Unleash live Blog

Automating Railway Track Inspections

Written by Unleash live | Jun 15, 2023 12:01:25 PM
 

Despite the fact that rail continues to be one of the safest forms of transport, accidents continue to cause fatalities, environmental damage, injuries and infrastructure costs.

Statistics are hard to get, but in 2015 alone, there were 567 incidents where Hazmat cars in the US were either damaged or derailed. In February that year, a derailment in West Virginia determined to be as a result of a broken rail track released nearly 400,000 gallons of oil and caused extensive environmental damage. Hazmat releases are not an uncommon event either, with the Federal Railroad Administration office of safety analysis quoting figures similar to 2015 occurring annually in the US since 2009.

In addition to environmental costs, infrastructure damage has also remained fairly significant with approximately 3% of all incidents having a total cost of over $1m.

Many rail passengers would be surprised to hear that rail tracks are still inspected visually, by an inspector walking the track and automated and monitoring systems used for softer conditions, such as track geometry deviations.

There are also 4 processes followed to determine flaws:

  • Portable test process: Operator pushes test device at walking pace
  • Stop/start process: A vehicle detects rail flaws at low speeds (typically under 20 k/ph)
  • Chase car process: A lead test vehicle performs the flaw detection in advance of a following verification vehicle
  • Continuous test process: A high speed vehicle based system tests non stop, and data is later downloaded and analysed for a fault from a centralised location

All processes utilise human assessment to some degree of interpretation and are therefore prone to the risks of error that come with human oversight.

The list of accidents on the Federal railroad administration database is long. Check it out yourself.

Train accidents by primary cause, Jan - Dec (2017 preliminary)


Train accidents can be caused by equipment error, human factors and track and signal issues

 

As part of the Rail Safety Improvement Act of 2008, a track inspection Time study report was commissioned and presented to congress in May 2011. The survey interviewed rail track inspectors whose jobs involved inspecting tracks from a hi-rail vehicle or on foot. Survey results indicated that track inspectors frequently worked beyond their 8 hour days, worked on their rest days and on average achieved slightly less than 80 miles of track inspections a day. 

A number of recommendations were made as a result of this report, one being expanded use of automated inspection systems to supplement visual inspections.

With improved cameras, compute power, new developments in AI and better network connectivity there is significant opportunity to improve accuracy and lower cost of inspections.

Unleash live is a powerful cloud based software platform delivering real time AI algorithms to live stream video for industries with vastly dispersed infrastructure to make fast and insightful decisions.

For railroad, the benefits of leveraging automation can be visualised with an example seen below.

 
Enabling Railway Gauge AI in Unleash live

 

The implementation of automated inspection systems will reduce the likelihood of missed defects that could cause derailment and provide inspectors the opportunity to conduct more thorough visual inspections. There is also the case where inspectors are deterred from reporting faults as they themselves are given the responsibility to repair the fault.

The implementation of automation brings with it, process change and the ability for skilled inspectors to focus on reducing the rate of faults per mile, while Unleash live can support in the following ways:

  1. Visual documentation, including exact location and record keeping

  2. Real time fault assessment — faster and more accurate

  3. The ability to isolate specific areas of interest, i.e. cracks in rail, missing bolts, or geometry alignment

  4. Ability to perform AI analytics on all video captured with intelligent search to find what you’re looking for fast

  5. Live collaboration with multiple stakeholders

Take item 2 above for example, below is a graph showing the relationship between inspection speed, defect size and the number of search objects. Coupled with the complex issue of inspector fatigue, distraction and vigilance; this is where the power of AI truly is divergent from their human counterpart.

Source: US Department of Transportation, Federal Railway Administration


When presented with AI powered solutions, we notice that inspectors and their supervisors as well as head office are delighted to share additional areas where video assisted AI inspections can significantly improve the rail infrastructure to keep us and the environment safe.

Get in touch if you’d like to discuss this topic in more detail, or wish to share your ideas for AI in railroad infrastructure.

Get Started with Unleash live

Visit unleashlive.com to learn more or book a time with one of our specialist team members to learn how your business can take advantage of live streaming video and AI analytics.