HomeMy WebLinkAboutL.I. Without Shoreham Power Plant - 1983 Technical Report A83-14/A
LONG ISLAND WITHOUT THE SHOREHAM POWER PLANT:
ELECTRICITY COST AND SYSTEM PLANNING CONSEQUENCES
TECHNICAL REPORT A
LONG RANGE FORECAST OF ELECTRICITY REQUIREMENTS
IN THE LILCO SERVICE AREA
July, 1983
ENERGY SYSTEMS RESEARCH GROUP,
120 Milk Street
Boston, MA 02109
(617) 426-5844
INC.
4
TABLE OF CONTENTS
LIST OF TABLES
LIST OF FIGURES
INTRODUCTION
1.1 Background
1.2 Forecasts of'E~e~g~ ~e~u~r~m~n~s'a~d'P~a~ 6e~a~d~
1.3 End-Use Approach . - . .
1.4 Comparison of Base Case and LILCO Forecasts .
2. OVERVIEW OF FORECASTING APPROACH
3. RESIDENTIAL SECTOR . . 3.1 Number of Units
3.2 Saturation Curves
3.3 End-Use Submodels
3.3.1
3.3.2
3.3.3
3.3.4
3.3.5
3.3.6
3.3.7
3.3.8
3.3.9
3.3.10
3.3.11
Refrigerators and Freezers
Electric Ranges
Lighting .
Television
Clothes Dryers
Clothes Washer and DiShWaSher
Electric Water Heaters
Air Conditioners
Electric Space He~t~n~
Heating Auxiliaries
Miscellaneous Applian~e~
COMMERCIAL SECTOR
4.1 Model Structure .
4.2 Commercial Floorspace
4.2.1 1975 Floorspace .
4.2.2 Floorspace Growth ndices"''
4.3 Electric Energy Intensities .
4.3.1 1975 Intensities
4.3.2 Future Intensities .
4.4 Energy Forecast .
5. INDUSTRIAL SECTOR
5.1 Model Structure .
5.2 Base Year ExperienCe''
5.3 Employment Growth .
5.5 Fraction Self-Generated
5.6 Energy Forecast
6. OTHER ENERGY REQUIREMENTS
7. PEAK POWER
- i -
Page
iv
1
1
1
2
5
13
18
19
19
2O
20
24
26
28
30
31
34
36
38
42
44
45
45
45
5O
50
52
52
52
56
57
57
6O
60
60
61
61
63
64
TABLE OF CONTENTS
(Continued)
Page
MODEL INPUTS AND FORECAST ASSUMPTIONS .
8.1 Energy Sales By Sector .
8.2 Residential Sector
8.2.1 Residential Customer Forecast .
8.2.2 Appliance Saturation and Unit Usage
8.2.3 Electric Space Heat and Water-Heating ~e~e~ritioAs
8.2.4 Thermal Integrity Impacts
8.2.5 Additional Data Requirements
8.2.6 Appliance Lifetimes .
8.3 Commercial Sector Data Inputs
8.3.1 1975 Floorspace .
8.3.2 Floorspace Growth' - ' ' ' 'IndiCeS,'High Case
8.3.3 Floorspace Growth Indices, Low Case
8.3.4 Electric Intensities and Saturation
8.3.5 Future Commercial Intensities and Saturation
8.4 Industrial Data Inputs .
8.4.1 Base Year Experience · ·
8.4.2 Employment Growth . . ~ .
8.4.3 Electric Energy In%enslty .....
8.4.4 Fraction of Electricity Self-Generated
8.5 Other Energy Requirements · .
8.6 Peak Power Model . . · ·
8.6.1 Peak Load Data . .
8.6.2 Load Management Impact
68
68
69
69
72
76
REFERENCES
APPENDIX A: ESRG High and Low Case Forecasts of Long Island
Lighting Company
76
78
82
83
83
84
85
86
87
89
89
90
91
94
94
95
95
96
97-100
- ii -
E S R G
Table No.
1.1
1.2
1.3
1.4
1.5
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
4.1
4.2
4.3
4.4
4.5
4.6
5.1
5.2
6.1
7.1
7.2
8.1
8.2
8.3
8.4
8.5
8.6
8.7
8.8
8.9
LIST OF TABLES
Page
FORECAST OF ENERGY AND PEAR DEMAND 3
DISAGGREGATED FORECAST BY SUBSECTOR COMPONENTS - LILCO. 4
FORECAST AVERAGE ANNUAL GROWTH RATES FOR LILCO 5
COMPARISON OF 1983 ESRG BASE CASE AND LILCO FORECASTS 9
COMPARISON OF 1981 APPLIANCE SATURATION LEVELS WITH
LILCO AND ESRG PROJECTIONS . 11
RESIDENTIAL END-USE SUBMODELS
SUBMODEL FOR REFRIGERATORS AND FREEZERS
SUBMODEL
SUBMODEL
SUBMODEL
SUBMODEL
SUBMODEL
SUBMODEL
SUBMODEL
SUBMODEL
SUBMODEL
SUBMODEL
FOR ELECTRIC RANGES
FOR LIGHTING
FOR TELEVISIONS
FOR CLOTHES DRYER
FOR CLOTHES WASHER AND DISHWSHER
FOR ELECTRIC WATER HEATER
FOR AIR CONDITIONERS
FOR ELECTRIC SPACE HEATERS
FOR HEATING AUXILIARIES
FOR MISCELLANEOUS APPLIANCES .
18
23-24
25
27
29
3O
32-33
35
37
39-41
43
44
COMMERCIAL MODEL END-USE,
COMMERCIAL CATEGORY
COMMERCIAL MODEL - FLOORSPACE .
SQUARE FOOTAGE MULTIPLIERS
ELECTRIC ENERGY INTENSITIES .
FRACTION OF LOAD SAVED
COMMERCIAL ENERGY FORECAST
BUILDING TYPES AND
46
48-49
51
54
55
56
STANDARD INDUSTRIAL CLASSIFICATIONS
INDUSTRIAL ENERGY FORECAST
58
62
OTHER ENERGY 63
PEAK POWER MODEL
END-USES USED IN PEAR POWER MODEL
66
67
ADJUSTED LILCO SALES BY SECTOR, 1982 69
COMPARISON OF POPULATION PROJECTION SOURCES, LONG
ISLAND 71
RESIDENTIAL CUSTOMER FORECAST, LONG ISLAND LIGHTING
COMPANY .... . 72
LILCO APPLIANCE SATURATION ASSUMPTIONS 73
RESIDENTIAL APPLIANCE USAGE AND EFFICIENCY IMPROVEMENTS 75
LILCO ELECTRIC SPACE HEAT PENETRATION 76
THE~4AL INTEGRITY CHARACTERISTICS, LILCO AIR
CONDITIONING . . 78
APPLIANCE LIFETIMES IN YEARS 83
1975 COMMERCIAL FLOORSPACE, LILCO SERVICE AREA 84
- iii
Table No.
8.10
8.11
8.12
8.13
8.14
8.15
8.16
8.17
8.18
8.19
8.20
8.21
Figure No.
1.1
1.2
2.1
2.2
2.3
3.1
4.1
5.1
8.1
LIST OF TABLES
(Continued)
HIGH CASE COMMERCIAL GROWTH INDICES, LONG ISLAND
LOW CASE COMMERCIAL GROWTH INDICES, LONG ISLAND
COMMERCIAL ELECTRIC USE COEFFICIENTS, LILCO
YEARS PAYBACK FOR 50% ACCEPTANCE .
FUTURE ENERGY PRICE ASSUMPTIONS (COMMERCI~L'SEC~O~)
LOW CASE .
LOW CASE CONSERVATION'=vEL'PENETRATIONi'-----' LI=O
LILCO INDUSTRIAL SALES MIX .
LILCO EMPLOYMENT GROWTH BY STANDARD INDUSTRIA~
CLASSIFICATION .
ELECTRICAL INTENS~T~ REGRESSION RESULTS, NEW YORK
PRODUCTIVITY TREND RESULTS, NEW YORE 1982-1990 .
1990 PROJECTED STATE ENERGY INTENSITY INDEX, NEW YORK.
PEAK MULTIPLIER FACTORS (PKFAC) - LILCO
LIST OF FIGURES
COMPARISON OF HISTORIC CHANGES IN 1990 PEAK FORECAST .
COMPARISON OF 1983 LILCO AND ESRG PEAK FORECASTS
AND GROWTH TREND IN PEAK
MODEL COMPONENTS
ENERGY FORECASTING MODEL SCHEMATIC
FORECAST SCENARIOS .
SCHEMATIC OF YEARLY ENERGY INCREMENTS BY END-USE
COMMERCIAL SECTOR MODEL SCHEMATIC
INDUSTRIAL SECTOR MODEL SCHEMATIC . .
GROWTH OF NASSAU AND SUFFOLK COUNTIES FROM 1950 TO
1982 .
- iv -
£ S R G
Page
84
85
86
87
88
88
89
90
92
93
94
96
Page
6
8
14
15
17
21
47
59
7O
1. INTRODUCTION
1.1 Background
This report presents ESRG's long-range forecast of electric
energy requirements and peak demand of the Long Island Lighting
Company (LILCO) service area. The report includes a full
description of the methods, assumptions, and results of the
forecast. This research was sponsored by the County Executive's
Office of Suffolk County, New York, as part of a series of five
documents examining the economic impacts of excluding the
Shoreham nuclear power plant from operation.*
The forecast of electricity demands for the LILCO service
territory is based on the best presently available utility, local
and national research data: the demographic, economic, and
energy use-related characteristics of the service territory and
state. The forecast is intended as a business-as-usual scenario,
with input assumptions that represent no substantial departure
from currently foreseeable trends in the forces which influence
electricity use characteristics in the region. A realistic "Base
Case" forecast for LILCO is formed by creating High Case and Low
Case forecasts reflecting the range of uncertainty in the
forecast assumptions and input data. From these, the midrange or
Base Case forecast is developed.
The remainder of this chapter presents the forecast results.
Chapters 2-7 describe the nature and details of the end-use
forecast model that is used by ESRG. Chapter 8 provides the
principal data inputs underlying this forecast for LILCO.
1.2 Forecasts of Energy Requirements and Peak Demands
Forecasts made by other electric utility companies as well as
Long Island Lighting Company have tended to decline in response
to the nationally experienced slow-down in growth since 1973,
changing expectations regarding the potential for energy
conservation, the increasing cost of electricity, and the
likelihood of slower economic and demographic growth. For the
case of LILCO, the downward adjustment in long-range load growth
is displayed in ]Figure 1.1 (along with ESRG's earlier forecasts).
The present forecast again projects somewhat lower growth rates
in energy and peak than does LILCO currently (Ref. 1)o**
* Other volumes include Long Island without the Shoreham Power
Plant: Electricity Cost and System Plannxn~ Consequences,
Surmuary of Findings; Technical Report B: Shoreham Operations
and Costs~ Technical Report C.' The Conservation Invest-
ment Option, and Technical Report D: Computer Output.
** A review and critique of LILCO's methods and assumptions is
offered in Section 1.4.
- 1 -
Base Case Forecasts for LILCO are presented in Tables 1.1
and 1.2. In these tables energy requirements in gigawatt-hours
(millions of kwh) and su~mler and winter peak demands in megawatts
(thousands of kw) are presented for the total system (Table 1.1) and
also on a disaggregated basis (Table 1.2) by customer class --
residential, commercial,, industrial and other (including losses) --
for the base year 1982, as well as for each year through the year 2000.
As described in Section 2 on the ESRG forecasting approach, the
Base Case forecast is formed from High and Low Case scenarios
embodying alternative assumption sets. Results of the High and Low
Case forecasts are set forth in Appendix A. Table 1.3 translates the
forecast data in Table 1.1 (and Appendix A) into average annual growth
rates for the ten year period 1982-1992 and the eighteen year period
of 1982 to 2000. The Base Case for the latter period yields a 1.03
percent growth rate in energy and .82 percent growth in su~muner peak.
1.3 The End-Use Approach
Electricity consumption at the system level is a composite
of the myriad end-use demands serviced by the given utility. The
end-use configurations vary within and between the major consuming
sectors (residential, comercial, institutional, transportation,
manufacturing, etc.). The model employed here is based on the
conviction that system requirements can best be understood and the
impact of the various factors driving growth best be computed if the
model itself is a composite of submodels for the major subsectors and,
to the extent possible, end-uses.
This approach is sometimes called the engineering end-use
approach to the extent that it identifies the actual physical
energy-using stock of equipment. It models its energy requirements
characteristics as consistently as possible given the data
constraints. In this way, the effects of definite changes in
equipment efficiencies, fuel mix, manufacturing processes, etc., are
tracked at the point where they actually occur. Further, the
disaggregated approach enables the user to flexibly and eclectically
draw from the best subsidiary data sources available, be they
optimization studies, market penetration analyses, or national and
state economic and demographic projections. For example, the effects
of a specific policy option, such as improved appliance efficiencies,
can be accounted for directly where they impact through appliance
submodels incorporating retirement schedules, vintages, and technology
changes. Econometrically measured influences on consumer behavior can
also be incorporated in the disaggregated framework.
- 2
~. S R G
TABLE 1.1
FORECAST OF ENERGY AND PEAK D~4AND
LILCO
SA~E
ENEROY IH OWH PEAK POWER LOAD IN MW
LILC083 RESIDENT. COMHER. INDUSTR. OTHER TOTAL SUHHER WINTER
1982 5574. 5340. 1205. 1637. 13757° 3070. 2471o
1983 5640° 5360. 1240° 1650° 13700° 3100o 2500.
1984 5700, 5380° 1280. 1670, 14040° 3120.
1985 5760, 5410, 1320o 1680. 14170° 3140, 2560.
1986 5800, 5480. 1350, 1700. 14330. 3160, 2600,
1987 5840° 5550. 1380, 1720, 14500. 3190. 2640.
1988 5880, 5620. 1420° 1740, 14660° 3220, 2680,
1989 5920° 5690. 1450. 1760. 14820, 3240. 2710o
1990 5960. 5760. 1480. 1770. 14980. 3270. 2750.
1991 5990. 5830. 1510. 1790. 25130. 3300, 2790.
2992 6020. 5910. 1540, 1810, 15280, 3320, 2820,
1993 6040. . 5980, 1580, 1830, 15430, 3350. 2860.
1994 6060. 6060, 1610, 1850, 15570, 3380, 2890.
1995 6080. 6130o 1640° 1870. 15720, 3410, 2?30.
1996 6100. 6210, 1670. 1890. 15870. 3430. 2960,
1997 6130. 6280. 1700. 1910. 16030. 3460. 3000.
2998 6160, 6360, 1740. 1930, 16190, 3490. 3030.
1999 6200, 6440. 1770. 1960, 16360, 3520. 3070.
2000 6230, 6520, 1800, 1980, 16530* 3560, 3110,
- 3 -
E S R G
TABLE 1.2
DISAGGREGATED FORECAST BY SUBSECTOR CO.M~P_0~E~TS
LILCO
LILC083 BASE CASE
i REFRIGERATORS
2 FREEZERS
RANGES
4 LIGHFING
5 TELEVISIONS
6 CLOTHES ORYERS
7 CLOTHES WASHERS
8 DISH WASHERS
9 WATER HEATERS
10 RSOH A/C
11 CENTRAL. A/C
12 SPACE HEAl'ERS
HEATINGAUXILIARY
MISCELLANEOUS
RESIDENTIAL SECIOR - E
1982
1339
268
82'2
36<5
424
62
290
342
29 ]
254
S59
1985 1988 1991
1348, 1332, 1295,
307, ~09,
276, 283, 289,
841. 858.
369. 375, S85,
441, 457°
64. 65. 67,
1~9, 1~¥. 13¥,
SOS. 307,
342, 387, SSS.
~()~. ~5/, 396,
354, 350, 545·
351, 384, 418·
NERJfY IN GWH
1994 1997
1236. 1175
300. 29l
295.
882 889
396 409
48~ 508
68 70
139 140.
315 321,
$81
436 472,
452 487.
LILC083
1: OFFICES
1: HEATING
' COOL.£NG
3: L. IGH] ING
4: AUX ~ POWER
21 RETAIL
1; HEATING
~,°' COOLING
3: LIGHTING
4: AUX ~ POWER
3: HOSPITALS
1: HEATING
2: CGOI.[NG
3: LIGHTING
4~ AUX ~ POWER
4~ SCHOOLS
1: HEATING
2: COOLING
3: LIGHTING
4: Al.IX & POWER
5 OTHER
I HEATING
2 COOL. ING
3 LIGHTING
4 AUX & POWER
2000
1151.
280.
306.
894,
423,
518.
142~
328.
329~
427·
507.
524.
BASE CASE - COMMERCIAL SECTOR FNER(½Y 1N (~WH
1982 1985 1988 1991 1994 1997 2000
49, 61, 73, 8.5, 98. 11~ · 124.
466, 461, 476, 491, 507. 528. 538,
387, 399. 421, 444. 468. 498~
23, 28, ~S, ~9, 44, 50. 56
330, 3S7. 348, 360, 371. SDS. ,~95.
1237, ]260, 1294, 1328, 1362, 1396, 14S1.
5, 6, 7, 8, 9, 10,
58, 51 · 51, 51, 51, 51, 51
146, J45, 146, 147, 148, 148, 149,
79, 79. 82, 85, 88, 90. 93.
16, 17, 21 , 24, 28, 31 .
76, 66, 67. 68, 69, 70.
~7, 277. 278, 278, 279, 279, 280.
187. 164, 169, 178, ~.77, 181, 185.
22, 27, 32, 38, 4~, 49, 56.
249, 255, 262, 268, 27b, 282. 289,
553, 581, 604, 627, 649. 672, 695,
i~64, ~89, 415, 442, 471, 502, 534.
LI[.C083 BASE CASE
INBLISTRIAI. SEC:10R - E
1982 1985 1988 1991
58, 65, 7], 77,
21, 22, 21. 21,
16, 18, 20. 22,
7. 9, 11. iS,
15, 16o i6, 17,
4/. 53, 57. 62.
90, 1i6, 142. ~69,
58. 59. 59· 59
5, 6. 6, 7
40, 48, 45, 48
73, 76, 78, 80
255, 280, ~OO, 1521
176. 184, 189. 19t
56, 66, 75, 84
3. 4. 4. 4
11. J2. 12. IS
118, 129. 137, L47
NERGY IN GWH
1994 1997
82 88,
21 20.
24 26.
15 17,
17 ]8,
66 70,
]96 224.
58 58.
8
50 52.
82, 84.
15~, 160.
342. 565.
192. 192.
94. 105.
13, i4,
156, 166.
S4,
20 F0OD
22 TEXTILES
23 APPAREL
24 LUMBER
25: FURNITURE
26: PAPER PRODUCTS
27: PRINTING ~ F'UBL,
28: CHEMICALS
PETROLEUM ~ COAk
33: PRIMARY METALS
34: FABRICAT. H~TALS
MACHINERY
ELECTRIC EQUIP.
TRANSPORFAI'ION
RUBBER ~ PLASTIC
LEAFHER
32t STONE,CLAY,GLASS
INSTROMENTS
OTHER
- 4 -
G
E
2000
19.
28,
19,
18,
74,
252.
57.
10.
54,
86,
167,
388,
190,
6.
177.
~4.
TABLE 1.3
FORECAST AVERAGE ANNUAL GROWTH RATES FOR LILCO (%/YEAR)
ENERGY PEAK
YEARS HIGH BASE LOW HIGH BASE LOW
1982-1992 1.91 1.06 .13 1.69 .82 -.16
1982-2000 1.81 1.03 .12 1.61 .83 -.07
1.4 Comparison of Base Case and LILCO Forecasts
Earlier versions of the ESRG disaggregated/end-use forecasting
model have been employed in producing Base Case forecasts of
electricity demand and energy in the LILCO service area since
1977. ESRG enjoys an excellent track record for its stable and
high confidence forecasts both on Long Island and in scores of
the other applications. These forecasts anticipated the
necessity for LILCO to radically revise its own long-range
forecasts during the late 1970s and early 1980s. The historic
pattern is illustrated in Figure 1.1 which shows the precipitous
drop in the LILCO forecast of the 1990 summer peak after 1974.
Earlier ESRG forecasts are also shown.
The dramatic decrease in LILCO's load growth forecasts (and
corresponding adjustments in plans for new power plant
construction) can be traced to several factors. First, the
slowdown in population and in economic growth prospects for Long
Island from the post-World War II boom levels was gradually
incorporated into the forecasts. The levelling off pattern in
population growth during the 1950-1982 period is shown in Figure
8.1.
Second, the changes in energy consumption patterns and
conservation initiatives ushered in by the 1973 oil embargo and
energy price jolts, as well as the changed regulatory and policy
context, were eventually recognized as alterations to the
long-term energy planning environment, rather than as temporary
aberrations.
Third, serious methodological and conceptual pitfalls in
LILCO's forecasting apparatus were identified. These were
largely corrected.
- 5 -
Figure 1.1
COMPARISON OF HISTORIC CHANGES IN 1990 PEAK FORECAST
0
0
,<
0
8000
7000
6000
5000
4000
3000
2000
1000
0
ESRG
YEAR OF FORECAST
- 6 -
E S R G
Despite the corrective alterations in their earlier forecast
procedures, to this day the Company retains a tendency to
forecast with unwarranted optimism. Figure 1.2 shows the Company
summer peak forecast to the end of the century.* The average
annual growth rate is 1.6 percent per year. The actually
experienced summer peak over the 1973-1982 period is also shown
for contrast (the growth rate was 0.5 percent per year), along
with a time trend based on that historic experience. Finally,
the forecast developed for the current invesigation (0.8 percent
per year) is shown on Figure 1.2. The comparisons are shown in
more detail in Table 1.4.
Analysis of the assumptions in LILCO's latest forecast
reveals that the higher LILCO forecast is traceable to certain
judgemental rapid-growth inputs and some continuing problems
with self-consistency in their methods. These are identified
below.
* The Company current long-range forecast is contained in the
1983 Report of Member Systems of the New York Power Pool
(Vol. 1) submitted to the New York State Energy Office (Ref. 1).
- 7 -
E $ R G
Figure 1.2
COMPARISON OF 1983 LILCO AND ESRG PEAK FORECASTS
AND GROWTH TREND IN PEAK
4000
3500
3000
2500
LILCO
ESRG
TREND
YEAR
NORMALIZED SU~ER PEAK EXPERIENCED (MW)
- 8 -
TABLE 1.4
COMPARISON OF 1983 ESRG BASE CASE AND LILCO FORECASTS
SALES (GWH) TOTAL PEAK
ENERGY DEMAND
RES. COMM. IND. (GWH) (MW)
ESRG 1982 5,574 5,340 1,205 13,757 3,070*
2000 6,230 6,520 1,800 16,530 3,560
% Growth Rate 0.6 1.1 2.3 1.0 0.8
LILCO 1982 5,556 5,216 1,303 13,713 3,070*
1999 7,427 7,013 2,063 18,748 4,015
% Growth Rate 1.7 1.8 2.7 1.9 1.6
*Normalized from the experienced value of 3,045 MW for weather
and time-of-peak effects.
Residential Customers: The Base Case forecast adopts the Company
projections of residential customers (p. 128).* However, the
LILCO forecast does not appear to analyze energy requirements in
terms of the different levels of usage for single family and
multifamily housing. But the evidence is that smaller household
units, (primarily multifamily units) are gaining a larger share
of the Long Island region's new housing market. These homes are
characterized by lower average usage levels for such major
appliances as electric space heating, central air conditioning,
water heating, and refrigeration. LILCO acknowledges that
smaller households are probable (p. 122) but fails to analyze the
effect of such decreases in household size on usage levels.
*Page references are to Ref. 1 throughout this subsection.
- 9 -
E $ R G
Appliance Saturations: Company projections of appliance satura-
tions (total units as a percentage of total households) in the
residential sector are developed by judgement. The 1999
projections (p. 134) are compared to ESRG estimates in Table
1.5.* It will be noted that the most significant deviations are
in the end-uses of ranges, TVs, dishwashers, room air
conditioners, electric space heaters and electric water heaters.
For perspective, simple time trending (based on regression
analysis on post-1972 saturation data in the LILCO service area)
to the year 2000 gives saturation results that are close to or
below the Base Case values: 42.8 for ranges, 259.0 for TVs, 40.7
for dishwashers, 105.3 for room A/C, 6.8 for ESH, and 13.0 for
EWH.
LILCO's growth assumptions are particularly questionable in
the energy intensive end-uses of electric space heating, air
conditioning, and electric water heating. The number of electric
heating customers is assumed to grow much faster than historic
trends would indicate. In LILCO's view, by 1990 and thereafter,
virtually all new customers will have electrically space heated
homes. Further, by 1999, new electric space heat customers will
exceed the number of new customers due to switching from oil or
gas heated homes to electricity. Similarly, electric water heat
is projected to grow from a 1981 saturation of 8.3% to a 1999
level of over 29%. The implication of such a strong saturation
shift, a penetration rate of 183 percent, is that all new
customers will instal], electric water heating during the forecast
period, while others will switch thereto.
As reflected in ESRG's projections, electric heat pump
technologies, with their year-round application, should
contribute to a strong electric space heat market.* However, the
drastic shift from historic saturations assumed by LILCO are
particularly questionable during a period when electricity prices
will increase substantially while fossil fuel price increases
have moderated. The same questionability applies to electric
water heating growth assumptions.
The assumed growth in central air conditioning (including
heat pump) saturations also results in over 100 percent penetration
levels, this in addition to assumed growth in room air-
conditioners. Growth in the heat pump, solar, and storage
technologies, as well as in traditional appliances growth, is a key
element in the LILCO forecast. But experience to date does not
warrant such optimistic assumptions of accelerating growth. ESRG
uses more cautious Base Case projections.*
*For the basis of ESRG's assumptions and projections, see Section 8.
- 10 -
E S R G
TABLE 1.5
COMPARISON OF 1981 APPLIANCE SATURATION LEVELS WITH
LILCO AND ESRG PROJECTIONS
SATURATIONS (Percent)
LILCO ESRG
1981 1999 2000
APPLIANCE ACTUAL PROJECTED PROJECTED
Refrigerators 113.0 121.5 117.1
Freezers 30.5 35.2 33.0
Ranges 47.1 57.2 49.5
Televisions 220.3 336.9 245.6
Clothes Washers 87.8 90.2 87.4
Clothes Dryers 55.5 64.4 58.3
Dishwashers 52.2 61.9 52.8
Room Air Conditioners 112.2 128.0 120.0
Central Air Conditioners 13.0 21.1 21.8
Electric Space Heaters 3.4 11.3 7.9
Electric Water Heaters 8.3 29.2 13.0
Space Heating Customer Usage: Growth in the residential sector's
energy requirements may be the result of an increase in customers
(households), numbers of appliances per household, or usage
levels per appliance. LILCO's forecast methodology includes a
short-term (five year) econometric forecast wed to a long-term
end-use forecast ,[cf. pp. 121, 122). In both methodologies,
customers and use per customer are projected separately, with the
forecasts of annual sectoral energy resulting from the product of
the two projections.
An unexplained discontinuity in use per customer occurs in
the transition from the short-term forecast to the long-term
forecast of "residential space and water" customers, resulting in
an illogical leap in projected residential sales. In comparing
the short-term and long-term sales forecasts (pp.131 and 143,
respectively), a number of anomalies arise. The actual annual
use per space heating customer in 1982 was 14,431 KWH (Ref. 17).
The short range forecast produces a 1982 level of 15,271 KWH,
growing to 16,222 KWH in 1983, then decreasing to 15,635 KWH by
1986. These fluctuations are not accounted for. In 1987, the
first year of the long-range forecast, this usage level jumps
over 2,500 KWH to 18,198 KWH/year. Taken in conjunction with the
assumption of unprecedented increases in the penetration of space
heating customers (discussed in the previous section), this usage
results in a forecast growing in moderate increments from 1983 to
1986, then jumping sharply in 1987, and growing at an accelerated
pace thereafter. The mathematical methods employed are producing
unreasonable results.
- 11 -
E S R G
Thermal Integrity Improvements: The LILCO forecast does not
directly adjust space conditioning usages to account for improved
building practices and insulation of building stock. There is
evidence that new homes are being built to considerably higher
levels of thermal integrity than that found in the existing
housing stock. These homes require lesser amounts of energy for
heating and cooling. Additionally, the use of energy management
systems in the commercial and industrial sectors may result in
even further reductions in building energy use. Such improue-
ments are incorporated in the ESRG forecast.*
Commercial and Industrial Sector Energy Forecasts: The divergence
between the Base Case and LILCO forecasts is less than in the
residential sector. However, there are a number of areas for
concern. The Company uses a statistical approach which attempts to
explain consumption by commercial subsectors as a function of
employment growth and, in some cases, electricity price growth
rates. However, in important sectors that account for 70 percent
of LILCO's commercial and industrial growth (finance, insurance,
real estate, manufacturing, and government), the price of
electricity has not been utilized in projecting future changes in
electricity demand. This is implausible, and builds in the rapid
growth dynamic embodied in the data base time series (which
begins in 1966), unmoderated for the forecast electric price
Sharp increases in electricity prices are likely to have an
effect on demand levels of existing customers in terms of heating
fuel selection, appliance usage behaviors and commitment to
conservation practices.. These prices may also affect decisions
concerning location of new businesses in the area. As a related
issue, the important finance, insurance, and real estate
subsectors do not include a conservation variable, nor do the
transporation, communications, or sanitary services subsectors.
In the post-1973 era, an increase in conservation practices is a
documented aspect of commercial energy use. The rapid growth in
electricity prices expected on Long Island will likely result in
an accelerated conservation effort. The ESRG end-use approach,
used in the Base Case and described in subsequent sections,
tracks the effects of conservation, building characteristics, and
equipment trends in greater detail than the aggregate time
trending approach used by LILCO.*
*For the basis of ESRG's assumptions and projections, see Section 8.
- 12 -
E S R G
2. OVERVIEW OF FORECASTING APPROACH
This section is restricted to a broad description of the
forecast model characteristics. The conceptual basis and mathe-
matical structures of the model are described in subsequent
sections.
The model forecasts are based on the aggregation of separate
forecasts for the major end-use components comprising system demand.
This allows for explicit incorporation of the impacts of differential
end-use-growth, energy policy, new technology and specific conser-
vation practices. For example, appliance efficiency improvements
are integrated indirectly into the appliance submodel rather than as
approximate adjustments on gross energy requirements.
The energy consumption for a given component is given by the
expression:
Energy Consumption in End-Use Category = End-Use Measure
x Energy Intensity
In other words, the energy consumption by end-use is the product
of the quantity of the end-use ("End-Use Measure") and the annual
average energy consumed per unit of the end-use ("Energy Intensity").
The measure of an end-use activity will be in units appropriate to
the sector being modelled. These are summarized in Figure 2.1.
The forecasting technique consists of three fundamental
steps: (1) analysis of base year energy-consuming stock in terms
of average measurE! levels and intensities, (2) specification of
growth in the end-use measures and (3) simulation of the factors
affecting the intensity of unit energy use. The actual mathematical
analogs chosen for energy consumption in the end-use models must be
wedded to the specific character of the end-use category. Further,
they must be constrained by limitations in available data. The
computational procedures selected are discussed in detail in
Sections 3 through 7.
The energy forecast model7 schematized in Figure 2.2, is the
heart of the system. It, in turn, is comprised of a ser~es of
submodels which produce forecasts of energy consumption disaggre-
gated by end-use. These are summed to give annual energy and are
input to the demand forecast model. The results for the utility
are combined to output system energies and peaks. (Additionally,
the energy forecasts broken down by end-use category may be
reported allowing for a clearer understanding of the structure
of total consumption and sensitivity to specific assumptions).
- 13 -
E $ R G
FIGURE 2.1
MODEL COMPONENTS
SECTOR END-USE ACTIVITY END-USE MEASURE ENERGY INTENSITY
Residential 14 Appliance Categories number of units average annual consumption
2 housing types per unit
5 building types floorspace square average annual consumption
Commercial 4 end-use categories footage per square foot
2 vintages (new & existing)
Industrial 19 manufacturing subsectors employment average annual consumption
per e~ployee
Chauges in
Efficiency,
Equipment,
qUnit Usage~
Residential
End-Use
Submodel
FIGURE 2.2
ENERGY FORECASTING MODEL SCHEHATIC
Indices
Building Type
and Vintage
Consumption
Per Ft~
Employment
Growth
Indices
Level by
Standard
Industrial
Intensity:
Consumption
Per Employee
by SIC
Commercial
End-Use
Submodel
{
~Energy by
TO PEAK LOAD MODEL
IMlx of put{
chased and
: self-
generated
energy
Industrial
End-Use
Submodel
Energy from
Residential,
Commercial,
and Industrial
Sectors
Other
Energy
Submodel
From one perspective, the model is a functional relationship
between a set of independent variables (data file) and selected
dependent variables (output forecasts). The computer program
designed for executing this mapping accepts a user-selected data
file and produces user-selected outputs. The inputs are of two
types: (1) data which characterizes the actual base year experi-
ence of a given utility and (2) assumptions on future values of
the independent variables which chart the changes in base year
values. The first type of data is developed and updated from
independent sources (utility surveys, industry load studies,
census information, etc.). The second type of input defines a
set of growth assumptions or "scenarios." Although one has guide-
lines fo~ estimating the growth variables e~tering the submodels
(historic patterns, independent national and state projections,
policy impacts, market penetration analysis, econometric equations,
etc.), uncertainty cannot be avoided.
This uncertainty is dealt with in the program in two ways.
First, a range of growth variable values are automatically required
in producing a forecast. The model is designed to accept from
the outset the uncertainty in the driving variables identified
by the user. The pro§'ram operates from "high" and "low" data
files associated with data choices for HIGH and LOW cases,
respectively. Although one cannot prophesize a given input item
with certainty, a realistic range of possible future values can
be given with some confidence. The high and low scenarios are
designed to bracket the set of possible futures. The Base Case
is defined in the model as the mid-range forecast illustrated
in .Figure 2.3. The uncertainty in the input data set is
reflected in the overall forecast uncertainty. The range of
uncertainty, is, of course, an increasing function of time.
The second method for treating uncertainty is through sensi-
tivity analysis. TSe program allows for temporary changes of an
input item (or set of items), permitting tests of the response
in forecast output to changes in data file input. The stability
of output to specific input variations can be computed and
utilized in assessing the validity of a given forecast.
- 16 -
E S R G
Figure 2.3
Forecast Scenarios
Variable
Forecast
I
time
Base Year
range of uncertainty
3. RESIDENTIAL SECTOR
This section describes the electrical energy demand forecast
model for the residential class of customers. The component end-
uses of residential energy consumption are treated in fourteen sep-
arate submodels. This level of detail allows the incorporation of
the central factors affecting overall demand which can be lost in
methodologies which forecast aggregate demand alone.
The fourteen residential end-uses for which submodels have
been developed are listed in Table 3.1.
TABLE 3.1
RESIDENTIAL END-USE SUBMODELS
End-Use
Input
1
2
3
4
5
6
7
8
9
10
11
12
13
14
These submodels will be described later.
level, annual consumption for end-use (i)
by the expression:
Refrigerator
Freezer
Electric Range
Lighting
Television
Clothes Dryer
Clothes Washer
Dishwasher
Water Heater
Air Conditioning - Room
Air Conditioning - Central
Space Heat
Heating Auxiliaries
Miscellaneous
At the most elementary
in year (t) is given
Et,i=Nt,i×Ct,£
where
(3.1)
Et,i = Total annual energy consumption of end-use (i) in year
Nt,i = Total number of corresponding units
Ct,i = Average annual energy consumption per unit
Then the total energy consumption in the resi~.ential sector
for year (t) becomes
~ Et,i
- 18
E $ R G
A glance at Equation 3.1 will show that the residential fore-
cast for each end-use can be viewed as a combined forecast of the
total nuz~ber of units, on the one hand, and the average consumption
per unit, on the other hand.
3.1 Number of Units
The number of units for a given end-use is computed as the
product of the number of households and the end-use saturation,
defined here as the average n~mber of units per household. The
number of household u~its is further divided into single family
u-nits (SF) and multifamily u~its (M_F). This breakdown is desirable
since appliance ownership and usage patterns may vary significantly
by housing type. A shift in the mix of SF and MF in the forecast
period thus affects ultimate demand.
5.2 Saturation Curves
Saturations enter the end-use submodels via the logistic
growth curve. This curve has the general form:
Ci~k
SATt,i,k ~ l+Bi,k×e-(Ai,k.T)
· for the saturation (SAT)
(k),
of a given end-use (i), and housing type
in year (t). The parameters are constrained by:
B>O, A>O, O<C<i.
(The indices are suppressed for notational convenience.)
Parameter C is called the ceiling, representing the asymptotic limit
of ~he dependent variable; the greater the value of A, the more
rapid is the approach to the ceiling. From the derivative
d SATt A
dt ~ ~ 'SATt '(C-SATt)
we see that the growth rate is proportional to both the level
already achieved and the increment remaining to the ceiling.
Ideally, the parameters would be estimated by fits to historic
saturation data. The data, however, is not sufficient to warrant
such a complete determihation. Instead, we have used base year
saturations (SBY) to determine one parameter, chosen values for the
ceiling or terminal saturation (STERM) according to scenario assump-
tions, and used historic data to fit the remaining variable A.
Rewriting Equation 3.2 in terms of STERM and SBY and fixing
the base year t=l as we do throughout the model, we arrive at the
form of the saturation curve as it enters the submodels:
(3.2)
(3.3)
(3.4)
- 19 -
E S R G
SATt= STERM
-A. (t-l)
1+ (STERM-SBY.] × e
[ SBY ]
(3~5)
3.3 End-use Submodels
The second term in Equation 3.1, the average annual energy
consumption for each end-use, incorporates a great deal of
complexity. Once the base year energies are established, the
time dependence of average energy consumption must be computed.
The major factors which can impact average energy use are:
· appliance efficiency increases
· thermal integrity improvements of building shells
· new technology market penetration
· population per household decreases
· energy conservation practices induced by
electricity price increases
The end-use submodels are designed to allow sensitivity to
assumptions on these trends. Consequently, overall forecasts
based on a range of reasonable input assumptions allow for the
development of a band of possible error within which lies the
"probable" forecast.
The submodels will be discussed in the sequence given
in Table 3.1. In each case, we give a brief qualitative descrip-
tion in the text and the system of equations in an accompanying
table. Although the end-uses have particular characteristics
which require unique model elements, the overall strategy dis-
played schematically in Figure 3.1 is used throughout. The
yearly increment in electrical energy consumption is calculated
by (1) subtracting the energy consumption of retired units,
(if any), (2) adding the energy consumption of replacements,
and (3) adding the energy consumption of additonal new units
due to customer and saturation growth. With this iteration
technique, we can, once the base year breakdown is established,
compute energy consumption for each year of the forecast under
a given set of assumptions on changes in saturation, customer,
technology mixes, efficiencies and use patterns.
3.3.1 Refrigerators and Freezers
The factors affecting demand for these two appliances are
quite similar so that the same algorithm for modeling growth
in energy consumption are employed. Variable definitions and
dynamic equations are summarized in Table 3.2.
In the case of decreasing saturations, the form of the curve
is given by:
SAT = STERM + (SBY-STERM) × e-A(t-1)
20 -
E S R G
FIGURE 3.1
Schematic of Yearly Energy Increments by End-use
Changes in:
Customers, Saturation,
Efficiencies, Equipment,
Use Pattern
Consumption
INew Additional Units
[Year t + 1
Consumption ~
Year t .
Consumption
Retired Units
Year t
· Consumption
L Year t + 1
Consumption
Replacement Units
Year t + 1
Changes in:
Efficiencies,
Equipment,
Use Pattern
- 21 -
E $ R G
The total number of appliances by housing type is obtained
by multiplication of saturation and households (Eq. 3.6). The
iteration procedure is initialized by computing base year consump-
tion as the product of the number of appliances on-line in the
base year and their average unit consumption (Eq. 3.7). There
is a great deal of variation in energy demand with brand, size
and model. Therefore, average usage may vary as a function
of regional appliance mix.
The iteration proceeds from year to year by subtracting
out the energy consumption of retired units and adding back
the energy from new units added (Eq. 3.8). The retired
energy is a product of the average number retired per year
(calculated as an approximation of the number of units coming
on-line one average lifetime before, Eq. 3.13a) and the average
unit consumption of the retired units (Eq. 3.9a). This last
factor must be treated with care. The 1960's saw an increase
in the average size of refrigerators and freezers and a rapid
penetration of the energy consuming frost-free feature. Then,
in the late 1970's these trends leveled off while the
efficiency of new units increased as a response to energy
price increases and government policies. Therefore, the
usage of retired units is modeled to changes over time, first
increasing and then decreasing, to reflect the changes in
historical vintages (Eq. 3.10).
New units, both replacements and net additions, are brought
on-line at current energy levels (Eq. 3.9b) , with new unit
average usage according to the efficiency improvements and
efficiency phase-in period assumed in a given model run
(Eq. 3.11).
- 22 -
~ S R G
TABLE 3.2
SUBMODEL FOR REFRIGEP~ATORS AND FREEZERS
Variable Code
t
i
k
TOTNUM
HSTOCK
SAT
UNNEW
UNAVBS
ALT
EFFIMP
EFFIMT
TEND
UNREP
NEWENI
RETENI
ENREU
YEAR
BY
UNOLD
YRMAX
UNMAX
RETNUM
NEWNUM
Year (base year = 1)
Appliance index (l=refrigerator and 2=freezer)
Housing type (SF=i, MF=2)
Total number of appliance
Households
Saturation
Average unit usage of new appliance
Average unit usage of base year stock
Average appliance lifetime
Efficiency improvement over base year models
Terminal efficiency improvement over base year
models
Final year of efficiency improvement phase-in
Average unit usage of replaced units
Energy use of new units
Energy use of retired units
Annual appliance energy demand
Year A.D. corresponding to t
Base year (A.D.)
Average unit usage of new units one average
lifetime prior to the base year
Historical year (A.D.) during which new
units had highest usage of all years
Average unit usage of new units in YRMAX
Number of re~ired units
Number of new units
Equations
Stock stream:
TOTNUMt,k,i
Initialize:
ENREUi,k,i
Iterate for t>l::
ENREUt,k,i.
where
RETENIt,k,i
SATt,k,i × HSTOCKt,k,i
TOTNUMi,k,i × UNAVBSk,i
ENREUt-i,k,i
UNREPt,k,i ×
- RETENIt,k,i
RETNUMt,k,i
+ NEWENIt,k,i
NEWENIt ,k,i
~ UNNEWt,k,i × NEWNUMt,k,i
(3.6)
(3.7)
(3.8)
(3.9a)
(3.9b)
- 23 -
E S R G
and
UNREPt ,k , i
UNNEWt,k,i
EFFIMPt, i
PJ~TNUMt ,k ,i =
NEWNUMt, k, i
'UNOLD + [YEARt-BY) / (YRMAXk,i + ALTi _ BY) ]
x (UNM~,i - UNOLDk,i)
for t ! YRMAXk,i + ALTi - BY
UNMAX + [ (YEARt - yp~AXk,i - ALT) / (BY-MAXk,i) ]
x (UNNEWi,k,i - UMAXk,i)
for YRMAXk,i + ALTi - BY<T <- ALTi
UNNEWt-ALTi for t · A~..Ti
- (1-EFFIMPt,i) × UNNEWi,k,i
for
EFFIMTi t > TEND
'TOTNUMi,k,i /ALTi for t -< ALTi
NEWNUMt_ALT,k,i for t > ALTi
TOTNUMt,k,i - TOTNUMt_i,k,i + RETNUMt,k,i
(3.10)
(3.11)
(3.12)
(3.13a)
(3~13b)
3.3.2 Electric Ranges
The determinants of growth for electric ranges are straight-
forward: saturation and customer increases, efficiency improvements
in new appliances, and market penetration of t_he microwave oven
feature which can decrease overall energy demand.
The total stock is given, as usual, as the product of
saturation and housing stock (Equation 3.14). Further dis-
aggregation by housing type is not necessary for this end-use
since available saturation.and energy demand data does not
distinguish between single and multi-f~m~ly usage patterns. The
iteration process is initialized with base year data (Equation
3.15) and proceeds with the characteristic subtraction of retired
units and addition of new units (Equation 3.16). Units are
re=ired at a rate equal to the inverse of the average life
time (Equation 3.17). The ~wo sources of new units, net additions
and replacements, are represented by the first and second terms
of Equation 3.18, respectively. Average usage of new units is
decremented by a factor derived from assumed efficiency targets
and phase-in times (Equations 3.19 and 3.20).
Finally, account is taken of the decreased energy usage
associated with microwave ovens used in association with electric
ranges. The total energy demand is a weighted factor of usage
without and with microwave ovens, the first and second terms,
respectively, in Equation 3.21.
E S - 24 - R O
TABLE 3.3
SUBMODEL FOR ELECTRIC RANGES
Variable Code
t
TOTNUM
HSTOCK
SAT
ENREU1
ENREU
UNAVB S
UNNEW
EFFIMP
TEND
RETENI
NEWENI
ALT
MSAT
EDF
Year (base year = 1)
Total number of appliance
Households
Saturation
Annual electric range energy demand w/o microwaves
Annual electric range energy demand with microwaves
Average usage base year stock
Average unit usage of new units
Efficiency improvement
Final year of efficiency improvement phase-in
Energy use of retired units
Energy use of new units
Average lifetime
Microwave oven saturation as a fraction of
electric ranges
Energy demand factor: ratio energy demand with
and without microwave oven
Equations
Stock stream:
TOTNUMt
Initialize:
ENREU11
Iterate for t>l:
ENREU1t
where
RETENIt
NEWENIt
= SATt × HSTOCKt
= TOTNUM1 × UNAVBS
= ENREUlt_1 - RETENIt
= ENREUlt_ 1/ALT
+ NEWENIt
(TOTNUMt - TOTNUMt_i) x UNNEWt
+ ITOTNUMt_i/ALTI×UNNEWt
and
UNNEWt = (1-EFFIMPt)
with { EFFIMT
EFFIMPt =
EFFIMT
× UNAVB S
× (t-l) / (TEND-l)
Microwave oven adjustment:
ENREUt
for ~ t<TEND
~ t>TEND
= ENREU1t × (1-MSATt) + MSATt × EDF × ENREU1t
(3.14)
3.15)
3.16)
3.17)
3.18)
3.19)
3.20)
3.21)
- 25 -
E S R G
3.3.3 Li~htin~
Lighting energy demand is represented as the product of average
annual energy usage per household and the number of households
(Equation 3.22). The household growth is developed outside the
submodel and inputted to it. There remains the anticipated changes
in lighting energy demand per household.
The model assumes that saturations are currently at 100%; i.e.,
all households have electric lighting and this shall remain true
throughout the forecast period. However, the intensity of lighting
use per household as well as the efficiency of conversion of
electric to light energy has in the past, and may well in the
future, vary with time. Future deviations from base year levels
is taken into account by the usage factor (Equation 3.23).
In the past, several factors have contributed to increases in
lighting energy demand, per household: shift in housing mix toward
larger SF residences, inexpensive electricity fostering purchase
of decorative lighting' and discouragement of household conservation
practice. These trends have generally reversed: family size is
gradually shrinking, MF dwellings are rising relative to SF, and
rising electricity costs are encouraging conservation.
It appears likely' that these shifting patterns will lead, at
least to some extent, to the market penetration of energy efficient
lightbulbs. These include improved incandescents and more fluores-
cents in the near term, followed possibly by commercialization of
the screw-in fluorescent in the 1980's. Possible impacts of such
technology shifts are incorporated in Equation 3.24.
- 26 -
TABLE 3.4
SUBMODEL FOR LIGHTING
Variable Code
HSTOCK
UNAVB S
UNAV
UF
MF
RELEFF
ENREU
Year (base year = 1)
Households
Average consumption per housing unit in the
base year
Average consumption per housing unit
Usage factor
Market fraction efficient bulbs
Efficiency improvement of nonconventional bulb
Annual energy demand for lighting
Equations
ENREUt
wi th
= UNAVt × HSTOCKt
UNAVt = UFt × UNAVBS
with efficient bulb capturing market fraction:
UNAVt = (1-MFt) × UNAVBS + MFt × (1-RELEFFt)
or
UFt = 1-MFt × RELEFFt
(3.22)
(3.23)
× UNAVB S
(3.24)
- 27 -
E $ R G
3.3.4 Television
The submodel for televison usage must contain sufficient
complexity to allow for (1) saturation and customer growth, (2)
changes in unit energy requirements, (3) changes in the mix of
black and white and color televisions,and (4) decreased usage per
unit in cases of multiple ownership. The last factor is due to
the nonproportionality between the number of televisions and the
viewing hours. That is, if, for instance, a family purchased a
second television, the hours of use will not simply double since
the redundan~ unit will be used to some extent in substitution for
the first.
The dynamics of television energy demand growth are presented
in Table 3.5. After defining the stock stream saturation and
housing stock with inputs from outside the submodel (Equation 3.25),
the iteration procedure is initialized with base year data
(Equation 3.26) and proceeds from year-to-year in the usual way
(Equations 3.27 to 3.31). Changing ratios of black and white to
color are allowed in the weighted averages for new units in
Equation 3.29. Finally, in the case of multiple average ownership,
the total energy is decremented by a decreased use factor for
second and third televisions (Equation 3.32).
- 28
E S R G
TABLE 3.5
SUBMODEL FOR TELEVISIONS
Variable Code
t
k
TOTNUM
HSTOCK
SAT
NEWCOL
NEWBW
EFIMCO
EFIMBW
TEND
FRBW
ALT
RETENI
NEWENI
UNAVBS
EN P~EU
DUF
Year (base year = 1)
Housing type
Total nLunber
Housing units
Saturation
Average unit usage of new color television
Average unit usage of new black and white television
Efficiency improvement color units over base year
Efficiency improvement black and white units
over base year
Final year of efficiency improvement phase-in
Fraction new units which are black and white
Average lifetime
Energy use of retired units
Energy use of new units
Average unit usage in base year
Arunual energy demand in (t,k)
Use factor for televisions beyond one per
household (DUF=i for full use, 0 for no use)
Equations
Stock stream:
TO~NUMt,k
Initialize:
ENREU1,k = TOTNUM1,k
Iterate for t>l:
ENREUt,k
where
RETENIt,k
NEWENIt,k
= SATt,k × HSTOCKt,k
UNAVBS
= ENREUt_i,k - RETENIt,k + NEWENIt,k
= ENREUt_i,k / ALT
= (TOTNUMt,k - TOTNUMt_i,k + TOTNUMt_i,k /AL~
× ( (1-FRBWt ) × NEWCOLt + FP. BWt × NEWBWt )
and
NEWCOLt = (1-EFiMCOt) x NEWCOL1
NEWBWt = (1-EFiMBWt) x NEWBW 1
~t<TEND
withEFIMCOt -%EFIMCOT-~(t-1 / (TEND-i). x EFIMCOT for % t>TEND
(similarly for EFIMBWt)
Decrease usage for multiple ownership (for SATt,k>i):
ENREU --- ENREUt,k x (i+DUF x (SATt,k -1) / SATt,k)
29 -
(3.25)
(3.26)
(3.27)
(3.28)
(3.29)
(3.30)
(3.31)
(3.32)
3.3.5 Clothes Dryers
The submodel for clothes dryers is quite simple. Demand
is primarily a function of saturation and customer growth since
efficiency improvement possibilities are small and substitute
technologies to conventional dryers are not on the horizon
(increased use of solar drying would be reflected in lower satura-
tions). Although predictions of changing unit usage intensity (such
as loads per week) are unrealistic, qualitatively, the decreasing
trend in population per household would suggest that current levels
should safely overestimate demand. The equation set (Table 3.6)
should by now be self-explanatory.
TABLE 3.6
SUBMODEL FOR CLOTHES DRYER
Variable Code
t
TOTNUM
HSTOCK
SAT
UNAVB S
ALT
EFFIMP
TEND
NEWENI
KETENI
UNNEW
ENREU
Year (base year = 1)
Total number
Households
Saturation
Average unit usage of base year stock
Average lifetime
Efficiency improvement over base year units
Final year of efficiency improvement phase-in
Energy demand of new units
Energy demand of retired units
Average unit usage of new units year t
Annual energy demand in year t
Equations
Stock stream:
TOTNUMt
Initialize:
ENREU1
Iterate for t>l:
ENREUt
where
RETENIt
NEWENIt
and
UNNEWt
with
EFFIMPt
= SATt × HSTOCKt
= TOTNUM1 × UNAVBS
= ENREUt_1 + NEWENIt
= ENREUt_1 / ALT
= (TOTNUMt - TOTNUMt_1
= (1-EFFIMPt) × UNAVBS
_~(t-1) /(TEND-i)
-~EFFIMT
(3.33)
(3.34)
- PETE}IIt (3.35)
(3.36)
+ TOTNUMt_1 / ALT)x UNNEWt
(3.37)
× ~ t<TEND
EFFIMT for t t~TEND (3.38)
- 30 -
E S R G
3.3.6 Clothes Washer and Dishwasher
Clothes washers and dishwashers are treated together since,
as we shall see, the algorithm for modeling demand is identical.
Each of these end-uses requires energy in two forms: (1) electric
energy to drive motors and auxiliary equipment and (2) thermal
energy in the form of hot water for process functions. Technology
shifts are in the offing which would effect each of these.
For the case of thermal requirements, the impact on overall
electrical energy is indirect. Specifically, changes in hot water
demand will "flow through" to effect the electricity demand in
the cases where hot water is produced in electric hot water heaters.
The submodel allows for changes in both the electrical and thermal
demands, saving the latter for input into the electric water heat
submodel.
Therefore, after running the usual iteration to develop
direct electrical energy demand (Equations 3.39 to 3.45), average
forecast hot water demand for each appliance is calculated as a
function both of overall saturation growths and unit demand
changes. (Equation 3.46 to 3.47). These results are incorporated
into the electric hot water heater submodel.
- 31 -
E S R G
TABLE 3. 7
SUBMODEL FOR CLOTHES WASHER AND DISHWASHER
Variable Code
t
TOTNUM
HSTOCK
SAT
UNAVBS
ALT
CWHW
DWHW
HWRECW
HWREDW
UNNEW
NEWENI
RETENI
ENREU
EFFIM_~
TEND
Year (base year = 1)
Appliance index (CW = 7, DW = 8)
Total number
Households
Saturation
Average base year unit electric energy usage
Average appliance lifetime
Clothes washer average hot water demand per
customer
Dishwasher average hot water demand per customer
Hot water reduced deman~ -- clothes washer
Hot water reduced demand -- dishwasher
Average unit electrical energy usage of new
appliance units
Energy demand of new units
Energy demand of retired units
Annual energy demand
Efficiency improvement over base year
Final year of efficiency improvement phase-in
Equations
Stock stream:
TOTNUMt,i
Initialize:
ENREUi,i
Iterate for t>l:
ENREUt,i
where
RETENIt,i
NEWENIt,i
and
UNNEWt,i
with
EFFIMPt,i
= SATt,1 x HSTOCKt,i
= TOTNUMlfi × UNAVBSi
= ENREUt_i,i + NEWENIt,i - RETENIt,i
= ENREUt_i,i/ALTi
=(TOTNUMt,i-TOTNUMt_i,i+TOTNUMt_i,i/ALTi)
x UNNEWt,i
= (1 - EFFIMPt,i)
= ~ (t-l/(TEND-1)
~EFFIMT
× UNAVBSi
× EFFIMT for
It<TEND
t>TEND
(Continued)
- 32
(3.39)
(3.40)
(3.41)
(3.42)
(3.43)
(3.44)
(3.45)
E R G
TABLE 3.7 (Continued)
Hot Water Demands:
New unit usage year t:
UCWHWI = 19 × UNAVBS7 × (1-HWRECWt)
UDWHWI = 4.6 × UNAVBS8 x (1-HWREDWt)
(factor 19 and 4.6 are ratios of hot water to elect=ic energy requirements
for clothes washer and dishwasher, respectively ~ef. l~i
with
~t<TEND
HWRECWt ={((t"i)/(TZND-1))HWRECT × HWRECT for ~t>TEND
Average unit usage:
U~It-I~t = (UC'WI~t_1 ×R-~,Mt + (TOTNUMt - RlgMt)
where REMt = remaining units from previous year = TOTNUMt_1
~( (1-1/ALTi)
Average usage per customer:
CWHWt = SAT~,7 x UCWHWt
And similarly for dishwasher.
x UCWq~.~It)/TOTNUMt ( 3.46 )
(3.47]
33
3.3.7 Electric Water Heaters
The electric water heater submodel is sensitive to a number
of time dependent factors affecting overall energy demand:
saturation
efficiencies
average residential hot water requirement
solar technology penetration
The number of electric water heaters is computed in Equation
3.48. First, in Equation 3.48a, the base year units are computed
from input data. For subsequent years, ~he total number is computed,
as the combination of the previous year's value (first term on the
right of Equation 3.4Sa) plus additions from two new markets. First,
all new electric space heaters are assumed to also have electric
water heaters. (This will slightly overstate growth.) This is
reflected in the second line on the right of Equation 3.48b (penetrations
of electric space heat also appear in the esh submodel, Section 3.3.9).
Second, new non-electric space heated homes (Equation 3.48b, third
line first bracket) are assumed to purchase electric water heaters
according to base year electric water heaters saturations in base
year non-electric space heated homes (Equation 3.48b, third line,
second bracket).
The hot water'energy demands of clothes washer and dishwasher
have been developed earlier and are used in Equation 3.49 to define
the demand from "other" uses. Possible reductions in this category,
such as widespread adoption of slow-flow shower heads, etc., which
are now on the market, are also allowed for in the last expression.
Average efficiencies of electric w~t~r.h~aters are expected
to improve with time primarily due to mlnlmlz~ng stand-by losses
through better insulation jackets. The iterative procedure in
Equation 3.51 weights new units (first term) with existing units
(second term). The unit electric energy demand is then given by
the ratio of hot water output (measured in KWH's) and the average
efficiency (Equation 3.52). If there is some penetration of solar
equipment to assist in ihot water production, this average must be
properly corrected by weighting in the fraction solar assisted at
reduced demand levels (Equation 3°52). The total electric energy
required for this energy then follows immediately as the product
of the total number on line and the average unit usage (Equation 3.53).
34
E S R G
TABLE 3.8
SUBMODE7. FOR ELECTRIC WATER HEATER
Variable Code
t
k
TOTNUM
HSTOCK
SBY
ESHSAT
UNAVBS
UNAV
ALT
CWHW
DWHW
OTHW
HWREOT
AVEFF
NUNEFF
FS
PCSOLW
ENP~U
PEN
Year (base year = 1)
Housing type (1 = SF, 2 = MF)
Total number
Households
Base year electric water heater saturation
Electric space heating saturation
Average base year unit electric energy demand
Average unit usage
Average lifetime
Clothes washer hot water demand
Dishwasher hot water demand
Other hot water demand
Hot water reduced demand for "other"
Average electric water heater efficiency
New unit average efficiency year t
Fraction electric hot water heaters solar assisted
Fraction supplied by solar in solar assisted units
Total energy demand year t
Penetration of esh in new construction
Equations
Stock stream:
TOTNUM1,k
TOTNUM%,k
= SBYk × HSTOCK1,k
= TOTNUMt_i,k + (HSTOCKt,k - HSTOCKt_i,k) × PENt,k
+ [(HSTOCKt,k - HSTOCKt_i,k) × (1-PENt,k)]
× [(SBY-ESHSATi,k)/(1-ESHSATi,k)]
"Other" water demand:
OTHWt = (UNAVBS × AVEFF1 - DWHW1 - CWHW1) × (1-HWREOTt)
Where DWHW and CWBW are from previous submodel, the first term in
parenthesis is the base year total hot water usage.
By definition
NUNEFFt =
where
EFFIMPt =
Average efficiency from:
TOTNUMt × .AVEFFt =
then,
UNAVt =
Finally,
ENPd~Ut =
AVEFF1/(1-EFFIMPt)
(t-1)/(TEND-1) x EFFIMT for
EFFIMT
(TOTNUMt - TOTNUMt_1
x NUNEFFt + (TOTNUMt_1
× AVEFFt_1
(DWHWt + CW~Wt + OTHWt)/AVEFFt
x (1-FSt + FSt x (1-PCSOLW))
TOTNUMt × UNAVt
t<TEND
t~TElqD
+ TOTNUMt_I/.ALT)
- TOTNUMt_i/ALT)
(w/o solar)
(w solar)
- 35 -
(3.48a
(3.48b)
(3.49)
(3.50)
(3.51)
(3.52)
(3.53
~ S R G
3.3.8 Air Conditioners
The two types of air conditioners -- room and central --
are treated as separate end-uses. For each, the final forecast
is a co-mingling of saturation and customer growths, efficiency
increases, and building shell-thermal integrity improvements.
It is tacitly assumed that average unit size will not increase
over the base year due to demographic trends toward smaller
family size and the decreased cooling load requirement that
accompanies improved insulation.
Energy demand is calculated by employing the usual iterative
sequence (Equations 3.54 to 3.60). The model assumes that in
cases of multiple room air-conditioner ownership, average energy
usage is additive. This may lead to a slight overestimate of
demand insofar as second and third window/wall units are used
substitutively to some extent. Such an effect is, however,
difficult to estimate.
The model allows for adjustments in the average thermal
integrity of building shells in the housing stock (Equation 3.61).
This is given as an average over changes in base year and new
construction units as indicated in Equations 3.62 and 3.62a. There
are two likely sources for improvements here: reinsulation in
the retrofit market and stricter conservation practices in new
building designs relative to historic design standards.
Consequently, the overall improvement over base year values
depends on estimates of several factors such as-current building
stock average characteristics, the degree of future reinsulation,
and the effects of anticipated building codes for new construction.
- 36 -
E S R G
TABLE 3.9
SUBMODEL FOR AIR CONDITIONERS
Variable Code
t
k
i
TOTNUM
ALT
HSTOCK
BYHSTK
HRET
TIIMP
TIE
TIN
EFFIMP
TEND
SAT
UNAVBS
UNNEW
NEWENI
RETENI
ENEUI1
ENREU
Equations
Stock stream:
Year (base year = r)
Housing type (1 = SF, 2 - MF)
End use index (10 = Room A/C, 11 = Central A/C)
Total number on-line
Average appliance lifetime
Housing units
Base year housing stock surviving
Housing unit removal rate
Average thermal integrity improvement
Thermal integrity improvement of base year housing units
Thermal integrity improvement of new construction units
Efficiency improvement over base year
Final year of efficiency phase-in
Saturation
Average base year unit consumption
Average unit usage of new units
Energy demand of new units
Energy demand of retired units
Annual energy demand w/o thermal integrity improvement
Annual energy demand
- RETENIt,k,i
= ENEUIlt_i,k,i/ALTi
(TOTNUMt,k,i - TOTNUMt_i,k,i
+ TOTNUMt_i,k,i/ALTi) x UNNEWt,k,i
TOTNUMt,k,i = SATt,k,i x HSTOCKt,k
Initialize:
ENEUIll,k,i = TOTNUMi,k,i x UNAVBSk,i
Iterate for t>l:
ENEUIlt,k,i = ENEUIlf_i,k,i + NEWENIt,k,i
where
RETENIt,k,i
NEWENIt,k,i
(3.54)
(3.55)
(3.56)
3.57)
3.58)
and
UNNEWt,k,i
with
EFFIMPt,i
= (1 - EFFIMPt,i) x UNAVBSk,i
x EF MT FOR it < TEND
Correct for changes in thermal integrity:
ENREUt,k,i = (1-TIIMPt,k,i) x ENEUIlt,k,i
where
TIII~t,k,i
and
BYHSTKt,k
3.59)
(3.60)
(3.61)
= [TIIMPt_I,k,i x HSTOCKt_i,k + TIEt,k,i x BYHSTKt,k
- TIEt_l,k,i x BYHSTKt_i,k + TINt,k,i x(HSTOCKt,k
- HSTOCKt-i,k + BYHSTKt,k - BYHSTKt-i,k) ~ (3.62)
/ HSTOCKt,k for t>l
= HSTOCK1,k x (1-HRETk) t-1 (3.62a)
- 37 -
E S R G
3.3.9 Electric Space Heating
The growth in the number of electric space heated (ESH) homes
is closely related to the decision on fuel use in new construction
markets or in converting existing households from fossil fuel
heating to electric. Consequently, it is analytically useful to
introduce the concept of "penetration" in developing the number
of housing units with ESH. In the model, the following definition
is used:
A electric space heat customerst
Penetrati°nt = A customerst
where t is the year label and "A" signifies the change from the
previous year. The historic values of the increments are readily
available from utility records providing useful information in
estimating future trends. With this definition, the yearly number
of ESH units can be computed through the iteration procedure of
Equation 3.63 of Table 3.10 with the initial number defined as
the product of base year saturation and household for each housing
type.
The ESH intensity (annual KWH consumption per unit) must be
represented as the combination of three distinct heating systems:
conventional resistance heating, electrically driven heat pump,
and solar augmentation (with or without heat pumps). The key
dynamic expression is the iteration formula, Equation 3.66,
which increments the previous year's total ESH energy demand
by the additional demand coming on-line. This additional demand is
the sum of the contributions from the system options considered:
conventional resistance ("direct"), heat pump and solar, respectively,
in Equation 3.66b. Each of these is in turn decomposed into the
product of new units in the ESH subcategory and usage per unit
Equations 3.67, 3.68, and 3.69.
A supplemental wood heat option reduces the electric energy
intensity in both existing and new dwellings (in Equations 3.66a
and 3.68a, respectively. Adjustments are also made for conservation
oriented changes in building envelope designs ("thermal integrity
factor") in new units relative to the base year mix of electrically
heated units, as shown in Equation 3.68a. Adjustments can also
be made over time for retrofit improvements in the building
envelopes of base year units, as described in Equations 3.66a and
3.70. Finally, the market share of each ESH option is given a
broken linear time dependence over the forecast period.
- 38 -
E S R G
TABLE 3.10
SUBMODEL FOR ELECTRIC SPACE HEATING
Variable Code
year
t
k
BY
TOTNUM
DELNUM
HSTOCK
PEN
PENIN
PEN90
ESHSAT
UNAVBS
RHAVBS
RHAVT
RSHUEX
RSHUNW
NEWENI
NESHDI
NESHHP
NESHSA
FHP
FHPBY
TEHP
COP
COPEI
TEFFI
FSA
TSSA
PCSOL
FWHE
FWH90
FWHA
FWHN
TIF
TIEHE
TIE
ENREU
Year (A.D.)
Year Index (base year = 1)
Building type Index (SF=i, MF=2)
Base year (A.D.)
Total number
New units in current year
Housing stock
Penetration in current year
Initial penetration
Penetration in 1990
Base year saturation of electric space heat
Average base year usage all units
Average base year usage resistance heating units
Average usage new electric resistance heating
units in current year
Energy demand of base year stock in current year
Energy demand of all units installed after base
year through current year
Energy demand of new units installed in current
year
Energy demand of new direct ESH
Energy demand of new ESH with heat pump
Energy demand of new ESH with solar assist
Fraction of new ESH units with heat pump
Fraction of base year ESH units with heat pump
T~e end of increasing heat pump fraction of
new ESH
Heat pump coefficient of performance in current
year
COP efficiency improvement
End year COP efficiency improvement
Fraction of new ESH units with solar assist
Time start of solar space heat penetration
Percent heating requirement due to solar in
solar assisted ESH units
Wood space heat fraction in base yaer
Wood space heat fraction in 1990
Wood space heat fraction current year
Wood space heat fraction for new homes
Thermal integrity factor adjusting new unit
demand from base year unit demand
Thermal integrity improvement for base year units
in year t = 20
Thermal integrity improvement for base year units
in. current year
Annual energy demand
- 39
E $ R G
TABLE 3.10
(Continued)
Equations
Stock stream:
TOTNUMI,k = ESHSATk × HSTOCK1,k (3°63)
TOTNUMt,k ~_ PENt,k -
= TOTNUMt 1,k + × (HSTOCKt,k HSTOCKt_i,k) (3.63a)
for t >1 and HSTOCK >
t,k HSTOCKt-1,k
TOTNUMt,k = TOTNUMt_i,k
Initialize:
ENREU1,k = TOTNUM1,k × UNAVB~ (3.64)
RHAVBS = UNAVBSk/(1 - FHPBYk + FHPBYk/COPl,k) (3~65)
Iterate:
where
RSHUEXt,k
for t > 1 and HSTOCKt,k ~ HSTOCKt_l,k(3o63b)
ENREUt,k = RSHUEXt,k + RSHUNWt,k
= TOTNUM1,k × UNAVBS 1 - FWHAtrk
1 - FWHEk
× (1 - TIEr,k)
RSHUNWt,k = ~ NESHDIt,k + NESHHPt,k + NESHSAt,k
t
Subcomponents of new demand:
= (1 - FHPt,k - FSAt,k) × DELNUMt,k × RHAVT
= FHPt,k × DELNUMt,k × RHAVT/COPt,k
= FSAt,k ~ DELNUMt,k × (1 - PCSO~/100)
NESHDIt,k
NESHHPt,k
NESHSAt,k
where
DELNUMt,k = TOTNUMt,k -
TO~t_l,k
1-FWHNk
1 - FWHEk
RHAVT = RHAVBS × TiFk ×
(3~66)
(3.66a)
(3.66b)
(3.67)
(3.67a)
(3.67b)
(3.68)
(3.68a)
- 40 -
TABLE 3.10
(Continued)
Linear phase-ins of time dependent variables:
FHPt,k =iFHPi,k + (t-1)/(TEHP-1)× (FHPTEHP,k
iFHPTEHP,k
t < TEHP
>
t - TEHP
FSAt'k = [(t-TSSA)/(21-TSSA)] x FAS21,k for
I FWHE + (k-'~H90 - k'~HE) x
FWHAt'k = [FWH90
[year < 1990
~year ~ 1990
PENt,k
~PENIN + (PEN90
~PEN90
- PENIN) ×
- FHP1,k) for
TSSA
TSSA
(t-1)/(1990-BY)
for
(t-2)/(1990-BY-1)
for
Iyear < 1990
>
[year - 1990
TIEt,k = TIEHE(k) × (t-1)/20
= ICOPi,k × [1 + COPEI
coPt'k [COP1,k × (1 + COPEI)
x (t-1)/(TEFFI-1)]
for
ft < TEFFI
t ~ TEFFI
(3.69)
(3.69a)
(3.69b
(3.69c
(3.70)
(3.71)
41 -
E S R G
3.3.10 Heating Auxiliaries
Heating auxiliaries refers to the electrically driven
equipment such as pumps and fans used in conjunction with oil
and gas home heating systems. Energy demand is simply the number
of fossil-fuel heating systems multiplied by the average unit
electrical demand for auxiliaries. With the assumption that all
customers have either fossil fuel or electric space heating,
the heating auxiliary saturations is given simply by one minus
the electric space heating saturations. This is used in
developing the yearly number on-line (Eq. 3.72). The expression
for annual heating auxiliary energy consumption (Eq. 3.75a) is
composed of contributions from surviving base year households
(defined in Eq. 3.74) and newly constructed units, reduced by
a factor to account for growth in supplemental wood heating.
Energy requirements for these are shown, respectively, in
Eqs. 3.73 and 3.75 where possible decrements in average units
usage due to improvements in the average thermal integrity of
residential buildings is accounted for. On the other hand, the
model does not explicitly include possible decreased energy
requirements due to heating system or electric motor efficiency
improvements.
42 -
E S R G
TABLE 3.11
SUBMODEL FOR HEATING AUXILIARIES
Variable Code:
t
k
ESHSAT
UNAVBS
TIIMP
TIE
TIN
HSTOCK
BYH-STK
HRET
TOTK
ENEUI
ENEUI1
ENEUI2
FWFFE
FWFFE90
FWFF
Equations
TOTKt,k
ENEUIlt,k
where:
BYHSTKt ,k
ENEUI 2t ,k
Finally,
ENEUIt,k
FWFF
Year (base year = 1)
Housing type (i=$F, 2=MF)
Electric space heat saturation
Average unit usage in base year
Thermal integrity improvement over base year
Thermal integrity improvement of base year
housing units
Thermal integrity improvement of new construction
units
Housing stock
Base year non~ESH housing units surviving
Housing unit removal rate
Total number of non-ESH housing units
Annual energy demand
Annual energy demand from base year housing stock
Annual energy demand frc~ newly constructed units
Wood space heat fraction in base year
Wood space heat fraction in 1990
Wood space heat fraction in current year
= (1 - ESHSATt,k) × HSTOCKt,k (3.72)
= BYHSTKt,k × (1-TIEr,k) × UNAVBSk (3.73)
TOTK1,k × (1-HRETk)t-1 (3.74)
(ENEUI2t_I,k + (1-TINt,k) × UNAVBSk ×
TOTKt,k - TOTKt_i,k + BYHSTKt_i,k -
BYHSTKt,k] (3.75)
(ENEUIlt,k + ENEUI2t,k) × [1-FWFFE]
'FWFFE + (FWFFg0-FWFFE) × ~1990_BASEYR3
FWFF90
(3.75a)
~year<1990
for (3.75b)
%year~1990
- 43 -
S R G
3.3.11 Miscellaneous Appliances
This category includes an enormous array of small appliances used
in the home for food preparation, entertainment, maintenance and personal
care. Since energy demand in this category consists of use in a large
variety of devices, each with low annual consumption, a disaggregated
computational scheme is inappropriate. Consequently, forecast energy
consumption is computed simply as the product of average demand per
housing unit and the number of housing units (Eq. 3.76). The average
unit usage deviates from base year values by a factor which is phased in
linearly over the forecast period (Eqs. 3.76a and 3.77).
Average u~e per customer of miscellaneous appliances had been
generally increasing prior to 1973 as part of the overall growth in energy-
intensive equipment fostered by a combination of rising real per capita
income, declining real electricity prices, and an explosion of small
convenience devices. Current trends can be expected to moderate grow~h~
Major factors are:
· increasing electricity costs
· substitution effects (e.g., cooking devices for ranges)
· decreased growth in disposable income
· energy conservation awareness
· smaller families
· market saturation
On the other hand, unanticipated new devices may appear in the marketplace
to refuel growth in average consumption. Consequently, there is a good
deal of uncertainty in use per customer trends over the twenty year
forecast. Actual scenario runs of the model encompass a range of values.
Variable Codes
t
HSTOCK
UNAVBS
UPCIN
UNAV
ENREU
Equations:
ENR~Ut
where
UNAVt
with
UPCINt
TABLE 3.12
SUBMODEL FOR MISCELLANEOUS APPLIANCES
Year (base year = 1)
Total number of housing units year t
Annual average usage per housing unit in base year
Use per customer increase
Annual average usage per household unit
Total annual energy consumption
= UNAVt x RSTOCKt
(3.76)
= (1 + UPCINt) × UNAVBSt
(3.76a)
= ((t-i)/20) × UPCIN21
(3.77)
- 44 -
E S R G
4. COMMERCIAL SECTOR
In modeling electrical energy consumption for the commercial sector,
the degree of analytic detail is constrained by the adequacy both of the
data base and current understanding of energy flows in the commercial
building sector. Over the past few years, however, substantial progress
has been made in quantitatively characterizing the components of
commercial demand which allows for considerably more refinement than
has been traditionally employed (~.~., Refs. 2-6, 8).
The importance of avoiding aggregate historical trending or
correlation analysis is underscored by the reversal or diminution of
the underlying factors that drove U.S. commercial energy growth at
over 5% per year in the twenty years preceding the oil embargo of
1973. These factors included: rapidly increasing population, per
capita income, and proportion of employment in services, combined with
decreasing energy costs.
The commercial model tracks energy demand for five building types
(BT), four end-uses (EU), or twenty BT/EU combinations each for
existing and new buildings. These are displayed in Table 4.1 along with
the commercial category allocated to each building type. Both demarca-
tions -- "building type" and "commercial category" -- will be useful
in constructing the commercial model.
4.1 Model Structure
AS discussed in Sec. 2, the underlying strategies in the commercial
and residential sectors are analogous. In the commercial sector, the
measure of energy using activity is the magnitude of floor space while
the energy intensity is expressed in average annual kwh/square foot
for each end-use, building type, and utility service territory. The
elements of the model are displayed schematically in Figure 4.1.
The specifications of base year floor space, average consumption
per square foot of each end-use ("electrical use coefficients"), and
saturations (fraction of floorspace with end-use) gives the base year
breakdowns. Folding in the time dependences of floorspace, conserva-
tion, and saturations, one arrives at the yearly forecasts.
The commercial forecast model, therefore, divides conceptually
into two separate submodels: one for floorspace and the other for
electric intensity. These will be discussed in turn.
4.2 Commercial Floors~ace
The floorspace computation is summarized in the first row of
Figure 4.1. Note that the floorspace analysis is disaggregated by
commercial category; these are then aggregated to building types
according to the allocations of Table 4.1. The reason for this
procedure is that while detailed growth forecasts are available
for the 14 commercial categories (e.g., Ref. 7), the latest intensity
E S - 45 - R G
TABLE 4.1
COMMERCIAL MODEL END-USE, BUILDING TYPES AND COMMERCIAL CATEGORY
Index
i End Use
1 Space-Heating
2 Cooling
3 Lighting
4 Aux. & Power
Index Index
k Building Type j Cc~nercial Category
1 .Office 1 Finance, Insttrance and Real
Estate
2 Federal Gove~ ~nt
3 State & Local Gove~nt
4 Professional Services
2 Retail 5 Re~c~i] and Wholesale
3 Hospitals 13 Hospi~m]s and Health Ralated
EstaBlishments
4 Schools 14 Schools and Education
5 Other 6 Trucking and Warehouse
7 Other Transportation Services
8 C~L,L~ulications
9 Lodging & Personal Services
10 Business & Repair Services
11 Amas~-nt & Recreation
12 Railroad
- 46
E $ R G
Electrical Use
Coefficients, by
Building Type
End-Use
Ai r-Condi tioninq
&
Electric Space Ileat-
ing Saturations
Temporal Factors
Floorspace
Growth and
Retirement by
Con~ercial Category
Conservation Tech-
nology Penetration
Rates and Energy
Savings by BT, EU
new & retrofit markets
Saturation in New
tion Growth in Existil%g BTs
Floorspace
Retrofit Market
by ~T & EU
-~'~oorspaco
New Construction
Market by BT & EU
Retrofit Market
Electrical Use Co-
efficients by BT & EU
New Construction Mar-
ket Electricial Use
Coefficients by BT & EU
Retrofit Market
Electrical Consumption
by BT & EU
New Construction Market
Electrical Consumption
by BT & EU
Co~anercial Sector
Electrical Consumption
by
BT & EU
Electrical
Consumption
by
liT & EU
FIGURE 4.1
COMMERCIAL SECTOR MODEL SCHEMATIC
Indices
t = 1,2 ....
j = 1 to 14
n = 1 to 2
k = 1 to 5
Variables
SQFTCC
SQFTBT
RSQFT
SPOP
UPOP
SAPOP
PARAM
EMP
C~{IND
OSQFT
NSQFT
AGG
TABLE 4.2
COMMERCIAL MODEL - FLOORSPACE
Year (1975 = 1)
Commercial category
Existing or new building
Building types
Square footage by commercial category
Square footage by building type
Annual retirement rate of base year floorspace
Statewide population
Population in forecast area
School age population in forecast area
Parameter used for floorspace growth
Statewide employees
Commerc£al index giving floor, pace ratios
in. successive years
Pre-1976 floorspace remaining in year t
New floorspace
Aggregation matrix from commercial category
to building type
Equations:
Growth parameters:
PARAMt,j = E}~t,j
for j = 1 to 12 and
PARAMt,13 = UPOPt
PARAMt,14 = SAPOPt
Growth indices:
COMINDt,j = PA~Mt,j
Iterate:
· SQFTCCt,j = COMINDt,j
x UPOPt/SPOPt
/PARAMt_ltj
× SQFTCCt_i,j
witlh SQFTCC1,j inputted.
Aggregate to building type:
SQFTBTt,k = ~AGGj,k x SQFTCCt,j
Breakdown to existing and new:
t
SQFTBTt,k = OSQFTt,k +t~=2NSQFTt',k
(continued)
E S - 48 - R
for t>l
for t>l
G
(4.1)
(4°2)
(4.3)
(4.4)
(4.5)
(4.6)
(4.7)
where
and
TABLE 4.2 (Continued)
SQFTBT1,k {t = 1
OSQFTt,k = for
(1-RSQFTk) x OSQFTt_i,k t > 1
NSQFTt,k = SQFTBTt,k - SQFTBTt_i,k
+ RSQFTk x SQFTBTt_i,k
(4.8)
(4.9)
- 49 -
E S R G
data and conservation penetration analysis are available on the
basis of building type (References 6 and 8). Floorspace is thus
treated on the basis of commercial category and then aggregated
to the building type demarcation.
The system of equations for the floorspace component of the
commercial model is given in Table 4.2. The model is based on a
year-to-year iteration (Equation 4.5). Two factors are involved:
1975 floorspace data to initialize the iteration and an annual
growth index.
4.2.1 1975 Floorspace
A separate computation was performed to generate the 1975
floorspace data. This was required by the paucity of data on
existing commercial building stock. Except for schools and
hospitals, initial floorspace estimates were derived as the product
of employment by SIC (Refs. 9, 29,36), adjusted to full-time eQui-
valents, and average floorspace per employee by SIC. These estimates
were then aggregated according to the groupings in Table 4.3. The
square foot multipliers, giving average footage per employee by SIC,
are displayed in Table 4.3. They are based on average values given
in the literature (Ref. 10~. School and hospital floorspace
estimates were derived by scaling national floorspace estimates
(Ref. 6) by the ratio of forecast area to national pupils and
hospital beds. (Ref. 27), respectively.
The 1975 floorspace are used as data in the floorspace module
of the main Commercial program (see Eq. 4.5). Total floorspace
is ultimately normalized to base year energies as discussed later.
4.2.2 Floorspace Growth Indices
The growth indices ("COMIND") give floorspace ratios in
successive years (Table 4.2, Equation 4.4). The growth indices
are equivalent to:
where
COMINDt,j
GRSQFTCCt,j
= (1 + G~SQFTCCt,j) (4~10)
average annual growth rate of
square footage in commercial
category and year t.
For the case of hospital and health related establishments (j=13),
population growth is the proxy for floorspace growth (Equation 4.2).
For the case of schools (j~14), floorspace growth is equated to growth
in school age population (Equation 4.3). For the other co~ercial
categories, the level of employment was taken at t_he best measure of
activity and, therefore, floorspace growth. Estimates of population
and employment growth used in the current forecast are postponed to
the data discussion below.
- 50 -
E S R G
T~mT,~ 4.3
.~UARE ~OOTAGE MULTIPT,~;~S
C~,~rcial Category
Corres?ondi n~ SIC' s
1. Finance, Insurance, Real Estate(FIRE)
2. Federal Govez~L~nt
3. State and Local Gove~%~nt
4. Professicc~l Service
5. P~tail & Wholesale
6. Trucking & Warehouse
7. Other Transportat/on
8. C~u~ication
9. Lodging and Personal Service
10. Business and Repair Service
11. ~t & Recreation Service
12.
60
61
62
63
64
65
66
67
91
92
93
81
83
89
50,51
52
53
54
55
56
57
58
59
42
41
44
45
46
47
48
70
72
73
75
76
78
79
84
86
4O
* The A~ntinistrative ~nd Auxiliary porticos of FI~E, retail and
wholesale, transportation, C~,~nication, and Utilities are
allotted 200 sq. ft. per employee. Source: P~f. 10.
Average
Square Feet
Per Employee*
155
214
176
149
149
390
187
156
189
183
393
211
216
312
682
987
271
509
5O2
532
878
270
444
3162
28O
139
809
8050
780
177
837
304
275
1422
270
777
871
2000
860
187
- 51 -
E S R G
4.3 Electric Energy Intensities
With floorspace estimates generated with the methodology just'
described, there remains the second element of the commercial fore-
cast: average electric energy consumption per square foot. As
shown in the lower two rows of boxes of Figure 4.1, the evaluation
of intensities again involves two phases: first, a specification of
initial values of electrical demand coefficients (defined as
average annual electrical consumption of a given BT/EU/service
territory combination) and end-use saturations; second, an estimation
of conservation penetration and saturation growth. We shall discuss
these two phases ~equentially.
4.3.1 1975 Intensities
Average electrical demands by end-use and building types have
been adapted from the "theoretical building loads" developed for
the Department of .Energy by Arthur D. Little, Inc. (Ref. 8).
The study combined engineering design parameters and survey research
to arrive at estimates of average building requirements for each
of the EU/BT combinations treated in ~he commercial model. The
adaptation of ADL's relevant regional building loads to unit
electricity demands (electrical use coefficients) by service
territory requires the adjustment of weather sensitive loads to
the prevailing climatic conditions.
4.3.2 Future Intensities
The computation of forecast year intensities is described
in Table 4.4. Intensities are, by definition, the product of
the saturation (fraction of floorspace with end-use) and the
electrical use coefficients (average annual kwh/ft2 of floorspace
with end-use). This is expressed mathematically by Eq. 4.13~ Note
that the intensities are specified by 4 end-uses and 10 building
types. In practice, however, many of the inputs are trivial.
(E.g., saturations are defined as 1 for i = 3 and 4). The time
dependence of the electric use coefficient ("EUC") is obtained
by incrementing the 1975 values by changes in end-use demands
due to conservation practices initiated in the post-1975 era.
In reference 6, three levels of efficiency improvements are
considered. The levels are defined by cost-effectiveness groupings,
i.e., level 1 changes have the shortest paybacks and level 3 changes
the longest (though all are cost-effective). The levels incorporate
bundles of design features, devices, measures and/or equipment in the
following categories:
e
Building thermal integrity, including passive solar
measures.
Heating, ventilating, and air conditioning systems
and controls.
- 52
E S R G
® Internal. loads and comfort conditions.
· Operation and maintenance provisions.
Measures in the last category, O&M provisions, tend to drop
out of the level 3 technology combinations, which are the most
capital-intensive of the three. These three groupings are labelled
"m" in Table 4.4. In addition, an exogenous growth may be specified
for the fourth category (Auxiliaries and Power) to capture such
effects as increased use of electrical and electronic equipment.
The energy savings that the technology and modifications associ-
ated with each conservation level would achieve are provided in
Reference 6 for each United States region. These savings are to be
applied against the base line loads discussed above. The matrix of
percentage efficiency improvements is given in Table 4.5 by level,
building type and end-use. They are also broken down by new buildings
and 1975 stock ("retrofit").
The overall savings are functions both of the energy requirement
reductions related to the conservation level and the penetration of
these levels. Here, level "penetration" is defined as the fraction
of floorspace in the given year and BT/EU combination at the given
level. The average savings are then given by the sum over levels of
the product o~)level penetration ("PENt,i,k,m") and percent improvement
("PIMPt,i,k,m as given in Eq. 4.12.
The time dependence of the electrical use coefficients can then
be written as the initial value multiplied by a decreased demand
factor (Eq. 4.11).. The penetration of the conservation'level
technology groupings is dependent on a number of factors: initial
costs, consumer preference, capital availability, payback time and
electricity costs.. The penetration levels are calculated by using
an economic model which applies the estimated payback period to S-
shaped market acceptance curves. The levels of penetration which
result are functions of inputted economic assumptions. Consequently,
the forecast scenarios can incorporate sensitivity to a range of'
assumptions on, ~ future fuel costs. The electrical intensities
require, in addition to the electrical use coefficients, "saturation"
estimates (Eq. 4.1.3).
An additional, factor must be taken into account for the electric
space heat end-use: the possible use of heat pumps. Penetration
analysis suggests that electric space heat with a heat pump is
cost-effective over conventional electric resistance heating. The
model allows for a market response delay by phasing in the fraction
of new electrically space heating buildings which have heat pumps to
a specified level in 1985. Additionally, the model incorporates the
cautious assumption that solar heating and air conditioning will have
an insignificant impact on overall load during the forecast period.
In the case of water heating, where electricity consumption is
relatively insignificant, solar energy would substitute primarily
for fossil fuels.
- 53 -
E S R G
TABLE 4.4
ELECTRIC ENERGY INTENSITIES
Indices
t
i
k
n
m
Year (11975 = 1)
Commercial end-use (i = 1 to 4)
Building type (k = 1 to 5)
Existing or new buildings (n = 1 to 2)
Conservation levels (m = 1 to 3)
Variables
INTEN
EUC
SAT
PEN
PIMP
PENSUM
HPFRAC
COP
AUPFAC
Electrical intensity (average annual KWH/FT2)
Electrical use coefficient (= INTEN with all
saturations - 1)
Saturation (fraction floorspace with end-use)
Market fraction ("penetration")
Fractional energy savings (i,k,n)
at given conservation level (Table 4.5)
Fractional energy decrease
Fraction new electrically heated buildings
Heat pump coefficient of performance
Fractional increase of terminal year auxiliary and
power intensity over base year
Equations
From definitions:
EUCt,i,k,n = (1- PENSUMt,i,k,n) x ~UCl,i,k,n
(4.11)
where Z (4.12)
PENSUMt,i,k,n'= m PIMPt,k,n,m x PENt,i,k,n,m
and
INTENt,i,k,n = SATt,i,k,n x EUCt,i,k,n (4.13)
except for Auxiliaries and Power, where growth is incorporated:
INTEN _Il +AUPFAC x
t,4,k,n -~ 25 ] (4.13)
and for new electric space heating building where heat pumps
are phased-in:
INTENt,i,k,2 = (HPFRACt/COP + (1-HPFRACt)) x
SATt,i,k,2 x EUCt,i,k,2
where HPFRAC is given the following linear parameterization:
for
t~ll
HPFRACll
(4.14)
(4.!5)
- 54 -
E S R G
TABLE 4.5
FRACTION OF LOAD SAVED
Conservation Level
Building Type End-Use Retrofit Market New Market
1 2 3 1 2 3
Dffice Heating .11 .15 .23 .25 .35 .40
Cooling .13 .17 .34 .20 .35 .47
Lighting .25 .50 .50 .15 .25 .25
Aux.&Power .17 .28 .38 .10 .16 .20
Retail Heating .08 .23 .25 .30 .42 .50
Cooling .12 .20 .20 .25 .37 .46
Lighting .13 .25 .25 .15 .24 .30
Aux.&Power .18 .36 .45 .10 .16 .20
Hospital Heating .07 .15 .16 .20 .32 .40
Cooling .07 .24 .28 .15 .25 .33
Lighting .08 .12 .17 .10 .15 .15
Aux.&Power i .19 .25 .30 .10 .15 .15
Schools Heating .14 .21 .29 .30 .42 .50
Cooling .16 .26 .56 .25 .35 .41
Lighting .12 .30 .42 .15 .20 .20
Aux.&Power .26 .33 .53 .20 .25 .30
Miscellaneous Heating .09 .15 .26 .30 .42 .50
Cooling .05 .12 .24 .25 .35 .40
Lighting .09 .15 .24 .15 .15 .20
Aux.&Power .14 .23 .32 .15 .20 .20
55 -
E S R G
4.4 Energy Forecast
The computation of commercial sector energies is a straight-
forward exercise once the forecasts for floorspace and electrical
energy intensity have been obtained. The expressions for average
annual energy consumption by end-use and building types are given
in Table 4.6.
Calibration to base year sales is performed on total sales:
~ (4.16)
Commercial Energy Sales, year t =i,k,n ENCEUt,i,k,n
The model is first run from 1975 (t=l) to the base year (t = l+base
year - 1975). The total floorspace is then adjusted to normalize
total sales in a given service territory to base year experience.
An overall square foot adjustment factor scales each term in the
energy s~/n (Equation 4.1.7). The necessity for such an adjustment
is traced to the use of national average square foot per employee
data. One finds, as anticipated, that such data closely approximates
state averages except in service areas dominated by land-scarce
urban centers.
TABLE 4.6
COMMERCIAL ENERGY FORECAST
Indices
t
Year ,[1975 = 1)
Co~ercial end-use (i = 1 to 4)
Building type (k = 1 to 5)
Existing or new buildings (n = 1 to 2)
Variables
ENCEU
INTEN
OSQFT
NSQFT
Annual energy consumption
Corresponding electrical energy intensity
(See Table 4.4)
Remaining 1975 building stock floorspace
(See Table 4.2)
New floorspace (See Table 4.2)
Equations
Retrofit market:
ENCEUt,i,k,1
New Construction:
ENCEUt,i,k,2
= INTENt,i,k,1 × OSQFTt,k
t
= ~
t' =2
INTENt' ,i,k,2 × NSQFTt' ,k
- 56 -
$ R G
(4.17)
(4.18)
5. INDUSTRIAL SECTOR
As with the residential and commercial sectors, industrial
energy consumption is broken down into the product of energy
using activities and energy intensities of those activities.
The measure of activity in the case of industrial energy consump-
tion is employment for each major manufacturing subsector. The
subsectors are chosen at the two-digit Standard Industrial Clas-
sification (SIC) level. Less detail would lose sensitivity to
differing growth and electricity use trends among industries;
more detail would require inputs beyond the capability of the
current data base. The SIC's included in the forecast are given
in Table 5.1.
The electric energy intensity for the industrial sector is
correspondingly defined as average electricity consumption per
employee. The growth in employment is related to the level of
economic growth and business activity in the state, while the
electric intensity is a function of several major factors: pro-
cess technology, labor productivity (production per employee),
pollution control requirements, conservation level, and fuel mix.
In past decades, electrical energy growth has been driven by in-
creases in employment and production levels, energy-intensive-
ness in manufacturing processes, and increased fuel fraction for
electricity on the one hand and a virtual absence of energy con-
servation on the other. The job of forecasting is to adequately
characterize historical experience and to incorporate a realistic
range of growth lin the demand-driving factors.
5.1 Model Structure
The model elements and their relationship are schematized in
Figure 5.1. Growths in base year electric energy consumption by
SIC are related to growths in employment and electric energy in-
tensity. The resultant electric energy demand must then be di-
vided into the amount purchased and the amount self-generated,
since it is the purchased energy which is ultimately identified
with utility sales. Changes in the fraction of electric energy
consumption supplied by self-generated electricity must also be
allowed for.
The forecast energy thus depends on the specifications of
base year experience, the forecast of employment growth, the trend
in electric energy intensity, and the changes in fraction self-
generated. These will be discussed, respectively, in Sections
5.2 to 5.5 and brought together in the energy forecast model de-
scribed in Section 5.6.
- 57 -
E S R G
TABLE 5.1
STANDARD INDUSTRIAL CLASSIFICATIONS
ESRG
Index
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
SIC Description
20
22
23
24
25
26
27
28
29
33
34
35
36
37
3O
31
32
38
39
Food and Kindred Products
Textiles
Apparel and Other Textile Products
Lumber and Wood Products
Furniture and Fixtures
Paper and Allied Products
Printing and P~blishing
Chemicals and Allied Products
Petroleum and Coal Products
Primary Metal
Fabricated Metal Products
Machinery (except electrical)
Electric Equipment
Transportation Equipment
Rubber and Plastics
Leather
Stone, Clay and Glass
Instruments, Related Products
Miscellaneous
E
- 58 -
R G
Base Year
Electric Energy
Consumption by SIC
!Self-Generated
Electricity by SIC
Purchased
Electricity by SIC
Temporal Factors
Growth in Employment
iGnr~nthsi~ Electric Energy[ I
Electric Energy SIC]
:iConsumption by
Change in Fraction
Self-Generated
Self-Uenerated [
Electric__lty by SIC ]
[Total Industrial Sales]
FIGURE 5.1
INDUSTRIAL SECTOR MODEL SCHEMATIC
5.2 Base Year Experience
The model requires inputs on base year industrial sales and
self-generated electricity by two-digit SIC. Statewide data is
available from public sources (see e.g., Reference 12). Fractional
breakdowns of base year sales by service territory and SIC are
generally available and may also be generated from statewide data
on the basis of county employment by industrial grouping (Refer-
ance 9) and on county to service area allocation matrices. This
is, of course, not necessary for statewide forecasts.
5.3 Employment Growth
The measure of industrial activity used in the model is the
level of employment. Employment is closely related to actual phys-
ical production and detailed employment data is much more readily
available than other measures, such as value added, value of ship-
ments, etc. Electrical intensity (discussed in the next subsection)
is then expressed in kwh per employee.
The Bureau of Labor Statistics, the Bureau of Census, and a
number of state and local agencies collect employment data and
predict future trends. The employment values used in the indus-
trial model are for the total production and non-production em-
ployees, which is generally typical of the reported data. Most
of the data sources tend to be consistent, with the variations
between sources providing a cross-check and a possible range for
alternative forecasts.
The employment values used in the model are the ratios of
future year employment to base year employment by industry class,
expressed as a State Employment Index (SEMPI). Consequently,
the absolute employment figures are not of importance to the
model.
5.4 Electrical Energy Intensity
Electrical Energy Intensity (SEI) is defined as electricity
consumption per employee. Its value, for any given two-digit SIC,
has changed over time as a result of the adoption of capital-in-
tensive production technologies aimed at increasing labor produc-
tivity and price trends of electricity (and other production in-
puts) that have tended to affect industrial energy management prac-
tices..
The electrical intensities for the major industrial classes
in each state are calculated by testing a variety of multiple re-
gressions on the historical data, using such explanatory variables
as labor productivity, electrical price, and time trends. The
best explanatory model is then used to predict future values of
the intensities. We discuss in Chapter 8 the details of these
methods.
- 60 -
E S R G
5.5 Fraction Self-Generated
Up to this point, industrial electricity consumption has
been forecasted on the basis of the total demand for electricity
on the customer's side of the meter:
Total kwh Demand = Employment x Intensity
where intensity is expressed in terms of unit employment require-
ments. Only part of this demand must be met by the utility, how-
ever, since many industries produce some of their electricity re-
quirements in-house.
Therefore, an additional factor--the fraction of total elec-
trical energy consumption which is self-generated--is necessary in
computing forecast industrial sales. This fraction may change over
present values as a result of national energy policy, developing
state interest in addressing regulatory and other barriers to such
investment, and renewed interest among,industrial planners in com-
bined energy systems as a result of the increasing costs of elec-
tricity. Therefore, the historic decrease in the fraction self-
generated is likely to reverse. The degree will depend on scen-
ario assumptions based on existing studies of cogeneration poten-
tial and on historic levels experienced in the state.
5.6 Energy Forecast
The basic elements required for the industrial sector have
now been described. They are brought together in the energy fore-
cast model summarized in Table 5.2. The fractional breakdown of
industrial sales (Equation 5.1) is used to define base year sales
by SIC. Total energy is derived from purchased energy using base
year values for the fraction self-generated (Equation 5.3). The
growth of Equation 5.4 is based on the growth in state employment
index (Section !5.3) and electric energy intensity (Section 5.4).
Finally, forecasted total energy consumption is decreased by the
self-generated component to arrive at the forecast for industrial
sales (Equation 5.5).
- 61 -
E S R G
TABLE 5.2
INDUSTRIAL ENERGY FORECAST
Indices
t
Year (base year = 1)
Industrial grouping by two-digit SIC (j ~ 1 to 19)
Variables
TESIC
PENSIC
ISALES
SEI
EMP
SEMPI
MIX
SGEN
Total electric energy consumption
Purchased electric energy consumption
Base year total industrial sector sales
Electric intensity
State Industry Employment
State employment index
Fraction base year industrial sales breakdown
Fraction self-generated
Equations
Initialize (t ~ 1):
SEMPIt,j ~, EMPt,j/EMP1,j
PENSIC1,j = MIXj x ISA/~ES
TESIC1,j ~ PENSICi,j/(1 - SGEN1,j)
Then for t · 1,
TESICt,j =, SEMPIt,j x(SEIt,j/SEII,j) x TESIC1,j
PENSICt,j = TESICt,j x (1 - SGENt,j)
(5.1
(5.2
(5.3
(5.4
(5.5
- 62 -
6. OTHER ENERGY REQUIREMENTS
The residential, commercial, and industrial sectors account for
the bulk of energy consumption. The residual categories are street
and highway lighting, transit systems, company use, losses, and sales
for resale. Of these, the last item represents KWH sales to other
electric utilities. Since we are interested in only demand for
electricity on the utility system (not on itself), this category can
be ignored.
The category "losses" refers to electric energy lost in the trans-
mission and distribution lines in the course of serving system customers.
Utility "company use" is the energy consumed by the electric utilities
themselves in business operations. These two categories -- losses and
company use -- are accounted for in the model by a fraction of total
sales (FP~LSt in Table 6.1). That is
FP~LSt = Ii°sses and c°mpany usel
total sales t
Total sales includes, in addition to the three main sectors discussed
in earlier sections, sales for transit systems and street and highway
lighting. Total energy from the "other" sector is then derived from
Equation 6.1. Yearly energy sales from the three main sectors are
inputted from the respective sectoral models, base year data for
"other" sales, losses and company use are readily available from
utility records. Deviations from base year values are provided by
Company forecasts .or can be independently estimated.
Indices
t
Variables
FRLSt
SUMt
OSALESt
OTHENt
Equation
TABLE 6.1
OTHER ENERGY
Year (base year = 1)
Losses and company use as a fraction of total sales
Sum of energy sales to residential, commercial
.and industrial sectors
Energy sold for street and highway lighting and
railroads
Energy sendouts in "other" category
OTHENt = OSALESt + FRLSt × (SUMt + OSALESt) (6.1)
- 63 -
7. PEAK POWER
In the preceding sections, we have concentrated on the
electrical energy forecasting model. Here, we shall turn
instead to the method for translating these results into
peak power demand forecasts ("demand" henceforth). Power,
being the rate at whic!h energy is expended, will be expressed
in units of 1000 KWh/hour or simply MW.
In developing the systemwide peak power forecasts, one
strives ideally to model the contributions of each end-6se
category separately. Peak power forecasts which are based
on gross load factor analysis (defined as average demand
divided by peak demand) lose the ability to adequately
track changes over time in the relationship between energy
and peak due to differences in growth rates between the
end-use categories and shifts in load pattern as a result,
for example, of load management programs.
While the current status of load data and research do
not permit a completely disaggregated treatment, sufficient
information exists to capture the primary effects. The
approach adopted here analyzes peak power dema~-ds as the
summation of the contributions of the various end-use
categories at the time of the system summer and winter peak.
Thus, the impact on peak of the end-use forecasts is treated
explicitly, as are the effects of any forecasted shifts in
usage pattern (resulting from time-of-use rate initiatives,
direct control of equipment and so on).
The computations are summarized in Table 7.1 The
structure is rather straightforward. The forecasted annual
energy requirements for the twenty-one consumption categories
listed in Table 7.2 are multiplied by a "peak factor" to
form the contribution of each to system peak. This is
indicated in Eq. 1 where the peak contributions, unadjusted
for possible future shifting of load patterns, are formed
as the product of the peak factor and forecast annual energy
consumption. The factor 8.76 (thousand hours in a year) is
included so that the peak factor will be defined as the ratio
of the end-use peak at time of the system peak divided by
its average peak load over a year (recall that loads are
expressed in MW and energies are expressed in GWh or 103 MWh).
For nineteen of the peak factors, substantial data is available,
and they are developed, exogenously (see S~¢. 8 for data
discussion). Additionally, the "other energy" category
which consists primarily of line losses may be computed via
Eq. 7.2. The heat pump category requires special treatment.
- 64
E S R G
T~e coefficient of performance of heat pumps decreases
with temperature. If the peak factors associated with
resistance heating were used directly, underestimates of
winter peak demand would result since more kwh of electrical
input per kwh of heating output are required on colder days.
This efficiency loss is reflected in Eq. 7.3 where the winter
heat pump peak factors (summer heating peak factors are of
course zero) are written as a factor times the corresponding
resistance heat peak factor. These coefficient of performance
correction factors (COPCR and COPCC for the residential and
commercial sectors, respectively) are defined as the ratio
of average to coldest day COP. The method and assumptions
for estimating the corrections are discussed in detail in
Sec. 8.
For most categories, the temporal variation of use, and
thus the peak factors, are fairly stable from year to year
since they reflect statistical averages of behavioral patterns.
For example, buildings use more electricity during business
hours, lighting requirements vary regularly with season and
hour, refrigerators respond systematically with annual
temperature variations, and so forth. In addition to these
basic peak factors, however, the model incorporates the option
to include the impacts of load shifting and peak reduction
impacts as well. These are represented by the peak reduction
variables in Eq. 7.4 which are applied to the estimated un-
adjusted peak demands. The ultimate level of peak reduction
assumed is phased linearly from the starting and the attain-
ment year which are also specified by the modeller. A number
of load management options are thus available for consideration
such as controlled water heaters, space heat storage, air
conditioner cycling, time-of-use rate response, voltage
reduction, etc.
Finally, total summer and winter system peaks are
calculated as the sum of the component contributions as
shown in Eq. 7.5 and appropriately normalized to experienced
base year peaks.
- 65
E $ R G
TABLE 7.1
PEAK POWER MODEL
Indices
Year (base year = 1)
Season (l=sununer, 2=winter)
End-use (see Table 7.2)
Variables
OTHEN
E
PKFAC
PR
UPEAK
PEAK
COPCR
COPCC
"Other" energy requirements, losses, etc.
Annual energy consumption for end-use
Contribution of end-use to peak
Peak reduction/load control factor
Total peak by end-use category, unadjusted
for load shifting
Total peak by end-use category, adjusted
for load shifting
System peak load
COP correction factor -- residential
COP correction factor -- commercial
Equations
Develop component peak contributions unadjusted
shifting:
for load
UPEAKt,p,i .= ~[PKFACp,i × Et,i)/8.76
where
21 21
PKFACp,21 = 1t.76 × (~UPEAKL .)/(~ Ei)
i=l u'P'~ i=l
and the heat pump peak factors for residential and
commercial sectors are:
PKFAC2,13 ~ COPCR × PKFAC2,12
PKFAC2,17 = COPCC × PKFAC2,16
Adjust for peak reduction
APEAKt,p,i ~ UPEAKt,p,i ×
Then
22
= ~APEAKt,n,i
PEAKt,p i=1 =
(1-PRt,p,i)
(7.1)
(7.2)
(7.3
(7.4
(7°5
$
- 66 -
G
TABLE 7.2
END-USES USED IN PEAK POWER MODEL
INDEX
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
SECTOR END-USE
Residential Refrigerators
" Freezers
" Ranges
" Lighting
" Televisions
" Clothes Dryers
" Clothes Washers
" Dish Washers
" Water Heaters
" Room A/C
" Central A/C
" Resistance Heating
" Heat Pump
Heating Auxiliary
" Miscellaneous
Commercial Resistance Heating
" Heat Pump
" Cooling
" Lighting
" Auxiliary and Power
Industrial All
Otlher Energy All
- 67 -
E S R
8. MODEL INPUTS AND FORECAST ASSUMPTIONS
The demand forecast model previously described in this study
defines electricity demand at the point of consumption. It
provides a framework for mapping input data by defining base year
experience, equipment ownership growth, technology and demographic
shifts, and fuel switching onto a set of output results (system,
sectoral, and end-use energy forecasts; and peak demands.) This
section describes the data utilized and assumptions made in
generating the High and Low Case forecasts of both energy and peak
demand in the LILCO territory. The Base Case forecast is the
mid-range of the High-Low band. These uncertainty bands in input
parameters are intended to represent the plausible range of
business-as-usual futures.
The computer program that ESRG has developed to implement the
model allows for flexibility in the choice of both inputs and outputs.
The data explication below is intended to correspond closely with the
model descriptions given in Chapters 3 through 7. The reader
should be able to understand fully the basis for the forecasts
reported in Chapter 1 with cross-referencing to the discussion of
the mathematical structure of the appropriate sub-model. The base
year, 1982, serves as the departure point for forecasting.
8.1 Energy Sales by Sector
Details of 1982 electricity sales by sector were available in
Refs. 1, 13, 15, and 17. These sales amounted to 12,519 GWH.
Including losses and Company use, total energy requirements
for the year amounted to 13,713 GWH. The present forecast seeks
to normalize the 1982 energy requirements upward to partially
counteract the effect of the recession year on sales. The
Company's normalized summer peak of 3,070 MW (Ref. 1) and a
representative load factor of .5115 (again Ref. 1) provided the
basis for an adjusted energy figure of 13,756 GWH for 1982. Sales
for each sector were increased by the ratio of this adjusted
energy figure to the experienced 13,713 GWH.
Experienced residential sales were 5,556 GWH. combined
commercial and industrial sales were 6,525. These were divided
into 5,324 GWH for commercial sector sales; and 1,201 GWH sales for
manufacturing and mining customers (the ESRG industrial sector).
Other sales, including street lighting, other public authorities,
sales for resale to Rockville Centre and Freeport, and Long Island
Railroad, amounted to 438 GWH in 1982. Each sectoral sales figure
was revised upward by a factor of 1.0031 as described above.
Table 8.1 provides these values.
- 68 -
E S R G
TABLE 8.1
ADJUSTED LILCO SALES BY SECTOR~ 1982
Sector
Sales (GWH)
Residential 5,574
Commercial 5,340
Industrial 1,205
Other Sales 439
Total Sales
12,558
8.2 Residential Sector
Here we will review the data inputs to the residential forecast
model described in Chapter 3. Broadly, growth in the residential
sector is a function of changes in customers (housing stock),
appliance saturations, and patterns of unit energy usage of
appliances. The following sections provide discussions of each of
these factors.
8.2.1 Residential Customer Forecast
Residential customer growth in a given service area derives
from growth in population and/or changes in household formation
patterns, where fewer persons per housing unit result in increased
numbers of homes. While the two Long Island counties and Rockaway
had experienced continuing population growth during the 1970's (Ref.
18), available data from LILCO research indicates a levelling and
slight decline thus far in the 1980's (Ref. 19). While Suffolk
County continues to grow modestly, Nassau County has experienced a
counteracting decline. Figure 8.1 below is taken from the LILCO 1982
Population Survey (Ref. 19).
None of the available forecasts for the region predicts zero
population growth for the Nassau-Suffolk region. The two most recent
region-specific population projetions available to the present
forecasters are the preliminary population projections by the New York
State Department of Commerce (Ref. 20) which will be subject to
revisions due to an update of the migration assumptions, and the
Nassau-Suffolk SMSA projections by the Bureau of Economic Analysis
(Ref. 7). The Company's present population assumptions are not
readily available from public documents. Table 8.2 provides a
summary of the two forecasts mentioned above.
- 69 -
FIGURE 8.1
GROWTH
1.5
1.0
POPULATION
(millions)
OF NASSAU AND SUFFOLK COUNTIES
FROM 1950 TO 1982
1960 1970 1980 1990
.5
0
1950
YEAR
TABLE 8.2
COMPARISON OF POPULATION PROJECTION $OURCES~ LONG ISLAND
Ref. Ratio to 1982 Population*
Source No. Area 1985 1990 1995 2000
1) 1980 Bureau 7 Nassau- 1.018 1.028 1.019 1.010
of Economic Suffolk SMSA
Analysis
2) New York 20 Nassau- 1.008 1.038 1.068 1.093
State Dept. Suffolk SMSA
of Commerce
(1983)
* Intermediate years were interpolated where specific data was not
given in source.
Household size on Long Island differs distinctly from both
national and statewide persons per household. LILCO estimates a
1982 value of 3.10 persons (Ref. 19), slightly less than the 3.16
Census estimate for 1980. This compares with 1980 Census values
of 2.70 persons per household in New York State as a whole, and
2.75 p.p.h, nationally. The U.S. Census Bureau is presently re-
vising its national projections of household size to include
actual 1980 Census values in their data base. Given the
difference between the historic regional and national household
formation patterns, the accuracy of these figures is questionable.
The present forecast of future households adopts the customer
projections prepared by LILCO and provided in Ref. 1. These
projections appear to be in line with recent experience in
customer growth in the service area. If population in the region
were to continue to decline, these household projections may be
high over the full forecast period, but there are no present
indicators of a major shift in the growth of households.
Given the differentiated usage levels of many appliances
between single and multifamily homes, the ESRG demand model
requires that base year and projected residential customers be
divided into housing type. While LILCO does not provide data by
housing type in its saturation survey results (Ref. 1), reasonable
estimates of present housing mix can be derived via comparison of
1980 housing mix by county (Ref. 25) with comparable 1970 Census
data. These sources indicate that 80.9% of homes in the region
- 71
E S R G
were single family in 1980. They also indicate that the net
effect of housing stock shifts between 1970 and 1980 in new homes
built and old housing stock retirements was that 60% of the net
housing added during the 1970s were single family. This statistic
as well as LILCO's estimate of new homes added during the 1980s
(Ref. 19) serve as a basis for a 1982 estimate of housing mix set
at 80.5% single family and 19.5% other. The 1970s differential
housing stock data also serves as a basis for projecting the mix
of new homes in the forecast period. The High Case assumes a 65%
single family, 35% multifamily split, while the Low Case assumes a
continuing trend to 55% single family and 45% multifamily.
Housing starts and permits data given in Ref. 40 indicate a
stronger percentage of state-wide multifamily homes can be
anticipated. However, since Long Island has historically lagged
behind the state, these data were rejected.
Based on the assumptions discussed above, the residential
customer forecast is provided in Table 8.3.
TABLE 8.3
RESIDENTIAL CUSTOMER FORECAST~ LONG ISLAND LIGHTING COMPANY
(102 Units)
1982 1985 1990 1995 2000
Housing*
Units Base Year High Low High Low High Low High Low
Single-
family 660.6 674.0 671.9 695.6 690 2 714.2 705.9 732.4 721.3
Multi- '
family 165.1 172.3 174.4 184.0 189.4 194.0 202.2 204.4 215.4
TOTAL 825.7 846.3 846.3 879.5 879.5 908.1 908.1 936.7 936.7
* Columns may not sum exactly due to rounding.
8.2.2 Appliance Saturation and Unit Usagm
Appliance saturation assumptions are s~marized in Table 8.4.
The table includes a growth parameter used to develop the S-shaped
saturation logistics curves for each end-use. The end-uses not
included in the table have predetermined saturation values. Lighting
and miscellaneous are each fixed at 1.0, while heating auxiliary
saturation is defined as 1.0 minus the electric space heat saturation.
- 72 -
E S R G
The base year saturations are based on regressions on Company
data given in the 1983 New York Power Pool Report (Ref. 1). Growth in
appliance saturations are developed by calculating an index ("A") for
each appliance using the logistic curve fitted to the historical
appliance saturation data given in the same refe=ence. Terminal
saturation levels are estimated by examining historical growth in
saturations via the regressions, present levels, LILCO assumptions of
future saturations, and the econometric relationships between
appliance saturations and the energy price/income variables. Ref. 14
(pp. 6-1 to 6-12) and Ref. 23 (pp. 17-19, 32, 33) provide fuel-price
elasticities which can be used to estimate appliance choice for many
of the residential appliances, given energy price growth assumptions.
In terms of price assumption inputs, the following growth rates were
assumed in estimating terminal saturations:
Electricity Natural
Income Price Gas Price
High 2% 0% 6%
Low 1% 2% 4%
Table 8.4 provides results of the appliance saturation analysis.
TABLE 8.4
LILCO APPLIANCE SATURATION ASSUMPTIONS
Base Year High Case Low Case
Terminal Terminal
Satura- Satura- Growth Satura- Growth
Appliance tion tion Paramete~ tion Parameter'
Refrigerator (SF) 1.22 1.22 0 1.22 0
Refrigerator (MF) 1.00 1.00 0 1.00 0
Freezer (SF) .37 .60 .023 .37 0
Freezer (MF) .09 .15 .015 .09 0
Range .47 .70 .016 .50 .076
Television (SF) 2.40 3.50 .043 3.00 .060
Television (MF) 1.10 1.50 .015 1.30 .025
Clothes Dryer .53 .90 .029 .53 0
Clothes Washer .86 .90 .067 .86 0
Dishwasher .52 .80 .005 .52 0
A/C Room 1o15 1.25 .162 1.15 0
A/C Central .13 .30 .075 .25 .084
- 73 -
E S R G
The demographic and saturation computations generate the number
of electricity-consuming units. The model then requires input as to
the use per unit KWH consumption. LILCO provides an analysis of
residential appliance saturation and usage in the NYPP Report (Ref. 1)
for future years except for refrigeration and freezer usages (for
which a detailed analysis has been performed. Backcasted extrapola-
tions of unit usage assumptions given by LILCO are the basis of the
values given in Table 8.5. These usages, as well as efficiency
improvements, are discussed in Section 8.2.5.
Consumption levels for single-family and multifamily homes were
adjusted to capture the effects of variations in the size of the living
space and number of inhabitants. The averages were disaggregated over
housing types based on the following typical SF-to-MF unit use
relationships: 4 to 3 for refrigerators, freezers, and water heaters;
2 to 1 for central air conditioners, heating auxiliaries, and electric
space heaters. The annual appliance usages based on the above
estimates are given in Table 8.5.
Table 8.5 provides the efficiency improvements assumed in the
forecast runs, as well as base-year average unit usages. The basis for
the latter efficiency is the FEA standards (Ref. 24), adjusted for
improvements already incorporated in recent appliance stock vintages.
The efficiency improvements employed are significantly less than those
used by LILCO. Usage levels for televisions are described below;
Section 8.2.5 provides usage assumptions for refrigerators.
· The FEA program achievement data was 2980. The forecasts
move these target dates back to 1984 and 1987 for the
low and high estimates respectively.
· Though not shown on the tables, decreases in hot water
requirements have also been included at FEA-targeted
levels (37% for clothes washers, 17% for dishwashers,
with 1984 and 1987 phase-in dates for the Low and High
Cases, respectively). Though such increased thermal
efficiencies do not directly affect electrical consumption,
they will impact indirectly on electric hot water
requirements.
· Energy reductions for televisions are not assumed to
follow the targets; the forecasts assume no further
improvement.* The present forecast assumes that
this phase-out of tube-type models is complete for
new units and no further efficiency improvement will
be experienced.
- 74 -
£ s n o
Efficiency improvements for air conditioners are based on
an analysis of shipment data through the year 1981 (Ref. 34
for central A/C and Ref. 37 for room A/C) and the proposed
DOE efficiency improvements for these appliances (Ref. 26).
It is assumed that no further improvements will be made on
new A/C appliances for the High Case; during the
forecast, the average unit usage will improve to the level
of models already on the market. Efficiency is projected
to improve for the Low Case to the levels proposed in
Ref. 26.
Lighting
Smaller family size, trend in housing mix toward smaller units,
conservation induced by increasing prices, more efficient fluorescent
bulbs and new lighting technology lead one to suspect that lighting
energy demands are decreasing. Such major manufacturers as General
Electric, sylvania, Westinghouse, Norelco, and Duro-Test are all
involved in the development and marketing of higher efficiency
light-bulbs, with typical energy savings in the range of 50-70%.
We have made the cautious assumption that none of these demand
decrementing factors or new products significantly impacts lighting
levels for the higher-growth scenario. We have included a modest
estimate of the impact of more efficient bulbs and energy decreasing
factors for the Low Case. An energy reduction level of 50% is
achievable for the more efficient bulbs currently on the market or in
advanced development. In the Low Case, it is assumed that penetration
levels reach 20% by the end of the forecast period, yielding an
overall 10% reduction in lighting energy use.
TABLE 8.5
RESIDENTIAL APPLIANCE USAGE AND EFFICIENCY IMPROVEMENTS
Base-Year Average Forecast Percent
Usage (KWH/Year) Energy Reduction
Appliance SF Both MF High Low
Eefrigerator 1440 1080 .05 .05
Freezer 1180 880 .05 .05
Range 690 .023 .023
Television 282 0 0
Clothes Dryer 970 .03 .03
Clothes Washer 87 0 0
Dishwasher 320 .14 .14
Water Heater 4200 3150 .09 .09
Room A/C 360 .07 .26
Central A/C 2970 1485 .12 .33
Electric Space
Heat 10130 5065 0 0
Heating Auxiliary 500 250 0 0
Lighting 995 0 .10
- 75 -
E S R G
8.2.3 Electric Space Heat and Water-Heating Penetrations
The forecast of electric space heat (ESH), electric water heat
(ESH), and heating auxiliary usages are treated differently from the
other appliances. These usages are treated as penetrations, i.e.,
fractions of new housing units possessing the end-use. Table 8.6
provides the base-year saturations and assumed penetration rates of
electric space heat for the LILCO service area. ESH penetration rates
during the past ten years have averaged approximately 24%. The 1982
saturation was assumed to be 3.594 based on data in Ref. 1. On the
basis of an historic two-to-three ratio of saturations between the two
housing types, this value was split into single family and multifamily
saturations. The High Case assumes a substantial increase in ESH
penetration levels due to the potential of increasing costs of
competitive home heating fuels and the desirability of using electric
heat pumps as year round thermal conditioning devices. The Low Case
assumes a levelling of growth in ESE based on its historic.penetration
level and the likely strong increase in the relative price of
electricity. In both cases, the initial penetration rate is assumed
to approximate the 1981-82 penetration rate of 37%. Multifamily units
are assumed to continue to experience higher penetration rates than
single family units.
The penetration of electric water heating customers is assumed to
be tied to new space heat customers, resulting in equally high growth
in EWH demand. The base year saturation of EWH was taken as 8.8%
based upon a regression of historic saturations of this end-use
(Ref. 1). Heating auxiliaries' saturation is taken to be the percent
of residential customers which are not space heating customers.
TABLE 8.6
LILCO ELECTRIC SPACE HEAT PENETRATION
1990 Rate of
Housing Base-Year Initial Rate Penetration
Units Saturation of Penetration High Low
Single-
family .032 .30 .50 .20
Multi-
family .048 .45 .75 .30
8.2.4 Thermal Integrity Impacts
The building envelope characteristics of a dwelling unit
influence the heating and cooling load which is to be satisfied by
heating and cooling systems. In terms of electricity consumption,
significant impacts of building thermal integrity are on three
temperature-sensitive end-uses: electric space heating, electrically
driven auxiliaries (fans or pumps) associated with fossil heating
systems, and air conditioning.
- 76
the
Naturally, other factors such as back-up systems, portable space
heaters, solar, and wood heat influence the electricity consumed for
these three end-uses. (See Section 8.2.5 below.) In addition, there
are the efficiency factors associated with the conversion devices
themselves (e.g., air conditioner and heat pump coefficients of
performance). These other factors are treated separately and
additively by the model.
ESRG's HOMES model was used to estimate the impacts of thermal
integrity improvements. This model performs heat loss calculations
according to standard ASHRAE procedures (Ref. 39). The location
specific thermal integrity data is based upon information from the
Long Island jobs study (Ref. 45, Appendix B) and the state insulation
surve~ (Ref. 46).
For electrically space-heated homes, the available sources
indicate that higher levels of insulation and thermal integrity will
exist in newer structures. The present forecast assumes that
electrically heated homes built during the forecast will use less
electricity for heating than existing ESH homes, with the reductions
being 10.6% for single family dwellings and 13.2% for multifamily
dwellings. For fossil heated homes, we applied a reduction to the KWH
usage of the electrical heating auxiliaries of fossil fuel heating
systems. The comparative savings are substantial for new homes.
Savings are set at 30.6% for single-family homes and 31.4% for
multifamily units. These thermal integrity improvements are applied
to new units.
The model also calculates a gradual Smprovement in existing
fossil and electrically heated homes. In principle, all homes
remaining in the housing stock could be retrofitted to a level close
to that of new units. In practice, it is necessary to make a
judgement of the extent of retrofit activity during the forecast
period. The existing housing stock was characterized by several
different categories of thermal integrity. Of these, only those homes
in the categories with inadequate thermal integrity are likely to
invest in improvements. It is also true that many of the homes with
inadequate insulation are likely to remain at their'present low
levels. The assumption made in this forecast is that 50% of the
existing single-family and 25% of multifamily homes are retrofit by
the end of the forecast period. The retrofits reduce KWH usage by
existing heating auxiliaries for fossil heated homes. This reduction
was calculated to be approximately 11% for single units, phased in
linearly from zero in the base year to the full fraction in the year
2000. The reduction is 6% for multifamily homes. No reduction was
assumed for weatherization retrofit of existing ESH buildings.
Using the prototypical units and thermal integrity levels from
the above analysis, fractional reductions in air conditioner usage
were also computed. The resultant fractional reductions are used as
given in Table 8.7.
- 77 -
E S R G
TABLE 8.7
THERMAL INTEGRITY CHARACTERISTICS~
LILCO AIR CONDITIONING
Housing Units
Single-family
Multifamily
Central A/C
New Existing
.025 .010
.032 .032
Room ~
A C
New Existing
.147 ! .o5o
For further discussion of the assumptions and methodology upon
which the thermal integrity improvements are based, see Appendix A in
ESRG's The Conservation Alternative to the Power Plant at Shorehamf
Long Island (Ref. 47).
8.2.5 Additional Data Requirements
The basic model structure utilizes base-year and forecast counts
of residential customers, saturations of the various end-uses, and
usage per unit to derive sectoral energy demand. However, there are
a number of additional factors peculiar to some appliance types which
influence usage and thus overall demand. The following is a brief
discussion of each of these factors, by end-use.
Refrigerators and Freezers
Due to the introduction of frost-free appliances in the 1960s and
1970s, followed by the rapid improvement in efficiency levels of new
appliances during the mid-1970s to the present, the Submodels for
residential refrigerators and freezers require additional inputs to
capture the complex unit usage characteristics. The methodology used
for both appliances establishes the unit usage of new appliances at a
point in one average lifetime before the base year (UNOLD), the unit
usage of new appliances in the year (YRMAX) of maximum usage (UNMAX)
due to the phase-in of the automatic defrost feature, and the unit
usage of new units in the base year (UNNEWT). Using these values, an
annual average usage of appliance units on-line in the base year
(UNAVBS) is then calculated as a weighted average of the usages of the
different vintage units.
Values for the representative size of 1962 vintage refrigerators,
the percent of new units featuring automatic defrost, and the relative
usage of these units and standard refrigerators are derived from
analyses provided in Refs. 28 and 30. Recent research (Ref. 21)
provides typical unit size and usage for 1975, which is set as the
YRMAX. These data are corroborated by Ref. 28.) For new units in
1982, research by the Association of Home Appliance Manufacturers
(AHAM, Ref. 22) cites an average 58% improvement in refrigerator
efficiency between 1972 and 1981. Given estimated unit usages in
1975, a 35% reduction in usage levels of new units was calculated by
1982. (This estimate is supported by the analysis in Ref. 21.) The
values for the refrigerator variables defined above are given as:
- 78 -
E $ R O
UNOLD UNMAX UNNEWT UNAVBS YRMAX
KWH/YR. 1040 1610 1050 1330 1975
These values were the basis for inputs into the Base Case runs
for LILCO, with one revision: these usage levels were recast in terms
of units in single family and multifamily homes on a 4 to 3 ratio.
The same methodology was applied to freezer usage level as for
refrigerators, and the same sources (Refs. 28, 30, 21, 22) provide
parallel analyses. The values established are as follows:
UNOLD UNMAX UNNEWT UNAVBS YRMAX
KWH/YR. 860 1440 890 1150 1975
Electric Ranges
The electric range submodel requires input data to simulate the
effects of microwave oven penetration. There are two inputs: what
fraction have microwaves and what is the effect of those that do on
energy use? Field studies cited in Ref. 32, p. G-23 indicate that
electric ranges with microwave ovens require 84% of the electric
energy which would otherwise be consumed. An energy demand factor is
thus set at 0.84 in the forecasts. Historic LILCO saturations (Ref.
1) yield a 1982 value of 8.2% fraction of range saturation. The New
York Power Pool Report (Ref. 1) also provides est/mates by LILCO of
future microwave saturations. ESRG adopted these values forJthe
present forecast. Both High and Low Cases were set at 11.3% in 1985
and 24.1% by 1995.
Televisions
The television submodel requires two additional items of data.
One item is the decreased use factor for adjusting the unit energy
requirements for second and third televisions. Estimates in the
literature vary widely. LILCO estimates that second units usages are
75% of primary set usages. The present forecast incorporates a
moderately high estimate of 50%. The second item is the mix between
black-and-white sets and color sets in future television sales. Based
on historic residential appliance saturations of LILCO (Ref. 1), the
model uses an estimate of 39% for the base-year fraction of
black-and-white sets. The forecast uses broad-ranging inputs in
assuming that black-and-white TV phases to 6% of units (a LILCO
assumption, Ref. 1) for a high-growth scenario, and r-mains constant
in the low-growth scenario.
- 79 -
~ S R G
Water Heaters
In addition to the assumptions already discussed, the electric
water heater submodel requires some additional data inputs. The first
characterizes the change in home hot water requirements for end-uses
other than dishwashers and clothes washers. (These are discussed
above.) This factor is capable of reflecting the effects of slow-flow
shower devices and energy-conserving faucets. Such plumbing fixtures
can up to 25% of energy for hot water. In the High Case, we assume no
move to such fixtures. In the Low Case, we project an ultimate market
factor of 25% for an energy savings of 9% (.35 x .25). The model
phases up to these full savings over a 20-year period beginning in
1982.
The other inputs concern the range of likely impact of
alternative technologies such as solar-assisted hot water, storage,
and heat-pump hot water. The LILCO forecast projects substantial
growth in these hot water technologies through 1999 (Ref. 1). The New
York State Energy Office also projects clear growth in solar-assisted
water heating (Ref. 38). ESRG has included moderate penetrations of
solar water heaters. In the Low Case, we assume 5% of new
electric water heat takes the form of the above technologies. In the
High Case, where strong penetrations of electric water heating are
projected, 30% of new water heating is assumed to be solar over the
forecast.
The electrical energy usage needed for backup of solar water can
vary regionally. Estimates used by other electric utilities
nationally vary from 25-75% of energy provided by solar; the remainder
by electrical back-up. The present forecast incorporates a value of
46% solar/54% 'electrical as the energy mix based on LILCO assumptions
in Ref. 1.
Electric Space Heat
There are a number of space heating-related elements in the
submodel that need to be specified in reference to solar,
heat-pump penetration and performance, and supplemental wood heating.
These will be discussed in turn.
While solar-assist features should make some inroads into the
water heating market, the economics and reliability of solar energy
does not tend to imply the same level of penetration in the
space-heating market. For the High Case, it is again assumed that
solar energy will not make an appreciable impact in the forecast
period. The Low Case assumes that by the year 2002, 5% of new ESH
penetration in single family homes and 10% of multifamily homes, will
have solar assist and/or storage features that will decrease electric
energy use 50%.
- 80 -
Areview of LILCO's projected appliance saturations in Ref. 1
underscores the assumption that heat pumps will provide a substantial
contribution to the growth of electric space heating in the service
area. Given that LILCO does not provide estimates of the penetration
to date of heat pump technology in the space heating market, data for
the Nassau-Suffolk SMSA was taken as an indicator for the region (Ref.
49). The fraction of existing ESH assumed to be heat pump was taken
as 15% in 1982, with the initial penetration rate of heat pump
as a fraction of new ESH also set at 15%. The High Case, predicated
on assumptions of strong ESH penetrations, assumes that the fraction
of new space heat units that are heat pumps will grow to 75% by the
year 2000. The Low Case assumes a heat pump fraction of 25% by the
year 1992.
Heat-pump coefficient of performance (COP) is defined as the
ratio of KWH heating .output to KWH electricity input (for operating
the compressor and fans). Data on unitary air-to-air heat pumps is
taken from an evaluation performed at Argonne National Laboratory
(Ref. 16) for typical models. We have assumed capacities of
approximately three tons and fifteen tons for the residential and
commercial sectors, respectively.
COP varies both with size and outdoor temperature. The model
requires average and low temperature values in order to estimate
average COP over the heating season in forecasting energy, and the
lower COP operating at the lower temperatures of the winter peak.
calculation based on Long Island temperature data (Ref. 42)
produced the following COPs:
A
Residential
Commercial
Average (for energy)
Low (for peak)
2.24 2.50
1.75 1.85
As with other appliances, efficiency improvements are anticipated
for heat pumps. The forecast assumes an improvement of 10% by the
year 2005 for the High Case and an improvement of 20% in the Low Case.
Wood Heating
The ESRG electric space heat and heating auxiliary modules
incorporate the effect of wood heat penetration on electricity use.
The model treats separately electric space heating customers,
deducting the electricity supplanted by wood usage, and customers
using other fuels via a reduction in usage for heating auxiliaries.
Each module uses two data points in time: the base year, and 1990
(with post-1990 being held at the level attained by 1990). In ESH,
retrofit and new are treated separately. Retrofit grows from base
year levels of usage to 1990 levels linearly. New ESH residences are
assigned a wood heat usage level which remains constant through the
forecast period.
- 81 -
E $ R G
The fractional savings attributable to wood heat for ESH
customers is the product of 1) saturation within electric space heat
and 2) the average percentage saved via wood heat usage.
The Company does not elicit information on saturations of wood
space heating in their appliance surveys (Ref. 1) and there is no
evidence of a strong potential in this area. The present forecast
assumes that there is negligible wood heat now and, in the High Case,
there will be no growth in wood heating in the future. On the basis
of the wood heat analysis given in the New York State Master Plan
(Ref. 38), the Low Case assumes a moderate penetration. By 1990, a
wood heat fraction of 2% is used for existing housing stock. New
homes are assumed to include no wood heat.
Miscellaneous Usage
There are a number of structural factors contributing to the
moderation of the historic: growth in miscellaneous appliance usage:
· decreasing population per household
· approach to market saturation
· increased efficiencies
· price-induced conservation
· substitutional effects (e.~., small kitchen
appliances precluding use of others)
Consequently, miscellaneous use per customer has not been
increasing rapidly in recent years. LILCO's analysis of residential
appliance usage (provided in Ref. 1) shows expected increases in small
appliances other than base use appliances. The present forecasts
assume an increase for the high range in miscellaneous use per
customer of 100%, phased in linearly over twenty years, while
the Low Case is held constant.
8.2.6 Appliance Lifetimes
Actual appliance lifetimes have been used rather than the
commonly employed United States Department of Agriculture figures for
average year of appliance possession by the first owner.
- 82 -
£ S R G
TABLE 8.8*
APPLIANCE LIFETIMES IN YEARS
Appliance
Lifetime
Refrigerator
Freezer
Range
Lighting
Television
Clothes Dryer
Clothes Washer
Dishwasher
Water Heater
Room A/C
Central A/C
Space Heat
Heating Auxiliary
Miscellaneous
2O
24.9
16.9
NA**
14.7
15.3
12.3
13.5
10
11
11
NA
NA
NA
* Source: Ref. 44, Appendix A-13.
** NA = Not applicable.
8.3 Commercial Sector Data Inputs
The following discussion of commercial forecast input assumptions
parallels the model description in Chapter 4. The base-year
commercial sales figure was 5236 GWH, as described in
Section 8.1.
The ESRG demand submodel for the commercial sector includes beth
small and large energy consuming commercial customers not involved in
the manufacturing or mining industries. As such, the non-
manufacturing industrial sales reported under other sectoral headings
by LILCO are included in ESRG's commercial sales figure.
8.3.1 1975 Floorspace
Estimates of the 1975 Long Island Lighting Company service
territory commercial floorspace are needed as input to the commercial
energy submodel. Sec. 4.2.1 provides a description of the basic
calculation method. The calculated values have been adjusted via
LILCO data on sales by building type. Estimates of the LILCO service
area 1975 commercial floorspace by category are given in Table 8.9.
- 83 -
E S R G
TABLE 8.9
1975 COMMERCIAL FLOORSPACEf LILCO SERVICE A~RA
1975 Floorspace
Commercial Category (106 ft.2)
F.I.R.E. 12.510
Federal Government 8.669
State/Local Government 39.597
Professional Services 5.122
Wholesale and Retail 62.154
Trucking and Warehousing 3.940
Other Transportation 6.009
Cc~ununications 1.842
Lodging and Personal Services 7.092
Business and Repair Services 13.002
Amusement and Recreation 13.987
Railroad .690
Health Services 8.077
Schools and Education 52.501
8.3.2 Floorspace Growth Indicesf High Case
Employment growth is used as the proxy for floorspace growth in
the first twelve commercial categories. Health and hospital
floorspace is projected to increase with population growth, while
school and educational floorspace is trended on the basis of
school-age population projections. In the High Case, the basis of
these-growth projections is the employment and customer projections
provided by LILCO in the 1983 New York Power Pool Report (Ref. 1, pg.
144), with 1975 backcast linearly. High Case population growth is
based on the projections for the Nassau-Suffolk SMSA by the New York
State Department of Commerce (Ref. 20), as are the school-age
population projections. High Case commercial growth ratios are
provided in Table 8.10.
TABLE 8.10
HIGH CASE COMMERCIAL GROWTH INDICES~ LONG ISLAND
Commercial Category
Finance, Insurance,
Federal Government
State/Local Government
Professional Services
Wholesale and Retail
Trucking and Warehousing'
Other Transportation
Communications
Lodging and Personal Services
Business and Repair Services
Amusement and Recreation
Railroad Transportation
Health/Hospitals
Education/Schools
1985 + 2000
1975 1975
and Real Estate
1.576 2.563
1. 063 1. 243
1.063 1.243
1. 291 1. 805
1.221 1.566
1.208 1.521
1.208 1.521
1.208 1.521
1.291 1.805
1.291 1.805
1.291 1.805
1.208 1.521
1.023 1.109
.765 .742
~; S - 84 -R G
8.3.3 Floorspace Growth Indices~ Low Cas~
Commercial floorspace growth indices for the Low Case are based
on OBERS employment and population projections for the Nassau-Suffolk
SMSA done by BEA (Ref. 7). Table 8.11 summarizes the growth factors
based on these state agency forecasts.
TABLE 8.11
LOW CASE COMMERCIAL GROWTH INDICES~ LONG ISLAND
Commercial Category
Finance, Insurance,
Federal Government
State/Local Government
Professional Services
Wholesale and Retail
Trucking and Warehousing
Other Transportation
Communications
Lodging and Personal Services
Business and Repair Services
Amusement and Recreation
Railroad Transport.
Health/Hospitals
Education/Schools
1985 + 2000
1975 1975
and Real Estate
1.493 1.592
1.177 1.253
1.097 1.050
1. 325 1. 451
1.182 1.171
1. 218 1. 208
1. 218 1. 208
1. 218 1. 208
1. 325 1. 451
1. 325 1. 451
1. 325 1. 451
1. 218 1. 208
1.066 1.058
.882 .921
The above assumptions for floorspace growth yield total esti-
mates of commercial floorspace in future years. However, the mix
of existing/new buildings changes over time also. The forecast
decreases existing stock at annual rates taken from regional
retirement rates given in Ref. 8. These retirement rates are:
Building TyDe
Retirement
Rate (%/Year)
Offices .647
Retail .647
Hospitals .648
Schools .757
Other .611
- 85 -
E S R G
8.3.4 Electric Intensities and Saturation
Electric intensity estimates are required to initialize the
commercial sector energy growth calculation. Given our 1975 commercial
estimates, we developed overall energy intensities (KWH/ft.2/year).
The intensities appropriate to building type and end-use were
developed by prorated Northeastern regional electric use
coefficients from Ref. 8 with weather-sensitive usage scaled, as
appropriate, by heating and cooling degree days. Weighted values for
degree days were developed using Long Island weather station data from
Ref. 42; the calculated value for heating degree days was 5415, and
for cooling, 740. The resulting coefficients are shown in Table
8.12 for both the existing 1975 stock and prototypical new floorspace.
Cc~unercial ESH saturations appear to follow residential
saturations, with some lag. On this basis 1975 ESH saturation was
estimated at 2.4%. Commercial floorspace air-conditioning saturation
was estimated to average approximately 80%, with hospitals set higher
and retail and schools set lower for 1975.
Future ESH saturations in new buildings were estimated in the
High Case to grow from 20% starting in 1975 to 40% by the year 2000;
in the Low Case comparable estimates are 10% and 20% respectively. In
terms of retrofit of existing floorspace, the High Case assumes that
ESH saturation reaches 12% by 2000; the Low Case assumes 8%.
Cc~unercial air conditioning is projected to reach 100% for all
new buildings by the year 2000 in the High Case. The Low Case assumes
air conditioning of new buildings will follow established saturation
levels in existing buildings, i.e., approximately 80% of new buildings
will be cooled. The High Case also assumes that all existing building
stock (except educational buildings) will be retrofitted to air
conditioning by the year 2000; the Low Case maintains present levels.
TABLE 8.12
COMMERCIAL ELECTRIC USE COEFFICIENTSt LILCO
(KWH/YEAR/Ft.&)
Existing New
Aux. Aux.
Building Heat- and Heat- and
Type lng Coolin~ Lightinq Power lng Cooling Lightin¢ Power
Office 9.01 5.94 7.00 5.30 12.77 4.13 7.00 4.40
Retail 4.06 6.72 ].8.20 6.40 6.34 4.52 18.20 5.90
Hospital 9.60 7.62 ].7.60 9.40 15.64 3.49 17.60 8.80
Schools 8.12 5.04 7.60 4.40 11.58 3.49 7.60 3.50
Other 4.65 6.72 10.00 6.40 6.93 2.58 10.00 5.90
- 86 -
E $ R
8.3.5 Future Commercial Intensities and Saturations
The methodology for incorporating future adjustments to
electrical intensities penetrates the conservation levels based on the
application of a payback analysis to S-shaped market-acceptance
curves. These are logistic curves which are defined in terms of 50%
acceptance levels; i.e., for a given payback period (appropriate for a
typical mix of owners of a given type of building) the conservation
option would be economically acceptable to 50% of the building owners.
If the payback period is shorter, the acceptance is proportionally
greater; if longer, the acceptance is less. The payback period is
analyzed for the marginal costs of moving to the next incremental
level of conservation technologies (see Section 4.3.2). Table 8.13
shows the 50% acceptance values used for the acceptance curves.
TABLE 8.13'
YEARS PAYBACK FOR 50% ACCEPTANCE
Building Type Office Retail Hospital School Other
Retrofit 3.7 2.6 3.5 4.0 2.6
New 3.7 2.8 4.0 4.0 2.8
* Source: Ref. 8.
The savings and costs are based on the electrical intensities and
reductions (discussed previously), the conservation costs (Ref. 6), and
the future price assumptions for electricity technology and fossil
fuels saved. These prices are shown in the following table, oil is
projected to experience zero real price growth in the Low Case.
Regional natural gas prices are assumed to grow at a real increase of
1% per year, while electricity is projected to experience a 2% annual
real growth. Further, the High Case conservatively assumes no
conservation technology penetration in the commercial sector over the
forecast period.
- 87 -
TABLE 8.14
FUTURE ENERGY PRICE ASSUMPTIONS (COMMERCIAL SECTOR)r LOW CASE
1985 2000
Fossil Fuel (1982 $/MMBTU) 7.19 7.52
Electricity (1982 ,C/KWH) 10.40 14.00
The derived penetrations used only in the Low Case are shown in
Table 8.15. Note that separate penetration matrices are developed for
the electric space heat end-use. The values in the table are
fractions of floorspace ,at these conservation levels. When the sum is
less than one, the remainder has no conservation above base year
levels.
TABLE 8.15
LOW CASE CONSERVATION LEVEL PENETRATIONSr LILCO
Electric S)ace Heat Other End-Uses
Existing New Existing New
Building
Year Type Level 1 2 3 1 2 3 1 2 3 1 2 3
Office .11 .41 .3~ .12 .79 .04 .11 .38 .36 .16 .77 .01
Retail 0 .84 0 .09 .66 .2~ 0 .85 0 .10 .69 .15
1985 Hospitals .20 .44 .2f .09 .67 .21 .21 .44 .26 .15 .81 .01
Schools .20 .01 .6S .19 .72 .06 .18 .62 .76 .27 .69 01
Other .25 .05 .34 .40 .55 02 .25 .14 .36 .58 .36 0
Office .07 .29 .55 .08 .73 .16 .07 .31 .53 .11 .80 .05
Retail 0 .89 0 .06 .50 .4G 0 .90 0 .07 .56 .33
2000 Hospitals .13 .33 .47 .06 .51 .42 .15 .33 .46 .10 .82 .05
Schools .13 .01 .8fl .12 .66 .2G .13 .01 .80 .19 .74 .04
Other .17 .12 .54 .26 .63 .08 .17 .12 .53 .43 .51 .02
- 88 -
Finally, an additional factor is introduced into the model to
capture the potential increase in energy use intensity (KWH/ft.2) of
commercial sites due to the market penetration of computers, copying
machines, word processors, etc. This factor increases the electrical
intensity of the auxiliary and power end-use category in the High Case
by 50%, phased in linearly from 1975 to the year 2000.
8.4 Industrial Data Inputs
The industrial sector model is described in Chapter 5. There are
three kinds of data requirements: base year experience, production
growth and electric intensity. The following sections describe
industrial forecast submodule inputs.
8.4.1 Base Year Experience
The figure for total industrial base year sales (1204.9 GWH) was
discussed in Section 8.1. This figure is taken to represent
normalized sales to the manufacturing and mining sectors. The mix of
industrial sales was taken from LILCO's EEI Uniform Statistical Report
(Ref. 17) for 1981.. Sales by Industrial SIC are expressed as a
fraction of total industrial sales in Table 8.16.
TABLE 8.16
LILCO INDUSTRIAL SALRS MIX
SIC
Fraction of Sales
20 1 .048
22 2 .018
23 3 .013
24 4 .006
25 5 .013
26 6 .039
27 7 .075
28 8 .048
29 9 .004
30 15 .047
31 16 .003
32 17 .009
33 10 .034
34 11 .061
35 12 .101
36 13 .212
37 14 .146
38 18 .098
21,39,Mining 19 .028
* j = ESRG Index (see Table 5.1).
4- Customers with demands below 30 KW.
- 89 -
8.4.2 Employment Growth
The model requires growth estimates of "State Employment Indices."
These growth assumptions are summarized in Table 8.17. They give
employment levels relative to the base year 1982 (base year = 1) for
two future years, 1985 and 1990.
The growth in industrial employment was selected in conjunction
with growth in commercial sector employment; the Low Case was based on
1980 OBERS employment projections by the Bureau of Economic Analysis
(Ref. 7). The High Case is based on New York State Division of the
Budget projections (Ref. 43), which run from 1982 to 1985,
extrapolated out to 1990.
TABLE 8.17
LILCO EMPLOYMENT GROWTH* BY STANDARD INDUSTRIAL CLASSIFICATION
EMPL 1985/EMPL 1982 EMPL 1990/EMPL 1982
SIC HIGH LOW HIGH LOW
20 1.002 1.002 1.006 .965
22 .950 .970 .862 .874
23 .997 1.013 .992 .969
24 1.026 .998 1.070 .959
25 1.026 1.008 1.070 .977
26 1.002 1.012 1.006 .960
27 1.022 1.029 1.058 ~.036
28 .981 1.012 .948 .958
29 1.044 1.064 1.117 1.133
30 1.044 1.070 1.117 1.138
31 1.044 .948 1.117 .848
32 1.015 1.008 1.039 .981
33 .976 1.011 .934 1.005
34 1.011 1.008 1.029 .948
35 1.001 1.059 1.002 1.113
36 1.028 1.041 1.075 1.032
37 .921 1.009 .791 .980
38 1.057 1.037 1.153 1.033
21~39~ Mininq .975** 1.033 .933** .970
* Employment growth is measured by State (SIC) Employment Index
(SEMP1) defined as SIC employment in a forecast year divided by
employment in the base year.
**Category grown at SIC 39 rates.
- 90 -
F. S R G
8.4.3 Electric Energy Intensity
Growth in sales to manufacturing and mining concerns may result
from an increase in the level of activity within the industry, but
also can result from changes in the energy use characteristics of that
industry. Electric energy intensities are designed to capture these
changing usage patterns.
Industrial electric intensity is defined as the consumption of
electricity per employee by two-digit manufacturing SIC. Changes in
electrical intensity are incorporated in the energy forecast.
The electrical intensities (electricity per employee) are
estimated by applying multiple regression analysis to the historical
data, using an exponential equation. The independent variables of
productivity, price and time are tested, with the best equation then
selected for the forecast. The general form of the equation is:
INTENSITY = C x PRODUCTIVITyA1 x PRICEA2 x eA3 x TIME
Equations are fitted separately for each SIC. Historic manufacturing
data are taken from Census of Manufactur~ and Survey of Manufactures
documents (Refs. 5 and 12). Table 8.18 shows the regressions selected
for each SIC. For SIC 29, where specific data for the State of New
York were inadequate, national values were analyzed and substituted.
In this study, productivity rather than time was found to be the
preferred explanatory independent variable for most industries. The
future value of productivity for each SIC is estimated, using a linear
relationship:
PRODUCTIVITY ~ A1 + A2 x TIME
Productivity is defined as product per employee, where product is the
sum of the value added and the cost of materials.
- 91 -
TABLE 8.18
ELECTRICAL INTENSITY REGRESSION RESULTS~ NEW YORK
Industrial Productivity Price Time
SIC Coefficient Coefficient* Coefficient R-Square
20 2.156 .809
22 2.161 - .63 .780
23 2.323 - .84 .891
24 - .26 .0765 .846
25 1.005 - .75 .632
26 2.319 - .48 .874
27 -1.09 .0763 .907
28 .750 - .20 .916
29+ 1.603 - .53 .914
30 2.375 - .26 .792
31 - .97 .0464 .954
32 3.048 .676
33 1.194 .796
34 3.060 .641
35 1.668 .884
36 2.644 - .26 .882
37 1.429 - .26 .880
38 .904 -1.04 .796
39 1.224 -.46 .824
* Since the function is exponential, the Price Coefficient is the
"price elasticity." The elasticity values utilized estimates
from national cross-sectional data for SIC's 22, 24, 26, 29,
30, 36, 37, while empirical N.Y. State values were used for
SIC's 23, 25, 27, 28, 31, 38, 39.
+ Regression based on national data; state data incomplete or
regression results not significant.
Two productivity trends are developed for each SIC; one based on
data going back to the 1960's, and another based only on data from the
1970's. Neither long nor short term trends consistently yield the
higher growth ratios. Subsequent decisions as to High and Low Case
industrial intensity growth were dependent on which trend was the
higher of the two. Table 8.19 contains the derived productivity
growth rates.
- 92 -
E S R G
TABLE 8.19
PRODUCTIVITY TREND RESULTS~
1982-1990
NEW YORK
Industrial Growth Rate (%/Year)
SIC Long Trend Short Trend
20 2.12 2.41
22 1.51
23 2.20 1.90
25 1.51 1.52
26 1.67 1.54
28 1.97 .03
29* 2.25 2.08
30 1.72 1.07
32 .87 .13
33 1.73 2.07
34 .72 -1.59
35 1.53
36 1.30 .26
37 2.24 1.78
38 2.58 2.36
39 1.97 1.45
* Based on national data.
Finally, an "energy intensity index" for 1990 (where 1982 = 1) is
obtained for each SIC using the desired regression coefficients and
appropriate values for the explanatory variables. Where price is
found to be a significant explanatory variable, annual real
1982-1990 growth rates for industrial electricity prices of 2%
in the Low Case and zero real growth in the High Case were used to
project energy intensity indices. High and Low Case intensities
were selected, on the basis of high and low outputs from the produc-
tivity trends' input into the intensity equations. Table 8.20 provides
the intensity grdwth factors used in the present forecast.
- 93 -
TABLE 8.20
1990 PROJECTED STATE ENERGY INTENSITY INDEXr NEW YORK
(1982 = 1.00)
SIC High Case Low Case
20 1.331 1.285
22 1.295 1.000
23 1.497 1.242
24 1.844 1.770
25 1.129 1.002
26 1.361 1.231
27 1.841 1.548
28 1.124 1.000
2!) 1.330 1.197
30 1.383 1.174'
31 1.449 1.243
32 1.235 1.032
313 1.216 1.178
34 1.191 1.000
35 1.224 1.000
38 1.315 1.013
37 1.288 1.174
38 1.202 1.004
39,21,
Mining %.127 1.042
8.4.4 Fraction of Electricity Self-Generated
Escalating costs of electricity and favorable governmental
actions have stimulated an increased interest in cogeneration
by many industrial and commercial decision-makers. According to the
1982 New York State Master Plan (Ref. 48), there are presently
a number of industrial/commercial sites on Long Island that generate
electricity. That same document indicated a number of additional
sites for potential cogeneration on Long Island contributing
electricity to both industrial and commercial sectors. In spite of
these potentials, the status of future cogeneration and its
contribution to electricity supply in the state remains unclear. The
present forecast conservatively assumes no new contribution from
self-generation during the forecast period, in either High or Low
Case inputs.
8.5 Other Energy Requirements
Sales to customers other than those in the three major service
sectors include: street lighting sales, sales for resale (Rockville
Centre and Freeport), public authorities, and LIRR sales. Separately
- 94 -
E S R G
the 1982 sales figures for these four subclasses were 182, 14, 62, and
180 GWH respectively, summing to 438 GWH (Refs. 1, 13). LILCO
forecasts each category separately in Ref. 1. The ESRG forecast
uses these forecasts, extrapolating linearly to the year 2000, rather
than preparing an independent forecast. (See Section 8.1 for a
discussion of normalization of base year energy.)
The other required input datum is line losses and the own-use of
Long Island Lighting Company. LILCO uses this category to include
sales for electric vehicles in their Ref. 1 forecast. Based on the
NYPP report (Ref. 1), the 1982 fraction was set at .095. Following
LILCO's assumptions on losses, own-use, and electrical vehicles, the
High Case increases this fraction to .103 by 1999. In the Low Case,
LILCO's assumptions on losses and own use are endorsed but electrical
vehicles are not assumed to make the High Case contribution, resulting
in a fraction of .096.
8.6 Peak Power Model
The peak power part of the model is driven by end-use energy
outputs from the other submodels with the exception of two kinds of
data= (1) the end-use profile of base year peak and (2) forecasts of
load management impact. These will be discussed in turn.
8.6.1 Peak Load Data
The 1982 s~mmer peak for LILCO was 30?0 MW (megawatts),
weather-normalized on July 19, 1982. The winter peak for 1982-83 was
2471 MW, occurring on January 19, 1983.
The system peak is the sum of the end-use contributions to
peak. There are two sets of load data require~ for the peak demand
simulation= the contributions of each end-use to su-~er peak and to
winter peak. The peak submodel utilizes a peak factor multiplier
(PKFAC) which is the reciprocal of the coincident load factor of that
end-use, and which derives the peak contribution of that end-use from
the annual energy de,and of that end-use. Peak factors used in su~er
and winter peak calculations differ. An analysis of residential and
commercial/industrial end-use contributions to peak was provided by
LILCO in Ref. 1. These data served as the bases for the development
of the peak factors in Table 8.21.
Peak factors for the industrial sector were initially devised
base~ upon the NEPOOL analysis of peak contribution by SIC (Refs. 32
and 35) and ASHR~ design temperatures for Long Island (Ref. 39). The
resulting peak factors are shown in Table 8.21.
- 95 -
E S R G
TABLE 8.21
PEAK MULTIPLIER FACTORS (PKFAC) - LILCO
Sector End-Use S-.~er PKFAC Winter PKFAC
Residential Refrigerators .870 .871
Freezers 1.102 .883
Ranges 8.113 9.203
Lighting 1.136 2.274
T.V. 1.586 3.174
Clothes Dryers 1,370 1.143
Clothes Washers 2.156 1.705
Dish Washers 2.453 2.455
Water Heaters .754 .830
Room Air Conditioners 10.543 0
.Central Air Conditioners 6.336 0
Space Heating 0 5.264
Heating Auxiliaries 0 1.491
Miscellaneous 2.920 1.697
C~mercial Heating 0 4.672
Cooling 4.563 0
Lighting 1.656 1.722
Auxiliary & Power .811 1.611
Industrial Ail 1.197 1.134
8.6.2 Load Management Impact
Additional data requirements for the peak submodel relate to the
impact of load management during the forecast period. This is
introduced by means of peak reduction factors which scale down the
end-use contributions to peak.
While LILCO provides discussions of rate policies and load
management strategies designed to reduce peak growth in their forecast
(Ref. 1), the peak analysis given in Tables 23 and 24 of the same
document show no load reduction in the residential sector and a 36 MW
reduction in the c~nmercial/industrial sector by the year 1999. The
present forecast incorporates only the reductions indicated by LILCO
for use in creating peak reduction factors of .022 in these two
sectors, while additional load management may be a desirable option
for this service area, no clear cut plan presently exists.
- 96 -
E $ R CS
REFERENCES
Report of Member Electric Syst~ of the New York Power Pool and
the ESEERCO Long Range Planf 1983 pursuant to Section 5-112 of the
Energy Law of New York State, April 1, 1983.
General Electric, C~_~ercial Sector Energy Consumption, Data Base
Development Pro~ect, for United States Department of Energy,
June, 1978.
Jackson, J. et al., Commercial Energy Use= A Disaggregation by
Fuel, ~Type _an~ End-Use, Oak Ridge National Laboratory,
O---~/CON-14, February 1978.
, The Cc~ercial Demand for Energy= ~ Disaggregated
Approach, ORNL/CON-15, April 1978.
United States Bureau of the Census, Census of Manufactures and
Annual Survey of Manufacturest "Statistics fo--r States,' 1963 to
1978.
Carhart, S. et al., Th__eBrookhaven Buildin~Energ¥ Conservation
Optimization Mo~el, Brookhaven National Laboratory, Formal
Report, January 1978.
United States Department of C~erce, Bureau of Economic Analysis.
OBERS Pro~ections of Economic Activity in the United States, Tape
#1= Total United States and 51 States, 1981, and Tape ~3 SMSA's.
Glesk, M., et al., Residential/Co~mercial Market for Energy
Technologies, prepared for Department of Energy by Arthur D.
Little, Inc., C-790470-03, August 1979.
9. United States Bureau of the Census, County Business Patternst
New York State, 1975.
10. Ide, Edward, et al., Estimating Land and Floor Area Implicit in
Employment Pro~ectionsf NTIS %PB200-069, for US DPt, July 1970.
11. United States Bureau of the Budget, SIC Classification Manual,
1967.
12.
United States Bureau of the Census, Survey o__f Manufactures and
Census of Manufacturest 'Fuels and Electric Energy Consumed' 1963
to 1980.
13. Long Island Lighting Cempany, Annual Report t__o Stockholders, 1982.
- 97 -
E S R G
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Data Resources, Inc., The Residential Demand for Energy, EPRI
Report EA 235, Vol. I, January 1977.
Long Island Lighting Company, Annual Report o__f LILCO to the State
o__f New York Public Service C~ission~ Form 182, for the year
ended December 31, 1982.
Christian, J., ~ Air-to-Air Heat Pumps, Argonne National
Laboratory, ANL/CES/TE 77-10, July 1977.
Long Island Lighting Company, Edison Electric Institute Uniform
Statistical Report - Year Ended December 31, 1982; April 1983.
United States Bureau of the Census, General Population Character-
istics~ New York, 1980 Census of Population, PC 80-1-B34, 1982.
Long Island Lighting Company, Population Survey 1982, Current
Estimates for Nassau and Suffolk County.
New York State Department of Co~erce, Preliminary Official
Population Pro~ections~ March, 1983.
David B. Goldstein, Efficient Refrigerators; Market Availability
an__d Potential Savings (Revised), National Resources Defense
Council, 1982.
Association of Home Appliance Manufacturers, Appliance Letter,
July-September 1981.
Hirst, E. et al., Fuel Choices in the Household Secto__r, Oak Ridge
National Laboratory, 0~~, october 1976.
FEA Efficiency Improvement Targets, Federal Register, vol. 43,
No. 70, April 11, 1978.
United States Department of Commerce, Bureau of the Census, 1980
Census of Housing~ General Housing Characteristics~ New York,
HCS0-1-A34, 1982.
Department of Energy, "Proposed Appliance Efficiency Rules", Federal
Register, Vol. 45, No. 127, June 30, 1980.
United States Department of Ccmmerce, Bureau of the Census,
County end City Data Book 1977, May, 1978.
Roskins, R.A. et al., Enerq¥ and Cost Analysis of Residential
Refrigerators~ oak Ridge National Laboratory, ORN/CON-6, January
1977.
Nl!mber of Employees of Class I Line-Haul Railroads, Association
~-~r~an Railroad~, Econ----~x"'cs and Finance Dept., unpublished
document.
Tansil, J., Residential Consumption of Electricity 1950-1970~ Oak
Ridge National Laboratory, ORNL-NSF-EP-51, July 1973.
- 98 -
£ S R G
31. Liepens, G. et al., Building Energy Use Data Bookt Oak Ridge
National Laboratory, ORNL 53-63, April 1978.
32.
33.
34.
35.
36.
37.
38.
39.
NEPOOL Load Forecasting Task Force and Battelle-Columbus
Laboratories, Report on Model for Long-Range Forecasting o__f
Electric Energy and Demand t__o the New England Power Pool. West
Springfield, Massachusetts: New England Power Planning June 30,
1977 (including subsequent NEPOOL model documentation).
40.
Stanford Research Institute, Business Development Strategy and
Market for a New General Purpose Lampt December 1976.
Air-Conditioning and Refrigeration Institute, Comparative Study
~f. Energy Efficiency Ratios~ January 1983.
New England Power Pool, Computer Printouts and Microfilm of
NEPOOL/Battelle Forecast Model Runs (Maine), May 3, 1980.
United States Department of Coa=~erce, Bureau of the Census,
Government Employment in 1975, August 1975.
Association of Home Appliance Manufacturers, Facts o__n Major Home
Appliance Energy Consumption and Efficiency Trends, 1983.
New York State Energy Office, New York State Energy Master Plan
and Long Range Electric and Gas Report, Final Report, March 1982.
American Society of Heating, Refrigerating, and Air-Conditioning
Engineers, Inc., ASHRAE Handbook and Product Directoryf 1977f
Fundamentals, 1977.
41.
U. S. Department of Commerce, Bureau of the Census, Housing
Authorized by Building Permits and Public Contractsf C40-13 Years
1970 through 1981.
42.
United States Civil Service Commission, Federal Civilian Work-
force Statistics~ Annual Employment by Geographic Area, SM68-10,
December 31, 1975.
43.
44.
United States Department of Commerce, mMonthly Normals of
Temperature, Precipitation, and Heating and Cooling Degree Days,
1941-1970,m August 1973.
New York State Division of the Budget, Private C~m~mication,
March, 1983, 1982-1985 Employment Values for New York State.
California Energy Resources Conservation and Development
Commission, Analysis of Residential Ener~ Usest 1977.
- 99 -
£ S R G
45.
46.
47.
48.
49.
50.
Steven Buchsbaum and James W. Benson, Jobs and Energy: The
Employment Impacts of Nuclear Power~ Conservation~ and Other
Energy Options. New York: Council on Economic Priorities,
1979.
New York State Electric Utilities, New York State Public
Service Commission, and New York State Energy Office, New
York State Residential Insulation Survey: Final Report.
Albany: New York Department of State, September 16, 1977.
Paul Raskin, et al., The Conservation Alternative to the
Power Plant at Shoreham~ Long Island~ ESRG 80-31, Energy
Systems Research Group, November, 1980.
New York State Energy Office, New York State Energy Master
Plan, March, 1980.
United States Bureau of the Census, Provisional Estimates of
Social~ Economic~ and Housing Characteristics~ Supplementary
Report PHC 80-51-1, March, 1982.
New York Gas Group, 1983 New York Gas Report (SEMP III),
submitted to the New York State Energy Office, April 1, 1983.
- 100 -
£ S R O
APPENDIX A
ESRG HIGH AND LOW CASE FORECASTS
OF
LONG ISLAND LIGHTING COMPANY
~ S R G
ESRG HIGH CASE AGGREGATE FORECAST
LILCO
HISH CASE
LILC883 ENEROY IN O~H
RESIDENT. COHHER. INDUSTR,
1982 5574. 5340. 1205o
1983 5660. 5450. 1250,
1984 $750. 5560. 1300.
1985 5830, 5660, 1350.
1986 5910, 5800, 1400,
1987 5980, 5950. 1450.
1988 6050. 6090. 1500.
1989 6120, 6230. 1550,
1990 6190, 6380, 1590,
1991 6250, 6520, 1640.
1992 6300. 6670, 1690,
1995 6360. 6820* 1740.
1994 6410, 6970* 1790,
1995 6460. 7120. ~840.
1996 6510. 7270. 1890,
1997 6570. 7420. 1950.
1998 6630. 7570. 2000.
1999 6690. 7730. 2050.
2000 6760. 7880. 2100,
PEAK PO~ER LOAD IN HU
0THER TOTAL 8UNHER elNTER
1637. 13757. 3070. 2471.
1670. 14030* 3130. 2520.
1700. 14310. 3180, 2580.
1730, 14580. 3240. 2630.'
17~0, 14870. 3300. 2690.
1790. 15170. 3350. 2750.
1830. 15460. 3410. 2810.
1860. 15760. 3470. 2870.
1890.' 16050. 3520. 2940.
1930. 16340. 3580. 3000.
1960. 16630, 3630. 3060.
1990. 16910. 3690. 3120.
2030. 17200. 3740, 3180.
'2070. 17480. 3800. 3240.
2100. 17780. 3860. 3300.
2140. 18070. 3910. 3360.
2180. 18370. 3970. 3420.
2220, 18680. 4030* 3490.
2260, 19000, 4090, 3550.
A-1
ESRG LOW CASE AGGREGATE FORECAST
LILCO
LOW CASE
LILCOB3
RESIDENT.
1982 5574.
:95J 5620.
1984 5660.
1985 5680.
1986 56?0.
1987 57i0.
I738 5720.
I989 5720.
1970 5730.
:991 5730.
1992 5730,
1993 5720.
1994 5710.
1995 5700.
1996 5700,
1997 5690.
1998 5700.
1999 5700,
2000 5710,
ENERGY XN GWH PEAK POWER LOAD IN MW
COMHER. INDUGTR. OTHER TOTAL SUHHER WINTER
5340, I205. 16~7. :3757. 3070. 2471.
5280. 1230. ~640, 15770. 3070. 2480.
5210. 1260. 1640, 13770, 3060. 2490°
5150. 1290. 1640. 13760. 3040. 2500,
5150. 1310, 1640. 13790, 3030. 2520,
5150. 1320, 1650, 13820. 3030. 2530.
5150, 1340. 1650, 13850. 3020, 2540.
5150. 1350. 1650. 13870. 3020. 2550.
5150. 1370. 1660. 13900, 3020. 2570.
5150. 1380. 1660. 13910. 3020. 2580.
5140o 1390, 1660, 139~0o 3020, 2590,
5140. 1410o 1660, 13960, 3020. 2600,
5140. 1420. 1670. 13950. 3020. 2610.
5150, 1440, 1670. 13960. 3020* 2610,
5150. 1450, 1680. 13970. 3020, 2620.
5150. 1460. 1680, 13990, 3020, 2630,
5150. 1480, 1690. 14010. 3020, 2640.
5150, 1490. 1690. 14030. 3020, 2650,
5150, 1510. 1700. 14060° 3030. 2660.
A-2
E S R G
ESRG HIGH CASE DISAGGREGATED FOP. ECAST
LILCO
LILC083
HIGH CASE - RESIDENTIAL SE'CTOR - FNFRGY IH GWH
1982 1985 1988 1991 1994 1997
1: REFRIGERATORS 1339. 1,3,50. 1336. 1299,
3: RANGES 2~8, 277. 285, 293,
4: I..I BHT ING 822, 842, 862. 881,
5: TELEVISIONS 366. 370,
6: CLOTHES LRYERS 424. 449. 47.~.
7: CLOTHES WASHERS 62, 64, 66. 68
8: B%SH WASHERS 137, 140, 140, 141
9: WATER H~ATERS 290, 307, J~15, 324
10: ROOM A/C 342, =~, ,
ll: CENTRAL A/C 29] , 330,
12: SPACE HEA] ..RS 254, .~0;,
13: HEATINGAUXILIARY 359. 356. ~52, ~48.
14: M I SCEL. I..ANEOUS 318. 375, 4~4~
1240,
314
300
898
413
519
69
332,
367,
429,
486,
557,
2(t00
1180, 1156,
308, :~00,
:~07, 314,
915, 9~2.
4~4, 457,
542, 566,
71. 72.
143o 144,
341. 350,
462, 494,
535, 579,
686,
LII. C083 HIGH CASE - COHHERC1AI.
1982 1985
1: OFFICES
1: HEATING
2: COOl. lNG
3: LIGHIlNG
4: A/JX ~ POWER
21 RETAIL
1: HEATING
2: COOt. lNG
3: LIGHTING
4: AO× ~ POWER
3: HOGF'ITALS
1: HEATING
2t C00t. ING
3: LIGHTING
4: AtIX & POWER
4: SCHOOLS
1: HEATING
2 COOL. lNG
3 LIGHTING
4 AUX & POWER
5 OTHER
1 HEATING
2: COOt. LNG
3: LIGHTING
4: AOX ~ POWER
57, 73,
337, 358,
478, 501,
402, 439,
28, 36.
~33, 357,
1204, 1273,
476, 527,
51, 51,
133, ]34,
81, 85,
17, 19,
81, 78,
307, 28],
203, 195,
26, 33,
246, 262,
510, 544,
366, 410,
S~CTOR - ENERGY IH BWH
1988 1991 1994 1997 2000
93. 114. 136, 158. 182,
383, 408, 433, 457. 481.
532, 563, 594, 625, 657,
484, 532, 582, 634. 689,
45, 55, 64, 74, 85,
1346, 1419, 1493, ].567, ]640,
583. ~41, 703. 767. 834.
8. 10, 11, 13, ]5.
52, 53, 54, 54, 55~
~36, 139, ~41, 144,
91, 97, 10~, [09, 115.
23, 28, 32, 3&, 41,
80, 82, 84, 87,
280, 278, 277, 276, 274,
205, 211. 219. 227, 234.
42, 51, 61, 71, 82,
279, 296, ~12, 329, 545,
584, 625, 665, 705, 745,
459, 51~, 568, 626, 687,
LILC083
20:
22:
23;
24:
25:
26:
27:
28:
29:
33
34
35
36:
37:
30:
31:
38:
39:
HIGH CASE - INBLISTRIAL
1982 1985
FOOD 58, 65,
TEXTII.ES 21. 2:5,
APPAREL. 16, 19,
I. UHBER 7,
FURNITIJRE 15. 16,
PAPER PRODUCTS 47, 54,
PRINTING ~ PUBL, 90, 121,
CHEN£CALS 58, 60,
PETROL£UH ~ COAL 5,
PRIMARY METALS 40, 4~,
FABRICAT. HETAL. S 73, 79,
MACHINERY 1~2, 133,
ELECTRIC EQUIP. 255, 293,
TRANSPORTATION 176, 179.
RUBBER ~ F't. ASTIC 5&. 67,
LEATHER 3, 4,
STONE,CLAY~GLASS 11, 12.
INSTRUMENTS 118. 135.
OTHER 33. 34,
SECTOR - ~N~RSY IN GWH
1988 1991 1994 1997 2000
73, 80, 87, 95, J02,
25, 24, 24, 24, 24,
22, 25, 28. 30, 33,
12. 14. 17. 20* 23.
I7, 19, 20, 21. 23.
60. 67, 73. 80, 87.
153, 186. 221, 257, 294,
61. 62. 6:5. 65, 66.
7, 8. 9. 10. 1].
45, 46, 48, 50, 51.
85, ~1, 98, 104, iii,
143, 153, 164, 174, 184,
~33, 375. 4~8. 46F. 510.
180, 178, 17~, 165, 154,
79, 91, 104, 118, 132,
5, ~* 6, 7, 8,
1~, 14, 16, 17, ~8,
152, 170, 190, 210, 232,
35, 35, ~6, 36, 36,
A-3
ESRG LOW CASE DISAGGREGATED FORECAST
LILCO
LILC083
LOW CASE - RESIDENTIAL SECTOR - ENERGY 1N GWH
1982 1985 1988 1991
1339, 1346. 1329. 1291.
268. 275, 281. 286.
822. 840. 854. 863.
366. 367. 370, 374.
424. 433° 441. 449,
62. 63, 65. 66.
137. 138. 137, 137.
290. 299. 298. 298.
291. 306, 311. 32].
359. 353. 348, 343.
318. 326. 334. 341.
1994
1232.
287.
29].
865.
378
456
68
137
298
295
333
385
348.
1 REFRIGERATORS
2 FREEZERS
RANGES
4 L. IOHFING
1ELEVISIONS
CLOTHES DRYERS
7: CLOTHES WASHERS
DISH WASHERS
WATER HEATERS
10: ROOM A/C
11: CENTRAL A/C
121 SPACE HEATERS
HEATINGAUXILIARY
14: MISCELLANEOUS
1997
]171
274
863
384
464
69
138
301
289
546.
410.
337.
354.
LIL.C083 LOW
1: OFFICES
HEATING
COOk. lNG
LIGHTING
AUX & POWER
RETAIL
HEATING
21 COOk. lNG
LIGHTING
4: AUX ~ POWER
HOSPITALS
HEATING
21 COOl. lNG
LIGHTING
AUX ~ POWER
4: SCHOOI_S
HEATING
2: COOLING
LIGHTING
AUX ~ POWER
5: OTHER
1: HEATING
2: COOLING
LIGHTING
AOX ~ POWER
CASE - COMMERCIAl.
1982 1985
SECTOR - ENERGY IN GWH
1988 i991 1994 1997
2000
ii45,
260.
856.
389.
47i.
70,
139.
286.
~60.
434.
LILC083
20: FOOD
22: TE×TII. ES
23t APPAREL
24: LUMBER
25: FURNITIJRE
26; PAPER PRODUCTS
27: PRINTING R F'UBL.
28: CHEMICALS
PETROLEUM ~ COAL.
PRIMARY MEI'ALS
FABRICAT, METALS
MACH£NERY
ELECTRIC EQUIP.
37: TRANSPOR[ATiON
RUBBER ~ PLASTIC
LEATHER
STON~,CLAY,GLASS
38: INSTRUMENTS
OTHER
2000
41. 48. 52. 56. 60. 63. 67,
337. 324. 320, 3[7, 313. 309. 305,
455. 420. 420. 420. 420, 420. 420.
372, 358, 357, 355. 353. ~52. 350~
17. 20. 21. 23. 24. 26, 27.
328. 317. 314. 310. :~07. 304. 300.
1270. 1248, 1242, 1237, 1233, 1226. ]22i.
414. 395. 393. 392, 390. 389. 388.
4. 5. 6. 6. 7. 7. 8.
54. 51. 50. 49. 48. 47. 47.
159. 156o 156. 155, 154. 153. 153.
78. 7~. 73. 73, 72. 72. 72.
15. 16. 18. 21. 24. 26. 2~.
70. 55, 54. 54. 54. 5~. 53.
326. 273. 275, 278. 280. 283. 285.
18. 21. 23. 24. 26. 28. 29.
252. 247, 244, 241, 238, 2~6. 233,
596. 618. 62~. 629, 634. 640. 645.
;~6~. 369. 371. 373. 373. 377. 380.
LOW CASE INDUSTRIAl.
1982 1985
58, 64,
21 21.
16 18,
7 9,
15 ]5.
47 52.
90 111,
58 59.
40. 44,
73. 73.
122. 129.
255,
176. 189.
56. 64.
3. 3.
11.
118. 123.
33. 35.
SECTOR - ENERGY IN GWH
1988 1.991 1994 1997
69. 73. 77.
19. 18. 17. 16.
19, 20. 20, 21,
10. 12. 13. 15
15. 15. 14. 14
54. 56. 58. &O
131. 151. 171. 191
57, 55, 53. 51
6, 7. 8. 8
46. 49. 5l. 54
71, 68. 66. 63
13~, 137. 141, 145
267, 267, 266, 266
197, 205* 212. 219
71. 77. 84. 92
3, 3, 3.
11. I1. 11. 11
123, 123, 123, 12~
34. 33, 33. 32
A-4
E S R O
2000
~5
15
22
16
14
61
49
9
56.
60 *
149.
266.
225 ·
100 ·
3.
122.
31.