Your browser is no longer supported

For the best possible experience using our website we recommend you upgrade to a newer version or another browser.

Your browser appears to have cookies disabled. For the best experience of this website, please enable cookies in your browser

We'll assume we have your consent to use cookies, for example so you won't need to log in each time you visit our site.
Learn more

How wireless wheel monitors could identify railway track faults

Perpetuum wireless sensors

Southeastern Railway has equipped its entire fleet with energy harvesting wireless sensors to predict wheel bearing failures. Now it wants to share the data generated by the technology with Network Rail to help it predict track faults.

Train wheel bearing failures used to be a big headache for Southeastern Railway’s engineering director Mark Johnson. The man who is responsible for the train operator’s entire rolling stock and some 550,000 passengers daily, used to dread a fault occurring while one of his trains was in service.

“Should it happen in the wrong place, a fault could wipe out the whole of the South East for 24 hours,” he says.

Perpetuum wireless sensors

Perpetuum wireless sensors

Configuring and cataloging the system

Because all of the wheel bearings in Johnson’s fleet sat in an enclosed system, there was minimal visual indication that a fault was about to occur. The bearings are designed to last for 2.2M.km but because of the harsh environment below the train, early-life failures were beginning to creep in at 960,000km. The consequences of failure were so significant that Johnson had no choice but take a safety-first approach to maintenance, scrapping an average of 64 bearings on a four-car train for his entire 148-train fleet at the half-life stage.

“Previously, a lot of the industry was focusing on applying sensing technology that showed you a temperature increase when the bearing was about to fail,” says Johnson. “But one of the fundamental problems with that is, once the thing gets hot, the failure has already started. Another problem was trying to retrofit something onto a train that had already been built.”

“We have been in meetings with our colleagues at Network Rail to ask what other information we can glean about the other half of the rail interface.”

Mark Johnson, Southeastern Railways

Looking to reduce the frequency with which his trains had to return to the maintenance depot and improve the reliability of his fleet, Johnson turned to a technology provider he had encountered in another application.

“We’d come across Perpetuum and the way they had developed an energy harvesting device to power computers remotely using vibration, and had a discussion with them,” says Johnson. “The brainwave came from these guys that you could combine that technology with an accelerometer to measure the acoustic vibrations and the temperature on a wheel bearing in a very small package.”

12-month trial

Perpetuum came up with a wireless energy harvesting sensor node (see box below) that could be attached to the bearing case cover on the underside of trains. Southeastern then ran a 20 unit trial over a 12 month period  to see if the technology could be used to predict wheel bearing faults.

“We wanted to build an understanding about which signals were the right ones to be looking for and how they change,” says Johnson. “We then overlaid that with what we were seeing in our physical inspections and evidence of the actual physical degradation [of the wheel bearings].”

The trial was such a success that the operator fitted it to the whole fleet.

“Now we only have to replace the bearings on condition,” says Johnson. “Since we have had the kit on board, we have not seen a single failure for bearings as a result of the bearing starting to disintegrate, and in years previously we would see six to eight.”

Perpetuum wireless sensors

Perpetuum wireless sensors

An asset manager viewing data about the condition of the wheel bearing

But Johnson thought the technology could also be used to help civil engineers involved with track maintenance.

“We realised that the wheel is only one half of the interface,” he says. “So in recent times we have been in meetings with our colleagues at Network Rail to ask what other information we can glean about the other half of the rail interface. Just because the piece of kit isn’t fixed to the solid piece of track that doesn’t move, is there anything we can pick up from our wheel sets?”

 Johnson thinks that the sensors on his trains could be employed to tell Network rail how its tracks are wearing and about the condition of its track beds.

“They’re all dynamic things and something that’s dynamic will always exert a frequency response,” he says.

Incomplete picture

Network Rail has traditionally used three to four track monitoring trains to monitor the condition of the roughly 35,500km of track that it is tasked with maintaining. Given that the monitoring trains have to cover the entire network, and strategic train lines are given priority, Johnson argues that the infrastructure owner has an incomplete picture of its network and, when things go wrong, this unfairly causes reputational damage to Southeastern.

“There’s nothing more frustrating than getting on a piece of track with a 70mph to 90mph speed limit but they’ve got a temporary speed restriction of 15 mph because there’s a broken rail or a soggy track bed,” he says.

Perpetuum wireless sensors

Perpetuum wireless sensors

Installation of the wireless sensor

“If we can start to understand what signal the track gives off when it starts to degrade, we have an opportunity to go in there before a speed restriction comes into place and make the repairs. It’s a win-win for everyone and we don’t have to tell our passengers that our train is going to be 15 minutes late.”

For Network Rail to take advantage of the 3.9GB of data that Southeastern is collecting through its wireless sensors, Johnson thinks both parties will have to come out of their silos and start to appreciate the railway and the trains operating on it as a single entity.

Systems approach

“It’s taking more of a systems approach towards the railway. The train operators are mainly interested in their wheel sets and how they optimise them; Network Rail looks at their tracks and how they can optimise them, when actually it’s two halves of the same circle,” he says. “Any change that we do to our side of the interface will impact their maintenance regimes and vice versa.”

Once such a ‘systems approach’ is adopted, Johnson thinks that Network Rail can begin to have more ‘mature’ discussions with its operators about the best way to work together.

“You could ask if it’s better for the wheel set to wear quicker and the rail head to not wear, thereby transferring more of the wear rate to the operator, or ask if the most cost effective way of doing things is to have harder wheels and replace the rails more often.”

Whether Network Rail will be expected to pay for Southeastern’s data has yet to be decided, but Johnson thinks a discussion about data rights is inevitable. “Both parties are willing to invest time and resources into the information this technology can generate, the next stage is to discuss rights,” he says. “How do we all make sure that the most is achieved?”

The sensor

The Perpetuum wireless sensor node is attached to a wheel bearing case cover on the underside of the train. Vibrations picked up by the sensors also perform a useful function in driving the energy harvesters contained in one half of the node. The other half contains a three axis accelerometer, a temperature sensor, wireless communications, electronics and capacitors.

“The sensor has a 453mHz radio transmission which takes the data off the node,” says Perpetuum commercial director Justin Southcombe. “The energy harvesting and management system that provides the electronics with a smooth power supply is the secret part of the technology. That allows the node to power itself, and when you add that to the fact that it communicates wirelessly, you realise it’s completely autonomous.”

The node collects packages of data every minute and sends them to a data concentrator situated elsewhere on the train. One data concentrator is sufficient for up to five cars. To watch a video about the technology, click on the image below.

Have your say

You must sign in to make a comment

Please remember that the submission of any material is governed by our Terms and Conditions and by submitting material you confirm your agreement to these Terms and Conditions. Please note comments made online may also be published in the print edition of New Civil Engineer. Links may be included in your comments but HTML is not permitted.

Related Jobs