Integrity monitoring of vertical and horizontal structures via visualization and statistical inspection of a mesh sensor network

Date of Publication


Document Type

Master's Thesis

Degree Name

Master of Science in Computer Science

Subject Categories

Computer Sciences


College of Computer Studies


Computer Science

Thesis Advisor

Arnulfo P. Azcarraga

Defense Panel Chair

Marnel S. Peradilla

Defense Panel Member

Marnel S. Peradilla

Arnulfo P. Azcarraga

Macario O. Cordel, II


Studies have been conducted to understand how mild but frequent tremors affect the integrity of built structures such as of buildings and bridges. These have been done through the analysis of data gathered using tremor sensors that are either attached on or embedded within the structures. However, current tools used to analyze the sensor data usually require in-depth knowledge of seismic data and expertise in structural engineering to be able to interpret the generated seismic graphs. Not only are the data difficult to interpret, but systems being used to gather and process these data tend to be very expensive and rely only on prevailing seismic activities. This study presents a novel approach in monitoring and analyzing the structural integrity of buildings located in earthquake-prone areas through the use of a mesh of tremor sensors. Data collected from these numerous sensors are analyzed so that specific areas of built structures that may have some structural defect may be identified. These identified probable anomalies in the structure may then be the subject of a more thorough investigation. Anomaly detection in the structures is done through the use of unsupervised machine learning techniques that yield the expected movement readings of each area in the building, given the movement in the surrounding areas. Aside from making it easy to visually inspect the movement of the built structure, this study also proposes the use of statistical tests, specifically the chi-statistic that is fed to the Kruskal-Wallis test, in order to isolate the specific locations in a building where a structural anomaly might be found. Experiments were conducted using actual 2013 earthquake data from Bohol, Philippines, and applied on hypothetical healthy and damaged buildings that were constructed using the ETABS simulation software.

Abstract Format






Physical Description

221 leaves


Sensor networks; Structural health monitoring; Visualization

Upload Full Text


Embargo Period


This document is currently not available here.