Numerical optimization of S-shaped intake performance using special cross-sectional profiles
S.H. Sadatpour, A. Madadi, R. Ahadian
Amirkabir University of Technology, Tehran, Iran
Keywords: S-shaped intake, genetic algorithm, artificial neural network, flow uniformity, pressure recovery
Abstract
An intake has to provide air for the engine uniformly with minimum total pressure loss. Nowadays, regarding the usage of S-shaped intakes, optimization of these ducts has been considered. Uniform distribution of flow at the compressor inlet directly influences the engine performance, and non-uniformity of flow increases surge occurrence possibility. Flow separation along the duct causes a reduction of pressure recovery and engine thrust force. This research has optimized an S-shaped intake to reduce the total pressure loss and flow distortion. The genetic algorithm and artificial neural networks have been combined to decrease the computational cost. Two optimizations, using different conditions, have been studied. In the first case, by modifying centerline coordinates and area ratio of sections, new geometries have been produced, which have caused an improvement of 32.5 % in pressure recovery coefficient and a decrease of 35.8 % in flow distortion. In the second optimization, the shape of each section has also been changed. Super ellipse, egg-shaped and circular profiles are considered as cross sections of the duct. The second optimization has improved the pressure recovery coefficient by 35.5 % and decreased flow distortion by 39.2 %.
|
