THINKING OUTSIDE THE BOX
In most industries the cost of maintenance accounts for around 10 to 25 percent of total operating expenses. Machine learning and AI techniques – fueled by 5G connectivity – can be the greatest enablers for measures of predictive maintenance and for tackling the problem. But how to start? Often, problems just seem too huge and the solutions just too complex to imagine and visualize. We know that to find the right solution, you sometimes need to get creative and think outside the box. This is why this time we used a LEGO® Education solution to make your technology shrink and your IoT vision rise!
The world of predictive maintenance
Predictive maintenance is not only about improving your maintenance processes. It can optimize your smart buildings and cities, shape the future of your utilities and oil & gas plants and be the game changer for your medical device.
Every manufacturing tool has a lifespan – so does a drilling head. This, however, highly depends on factors like materials and products.
Not being able to predict maintenance needs, means: unplanned downtimes of production facilities (e.g. due to broken tools) and quality damages – risking productivity and your company's reputation.
Without IoT, production plannings are often based on assumptions only. Additionally, in reality many tools get replaced without being damaged, just because of static maintenance schedules.
So how did we approach this challenge?
Usually we would go for a digital twin to visualize the challenge. But this time we tried something new: we put our greatest developers into a room and let them build a mini version of the use case using LEGO® MINDSTORMS® Education EV3: we monitored and collected telemetry data/KPIs on the tool length using 5G connectivity powered by Vodafone in a Cloud. In real life, our data would even include more different given machine data (e.g. temperature, speed, torque, length, pressure). Out of this data we developed a Cloud-based data lake and trained an AI model to predict the lifespan of the drilling tool using AZURE AutoML.
The AI algorithm can then predict and give information on:
- the current status of the drilling head
- the point in time to install a new tool
- the number of products produced before the tool will reach the end of its lifespan
- the estimated quality of the product (due to current status of tool)
This way, maintenance plannings can significantly be improved: saving time and money and ensuring smooth operations.
ENABLING IOT USING A LEGO® EDUCATION SOLUTION AND 5G CONNECTIVITY
We rebuilt IoT challenges using a LEGO® Education solution and connected data streams with Vodafone, a pioneer in 5G technology and connectivity in the smart factory environment. Learn about our ecosystem and Vodafone's smart factory offering.