Predictive maintenance involves complicated algorithms, artificial intelligence, and sensors, but the purpose is pretty simple: fix small problems before they become big ones. Let's examine a common way a driver for heavy-haul fleet United Road might benefit from using the technology.
Maintenance isn't at the forefront of drivers' minds as their big rigs barrel down the highway. They are more concerned about turning wheels and earning money than worrying about maintenance issues. Suddenly, dispatch alerted a driver to a potential NOx sensor failure and to come in for service as soon as possible.
That’s odd, he thought. Nothing on the instrument panel alerted the driver to a problem.
In this scenario, scheduled maintenance wasn’t set to occur for three weeks, but the driver knows better than to argue. After all, the fleet did integrate an AI-enabled predictive maintenance software that he heard has helped other drivers stay on the road longer. After delivery, the driver brought the truck into the shop. Technicians confirmed the looming failure of the NOx sensor 21 days before the truck’s regular service call.
It was another win for the fleet's predictive maintenance. Problems get fixed before they become problems. As a result, no road call, no expensive tow fees, no breakdown docked driver pay, no late delivery, no disappointed customer.
That’s the argument for predictive capabilities and why fleet maintenance programs of all sizes are infusing their preventive maintenance with the technology. However, for fleet teams that want to get into predictive maintenance, a few things can block progress. Fortunately, for every obstacle, there is a workaround.
Data is essential but can be overwhelming
Predictive maintenance centers on developing insights based on real-time and historical data collected from the trucks. There is no shortage of data sources, including telematics, fluids, and components. There are also contextual data points, such as weather conditions, driving conditions, and operator behavior.
Fleet maintenance has its own data challenges with OEM fault codes and sensors firing. Dealing with the reporting data combined with source data can be overwhelming. Maintenance teams lack the time and tools to analyze all the available data. The overload contributes to shops getting stuck in a cycle of reactive maintenance.
Predictive maintenance analyzes a variety of data sources and can distinguish what is useful and meaningful. The more relevant the data, the better for the accuracy of the insights generated by the data models. In short, input improves output.
Older trucks require predictive maintenance
Before the supply chain went haywire and diesel gas prices skyrocketed, fleet maintenance could carry on by following OEM recommendations and performing scheduled maintenance.
With the shortage of new trucks, fleets are maintaining trucks for longer periods of time. Trucks replaced every four years are being taken off the road in five or six years. That’s a huge burden on fleet maintenance to ensure these older trucks stay on the road.
The influx of sensors and fault codes adds pressure to maintenance teams. Breakdowns on the road and unscheduled service calls are more likely with older trucks. Shops are already seeing more exhaust system repairs, a sure sign of aging equipment.
By adding predictive maintenance to your preventive maintenance program, you can leverage insights to better manage maintenance across the fleet. Technicians are also more efficient with opportunities like combining predictive and preventive maintenance on a single service call.
Any solution must integrate with your workflow
Every fleet’s maintenance department has its own workflow system and solutions, whether it’s enterprise asset management (EAM) software, computerized maintenance management software (CMMS) or general transportation software.
Maintenance technicians are busy all day, taking care of work orders and responding to tasks right in front of them. It’s a whirlwind of reactive maintenance combined with regular service. Data plays a role, but it’s more related to diagnosing issues from sensors and fault codes. For technicians, their workflow system is what they live by. Fleet management can add solutions, but adopting and using any solution is a different challenge.
That’s why any solution designed to facilitate predictive maintenance must work seamlessly with your fleet’s technology. It needs to be as simple as going from insight to actual work order automatically entered into the fleet’s workflow system.
Insights without actions are just nice to know
Fleets are no strangers to technology and data. If predictive maintenance appeals to you, here’s a word of caution. It can be tempting to talk up insights. But without a process for putting insights into action, it’s just nice to know information.
A perfect example of this involves fluid analytics. Fleets often contract with labs to conduct oil and coolant analysis. Such analysis reveals degradation and contamination, indicating the asset’s state of health. Many labs use a color-coding system of red, yellow, and green for presenting results. Red should be addressed. Green is good. Yellow is the bugaboo. Too many yellows are inconclusive. It’s an insight that you can’t take action on.
Predictive maintenance technology that utilizes data science algorithms can give you the confidence to turn yellows into green and a few into red. There’s less gray in what you should do. You can move beyond a nice-to-know insight into an action-oriented insight.
Predictive maintenance’s role in preventive maintenance
Preventive maintenance is a fleet’s standard operating procedure. For fleets that need an added advantage, predictive maintenance can help cut maintenance costs, increase uptime across the fleet and improve fuel efficiency. Technology that is purposeful and flexible can help.
The fleet leader that preempted a NOx sensor failure estimates an annual value of $3,387 per vehicle from its predictive maintenance program. What could your fleet maintenance program accomplish by investing in predictive maintenance?