Probabilistic vibration serviceability assessment of stadium grandstands using season-scale monitoring data

Abstract

Crowd-induced vibrations in stadium grandstands remain a persistent serviceability concern, yet assessment practice is largely event-based and deterministic, limiting quantitative comparison across seating sections and seasons. This study presents a monitoring framework for stadium structures that integrates comfort standards with probability-of-exceedance (PoE) analysis and a proposed weighted Criticality Index (CRI) to rank sections by vibration demand. The framework was applied to a large-scale field campaign in which 116 accelerometers instrumented four seating sections during five football games in one season. Results quantitatively demonstrate substantial spatial and temporal variability in human–structure interaction. The motoring data from the most critical section of the stadium was modeled using extreme event probability as its response followed a heavy-tailed Generalized Extreme Value distribution, reflecting transient synchronized crowd excitation, while other sections were well described by lognormal behavior. CRI rankings were robust across weighting schemes and statistically separated the most demanding section from the others. Single-game assessments were shown to overestimate season-level exceedance probabilities by up to 77% compared to season aggregate performance, quantitatively demonstrating the necessity of multi-game monitoring to capture true serviceability demand. The proposed framework provides a repeatable, standards-compatible methodology for transforming large-scale stadium monitoring data into objective, risk-informed serviceability metrics, advancing the quantitative understanding of human–structure interaction in operational sports venues.