Predictive maintenance: future gazing04 May 2020

Seven years ago Volvo took its first tentative steps towards predicting when a truck component might fail. Today, thanks to the data that it is collecting from ‘connected trucks,’ Volvo is boosting its customers’ uptime, writes Brian Weatherley

Readers who worry about hidden surveillance should look away now. But those operators that are keen to eliminate unplanned downtime should read on. The fact is, there’s nothing ‘new’ about remote vehicle conditioning monitoring (RVCM) or predictive maintenance – the ability to track a vehicle in real time, flagging up in advance when a component might fail, thereby allowing it to be replaced before it causes an expensive breakdown.

Volvo can justifiably claim to be a pioneer of truck monitoring, having originally created early predictive maintenance algorithms based on the five-year analysis of the service history data of over 3,500 Euro IV and V customer vehicles built-between 2007-2008. Indeed, the Swedish truck-maker has been ‘connecting’ to its trucks since 2013, though initially technicians could only remotely monitor four, albeit important, components: brake pads, batteries, clutches and air dryer filters on FH, FM and FMX models.

It’s come a long way since then. Today, Volvo’s Truck Monitoring Centre (TMC), the organisation developing its connected service, is an integral part of its Uptime Centre, based at the manufacturer’s Ghent assembly plant in Belgium. There are now 60,000 monitored Volvo trucks operating across 24 European countries, with that figure expected to hit 100,000 by the end next year. Of that total, more than a third, 23,000, are ‘actively monitored’ – of which some 4,000 of those are in the UK.


The principle behind truck monitoring is obvious: prevention is better than cure. This philosophy undoubtedly guides Pieter Gheeraert, remote monitoring services manager at Ghent, who says his goal within the Uptime Centre is to change the whole process of vehicle utilisation and breakdown support from reacting to problems to anticipating them. “For years we’ve been waiting for a vehicle to break down, or for a dealer to have an issue to call us here. Three years ago we started thinking, ‘how can we leverage all the technical and customer data that resides here into new services? How can we reduce unplanned stops?’”

Since 2016, Volvo’s TMC team has been crunching some very big numbers, capturing vehicle data from Volvo Action Service (the manufacturer’s roadside breakdown service), along with the service records of those connected customer vehicles built since 2015 and covered by its ‘gold standard’ repair and maintenance contract. Gheeraert explains: “We targeted VAS cases – but not to make another part of the Volvo Uptime Centre obsolete! It’s to get a complete picture of our customers. I still believe we’ll need a breakdown support service, but we want to complement that with proactive services.”

The first step in that journey is real-time data capture: “We’re looking for trigger-based patterns on those vehicles,” says Gheeraert. “When something occurs on them, the information ends up in our system. We’re talking about millions and billions of ‘data points’ we’re gathering.” The next step is to marry up that data with information generated by a vehicle’s own on-board diagnostics, utilising Volvo’s remote diagnostics system to identify and detect faults before they cause a problem.

“Because vehicles ‘talk’ all the time, it’s our job to listen carefully and filter out the noise to really pinpoint what’s important. Otherwise you end up trying to fix something that’s not a problem,” stresses Gheeraert. An example of that filtering process might be the ability to recognise a spurious fault code for Lane Keeping Assist caused by a build-up of snow obscuring the system’s sensors, rather than a defective component.

However, the key task is reading all the data pouring into Ghent and translating that into requests for actions that are forwarded to an operator’s home dealer workshop. They can then speak to the customer, advising them of a potential fault and booking time to have it rectified. Gheeraert describes it as building ‘intelligence’ into the service.


Thanks to ever-increasing computer power, handling all the incoming data has become progressively easier, reports Gheeraert: “When we started three years ago, it took two guys one day to process the component data on 600 vehicles. We can now do multiple components on 60,000 vehicles in 30 minutes. It runs through a computer and is [updated] in real-time.” Today, 12 full-time people at Ghent collate and analyse the data from the monitored fleet.

Being able to monitor condition in real-time – and not only recognise when an individual component might fail, but also spot the potential broader trends in its overall life-cycle – places Volvo in a very powerful position when it comes to tackling unplanned downtime. And it’s no longer individual components that are being tracked. Gheeraert reports that the TMC team is now applying a more holistic approach. “Three years ago we looked at four components. Today we’re moving more and more away from individual component monitoring to looking at and mastering entire systems– like aftertreatment and fuel systems.”

Further refining of the predictive algorithms will, for example, allow Ghent to go beyond a simple ‘this truck has an air leak’ remote diagnosis. Rather, based on an analysis of the operating cycle of the vehicle’s compressor, it can determine whether the fault lies with the truck or the trailer it’s pulling – saving workshop time trying to isolate a leak.

“And it often is the trailer!” reckons Gheeraert. “If the dealer signed off the truck in its service with no air leaks, we’ll know that and can pinpoint the trailer.”


As the use of electronics on a truck has increased exponentially, so the nature of breakdowns has also changed. At one time, most VAS call-outs involved mechanical rather than electrical problems. As the importance of sensors has increased, that is no longer the case, Gheeraert confirms. He says: “Twenty years ago electronic component failure was very low, maybe 3% at most. Now it’s more two-thirds mechanical and a third electronic or software issues. And that shift will continue.”

An oft-repeated complaint from operators is that faulty sensors can trigger a fault code which in some cases initiates a breakdown call from a driver unwilling to ignore a red dash alarm light for fear of the damage that might be done. “It happens”, admits Gheeraert. Fortunately, the TMC team is already developing ways to ‘sense’ the sensor. “It’s the second stage of the service,” he promises, adding “We now not only have a model that says, ‘Your battery is broken,’ but where we can now say, ‘No, it’s the battery sensor which is broken’, and it’s an action case for that sensor that’s launched.”

One of skills required within the TMC team is being able to spot the difference between a small number of random, unrelated breakdowns, and potential trends involving many more trucks. This what Gheeraert calls forensic research, and he likens it to TV crime dramas. “It starts with a dead body. Then, if we start to get 10 similar dead bodies, we have a very thin pattern. That’s enough to give it over to the data guys who will find a true pattern in there; then it’s turning that ‘reactive pattern’ into actual numbers of vehicles that are currently driving around.”

Meanwhile, the ultimate goal is to create a truly dynamic system capable of predicting component replacement cycles based on actual, real-time individual vehicle usage, rather than a blanket ‘generic’ solution based on the wear rate of a particular component derived from the analysis of thousands of vehicles with a broadly similar operating profile. That discrimination between wear and usage will deliver more unique predictions and advice for each customer, promises Gheeraert.

Offering a timescale for repair that allows a truck to stay on the road for as long as possible, without allowing a fault to actually cause a breakdown, is clearly a fine balancing act. When the TMC team at Ghent identifies a problem, it gives the operator’s home dealer a recommended repair ‘threshold’ – typically between 21-28 days – within which the faulty component should be replaced or repaired. However, the period will vary depending on the nature of the fault; it can also be shorter.

Gheeraert explains: “We look at the fault codes, we look at the service records where the details are in common, and we make a conclusion and say this truck has an 88-90% chance of breaking down within so many days…that’s where we put the threshold; then the truck is called in.”

Alerting the dealer first is crucial, says Gheeraert. “The initial set-up was direct to the customer. However, we want to follow the Volvo strategy of going through the dealer workshops first, as they’re the touchpoint with the customer. The benefit of involving the home dealer at the start is that it can forward-plan workshop time and do the repair itself.” Once the repair finishes, information on the faulty component or system is sent back to Ghent ‘to feed the data machine’, he adds.


After remote conditioning monitoring, what comes next? Remote fixing? It’s already happening within the Volvo Group. In the US, Mack Trucks has used its GuardDog Connect system to provide remote software updates. Volvo says it will soon release a system to protect remote downloads. “It needs to be a secure world,” says Gheeraert, before adding, “It will download first on the vehicle and then on to the ECU in a controlled manner. In the future it can be done like that. So it will avoid a truck coming into the workshop again.”

Should anyone doubt the potential for RCVM and its ability to avoid breakdowns, three years ago the TMC team were challenged to show significant improvements in the uptime of those connected vehicles it monitored, based on VAS call statistics. Today, Gheeraert reports: “A vehicle with a gold service contract that is monitored has a considerable reduction in unplanned stops. The chances of an unplanned breakdown for these vehicles is 50% lower than for those non-contract vehicles we don’t monitor. Based on the 23,000 trucks currently being actively monitored, we have predicted and prevented almost 4,000 unplanned stops in the past few years.”

Along with current FH16/FH and FM/FMX models, Volvo is ultimately looking to connect FE and FL ranges too. More importantly, it has stepped up the marketing of its connected service, actively promoting it as an integral part of its gold standard service contract, rather than an add-on. How much more information will end up coming back remotely from a truck is anyone’s guess.

Nevertheless, Gheeraert is adamant real-time monitoring is something that operators do want. “It has real value. Our customers tell us that unplanned events bite them hard, because they can’t deliver their loads, which jeopardises their contracts with their customers. They want to deliver those loads, so we need to take away the unplanned downtime.”

William Dalrymple

Related Companies
Volvo Group UK Ltd

This material is protected by MA Business copyright
See Terms and Conditions.
One-off usage is permitted but bulk copying is not.
For multiple copies contact the sales team.