Appendix B 329
from expectations. Snap-shot measurements will significantly improve
simulation accuracy. Where resources are limited, on/off tests can be
used to determine snap-shot end-use measurements of lighting, recep-
tacle plug loads and motor control centers. These tests can be performed
over a weekend using a data logger or EMCS to record whole-building
electricity use, usually at one-minute intervals, and in some instances
with inexpensive portable loggers that are synchronized to a common
time stamp (Benton et al. 1996, Houeek et al. 1993, Soebarto 1996).
Following collection of as much calibration data as possible, the
steps in calibrating the simulation are as shown below.
1 Assume other input parameters and document them.
2 Verify that the simulation predicts reasonable operating results
such as space or process temperature/ humidity.
3 Compare simulated energy and demand results with metered data,
on an hourly or monthly basis. Use actual weather data when
conditions vary significantly from average year weather data. As-
sess patterns in the differences between simulation and calibration
data. Bar charts, monthly percent difference time-series graphs and
monthly x-y scatter plots give visual presentations which aid the
identification of error patterns.
4 Revise assumed input data in step 1 and repeat steps 2 and 3 to
bring predicted results reasonably close to actual energy use and
demand. More actual operating data from the facility may also be
needed to improve the calibration.
Buildings types which may not be easily simulated include those
with:
- large atriums,
- a significant fraction of the space underground or ground coupled,
- unusual exterior shapes,
- complex shading configurations,
- a large number of distinct zones of temperature control.
Some building ECMs cannot be simulated without great difficulty,
such as: