Condition monitoring for bridge cables using time-history area of cable forces under stochastic traffic flow

Damage or condition anomalies of stay cables will directly affect the safety of bridges in operation. As an important load type, the cyclic action of stochastic traffic flow (STF) is a main cause of cable performance degradation. However, due to its time-varying and complexity, there have been few studies on the anomaly diagnosis of stay cables under STF loading. Therefore, this paper proposes a novel condition monitoring method for stay cables based on the time-history area (THA) of cable forces under STF loading. First, mechanical analysis was performed to establish the theoretical relationship between THA and STF. The feature vector and matrix to characterize the normal condition of stay cables were developed based on the above relationship. Second, an online diagnosis model was established using the constant correlation between feature vector elements in the health state. On this basis, the Mahalanobis-squared distance (MSD) is employed to define an early warning index, which is used to inspect abnormal changes in cable performance. Third, according to numerical analysis, the rationality of the theoretical derivation was further confirmed and the feature vector variations in different simulated cases were presented. Finally, the effectiveness and practicability of the proposed method was verified on an actual cable-stayed bridge. The results demonstrate that the established online model has satisfactory diagnostic capabilities, which can accurately monitor the service condition of stay cables.