TRB Pubsindex
Text Size:

Title:

Noncontact Ultrasonic Guided-Wave Detection of Rail Defects
Cover of Noncontact Ultrasonic Guided-Wave Detection of Rail Defects

Accession Number:

01206828

Record Type:

Component

Availability:

Find a library where document is available


Order URL: http://worldcat.org/issn/00978515

Abstract:

In an effort to overcome some of the issues associated with ultrasonic wheel inspections, some research groups, including the University of California, San Diego (UCSD), are investigating ultrasonic guided waves for rail inspections. There are several reasons for this. First, guided waves propagate along, rather than across the rail, and are thus ideal for detecting the critical transverse defects (TDs). Second, guided waves increase the inspection coverage (length of rail inspected at once) thereby relaxing limits on the achievable speed. Third, since guided waves penetrate a finite depth of the rail surface, they can travel underneath shelling and still interact with internal defects allowing for their detection. Fourth, since guided waves travel in the mid-frequency range, between 20 kHz and 1 MHz, they can penetrate alumino-thermic welds, hence potentially targeting weld cracks/discontinuities. The advantages of guided waves come with the difficulty in managing their complicated propagation behavior. UCSD has utilized a semianalytical finite element (SAFE) method that combines theoretical and numerical formulations to allow the study of high-frequency guided waves in rails in a computationally efficient manner. These studies have allowed an optimized design of the rail defect detection prototype. The UCSD/FRA prototype uses non-contact means of generating and detecting the guided waves in the rail. The solution of choice is a combination of laser and air-coupled sensors, which were first proposed in 2000 for noncontact rail probing. The prototype also uses a statistical pattern recognition algorithm for detecting and classifying defects. The uniqueness of this algorithm is that it does not require previous knowledge of defects (i.e. it does not require a “training” phase as in neural network-based classificators), and it is hence very practical. This algorithm was successfully tested in the field with excellent results. This paper describes first SAFE models of guided waves in rails. It then discusses the status of the UCSD/FRA rail defect detection prototype including the results of the last field tests performed in March 2008.

Monograph Accession #:

01206823

Language:

English

Authors:

Coccia, Stefano
Bartoli, Ivan
Salamone, Salvatore
Phillips, Robert
Lanza di Scalea, Francesco
Fateh, Mahmood
Carr, Gary

Pagination:

pp 285-296

Publication Date:

2010-7

Serial:

Transportation Research Circular

Issue Number: E-C145
Publisher: Transportation Research Board
ISSN: 0097-8515

Conference:

Joint International Light Rail Conference: Growth and Renewal

Location: Los Angeles California, United States
Date: 2009-4-19 to 2009-4-21
Sponsors: Transportation Research Board; American Public Transportation Association

Media Type:

Web

Features:

Figures (5) ; Photos (3) ; References (17) ; Tables (1)

Uncontrolled Terms:

Subject Areas:

Maintenance and Preservation; Public Transportation; Railroads; Terminals and Facilities; I22: Design of Pavements, Railways and Guideways

Files:

TRIS, TRB

Created Date:

Oct 7 2010 3:40PM

More Articles from this Serial Issue: