Florida's Phased Deep Retrofit Project
What are the realistic energy savings that can be obtained in homes in a hot-humid climate? As part of U.S. DOE’s Building America Program, phased deep retrofit (PDR) project, the Florida Solar Energy Center (FSEC) is retrofitting 58 Florida homes. FSEC is collecting detailed pre- and post-retrofit audit data in order to evaluate potential energy reductions. Since February 2013, a series of shallow retrofit measures have been installed. As of the end of May, there were enough of these completed to allow an estimate of the pre- and post-retrofit savings in about half of the homes. We will cover the progress to date in this article.
The project’s ultimate goal is to evaluate a “phased approach” to energy retrofits, where simple pass-through measures are installed at the time that data is collected on the house and systems. Later, armed with the information, we plan to evaluate much more significant measures, or deep retrofits, installed in a sub-sample of the homes. The homes are highly instrumented to allow us insight into the various measures installed.
Shallow Retrofit at Site #7—March 28, 2013
Table 1. One Month Pre- and Post-Retrofit: Measured Savings
The measures being installed in the shallow retrofits at the homeowners’ discretion are:
- Change all incandescent lighting to CFL or LED lighting
- Add exterior insulation tank wrap to hot water tank
- Replace shower fixtures with high-efficiency heads if measured flow is greater than 2.2 gpm
- Set pool pump hours to no more than 5 hours per day
- Clean refrigerator coils if dirty
- Provide smart power strip to any standby power loads greater than 10 Watts continuous
Based on project results so far, the latter measure was very seldom installed, and as a result, we don’t report on it in this early evaluation. The limiting issues with this particular measure will be further described in a later report.
To complete the evaluation, we examined the monitored power for the hot water, pool, refrigerator, and lighting. We calculated energy use for lighting, ceiling fans, and plug loads by subtracting all of the end uses from the measured total power. This residual is the energy use of plugs, lighting, and ceiling fans. We also obtained total building power to see if the shallow retrofits were visible to homeowners who simply have utility bills to observe.
The retrofit savings estimation methodology was to evaluate power use on each potentially affected circuit for one month before and one month after the retrofit. Since the measurement period spanned two months and most of the measured appliances and end-use loads have some natural seasonal variation, we also measured the average change in the end-use load over the same period in the homes that had not yet received the retrofits. In this way, it was possible to provide a control for weather-related influences.
Figure 1 shows an example of what we observed in daily power for the affected end uses at Site #7 one month before and after energy savings measures. The measures were installed from March 14–May 1, 2013 with recorded data on the retrofit particulars: number of lighting fixture changes and types, measured flows of showerheads, pool pump on time, and hot water tank location and status. Much of this data will later be summarized in a full project report.
In this preliminary evaluation, we examined the average reduction in each affected end-use load of the pre- and post-retrofit groups. A t-test of paired mean averages before and after the retrofits was used to establish statistically significant differences.
As all of the houses are all-electric ones, as is common in Florida, the first thing to note is that when looking at total energy use, it tended to increase in the month after the retrofit. Negative savings? Why is that? As the retrofits were undertaken during the spring, air conditioning was increasing rapidly in the homes at the end of the monitoring period. Thus, the natural increase in energy use associated with cooling masked the actual savings that could be observed by examining the circuits directly affected.
Further, it should be noted that in order to control for seasonal fluctuations, the kWh pre-values in Table 1 have been adjusted based on the end-use load measurements in the control (non-retrofit) homes during the same period (see “Non-Retrofit Sample Data”). This is necessary so as to not underestimate some measures where energy use is naturally increasing with higher seasonal temperature (for example, refrigerators) or to overestimate savings from others that are naturally dropping (water heating).
The estimated average energy savings in the affected loads in the overall sample of 26 retrofit homes was 3.68 kWh per day for a 9% savings. This savings in the affected loads was hidden in the total loads by an increase in cooling over the period as air conditioning increased with warmer late spring conditions.
The savings rate was strongly affected by the presence of swimming pools and by installation of measures. For instance, the savings in homes with swimming pools averaged 5.01 kWh/day (11%) versus 1.75 kWh/day (4%) in homes without pools. Although the swimming pool pump time reduction had the largest impact, it was also highly variable, as four homes already operated at 5 hours or less per day and this was not changed. However, the other six homes where hours were reduced saved an average of 6.06 kWh/day each.
Weather-adjusted savings to hot water energy use averaged about 0.5 kWh/day or about 9% of hot water energy use. We examined the impact of changing showerheads versus tank wraps, but the data did not support statistical determination of specific impacts. Smart power strips were installed on five of the 26 sites, although dedicated power circuits were not available to see their activity.
Non-Retrofit Sample Data
The unaltered non-retrofit sample saw loads change over the month as follows:
- Total household electricity consumption increased by 12% (37.1 to 41.4 kWh)
- Water heating dropped by 9.8% (unadjusted DHW savings were 18%)
- Refrigerator kWh increased by an average of 6.6% (no savings in the unadjusted sample)
- Pool pump energy increased by 6.2%
(unadjusted savings were 24%)
- Lighting/ceiling fans and plug loads grew by 6.5% (unadjusted savings were approximately 10%)
The variable COVAR (Table 1) is the coefficient of variation of the savings achieved (standard deviation over the savings level). As such, low values of this parameter indicate reliable savings, while values >0.5 indicate declining reliability. This clearly shows that while swimming pool pump adjustment had the largest impact, it also varied very significantly from one house to the next. Since the sample was very small (n=10), it was not statistically significant. However, we expect that as the shallow retrofits are completed, this will change and pool pump adjustment will emerge as a robust and very effective measure for saving energy.
Not surprisingly, the lighting retrofit change was the most reliable measure, although not all homes could take advantage of this improvement. Indeed, one home had all LED lighting, whereas four others had mostly CFLs and needed few fixtures changes. If homes that had fewer than 20% of fixtures changed were eliminated from the analysis, the savings of the lighting retrofit measure increases from 1.11 to 1.54 (+0.43) kWh/day. The refrigerator coil cleaning was also reliable in savings, although the very small savings level (4% or 0.11 kWh/day) is not detectable unless one compares the group of refrigerators not receiving coil cleaning over the same period.
Overall, the most important of our energy-saving measures are pool pumps followed by the lighting retrofit and then water tank blanket.
One issue we discovered regarding customer expectations with the shallow retrofits is illustrated in Table 1. The savings level is small enough (approximately 10% of pre-retrofit consumption) that focus on the overall billing data (Total kWh/day) before and after the retrofits will not reliably reveal the savings. This is reflected in the averages in the far right column. Note that in the sample of 26 homes, the average total consumption was 1.6 kWh higher in the month post-retrofit than the one before it. This is because changes in heating and cooling energy can easily swamp the savings level generated from the shallow retrofit measures, depending on when they are installed.
Because the retrofits in this pilot were being completed in the spring, when air conditioning is increasing, the rising level of space cooling in the post period generally hid the generated savings, which instead appeared as a lower-than-expected consumption increase in the post-retrofit month. (In the non-retrofit sample, the consumption was 4.3 kWh higher in the post month). Of course a similar issue will be encountered for retrofits done in the late summer, as the post-retrofit reduction of cooling will exaggerate savings.
One way around this issue is to examine the annual 12-month utility records before and after the retrofits. Such an analysis is planned within this project. These results, however, will have to await the accumulation of sufficient post-retrofit period data—a year is needed with no further alteration of the homes. This point should come in the early summer of 2014.
After adjusting for weather-related changes, savings of the overall shallow retrofits in the first 26 homes averaged 3.7 kWh/day or about 9% of pre-retrofit monthly consumption. Potential for reducing pool pumping hours is large and is a key factor in an effective program. Homeowner information also may be useful, as most do not know pumps are costing $50 a month when operated 8 hours per day. The lighting retrofit measure is effective. Hot water savings are large enough to continue to emphasize tank wraps and showerhead change-outs (approximately 190 kWh/yr).
Parker, D.; J. Cummings, D. Chasar, J. Montemurno, D. Hoak, B. Amos, and J. Nelson. Pilot Demonstration of Phased Retrofits in Existing Homes in Florida: Preliminary Data Collection and Project Status Update. Prepared for Building America Program with the U.S. Department of Energy, Florida Solar Energy Center, Cocoa, Fla. Prepared for Florida Power and Light, Company, Miami, Fla.
Savings from refrigerator coil cleaning were very small and could be ignored in a mature program. Rather than cleaning, refrigerators should be monitored over a multi-day period. Those found be using greater than 3 kWh/day in electricity should probably be slated for replacement with newer, more efficient models. While savings from a shallow retrofit program in Florida are effective, reductions are small in magnitude, and customers may not see their impact when comparing month-to-month billing data.
The above results will be revisited when the overall shallow retrofit sample is complete and sufficient post-retrofit data is available in late summer. We will also update the analysis methods with whole-year billing data, pre- and post-retrofit, after one year’s time. Look for more results of the phased deep energy retrofit project in a future Home Energy.
Enter your comments in the box below:
(Please note that all comments are subject to review prior to posting.)
While we will do our best to monitor all comments and blog posts for accuracy and relevancy, Home Energy is not responsible for content posted by our readers or third parties. Home Energy reserves the right to edit or remove comments or blog posts that do not meet our community guidelines.