DAYLIGHTING PERFORMANCE SIMULATION: PREDICTION ACCURACY/PROCESSING SPEED TRADE-OFF

Abstract

In daylighting performance simulations of façade systems, a trade-off is often required between processing speed and prediction accuracy. This is particularly relevant at design onset, where plausible simulation outcomes are essential to drive decisions between several alternative façade configurations. To help address this trade-off, this paper presents a sensitivity analysis evaluating the influence of key input parameter settings, namely ambient bounces and grid size, on the convergence of performance outcomes and on simulation run times. The results provide statistical evidence that, although lower precision settings mostly accelerate calculations, they decrease the accuracy of prediction estimates, particularly for complex façades. Conversely, the relative increased accuracy resulting from higher precision simulations might reach a point where differences have a negligible practical impact. The paper concludes with a range of recommendations to support the early-stage selection of parameter settings and contributes to more robust simulation outcomes towards reducing the gap between simulated and measured data.