FAQ

1. The minimum degree-day/day setting should be left at its default of 0 unless one of the exceptional conditions applies, as described in the next statements

2. If the regression is done for HDD (or CDD only), and if the Usage (or Demand) vs. Degree Days graph shows the bills clustering around a fit line, plus several bills clustering on the y-axis, then:

   1. If in the Usage (or Demand) vs. Outside Temperature graph you see a linear dependence on outdoor temperature, that "dies" into the x-axis, then set the minimum degree-day setting to 0;

   2. If, on the other hand, you see both a temperature-dependent region as well as a plateau, where bills level off at a non-zero level, because of a significant non-temperature dependent load (e.g. summer gas usage due to water heating), then:

   1. If the usage vs. DD plot shows the cluster on the y-axis falling more or less on the intersection of the regression line with the y-axis, i.e., as many points fall above as below of the intersection point, and the points are close to the regression line. In this case, the non-temperature sensitive baseload appears to be constant year-round. Consequently, the threshold is not necessary, and can be left 0.

   2. If the usage vs. DD plot shows the cluster on the y-axis falling mostly above or mostly below the intersection of the regression line with the y-axis, or, while "centered" on the intersection, there could be "large" up and down deviations. All of these cases are, in essence, caused by a variable baseload, e.g. water heating of a swimming pool with seasonal operating hours. In this case you may need a minimum degree-day / day threshold to prevent the points clustering around the y-axis from adversely affecting the regression. There should also be some additional modification for those points (that cluster on the y-axis). Failing that, you will have a large net bias and a bad RMSE, in short, an incomplete model.

      The best way to address this is through a second independent. variable, e.g. pool operating hours. If properly chosen, this second variable will greatly improve the overall model fit and eliminate the need for a threshold.

      For those users unable to obtain, or rationally construct, such an independent variable, bill modification for the excluded points remains the last resort .

      This last case should be avoided whenever possible (according to GPC14P, who frown on that kind of ad-hoc fudging).But if bill modification of excluded points is necessary, then a threshold value greater than zero should be used to exclude those bills that are most influenced by the non-weather-sensitive baseload.

I can’t find the answer to your question ? Try checking our User Forums.

Otherwise feel free to contact our Tech Support staff at (805) 329-6565, or via email at helpdesk@abraxasenergy.com.