Probabilistic Serviceability Assessment for Pedestrian Walkways: A Case Study on a 52-m Span Footbridge

The increasing adoption of lightweight, long-span pedestrian walkways has heightened concerns regarding vibration serviceability due to reduced natural frequencies and damping ratios. Traditional evaluation methods, such as root mean square (RMS) acceleration and maximum transient vibration value (MTVV), are widely used in design standards. However, signifi cant discrepancies exist between threshold values specifi ed in guidelines and real-world vibration experiences, leading to inconsistencies in serviceability assessments. This study proposes a probabilistic methodology that evaluates walkway vibrations based on in-situ measurements rather than predefi ned standard limits. By examining resonance-induced pedestrian discomfort, site-specifi c discomfort thresholds are established. A weighted MTVV (WMTVV) approach is introduced, integrating probabilistic modeling to enhance accuracy in real-world applications. To validate this framework, an experimental study was conducted on a 52-m span footbridge, incorporating long-term ambient vibration monitoring and controlled resonance experiments. The results reveal that existing vibration assessment methods often yield subjective and overly conservative or inadequate criteria. The study highlights the necessity of data-driven, probabilistic methodologies tailored to structure-specifi c conditions, thereby improving accuracy, reliability, and practical applicability in pedestrian walkway vibration evaluation.