AI models for device health status prediction

Understanding the status of the components or of a device can be hard in scenarios where it is not possible to access the devices easily, i.e., components integrated inside motors. Multiple factors can affect this health measurement, making it complex to determine. However, it is an important parameter since it can be used for predictive maintenance, ensuring a better performance of the system. In COMPAS project, Eesy innovation has developed multiple AI models to predict the health status for Infineon, NXP and SIEMENS based on some dataset provided by those partners.

These AI models have been coded in Python language to make use of most relevant AI frameworks. However, making use of containers, these models can be deployed in numerous systems making it more accessible for customers. The data preprocessing techniques have also been integrated in the AI pipeline to make it simpler for users. With some dataset from new users/customers, the designed algorithms could be adapted to work in different scenarios.

The trained AI models have been validated using test datasets (not used during the training phase) to ensure a correct validation of the models.

In order to use these AI models, a first phase of model fine tuning is required for new customers. During this phase, new datasets from the new scenarios would be required.

Infineon, NXP, SIEMENS and other companies/customers that need information about the health status of their components/devices.

These models will be used by Eesy innovation as examples for new customer as well as initial approaches for other health status prediction models in future projects or consultant services.