Search Results

During the development of an Internal Combustion Engine-based powertrain, traditional procedures for control strategies calibration and validation produce huge amount of data, that can be used to develop innovative data-driven applications, such as emission virtual sensing. One of the main criticalities is related to the data quality, that cannot be easily assessed for such a big amount of data. This work focuses on an emission modeling activity, using an enhanced Light Gradient Boosting Regressor and a dedicated data pre-processing pipeline to improve data quality. First thing, a software tool is developed to access a database containing data coming from emissions tests. The tool performs a data cleaning procedure to exclude corrupted data or invalid parts of the test. Moreover, it automatically tunes model hyperparameters, it chooses the best set of features, and it validates the procedure by comparing the estimation and the experimental measurement.

The Selective Catalytic Reduction (SCR) system installed on the exhaust line is currently widely used on Diesel heavy-duty trucks and it is considered a promising technique for light and medium duty trucks, large passenger cars and off-highway vehicles, to fulfill future emission legislation. Some vehicles of these last categories, equipped with SCR, have been already put on the market, not only in the US, where the emission legislation on Diesel vehicles is more restrictive, but also in Europe, demonstrating to be already compliant with the upcoming Euro 6. Moreover, new and more stringent emission regulations and homologation cycles are being proposed all over the world, with a consequent rapidly increasing interest for this technology. As a matter of fact, a physical model of the Diesel Exhaust Fluid (DEF) supply system is very useful, not only during the product development phase, but also for the implementation of the on-board real-time controller.

To optimize the tyre contact patch in a sports car, Ferrari has developed an active camber and toe (ACT) system comprising of 4 actuators for the rear axle. This complex and completely new system is difficult to model accurately and for this reason, it was decided to combine a physical prototype with a full vehicle model to carry out the functional tests. The method of combining a virtual model with a physical test is known as hybrid simulation. This functional testing of both the actuators and the vehicle dynamics logic will be performed on an MTS 6DOF bench test prior to physical track testing on a prototype vehicle using Ferrari facility in Maranello, Italy. In support of this functional testing, we will use hybrid simulation techniques with software and methods specifically developed. The planned hybrid test system described in the paper will allow dynamic coupling between the physical bench test and a modified full vehicle simulation model.