Date of Publication

3-2022

Document Type

Master's Thesis

Degree Name

Master of Science in Computer Science

Subject Categories

Computer Sciences | OS and Networks

College

College of Computer Studies

Department/Unit

Computer Science

Thesis Advisor

Clement Y. Ong

Defense Panel Chair

Marnel S. Peradilla

Defense Panel Member

Gregory G. Cu
Clement Y. Ong

Abstract/Summary

Civil structures such as buildings, bridges and dams are subject to degradation over time due to environmental stresses, both natural and man-made. Of these, vibrations (i.e. earthquakes, wind, road traffic) are of particular significance. The structures are designed and built to bend and twist in response to the everyday stress and use. By monitoring these movements, it is possible to ascertain their integrity. A Structural Health Monitoring (SHM) system can be deployed to these civil structures to sense and record vibrations around different points of the structure over time. However, when deploying SHM systems to large-scale structures, the sensor data needs to travel over long distances to reach the gateway. Moreover, if cables are used to connect the network of sensors, it brings with it the problem of having to deal with the cost of the cables and having a limited size and span of the network. Therefore, by deploying an SHM system as a wireless mesh, the aforementioned cost and scalability issues can be addressed. This gives engineers the freedom to deploy as many sensor nodes around the building, which will allow for more data points that can be used for the structural health analysis. Therefore, this research aims to develop a mesh of low-cost wireless sensors for vibration-based structural health monitoring. The WSN, with a minimum of 20 nodes, will be deployed to a civil structure, and should be designed such the data being collected can be used for the identification of the modal parameters. Time synchronization and power management protocols will also be developed to ensure that no two nodes in the network have clock offsets of more than one millisecond and that each node can last on battery power for more than one week. The resulting prototype network consists of electronic components that are carefully chosen to ensure a balance between performance capabilities and power consumption. These components were first individually characterized to determine its performance characterization limitations across different modes. From these tests, it was discovered how their embedded features and functionalities should prioritizing for use when designing and developing the network packets, exchanges, and protocols. Overall, the nodes of the network are able to gather acceleration data and send them over to the cluster head for data analysis. The time synchronization algorithm is able to keep the local clocks synchronized to the reference with a clock offset between -0.25ms to 0.40ms. This performance is significantly effective because after just an hour without it, the nodes clearly show that they do not sample and timestamp accelerometer data synchronously.

Abstract Format

html

Language

English

Format

Electronic

Physical Description

132 leaves

Keywords

Wireless sensor networks; Structural health monitoring

Upload Full Text

wf_yes

Embargo Period

3-6-2022

Download

Share

COinS