Abstract
Abstract Accurate and efficient predictions of the steady and unsteady flow responses due to the blade-to-blade variation as well as due to the nonaxisymmetric inlet distortion have been continually pursued. Computation of two problems concurrently has been rarely done in the past partly because of the need to perform whole annulus bladerow simulations, despite the advances in the current state-of-the-art methods with the phase-shift single passage simulations. The current work attempts to deal with this challenge by developing a new computational approach based on the principle of the multiscale method in the framework of a commercial solver (CFX). The methodology formulation relies on summation of the constituent source terms, each of which corresponds to a particular flow perturbation. The source term element corresponding to the blade-to-blade variation effect is linearly superimposed as in the classical Influence Coefficient Method. The unsteady flow field around a blade at any time instant depends only on its relative position to all its neighboring blades, so that the influences of an arbitrarily mis-staggered bladerow can be computed efficiently. In addition, the source term arisen due to the inlet distortion is calculated based on the spatial Fourier transform. A key enabler is that the source terms can be precomputed using a small set of identical blade passages. The source term is then propagated to different spatial and temporal locations depending on the combination of the mis-staggering pattern and the inlet distortion. The multiscale treatment makes it possible to predict a high-resolution flow field effects on the base coarse mesh as if a fine mesh were locally solved, while achieving a considerable computational efficiency gain. The proposed influence-coefficient and source term based method has been validated for test cases with a uniformly staggered bladerow, and for an arbitrarily mis-staggered bladerow, under a clean inflow condition as well as that subject to an inlet distortion.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have