Fatigue Assessment of Suspension Bridges under Non-Stationary Buffeting Using TSD-RFC Method

Abstract

Suspension bridges experience significant stress cycles from wind loads, leading to fatigue damage. Traditional rainflow counting methods perform poorly under non-stationary wind loads, especially when stress responses exhibit abrupt state changes, often causing incomplete cycle identification or amplitude misjudgment. This paper proposes a Time- Domain-State Dual-Layer Rainflow Counting (TSD-RFC) method. The method decomposes the non-stationary response into multiple quasi-stationary segments, applying classical three-point rainflow counting within each segment to identify primary fatigue cycles. Additionally, for state transition segments, an independent cycle counting layer captures large-amplitude stress cycles induced by wind speed fluctuations. Compared with traditional rainflow counting, TSD-RFC identifies stress cycles more accurately, particularly in high-amplitude regions, and effectively distinguishes cycles caused by state transitions. In fatigue assessment at a wind speed of 20 m/s, TSD-RFC yields approximately 11.39% higher fatigue damage than traditional methods, demonstrating its clear advantage under non-stationary wind loads.