Quality control in daylight simulation

Abstract

From my own experience there is always some degree of uncertainty in trusting the daylighting information resulting from a computer simulation. There is no current method to enable me to have complete confidence in simulation output other than relying on my current daylighting knowledge and experience. Added to this, daylight simulation programs are becoming increasingly easier to use with more user-friendly graphical user interfaces being implemented into new software. No longer does an inexperienced user need to possess adequate photometric knowledge or knowledge of simulation principals to create a simulation model and generate output. The standard argument about software used for any building simulation is that it will have difficulty predicting exactly what a building will do but that the difference between the predicted performance for two different cases of the same building can be relied upon. What this research is trying to achieve is a means of assuring people that a building performance derived from software produces reliable predictions of the performance differences between different cases. More specifically, this research investigates whether it is possible to develop a series of tests to perform to a daylight simulation which would enable daylight simulation users to have more confidence that what is calculated by simulation is indeed what would be expected in reality.