Comparing Fuel Consumption Stats Season to Season

FJL

I'm trying to figure what conclusions to draw about my fuel consumption based on the statistics that my gas company gives to me. I presume I should compare fuel usage per degree day when comparing one heating season to another, as opposed to usage per day, because degree days is a statistic designed to take into account differences in temperature between each heating season. If there is a better way, please tell me.

If usage per degree day is the best way to compare fuel usage among heating seasons, how much variance can I expect to find when comparing usage per degree day from one season to another. If two heating seaons with an identifical number of degree days, how wide a fluctation could I expect in usage per degree day if everything else is the same? I assume that one day w/an avg temp of 25 might require more or less fuel consumption than another day w/an avg temp of 25.

Here's why I'm asking. During the first 60 days of this heating season, my building used .52 therms per degree day, as compared to .56 last year. How much of that difference can I attribute to differences in a lower t-stat setting (70 vs 72), longer set back period (night and day vs night only), etc ..., vs. what would be a natural fluctation in usage for the same number of degree days?

Thanks for sticking with me.

Jamie Hall

Mark Twain once said...

there are lies, d***ed lies, and statistics. Now that that's off my chest...

To begin with, yes you do indeed need to start off with the the degree day values as your principle (most significant) independent variable. How much of the variation in the amount of fuel used will be accounted for by the degree days? Good question. 'Quite a bit' is not really a very good answer, I know, but... in the building for which I am the superintendent (where I've been doing this for years, as it impresses the bean counters) the standard deviation on the coefficient relating heating use to degree days is about a third of the value of the coefficient (statistics on usage have been kept for 15 years so far) -- which suggests that more than half the variation is attributable to degree days. Now mine is a particularly draughty old building, and much of the rest of the variation is attributable to how hard the wind blows. Another significant variable is the amount of sunshine -- even in a non-solar building it makes quite a difference. Then another source of variation is in the degree-days readings themselves (unless you are measuring them properly at your location).

In my particular building (as the EPA says, your results may vary!) changing the thermostat setting from 72 to 70 (keeping the same setback) would save about 7% on fuel, which isn't too far from the change you are enquiring about. A longer (but not deeper) setback will also save fuel.

All that help any?

FJL

Maybe A Better Way . . .

for me to understand is to ask it this way, because I got lost a bit when you started talking about coefficients.

Lets say in Year 1 my usage is .55 per degree day. In Year 2, assume the number of degree days is the same. What is the range of usage per degree day that I could expect to see if everything else between years 1 and 2 is equal?

Jamie Hall

hard to say

but with a moderately typical house (typical infiltration levels etc.) I wouldn't be at all surprised to see at least a 10% variation between two weeks with the same degree day totals. Over the years it should even out -- but there will still be variation.