Xu, Zidan (SAIC Motor R&D Innovation Headquaters)

Machine Learning Based Pre-Ignition Detection for Hydrogen Internal Combustion Engines

Scheduled for presentation during the Regular session "Fault Diagnosis and Simulaton Technoqies " (ThA1), Thursday, October 31, 2024, 10:50−11:10, Room T1

7th IFAC Conference on Engine and Powertrain Control, Simulation and Modeling, Oct 30 - Nov 1, 2024, Dalian, China

This information is tentative and subject to change. Compiled on May 16, 2025

Abstract

In response to the urgent need for low-carbon and sustainable societal development, hydrogen fuel has become a primary focus for internal combustion engine applications. However, abnormal combustion events, such as pre-ignition, have been a major hurdle in its development. Currently, the most effective method for detecting pre-ignition is the use of in-cylinder pressure sensors, which provide high precision and sensitivity but are also costly.

In this work, we propose an machine learning based pre-ignition detection method for Hydrogen combustion engine with production level equipments. The results show that the proposed method achieved 97% accuracy and 80% recall score.