Experimental investigation of suddenly expanded flow at sonic and supersonic Mach numbers using semi-circular ribs: a comparative study between experimental, single layer, deep neural network (SLNN and DNN) models

Abstract In this work, we present the findings of experimental study conducted in a rectangular duct at sonic and supersonic Mach numbers using passive control form semi-circular ribs. Tests are number four numbers. The 1.5, 1.8, 2.2, 2.5. flow from nozzles is discharged into enlarged duct. ribs placed 28 mm (1D), 56 (2D), 84 (3D), 112 (4D) base to find effect mechanism on field pressure. 6, 8, 10 diameter used pressure ultimately drag. At 2.2 2.5, not effective because over-expanded. These results reiterate literature that whether or active when under influence favorable gradient. same evident 1.5 1.8. NPRs these such under, correctly, expanded. When operated for expanded case, an increase employed. highly complex data predicted single-layered neural network deep-layer save time make it cost-effective, which shows can be with accuracy 0.88–0.99. proposed models predict sensitive terms aerodynamic flows.