基于新奇检测技术的斜拉索状态评估

Condition Assessment for Stay Cables of Cable-stayed Bridge Based on Novelty Detection Technique

探讨了应用基于BP神经网络的新奇检测技术进行斜拉索状态评估的方法。通过对监测系统采集数据分析处理,生成训练神经网络需要的样本数据,按要求训练网络,建立了基于新奇检测技术的多阶段状态评估的神经网络模型,实现了斜拉索状态评估的两个阶段:状态预警、状态异常位置识别。状态异常位置识别采用逐步分区识别的方法,最终将损伤拉索的位置确定在较小的范围内。用有限元模型和实测数据进行了检验,结果表明,在不同的环境温度条件下,该方法能准确进行状态预警,有效地识别出状态异常的位置。

A method of condition assessment for stay cables by novelty detection technique based on BP neural network was discussed.The sample data for training neural network were generated by analyzing and processing the data collected by monitoring system.Trained networks as required,a multi-stage neural network model for condition assessment based on novelty detection technique was established.Thus the condition assessment was divided into two stages： condition alarming and locating of abnormal condition.The location of abnormal condition can be identified by dividing zones step by step,and the damaged cable can be identified in a small zone at last.It was tested by finite element model and field measurement.The results show that the method can accurately alarm and effectively identify the location of damaged cable at different environmental temperatures.