Sappi is the leading European producer of paper and packaging. Sappi collects real time sensor data from its manufacturing processes. Regretfully, sensors not always function correctly. Sensor failure causes costly interruptions. Using the Scailable platform, this sub-project will validate anomaly detection models (provided by EFPF service: https://docs.efpf.linksmart.eu/projects/data-analytics/) to detect sensor failures in the challenging production environment of a Sappi mill.

 

This sub-project demonstrates the value of deploying AI “on the edge” in Industrial IoT. Edge deployment (i.e. in-situ as opposed to the cloud) reduces latency and energy consumption. In the context of a Sappi’s mill that would mean quicker detection of sensor failures and therefore avoiding costs due to machines’ downtime and pre-emptive maintenance when and where necessary.

​

The objective of this sub-project is to use EFPF’s service to detect sensor failures directly on the edge for a machine in Sappi’s Lanaken mill in Belgium.

 

Edge deployment (i.e. in-situ as opposed to the cloud) reduces latency and energy consumption. In the context of a Sappi’s mill that would mean quicker detection of sensor failures and therefore avoiding costs due to machines’ downtime and pre-emptive maintenance when and where necessary.

 

This sub-project will a) validate anomaly detection models by using these in a demanding industrial context, and b) (further) validate the EFPF services “on the Edge” and Scailable’s platform for deployment of trained AI models in the context of industrial paper making.