A flow control method based on an active jet is developed to restart hypersonic inlets. The dynamic restarting process is numerically reproduced by unsteady Reynolds averaged Navier-Stokes(RANS) modeling to verify the...A flow control method based on an active jet is developed to restart hypersonic inlets. The dynamic restarting process is numerically reproduced by unsteady Reynolds averaged Navier-Stokes(RANS) modeling to verify the effectiveness and reveal the influence of jet conditions. The active jet improves the inlet unstart status by drawing the high-pressure separation bubble from the internal compression duct and performing a full expansion to alleviate the adverse pressure gradient. Moreover, the favorable pressure gradient in the inlet caused by jet expansion allows for a successful restart after turning off the jet. The influence of the jet momentum ratio is then analyzed to guide the design of the active jet control method and choose the proper momentum ratios. A low jet momentum does not eliminate the high-pressure separation bubble, whereas an excessive jet momentum causes severe momentum loss due to the induced shock. The general rule in restarting the inlet using an active jet is to allow a full jet expansion downstream of the jet slot while avoiding excessive momentum loss upstream and preventing the thick low-speed layer.展开更多
Nowadays, media cloud and machine learning have become two hot research domains. On the one hand, the increasing user de- mand on multimedia services has triggered the emergence of media cloud, which uses cloud comput...Nowadays, media cloud and machine learning have become two hot research domains. On the one hand, the increasing user de- mand on multimedia services has triggered the emergence of media cloud, which uses cloud computing to better host media servic- es. On the other hand, machine learning techniques have been successfully applied in a variety of multimedia applications as well as a list of infrastructure and platform services. In this article, we present a tutorial survey on the way of using machine learning techniques to address the emerging challenges in the infrastructure and platform layer of media cloud. Specifically, we begin with a review on the basic concepts of various machine learning techniques. Then, we examine the system architecture of media cloud, focusing on the functionalities in the infrastructure and platform layer. For each of these function and its corresponding challenge, we further illustrate the adoptable machine learning based approaches. Finally, we present an outlook on the open issues in this intersectional domain. The objective of this article is to provide a quick reference to inspire the researchers from either machine learning or media cloud area.展开更多
基金supported by the National Key Research and Development Program of China (No.2021YFA0719204)the National Natural Science Foundation of China (No.12272387)。
文摘A flow control method based on an active jet is developed to restart hypersonic inlets. The dynamic restarting process is numerically reproduced by unsteady Reynolds averaged Navier-Stokes(RANS) modeling to verify the effectiveness and reveal the influence of jet conditions. The active jet improves the inlet unstart status by drawing the high-pressure separation bubble from the internal compression duct and performing a full expansion to alleviate the adverse pressure gradient. Moreover, the favorable pressure gradient in the inlet caused by jet expansion allows for a successful restart after turning off the jet. The influence of the jet momentum ratio is then analyzed to guide the design of the active jet control method and choose the proper momentum ratios. A low jet momentum does not eliminate the high-pressure separation bubble, whereas an excessive jet momentum causes severe momentum loss due to the induced shock. The general rule in restarting the inlet using an active jet is to allow a full jet expansion downstream of the jet slot while avoiding excessive momentum loss upstream and preventing the thick low-speed layer.
文摘Nowadays, media cloud and machine learning have become two hot research domains. On the one hand, the increasing user de- mand on multimedia services has triggered the emergence of media cloud, which uses cloud computing to better host media servic- es. On the other hand, machine learning techniques have been successfully applied in a variety of multimedia applications as well as a list of infrastructure and platform services. In this article, we present a tutorial survey on the way of using machine learning techniques to address the emerging challenges in the infrastructure and platform layer of media cloud. Specifically, we begin with a review on the basic concepts of various machine learning techniques. Then, we examine the system architecture of media cloud, focusing on the functionalities in the infrastructure and platform layer. For each of these function and its corresponding challenge, we further illustrate the adoptable machine learning based approaches. Finally, we present an outlook on the open issues in this intersectional domain. The objective of this article is to provide a quick reference to inspire the researchers from either machine learning or media cloud area.
