Improving the accuracy of circadian lighting simulation with field measurement

Abstract

For improving the simulation accuracy of circadian lighting, a workflow was established to first calibrate the simulation model under electric lighting followed by validation with integrated daylighting and electric lighting. A case study was conducted in a daylit office, with the evident impact of varying lighting parameters (light level, intensity, SPD, reflectance) on the simulation accuracy. In the fully calibrated model, the simulated and field-measured SPDs are highly correlated (rSPD ≥ 0.9), providing the acceptable accuracy of the simulated melanopic/photopic(M/P) ratio with the error rate (MBErel) less than 4% and residual deviation (RMSErel) within 7%. The MBErel and RMSErel of corneal illuminance (EV), EML, and CS are within 10% and 20%, respectively. The ratio of data points with the residual rate (e) less than 20% is over 75% in the calibration and validation phases. Conclusively, the proposed procedure could help prepare for further analysis of circadian light exposure.