Automatic anomaly detection and processing for long-term tension monitoring of stay cables based on vibration measurements

Abstract

Stay cables are the crucial force-transmitting components of cable-stayed bridges. Consequently, effective cable tension monitoring is typically regarded as the most important issue in structural health monitoring for this type of bridges. In long-term monitoring of structures, the occurrence of abnormal signals due to different reasons such as sensor faults, system malfunction, data loss in transmission, or instability of power supply, is almost unavoidable and may strongly deteriorate the quality of followed analyses. The current paper presents the first phase of a project to develop an automatic monitoring system of cable tension based on ambient vibration signals for Mao-Luo-Hsi Bridge. The results demonstrate that automatic anomaly detection and processing for cable vibration measurements can be conveniently achieved with a relatively simple algorithm merely using an appropriate threshold of maximum response value. In addition, elimination of anomaly in the testing stage of installing the monitoring system is also reported.