COLD CLIMATE ULTRASONIC RAIL FLAW TESTING  
for Transport Canada  

 
Prepared by Survesh Shrestha, Anish Poudel, MxV Rail 
Glenn Washer, University of Missouri 

 
Proprietary Report 
P-23-020 
June 30, 2023 

 
 
 
 
 
 
 
 
 
 

 

 

 
MxV Rail 

A subsidiary of the Association of American Railroads 
350 Keeler Parkway | Pueblo, Colorado USA 81001 

www.mxvrail.com 



 

 

 
 
 
 
 
 
 
 
 
 
 
 

Disclaimer: This report was prepared for Transport Canada Transportation Technology Center, Inc. (dba MxV Rail), 
a subsidiary of the Association of American Railroads, Pueblo, Colorado. It is based on investigations and tests 
conducted by MxV Rail with the direct participation of Transport Canada to criteria approved by them. The contents 
of this report imply no endorsements whatsoever by MxV Rail of products, services, or procedures, nor are they 
intended to suggest the applicability of the test results under circumstances other than those described in this report. 
The results and findings contained in this report are the sole property of Transport Canada. They may not be released 
by anyone to any party other than Transport Canada without the written permission of Transport Canada. MxV Rail is 
not a source of information with respect to these tests, nor is it a source of copies of this report. MxV Rail makes no 
representations or warranties, either expressed or implied, with respect to this report or its contents. MxV Rail 
assumes no liability to anyone for special, collateral, exemplary, indirect, incidental, consequential, or any other kind 
of damages resulting from the use or application of this report or its contents.



 

iii 

Executive Summary 

From December 2022 to June 2023, MxV Rail investigated the effects of extreme cold 
temperatures on ultrasonic rail flaw testing and signal attenuation. This report documents the 
testing and findings of the research performed under Transport Canada contract T8009-
180251/001/TOR. The study has led to an improved understanding of the interaction between cold 
temperatures and the ultrasonic testing (UT) of rail material. MxV Rail conducted the UT using 
rails with internal fatigue defects at temperatures ranging from 21°C to -40°C. The ultrasonic 
signal attenuation measurements were conducted by the University of Missouri-Columbia team at 
temperatures ranging from 20°C to -50°C. Some of the key findings of this research include: 

Ultrasonic Rail Flaw Testing 
• Results from handheld flaw detector tests showed a decrease in signal amplitudes from 

rail samples with a decrease in temperature. Due to the short water path (distance that an 
ultrasonic wave travels through a water medium before reaching the object being 
inspected) present in the contact approach, the cold temperature possibly had a minimal 
effect on the signal amplitude. 

• Results from the walking stick flaw detector test (using a roller search unit [RSU]) also 
showed a significant decrease in the signal amplitude from the 3.175-mm (0.125-inch) 
side-drilled hole (SDH) at extreme cold temperatures. The signal decreased from 80 
percent at room temperature to 36 percent when measured at -40°C. 

• The signal amplitudes from the SDH gradually decreased as the RSU was exposed to the 
cold temperature environment for the duration of testing (typically about 20 mins). 

• When compared to the signal amplitudes obtained from calibrating the equipment at 
room temperature, the signal amplitudes from defects in the rail samples improved when 
the equipment was calibrated at the same temperature as the test. 

• For the rail sample with transverse defect (TD) # 1, a 26 percent signal amplitude 
difference was observed between the results from room temperature calibration and cold 
temperature calibrations at -35°C, and a 22 percent difference was observed at -40°C. 
Similarly, a 14 percent difference in signal amplitudes was observed for the rail sample 
TD # 2 at -40°C. 

• In a perfect scenario, calibrating the equipment at a given cold temperature could 
minimize the loss of ultrasonic signals. 

 
Ultrasonic Signal Attenuation Measurements 

• The effective attenuation in the rail steel increased as a function of decreasing 
temperatures at a rate of 0.0005 to 0.0007 dB/mm/°C. Therefore, when calibration is 
conducted at extreme cold temperatures, some loss of sensitivity may result. The 
significance of this loss in practical terms needs further study. 

• The rail material absorbed more acoustic energy as the temperature decreased. 
• Signal amplitudes from defects or discontinuities in the material would be increased at 



 

iv 

lower temperatures if the calibration were conducted at room temperature. 
• The longitudinal wave velocities in the two specimens, fabricated from two different rail 

samples, were very similar with a difference of only <0.1 percent. The velocities 
increased as temperatures decreased at an average rate of 0.65 m/sec/°C. 

• Longitudinal wave velocities in the ethanol bath also increased as the temperature 
decreased at a rate of 4.1832 m/sec/°C. 

 
  



 

v 

Table of Contents 

1 Introduction ........................................................................................................... 7 
1.1 Background .......................................................................................................... 7 
1.2 Ultrasonic Rail Flaw Testing .............................................................................. 7 

2 Project Objective and Scope ................................................................................ 9 
3 Experimental Setups ............................................................................................ 9 

3.1 Rail Samples ....................................................................................................... 9 
3.2 UT Equipment ................................................................................................... 10 
3.3 UT Test Setups ................................................................................................. 11 
3.4 UT Calibration ................................................................................................... 12 
3.5 Testing in Environmental Chamber (Cold room) .......................................... 13 
3.6 Ultrasonic Signal Attenuation Measurement ................................................. 14 

3.6.1 Equipment .................................................................................. 14 
3.6.1.1 Immersion Chiller Bath ................................................. 14 
3.6.1.2 Ultrasonic Transducer ................................................... 15 
3.6.1.3 Delay Lines ..................................................................... 15 

3.6.2 Test Specimen ........................................................................... 16 
3.6.3 Methodology ............................................................................... 17 

3.6.3.1 Longitudinal Wave Velocity  
Measurements in Couplant .......................................... 17 

3.6.3.2 Shear Wave Velocity in Steel ....................................... 17 
3.6.3.3 Longitudinal Wave Velocity in Steel ............................ 18 
3.6.3.4 Timing ............................................................................. 19 
3.6.3.5 Amplitude Measurements ............................................. 19 

4 Results ............................................................................................................... 21 
4.1 Ultrasonic Rail Flaw Testing ............................................................................ 21 

4.1.1 Handheld Flaw Detector ............................................................. 21 
4.1.2 Walking Stick Flaw Detector ....................................................... 22 

4.2 Ultrasonic Signal Attenuation Measurements ............................................... 25 
4.2.1 Velocity Measurements .............................................................. 25 
4.2.2 Attenuation ................................................................................. 27 

5 Conclusions ........................................................................................................ 30 
6 Guidelines for cold temperature testing .............................................................. 31 
References .................................................................................................................... 33 
 

  



 

vi 

List of Figures 
Figure 1. Ultrasonic transducers configurations in RSUs7 .......................................... 8 
Figure 2. Rail samples with internal TDs .................................................................... 9 
Figure 3. Engineering drawing of the calibration rail specimen (units in mm) ........... 10 
Figure 4. UT Test equipment. Handheld UT flaw detector (left) and  

walking stick with RSU (right) .................................................................... 11 
Figure 5. Temperature sensor attached to a rail sample .......................................... 12 
Figure 6. IIW Type 1 reference block ....................................................................... 12 
Figure 7. Test setup for the cold room test ............................................................... 13 
Figure 8. Schematic showing the test setup using the chiller and  

immersion tank bath .................................................................................. 14 
Figure 9. A chiller bath with transducer and test specimen in  

immersion non-contact setup .................................................................... 15 
Figure 10.  Delay lines used for ultrasonic transducer................................................. 16 
Figure 11.  Engineering drawing of test specimens (units in mm) ............................... 16 
Figure 12.  Schematics showing the timing measurements for ultrasonic velocity 

calculation in liquids .................................................................................. 17 
Figure 13.  Schematics showing measurement parameters in rail test specimen for 

velocity calculations .................................................................................. 18 
Figure 14.  A-scan signal amplitude measurements for TD 1 ...................................... 21 
Figure 15.  A-scan signal amplitudes measurements for TD 3 .................................... 21 
Figure 16.  A-scan signal amplitude measurements  

for 3.175-mm (0.125-inch) SDHs .............................................................. 23 
Figure 17.  Comparison of signal amplitudes for TD 1 before and after calibrating the 

flaw detector at given cold temperatures ................................................... 24 
Figure 18.  Comparison of signal amplitudes for TD 2 before and after calibrating the 

flaw detector at given cold temperatures ................................................... 24 
Figure 19.  Velocity measurements for ethanol in cooling tank ................................... 25 
Figure 20.  Longitudinal wave velocities measurements for  

rail specimens A and B ............................................................................. 26 
Figure 21.  Shear wave velocities measurements for rail specimens A and B ............ 27 
Figure 22.  Signal amplitude measurements for three SDHs ...................................... 28 
Figure 23.  Ratio of signal amplitudes (dB) for three SDHs ......................................... 28 
Figure 24.  Attenuation (dB/mm) measurements for rails with SDHs .......................... 29 
Figure 25.  Signal amplitudes for calibration at 20°C and calibration at  

inspection temperature .............................................................................. 30 
 
List of Tables 
Table 1. Transducer orientation and targeted flaw types7 ............................................... 8 



 

7 

1 INTRODUCTION 
Transport Canada’s (TC) Innovation Centre (IC) awarded MxV Rail a Phase II contract to 
continue researching “cold climate railway technologies.” The tests and findings of the research 
conducted between December 2022 to June 2023 under Transport Canada contract T8009-
180251/001/TOR are documented in this report. The work conducted during this phase continued 
investigating the effect of extreme cold temperatures on the ultrasonic rail flaw testing studied in 
the previous phase1 and led to an improved understanding of the interaction between cold 
temperatures and the ultrasonic testing of rail material. 

1.1 Background 
Rail defects come in various forms (shapes, sizes, locations, and orientations) and are well 
described in the United States Department of Transportation (USDOT) Federal Railroad 
Administration (FRA) rail defect manual and in various railroads and rail service providers’ rail 
defect manuals. Minute anomalies or defects can act as initialization sites for internal rail cracks 
that can grow with accumulated train tonnage.2-4 Broken rails and welds pose significant 
challenges for railroads, especially those operating in extreme cold weather conditions.5 One of 
the main reasons for increased broken rails in the extreme cold winter is due to the rail 
temperature decreasing by a significant amount from the rail neutral temperature, thereby 
causing the rails to experience high tensile forces6 and subsequent failure at weak spots, such as 
fatigue defects that form inside the rails and welds. 

North American Class I railroads rely primarily on ultrasonic testing (UT) non-destructive 
evaluation (NDE) technology to routinely detect and characterize rail defects. However, even 
with such advanced NDE technology, some defects are only discovered after the rail breaks.7 
Therefore, improving railroad safety through the reduction of rail failures and the associated 
risks of train accidents while achieving higher reliability with ultrasonic NDE inspection for rail 
flaw detection and characterization is the primary goal of rail defect detection. 

1.2 Ultrasonic Rail Flaw Testing 
The in-motion UT technology for rail flaw testing is usually implemented on a hi-rail vehicle 
platform (referred to as “rail detector cars”). These cars use fixed-angle piezoelectric transducers 
housed in a liquid-filled membrane (tire) called a roller search unit (RSU) to generate/emit 
ultrasonic waves in the rail. Figure 1 shows a diagram of a typical RSU configuration running on 
the rail. The ultrasonic signal path is displayed as a single line, but this line indicates an array of 
the ultrasonic beams being transmitted and received. Similarly, Table 1 shows different 
transducer orientations that can be used to finding different types of internal rail defects.  

The peak testing speed that rail detector cars cannot exceed is 30 km/h on North American 
railroads, but the start-stop “hand verify” operation often limits average test speeds to around 12 
km/h.7 In some countries, inspection speeds are reported to be as high as 100 km/h.8, 9 In many 
cases, a detector car does not stop to verify defects. Instead, the inspection data is recorded and 
analyzed later in the back office. For North American Class I railroads, the start-stop hand verify 
operation requires NDE operators to stop the detector car, step down from the vehicle, and 
manually hand verify flaws in the rail using the portable ultrasonic system to see if any relevant 



 

8 

indication can be detected. Hand verification is usually done to confirm a detector car’s findings 
and size the defect so proper remedial action can be taken. 

 

 
Figure 1. Ultrasonic transducers configurations in RSUs7 

 
Table 1. Transducer orientation and targeted flaw types7 

Transducer 
Orientations Target Flaw Types 

0-degree Flaws oriented horizontally -shells, horizontal split head (HSH), split web 

37.5- or  
45-degree Bolt hole cracks, web defects 

70-degree Transverse defects (TDs), vertical split heads (VSH) weld defects (porosity, 
inclusions, etc.) 

 
The FRA has recently started to allow North American railroads to conduct continuous rail 

testing using UT augmented with global positioning system (GPS) technology. In this approach, 
rail inspection equipment can collect and transmit inspection data nonstop to remote locations for 
detailed analysis.10 If a suspected defect is verified, FRA regulations require the railroad to 
immediately apply the proper remedial action. In some cases, the proper remedial action means 
repairing or replacing the defective rail, slowing trains over the defect, or removing the track from 
service to stop train movements until repairs can be made.11  



 

9 

2 PROJECT OBJECTIVE AND SCOPE 
The main objectives of this research were to both investigate and understand the effect of cold 
temperatures on ultrasonic rail flaw testing capabilities. As part of this research, the following 
tasks were performed: 

• Conduct UT on rail samples with internal flaws using handheld and walking stick 
ultrasonic flaw detectors at cold temperatures 

• Quantify ultrasonic signal loss and attenuation in rail steel as a function of temperature. 
 
3 EXPERIMENTAL SETUPS 
The cold climate tests were conducted in three different test environments. Rails with internal 
fatigue defects at MxV Rail’s facilities were inspected in both a cold ambient environment using 
both the contact UT approach (handheld transducer) and the non-contact UT approach (RSU), 
and in an environmental chamber (cold room) using the non-contact approach. The ultrasonic 
signal attenuation measurements were conducted by the University of Missouri-Columbia team 
using a non-contact approach in an ultrasonic immersion tank. The following subsections 
describe the rail samples, the UT equipment, and setups used for the testing. 

3.1 Rail Samples 
The rail samples considered for testing had internal TDs that occurred naturally from years of 
service. Three sections of rail samples with TDs of different sizes and orientations were selected, 
with one of the rail samples containing two TDs close together. The defects in the rail samples 
were labeled TD 1, TD 2, TD 3, and TD 4, as shown in Figure 2.  

 

 

  
Figure 2. Rail samples with internal TDs 

TDs 1 & 2 Calibration Rail 

TD 3 TD 4 



 

10 

Two identical rail calibration specimens were fabricated for the RSU transducer’s calibration. 
Each calibration rail specimen had a 3.175-mm (0.125-inch) diameter side-drilled hole (SDH) 
12.7 mm (0.5 inch) below the top of the rail head. The engineering drawing of the calibration rail 
is shown in Figure 3. One of the calibration rail specimens was placed in the lab at room 
temperature (~21°C) and used for the room temperature calibration. The other calibration rail 
specimen was placed outdoors in the cold ambient environment near the defective rail samples 
for the cold temperature calibration. 

 

 Figure 3. Engineering drawing of the calibration rail specimen (units in mm) 
 

3.2 UT Equipment 
A handheld UT flaw detector and a walking stick system that uses an RSU (Figure 4) were used 
for the ultrasonic signal measurements. A 2.25 MHz 70-degree ultrasonic transducer was used 
with a handheld flaw detector (Olympus Epoch 600) to scan and measure the ultrasonic signals 
from the rail defect samples. Similarly, a 70-degree transducer in the RSU was used with the 
same flaw detector to scan and measure the ultrasonic signals from rail defect samples. 



 

11 

   
Figure 4. UT Test equipment. Handheld UT flaw detector (left) and walking stick with RSU (right) 

 
A couplant is a medium applied between the transducer (probe) and the test specimen in UT. 

The primary purpose of using a couplant is to facilitate the transmission of ultrasonic waves from 
the transducer into the specimen and back. EchoPure water-free, medium viscosity couplant 
was used for the contact UT measurements. Similarly, a 50/50 mixture of propylene glycol and 
water was used as a couplant for the non-contact UT measurements. Propylene glycol has a 
melting point of -60°C and acts as an antifreeze that helps facilitate testing in temperatures below 
0°C. It is readily biodegradable in water and soil making it favourable to use it in tracks.  

3.3 UT Test Setups 
The ultrasonic rail flaw testing was conducted using defective rail samples on a track panel in a 
cold ambient environment at temperatures ranging from approximately 0°C to -22°C. The rail 
samples were allowed to cool down naturally in the cold ambient environment. The UT was 
conducted using both contact and non-contact approaches. After completing the tests in the cold 
ambient environment, additional tests were conducted in a cold environmental chamber using the 
same approach. In all the tests conducted, the equipment used was exposed to the extreme cold 
environment only for a short duration while conducting the tests. Exposing the ultrasonic flaw 
detector equipment to an extreme cold environment for an extended period could significantly 
affect the results, and any such effects have not been considered in this research. A resistance 
temperature detector (RTD) was attached to the surface of one of the rail samples to monitor the 
temperature of the rail samples (Figure 5). 

 



 

12 

 
Figure 5. Temperature sensor attached to a rail sample 

 

3.4 UT Calibration 
A 2.25 MHz 70-degree shear wave transducer was used with the handheld UT flaw detector and 
calibrated using the 1.5-mm (0.06-inch) diameter SDH in the IIW Type 1 reference block in a 
laboratory environment at room temperature (21°C) (Figure 6). The amplitudes of the ultrasonic 
signals from the TDs in the rail samples were measured at room temperature and used as the 
baseline. The rail samples were then placed outdoors in a cold ambient environment at 
temperatures ranging from approximately 0°C to -22°C. The ultrasonic signal amplitudes from 
the rail sample defects were measured at different cold ambient temperatures ranging from 
approximately 0°C to -22°C and compared with the baseline measurements. 

 
Figure 6. IIW Type 1 reference block 



 

13 

The 70-degree shear wave transducers in the walking stick UT system (RSU) were calibrated 
using the calibration rail specimens. Baseline amplitude measurements from the rail samples 
were initially recorded with the walking stick UT system (RSU) calibrated at room temperature. 
Also, while performing the tests, the walking stick UT system (RSU) was first calibrated using 
the calibration rail at room temperature, and the signal amplitude readings from the defective rail 
samples were measured at a given cold temperature. Then, the walking stick (RSU) was again 
calibrated using the other calibration rail at the given cold temperature, and the signal amplitude 
readings from the defective rail samples were measured again. The two data sets were used to 
compare the signal amplitude differences from calibrating at room temperature versus the given 
cold temperature.  

3.5 Testing in Environmental Chamber (Cold room) 
Tests were also conducted in a controlled environment inside an environmental chamber (cold 
room). The cold room had an internal floor area of 2.74 m×2.74 m (9 ft.×9 ft.) and was capable 
of reaching -65°C. However, due to the scope of this research, the environmental chamber was 
only used for tests conducted at temperatures ranging from 0°C to -40°C. The rail samples used 
for the cold room test were TD 1 and TD 2. Tests were conducted only using the walking stick 
system. The handheld flaw detector was not used for testing due to the safety hazards posed by 
extreme cold temperatures. The rail samples and the calibration rail specimen used for cold 
temperature calibration were placed in the cold room as shown in Figure 7. Testing was 
conducted using the approach described in Section 3.4. The measurements were taken at every 
5°C interval between 0°C and -40°C. 

 
Figure 7. Test setup for the cold room test 

 



 

14 

3.6 Ultrasonic Signal Attenuation Measurement 
The ultrasonic signal attenuation study was conducted by the University of Missouri-Columbia 
team in their laboratory. The equipment, test specimens, and procedures used for this study are 
described in this section. 

3.6.1 Equipment 
3.6.1.1 Immersion Chiller Bath 
Ultrasonic signal attenuation measurements were conducted in a chilled bath placed in an 
ultrasonic immersion tank, as shown in Figure 8. The test arrangement utilized an external chiller 
(Neslab Ult 80) that was able to achieve -80°C in the chiller itself. Due to its very low freezing 
temperature, ethanol was used as the liquid in the chiller bath. The primary reasons for using 
ethanol instead of a glycol/water mixture are 1) that the glycol/water mixture gets too viscous for 
UT and 2) that it absorbs much of the ultrasonic energy as the temperature drops below -10°C. 
The high viscosity of the glycol/water mixture also makes it more difficult for the recirculating 
pumps to move the fluid, possibly affecting how cold the bath can get. 

 

Figure 8. Schematic showing the test setup using the chiller and immersion tank bath 
 

The chiller ethanol was piped through Tygon tubing to a copper heat exchanger in a cooling 
tank, as shown in Figure 9. The cooling tank also contained ethanol as the chiller liquid. A 
circulating pump was used to circulate the ethanol in the cooling tank to ensure temperature 
uniformity in the cooling tank. This test arrangement provided a stable minimum temperature of 
-50°C. Figure 9 also shows the location of the ultrasonic transducer mounted on the gimbal used 
to adjust the transducer angle. The figure also illustrates the location of the cooling tank within 
the stainless steel ultrasonic immersion tank. The immersion tank is operated by a MISTRAS 
four-channel ultrasonic pulser-receiver that also controls the motion of the motor-driven control 
system that allows programmed scanning in the x, y, and z axes and changes the incidence angle 
of the transducer. 



 

15 

 
Figure 9. A chiller bath with transducer and test specimen in immersion non-contact setup 

 
3.6.1.2 Ultrasonic Transducer 
The ultrasonic transducer used was a 2.25 MHz, 12.7-mm (0.5-inch)-diameter ultrasonic 
immersion transducer (model V306-SU) manufactured by Olympus America, Inc. The 
transducer is broadband, produces a relatively short waveform duration, and is recommended for 
situations where good axial resolution is important, such as timing measurements for 
determining velocity. A single transducer was used for all measurements. 

3.6.1.3 Delay Lines 
Delay lines were manufactured from a 31.75-mm (1.25-inch) cast acrylic rod to buffer the cold 
temperatures experienced during testing. A stainless-steel collar was fabricated to fix the delay 
lines to the front of the transducer and apply suitable pressure to enable the efficient transfer of 
acoustic energy. The delay line was coupled to the transducer using Echo Ultrasonic Forever 
wedge couplant. Figure 10 shows a photograph of the three delay lines. The delay lines measured 
6.35 mm (0.25 inch) long, 19.05 mm (0.75 inch) long, and 31.75 mm (1.25 inches) long. 



 

16 

 
Figure 10. Delay lines used for ultrasonic transducer 

3.6.2 Test Specimen  
Testing was conducted on two identical test specimens fabricated from different 136RE rail 
samples. The specimens were sectioned from the head and web area of the rail and arbitrarily 
labeled Specimen A and Specimen B. Figure 11 shows the engineering drawing of the 
specimens. Each specimen was fabricated with a 50.8-mm (2-inch) radius and three 1.5-mm 
(0.06-inch) SDHs at depths of 12.7 mm (0.5 inch), 25.4 mm (1 inch), and 38.1 mm (1.5 inch) 
from the surface of the specimen and labeled as SDH-1, SDH-2, and SDH-3, respectively. 

 
Figure 11. Engineering drawing of test specimens (units in mm) 

 
  



 

17 

3.6.3 Methodology 
3.6.3.1 Longitudinal Wave Velocity Measurements in Couplant 
The longitudinal wave velocity in ethanol was measured by changing the elevation of the 
transducer and measuring the time of flight for temperature ranging from 20°C to -50°C. The 
time of flight refers to the time it takes for an ultrasonic wave to travel from the transducer to a 
target and back to the transducer. Figure 12 shows a schematic of the measurement procedure. 
The velocity was calculated from Equation 1: 

 𝐶𝐶 =
𝑑𝑑1 − 𝑑𝑑2
𝑡𝑡1 − 𝑡𝑡2

  (1) 

Where: 

d1 and d2 are the two distances shown in Figure 12, and 

t1 and t2 are the measured travel times of the ultrasonic pulse. 

 
Figure 12. Schematics showing the timing measurements for  

ultrasonic velocity calculation in liquids 
 

3.6.3.2 Shear Wave Velocity in Steel 
Shear wave velocities were determined by placing test specimens in the cooling tank and scanning 
the surface of the specimen, as shown in Figure 13. The following dimensions were as-machined: 
x0 = 38.1 mm (1.5 inches), d1 = 12.7 mm (0.5 inch), d2 = 25.4 mm (1 inch), and d3 = 38.1 mm (1.5 
inches). The focal point of the 50.8-mm (2-inch) radius is shown in the figure. To locate the 
position of the transducer in the scanning tank, the maximum amplitude of the reflection from the 
50.8-mm (2-inch) radius was located and used as the reference during scanning. Locations x1 and 
x3 were determined from the maximum amplitude reflections from the 1.5-mm (0.06-inch) SDHs. 



 

18 

The angle measurement was calculated as shown, for example, in Equation 2: 

 tan𝜃𝜃 =
𝑥𝑥3 − 𝑥𝑥0
𝑑𝑑3

  (2) 

 
The ultrasonic velocity was determined by scanning across the specimen and locating the 

maximum amplitude reflections from the SDHs at depths d1 and d3. The velocity was then 
determined as follows: 

 

 𝐶𝐶 =
𝑃𝑃3 − 𝑃𝑃1
𝑡𝑡3 − 𝑡𝑡1

  (3) 

Where: 

t3 is the timing of the reflection from the SDH at d3,  

t1 is the timing of the reflection from the SDH at d1,  

P1 and P3 are the path lengths for wave travel in the steel as determined from the geometry of 
the depth of the SDH and the horizontal position of the transducer (i.e., x3 – x1).  

In this way, system delays were eliminated from the velocity measurement. 

 

 
Figure 13. Schematics showing measurement parameters 

 in rail test specimen for velocity calculations 
 

Testing was conducted at 20º to -50ºC at intervals of 10ºC. This required that the temperature 
in the cooling tank be set at the appropriate level for the desired measurement. The specimen was 
allowed to stabilize at the given temperature for approximately 60 minutes. 

3.6.3.3 Longitudinal Wave Velocity in Steel 
Measurements of the longitudinal wave velocity were performed by measuring the time 
difference between the front wall and back wall reflections through the thickness of the 



 

19 

specimens shown in Figure 13. The path length of the longitudinal wave was approximately 
50.8 mm (2 inches) through the specimen (101.6 mm [4 inches], two-way path length). 

3.6.3.4 Timing 
The timing of ultrasonic waves was completed during the data postprocessing of waveforms 
captured during scanning. Waveforms were captured at a 50 MHz data acquisition rate 
(20-nanosecond intervals). Waveforms were collected as the transducer, set at the appropriate 
incidence angle, scanned over the specimen’s surface. Data was collected at 0.1 mm (0.004 inch) 
steps during the scan to produce a B-scan showing three SDHs and the radius of the specimen. 
Waveforms were selected by identifying the maximum reflection from each SDH, and these 
waveforms were analyzed to determine the shear wave velocity.  

A timing algorithm that selects the rising first peak of the waveform to be timed was used. 
Data points above and below the zero-crossing of the waveform were used to interpolate the 
zero-crossing between the points to determine the time of flight for the waveform. For the 
longitudinal wave measurements, the time-of-flight from the front wall reflection to the back 
wall reflection was determined to calculate the velocity of the wave through the steel. 

3.6.3.5 Amplitude Measurements 
Amplitude measurements were collected at different temperatures based on reflections from the 
SDHs in specimens A and B. The maximum amplitude from the SDH was determined by 
scanning across the specimen and recording the A-scan signals at each step along the scan line. 
A portion of the A-scan was analyzed using the gate feature to determine the maximum 
amplitude from the hole. 

The attenuation in the material itself was determined by comparing the amplitude of SDH-1 
at a depth of 12.7 mm (0.5 inch) to the amplitude of the SDH-2 and SDH-3 at depths of 25.4 mm 
(1.0 inch), and 38.1 mm (1.5 inches), respectively. For these measurements, the signal amplitude 
was calculated relative to the SDH-1 value at each temperature: 

 𝑑𝑑𝑑𝑑𝑇𝑇 = 20 ∗ 𝑙𝑙𝑙𝑙𝑙𝑙
𝐴𝐴𝑇𝑇,𝑖𝑖

𝐴𝐴𝑇𝑇,1
  (4) 

Where: 

dBT = Relative amplitude of SDHi at temperature T (°C), 

AT,i = Amplitude (V) of reflected signal from SDHi at temperature T (°C), and 

AT,1 = Amplitude (V) of reflected signal from SDH-1 at temperature T (°C). 

In this way, temperature-dependent variations in the incident wave amplitude, 
transmission/reflection coefficient, and attenuation in the ethanol are eliminated from the 
measurement since these factors affect both the amplitudes from SDH-1 and the other SDHs in 
the specimens equally. Therefore, the resulting differences result from changes in the material 
itself. These changes include the attenuation characteristics of the material as well as variations 
caused by temperature-dependent changes in the diffraction of the beam to the extent these 
changes may exist. Such changes in beam diffraction were outside the scope of the current 



 

20 

research. However, these effects would also be part of a practical testing scenario, despite slight 
variances in frequency, transducer diameter, and material anisotropy.   

An attenuation factor α` was calculated by dividing the amplitude dBT by the one-way path 
length of the refracted 45-degree wave, resulting in units of dB/mm. The calculated attenuation 
factor for the fixed frequency of 2.25 MHz as measured within the test conditions is described in 
this report.  

Additional analysis was completed to compare signal amplitudes for SDH-2 and SDH-3 to a 
calibration performed at room temperature: 

 𝑑𝑑𝑑𝑑𝑇𝑇 = 20 ∗ 𝑙𝑙𝑙𝑙𝑙𝑙
𝐴𝐴𝑇𝑇,𝑖𝑖

𝐴𝐴20℃,1
  (5) 

Where: 

A20°C, ,1 = Amplitude (V) of reflected signal from SDH-1 at room temperature of 20°C 

Equation 5 provides a measure of the impact of calibrating at room temperature and 
conducting inspections at colder temperatures. This data is included to provide an illustration of 
how these changes might affect a practical test scenario.   

The data reported herein for attenuation in the rail specimens was collected at a refracted 
angle of 45-degrees. The incidence angle was adjusted for each temperature to compensate for 
changes in the longitudinal velocity in the ethanol bath and the steel itself. As a result, each 
temperature for which data is reported had a different incidence angle in order to produce a 
45-degree refracted angle in the steel.   

The data was collected by performing three consecutive scans of specimen A and specimen B 
at each test temperature ranging from 20 °C to -50 °C. The transducer was lowered into the 
coolant and then removed after each of the scans. The transducer and delay line were completely 
removed from the ethanol bath between each scan, held at room temperature air for several 
minutes, and then reinserted into the ethanol bath. At low temperatures, it was observed that the 
overall signal amplitudes varied somewhat between subsequent tests. This variance was likely 
due to delay line material cooling as a result of the short-term immersion in the ethanol. 
However, these small variations are normalized in the attenuation results because of the 
analytical approach used. It is noted herein for future researchers developing procedures for this 
type of testing. 

  



 

21 

4 RESULTS 
This section describes the ultrasonic signal amplitude measurements obtained from the ultrasonic 
rail flaw testing and the signal attenuation measurements at extreme cold temperatures. 

4.1 Ultrasonic Rail Flaw Testing 
4.1.1 Handheld Flaw Detector 
The maximum signal amplitudes obtained from the TDs in the rail samples at cold temperatures 
were measured and plotted. The lowest ambient temperature recorded at the test site 
was -21.7°C. Therefore, the measurements were only conducted between 21°C to -21.7°C. The 
results from TD 1 and TD 3 are shown in Figure 14 and Figure 15. 

 
Figure 14. A-scan signal amplitude measurements for TD 1 

 
Figure 15. A-scan signal amplitudes measurements for TD 3  



 

22 

 
The data was observed to be highly scattered. The linear regression lines from the plots show 

decreasing signal amplitudes with decreasing temperatures. One of the challenges faced during 
the on-track testing was the inconsistency in the data. Various factors could have affected the 
signal amplitude measurements. These factors include the person conducting the test, the 
calibration of the equipment, the placement of the ultrasonic transducer on the sample, and the 
amount of couplant present between the transducer and the sample during the test. These factors 
could have been the reason behind the inconsistent and scattered data. 

The signal amplitude measurements obtained from the TDs at colder temperatures were 
mostly found to be close to the signal amplitudes measured at room temperature, except for a 
few measurements that are below 90 percent signal amplitude in the plots shown in Figure 14 
and Figure 15, which could be due to negligible water path distances in the contact approach, and 
the cold temperature could have had a minimal effect on the signal amplitudes. 

4.1.2 Walking Stick Flaw Detector 
Contrary to the contact approach using a handheld transducer, the RSU in the walking stick had a 
more significant water path due to the liquid inside the RSU. Also, the tests conducted in the 
cold room were in a controlled environment; therefore, the results obtained from the cold room 
tests were more reliable than those obtained from the ambient environment testing. Therefore, 
the results obtained from the cold room testing are presented in this section. 

The signal amplitudes from the 3.175-mm (0.125-inch) SDH in the calibration rail specimen 
were measured between 21°C (room temperature) and -40°C to study the effect of extreme cold 
temperatures on the ultrasonic signal. Three measurements were taken at room temperature and 
at every 5°C interval between 0°C to -40°C. Figure 16 shows the plot of average signal 
amplitudes from the SDH measured at temperatures ranging from 21°C to -40°C. Researchers 
observed that the amplitude decreased linearly as the temperature decreased. A significant drop 
in the ultrasonic signal from the 3.175-mm (0.125-inch) SDH was observed at -40°C. The signal 
amplitude decreased from 80 percent at room temperature to 36 percent at -40°C. A similar trend 
in the signal amplitude from the 3.175-mm (0.125-inch) SDH was observed for the ambient cold 
temperature testing. During these tests, it was also observed that the signal amplitudes from the 
SDH gradually decreased as the RSU was exposed to the cold temperature environment for a 
more extended period. 

 



 

23 

   
Figure 16. A-scan signal amplitude measurements for 3.175-mm (0.125-inch) SDHs 

 
The effect of the cold temperature on the ultrasonic signals can cause challenges during defect 

detection. One of the things that could be considered to minimize the effect of cold temperature on 
the ultrasonic signals is the calibration of the flaw detector at a given cold temperature. Hence, a 
calibration rail specimen was used in the test to calibrate the flaw detector at a given cold 
temperature. The results obtained from the tests after calibrating the test equipment at given cold 
temperatures were compared with those obtained from the calibrated equipment at room 
temperature. Figure 17 and Figure 18 show the comparisons of the results obtained from 
calibrating the equipment at room temperature and at given cold temperatures for TD 1 and TD 2. 
It was observed that the signal amplitudes from both the samples, TD 1 and TD 2, improved to 
some degree after calibrating the equipment at the actual inspection temperatures. The results 
obtained from TD 1 show that the signal amplitudes increased as the temperature decreased. 
However, this was different for TD 2 because the signal amplitude trendline obtained from TD 2 
was almost parallel to the x-axis. The defects in TD 1 and TD 2 were in different orientations, 
which could have affected the amount of ultrasonic signals reflected back to the transducer. 

A 26 percent difference in the signal amplitudes from TD 1 between the results from room 
temperature and given cold temperature calibrations was observed at -35°C, while a 22 percent 
difference was observed at -40°C. Similarly, a 14 percent difference in signal amplitudes was 
observed at -40°C for TD 2. Based on this information, it can be concluded that, in a perfect 
scenario, the loss of ultrasonic signals can be minimized by calibrating the equipment at a given 
ambient cold temperature. 

 



 

24 

 
Figure 17. Comparison of signal amplitudes for TD 1 before and after calibrating the flaw detector 

at given cold temperatures 

 
Figure 18. Comparison of signal amplitudes for TD 2 before and after calibrating the flaw detector 

at given cold temperatures 
  



 

25 

4.2 Ultrasonic Signal Attenuation Measurements 
4.2.1 Velocity Measurements 
In order to calculate the ultrasonic signal attenuation, ultrasonic wave velocities in ethanol and 
the rail specimens were initially determined. First, the velocity of ultrasonic waves in the ethanol 
liquid was determined (Figure 19). The linear regression indicated that the increase in velocity in 
the ethanol bath was -4.1832 m/sec/ºC. The coefficient of determination (R2) was found to be 
0.9794. The data illustrates the increase in velocity in ethanol as temperatures decrease. 

 
Figure 19. Velocity measurements for ethanol in cooling tank 

 
Second, the longitudinal wave velocities in rail specimens A and B were determined and are 

shown in Figure 20. The longitudinal wave velocities were very similar for the two samples 
tested. The regression data indicated the change in wave speed of -0.6272 m/sec/ºC for specimen 
A and -0.6642 m/sec/ºC for specimen B. 

 



 

26 

 
Figure 20. Longitudinal wave velocities measurements for rail specimens A and B 

 
Third, shear wave velocities were determined for specimens A and B and are shown in 

Figure 21. There were noticeable differences in the shear wave velocities observed through the 
experimentation. However, close examination of the data indicated that the quality of the beam 
profiles was diminished by the effects of temperature on the material, making exact timing of the 
signals problematic. Time precluded re-analysis of data subsequently collected with improved 
beam profiles. Regardless, the trend of the data illustrates that the shear wave velocities also 
increase as the temperature decreases. 



 

27 

 
Figure 21. Shear wave velocities measurements for rail specimens A and B 

 

4.2.2 Attenuation 
After determining the velocity of ultrasound in rail and ethanol, attenuation measurements were 
conducted. The summary of results included here focuses on the attenuation in the rail steel at 
different temperatures as determined using Equation 4 described earlier. Extensive data analysis 
was completed to assess the behavior of the rail steel at extreme cold temperatures, and this 
analysis indicated that there was no significant difference between the behavior of specimen A 
and specimen B. Therefore, the data included in this report combined measurements from 
specimen A and specimen B into a single data set. 

Figure 22 shows the average signal amplitudes (V) for each of the SDHs. This data is based 
on three consecutive scans at each temperature, and the sample standard deviation is shown as 
whiskers for each data point. It can be observed in the figure that the amplitudes from SDH-1 
were increasing as the temperature decreased. Signal amplitudes from SDH-2 and SDH-3 also 
increased but at a much lower rate. 

The overall increase in signal amplitude reflects that temperature-dependent increases in 
wave amplitude as the material (delay line, ethanol, and steel) increased in stiffness. Given that 
these changes affect each of the SDH reflections equally, the amplitude values for each SDH 
should remain parallel across the test temperatures. However, the values are not changing in 
parallel due to changes in the material characteristics changes (i.e., attenuation) due to lower 
temperatures. 



 

28 

   
Figure 22. Signal amplitude measurements for three SDHs 

 

To analyze this effect, the data was analyzed assuming that SDH-1 was the calibration 
article. The amplitudes from SDH-2 and SDH-3 were compared with this value at each 
temperature as shown in Figure 23. This data clearly indicates that the attenuation in the material 
is increasing as the temperatures decrease, resulting in the signal from SDH-2 and SDH-3 being 
smaller as compared with SDH-1 as the temperature is decreased. 
 

   
Figure 23. Ratio of signal amplitudes (dB) for three SDHs 



 

29 

This data was used to calculate the effective attenuation factor α` by dividing the relative 
amplitudes from SDH-2 and SDH-3 by the respective one-way travel paths. This data is shown 
in Figure 24, and it indicates that the rate of attenuation increase is 0.0005 to 0.0007 dB/mm/°C 
(0.0127 to 0.01778 dB/in/°C). This data indicates that the rail steel is absorbing more energy as 
the temperature decreases, meaning the signal from defects or discontinuities would be of lower 
amplitude at extreme cold temperatures as compared with typical ambient temperatures, 
assuming the calibration is conducted at the same temperature as the test. 

   
Figure 24. Attenuation (dB/mm) measurements for rails with SDHs 

 
To further examine the phenomenon of decreasing defect signal amplitude at extreme cold 

temperatures, data was analyzed assuming that the calibration was conducted at room temperature, 
and then inspections were conducted at extreme cold temperatures. In this case, the amplitude 
values (V) from SDH-2 and SDH-3 at each temperature step were compared with the value (V) of 
SDH-1 at 20°C and are shown in Figure 25. This data is revealing because it demonstrates that the 
apparent attenuation in the material is decreasing with temperature (i.e., signal amplitudes are 
increasing as compared with the room temperature amplitude). However, this decrease in signal 
attenuation in the material with decreasing temperature results from increased signal amplitudes 
overall as the temperatures decrease and material stiffness increases. In other words, defect or 
discontinuity signals would be increased at colder temperatures (relative to calibration at 20°C), 
suggesting increased sensitivity for defect detection. Because the input function (acoustic energy) 
increases as the temperature decreases, the resulting signal would be greater for the same defect 
detected at -50°C as compared with the defect detected at 20°C. 



 

30 

   
Figure 25. Signal amplitudes for calibration at 20°C and calibration at inspection temperature 

 
5 CONCLUSIONS 
The Canadian railway system faces significant challenges due to the extreme cold winter 
conditions. Monitoring the rails in such extreme cold conditions is very difficult. Therefore, 
research has been focused on advancing technologies to facilitate rail inspection in extreme cold 
weather. This research investigated the effects of extreme cold temperatures on ultrasonic rail 
flaw testing and ultrasonic signal attenuation. The ultrasonic rail flaw testing was conducted on-
track in a cold ambient environment and in an environmental chamber (cold room). The 
ultrasonic signal attenuation measurements test was conducted in an ultrasonic immersion tank. 

The measurements in this research were intended to provide signal amplitude data from rail 
flaw samples at temperatures ranging from 21°C to -40°C and ultrasonic wave velocity and 
attenuation data at temperatures ranging from 20°C to -50°C. Some of the key findings of this 
research include: 

Ultrasonic Rail Flaw Testing 
• Results from the test using a handheld flaw detector showed a decrease in signal 

amplitudes from the rail samples with the decrease in temperature. Due to the short water 
path present in the contact approach, the cold temperature possibly had a minimal effect 
on the signal amplitude. 

• Results from the test using a walking stick flaw detector showed a significant decrease in 
the signal amplitude from the 3.175-mm (0.125-inch) SDH at extreme cold temperatures. 
The signal decreased from 80 percent at room temperature to 36 percent when measured 
at -40°C. 



 

31 

• It was also observed that the signal amplitudes from the SDH gradually decreased as the 
RSU was exposed to the cold temperature environment for a more extended period. 

• Signal amplitudes from defects in the rail samples improved when the equipment was 
calibrated at a given cold temperature compared to the signal amplitudes obtained from 
calibrating the equipment at room temperature. 

• For TD 1, there was a 26 percent signal amplitude difference between the results from 
room temperature and cold temperature calibrations at -35°C, and a 22 percent difference 
was observed at -40°C. Similarly, a 14 percent difference in signal amplitudes was 
observed for TD 2 at -40°C. 

• In a perfect scenario, calibrating the equipment at a given cold temperature could 
minimize the loss of ultrasonic signals. 

 

Ultrasonic Signal Attenuation Measurements 
• The effective attenuation in the rail steel, α`, increased as a function of decreasing 

temperatures at a rate of 0.0005 to 0.0007 dB/mm/°C. Therefore, when calibration is 
conducted at extreme cold temperatures, some loss of sensitivity may result. The 
significance of this loss in practical terms needs further study. 

• The rail material absorbs more acoustic energy as the temperature decreases. 
• Signal amplitudes from discontinuities would be increased at lower temperatures if the 

calibration were conducted at room temperature. However, this was true only for the 
experiments conducted using ethanol as a couplant. In real-world applications, the use of 
ethanol as a couplant is not possible. 

• The longitudinal wave velocities in the two specimens were very similar, with only a 
difference of <0.1 percent. The velocity increased as temperatures were decreased at an 
average rate of 0.65 m/sec/°C. 

• Longitudinal wave velocities in the ethanol bath also increased as the temperature 
decreased at a rate of 4.1832 m/sec/°C. 

 

6 GUIDELINES FOR COLD TEMPERATURE TESTING 
The following guidelines could be helpful while conducting ultrasonic inspection in extreme cold 
weather conditions: 

• The equipment to be used for ultrasonic inspection of rails needs to be calibrated at the 
actual inspection temperature using a proper calibration rail standard as described in the 
American Railway Engineering and Maintenance-of-Way Association (AREMA) Manual 
Chapter 4 - Rail. To further improve the resolution, sensitivity, signal-to-noise ratio 
(SNR), penetration depth, ultrasonic beam spread and beam width, the equipment can be 
calibrated using a smaller size (3 mm [0.12 inch] or smaller) SDHs. 

• The couplant used for ultrasonic inspection in cold weather should not be water based. 
An anti-freezing agent such as propylene glycol should be added to the couplant. Based 



 

32 

on MxV Rail’s previous experience with cold weather testing, a 50/50 mixture of 
propylene glycol and water is recommended to prevent the couplant from freezing. 

• If possible, minimize the exposure time of the ultrasonic test equipment to extreme cold 
temperatures. We observed that the signal amplitude from the defect decreases gradually 
with the increase in exposure time of the test equipment to the cold temperature. 

• If conducting a handheld ultrasonic inspection, avoid exposing the ultrasonic transducer 
and flaw detector to temperatures below the minimum operating temperature 
recommended by the manufacturer for an extended period to avoid device failure. 

• Ice layers accumulated on top of the rail surface could affect the defect detection 
capabilities of the ultrasonic testing method to some degree. Removing the icy top layer 
from the rail surface is recommended to avoid ultrasonic signal loss. Mechanisms to melt 
the top ice layer from the rail surface could be employed. 

• Inspection vehicles should be modified to include ice-melting mechanisms and keep the 
RSU warm. 

 

  



 

33 

REFERENCES 
1. Poudel, A., Shrestha, S., Lindeman, B., and Washer, G., 2022, “Ultrasonic Rail Flaw 

Testing Parameters in Extreme Cold Temperature,” Proprietary report for Transport Canada, 
P-22-018, Pueblo, CO: MxV Rail. 

2. Garcia, G. and Jeong, D. Y., 2003, “Rail Defect Growth Under Heavy Axle Loads,” 
Railway Track and Structures, Simmons-Boardman Publishing Corporation, New York, 
NY, pp. 17–19. 

3.  Orringer, O., Morris, J. M., and Jeong, D. Y., 1986, “Detail Fracture Growth in Rails: Test 
Results, ” Theoretical and Applied Fracture Mechanics, 5(2), pp. 63–95. 

4.  Banerjee, A. and Morrison, K., 2019, “Fatigue Crack Growth Rate Properties of Rail Steels 
and Their Influence on Rail Life,” Technology Digest, TD19-008. AAR/TTCI, Pueblo, CO. 

5. Stenström, C., Famurewa, S., Aditya, P., and Galar, D., “Impact of Cold Climate on Failures 
in Railway Infrastructure,” Proc. 2nd International Congress on Maintenance Performance 
Measurement & Management. 

6.  Ladubec, C. and Magel, E., “Winter Railroading in Canada - A Review of Track and 
Rolling Stock Challenges,” Proc. International Heavy Haul Association Conference. 

7.  Poudel, A., Lindeman, B., and Wilson, R., 2019, “Current Practices of Rail Inspection Using 
Ultrasonic Methods: A Review,” Materials Evaluation, 77(7), pp. 870–883. 

8.  Alahakoon, S., Sun, Y. Q., Spiryagin, M., and Cole, C., 2017, “Rail Flaw Detection 
Technologies for Safer, Reliable Transportation: A Review, ” Journal of Dynamic Systems, 
Measurement, and Control, 140(2). 

9. Robin, C., 2004, “Rail Flaw Detection: Overview and Needs for Future Developments, ” 
NDT & E International, 37(2), pp. 111–118. 

10.  U.S. Code of Federal Regulations, 49 CFR Part 213.240, 2020, “Continuous Rail Testing, ”, 
Federal Railroad Administation, ed., USDOT, Washington DC. 

11.  U.S. Code of Federal Regulations, 49 CFR Part 213.113, 2020, “Defective Rails, ”, Federal 
Railroad Administation, ed., USDOT, Washington DC. 

 


	Cover
	Executive Summary
	Table of Contents
	List of Figures
	List of Tables
	1 Introduction
	1.1 Background
	1.2 Ultrasonic Rail Flaw Testing

	2 Project Objective and Scope
	3 Experimental Setups
	3.1 Rail Samples
	3.2 UT Equipment
	3.3 UT Test Setups
	3.4 UT Calibration
	3.5 Testing in Environmental Chamber (Cold room)
	3.6 Ultrasonic Signal Attenuation Measurement
	3.6.1 Equipment
	3.6.1.1 Immersion Chiller Bath
	3.6.1.2 Ultrasonic Transducer
	3.6.1.3 Delay Lines

	3.6.2 Test Specimen
	3.6.3 Methodology
	3.6.3.1 Longitudinal Wave Velocity Measurements in Couplant
	3.6.3.2 Shear Wave Velocity in Steel
	3.6.3.3 Longitudinal Wave Velocity in Steel
	3.6.3.4 Timing
	3.6.3.5 Amplitude Measurements



	4 Results
	4.1 Ultrasonic Rail Flaw Testing
	4.1.1 Handheld Flaw Detector
	4.1.2 Walking Stick Flaw Detector

	4.2 Ultrasonic Signal Attenuation Measurements
	4.2.1 Velocity Measurements


	4.2.2 Attenuation
	5 Conclusions
	6 Guidelines for cold temperature testing
	References