Validating a Computer Program
Using Physical Models - Abstract

Ibrahim Al-Turki
P.O.Box 59035, Riyadh 11525
Saudi Arabia

and

Marc Schiler, Assoc.Prof.
School of Architecture
University of Southern California
Los Angeles, CA 90089-0291

INTRODUCTION:

This paper is to validate a computer program called "LIGHTSUM" which calculates the available illuminance inside atria at the ground level using IES, Gillette, and Dogniaux sky models. In addition, it creates form factors for reflective surfaces.

DESCRIPTION OF METHOD:

A physical model of a scale of 1'=1" was used to validate the LIGHTSUM computer program. An atrium of 20' width, 20' length, 40' height, with walls and ground of 30% reflectance is used. The physical model was tested under exclusively clear sky conditions every hour from 9:00 a.m. to 3:00 p.m. on the 21 st of each month. Licors LI-210-S illuminance meters were cconnectedto a portable computer to read the illumination level from each licor. Seven licors were located inside the atrium, and one was located outside the atrium to measure the Total Horizontal Illumination.

To validate the LIGHTSUM computer program, all the input data had to be analogous to and under the same conditions that the physical model was tested. Therefore, a TMY (Typical Meteorological Year) weather file was written for the entire year to read only clear sky condition data.

ANALYSIS & CONCLUSIONS:

The data collected from LIGHTSUM for both IES and GILLETTE recommended models were calculated in Kilo-lux; so, to compare the output data to it in the physical model, all the data were converted to footcandles and then to DF to facilitate the comparison. The illumination level was measured in the atrium, and calculated in LIGHTSUM in several chosen reference points.

The physical model does not agree with the computer program in three major facts:

  1. The daylight factor calculated for each reference point seems significantly higher than the DF measured from the physical model.

  2. LIGHTSUM does the inverse of what is shown on the graphs from the physical model, high daylight factor in winter and low daylight factor in summer and spring for both IES and Gillette models.

  3. LIGHTSUM shows very low exterior horizontal illumination in winter and that causes the high daylight factor. However, the daylight factor in summer and spring is twice as high as in the physical model.

Therefore, the physical model tests do not validate the computer program.

Back to Papers Published