Rail geometry information is a mainstay of the railway industry. Gauge, alignment and cross-level measurements are used to monitor and model track safety under varying load conditions, curvatures and gauge widths.
While rail geometry measurements typically require a separate dedicated inspection system, Railmetrics™ LRAIL™ is able to perform these tasks simultaneous with other important tasks such as clip and spike inspection, ballast inspection, cross tie inspection and others.
The LRAIL uses Artificial Intelligence to first correct 3D scans for measurement vehicle motion using built-in IMUs. Then 3D data is analyzed in order to detect rails, locate the top of rails for cross-level measurements and the gauge measurement point (5/8″ below top of rail) for both gauge and rail alignment measurements. Chord lengths are customizable in order to match the customer’s specific standard.
Rail heads are regularly exposed to heavy loads from long train consists as well as localized stress associated with both acceleration and breaking. These stresses deform the surface of the rails resulting in head and side wear.
While rail head wear typically requires a separate dedicated inspection system, Railmetrics’ LRAIL is able to perform this task simultaneous with other important tasks such as clip and spike inspection, ballast inspection and cross tie inspection for example.
The LRAIL algorithms analyze 3D data in order to measure top, side and corner rail wear. Rail wear measurements can be made for both standard rail profile as well as grooved (metro) rail.
Additionally, through the LRAIL’s Change Detection algorithms, changes in wear can be automatically detected and reported pointing track maintenance crews to the portions of the network that are most in need of attention.
Spikes play a crucial part in the safe and proper functioning of railway fastening systems as they are used to secure rails and base-plates to cross ties (sleepers). Spikes come in a variety of shapes and types with some of the most common being cut spikes, dog spikes and chair screws.
Missing, broken and high spikes degrade the overall performance of the fastening system and thus their detection is an important railway inspection task.
However, spike inspection can be a challenging task for traditional 2D machine vision systems due to their lack of 3D data. 3D data is needed to detect high spikes and often to correctly identify spike heads which are difficult to separate from ballast particles due to their similar size and shape in 2D images.
Railmetrics’ LRAIL simplifies the spike inspection task by using Artificial Intelligence to analyze both 2D and 3D data in order to more accurately detect present and missing spikes, spiking pattern, high spikes and broken spikes.
Additionally, through the LRAIL’s Change Detection algorithms, changes in spike inventory and condition can be automatically detected and reported pointing track maintenance crews to the portions of the network that are most in need of attention.
Clips play a crucial part in the safe and proper functioning of railway fastening systems. Clips are responsible for fastening rails to the underlying base-plate which ensures that rails are securely connected to underlying cross ties (sleepers).
Missing, loose and damaged clips degrade the overall performance of the fastening system and thus their detection is an important railway inspection task. However, clip inspection can be a challenging task for traditional 2D machine vision systems as they come in a wide variety of shapes, sizes and designs thus making them difficult to detect from 2D data alone.
Railmetrics’ LRAIL simplifies the clip inspection task by using Artificial Intelligence to analyze both 2D and 3D data in order to more accurately detect both present and missing clips (and components), clip type, loose clips, damaged clips and covered clips.
Additionally, through the LRAIL’s Change Detection algorithms, changes in clip inventory and conditions can be automatically detected and reported pointing track maintenance crews to the portions of the network that are most in need of attention.
Track ballast plays an important role in the proper functioning, and ultimately safety, of a railroad track. Ballast is used to transmit and distribute the load of the track and rolling equipment to the subgrade, restrain the track laterally, longitudinally and vertically, facilitate drainage of water, maintain proper cross level, surface, and elevation, as well as inhibit the growth of vegetation.
However, due to weather, train loading, and differences in ballast material quality, ballast can degrade in condition and become less effective over time. For example, heavy rains can result in ballast being washed away and leaving railway ties (sleepers) exposed and more likely to skew.
Ballast can also become fouled due to the fracture and abrasion of ballast particles, infiltration from underlying layers, and spillage from containers in transport. Fouling is often exacerbated by dry-wet cycles and fouled ballast can result in poor drainage and ultimately track surface deviation due to increased layer stiffness.
However ballast inspection is often a very subjective process due to the lack of a formal definition for when ballast is considered fouled and the often manual process of inspection.
Railmetrics’ LRAIL simplifies the task of ballast inspection by using Artificial Intelligence to analyze both 2D and 3D data in order to precisely and objectively measure ballast height and ballast fouling. Ballast height is automatically measured at the end of ties as well as in the crib area and locations with too little or too much (based on user-definitions) can be flagged. Ballast fouling in the gauge side as well as in the field side can also be automatically detected and quantified objectively. Changes with regard to ballast height and fouling level between repeat runs can be automatically detected and reported.
Railroad cross-ties (sleepers) play an important role in the function of a railroad by transferring loads to the track ballast and sub-grade, by holding rails upright, and by helping to maintain proper gauge thus preventing derailments.
However, wooden (timber) and concrete tie systems degrade over time and thus require efficient and objective inspections in order to ensure their integrity. Wooden ties can develop increasingly large splits, become rotten and fail to properly support the track. Concrete ties can develop both longitudinal and transverse cracking, and chipping from derailments and tamping operations. As well, both wooden and concrete ties can become skewed (no longer perpendicular to the rails).
However, tie inspection can be a challenging task for traditional 2D machine vision systems as they are limited to tie condition assessment based on merely the appearance of the tie condition as opposed to 3D measurements of the length, width and depth of splits, cracks and chips. Additionally, the surface of ties can become discolored over time reducing the already limited condition information available from 2D systems even further.
Railmetrics’ LRAIL simplifies the tie inspection task by using Artificial Intelligence to analyze both 2D and 3D data in order to more accurately detect and report tie count, position, material, surface cover, tie skew, wooden tie grade and concrete tie grade.
Additionally, through the LRAIL’s Change Detection algorithms, changes in tie grading and skew between repeat runs can be automatically detected and reported pointing track maintenance crews to the portions of the network that are most in need of attention.
Until recently, the railway industry has tended to focus on the measurement of discrete performance parameters in order to determine if a maintenance or safety threshold has been met.
However, this approach is limited to flagging sections of track only once they meet a defined minimum threshold despite the fact that the same section of track may be steadily deteriorating up until that point.
Railmetrics’ LRAIL turns this process on its end by going beyond discrete measurements to the detection of change in key regions of interest along the track. This method allows track problems to be detected as they develop; well in advance of meeting maintenance or safety thresholds which could result in a closure or speed warning.
The LRAIL’s Artificial Intelligence automatically aligns repeat-runs and then analyzes both 2D and 3D data in order to detect changes between them.
Both positive changes, due to the performance of maintenance, as well as negative can be detected. Example changes that can be detected include:
Clip inventory changes including the number of missing or broken clips
Spike inventory changes including the number of missing spikes, the number of high spikes, the number of damaged spikes and changes in spike patterns
Wooden tie grading and skew angles
Concrete tie grading and skew angles
Ballast height and fouling level
Joint gap and joint bar bolting
Changes in gauge, cross level and alignment
Changes in rail surface condition
Get in touch with us to get further information about the variety of railway asset properties and defects our technology can automatically inspect.
Our website uses cookies to provide you with the best possible online experience by securing your navigation, improving the operation of our website, measuring our audience, providing you with relevant and personalized marketing content. You have full control over whether or not cookies are enabled. You can accept cookies by clicking the "Accept" button. You can also refuse all unnecessary cookies by cliking on X.