铁路站房结构损伤与传感器故障识别研究

Research on the Identification between Structure Damage and Sensor Fault in Structural Monitoring of the Railway Station

研究目的:传统站房结构健康监测系统的设计假设传感器都能够正常工作,然而传感器故障却经常发生,监测系统会将由传感器故障导致的监测数据变化认为是结构损伤,因此需要在监测系统的设计中考虑传感器故障,以便降低监测系统的误报率。研究结论:(1)传感器失效使采样数据的维度或者数据规模降低,传感器故障导致采样的均值发生偏移,依据已生成的控制界限进行监测将使系统失控运行长度增加;(2)在训练数据处理中,通过协整关系构建传感器输出数据的正常范围,不需要大量测试数据,还可以消除温度、环境和载荷等对结构损伤识别的影响;(3)提出两层数据处理模式,将非基于模型的传感器故障诊断放在数据预处理模块,将结构损伤识别放在故障诊断模块,在系统实现上比较方便;(4)该研究结果可为降低铁路站房结构健康监测的误报率提供参考。

Research purposes： Structural health monitoring system usually assume that the sensor can work normally, but sensor malfunctions often happen. Monitoring system will consider the data change caused by sensor malfunctions as structural damage, therefore, it is necessary to take into account in sensor malfunctions in order to reduce false alarm rate. Research conclusions ： （ 1 ） Sensor failure makes the dimension of the sampling data or data size reduce, the sensor fault cause a drift of the sample average, that cause the run length out of control in system increased according to the generated control limit. （2） In processing of the training data, the normal range of sensor output data can be got by building cointegration relationship, eliminate the influence of temperature, environment and load effect on identification of structure damage, and need not a large number of test data. （3） Sensor fault diagnosis strategy and double layer data processing model are put forward based on the analysis. Put the sensor fault diagnosis in data preprocessing module, and put structural damage identification in the diagnosis module, which is more convenient to system development. （4） The research results can provide reference for reducing the rate of false alarm of structural monitoring of the railway station.