Thanks a lot.

Virginia Lindahl
Alexandria, Virginia

Bruce Nordman, principal research associate at Lawrence Berkeley National Laboratory, responds:

There is no reason, other than a person's patience, not to turn your computer off if it will be sitting idle for any period longer than a few minutes. In fact, from a physics perspective, taking more energy to boot up a computer than to leave it on can't be true. Here's why. A computer and monitor typically use 150 watts between them, which is about 1/12 of the electrical capacity of a typical 15-amp electrical circuit. If we assume that a personal computer (PC) uses enough energy during booting to compensate for not turning the computer off for, for example, 12 minutes that it would otherwise have been idle, then the PC would have to use electricity at 12 times its normal rate during booting. At this rate the PC would use all the capacity of the electrical circuit. If the PC used any more power to boot up, this would trip the circuit breaker. So to save energy, you are best off powering the machine down or off as often as possible. In small offices where the printer is hooked up to the computer and gets switched off when the computer does, even more energy will be saved every time the computer is turned off.

As to downsides, there is a theoretical possibility of weakening the solder joints by frequently turning the PC on and off, which could lead to premature hardware failure; but in practice this isn't a problem. Disk drives used to be sensitive to powering on and off, but this hasn't been true for over a decade.

The practical limit to power management is people's patience. Most office workers won't want to boot up their PC more than once a day, due to the waiting time. They may be willing to turn the monitor off several times during the day (for example, during the lunch break or while in meetings), as monitors turn on quite quickly. Better yet, using the power management capability built into most PCs and monitors today, the machine can go into a low-power sleep mode when not used for a time, then recover within seconds when a key is pressed or the mouse is moved. This capability can reduce the power draw by about 50% for PCs and up to 95% for monitors.

So, yes, the idea that turning things off ends up using more energy than leaving them on is an urban myth. Using whatever energy-saving techniques are available, as frequently as possible, can result in significant energy savings.

A matched-pair study design is used to measure savings; however, there is not enough detail in the report to establish whether the savings reported by the author are due to anything other than chance. We are not told the actual R-values of the components in the houses, the window areas and U-values, or the infiltration rates. We must assume they are identical (save for the wall type); the study does not document that they are. There is no mention of the types of heating system in the matched pairs; that is, one could contain a ducted heating system and the other zonal heat. The climates used for comparison are not carefully described, so a matched pair in the Canada district could conceivably pair a house in a maritime climate with a house on the open prairie. There are many other potential sources of error that are not carefully addressed in VanderWerf's report.

The Portland Cement Foundation (sponsors of the report) would have better spent their money in a more closely controlled prototype analysis in a few climates. In fact, guarded hot-box tests have now been performed on several ICF walls and have found R-values in a range that seems reasonable (R-17 to R-25), given the wall materials and my own experience modeling several of these walls with contemporary simulation programs (under hire to various ICF manufacturers).

I know Alan Meier advised caution on ICFs in his editorial of that issue, but I think Home Energy should be more careful with the technical review of what is passed off as research.

Bob Davis
Ecotope
Seattle, Washington

Dr. Pieter VanderWerf responds:

Many thanks to Bob Davis for raising important and useful issues. Some of Mr. Davis's concerns are covered in our report. We tested the possible role of chance by the conventional method, the confidence interval. This indicated that the savings, with over 97% confidence, were not the result of chance alone. Air infiltration is not assumed to be constant. In fact, we know from blower door tests that it is significantly lower in homes with ICF walls. Reduced infiltration is actually believed to be the greatest single factor accounting for the lower energy consumption of ICF homes.

A few of Mr. Davis's concerns were unfortunately not covered because of the summary nature of the report. During our study, we did determine several details of the HVAC systems used and we corrected for differences. Matched pairs of houses were within 2 miles of each other in 80% of all cases; in a few situations we were forced to go as far as 20 miles to get a fair match.

A few of Mr. Davis's other concerns are about statistical research in general. In any research, it is impossible to control for all variables. Monitoring and measurement studies cope by eliminating or controlling as many sources of variation as are practical. Statistical studies measure and correct for some variation, and determine whether other sources of variation might introduce biases. The researcher checks independent data or a subsample to make sure that uncontrolled variables (for example, fenestration) do not tend to be significantly different for some groups (for example, ICF houses) than for others. We did this and discussed it in our report. With these conditions, statistical theorems show that point estimates (for example, average energy savings) will still be valid. If one does not accept this principle, one should reject our study along with all statistical field research.

Three implied statements in the letter trouble me. The first is that the energy performance of ICF walls can be characterized by their R-value. Work at Oak Ridge National Laboratory and elsewhere shows that several other factors are of comparable or greater importance: reduced air infiltration, thermal mass, and (possibly) conduction of geothermal energy. The energy modeling of ICF structures to date is unfortunately not of much practical significance because it represents the effects of only one or two of these mechanisms.

The second is the implication that controlled prototype analysis is adequate to fully characterize the energy performance of a wall system. The Portland Cement Association and other organizations have sponsored some projects to monitor the energy consumption of side-by-side prototypes. But they have also sponsored laboratory testing, computer simulations, and statistical field studies such as this one. I hope that building science follows the lead of the other sciences by demanding that final analysis of any important phenomenon rest on results from a variety of different research methods, not just one. Each has its unique lessons to offer.

The third is the implication that peer-refereed research by scientists is the only source of useful information. In an 11-year career as a university professor writing almost nothing but refereed journal articles, it has become obvious to me that this attitude, if pervasive, would halt progress as we know it. Virtually none of the important energy-saving products we rely on today had unequivocal, fully documented performance when the first engineers designed with it, the first architects specified it, the first contractors installed it, and the first building officials approved it. It is true that there will always be charlatans who overstate the performance of their products. But unless we can rely on the training, observations, common sense, and judgment of the practitioner to assess product claims and preliminary research, nothing new will ever get off the ground. Let us in the research community take our decade to achieve full precision and consensus on a new phenomenon. Energy professionals have to estimate paybacks, and HVAC contractors have to size equipment now.

Current MECcheck versions do not address the International Energy Conservation Code, the 1998 version of the Model Energy Code, as described in Energy Code Goes International in the same issue (p. 7). MECcheck does evaluate low-rise residential buildings for compliance with the 1992, 1993, or 1995 versions of the MEC. Our current software versions allow the user to select which MEC edition is applicable to her or his residential design.

MECcheck software runs from any Windows platform (3.1, 95, or NT), as well as DOS-only platforms. Users should order or download any MECcheck version designated 2.01 or higher to ensure that they are getting a Windows application. (Version 2.07, release 3, is the most current version.) Residents of Arkansas, Massachusetts, Minnesota, and Vermont can obtain MECcheck versions specifically applicable to their state.

Your Software Contacts sidebar should have included the Web site where MECcheck software (and other nonsoftware compliance tools for the MEC) can be downloaded, free of charge. DOE's Building Standards and Guidelines Program site is at http://www.energycodes.org.

Those persons who do not have Internet access can obtain MECcheck software and other products by calling DOE's Hotline at Pacific Northwest National Laboratory: (800)270-2633. Hotline orders require a small fee to cover duplicating and mailing expenses.

Stephen Turchen
Office of Codes and Standards
U.S. Department of Energy
Washington, D.C.
 

Home Energy can be reached at: contact@homeenergy.org
Home Energy magazine -- Please read our Copyright Notice