This content is not included in your SAE MOBILUS subscription, or you are not logged in.

Artificial Intelligence Strategies for the Development of Robust Virtual Sensors: An Industrial Case for Transient Particle Emissions in a High-Performance Engine

Journal Article
03-17-02-0014
ISSN: 1946-3936, e-ISSN: 1946-3944
Published September 08, 2023 by SAE International in United States
Artificial Intelligence Strategies for the Development of Robust
                    Virtual Sensors: An Industrial Case for Transient Particle Emissions in a
                    High-Performance Engine
Sector:
Citation: Pulga, L., Forte, C., Siliato, A., Giovannardi, E. et al., "Artificial Intelligence Strategies for the Development of Robust Virtual Sensors: An Industrial Case for Transient Particle Emissions in a High-Performance Engine," SAE Int. J. Engines 17(2):237-253, 2024, https://doi.org/10.4271/03-17-02-0014.
Language: English

Abstract:

The use of data-driven algorithms for the integration or substitution of current production sensors is becoming a consolidated trend in research and development in the automotive field. Due to the large number of variables and scenarios to consider; however, it is of paramount importance to define a consistent methodology accounting for uncertainty evaluations and preprocessing steps, that are often overlooked in naïve implementations. Among the potential applications, the use of virtual sensors for the analysis of solid emissions in transient cycles is particularly appealing for industrial applications, considering the new legislations scenario and the fact that, to our best knowledge, no robust models have been previously developed. In the present work, the authors present a detailed overview of the problematics arising in the development of a virtual sensor, with particular focus on the transient particulate number (diameter <10 nm) emissions, overcome by leveraging data-driven algorithms and a profound knowledge of the underlying physical limitations. The workflow has been tested and validated using a complete dataset composed of more than 30 full driving cycles obtained from industrial experimentations, underlying the importance of each step and its possible variations. The final results show that a reliable model for transient particulate number emissions is possible and the accuracy reached is compatible with the intrinsic cycle to cycle variability of the phenomenon, while ensuring control over the quality of the predicted values, in order to provide valuable insight for the actions to perform.