Anyone who has calibrated a building energy model for M&V or ASHRAE 14 knows that the real work begins after the first simulation run.
Calibration is rarely a one-and-done process. Instead, it is an iterative workflow where the model is run repeatedly while making small adjustments to assumptions, schedules, equipment efficiencies, or loads. After each run, the results are compared to measured utility data to determine whether the model is moving closer to reality.
In practice, this often means 20 to 30 simulation runs before the model begins to align with the measured data.
The frustrating part is not the modeling itself. It is the repetitive process of extracting results from the simulation output, organizing the monthly values, calculating the comparison metrics, and updating charts to see whether the model improved or got worse.
That repetitive process can consume a surprising amount of time during calibration.
The Iterative Nature of Calibration
A typical calibration cycle looks something like this:
- Run the energy model
- Extract monthly energy consumption from the output file
- Compare the model results to the measured utility bills
- Review the charts and calibration metrics
- Adjust model assumptions
- Run the model again
And repeat.
Each time the model is run, the modeler needs to quickly determine:
- Did the iteration improve the calibration?
- Did the model move closer to the utility data?
- Did a change make things worse?
