Cooling Degree Day Normalization

    The concept of cooling degree day (CDD) normalization is used by utilities and energy auditors to compare one year to another and one geographic area to another.  It is a "low precision" tool but it does have a place in such use.  It is based on a very simple calculation for each day.  The highest temperature and the lowest temperature for a given day are averaged and 65 degrees is subtracted from the result.

     The number cannot be negative. As an example, a day with a high of 95 and a low of 75 is averaged to 85 degrees.  Then 65 is subtracted from 85 and the result is 20 degrees for this individual day.  The individual days are added and each week, month and season "earns" a total number of cooling degree days.

     Over the course of an entire season, using CDD normalization for comparing residential air conditioning strategies has some merit.  For example, a given summer "earns" 2,000 CDDs while we are using air conditioning strategy A.  The next summer earns 2,500 CDDs while we use strategy B.  The CDD technique says the second summer had 500 more cooling degree days; it was 500 divided by 2,000 times 100% or 25% warmer and under "normal" conditions we should use 25% more energy if all other factors are equal.  If strategy B in the second year gives a result that increases costs by less than the 25% increase in CDDs, it can be assumed to have saved energy and that strategy is assumed to have merit. 

    So far, so good!  But here is where things do not work so well.  We recently saw a test from a Nebraska Pizza Hut and the tests were normalized using CDDs.  The week before the test was quite warm and after IceCOLD was installed the weather got cooler.  Using CDD normalization technology, the surveyor got no meaningful results for one of the two AC units.  The indication was that IceCOLD did not work in one system but did work in another.  The problem was not the product but the testing. Cooling degree days use a baseline of 65 degrees. The assumption is that little or no cooling is required below 65.  In a Pizza Hut it certainly is.

    Reviewing the data we saw that a considerable amount of cooling was required below 65 and at temperatures down to the lowest available during the test.  Nonetheless, three of the five days of the testing after installation, "post-test," earned no degree days at all.  The daily average for these days was below 65 degrees.  A far more realistic concept for commercial air conditioning is "growing degree days" (GDD.)  This technique uses a base temperature of 50 degrees and GDD information is readily available.  At this restaurant, even this did not work, two of the five days in the post-test period earned no GDDs, either.

    The Pizza Hut trial demonstrates the short-comings of CDD or even GDD techniques for very short tests like this one.  Even using GDDs, there were two days where the cooling ran, but no "DDs" were earned.  The greatest concern about any such technology is that neither gives us real information about anything but the "relative" warmth of the weather.  As an example, let's create an imaginary day with very high humidity that has a low of 75 degrees in the morning and it warms to 90 in the afternoon.  About closing time a storm moves in and the temperature drops to 60 degrees outside.  This day earns 15 degree days just like a much milder day that never got hot or humid.  Nothing so far reveals why one of the units used more power after the installation than it did before.  Any normalization insists that it was cooler, regardless of how it is calculated.

    There are other techniques that do give us reliable information.  In our next post we will discuss alternatives to cooling degree day normalization and whether there may other ways to look at IceCOLD and get meaningful data.  We will look again at this Pizza Hut with a very different approach.  HINT: It's GOOD