Abstract

ABSTRACT The authors have studied the wind-shielding effect on circular elevated tanks with computational fluid dynamics (CFD) analyses, caused by rectangular plan-shaped buildings, which alter wind load significantly. Velocity- streamlines have been observed to understand the effect of wind flow on the pressure due to building interference. The present CFD model has been validated with experimental literature. Supervised learning is done for different percentages of CFD data (33%, 55%, 66% and 88%) using various artificial neural network (ANN) architectures, among which the performances of three combinations (one, two and three hidden layers) are provided. Input variables are spacing between structures, interference height, wind angle and peripheral angle. Output parameters are time-average and fluctuating coefficients of wind pressure, distributed circumferentially around tankwall. Data ≥55% are sufficient for precise ANN training. The proposed ANN model efficiently predicts mean and fluctuating circumferential wind pressure coefficients on tank-wall for building-interference, in very little time, with 98.15% and 98.43% accuracy, respectively, benefitting structural designers.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call