铁路站房结构损伤与传感器故障识别研究

发布时间：2018-07-08 15:41

本文选题：铁路站房 + 传感器故障　；参考：《铁道工程学报》2017年05期

【摘要】：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。
[Abstract]:Research purposes: the design of the traditional station structure health monitoring system assumes that sensors can work normally, but sensor failures often occur. The monitoring system will change the monitoring data caused by sensor failures as structural damage. Therefore, the sensor failure should be considered in the design of the monitoring system in order to reduce the monitoring system. Research conclusions: (1) sensor failure causes the dimension or data size of sampling data to be reduced, sensor failures lead to the shift of the mean value of the sampling, and monitoring will increase the runaway running length of the system according to the generated control limits; (2) the output data of the sensor is constructed by cointegration relationship in the training data processing. The normal range does not require a lot of test data, it can also eliminate the influence of temperature, environment and load on the identification of structural damage. (3) a two layer data processing model is proposed, and the non model based sensor fault diagnosis is placed in the data preprocessing module, and the structural damage identification is placed in the fault diagnosis module, which is more convenient in the system implementation; (4) The research results can provide reference for reducing false alarm rate of railway station structure health monitoring.
【作者单位】：石家庄铁道大学;
【分类号】：TU317

【参考文献】