Automated stochastic noise mapping for construction noise prediction and risk assessment

Construction noise is a dynamic occupational and environmental hazard that is difficult to assess using point-based monitoring alone. This study develops an automated stochastic noise mapping framework that predicts spatial noise distributions and delineates noise risk zones for construction sites. The framework digitises stochastic prediction-chart data for moving equipment and uses site-specific geometry, machine sound power levels, operating sub-areas, background noise and barrier effects. Automated algorithms generate two-dimensional L Aeq contour maps and classify areas into noise risk zones. The framework was validated using field measurements from four construction activities, such as excavation, slope cutting and earth moving, foundation piling and road excavation. Across 16 control points, predicted and measured L Aeq values showed good agreement, with an average absolute difference of 1.7 dBA, a correlation coefficient of 0.9580 and r 2 = 0.9177. The maps identified higher noise zones around excavators, trucks and piling equipment, supporting hearing-protection enforcement, equipment-layout planning and temporary barrier decisions. The noise risk zones provide activity-specific spatial screening information for construction noise management. The proposed framework converts stochastic noise prediction into a practical decision-support tool for proactive construction-noise risk management.