Application of data driven methods to support the customer in the understanding the physical phenomena that lead to the breakout of electromagnetic actuators in a sport car exhaust system due to overheating.

Objectives

The approach implemented by SATE allowed the customer to identify the triggering conditions that lead to anomalous behaviour and breakout of the component and the working conditions that resulted in the increased stress of the component. SATE allowed the customer engineers to understand the problem which was not identified by them relying only on a-priori knowledge and physical/engineering approaches, due to the complex geometry and cooling paths of the system. The goal was also to identify boundaries of performance parameters that favour the phenomenon, thus to be avoided or limited for excessive durations.

Benefits

Indications on particular effects not previously considered

Indications on particular effects not previously considered

The identification of the conditions that may lead to overheating of the actuators may suggest the presence of a driving behaviour that have unexpected effects to be further investigated.

Indications on mitigation actions to be performed

Indications on mitigation actions to be performed

Thanks to the identification of the conditions that may lead to component failures, the customer engineers can extract indications on possible mitigation actions to be adopted in the components/vehicles.

Features

Machine learning models

Machine learning models

Models developed for the objectives achievement are based on machine learning algorithms trained to distinguish the conditions that led to overheating from those that did not.

Application of systems engineering approach

Application of systems engineering approach

The activities are performed following the systems engineering approach allowing a deep investigation of the entire system functionalities and not limiting the study to the available data.