A novel method for evaluating stack pressure in real high-rise buildings: Optimization of measurement points

Abstract

Excessive stack pressure in a high-rise building always causes various problems, and the most direct way to understand the extent of the stack pressure across building components is to derive pressure distribution profiles for the entire buildings. As an alternative to measurement process, which is generally time-consuming and labor-intensive, an optimal field measurement method is proposed to investigate the real-time pressure distribution in operating high-rise buildings. The absolute pressure at unmeasured points can be accurately predicted from measured values at selected key points, both horizontally and vertically. Robust linear regression and the MARS model are applied to develop the related prediction models, and the number and positions of measurement points in the horizontal and vertical directions are optimized sequentially. Two high-rise buildings were selected to validate the applicability of the method, and five and six key pressure measurement points were finally determined, respectively. The measured and predicted absolute pressure values at unmeasured points were further compared, and the results showed a high goodness of fits as the R2 values were higher than 0.98 and the average MAE values were less than 2.0 Pa for all models. It is observed that the proposed method is suitable for engineering applications and demonstrates a practical method for investigating stack pressure distribution within high-rise buildings.