Session: 01-16-02: Congress-Wide Symposium on NDE & SHM: Ultrasonic Waves for Material Characterization and Damage Assessment

Paper Number: 113918

113918 - Improved Non-Contact Ultrasonic High-Speed Structural Condition Monitoring of Rails Using a Controlled Acoustic Source and Random Wheel Generated Excitations 

Internal defects in rails are a major contributing factor in train derailment related accidents. These defects, if left unattended, can ultimately lead to crack propagation and broken rails. Continuous monitoring of such defects is, therefore, of particular interest to the railroad maintenance community. Existing techniques for rail inspections require specialized test cars that operate at slow speeds, requiring careful scheduling and traffic disruptions. A non-contact ultrasonic passive rail defect detection system is currently being developed at the Experimental Mechanics, NDE, and SHM Laboratory at the University of California San Diego. This technique has the potential to operate at high speeds and can be mounted on a regular train, thereby facilitating continuous condition monitoring of rails. Previous field tests at speeds of up to 80 mph have shown the potential of the technique in detecting internal rail defects and discontinuities (joints and welds). This paper presents an improved version of the ultrasonic non-contact passive rail defect detection system with the introduction of a non-contact active acoustic source. The existing version of the system utilizes air-coupled capacitive ultrasonic transducers in receiving-only mode to passively extract a defect-sensitive impulse response function of a rail segment using rail-wheel interaction as the random source of excitation. The presence of internal defects (such as transverse defects) or discontinuities alters the characteristics of the reconstructed impulse response function. Defect-sensitive features from the impulse response function are extracted, and a statistical multivariate outlier analysis is performed to calculate a metric representative of damage. Tests from previous generations of the system highlighted the importance of rail-wheel interactions in the defect detection performance; stronger excitation of the rails resulted in more reliable impulse response functions and better defect detection performance. Therefore, the use of an active and controlled non-contact acoustic source is proposed for the improved system that aims to enhance the signal-to-noise ratio at the transducer array independent of testing condition variabilities. Testing included various types of non-contact sources, source operating parameters, and signal processing routines to optimize the improved system. Miniature electrostatic transducers were investigated as a possible alternative to the capacitive transducers to potentially reduce the size of the prototype. Tests were performed at the Transportation Technology Center (TTCI) with the improved system at speeds of 25 mph, 33 mph, and 40 mph. Performance was evaluated through receiver operating characteristic curves by computing the rate of false alarms compared to true detections. Preliminary results indicate that defect detection performance improves when a controlled acoustic source is used.

Presenting Author: Diptojit Datta University of California San Diego

Presenting Author Biography: Dr. Datta is currently working as an Engineering Consulting Associate in the Mechanical Engineering Practice at Exponent Inc. in the Denver office. He has particular expertise in developing, testing, and evaluating structural health monitoring and non-destructive evaluation techniques for critical structural components in the civil, railroad and aerospace industries.

During his PhD at UC San Diego, Dr. Datta developed high-speed non-contact ultrasonic techniques for detecting internal defects in rails and degrading ballast support conditions in railroad ties. His PhD culminated with the development of two working prototypes with real-time data processing capabilities for autonomously detecting structural defects in rails and railroad ties during regular train service runs. Dr. Datta has experience with beamforming techniques for impact and structural defect localization in plates. Other research projects Dr. Datta has worked on include delamination detection in composite plates using infrared thermography and vibration-based stiffness degradation monitoring of structures subjected to earthquakes.

Dr. Datta has a breadth of experience in teaching and industry. At UC San Diego, he worked as a course instructor and taught three undergraduate courses on Solid Mechanics. Prior to beginning his PhD, Dr. Datta worked as an Assistant Professor of Civil Engineering at Assam Engineering College, India where he taught courses such as Design of Steel Structures and Structural Analysis. As an assistant project engineer at Indian Institute of Technology Guwahati, Dr. Datta developed a user-friendly MATLAB GUI based structural health monitoring toolbox incorporating various techniques for system identification using time and frequency domain analyses. He has also worked at Larsen and Toubro Construction Limited, in India, as a quality control site engineer.

Authors:

Diptojit Datta University of California San Diego
Ali Zare Hosseinzadeh University of California San Diego
Izabela Batista University of California San Diego
Francesco Lanza Di Scalea University of California San Diego