– Using Machine-Learning for Real Time Control of the Chemical Quality of Boring Machine Cuttings

 - 

ENVISOL has had the opportunity to conduct a project with a strong AI component. The main objective of the talk is to give a feedback on the key steps of the project : from the data acquisition, the measurements methodology and the first statistical analysis to the implementation of Machine Learning algorithms that give away the final output.

Machine learning fulfilled a need in gaining time, money and efficiency in the chemical characterisation of cuttings from boring machine. “Random-Forest” algorithms, trained on built datasets, make possible a real-time classification of those cuttings.

This talk, which aims to be accessible to everyone, will go through the main steps while avoiding going into unnecessary details. Limitations and difficulties encountered will also be developed to give a global picture of the use of Machine Learning in that practical and specific application.