Abstract:The last years have witnessed an enormous interest in the use of artificial intelligence methods, especially machine learning algorithms. This also has a major impact on aerospace engineering in general, and the design and operation of liquid rocket engines in particular, and research in this area is growing rapidly. The paper describes current machine learning applications at the DLR Institute of Space Propulsion. Not only applications in the field of modeling are presented, but also convincing results that prove the capabilities of machine learning methods for control and condition monitoring are described in detail. Furthermore, the advantages and disadvantages of the presented methods as well as current and future research directions are discussed.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is to improve the design and operation of liquid - rocket engines through machine - learning methods. Specifically, the paper explores the application of machine learning in the following aspects: 1. **Modeling**: - **Prediction of the maximum wall temperature in the cooling channel**: A surrogate model based on neural networks (NN) was developed to predict the maximum wall temperature in the cooling channels of liquid - rocket engines. This model uses data extracted from computational fluid dynamics (CFD) simulations for training and can provide accurate predictions at least 1,000 times faster than three - dimensional CFD simulations. - **Fatigue life estimation**: Neural networks are used to predict the fatigue life of reusable rocket thrust chambers. Training data is generated by finite element method (FEM) calculations, and combined with Coffin - Manson theory and ductility failure analysis, high - precision fatigue life prediction is achieved. 2. **Control**: - **Early detection of thermo - acoustic instabilities**: It was studied how to use support vector machines (SVM) combined with combustion noise characteristics to construct reliable early - warning signals for detecting thermo - acoustic instabilities in rocket thrust chambers. - **Startup control of gas - generator engines**: Deep reinforcement learning (DRL) is used to optimize the continuous startup process of gas - generator engines. A neural - network controller is trained through the EcosimPro simulation tool, achieving fast and stable startup control. - **Set - point control of expansion - bleed engines**: The method of using deep reinforcement learning for transient control of expansion - bleed liquid - rocket engines was studied. A neural - network controller is trained through the EcosimPro simulation environment, achieving fast and non - overshooting set - point tracking. 3. **Current and future research directions**: - **Integration of physical laws**: It was studied how to pre - integrate physical laws into machine - learning models to reduce the amount of required training data and improve prediction quality. - **Dynamic modeling based on machine learning**: The exploration of using methods such as Gaussian processes and recurrent neural networks (RNN) to model complex non - linear dynamic systems. - **Uncertainty quantification**: It was studied how to estimate the prediction accuracy of machine - learning models, especially in safety - critical applications. - **Life - extension control**: It was studied how to implement control strategies that achieve high performance without damaging the system through machine - learning methods, especially for reusable rocket engines. - **Fault detection and diagnosis**: Develop a real - time fault detection and diagnosis system to monitor the abnormal behavior of rocket engines and trigger emergency shutdown. - **Fault - tolerant control**: It was studied how to take optimal responses and remedial measures after detecting a fault. - **Application of machine learning in safety - critical situations**: It was studied how to ensure safety when using machine - learning algorithms, including software verification, validation, and testing. - **Embedded systems for modern rocket - engine control**: The implementation of embedded systems suitable for space applications was studied, with special attention to the special challenges of hardware and software in the harsh space environment. In general, this paper shows the broad application potential of machine learning in the design and operation of liquid - rocket engines through multiple cases and proposes future research directions.

Machine Learning Methods for the Design and Operation of Liquid Rocket Engines -- Research Activities at the DLR Institute of Space Propulsion

A Reinforcement Learning Approach for Transient Control of Liquid Rocket Engines

Reframe the Field of Aerospace Engineering Via Machine Learning: Application and Comparison

Data-Driven Aerospace Engineering: Reframing the Industry with Machine Learning

Optimizing Solid Rocket Missile Trajectories: A Hybrid Approach Using an Evolutionary Algorithm and Machine Learning

Enhancing airflow dynamics in airjet spinning: A machine learning approach to optimize nozzle design

Machine Learning in Aerodynamic Shape Optimization

Prediction of Particle Properties in Plasma Spraying Based on Machine Learning

Improving aircraft performance using machine learning: a review

Predicting Wall Pressure Fluctuations on Aerospace Launchers Through Machine Learning Approaches

Investigation of the aerodynamic optimization design of fluid machinery based on machine learning

Shape Optimization of a Diffusive High-Pressure Turbine Vane Using Machine Learning Tools

Machine Learning for Fluid Mechanics

Rotor37 Aerodynamic Optimization: A Machine Learning Approach

Reinforcement Learning for Engineering Design Automation

Machine-Learning and Physics-Based Tool for Anomaly Identification in Propulsion Systems

Liquid Rocket Engine Performance Characterization Using Computational Modeling: Preliminary Analysis and Validation

Aero-Engines AI - A Machine-Learning App for Aircraft Engine Concepts Assessment

Early Detection of Thermoacoustic Instabilities in a Cryogenic Rocket Thrust Chamber using Combustion Noise Features and Machine Learning

Principles of Machine Learning and Its Application to Thermal Barrier Coatings