Numerical study of the injection process in a transonic wind tunnel: The numerical details

Abstract

Numerical prediction of an injection process was successfully performed. The physical problem corresponds to the mixing of a number of parallel supersonic jets with a subsonic main stream, in the presence of solid walls. In practice, this arrangement is to be used as a boosting device in a transonic wind tunnel, with the ultimate objective of extending the tunnel’s envelope without penalizing the main compressor. Five supersonic nozzles are installed at the floor and five at the ceiling of the transition module’s entrance section. Due to the great difference between cross-sectional typical lengths of nozzles and tunnel, the numerical tool has to have a three-dimensional capability. The core of the code is an adaptation of the finite-difference diagonal algorithm, and turbulence effects are properly tackled by the use of the Spalart and Allmaras one-equation scheme. Some simplifications were adopted in order to render the problem minimally tractable, especially in this initial numerical simulation stage, which is the reporting objective of this article. Albeit this, and due to the magnitude of the problem in hand, these simplifying initiatives were not sufficient to bring the calculations down to a scale of reasonable computer costs. Hence, a sequence of grids, coupled with a division of the computational domain, was adopted. Boundary conditions were thoroughly worked, especially the ones related to the domain of calculation’s entrance plane. This plane is important because of the many viscous gradient regions that project against it. Among the many settings under which a tunnel can operate there is one defined as the “design condition”, or else as the “design point”, as it is sometimes called. For a tunnel equipped with an injection system, the design point asks (besides many other specifications) for equal static pressures at the section where the supersonic jet meets for the first time the main stream coming from the tunnel circuit. We have simulated five different operating settings of the tunnel and among them the design condition. Therefore, we have results for the design as well as for many off-design points. The idea was to simulate the many tunnel settings and also to test the code’s ability to handle all these different situations. The code was duly validated and verified, and in the sequel the steady flow field in the mixing region was calculated. Many interesting and very important results were obtained, among which we would point out the existence of compression and expansion domes at the supersonic regions, the calculation of the many engineering parameters related to the injection process, and especially the determination of the injection’s gain threshold.