Share:

Breght Van Baelen

Breght Van Baelen is an all-round data scientist at Cegeka, interested in both data strategy as well as deep-diving into technical details. He believes businesses are drowning in information but starved for knowledge. The practice of knowledge extraction has proven to reduce business costs and increase productivity, making businesses more efficient and effective.

Becoming an active member of the data community, Breght is excited to present an end-to-end solution for predictive maintenance at dataMinds connect 2018. His solution will reduce maintenance & transportation costs and help companies transition from a break-fix model to a prevent-optimize one.





Presenting

Real-time predictive maintenance with Azure ML Studio & PowerBI

Maintaining industrial machines is expensive: sensor data needs to be monitored and analyzed, technicians need to be sent out and defects need to be repaired. Predictive maintenance reduces these costs by transitioning from a break-fix model to a prevent-optimize one. To this extend, We propose an end to end Microsoft Azure solution. First, representative sensor data for industrial machines is ingested into Azure with Event Hubs. Using Streaming analytics and Azure ML Studio, we'll predict whether the machine approaches overheating in real-time. If overheating is predicted, the machine will be automatically shut down until cooled again to prevent hardware damage. The maintenance department will be notified by phone or mail and can further analyze the prediction or oppose the automatic shut down in a PowerBI dashboard.

Saar Gillis  400