ENERGY & ENVIRONMENi;;~, BC O'Regan, BS Wagner, JB Dickinson

Tags: house doctor, Walnut Creek, consumption, infiltration, natural gas, Treatment Group, houses, consumption data, temperature, natural gas consumption, energy savings, additional savings, water heater thermostat, water heater, electric baseboard heaters, Water heater blanket, infiltration losses, control group, Percent Change, person-hours, initial infiltration, house doctors
Content: LBL-15083 UC-95d c.::;)-
Lawrence Berkeley Laboratory
UNIVERSITY OF CALIFORNIA
ENERGY & ENVIRONMENi;;~ ;"'~~s3
DIVISION
L\BRAf-RESULTS OF THE WALNUT CREEK HOUSE DOCTOR PROJECT
B.C. O'Regan, B.S. Wagner, and J.B. Dickinson
November 1982
TWO-,WEEK LOAN COPY Thi~ is a Library Circulating Copy which may be borrowed for two weeks. For a personal retention copy~ call Tech. Info. Division~ Ext. 67B2.
~ J
)
ENERGY AND ENVIRONMENT DIVISION
Prepared for the U.S. Department of Energy under Contract DE-AC03-7?SF00098
DISCLAIMER This document was prepared as an account of work sponsored by the United States government. While this document is believed to contain correct information, neither the United States Government nor any agency thereof, nor the Regents of the University of California, nor any of their employees, makes any warranty, express or implied, or assumes any legal responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by its trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof, or the Regents of the University of California. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof or the Regents of the University of California.
LBL-15083 EEB-BED-82-15 RESULTS OF THE WAL~T CREEK HOUSE DOCTOR PROJECT B.C. O'Regan, B.S~ Wagner, J.B. Dickinson Buildings Energy Data Group Lawrence Berkeley Laboratory University of California · · Berkeley, CA 94720 . November 1982 ABSTRACT In this report .we present the results of a joint Lawrence Berkeley Laboratory and Pacific Gas and Electric Co. experiment designed to measure the additional energy savings achieved by adding two person-days ..·. .8-( .house doctoring to a standard energy audit. We compare a house doc7 ~-to'r and audit treatment to an audit alone and . to a passive control :·.c·, group. The results of a fourth treatment, house doctoring, audit and contractor retrofits, have not y.et been analyzed. The treatments were ~'pplied to randomly selected groups of 10 houses each in Walnut Creek, California. The difference in energy savings between the treatments, based on monthly utility bills, were not statistically significant due to wide variation in savings and the loss of several houses from each group. Predicted energy savings, based in part on measured air leakage area reductions, indicated that the retrofit package completed during house doctoring had a Cost of Conserved Energy of 42ў/therm. This work was supported by the Assistant Secretary for Conservation and Renewable Energy, Office of Building Energy Research and Development, Building Systems Division of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098, and by the Pacific Gas and Electric.Com~ pany. -1-
INTRODUCTION
Recent advances in the study of residential energy use have identified important paths for heating and cooling losses that previously eluded diagnosis and repair. Hany such paths, once' identified, can be repaired by an appropriate low-cost measure. A complete energy conservation retrofit, including repair of these recently discovered heat loss paths, can reduce the energy used to heat a house by 50% or more. 1 Unfortunately, skilled contractors able to carry out these new measures are not, as yet, readily available to homeowners.
"House Doctoring," a procedure designed·to commercialize these new retrofit techniques, is now being developed at several .institutions and small businesses. House doctoring focuses on using new instrumentation to find and fix problems otherwise difficult for a homeowner to diagnose and, while in the house, installing as many other low-cost measures as is possible. Two of the important diagnostic tools used by a house doctor are an infrared scanner and a "blower door," (a device which can slightly pressurize and depressurize the house). With these instruments infiltration rates can be quantitatively measured and otherwise hard-to-find leaks quickly identified and repaired. In addition, the effectiveness of an attempt to reduce leakage can be· checked immediately. Other measures generally undertaken by house doctors include insulating water heaters, installing low-flow showerheads, and replacing furnace fi1ters. Accomplishing these measures in one visit is potentially much more cost effective than visits by separate contractors for each kind of measure.
Although data on savings achieved in research houses are available,
reliable data on savings generated by "real world" house doctor programs
on occupied houses are scarce. To help rectify this scarcity LBL, in
co-operation with Pacific .Gas and Electric Company (PG&E), tested a
pilot house doctor program modeled after the Princeton Modular Retrofit
ii
Experiment (MRE). 2 PG&E supplied the billing records, provided and paid
the energy auditors. used as .house doctors, paid ':for the retrofit materi-
als, and helped defray LBL's costs. LBL provided data analysis before
and after the experiment, trained the house doctors, selected the
-2-
retrofit materials, provided instrumentation, and organized the materi-
als, house doctors,, and transportation during the retrofit period. This
report will examine the results of the program, specifically the energy
savings as determined from utility bills and engineering calculations,
,.
and make recommendations for improvements in house doctoring.
DESIGN' OF EXPERIMENT We designed our experiment to measure the incremental costs and energy savings resulting from adding house doctoring to a conventional energy audit. The experiment,was also intended to provide more information about the sources of infiltration in California housing and the best methods for reducing leakage rates in these houses. To these ends we planned to compare the energy savings resulting from four treatments: Full Retrofit (Group A): This group received a PG&E Home Energy Survey, approximately two person days of house doctoring, and conventional contractor re.trofits. * House Doc~ori:ng (Group B): This group received a PG&E Home Energy Survey and about 2 person days of house doctoring. Audit Only (Group C): This group received a PG&E Home Energy Survey alone. Blind Control (Group D): This group received no treatment. * A ''Home Energy Survey" was the pre-RCS computerized audit used by PG&E at the time of the experiment. For a complete list of measures involved in house doctoring see Table 1. The contractor retrofits included attic insulation, double pane windows, outside combustion air intakes for furnaces, fluorescent light fixtures, and duct taping and insulation. -3-
As the study site, we chose a subdivision in Walnut Creek CA, a · subUrban Community 30 miles east of San Francisco. Walnut Creek has an average of 2900 heating degree-days per year (base 65°F), close to the average for PG&E's service territory (2750 HDD). We then.selected 615 houses on three contiguous meter-reading routes as potential participants in the experiment. The analysis used to calculate .savings (see Analysis section) requires a good correlation between weather and energy use. Accordingly, for each house we calculated the least-squares fit between monthly gas use from utility bills and local degree-day data and elim- .' inated the 400 houses for which this fit had a correlation coefficient (r2) of less than 0.90. This first screening a.ccomplished a partial elimination of factors (such as heated swimming pools, electric heat, long unoccupied periods, etc.) ·that distort the gas-use to weather correlation. After the r 2 elimination, PG&E subcontracted for a phone survey of the remaining houses to determine the homeowners willingness to participate in an energy conservation experiment. Homeo~ers interested in participating were randomly assigned to a treatment group. Because we did not want participants to be aware of the other experimental groups, we described the experiment to each homeowner as involving only the treatment for which they had been selected. The homeowners were then asked a series of questions regarding occupancy changes, heating fuels, pools, etc. (The complete questionnaire is reproduced as Appendix A.) The houses, (now fewer than 54 in each group due to disinterested homeowners and those we could not contact) were scored 0 to 9 on the basis of their answers to the above questions. Within each group, the ten hou'ses with the highest scores were selected to participate. This second screening reduced the 215 houses to the 40 which took part in the experiment. Because they were to be passive controls we did not describe the experiment or conduct the occupancy and energy use portion of the telephone survey \'Then selecting the group D houses. The final sample of group D houses is thus from a -slightly different population, with -4-
respect to suitability for the analysis· of savings, than the other groups. An analysis of varianc~, however, shows all groups to be representative of the 215 potential participants with respect to initial energy use. Six energy auditors from local PG&E offices .were trained by LBL at a two week seminar covering use of the blower door, scanner, and retrofit techniques used in house doctoring. Each trainee worked in at · least three different training houses before working on the experimental houses. The house doctoring and conventional audits of the experimental houses took place in November and early December 1980. Two house-doctor teams of two persons· operated each day during this. period using a van containing the necessary tools and materials. A third team was available to purchase needed supplies as those in the vans were deplet~d or as the house doctors encountered new situations. The house doctors began each visit by measuring and photographing the house and doing an initial pressurization test. They then turned to reducing infiltration in the attic, crawlspace, and interior which usually required most of the day. After infiltration reduction they installe~ low flow showerheads, . faucet aerators. and new furnace filters as needed. Table 1 presents a more complete list of the measures. * The visit concluded with a final pressurization test and the audit recommendations. Insulating water heaters and installing intermittent ignition devices (IIDs), both possible house doctor tasks, were,· in this study, performed by other PG&E employees for legal and safety reasons respectively. The house doctors spent an average of approximately six hours in each house. Long commutes for some of the house doctors reduced the average time spent on the houses below the 8 hours originally planned. However, the time expended by the workers who installed the. IIDs and water heater blankets brought the total time expended by all workers to approximately 2 person-days per house. * For a complete description of the materials and procedures used by a house doctor see The House Doctor Himual3. -5-
Table 1. Measures Typically Installed During a House Doctor Visit. Items with asterisks are measures .emphasized during the Walnut Creek project.+ Hot Water System * 1. Install low-flow showerhead(s). * 2. Install faucet aerator(s). 3. Insulate water heater. * 4. Turn water heater thermostat down to 120 oF, or 140 °F if house has an automatic dishwasher. 5. Insulate first 5 feet of hot water pipe from water heater.
Furnace System
**
6. Replace air filter, if necessary. 7. Test and adjust to maximum steady-state
efficiency.
8. Set fan "off" control to 90 oF. * 9. Seal leaky ducts in attics and basements.
10. Install clock thermostat.
Heat and Air Leakage
* 11. Seal over dropped ceilings. * 12. Seal around pipes, electrical wires, and exhaust
vents in attics. * 13. Stuff openings around furnace flue and chimney
with fiberglass or caulk. * 14. Weatherstrip attic and basement/crawlspace hatch
or door.
,
* 15. Insulate attic hatch or door with R-19 fiberglass.
* 16. Seal around pipes, wires, and chimneys in unheated
basement/crawlspace.
* 17. Install a fireplace plug if no damper is present. 18. Install foam gaskets behind leaky switchplates
and outlets. * 19. Caulk around windows, doors, and window A/Cs.
20. Caulk baseboards and around electric baseboard
heaters. * 21. Seal holes behind sink and bathroom fixtures.
22. Insulate band joist
23. Caulk mudsill (only in heated basements).
Appliances 24. Install retrofit fluorescent lamps in much-used areas. 25. Turn on "power miser" switch or turn off "humidity" switch on refrigerator. 26. Give homeowner sample of cold-water detergent. +For a ~ore complete description of each measure see the House Doctor Manual. -6-
ANALYSIS
One goal of our ·analysis was to calculate the energy savings,
defined as the change in average yearly energy consumption, resulting

from the different treatments. To this end we calculated the Normaliz_ed
Annual Consumption (NAC) for each house. The NAC is the energy consump~
...
tion predicted for a year with average weather based_on consumption data
from any particular year. The prediction is based on the correlation
between fuel consumption and outside temperature~ We used natural gas
consumption from utility bills, and daily temperatures recorded at Saint
Marys College, located seven miles from the study area. ·The Saint Marys
data were adjusted to better represent the weather in Walnut Creek by
using a correlation based on 18 months of data collected at both loca-
tions in 1973-74. In order to compare pe~iods of different lengths, we
converted both the gas and temperature data to daily averages for each
billing period. The data points, 10 or 11 months in our case, were fit
to-a linear model:
where: Gj qverage daily gas consumption over period j (therms/day) R = the reference temperature (base temperature) used to calculate the degree days. (DDR) j =. average daily degree-.days per day over period j ·. (calculated using reference temperature R) The term c( (therms/day) is an estimate of the component of gas use not 'directly influenced by the weather, i.e. the gas used for cooking, waterheating, drying clothes, etc. The other term, ~(DDR)j, represents the space heating component of gas usage. .~ (therms/°F-day.) is thus an estimate of the total thermal conductivity (including infiltration) of the house divided by the efficiency of the heacing system · -7-
For each house we ran least-squares regressions
of
G. J
and
(DDR) j
with R taking on integral values from 35 to 75°F. The regression with
the best least-squares squares fit was used to establish the "best"c{, ~
and R. With these parameters we then calculated the Normalized Annual
Consumption as:
NAC = 365 [c{ + ~ (DDR)yr ] where (DDR)yr is the average daily degree days of an average year,. with respect to reference temperature R. We calculated the average year using temperature data from 1949 to 1979.
We calculated an NAC for each house before and after treatment. We expressed the effect of treatment as the average change in NAC of the treatment group. We also calculated NACs and changes therein for the average Walnut Creek household using PG&E's data on average natural gas consumed by all individually metered residences in Walnut Creek. The uncertainty for the average Walnut Creek NAC was calculated according to a method proposed by Fels and Goldberg (Princeton). 4
No electricity consumption model we tested enabled us to calculate an electric NAC with sufficient accuracy to compare the total electricity use of a house in different years. Because the main difficulty was in modeling energy consumed by air conditioners (which in this climate are often controlled manually rather than by a thermostat setting) we used the electricity_consumed for uses other than cooling, estimated as follows: We found the. average daily consumption for the months December to May, the months where no cooling occurred and for which we had consumption data for both the pre- and post-retrofit period. Assuming the average daily non-cooling use did not change for other months we multiplied the December-to-May figure by 365 to find the yearly non-cooling electricity use. We used the changes in this value between periods and groups to examine the effect of the treatment on electricity consumption and to check for fuel switching.
-8-
In addition to the NAC analysis, we calculated monthly infiltration rates for each house using a model developed by Sherman and Grimsrud (LBL). 5 This model calculates the leakage area (a parameter approximating the total cross-sectional area of air leaks) from the results of the pressurization test and calculates monthly infiltration rate from the leakage area and local monthly average windspeeds and temperatures. We ... then averaged the infiltration rates from November to March to find the heating-season infiltration. Using the change in heating-season infiltration for a particular house, and the average degree-days per year for the best-fit reference temperature for that house, we calculated the savings in gas consumption that might be expected from the infiltration reduction portion of the house doctor visit. This value, however, will not show the added savings that can be obtained when infiltration reduction also reduces convection losses, such a~ in the attic where the effect of . ~ov" ~r.:.:i,: .' .p..g·' dropped ceilings or an open wall cavity is to break·a convective'loop ·· · carrying heat from an uninsulated surface to the attic space. RESULTS AND DISCUSSION Of the 40 houses originally selected and treated, we eliminated 16 from the savings analysis. Nine houses with heated poois, spas or hot tubs were included in the original sample because of a poorly worded question in the phone survey (see Appendix A). We eliminated these houses because the pattern of natural gas use by these items renders calculation of an NAC unreliable. We also dropped houses due to change of owners or obvious meter reading errors. From the infiltration analysis we eliminated only three houses, each because of problems with the blower door data. Table 2 presents age, size, and occupancy data for the remaining houses in groups A, B, and C. Most were one-story ranch-style tract houses. A few were two-story tract houses. All houses had very low attics (4, or less at the ridge) with internal cross-bracing and 18" crawl spaces. All houses had gas-fired forced-air furnaces, gas water -9-
Table 2. General Characteristics of the Walnut Creek Experimental ·Houses by Treatment Croup.
Treatment Croup (A) House Doctor
Avg. Si~l (ft ) 1950±550
Avg. Vintage 1962±6
. Avg. II Residents 4.0±1.3
Total II in Final Sam lea 6
1-story 5
2-:story t' 1
(B) House Doctor
2490±520 1966±4 3.5±1.4
7

3
(C) Audit Only
2500±210 1968±5 4.5±0.8
6
5
1
aAll averages are based on the final sample of houses used to find the average energy savings, not the initial.sample of ten.
-10-
heaters and central.or room air conditioners. Gas was the more common fuel for both stoves and clothes dryers. These 24 houses, judged acceptable for the experiment by all criteria, had an average r 2 for the gas and weather correlation of 0.98. This high r 2, resulting from the use of the phone survey and post visit eliminations, implies that .90 was too low a cut-off value to use in the original r 2 elimination (see Design secti~n). Using a cut-off value of about .96 would have reduced the number of houses called during the phone survey, saving time and expense expended on that process, without eliminating any of the houses we picked for the final sample using the original method. Table 3 presents the average pre-retrofit NAC, post-retrofit NAC, and savings for each treatment group. The savings presented for group A show only the effects of the audit and house doctoring. because the contractor measures were installed at the very end of the post retrofit data collection period used for this paper. In other words, with respect to the results presented here, houses in groups A and B received essentially identical treatments and will be referred to collectively as the house doctored groups. As a further comparison we have included the "pre-ret;rofit" and "post retrofit" NACs for the average of all single family residences in Walnut Creek, calculated using data from the same months used for the test groups. The average natural gas savings (average percent change in NAC) for the two groups receiving house doctor visits are 11.3%±7.3% for group A and 11.5%±9.8% for group B. (The uncertainties given are 95% confidence intervals.) The savings for the audit-only group are 9.4%±9.8% and those for the blind controls 6.8%±20%. The NAC for the average of Walnut Creek single family residences showed a reduction of 7.0%±12. The average initial consumption for this group is smaller than that for ·the experimental groups due to inclusion of condominiums and some individually metered apartments. -11-
Table 3. Natural Gas and Electricity Savings by Treatment Group.
Natural Gas: Normalized Annual Consumption(NAC)a
Group
Pre Visit avg. NACa (therms/yr)b
(Post Visit ::::m~A~:) b
Change .in avg. NAC (therms/yr)c
Avg. Percent Change in NAC (%)c
Number of Houses
(A) House Doctord 1220±375
1075±313
-145±126
-11.3±7.3
6
(B) House Doctor
1335±259
1177±241
-158±127
-11.5±9.8
7
(C) Audit Only
1346±391
1204±306
-142±156
-9.4±9.8
6
(D) Control
1377±535
1262±434
-115±213
-6.8±20.
5
WC Resid. Avg.f
878±97
-816±77
-61±113
-7.0±12.
Group (A) House Doctor
Electric: Non-Cooling Consumptione
Pre Visit avg. ele),· fmwh/yr)
Post Visit avg. ele'b. (mwh/yr)
Change in (mavwgh.. /yerl).ecc.
10.3±5.1
10.3±3.6
0.0±5.7
Avg. Percent Change in elec. (%) o.o
Number of Houses 6
(B) House Doctor
9.5±4.2
8.7±4.3
-0.8±2.8
-5.5
7
(C) Audit Only
12.1±3.1
11. 7±4. 3
-0.4±0.5
-4.0
6
(D) Control
10.8±5.2
9.8±3.9
-1.0±2 .1
-9.4
5
WC Resid. Avg.f
7.0
6.7
-.3
-4.0
aAn NAC is calculatEd by normalizing any specific consumgtion data to a year with average weather. Mean value ± standard deviation. Mean value ± 95% confidence interval. Because group A did not receive the extended retrofits until July 1980, the savings shown are almost entirely from house doctoring. eWe calculated the noncooling electricity consumFtion by normalizing the average daily consumption from Dec-May to a full year. The "Walnut Creek Residential Average" is based on PG&E's total consumption for individually metered residential units in Walnut Creek. The gas category has 1800 meters of which 15% are multifamily and the electric, 2450 meters, of which 37% are multifamily. In both cases, most of the individually metered multifamily units are condominiums. -12-
Although each of the treated groups (A, B, and C) did show a sta-
tistically significant drop in natural gas consumption between the pre-
and post-retrofit periods, the overlapping confidence intervals make it
impossible to decide whether or not additional savings resulted from
·"
adding house doctoring to an audit. Subtracting the audit savings
(group C) from the house-doctor savings (group B) one gets 2.1%±12%.
Neither the average savings for the audit group nor those for house doc-
tor group differ significantly from the average savings of the blind
control group. The average savings of electricity consumed for non-
cooling uses also shown in Table 3, are not significantly different from
zero or from each other. We expected this finding as none of the measures we undertook specifically addressed:thi~ electric usage.
The la.rge uncertainty in the natural gas savings of the treatment groups is due to the wide variation in the individual savings (see Fig. 1) and to the small final sample size. (Appendix B pres~nts the savings for each ho~se.) One possible driving force for this variation was the large increase in energy prices during the experiment. Gas. prices rose from 29ў to 44ў/therm (+52%) and electricity from 4.3 to 9.4ў/kWh (+118%) during the course of the experiment. * Possible mechanisms by which this price increase may have caused variation in savings are:
1. Large changes in thermostat settings, hot water use, appliance use, etc. (i.e., the "level of service" purchased by the customer) in some houses and possibly no change in others. In . this moderate climate, even small changes in the thermostat setting can cause large changes in energy use due to the· high sensitivity of heating load to indoor temperature.
2. Variation in response·to the audit suggestions among treated houses arid completion of typical audit measures, or other measures, by some of the blind control houses.
* These are the prices per kWh and therm paid by a homeowner using 200 therms/month and 1000 kWh/month during a winter month. -11-
6
HOUSE DOCTOR
5
HOUSE DOCTOR
5
(Group A)
(Group 8) Mean savings ( 12 %)
4 3 2
Mean savings (II%)
3
·
2
·
.. .
en
Q) en ::l -0 s::; 0
0
10 2
-I
I
02
2~
I
10 Q 10
- - - I
I
00
I 0
0 10 0
I
I
- - - - 10 0
0
·~ ~
0
0
0 10 Q 10
10 N .2 0 N
0 -~ 0 10 N
'-
Q)
.D E
6
z::l 5
AUDIT ONLY
(Group c) 4
Mean savings ( 9 %)
3
~.
2
0
~ -I 0 0
Q -I 0 10
10 -I 0 0
10 .2 0
I
I
- -0
10
0
0
10 Q 10
10 -N 0 0
0 ~ ~ 10
NN
6
CONTROL
5
(Group D)
4
Mean savings (7%)
3 2
·
0 ~ Q 10
- - I 0
I
I 0
~
0- I
10 0 I
10 .2 0
-0 0 10
-10 0 Q
0 -N 0 10
-10 0 0 N
0 -~ 0 10 .N
0
10 0
-N N
I
I
0~
0 10
~ I .2 Q
N I
II
- - - Q I 0
10 I 0
10 0 0
Q ~
10 0
10
I
-~ ~ 0 .2 Q~
10 0
- - N
~
0
0
0 10
NN
Percent savings in total natural gas consumption
XBL 829- 1141
Figure 1. Histograms of Energy Savings by Treatment Group.
-14-
'· .
I
I
I
I
I
20 1- 0 = House doctoring.. (groups A 8 8)
,,
-
0 = Audit only houses
..·
~group. C)
- 15 1~
~
.-...I
C\1
..... '!
··'
- - ........ ::l 10 CD
0
-
0
0
,·,· 0
0 0
0
-
0
0
-0 5. t-
·' 00
0 0
-
(J)
0
0c ' :> e0 n
0
0
0
0 0
0
c
.I . 40
0
I
I
I
60
. 70
80
90
Initial consumption (1000 Btulft 2 -yr)
··
XBL 829- 1142
Figure 2. Scatter Plot of Natural Gas Savings· per .square Foot Versus Initial Natural Gas Consumption Per Square Foot.
-15-
Additional sources of variation, not specifically affected by price changes, are: 3. Changes in the number of occupants (which occurred despite the phone survey). 4. Variation incthe success of the house-doctor retrofit efforts due to the quality of the work or the type 'of house. 5. Variable errors in our corrections for weather differences between pre- and post-retrofit years. For instance, one winter significantly sunnier than another would not only cause a bias in the NAC calculation, but the bias would be different from house to house. (As our pre-retrofit and post~retrofit data periods differ by only 9% in total degree-days this effect is unlikely to have been important.) We did not measure several parameters that would have allowed us to determine which of the possible mechanisms listed above were responsible for the variation and the savings observed. For example, measuring inside temperatures and submetering furnace gas use would have allowed us to compare the effects of efficiency changes (changes in the building shell and heating -equipment) to the effects of changes in thermostat habits. With a record of all conservation measures completed or purchased by each homeowner we could have distinguished the importance of these actions relative to house doctoring. Using accurate data on these parameters to adjust the NAG for the effects of unintended changes in the households we could have reduced the variation in the measured savings and - perhaps calculated a more certain value for the savings from house doctoring. Alternatively, an experiment using a larger initial sample could use such data to eliminate houses where the conditions changed beyond previously selected limits. Although we collected some of this information by telephoning homeowners, the "self-reported" data did not correlate well with the observed consumption changes, and we did not make any adjustments on this basis. -16-
Although our NAC calculations have not .allowed us to quantify the
savings resulting from house doctoring, our other results do permit us
to suggest improvements in house doctoring. Based on our infiltration
calculations and predicted energy savings for the various measures com-
pleted by the house doctors (see Analysis section) we will examine the

relative cost effectiveness of these measures ·
Table 4 presents the average infiltration rate in air changes per hour (ach) before and after treatment for groups A,\ B, and C. The houses in our experiment had low initial infiltration rates, the average of all houses pressurized was .49 ach. The national average for residential infiltration is thought to be between 0.7 and. 1.0 ach (A sample of '400 homes, biased in favor of energy.efficiency has shown an average of 0. 70 ach. 6 ) The two house doctor groups differed in average initial infiltration and, apparently, in suitability for house doctoring. The average reduction was .07 ach for group A and .20 ach for group B. (The individual results of pressurization tests and infiltration calculations are presented as Appendix C.)
Table 5 presents the predicted savings from infiltration reduction and non-infiltration measures completed by the house doctors in the group B houses. As mentioned in the methods section the infiltration savings will not include the additional savings resulting from reduction in convective loss through sealing of bypasses but, due to the design of .the houses, we feel these savings were small in this experiment. We assumed that water heater turndown was equally likely to result from an audit as from house doctoring and did not include the savings as part of house doctoring. The predicted average savings for group B, from all measures included on Table 5, is 1.44±29 therms/yr, or approximately 11% of the pre-retro~it average NAC.
In Table 6 we present a cost breakdown of a house doctor visit in which all the measures emphasized in our program were installed. Because each house did not receive each measure, the actual cost per house varied. The average cost for the house doctor visits to the group B houses was $457. The visits to the audit only houses cost $90 thus the average incremental cost of adding a house doctor visit to an audit
-17-
Table 4. Pre- and Post-retrofit Infiltration Characteristics of Walnut Creek Houses by Treatment Group.
Treatment Group
Leakage Area(LA)
Pre Visit Post V~sit
(cm2)
(em )
Pre Visit Spec~fi2 LA (em /m ) ·
Post Visit Spec~fi2 LA (em /m )
Pre Visit Infil.a (ach)
Post Visit Infil.a (ach)
Change in ach
Number of Houses
(A) House Doctorb (final sample)c
1104±317
940±257
.6.0±1.5
5.1±1.4-- .48±.11 ~ .41±.10
-.07
10
(.53±.06) ( .44±.08) (-.09)
(6)
(B) House Doctor (final sample)c
..I... 00
(C) Audit Only
I
(final sample)c
1495±444 1109±202
969±219
6.7±2.1 4.7±1.2
4.4±1.4
.58±.19
.38±.12
-.20
9
(.56±.19) (.40±.14) (-.16)
(7)
.39±.07
8
''( .37±.07)
(4)
aValues given are average infiltration during the heating season, presented as group mean ± standard deviation. bThe infiltration reduction presented for the A group is due to house doctoring only. CValue's in parentheses apply to houses in the final sample used for energy savings analysis: see text.
Table 5. Predicted Energy Savings, Group B Houses.
Measure
Average Savings (Therms/yr)a
Number Houses Retrofitted
Total Savings (Therms/yr)
Estimated Uncertainty (%)
Infiltration
Reduction
51
...
Duct Sealing
29
7 (all)
357
40
7 (~11)
203
so
liD
80
3
240
30
Water Heater
Blanket ·
31
(
4
i24
30
Low-flow
Showerhead
43
2
86
so
TOTAL
1010
20
AVERAGE
144±29
aThe savings gained from installing low-flow shower heads are based on a reduction of 2 gallons per minute at 11 minutes/day use~ The savings for sealing ducts are based on an average decrease in duct losses from 12% to 9%. Savings for water heater blanket and liD are from reference 7.
-19-
Table 6. Cost Breakdown for a House Doctor Visita
HOUSE DOCTOR AND AUDIT
1. Labor ($12/hour)
a. Infiltration Reduction, 10.person-hours b. Water heater insulation and low-flow showerhead, 1 hour c. Audit, 1.5 person-hours d. Set up/take down, 2 person-hours e. Installation of intermittent ignition device (IID), 2 person-hours f. Travel Time, 1.5 person-hours
Total
18 person-hours
2. Haterials
a. Water heater blanket · b. Infiltration reduction materials c. Showerhead d. Intermittent ignition deviceb
Cost/house $120 12 18 24 24 18 $216 $15 60 10 60
Total 3. Grand Total 4. Grand Total + 50% Overhead 5. Infiltration reduction alone (labor, materials, and 50% overhead)
$145 $361 $542 $270
AUDIT O~~y (includes blower-door test)
1. Labor ($10/hour)
a. Set up/take down, 2 person-hours b. Audit, 1.5 person-hours c. Travel time, 1.5 person-hours
Total
5 person-hours
2. Total + 50% Overhead
Cost/house $24 18 18 $60 $90
a This cost break. down uses a labor rate slightly above that paid to the house doctors in our project (approx. $10/hr), and assumes all measures emphasized in the Walnut Creek project were successfully installed. Because not all measures were installed in every house, the average cost for a house doctor visit in the Walnut Creek project, using $12/hr, was $457. bin this study intermittent ignition devices (IIDs) were installed by PC&E furnace servicemen. The $60 cost quoted for liDs is the bulk rate paid by PG&E. -20-
was $367. The average cost for the infiltration reduction portion of the visit was $270 per house, the non-infiltration retrofit measures averaged $97. These last costs includ~ labor, materials, and overhead, but exclude transportation arid set up times as we assumed th.e house doctor.was already in the house for the audit. The costs for non~infiltration measures include the labor and materials for each.house where the measure was installed and, because the house doctor must spend a little time on a measure even to discover it cannot be installed, 1/2 of the labor for each house where it was not installed. In calculating all costs, we used a labor rate 30% above that paid to the house doctors in our experiment. This should remain reasonable even given the greater amount of training and skill that would be necessary to implement 'the suggestions below. The Cost of Conserved Energy*· (CCE) for the infiltration reduction portion of the visit (including the effects of duct sealing), based on predicted savings, is 48ў/therm ,(group B only). The CCE of the noninfiltration related house doctor activities, based again on predicted savings, was 21ў/therm (group B only). For each of the non-infiltration measures, there were houses in which that measure could not be installed without more materials, training, or time. The.lower CCE of the noninfiltration measures indicates that the overall cost effectiveness of house doctoring could be increased, at least in this type of housing and climate, by diverting some time and expense from infiltration reduction to .achieve a greater success rate with other measures~ For example, more time spent explaining the use and benefits of low flow showerheads, and stocking the vans with more plumbing equipment in order to install them in more situations, would increase the savings from this measure without raising the CCE above that for infiltration reduction. Additional measures, such as installing reflective window screens for cooling reduction or retrofit switches for "i.n. stant-on" TV. sets , could be added to . the * The Cost of Conserved Energy is equal to the cost of the treatment, multiplied by a capital recovery factor, and divided by the therms saved per year. The capital recovery factor is calculated on the basis of the' expected useful life of the retrofits and the real interest rate. We have used 10 years, a conservative estimate, and. 6% per year. -21-
house doctor repertoire. In colder climates, however, where each cm2 of leakage area causes more energy use, and in leaky houses, infiltration reduction will likely remain an important function of house doctoring. Based- on pressurization data, pre-retrofit infiltration losses in our houses were responsible for 10% of total fuel consumption, on the average, with a range of I 7.5% to 18%. The predicted average percent savings from infiltration reduction (group B) was 4% of the total consumption, with a range of 1% to 7%. The small contribution of infiltration to the fuel bill limited the percent savings available from infiltration reduction and the mild climate limited the absolute savings. Our research indicates that in other parts of the u.s. housing stock, including areas of mild climate, infiltration levels may be higher and the leaks easier to correct than'in our Walnut Creek sample. "' Scattered pressurization tests done by LBL have found many houses leak- ier than those in Walnut Creek. We have measured infiltration rates up to 2 ach. In addition to low initial infiltration rates, the houses in the Walnut Creek project had low crawlspaces and attics which hampered the detection and sealing of leaks. Much of the U.S. housing is older than that in Walnut Creek and, due to designs common in older housing, more suited to house doctor in< filtration reduction. * This evidence, and the positive correlation between the reduction in infiltration and the initial infiltration rate (Figure 3), suggests that more than twice our average savings . of .2 ach (group B) could be expected in a significant fraction of houses .across the country. *In addition to roomier attics and crawl spaces, older housing tends to have problems such as dropped ceilings and other similar bypasses. These localizeD. Problems (usually fixed by covering with a layer of plastic and insulation) take perhaps an order of magnitude less effort to correct than leaky ceiling-to-wall joints. (one common design in our study had over 150 feet of such cracks) for probably the same infiltration savings. -22-
0 ')
0
0
0 0 00
0
000
..
0
·o..,
00
0'
0
0 0
0.5
1.0
Initial infiltration (ach)
XBL 829- 1141
Figure 3. Reduction in Infiltration Rate Versus Initial Infiltration Rate for Walnut Creek House Doctored Houses (Croups A and B). The values presented represent average infiltration for the heating season. ·
· '. -23-
Three tasks remain in the analysis of the Walnut Creek data. The billing data collected for the extended retrofit group after the completion of the contractor retrofits will be analyzed by PG&E and the savings compared to the house doctor and other groups. Analysis of the second year of post retrofit data for the house doctored houses will reveal the durability of the observed decrease in gas consumption and perhaps give a better idea of the source of the savings. Submetered furnace consumption is now available for some of the houses. Analysis of this data will provide a check on the accuracy of the NAC calcula-. tions. CONCLUSIONS In this study the house doctored, audited, and control houses all showed savings in natural gas consumption but, due to the small sample size, no statistically significant difference was measured between any of the treatments. Although our experiment was inconclusive with respect to the original question, whether adding house doctoring to an conventional audit is cost effective, the results do allow us to draw some valuable conclusions. As noted in the discussion, experiments that attempt to measure energy savings in occupied houses, especially those with small sample size, require end-use submetering, indoor temperature monitoring, and careful attention to homeowner conservation actions. Without such data even statistically significant changes in energy use may not allow the experimenters to attribute the savings to specific causes. Only thorough data collection can resolve such questions. Fortunately, several lowcost systems for multi-channel data monitoring and storage are being developed. Houses vary widely in their suitability for house doctoring. House doctoring every house, or every house in an experimental sample, is similar in concept to adding R-19·insulation to ceilings regardless of how much insulation is already present, or whether the house has an attic or a cathedral ceiling. Diagnosing a house before doctoring is, -24-
therefore, a part of any reasonable house doctor program. The diagnosis, most likely a pre-visit blower door test and walk-through, will clarify whether major infiltration reduction, completion of noninfiltration measures, or no treatment at all, is appr.opriate.
We believe that the retrofit package completed during house doctor-

ing can be cost-effective, even with the problems we experienced ·
Because results from better monitored houses8 ' 9 .and laboratory experi-
ments support the methods used to calculate the predicted savings, and
because our measured savings are not incompatible with the predicted
savings, we are confident that the predicted savings for the retrofit
package are a fairly good estimate of the savings we would have measured
had no other changes occurred in the houses.· The CCE for our house doc-
tor visits, based on the predicted savings from the infiltration and
non-infiltration measures, and costs including transportation, was
42±8ў/therm. This is cost effective relative to the marginal price of
gas to the homeowners in the experiment (67ў/therm, Jan. 1981) and to
the price of new natural gas supplies in most locations. A similar
house doctor program employing a pre-house-doctoring diagnosis and/or
operating in a' colder climate should have a significantly lower CCE.
House doctoring can only become increasingly important as , conventional retrofits (e.g. attic insulation) reach saturation and the less obvious measures constitute a growing fraction __ of the conservation potential. Many questions remain to b.e resolved, however, before beginning a large scale utility house 4octor and audit program. Combined with an audit, pr as a separa·te service, house doctor procedures need refinement and house doctor savings need better measurement. We strongly encourage further research into these areas.
ACKNOWLEDGEMENTS This work was supported by the Assistant Secretary for Conservation and Renewable Energy, Office of Building Energy Research and Development, Building Systems Division of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098, and by the Pacific Gas and Electric Company. -25-
REFERENCES
1. Wall, L.W., et al. Building Energy Compilation and Analysis (BECA)
Part != Existing North American Residential Buildings. Lawrence
.·.
Berkeley Laboratory report, LBL-13385. July 1982.
2. Dutt, G., et al. The Modular Retrofit Experiment: Exploring the

House Doctor Concept. Princeton University/Center for Energy and
Environmental Studies, Report No. 130 June 82.
3. Diamond, R.C., et al. The House Doctor's Manual. Lawrence Berkeley Laboratory report, PUB-3017. February 1982.
4. Fels, M.F., Goldberg, M.L. Calculating Savings on the Standard Living Cycle. Princeton University/Center for Energy and Environmental Studies, Report # 131 1982.
5. Sherman, M.H., and D.T. Grimsrud. Measurements of Infiltration Using Fan Pressurization and Weather Data. Lawrence Berkeley Laboratory report, LBL-10852. October 198~
6. Grimsrud; D.T., R.C. Sonderegger, and M.H. Sherman. Infiltration Measurements in Audit and Retrofit Programs. Lawrence Berkeley Laboratory. LBL-12221. April_1981.
7. Wright, J., et al. Supplying Energy Through Greater Efficiency, The Potential for Conservation in California's Residential Sector. Lawrence Berkeley Laboratory report, LBL-10738. January 1981.
8. Dickinson, J.B., et al. Results of the Bonneville Power Administration Weatherization and Tightening Projects at the Midway Substation Residential Community. Lawrence Berkeley Laboratory report, LBL12742. February 1982.
9. Socolow, R.H. ed., Energy Savings in t9e Home: Princeton's experiments at Twin Rivers, Ballinger, Cambridge Mass. 1978.
-26-
Appendix A. List of telephone survey questions for prospective participants.
Question
Desired Answer
1. Have you livedin this house for the
past two years?
Yes
2. Do you plan to move from this house
anytime in the next two years?
No
3. Do you have and use any of the
following?a
Wood. stove
Major electric space heater
No
Hot tub or Sauna
Heated Swimming Pool
Solar. heating
4. Have you ever had an Energy Audit,
or do you have one now scheduled?
No
5. Do you have and use air conditioning
in your hoine?
Yes
6. Have you added a room, or made any
other major structural changes to
your house in the last two years?
No
7. Has there been any change in the number
of people living in your house in the
past two years?
No
8. Do you expect a change in the number
of people living in your house in the
next two years?
No
9. Have you taken any energy conservation
measures in the last two years?
For example:
Caulking
Insulation
No
Furnace Modifications
Other
ain our study this question was mistakenly worded: "In the last two years have you purchased any of the following?".
-27-
Appendix B. Meas'.lred savings in the Walnut Creek test houses.
Treatment ID Group
Pre-visit NAC (therm/yr)
Measured Change in NAC. (therm/yr)
Measured Change in NACa (%)
A1
House
A3
Doctor
and
A6
Extended A7
Retrofitb
A9
A10
1572 1062 1086 1790 824 986
-54 -93 --192 -382 -90 -58
-3.4 ±12. -8.7 ±5.9 -17.7 ±9.1 -21.3 ±6.0 -10.9 ±10. -5.9 ±5.8
B1
1123
B2
1115
House
B6
Doctor
B7
1221 1722
B8
1642
B9
1116
B10
1405
+76 -100. -300 .-322 -133 -222 -102
+6.8 ±7.0 -9.0 ±11. -24.6 ±9.3 -18.7 ±7.3 -8.1 ±11. -19.9 ±6.3 -7.2 ±8.3
Audit Only
C1
982
C3
1133
C5
1155
C6
2069
c9·
1487
C10
1252
+2 -60 -120 -260 -381 -31
+0.2 ±8.2 -5.3 ±6.6 -10.4 ±9.1 -12.6 ±8.2 -25.6 ±5.0 -2.4 ±10.
D2
Blind
D6
Control
D7
D8
D9
1195 1997 1886 814 991
-195 -58 -286 156 -186
-16.3 ±20. -2.9 ±10. -15:2 ±8.7 19.1 ±18. -18.8 ±7.5
a The 95% confidence interval presented with each savings was calculated
with the method used by ~els and Goldberg (Princeton) on monthly aggre-

gate data for communities · However, we have some doubt that the monthly
data for individual houses have the necessary independence of error
terms for this method to be rigorously applied. For this reason the
variance of each individual saving~ was not used in finding the· group
average presented in the main text. The savings shown for the extended
retrofit group are from house doctoring only. See main text.
-28-
Appendix C. Comparison of infiltration characteristics of Walnut Creek test houses before and after house doctoring.
Pre-visit Post-visit Pre-visit Post-visit
Treatment ID Floor Leaka§e
Leaka§e
Infil-
Infil-
Percent
Group
Ar~a
Are~
Are~
trat"on
trat~n
Change
(m )
(em )
(em )
ach
ach
in ach
Al
245
1518
1298
.48
.41
-15
A2
176
993
800
.43
.35
-19
·~
A3
145
980
906
.51
.47
-8
House
A4
174
1020
1126
.45
Doctor
AS
152
959
611
.49
Extended A6
142
941
802
.52
Retrofitc A7
245
1817D
1309D
.58
A8
316
866.
945
.22
A9
172
815
567P
.48
AlO 139
1132P
1040
.63
.50
+11
.31
-37
.44
-15
.42
-28
.24
+9
.34
-29
.58
-8
Bl
202
1672P
1269D
.85
B2
213
1491
1055
.54
B3
241
.65
-24
.38
-30
B4
261
2132
820
.83
.32
-61
House
B5
204
1170
807
· 45
.31 .
-31
Doctor
B6
179
1165
892
.so
.38
-24
B7
263
948
813
.37
.32
-14
B8
248
2048
1236
.64
.39
-39
B9
196
1807
1171
.72
.47
-35
BlO 317
1026
660
.33
.21
-36
Cl
200
1171
~45
C2
220
934D
.33
C3
206
C4
179
1145
.49
Audit
cs 223
760
Only
C6 . 243
905
.29
C7
251
1208
.37
C8
219
1261
.44
C9
259
1'401
.39
ClO 261
1200
.35
aA "P" follo. wing a leakage area value indicates that value is based on pressurtzation data only. A "D" indicates a value based on depressurization data only. ·The model used to calculate infiltration rates is considered to be accurate to within 20%. cThe savings shown for the extended retrofit group are from house doctoring only. See main text.
-29-
This report was done with support from the Department of Energy. Any conclusions or opinions expressed in this report represent solely those of the author(s) and not necessarily those of The Regents of the University of California, the Lawrence Berkeley Laboratory or the Department of Energy. Reference to a company or product name does not imply approval or recommendation of the product by the University of California or the U.S. Department of Energy to the exclusion of others that may be suitable.
TECHNICAL INFORMATION DEPARTMENT LAWRENCE BERKELEY LABORATORY UNIVERSITY OF CALIFORNIA BERKELEY, CALIFORNIA 94720 ) / !

BC O'Regan, BS Wagner, JB Dickinson

File: energy-environmeni.pdf
Author: BC O'Regan, BS Wagner, JB Dickinson
Author: Kodak Capture Software
Published: Wed Sep 26 19:01:57 2012
Pages: 33
File size: 1 Mb


Over the rainbow, 17 pages, 0.33 Mb
Copyright © 2018 doc.uments.com