The 1995 validation under measured sky luminance patterns is generally considered a benchmark, at least for ordinary glazing. It is described in exhaustive (exhausting?) detail here:
https://repository.lboro.ac.uk/articles/Daylight_simulation_validation_sky_models_and_daylight_coefficients/9460817
Related papers include:
J. Mardaljevic. Validation of a lighting simulation program under real sky conditions. Lighting Research and Technology, 27(4):181–188, 12 1995.
J. Mardaljevic. The BRE-IDMP dataset: a new benchmark for the validation of illuminance prediction techniques. Lighting Research and Technology, 33(2):117–134, 2001.
This paper describes the confounding factors that often come into play when attempting ‘real world’ validation:
J. Mardaljevic. Verification of program accuracy for illuminance modelling: Assumptions, methodology and an examination of conflicting findings. Lighting Research and Technology, 36(3):217–239, 2004.
And, of course, there have been other validation studies, e.g:
C. F. Reinhart and M. Andersen. Development and validation of a Radiance model for a translucent panel. Energy and Buildings, 38(7):890 – 904, 2006.
I don’t hold much store by the CIE test cases since they are so idealised. The reason Radiance performed poorly in one of them is because the scenario was effectively a highly reflective integrating sphere. Radiance could be tweaked to solve for this, but it would require a recompile with the AVGREFL macro redefined to, say, 0.95 IIRC (not something you’d want for any real-world application). That is in addition to the errors spotted by Ian Ashdown.
Cheers
John
John Mardaljevic PhD FSLL FIBPSA
Professor of Building Daylight Modelling
School of Architecture, Building & Civil Engineering
Loughborough University
Loughborough
Leicestershire
LE11 3TU, UK
Tel: +44 1509 222630 (Direct)
Aperture-Based Daylight Modelling: a new direction for daylight planning
[email protected]
Personal daylighting website:
http://climate-based-daylighting.com
