|Metrix » Tuning|
Baseload can be defined as the residual daily energy consumption when all variables (DD and others) are zero. In a single-variable baseline model, such as "Heating energy vs. HDD", the baseload is simply the y-intercept. For example in the tuning screen below, the baseload is 535.60 therms per day.
Now comes the rub. "Excluded months" (from the regression, not from the model!) are not used to determine the baseload value. Indeed, you could easily imagine a school with a conventional "Gas vs. HDD" relationship during 9 school months of the year. Let us assume for clarity that these 9 months all have HDD > threshold, and the remaining 3 months have HDD < threshold. The regression will likely indicate a positive intercept, e.g. 100 Therms/day, indicating that during an exceptionally warm spring week with HDD < threshold, the school would still be using 700 Therms (7 days X 100 Therms/day).
Now go to the 3 summer months, where the school is shut down or nearly so. Just looking at the summer bills will indicate that the "baseload", if any, is considerably less than 100 Therms/day. For sake of argument let's assume the 3 summer bills are zero.
So what's wrong? Only that we have failed to distinguish between the terms "regression equation" and "baseline model". In this case, the full baseline model is the regression equation PLUS a summer modification whose factors are approximately -3,000 Therms (100 Therms/day X the number of days in each of the 3 summer months).
Baseline Therms = c1 X #days + c2 * #HDD + Offset
Another way is to expand the Regression equation to two variables: HDD and school-in-session days. Remember that typically *no* months are excluded by the regression if you have more than one independent variable.
Baseline Therms = c1 X #days + c2 * #HDD + c3 * #SchoolDays
A third way is sometimes advocated by some purists who would like two separate models to be established for a case like that school: an "in-session" model for 9 months of the year and a "vacation" model for the other 3. I believe Metrix does, de-facto, the same thing, though we talk about only one, integrated model.
Otherwise feel free to contact our Tech Support staff at (805) 329-6565, or via email at email@example.com.