Background
Railroads are a critical component of transportation infrastructure, facilitating the movement of goods and people. As train speeds and axle loads increase, so does the strain on aging and deteriorating railroad tracks, leading to a rise in rail defects. One of the most hazardous types of defects is transverse defects located inside the rail head, which can cause rail breaks if they reach a critical size. According to the Federal Railway Administration, crack-type defects in the rail head were responsible for a significant percentage of railway track-related accidents in the United States between 2011 and 2020. To ensure the safety and reliability of railroads, there is a pressing need for effective rail defect detection technologies.
Current approaches to non-destructive evaluation (NDE) of rail defects, such as conventional ultrasonic testing, face several challenges. These systems often use piezoelectric transducers in a pulse-echo configuration, but surface defects can obstruct the ultrasonic waves, making it difficult to detect subsurface defects. Additionally, phased array ultrasonic testing, while capable of characterizing defects, requires precise positioning of transducers and suffers from limitations related to contact between the rail and the probe.
Non-contact inspection systems based on guided waves have been developed to address some of these issues, but they still face challenges such as the need for high-frequency propagation and the complexity of integrating various damage-sensitive features into a reliable detection framework. These limitations highlight the need for improved technologies that can provide accurate, non-contact, and efficient rail defect detection.
Technology overview
A rail defect detection system based on laser Doppler vibrometer (LDV) measurements has been developed to enhance the safety and reliability of railroads. This non-contact system is mounted on a rail car and uses two LDVs to record ultrasonic guided waves induced by random rail-wheel interaction forces. The system processes these signals through a damage detection framework that filters time-varying mean and impulsive noise, and then quantifies changes using moving standard deviation.
In the post-processing stage, two damage features based on the relative changes in the LDV signals and transfer functions between measurement points are combined using multivariate statistical analysis. This creates a damage index that identifies rail segments affected by defects. The system's effectiveness was validated through field tests at the Transportation Technology Center in Pueblo, CO, where it successfully detected a welded joint used to simulate a rail defect at speeds up to 48 km/h.
This technology is differentiated by its non-contact, remote, and compact design, which allows it to be mounted on operational rail cars without interrupting routine railroad operations.
Unlike conventional ultrasonic testing methods that require contact and can be obstructed by surface defects, the LDV-based system can detect internal defects without being affected by surface irregularities. The use of guided waves and LDV measurements ensures high spatial resolution and the ability to operate at higher speeds. The incorporation of multivariate statistical analysis in the damage detection framework further enhances the system's accuracy and reliability.
This innovative approach addresses the limitations of existing rail inspection methods, offering a more efficient and effective solution for monitoring rail integrity and preventing accidents caused by rail defects.
Benefits
- Non-contact rail damage detection
- Uses laser Doppler vibrometer (LDV) measurements
- Mounted on a rail car for continuous monitoring
- Detects changes in wave amplitude caused by defects
- Validated through field tests at various speeds
- Capable of identifying rail segments affected by defects
- Reduces false alarms compared to traditional methods
- Works without interrupting routine rail operations
- High-spatial resolution and wide frequency range
- Improves safety and reduces maintenance costs
Applications
- Railway track maintenance
- Rail defect detection
- Non-contact measurement systems
- Rail safety monitoring
Patent
US patent application serial number 19/184,364, “A damage detection framework to detect rail defects using laser doppler vibrometer measurements collected from a moving platform”
