Deformation Monitoring of Rail Tracks for Pipeline Crossings

Safety
According to the Transportation Safety Board of Canada, between the years 2010-2019, there were 26 employee fatalities. During a pipeline operation, the railway that is being operated on is often not shut down during the drilling, meaning that trains can pass by at any time. To address this, railway owners have implemented strict safety rules when it comes to being on the tracks. These rules can cause large logistical issues in regards to getting clearance in a timely manner. The laser scanning solution allows for very minimal/no need for a person to be on the tracks. Unlike conventional leveling, the scanning process does not require a person holding a rod on the tracks. The only reason a person would need to be on the tracks is to remove rocks/debris from the scan area.

Fewer People
When it comes to the conventional method of rod and level observation(s) on a rail track, there typically needs to be 4 employees. One observer, one rodman, and two train watchers. For laser scanning however, there is no rod being observed, and since no people are on the track there is no need for train watchers. That means only one person operating the laser scanner is needed for our solution. Each method only requires one person in the office for data processing as well.

Observation Frequency & Efficiency
Typically during a drilling operation, monitoring of the railway is needed twice daily, once in the morning and once in the afternoon. Thankfully, the laser scanning method is able to reach this requirement. In our field test from 15 scans the average time of observation was approximately 15 minutes. This is not dependent on how many points of interest there are. For example if there were 10 points or 100 points of interest in the area we scanned, both cases would only take 15 minutes. Whereas with leveling, we averaged a time of 2.7 minutes/point observed. This means 10 points would take 27 minutes, 100 points would take 4.5 hours.

Observation Complexity
The method of laser scanning is actually quite straight-forward. Trimble has very simple instructions for a basic scan:
1) Select measure/scanning
2) Select the scan density
3) Enter the approximate target distance
4) Enter the required field of view
5) Capture a panorama image (select panorama checkbox)
6) Tap start
These are 6 steps that anyone can perform to generate a basic scan.

The Trimble SX10 is a combination of a laser scanner and a total station. Through our research, we are unaware of any other group that has attempted to use a laser scanning total station for a purpose like this.

One major benefit of using a hybrid device like the SX10 is it cuts down data processing time as points are automatically registered upon uploading them to a data processing software and do not need to be geo-referenced. In the field, the SX10 uses its total station functionality to be in orientation with known point(s) during the scan. Then when it is time for data processing, scans for each epoch are also in that orientation making the comparison process much simpler.

Areas Monitored Instead of Points
In a conventional levelling loop, points of interest along the track need to be set up in a control network such that those locations can be monitored. The location of these points needs to be stable throughout the entirety of the drilling process which could take weeks. Setting up those points can be tedious, and their spacing may not be adequate. For example, if the points are too spread out deformation may happen between the points but would never be observed.

The laser scanning method however creates entire surfaces for comparison. These surfaces can be generated for an entire section of rail and a cut & fill grid comparison can be performed as shown on the left. This means deformations are more likely to be noticed at locations that aren’t specified by a point like in the leveling network.

Field Work Simplicity
One of the most annoying things for a field crew is a complicated system/device. Thankfully, the use of a laser scanning total station provides observations with ease. Firstly, the total station portion of the SX10 is not far removed from a regular total station. This is the same with the scanning portion in comparison to regular laser scanners. There is not a steep learning curve and most manufacturers provide in-depth documentation, videos, and tutorials for all skill levels.

One of the ways we validated our solution was a dynamic field test. This was performed on an acreage in Okotoks. To do this a home built railway model was constructed using a deck railing and strips of angle iron, as we needed a device to deform the rails.
The railings were 3 metres long and 0.8 metres wide. Two pieces of 1.5” x 1.5” x 0.125” angle iron were used to simulate the rails on an actual track as the laser scanner may return different results when scanning metal as opposed to wood. The angle iron was affixed to the tracks using wood screws through holes that were drilled in the angle iron. Six drywall screws were screwed into the track wood and were left approximately 1.5 cm above the surface of the wood so that a levelling rod could be held on them. These drywall screws are the rail points (R1-R6) that were observed during the levelling loops on the tracks. The image on the right shows these points as red marks.

A control network was set up in the cul-de-sac and ditch west of Elliott’s acreage. Five control points were established. The control points that were on the asphalt (1 and 2) were 2” concrete nails with flagging that were left above the asphalt approximately 1 cm to ensure that a levelling rod could be used on them. The control points in the ditch were ~10” long pieces of rebar with rounded tops and punch marks on the top to make it easier to centre the equipment on the points, which were left above the ground by a few centimetres for the same reason as the concrete nails.

Points 1-6 are the network control points, whereas points R1-R6 are the rail model points.

Conventional Levelling
In total, 12 levelling loops were performed on the rail model in groups of 6. Of the 6, the first 3 before deformation were adjusted, averaged, and used as a baseline. After deformation, 3 loops were compared to the baseline and used to find the mean deformation of points R1-R6. This test was performed as a comparison for the laser scanning results. We know the levelling method works, as it is the current professional standard for this type of deformation monitoring; and we can see how the scanning results match up.

Laser Scanning
Two types of test were performed with the laser scanner: a close test, and a far test. For the close test, the scanner was set up on point 2 about 9 metres away from the model. Coordinates of points 1, 2, 4 were entered into the data collector, point 4 was back-sighted, and point 1 was used as a check shot. The scans were then performed, 3 before deformation, and 3 after. In the far test the scanner was set up on point 3, point 1 was the backsight, and point 4 was the check shot. These scans proved to not be useful as the distance was about 114 metres away,  the scanner could really only see the north face of the tracks and only record about 21,000 points as opposed to the close tests 550,000 points.

This static test allowed us to compare multiple scans to a reference scan. This was done by creating a cut-fill map in Trimble Business Center. The numbers on the grid show the elevation differences between a monitoring scan and the reference scan. Our main takeaways from the static test is that: Other than a few outliers, the cuts and fills are small enough to consider no deformation being detected. The scanning method can detect if the rails remain stationary to an acceptable margin considering the scanner’s manufacture precision.