With 64 million units having to be taken off the market, 2014 was one of the worst years for the automotive industry in terms of the number of vehicles recalled. The fact that in 2017 the number dropped to only 10 million was an improvement, but still not cause for celebrating. Just a few weeks back, in March 2018, Tesla recalled 123,000 Model S cars over faulty steering components. Shares of the company fell 4% right after the announcement. Vehicle recalls continue to make headlines, costing car manufacturers millions of dollars and great reputational damage. This is all despite the fact that the automotive sector is one of the industries where manufacturing processes are most stringent in terms of quality checks, regular maintenance and monitoring.
With technology turning vehicles into platforms of innovation, through using security, efficiency and computing power performance, it’s time for manufacturers to become “smarter” when it comes to deploying the right technologies throughout the production and post-purchase lifecycle. This is where AI, machine learning and predictive analytics can enable manufacturers to gain full visibility and control of manufacturing processes.
Predictive maintenance from factory floor to out in the field
They say that the best time to correct an error is yesterday, and that is true in the automotive sector. While AI and machine learning cannot — yet — turn back time, cognitive technologies can analyze data in ways that were previously unattainable. While manual modeling on past failure patterns executed by data scientists is nothing new, data analysis performed by AI-powered platforms builds cognitive learning which not only can learn from past failure patterns, but more importantly learn to detect issues not known or seen before. This is very important as the majority of the issues that happen in the manufacturing process that eventually cause a recall are not repeat issues but new ones. Cognitive applications are able to teach themselves, with data from sensors, the various normal operating conditions and the environmental influences on a machine at a fine-grained level, going beyond the macro-patterns that the human brain tends to spot. This means that micro-anomalies and small changes that can go undetected in the quality check process can now be identified automatically as they occur and be responded to before they lead to current or future defects. In this way, breakdowns or faults can be predicted ahead of them occurring and sparking recalls or causing downtime.
To be most effective, predictive maintenance systems should be deployed across different production touchpoints. During the manufacturing process, cognitive predictive maintenance can identify and share alerts on in-line defects, which can then be corrected early on, while providing information to industrial engineers on the nature of the defect. In this way, defects are addressed before the product reaches the market and costly recalls can be avoided. This can also dramatically reduce scrap and rework in the product lines and thereby save significant costs too.
With industrial IoT and sensors, we now have the ability to understand the health of a machine in minute detail. This is achieved through the creation of digital twins, granular digital representations of the parts and the dynamics in the manufacturing process, which are enabled by machine learning and artificial intelligence. These digital models of physical assets provide a representation of their physical counterpart’s current status, working condition and functionality to help understand and predict the health and readiness of these parts. By creating digital twins, insights can be garnered to automatically detect and address the tiniest of issues that would otherwise be missed during a manual inspection process.
Once the vehicle goes to market, AI and predictive maintenance can work their magic. Service engineers can use cognitive analysis of multiple sources of data collected from sensors and service records to identify and solve problems in vehicles much before the “check engine” light goes on, eliminating safety risks and costly repairs, and delivering a superior customer experience. Also, it helps the engineers plan inventory of parts required for maintenance much better since they know of the problems ahead of time. Both the driver and the service engineers are empowered with proactive alerting systems that flag potential problems before they escalate into safety risks or costly damages, and at the same time move beyond just the schedule-based preventative maintenance to a need-based and timely predictive maintenance. Manufacturers can also gather useful insights from the field data to improve their product design as well as the manufacturing process.
Cognitive-first technology saves the day for the automotive industry
Predictive maintenance is not a layer of monitoring and checks that is added on current control systems. It is, in fact, an integrated cognitive and machine-first technology that runs end to end in the manufacturing and post-purchase lifecycle, ensuring that these processes can run like clockwork. With smart and connected cars changing the way we see vehicles, automotive manufacturers need to embrace what AI and machine learning have to offer to break a new record — that of zero car recalls!
All IoT Agenda network contributors are responsible for the content and accuracy of their posts. Opinions are of the writers and do not necessarily convey the thoughts of IoT Agenda.