Process in the application of machine learning in ejector research
The ejector is a widely used mechanical device with advantages such as simple structure,low initial cost,easy maintenance,and reliable operation.It is widely applied in fields such as refrigeration,desalination,chemical engineering,fuel cells,aerospace,etc.The ejector does not directly consume mechanical energy,which allows for energy-saving purposes,making it more important and attractive with the national goals of"carbon peaking and carbon neutrality"in China.Machine learning methods,as a data-driven automated analysis approach,can be used for analyzing the internal flow characteristics of ejectors and optimizing ejector performance.In recent years,a small number of scholars have already applied machine learning methods to the study of ejectors in various applications,aiming at improving the ejector performance and the system performance.But the research in the open literature is currently scattered,and the state of the art is not yet clear.The present work comprehensively reviewed the literature on the application of machine learning methods in the study of ejectors for different applications,analyzed the current research status,summarized the machine learning methods utilized in the open literature,and pointed out that in the future machine learning methods can be applied to the study of internal flow characteristics of ejectors,providing a basis and guidance for improving the efficiency and performance of ejectors.Machine learning methods can be applied to the study of ejector performance under variable operating conditions,constructing a pathway from automated design to real-world application of ejectors.Constructing more suitable algorithms and proposing a series of targeted solutions.
国家科技期刊平台
NSTL
万方数据
