AI based predictive maintenance of the Cloud-Edge continuum, DiagFit tool, two pilot deployment cases
The objective of the E2CC_PM project is to develop and implement an industrial predictive maintenance solution, based on AI, that can be flexibly deployed on the Edge-Cloud continuum according to operational constraints: for example, the processing of raw data at the Edge as close as possible to the sensors – to reduce the environmental impact and overcome the hindrance of transmission costs – and the monitoring of alarms by the operator via the Cloud.
DiagFit, the software product of Amiral Technologies, allows to build and run anomaly detections on equipment/systems/processes, based on machine learning algorithms applied to time series. Models can be built quickly, with little or no historical failure data, by domain experts via a no-code interface. The software then performs anomaly detection and diagnosis with the goal of predicting failures.
To meet the Edge to Cloud objectives, the scope of the E2CC_PM project is to facilitate access to data, thus also including the required preprocessing algorithms, to adapt the application architecture to cope with the Ege and Cloud modules, and to reduce the power and memory requirements to allow the deployment of the Edge module on small footprint electronics.
The E2CC_PM project brings to the E2CC project a "Predictive Maintenance" brick by artificial intelligence in the entire consolidation of this Edge to Cloud continuum and contributes to providing a solution to the two constraints and limitations mentioned above.