Master Thesis: “AI-based Diagnosis of Combustion Anomalies in Hydrogen Engines”</a
Author: Arun Sai Thunga
Supervisor: Prof. Enrique Zuazua
Date: June 30, 2024
In the field of automobile maintenance, a wide range of sensors are useful for capturing the signals inside the combustion chamber of engines. Addressing the intricacies of anomaly detection in H2 engine combustion chambers using numerical data requires a tailored approach that leverages the power of machine learning (ML). It is difficult for standard data processing techniques to distinguish between good signals and abnormal signals. This cutting-edge technology captures even the minute deviations from the ground truth signal. However, the data related to the H2 Engines, faces a data deficit. The lack of data makes training and generalizing models particularly challenging. Therefore, the over sampling techniques are employed to produce synthetic data.
Later, multi-class, multi-output classification models utilizing machine learning and deep learning algorithms are proposed. These proposed approaches simultaneously analyzes the relationships between various parameters, offering a more comprehensive understanding of anomalies that may be overlooked by traditional, parameter-centric approaches. Both machine learning and deep learning introduces a paradigm shift by allowing the system to understand deep patterns and make nuanced decisions. Implementing a multi-output, multi-class classification system enables the integration of these advanced techniques, providing a more nuanced and accurate approach to anomaly detection.
(…) Ultimately, the goal is to enhance the efficiency of anomaly detection procedures during combustion chamber maintenance by combining synthetic data generation with machine learning techniques. This method addresses data constraints and enables the model to accurately identify anomaly configurations observed in real- world situations.
See all details at:
Master Thesis: “AI-based Diagnosis of Combustion Anomalies in Hydrogen Engines”, by Arun Sai Thunga (June 30, 2024)
Supervisor: Prof. Enrique Zuazua
