Progress and prospects of artificial intelligence development and applications in supersonic flow and combustion

Abstract

Due to the significant improvement in computing power and the rapid advancement of data processing technologies, artificial intelligence (AI) has introduced new tools and methodologies to address the challenges posed by high nonlinearity and strong coupling characteristics in traditional supersonic flow and combustion. This article reviews the considerable progress AI has made in applications within the fields of supersonic flow and combustion, covering three main aspects: intelligent turbulence combustion simulation, supersonic flow field intelligent reconstruction based on deep learning, and the intelligent design of the full-flow passage of supersonic engines. In recent years, the field of turbulent combustion has seen the utilization of large volume of data combined with implementation of advanced machine learning models, enabling accurate predictions of combustion efficiency and optimization of the combustion process. Flow field intelligent reconstruction employs deep learning networks to accurately reconstruct the detailed information of the entire flow field from limited observational data, enhancing the capacity to analyze and predict supersonic flows. The intelligent design of the full-flow passage of supersonic engines has led to efficient design and optimization of complex flow systems through the integration of advanced optimization algorithms and AI technology. These advancements have driven the development of supersonic flow and combustion theories and provided innovative solutions for related engineering applications. Finally, the challenges and future applications of machine learning in combustion research are discussed.