Adhere & V/T SIC: Trends in rail management

In December 2023, there was much adverse criticism of the railway when there were several incidents on the Great Western Main Line caused by rail breaks, compounded by a major delay caused by damage to the overhead line equipment. Against this background, Rail Engineer was keen to hear from Brian Whitney, Network Rail’s Engineering Expert for Track and Switches & Crossings at the recent Vehicle/Track System Interface Committee (V/T SIC) seminar. He has frequently spoken about the number of rail breaks and defects in GB’s approximately 64,000 km of rail.

For interested readers, his previous report on rail breaks was in Issue 197 (July/August 2022), while in Issue 202 (May/June 2023), Brian covered how hot weather issues affect the track. In the same issue, Bob Hazell described many of the tools and techniques used to identify and fix rail defects.

A record low

As always, things move on and Brian was able to report that the number of rail breaks in the period April 2023 to March 2024 was likely to be around 60, a record low despite the adverse publicity from the incidents on the Great Western mentioned above. Analysis of broken rails over the last 10 years shows that November, December, and January are the peak months, and generally the numbers are small in summer months.

At the time of his report there had been 56 broken rails allocated to 29 different cause codes.

The most numerous were 17 caused by rail foot corrosion leading to a stress raiser type crack leading to the fracture.

The most numerous of the other causes were three in the year but there was just one break in most categories. In context, there are typically over 30,000 defects reported per year of which an average of around 50 per month are serious, requiring a block to traffic or a severe speed limit.

Brian demonstrated why rail foot faults are not always revealed by ultrasonic inspection. A rail pad may look satisfactory but may have become worn under the rail, trapping water, leading to corrosion. Pitting on the rail foot might provide a confused test trace and the situation might be worse if the rail pad, whilst appearing intact, is fully worn under the rails.

Generally, though, it is high loads locally that finally cause the break, but more likely in winter where continuously welded rails are under tension due to thermal forces. Brian illustrated this where rails broke adjacent to dipped welds. As wheels pass over the dip they become slightly unloaded, following which the wheel rail vertical force increases rapidly.

Another reason, often on switches and crossings, is where there are dips or voids in the track support. In these locations, the forces arising from poor support are compounded by high peak loads as the wheels negotiate the discontinuities of the crossing. Brian showed an example of a vertical crack in a crossing originating from a stress raiser caused by these loads and the sudden change in section, illustrating how a minor redesign in this area might reduce the risk of cracking.

Identifying precursors

As an experienced track engineer, Brian often investigates rail breaks using data from various sources and then looks at photographs which clearly show a rail break precursor, such as the approach to a bridge or a site of pumping ballast. He has talked fondly of the day when data from different sources might be combined with photographic information to alert maintainers to the circumstances that might lead to a rail break, so that remedial action might be taken. He was mildly optimistic that the day is almost with us. The many measurements taken by the high-speed measurement train can now be accurately aligned so that experts can see how track is changing, suitably colour coded to show the good, the bad, and the ugly.

However, Brian had always been keen to understand the context – the combination of features, all within specification, that together might lead to a fault. As is often said, a picture is worth a thousand words, and the ability to link images to the data has the potential to be transformational. Not only that but artificial intelligence/machine learning is starting to be used to identify track features and rail defects.

As anyone who has carried out on site track inspections will tell you, it’s easy to miss things, especially at night. It’s even worse when watching video streams of track from a comfortable armchair. Brian reported that One Big Circle’s Automated Intelligent Video Review (AIVR) has been trained to identify track features and has accurately identified all the rail welds on a large sample of track. On a particular ultrasonic shift, it has also identified over 550 squats.

There was more. Brian illustrated the power of AIVR with an image of an area of light-coloured ballast, suggesting that there was lack of support.

This was obvious to the naked eye, but what wasn’t obvious was shown on a heat seeking camera which showed that an electrical rail bond across a rail clamp was exhibiting hot spots that needed investigation (see below).

An AI future

With the data from track recording, ultrasonics, and from video all aligned to time and location, Brian was optimistic that soon artificial intelligence will be able identify precursors to rail breaks. The challenge will be to determine what to do prior to failure as there are sure to be far more sites identified than actually break, at least in the early days.

It looks as though there is light at the end of the tunnel and that huge data sets might be capable of being turned into useful information. At last, the image of Brian Whitney, ‘burning the midnight oil’, surrounded by monitors showing data streams, might be consigned to history, although Rail Engineer is sure there will be new problems for him to solve.

Image credit: Network Rail