Content area
Abstract
In the present research flow issuing from a rectangular nozzle onto a wall (the wall jet) is investigated. This flow has a very interesting feature, an anisotropic spreading rate, faster spreading in the lateral direction than in the transverse direction. The thesis identifies flow features, i.e., organized structure within the flow, responsible for the spreading. Structure is defined, and investigated in a number of different flows.
Two new methods are introduced for flow investigation. The first is the Gram-Charlier (GC) technique. This method allows the reconstruction of the spatial and temporal cross-correlations, $R\sb{ij},$ with Hermite polynomials in order to remap a sparsely sampled dataset onto a grid with increased resolution. Gc was applied to the wall jet, as well as the mixing layer (8). The Hermite polynomials reconstruct well the temporal cross-correlations. The even and odd polynomials reconstruct characteristics of the mixing similar to those found with the proper orthogonal decomposition (POD). A new two-dimensional formulation is shown to reconstruct the overall spatio-temporal features found in the cross-correlation.
The second method introduced is Enhanced variable Interval Time Averaging (EVITA). EVITA filters a flow and examines it at a number of different scales. The technique is compared with VITA, and shown to be a significant improvement. The method is applied as one and two dimensional investigative functions. The one-dimensional function is found to extract predefined features quite well. In the channel (58), the two-dimensional form extracts near-wall structures with minimal predefinition. EVITA is shown to act as a "mathematical microscope" by varying the scale of the function and location.
The experimental work involves multipoint measurements made in the wall jet. The results are compared with Schwab (68) and good agreement is found with the present results, despite significant differences in Reynolds number. GC is combined with Linear Stochastic Estimation (LSE) to remap the experimental data onto a finer grid, and the resulting statistics compare well with the present measurements and those of Schwab (68). Using LSE, reconstructions of the instantaneous field were done. The constructions show regions of organized flow within the field. Wavelet power spectral densities (PSD) are introduced and used to determine scales important in the long-time mean. The results of the wavelet PSD show the evolution of structure from the initial measurement region to the final measurement region.
From the results a mechanism for the growth of the wall jet has been postulated. The jet exits from the nozzle as either a vortex ring or a vortex array. As it moves downstream and is constrained by the wall the ring flattens out, and develops into a large hairpin on the periphery of the jet. This structure acts to constrain growth normal to the jet and enhance growth parallel to the wall.
