Within a scramjet, the location of the shock train will change as the backpressure changes. When the shock train leading edge moves out from the entrance of the isolator, the engine usually goes into unstart. A precise shock train leading edge detection technique is indispensable for controlling the shock train within the isolator. In this Paper, a reconstruction-flowfield-based shock train leading edge detection method is proposed. A flowfield reconstruction model based on a convolutional neural network provides instantaneous flowfield structure, and then image processing is carried out on these flowfield images to obtain the shock train leading edge. A supersonic direct-connect wind tunnel was used to collect all the data, which were then postprocessed to train the convolutional neural network model. The trained model successfully reconstructed the flowfield structure based on the measured wall pressure, and the reconstruction effects under different pressure transducer arrangements were compared. The flowfield reconstruction model provides a richer information source for shock train leading edge detection, and the detection accuracy can be greatly improved. The proposed method was compared with the pressure ratio method and the pressure increase method, and the detection performance under different pressure transducer arrangements was examined.
Flowfield Reconstruction and Shock Train Leading Edge Detection in Scramjet Isolators
AIAA Journal ; 58 , 9 ; 4068-4080
2020-09-01
Article (Journal)
Electronic Resource
English
Scramjet Isolator Shock-Train Leading-Edge Location Modeling
British Library Conference Proceedings | 2011
|Shock Train Leading-Edge Detection in a Dual-Mode Scramjet
Online Contents | 2008
|