Autonomous and As-A-Service Models Will Rely on Predictive Maintenance

Data will drive business models of next generation commercial vehicle suppliers

Predictive maintenance has been used in some sectors, aircraft for example, for decades. The concept is simple; using data from real-world activity, predictions can be made as to when parts will need servicing or replacing. As both sensor technologies and data analytics have improved, the accuracy, scope and scale of what can be accomplished with predictive maintenance have all expanded. Now dramatically falling cost per query for analytics combined with evolution of business models is making predictive maintenance not just a reality, but a necessity in many sectors. The commercial vehicles sector is one example of where technical and business model innovation are driving the need and the opportunity for data-driven predictive maintenance to dramatically improve business outcomes. 

AUTONOMOUS OPERATIONS ARE HERE – IN PLACES
Amidst all the hype and inflated expectations of autonomous driving and important shift is occurring. Although it could be years (some say decades) before fully autonomous vehicles are common on our roads, they are already in regular use in some industrial scenarios. For example, many mines, harbours, warehouses and intralogistics operations already incorporate some autonomous elements. As well as freeing human drivers from mundane and routine task like shuttling goods around site these autonomous operations collect vast amounts of data. These data are vital for variety of reasons. Clearly it takes lots of data from onboard sensors to enable autonomous operation in the first place. Vehicles must be able to calculate their position precisely in real time, they need to report status of multiple indicators, from loading weight to engine temperature, wheel odometry to tyre pressures, and they need to accurately measure their distance and speed relative to the world around them. 

These data are also vital to understand how things go wrong when errors or incidents occur. Deep analysis of vehicle data, as well as external data on anything from weather conditions to changes to route or physical infrastructure. Analytical modelling of multiple data sets will help determine root cause of accidents and suggest potential mitigations to prevent the same thing happening again. Spotting indicative events that suggest a failure is imminent allows action to be taken in advance to avoid it. Vehicles can be taken out of service at a convenient time for preventative maintenance that lowers costs and avoids expensive breakdowns. 

Increasingly, these data are important for audit and reporting as well as liability purposes. Instead of subjective statements from witnesses, data from vehicles will increasingly be used to determine who, or what, was at fault in accidents and incidents. Similarly, using that data to predict potential incidents allows maintenance interventions that avoid them and potentially costly legal actions.  

TAKING THE GUESSWORK OUT OF AS-A-SERVICE
At the same time, forward looking suppliers of commercial vehicles are relying on data to shift to ‘as-a-service’ models. Customers want to move beyond traditional leasing agreement under which they pay a flat monthly ‘rental’ for a piece of equipment irrespective of how much it is used. Instead, they want to pay for the service – ie the work done by a vehicle. This shifts the risk to the equipment provider as they become responsibility for a guaranteed level of service and are only paid for the time equipment is used. Rolls-Royce coined the phrase ‘Power by the hour’ in the 1960’s but since then sensor and data processing capability has made it an option for a wide range of commercial vehicle businesses. 

These approaches have the potential to be more flexible and cost effective for customers whilst allowing commercial vehicle leasers to maximise the return from their investments. But without accurate, granular data, service providers are just guessing when it comes to setting prices and service level agreements for these agile contracts. And breakdowns shift from being an inconvenience for customers to immediate loss of revenue for the service provider. Adequate data is necessary to price contracts, accurately invoice and plan maintenance to maximise availability. 

THE NEED FOR A DATA FABRIC
The common thread to both these emerging trends is the need for complete, accurate data at speed and scale. Individual systems that collect or monitor data from single sources (ie a single vehicle) in isolation are not valuable in planning predictive and preventative maintenance. It is the interaction of vehicle, task, environment and numerous other factors that are predictive of potential incidents that could affect its operation. Collecting, integrating and analysing data on all the variables is essential to the effective, efficient and profitable autonomous and as-a-service commercial vehicle operations. 

Creating a common data fabric that connects every aspect of commercial vehicle use, from telematics and operations to service teams, parts and logistics chains allows the automation of predictive maintenance. Data must be combined across every aspect of the organisation so that issues are before they become critical and that service schedules can be booked, parts ordered and delivered, and mechanics’ time utilised effectively to minimise equipment downtime.

Providing early warning, and the ability to take the vehicle out of service before breakdowns protects revenue. It can also pivot maintenance from necessary evil to value-adding service. By providing information to end-user, whether in house fleet managers or end customers, maintenance can be planned to avoid downtime. It can be the heart of reimagined service provision as we’ll see in next blog.