Credit: CC BY-SA 4.0 BY DANLAW INC
Predictive maintenance can help avert automotive downtime
In most commercial sectors where delivery of goods and services is essential, reliable road transportation is key. For organizations with a large vehicle fleet, staying on top of maintenance schedules is a well-established challenge. In the trucking industry, for example, vehicle downtime can cost between $448-$760 per day, per vehicle. To avoid unplanned maintenance downtime, companies often opt to plan ahead by adhering to preventative maintenance schedules. And for good reason. Results from a 2017 study from Element Fleet Management indicate that vehicle fleets participating in preventive maintenance programs experience roughly 20 percent fewer maintenance-related downtime days than those that do not..
Many companies, convinced that predictive maintenance is both economically and logistically prudent, now hope to further optimize it through the adoption of Internet of Things (IoT) and machine learning data techniques. If integrated correctly, these advances will allow them to accurately and precisely pinpoint when vehicle maintenance is needed. Addressing potential issues proactively can avert the risk of a vehicle being out of commission unexpectedly.
Though commercial transportation organizations have been eager to adopt IoT-enabled predictive maintenance, individual consumers also see an advantage to connected vehicle technology—as the data it can gather has the potential to prevent automotive surprises.
What is IoT predictive maintenance?
The goal of IoT-enabled predictive maintenance is to use time series data to identify the time that the automotive equipment is likely to fail. In order to collect this data, sensors are used and housed within mechanical and electronic automotive systems to monitor efficacy and efficiency. After data is collected it can be transmitted to a reporting system, be it a vehicles’ fleet manager or a service provider.
Sensors can be used to monitor engine performance, exhaust systems, and transmission function. IoT technology can also evaluate things like tire pressure and oil levels—which are less complex systems—but can still cause significant problems if they were to fail.
In short, IoT predictive maintenance differs from preventative maintenance since it is responding to the current condition of parts and equipment rather than its estimated performance and longevity.
How does predictive maintenance work?
Most 21st century cars are already equipped with embedded computer systems that can deliver data to technicians for diagnostic purposes through a two-way communication process where technicians connect their service computer to a vehicle’s onboard computer to deliver form diagnostic codes to the service center. Automated onboard notifications that signal maintenance needs are also already present in most cars manufactured in the last two decades. We see this anytime an oil light or a service notification appears on the dashboard console. Some cars even have sensors that signal when tire pressure is low, reminding the user to act accordingly.
While these methods can spur needed maintenance resolution, service contact must be first initiated by the vehicle user, which is something that IoT solutions can change. If, for example, vehicle system data was shared with a consumer’s service center, it could allow for timely service or maintenance reminders that are customized for the user’s vehicle profile. On the service station side, technicians could periodically pull data from customers’ vehicles to perform predictive analysis and reach out to customers to schedule service appointments if needed. Push notification alerts could also delivered onboard via dash screens or to mobile devices, both methods encouraging immediate operator response.
For commercial vehicles, data sharing could support advanced analytics, wherein companies could make long-range maintenance projections based on historical data. This type of data analysis could be useful for automotive manufacturers as well, as they could have a better sense of failure trends for particular parts or systems.
It all sounds simple, but in order for IoT-enabled predictive maintenance to work, the algorithms used to analyze data need to be agile. Wael Elrifai, senior director of sales engineering and data science at Hitachi Vantara tells Network World that the models used to analyze equipment system behavior must change over time because as parts age, they will behave differently. It is common to see a lot of failure early on in a system’s life cycle and again toward the end. This is what Elrifai refers to as the “bathtub curve” of equipment failure. Given that, maintenance schedules—and the IoT automotive devices that map to them—must be adjusted at all stages of a vehicle’s life cycle.
Predictive maintenance has far reaching value
Predictive maintenance can help companies and individual consumers avoid vehicle downtime and can offer other insights regarding automotive part wear and usage patterns. Long term, it also has the potential to offer cost savings—not simply because vehicles can stay in operation reliably, but also because it can reduce repairs and improve safety outcomes. If vehicle maintenance is performed before equipment loses performance, unplanned service downtime may be a thing of the past.