Manish Deshpande

Cavity Flow Predictions Based on the Euler Equations

An Euler solver based on artificial-compressibility and pseudo-time stepping is developed for the analysis of partial sheet cavitation in two-dimensional cascades and on isolated airfoils. The computational domain is adapted to the evolution of the cavity surface and the boundary conditions are implemented on the cavity interface. This approach enables the cavitation pressure condition to be incorporated directly without requiring the specification of the cavity length or the location of the inception point. Numerical solutions are presented for a number of two-dimensional cavity flow problems, including both leading edge cavitation and the more difficult mid-chord cavitation conditions. Validation is accomplished by comparing with experimental measurements and nonlinear panel solutions from potential flow theory. The demonstrated success of the Euler cavitation procedure implies that it can be incorporated in existing incompressible CFD codes to provide engineering predictions of cavitation. In addition, the flexibility of the Euler formulation may allow extension to more complex problems such as viscous flows, time-dependent flows and three-dimensional flows.

Numerical Modeling of the Thermodynamic Effects of Cavitation

A Navier-Stokes solver based on artificial compressibility and pseudo-time stepping, coupled with the energy equation, is used to model the thermodynamic effects of cavitation in cryogenic fluids. The analysis is restricted to partial sheet cavitation in two-dimensional cascades. Thermodynamic effects of cavitation assume significance in cryogenic fluids because these fluids are generally operated close to the critical point and also because of the strong dependence of the vapor pressure on the temperature. The numerical approach used is direct and fully nonlinear, that is, the cavity profile evolves as part of the solution for a specified cavitation pressure. This precludes the necessity of specifying the cavity length or the location of the inception point. Numerical solutions are presented for two-dimensional flow problems and validated with experimental measurements. Predicted temperature depressions are also compared with measurements for liquid hydrogen and nitrogen. The cavitation procedure presented is easy to implement in engineering codes to provide satisfactory predictions of cavitation. The flexibility of the formulation also allows extension to more complex flows and/or geometries.

The Effect of Swirl on Mixing Characteristics of Turbulent Shear Layers

Simulations utilizing a model based on the FavreAveraged Navier Stokes equations coupled with a k — e turbulence model have been carried out to investigate the impact of swirl on the mixing characteristics of reacting turbulent shear layers. A number of different coaxial injector configurations have been used in the study for both reacting and nonreacting flows. It has been attempted to incorporate the various attributes of swirling motions and gauge the impact that each of these variances would have on flame-dynamics and mixing characteristics. The various parameters studied include swirl intensity, initial swirl distribution, expansion ratio and splitter plate thickness.

Application of a distributed network in computational fluid dynamic simulations

A general-purpose 3-D, incompressible Navier-Stokes algorithm is implemented on a network of concur rently operating workstations using PVM and com pared with its performance on a CRAY Y-MP and on an Intel iPSC/860. The problem is relatively computa tionally intensive, and has a communication structure based primarily on nearest-neighbor communication, making it ideally suited to message passing. Such problems are frequently encountered in CFD, and their solution is increasingly in demand. The communica tion structure is explicitly coded in the implementation to fully exploit the regularity in message passing in order to produce a near-optimal solution. Results are presented for various grid sizes using up to eight pro cessors.

Implementation of a parallel algorithm on a distributed network

The objective of this research is to investigate the potential of using a network of concurrently operating workstations for solving large computer-intensive problems typical of Computational Fluid Dynamics. Such problems have a communication structure based primarily on nearest-neighbor communication and are therefore ideally suited to message passing. The implementation of a 3D Navier-Stokes code on a network of IBM RISC/6000s is described. The performance of this code is compared with that on conventional high-performance machines such as the Cray-YMP and the Intel iPSC/860. The results suggest that a cluster of workstations presents a viable and economical resource for this purpose. Additionally a workstation network has the advantage of easy accessibility, fault tolerance and a simple environment for program development. 12 refs.

Practical Applications of Parallel Processing in Computational Fluid Dynamics

The onset of parallel computing has had a profound impact on CFD, both in terms of its utility as an analysis and design tool, and its capability in adequately resolving flowfields that exhibit complex physical phenomena. The resources provided by a distributed network of processors cohesively working together provides an opportunity to tackle some of the most challenging problems in fluid dynamics, that have proved elusive to serial supercomputing. Turbomachinery and combustion are two such areas of fluid dynamics that have benefited immensely from the increase in resources afforded by parallel machines. In this paper, we discuss the parallel implementation of conventionall y popular implicit and explicit serial CFD algorithms and utilize them to solve problems of increased complexity in turbomachinery and combustion. Specific applications in turbomachiner y that are discussed in this paper include the rotor-stator interaction problem, analysis of tip vortex generation, multiple passage simulations for non-periodic distortions and cavitation. Also discussed in the paper are problems related to combusting flowfields in rocket injectors that require fine local resolution necessary to capture the wide range of dominant characteristic scales.