University of Twente Student Theses

Login

Smartphone-Based Vibration Analysis for Bridge Health Monitoring

Avnon, R. (2022) Smartphone-Based Vibration Analysis for Bridge Health Monitoring.

[img] PDF
2MB
Abstract:The relation between bridge dynamic responses to load is the essence of vibration analysis. Bridge dynamic properties could serve as a reference for its “healthy” condition. An observed change in bridge response could be a sign of degradation or even damage. Currently, researchers usually use fixed sensor networks for monitoring purposes. These sensor networks consist of expensive sensors which require complex installation and high maintenance requirements. The management and aggregation of data from such networks is a challenging task. Alternatively, smartphones, which are equipped with networks, sensors, and storage capacity, might be useful for vibration testing, reducing cost and the need for expert installation. The outspread of smartphones has shaped the opportunity to utilize crowdsourcing for structural monitoring purposes. Crowdsourcing networks are widely desirable for their potential to generate Bigdata, cost-effectively and collectively. A network in which smartphones serve as mobile sensors, sharing structural vibration and location data to a cloud server. This BSc thesis presents innovative applications of smartphones to measure dynamic bridge responses and practical information for SHM. Three questions were investigated: (ⅰ) methods for extracting natural frequencies from smartphone measurements, (ⅱ) achievable frequency accuracy, and (ⅲ) ways for using gyroscope and GPS data for bridge monitoring. The smartphone measures acceleration, gyroscope, and GPS data while carried in the pocket, demonstrating crowdsourcing. The bridge is excited by walking over it and performing heel drops on its mid-span. One method enables the extraction of the bridge's eigenfrequencies from the walking measurements. It involves splitting the acceleration data into measurements on the bridge, and reference walking before and after the bridge, computing the Fast Fourier Transformation (FFT) for each data set individually, subtracting the outputs, and manually inspecting the resulting Power Spectral Density (PSD) difference graphs. The smartphone’s gyroscope captures the walking frequency and gait cycles. Dominant frequencies computed with gyroscope data could be eliminated as bridge response candidates. GPS measurements were insufficiently precise to detect exact location and assist in pairing frequencies with their mode shapes. Natural frequencies estimated from smartphone measurements appear to be comparable to vision-based systems frequency measurements. However, smartphone measurements seem to assist in computing more modal frequencies than cameras by providing 3D spatial data. The vibration analysis suggests that smartphones can measure bridge dynamic responses while placed inside a pocket. Natural frequencies can be computed by removing the walking frequency and adjusting the power spectral scale. Averaging small data sets from the crowd could serve as a database to monitor our infrastructure without installing and maintaining complex fixed sensor networks.
Item Type:Essay (Bachelor)
Faculty:ET: Engineering Technology
Programme:Civil Engineering BSc (56952)
Link to this item:https://purl.utwente.nl/essays/93552
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page