, county property tax assessors and other public service entities can be used to determine energy use intensity by region, building use, type of construction, age and many other market, climate and building characteristics. As a result, government agencies, utilities and energy service companies (ESCOs) can tailor conservation alternatives to specific markets, and thereby maximize the environmental impact and costeffectiveness of a demand-side management (DSM) program. The use of existing datasets however, may present limitations. Utility information systems and property tax assessor records are not generally designed to be merged. In fact, as shown in this

11 study, the migration from energy use history to building use, size and other characteristics required the merger of several databases. Minor discrepancies or deviations in data fields from one database to another resulted in the data being mismatched or eliminated. In addition, the integrity of data from records compiled over many years invariably introduces error. Markets with extended hours of operation or those with disproportionately high exterior energy loads and relatively little floor area (e.g. gas stations, car sales lots, etc.) will invariably show high energy use intensity. For these and other reasons, caution should be exercised when comparing energy use intensities between markets. In addition, REFERENCES Demand-Side Management Fact Sheet (2000). Gainesville Regional Utilities (GRU) Alternatives for Meeting Gainesville s Electric Requirements through 2022: Base Studies and Preliminary Findings. Gainesville, pp. F-11. National Renewable Energy Laboratory (NREL) Issues and Methods in Incorporating Environmental Externalities into the Integrated Resource Planning Process. U.S. Department of Energy, Washington, D.C. National Renewable Energy Laboratory (NREL) FY 2001 Sustainability Report. U.S. Department of Energy (DOE), Washington, D.C. Regan, E., et al Gainesville Regional Utilities Long Term Electric Supply Plan, Commission Report, Gainesville. Sonnenblick, R Is Demand-Side Management Economically Justified? CBS Newsletter, Summer, energy use intensities within a market can vary. Although within the same market, healthcare administration and non-emergency walk-in clinics for example, have dramatically different energy use characteristics and hours of operation than inpatient healthcare and hospitals. Changes in building use over time, especially within commercial lease space can also contribute error. However, as in survey research, an adequately sized and representative sample population can be used to control error. With more than 30% of the GRU commercial market screened and qualified for this study, error is estimated to be less than +/-5% at 95% or greater confidence. U.S. Census Bureau Annual Value of Private Construction Put in Place. Manufacturing, Mining and Construction Statistics. U.S. Census Bureau Profile of Selected Housing Characteristics: Publication DP-4, U.S. Department of Energy (DOE) Environmental Externalities: Case Studies. Energy Information Administration, of Coal, Nuclear and Alternate Fuels, Washington D.C. U.S. Department of Energy (DOE) Official Energy Statistics. Energy Information Administration. U.S. Energy Information Administration (EIA) U.S. Energy Use Intensity. U.S. Department of Energy (DOE), Washington D.C. U.S. Energy Information Administration (EIA) U.S. Electric Utility Demand-Side Management: Trends and Analysis. U.S. Department of Energy (DOE), Washington D.C., pp