Transition prediction for three-dimensional boundary layers in computational fluid dynamics applications

A method is presented for estimating the laminar/turbulent transition location in three-dimensional boundary layers for computational fluid dynamics (CFD) applications requiring numerous transition estimates with no user intervention. Given the Reynolds number and the [C.sub.P] distribution, the location of transition is estimated based on the attachment-line state, the potential for relaminarization, the occurrence of laminar separation, and the growth of instabilities. Transition caused by instability is estimated based on N factors calculated for Tollmien-Schlichting waves and for stationary crossflow instabilities. A neural network is used (in place of solving the Orr-Sommerfeld stability equation) for determining the instability growth rates. The current version of the method assumes incompressible flow. The boundary-layer flow and instabilities are calculated based on an infinite-sweep (strip boundary layer) approximation; the instability calculations also employ the parallel-flow approximation. Comparisons with traditional stability codes show good N-factor agreement over a practical range of [C.sub.P] distributions. The method is several hundred times faster than traditional stability calculations, and it is robust enough to function as a simple 'subroutine' in CFD codes. The method is biased toward application efficiency and simplicity as balanced against improvements in the detailed physical modeling.