Content area
Full Text
ABSTRACT
This paper introduces and provides results from a rigorous, scientific testing methodology that allows pure building model calibration systems to be compared fairly to traditional output errors (e.g., howwell does simulation output matchutility bills?) as well as input-side errors (e.g., how well, variable-by-variable, did the calibration capture the true building's description?). This system is then used to generate data for a correlation study of output and input error measures that validates CV(RMSE) and NMBE metrics put forth by ASHRAE Guideline 14 and suggests possible alternatives.
INTRODUCTION
In previous work (New et al. 2012; Garrett et al. 2013; Garrett and New 2015), the Autotune calibration system was used to calibrate a building energy model (BEM) to a highly instrumented and automated ZEBRAlliance research home (Biswas et al. 2011) by fitting measured monthly load and electrical data. This research showed that the evolutionary computation approach to automatic calibration was effective in fitting EnergyPlus output from the calibrated model to the measured data. This calibration included time-varying parameters such as occupancy and equipment schedules necessary for practical application. There are other detailed studies which have compiled approaches to calibration (Reddy et al. 2006) and the performance of many calibration methods (Coakley et al. 2014).
However, because the tuning was applied to a real building with unknown model parameters (thus the need for calibration), it was impossible to determine exactly how well the tuned model matched the actual building in terms of model parameters over the course of a year. Even with costly laboratory-controlled research homes involving perfectly repeated automated occupancy, measurement of materials entering the building, and documentation of the construction process, it is still impractical to track the exact value of physical parameters for all materials throughout the building as they change with time.
Therefore, in this work, rather than attempting to calibrate existing buildings to match measured data, we instead attempt to calibrate fully specified U.S. Department of Energy (DOE) commercial reference buildings to match EnergyPlus (DOE 2012) output generated from altered versions of those buildings (where the altered model parameters are known to calibration test designers but unknown to calibrators) using the pure calibration technique described in BESTEST-EX (Judkoff et al. 2011). This allows one to test a calibration process's accuracy on both...