Passive Wireless Strain Sensing Based On UHF Radio Frequency Identification Technology And Variational Bayesian Data Modeling

Structural health monitoring is one of the important research directions in the field of civil engineering,which is very important to ensure the safe and efficient operation of engineering structures.All kinds of structural response data obtained through advanced sensing technology is an important basis for the study of structural health monitoring,among which strain response contains a wealth of structural state information and is a key index for structural state identification and performance evaluation.Most of the traditional strain sensing technologies are wired measurement methods,which have limitations such as lengthy wiring,complex device and high cost.However,the strain sensing method based on Radio frequency identification(RFID)technology,The wireless communication between measuring device and receiving device can be realized through the space coupling of high frequency electromagnetic wave,and the contactless information transmission can be realized.It has the characteristics of high identification accuracy,simple and easy operation,and is an important way to realize intelligent monitoring.Therefore,this thesis studies the structural passive wireless strain sensing method based on UHF RFID technology.Specifically,there are two ideas.The first idea proposes an integrated RFID strain sensor based on semi-passive tags,which converts the electric signal of Wheatstone bridge into radio frequency signal to realize wireless strain measurement.Another idea is a passive RFID microstrip antenna strain sensor based on passive tags,which uses the linear relationship between the resonant frequency of the antenna and the strain to realize passive wireless strain measurement.The actual performance of two RFID strain sensors is tested and analyzed on various materials and structures.Further,based on the experimental measurement data of RFID strain sensor and the field measured response data of large-scale engineering,the modeling research of structural health monitoring data was carried out through the variational Bayesian inference method,and the accurate estimation of structural response and structural status recognition results were obtained by combining linear regression and logistic regression algorithms.The specific research content of this thesis is as follows:(1)Based on semi-passive RFID tags,an integrated RFID strain sensor is designed.Wheatstone Bridge is responsible for sensing signal generation,and the tag chip is responsible for analog to digital conversion and wireless data communication.In order to solve the problem of low accuracy of sensor acquisition,this study improved the accuracy of strain sensing by adding a signal amplification device in front of label signal reading.In addition,the sensor data acquisition system is improved so that the reader can automatically record and store data.The designed integrated RFID strain sensor has high accuracy and strong adaptability.Its sensing performance is tested in various experimental scenarios such as aluminum alloy beam,reinforced concrete beam,concrete structure column.(2)Based on passive RFID microstrip antenna tags,a completely passive wireless RFID strain sensor is independently designed.The linear relationship between the resonant frequency of the antenna and its strain is used to achieve strain characterization.The acquisition of resonant frequency is completed by reading the threshold power of the label by the reader.In this thesis,the microstrip antenna with short circuit truncated structure is designed and optimized by HFSS software,so that it can achieve impedance matching with the chip.The miniaturization design of microstrip antenna is realized by the form of edge slot.The mechanical deformation of the antenna was simulated in COMSOL multi-physical field simulation software and the resonant frequency was scanned.The linear relationship between the resonant frequency and strain was verified.The designed microstrip antenna was fabricated and the RF chip was welded.The strain sensing properties of the microstrip antenna were tested on steel plate,aluminum alloy beam and concrete beam.(3)Based on the measured structural response data of RFID strain sensor and the structural health monitoring data of large-scale actual engineering site,the structural response data modeling and analysis were carried out through the variational Bayesian inference method.The Bayesian linear regression algorithm was used to regression model the measured strain data obtained,and the optimal data model and accurate regression results were obtained.The strain response forecasting was carried out based on the Bayesian linear regression model.In addition,data classification based on variational Bayesian logistic regression was carried out.For monitoring data under different structural states and external incentives,training learning and feature extraction were carried out through variational Bayesian logistic regression algorithm to obtain high-precision structural response classification and damage state recognition results.The research results can provide scientific support for safe operation and timely maintenance of structure.