A Systematic Review and Analysis of Machine Learning models in Liquid Rocket Engine Control

This paper explores machine learning (ML) methods that enhance performance, design, health, and operation of liquid rocket engines. Various ML approaches, including reinforcement learning (RL), supervised, and unsupervised learning can potentially transform rocket propulsion technologies, essential for critical interplanetary missions. Specifically, this study reviews neural network-based models for health monitoring of rocket engines, RL for control of engine ignition and operation, and ML techniques for anomaly detection. The application of these algorithms leads to significant advances in analyzing and predicting rocket engine system performance. While such techniques enhance efficiency, they also present challenges, which are discussed herein. Subsequently, this study provides a comprehensive synthesis of ML techniques as applied to rocket engine diagnostics and prognostics. This approach not only addresses the existing research gap but also lays the groundwork for future explorations into the predictive maintenance and optimization of space propulsion systems.