Identification of Adequate Combustion in Turbulent Jet Ignition Engines using Machine Learning Algorithms. Issue 10 (2021)

• Record Type: Journal Article
• Authors: Novella, Ricardo; Pla, Benjamín; Bares, Pau; Aramburu, Alexandra
• Is Part Of: IFAC-PapersOnLine. Volume 54:Issue 10(2021)
• Page Start: 102
• Page End: 107
• Publication Date: 2021
• Subjects: Turbulent jet ignition -- extreme learning machine -- combustion stability -- misfire -- prediction; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2021.10.148
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: Turbulent jet ignition (TJI) engines might substantially improve the efficiency of spark ignition (SI) engines by locating the spark in a pre-chamber. Concretely, the passive pre-chamber concept could be directly applied in commercial SI engines with minor modifications. The pre-chamber burns a small quantity of fuel providing the rest of the chamber with sufficient energy for a fast and efficient combustion, however, the combustion suffers from controllability and the phenomena involved are difficult to predict. The present paper aims to develop a suitable algorithm for combustion prediction in TJI engines by exploring machine learning techniques to identify the more appropriate operating conditions. The algorithm differentiates between stable and unstable combustion using online data to keep updated in case of a non-expected bias. Experimental data from a research TJI engine demonstrate the capacity of the algorithm proposed.