Structural Health Monitoring and Seismic Response Assessment of Civil Infrastructure using Target-Tracking Digital Image Correlation
AuthorNgeljaratan, Luna N.
AdvisorMoustafa, Mohamed A.
Civil and Environmental Engineering
AltmetricsView Usage Statistics
The deployment of wireless or non-contact sensing emerges as an innovativemonitoring alternative in structural health monitoring (SHM) which facilitates monitoringof remote areas such as under bridges crossing waterways, and most importantly, operatesat lower cost due to quicker and cheaper installation of the sensor and data acquisitionsystem. Several previous studies investigated the capabilities and limitations of wireless typesensors. However, there have been issues in data transmission, limited measurementdistance, or limited measurement points. To cope with the problems of existing wirelesssensor, vision-based measurement systems have been developed and they are currentlyemerging in the field of SHM. The measurements of vision-based system use images andtracks multiple targets motion between image sequences. Some relevant works concludedthat, vision-based using digital image correlation (DIC) technique is viewed as a platformin which mobile computing and wireless-communicating elements converges with themonitoring sensors. Previous work also showed the potential of DIC to lead to aneconomical and robust system for obtaining direct simultaneous measurements at severallocations of realistic infrastructure systems undergoing complex 2D and 3D deformations.Although major efforts have been dedicated in deploying target-tracking DIC inSHM, at least from the academic point of view, robust standard of practice or guidelinesfor monitoring civil structures and infrastructure systems is not yet available because ofseveral research gaps that require more research studies. One important gap is associatedwith the lack of comprehensive studies characterizing displacement measurementaccuracy, and more specifically for dynamic response monitoring, of target-tracking DICsince most of the classical work focused more on quantifying the error of strainmeasurements for stochastic pattern DIC. Moreover, although considerable theoreticalwork has been done in both image correlation and stereovision, only very few studiesconsidered large-scale 3D experimental validation and verification (V&V) viacomparison with mechanical sensors while focused on civil structures. Therefore, one ofthe major objectives of this dissertation is to fill the identified gap above through dedicatedV&V experimental testing with several large-scale applications of dynamic responsemonitoring under seismic loading. Quantifying displacement measurement errorsassociated with calibration or DIC post-processing, which was independent frommonitoring application or target measurements, was needed.Target-tracking tracking DIC is a new member in the wireless and non-contactsensor family, and just like the other wireless sensor types, target-tracking DIC faces thepossibility of losing data in its raw data signals. If other wireless sensor types may possiblybe losing the signal during the transmission, target-tracking DIC data loss is mainlybecause of overexposure and the motion blur. Another challenge of deploying target trackingDIC in SHM is the trade-off among field of view, sampling rates, recording timeand exposure time such that one setting dictates the other settings. Selecting a lower fpsrate for monitoring might cause a significant issue when the monitored structure vibratedat high frequencies; higher than half of the adopted fps rate. When higher frequencies ofinterest need to be captured, cameras with higher fps recording capabilities will be requiredor otherwise the measurement will be incorrect. Considering these challenges, anothermajor objective of this doctoral work is to explore the feasibility and implementcompressive sensing techniques for target-tracking DIC signal reconstruction, signalrecovery from data loss, and sampling rate improvement by implementing. To this end, arealistic SHM case study was employed where a pedestrian footbridge at the University ofNevada, Reno campus was monitored as it was excited by pedestrian loading to conductsystem identification. The results of system identification of the reconstructed, recoveredand improved signals were compared to accelerometers and original DIC along with errorestimations.The last part of the dissertation focused on several large-scale applications of civilinfrastructure monitoring using target-tracking DIC. In general, no comprehensive studieshave implemented target-tracking DIC for monitoring the dynamic and seismic responseof civil structures, e.g. bridges, under extreme events such as earthquakes and used it forpost-event condition assessment. Thus, this last part of the study focused on demonstratingthe validity of target-tracking DIC measurements in capturing seismic response as well asin identifying structural modal parameters through system identification of several bridgemodels tested at the Earthquake Engineering Laboratory at UNR. The novelty of this studywas in the application as no previous studies used full 3D target-tracking DIC for systemidentification and monitoring of full structural systems.