AI Article Synopsis

  • The paper introduces a non-destructive testing system using tactile transducers and accelerometers to assess the condition of foundation piles and utility poles without physical contact.
  • Unlike traditional methods that use impact hammers, this system generates controlled narrow-band frequency stress waves, allowing for targeted testing.
  • Experimental validation on damaged timber and concrete poles showed that using advanced signal processing and support vector machines can classify damage types with an accuracy of 92.5% ± 7.5%.

Article Abstract

This paper presents a novel non-destructive testing and health monitoring system using a network of tactile transducers and accelerometers for the condition assessment and damage classification of foundation piles and utility poles. While in traditional pile integrity testing an impact hammer with broadband frequency excitation is typically used, the proposed testing system utilizes an innovative excitation system based on a network of tactile transducers to induce controlled narrow-band frequency stress waves. Thereby, the simultaneous excitation of multiple stress wave types and modes is avoided (or at least reduced), and targeted wave forms can be generated. The new testing system enables the testing and monitoring of foundation piles and utility poles where the top is inaccessible, making the new testing system suitable, for example, for the condition assessment of pile structures with obstructed heads and of poles with live wires. For system validation, the new system was experimentally tested on nine timber and concrete poles that were inflicted with several types of damage. The tactile transducers were excited with continuous sine wave signals of 1 kHz frequency. Support vector machines were employed together with advanced signal processing algorithms to distinguish recorded stress wave signals from pole structures with different types of damage. The results show that using fast Fourier transform signals, combined with principal component analysis as the input feature vector for support vector machine (SVM) classifiers with different kernel functions, can achieve damage classification with accuracies of 92.5% ± 7.5%.

Download full-text PDF

Source
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC5751589PMC
http://dx.doi.org/10.3390/s17122938DOI Listing

Publication Analysis

Top Keywords

tactile transducers
16
condition assessment
12
foundation piles
12
piles utility
12
utility poles
12
network tactile
12
support vector
12
testing system
12
vector machines
8
damage classification
8

Similar Publications

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!