Use case Nuova Simonelli


Predictive Maintenance

The use case involves the development of new "smart" models of use and maintenance of coffee machines (Nuova Simonelli). This approach allows a management of the operation (implementation of actions to restore and / or correction of operating modes) even remotely because the machines are constantly under control.
The analysis of the parameters detected and associated with a specific fault of the machine in use allows to derive a predictive algorithm based on artificial intelligence and machine learning so that it is possible to intervene on a machine before the fault by replacing the components whose failure is expected soon. These activities enable new levels of process integration and advancement by opening them up to new capabilities and services.
The topics that have the main relationship with this use case are: AI and big data in the perspective of predictive maintenance.
Integrations with respect to the objectives of the use case: two models of coffee machines produced by Nuova Simonelli were selected with the related post-sales data. The relative test datasets were also considered in order to search for possible correlations between the maintenance of coffee machines and failures during operation at the customer's premises. The datasets represent the essential elements to contextualize the scope of the research regarding the Artificial Intelligence techniques that can be applied and the correct choice of data analysis techniques.
Development of the use case: the datasets of testing and post-sales of coffee machines have been integrated into the data analysis processes in order to train predictive models. These processes are characterized by the engineering phases and the identification of the features in the data that best describe the operating parameters of coffee machines both during testing and operation. Thanks to the analysis of correlation with respect to failures verified over time, it was possible to train Machine Learning models
to be able to automatically recognize such conditions and anticipate them over time.
Valorisation of the results: the models and techniques used to analyse test and production data on coffee machines have made it possible to define new approaches to AI techniques in the industrial context. This makes it possible to improve and make efficient maintenance interventions over time with respect to classic scheduled interventions.


Lube 1