Predicting transient performance of a heavy-duty gaseous-fuelled engine using combined phenomenological and machine...

Area of Science:

• * Engineering
• * Environmental Science
• * Computational Modeling

Background:

• * Long-haul transportation heavily relies on fossil fuels, contributing significantly to greenhouse gas (GHG) emissions.
• * Natural gas (NG) engines offer a near-term solution for reducing carbon content in fuels while maintaining engine efficiency.
• * Accurate prediction of engine performance and emissions is vital for evaluating NG technology in various applications.

Purpose of the Study:

• * To develop and validate a coupled engine and machine learning model for predicting the performance and emissions of a high-pressure direct-injection NG engine.
• * To assess the model's accuracy in predicting engine parameters, nitrogen oxide (NOx), and methane (CH4) emissions over transient drive cycles.
• * To evaluate the potential of NG engines for GHG reduction in long-haul transportation.

Main Methods:

• * Adaptation of the GT-SUITE™ DI-Pulse model for a 6-cylinder NG engine, incorporating an air exchange system for transient analysis.
• * Coupling the 1-D engine model with a machine learning model to enhance the prediction accuracy of CH4 emissions.
• * Validation of the model using transient engine dynamometer data and application to regulatory and real-world drive cycles.

Main Results:

• * The developed model accurately predicts engine performance with less than 5% error in steady operation.
• * Cumulative errors for NOx and CH4 emissions were below 10% across studied drive cycles.
• * Methane (CH4) constituted 0.2% to 1.4% of total GHG emissions, depending on the drive cycle.

Conclusions:

• * The combined 1-D engine and machine learning model effectively predicts NG engine performance and emissions, including CH4 and NOx.
• * This modeling approach is valuable for evaluating NG engine technologies and their GHG reduction potential in long-haul transport.
• * The model demonstrates significant potential for engine evaluation studies, especially when integrated with emerging technologies.