Anomaly Detection Algorithm for Stay Cable MonitoringData Based on Data Fusion

In order to improve the accuracy and consistency of data in health monitoring system, an anomalydetection algorithm for stay cables based on data fusion is proposed. The monitoring data of Nanjing No.3Yangtze River Bridge is used as the basis of study. Firstly, an adaptive processing framework with feedbackcontrol is established based on the concept of data fusion. The data processing contains four steps ： dataspecification, data cleaning, data conversion and data fusion. Data processing information offers feedback to theoriginal data system, which further gives guidance for the sensor maintenance or replacement. Subsequently, thealgorithm steps based on the continuous data distortion is investigated,which integrates the inspection data andthe distribution test method. Finally, a group of cable force data is utilized as an example to verify theestablished framework and algorithm. Experimental results show that the proposed algorithm can achieve highdetection accuracy, providing a valuable reference for other monitoring data processing.