论文部分内容阅读
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: The design of traditional station building health monitoring system is based on the assumption that sensors can work normally, however, sensor failures often occur. The monitoring system considers the monitoring data changes caused by sensor failures as structural damage and therefore needs to be monitored in the design of the monitoring system Consider sensor failure to reduce the false alarm rate of the monitoring system. The conclusions are as follows: (1) The sensor failure causes the dimension or data size of the sampled data to decrease, and the average value of the sample sensor offset leads to an overshoot. Monitoring based on the generated control limits will increase the runaway length of the system. (2) In the process, the normal range of the sensor output data is constructed through the cointegration relationship without the need of a large amount of test data, and the influence of temperature, environment and load on the structure damage identification can also be eliminated. (3) A two-layer data processing mode is proposed, The sensor fault diagnosis of the model is put in the data preprocessing module, and the structural damage identification is put in the fault diagnosis module, which is more convenient in the system implementation. (4) The results of this study can provide a reference for reducing the false alarm rate of the railway station structure health monitoring .