Railway maintenance decisions depend on reliable inspection data. If measurements collected today cannot be confidently compared to measurements collected six months from now, it becomes difficult to determine whether a turnout's condition is actually changing or whether the difference is simply the result of the inspection process itself.
This challenge highlights the importance of repeatability in automated turnout inspection. For railways pursuing data-driven maintenance strategies, repeatable measurements are essential for tracking deterioration trends, prioritizing maintenance activities, and allocating resources effectively.
To evaluate whether automated turnout inspection can deliver the level of consistency required for long-term asset management, the National Research Council of Canada (NRC), CEFRAIL, VIA Rail, and Pavemetrics®, an Eddyfi Automated Vision Systems solution, conducted a field validation project under real operating conditions.
Test vehicle and field crew used during the turnout inspection validation project conducted with NRC, CEFRAIL, VIA Rail, and Pavemetrics in Glen Robertson, Ontario
Repeatability refers to the ability to obtain consistent measurements when inspecting the same asset multiple times.
For turnout inspections, repeatability is particularly important because maintenance decisions are often based on trends rather than individual measurements. Asset managers need confidence that changes observed in the data reflect actual asset deterioration and not variability introduced by the inspection method.
Without repeatable measurements, it becomes difficult to establish reliable condition baselines, monitor degradation over time, or evaluate the effectiveness of maintenance interventions.
Traditional turnout inspections typically rely on a combination of visual observations and manual measurements taken at selected locations.
While these methods remain widely used, they present several challenges for repeatability.
During the validation project conducted with NRC, CEFRAIL, and VIA Rail, we observed several challenges associated with manual turnout measurements. Because inspectors must manually locate measurement points during each inspection, it is difficult to reproduce the exact same measurement location over time. In addition, manual tools provide a limited number of measurements and results can vary depending on the inspection process.
As a result, comparing measurements collected months or years apart can be difficult, especially when trying to identify subtle changes in asset condition.
In October 2023, the National Research Council of Canada (NRC), CEFRAIL, VIA Rail, and the Pavemetrics team conducted a field validation project on VIA Rail's Alexandria Subdivision near Glen Robertson, Ontario.
VIA Rail's Alexandria Subdivision near Glen Robertson, Ontario
The objective was straightforward: evaluate the repeatability of measurements collected using the LRAIL™ Automated Turnout Inspection Module.
To assess consistency, the project team performed multiple scans through the turnout under real railway operating conditions.
The testing program included:
This methodology allowed the team to evaluate whether measurements remained consistent over time and whether scan direction affected measurement quality.
The automated inspection system generated measurements for several critical turnout parameters, including:
Using high-resolution 3D laser scanning and automated component recognition, the system continuously evaluated turnout geometry and component condition throughout the inspection process.
One of the key objectives of the validation project was to determine whether measurement quality remained consistent regardless of scan direction.
Measurements collected during forward scans were compared against measurements collected during reverse scans of the same turnout.
The results demonstrated that the automated inspection process produced consistent measurements independent of scanning direction. This finding is important because it confirms that the inspection results are not dependent on how the vehicle traverses the turnout.
For railways implementing network-wide inspection programs, this consistency supports more reliable condition monitoring and long-term trend analysis.
Automated turnout inspection can provide repeatable measurements by combining 3D laser scanning with automated measurement extraction.
During the VIA Rail validation project, the system generated more than 4,100 measurements on a single turnout and demonstrated consistent results across repeated forward and reverse scans. This level of measurement density and consistency provides railways with a reliable foundation for condition monitoring and maintenance planning.
During the validation project, more than 4,100 automated measurements were generated on a single turnout, covering both switch and crossing components
Because measurements are generated using the same process during every inspection, railways can compare data collected over time with greater confidence and reduce dependence on subjective observations.
The value of repeatability extends far beyond the inspection itself.
When railways can trust that measurements are consistent from one inspection cycle to the next, they can:
Rather than relying solely on inspection snapshots, maintenance teams gain access to a consistent dataset that supports more informed decision-making.
The validation project demonstrated how automated turnout inspection can deliver repeatable measurements under real operating conditions.
By combining 3D laser scanning with automated measurement extraction, the system generated thousands of objective measurements while maintaining consistency across repeated inspections and scan directions.
For railways looking to strengthen condition monitoring programs and support data-driven maintenance planning, repeatability is not simply a technical requirement, it is the foundation for making confident maintenance decisions.
Discover how the LRAIL™ Automated Turnout Inspection Module helps railways collect objective, repeatable turnout measurements and support data-driven maintenance planning.