A computational fluid dynamics (CFD) model that includes representations of effects of unsteady cavitation and associated dynamic loads has been developed to increase the accuracy of simulations of the performances of turbopumps. Although the model was originally intended to serve as a means of analyzing preliminary designs of turbopumps that supply cryogenic propellant liquids to rocket engines, the model could also be applied to turbopumping of other liquids: this can be considered to have been already demonstrated, in that the validation of the model was performed by comparing results of simulations performed by use of the model with results of subscale experiments in water.

The need for this or a similar model arises as follows: Cavitation instabilities in a turbopump are generated as inlet pressure drops and vapor cavities grow on inducer blades, eventually becoming unsteady. The unsteady vapor cavities lead to rotation cavitation, in which the cavities detach from the blades and become part of a fluid mass that rotates relative to the inducer, thereby generating a fluctuating load. Other instabilities (e.g., surge instabilities) can couple with cavitation instabilities, thereby compounding the deleterious effects of unsteadiness on other components of the fluid-handling system of which the turbopump is a part and thereby, further, adversely affecting the mechanical integrity and safety of the system.

Therefore, an ability to predict cavitation-instability-induced dynamic pressure loads on the blades, the shaft, and other pump parts would be valuable in helping to quantify safe margins of inducer operation and in contributing to understanding of design compromises. Prior CFD models do not afford this ability. Heretofore, the primary parameter used in quantifying cavitation performance of a turbopump inducer has been the critical suction specific speed at which head breakdown occurs. This parameter is a mean quantity calculated on the basis of assumed steady-state operation of the inducer; it does not account for dynamic pressure loads associated with unsteady flow caused by instabilities. Because cavitation instabilities occur well before mean breakdown in inducers, engineers have, until now, found it necessary to use conservative factors of safety when analyzing the results of numerical simulations of flows in turbopumps.

The model has been implemented within CRUNCH CFD, which is a proprietary CFD computer program that has been extensively tested and validated for predicting mean pump performances. The provision of the capability to simulate cavitation instabilities involved two major enhancements of CRUNCH CFD: (1) incorporation of a capability to model the varying properties of real cryogenic fluids in the presence of cavitation and (2) development of more-sophisticated physical submodels that, in comparison with corresponding prior models, represent more accurately the rates at which vapors are created or condensed back to liquids.

The model has been demonstrated to provide accurate estimates of the magnitudes and frequencies of unsteady pressure loads on inducer blades. The model has also been demonstrated to enable estimation of loads on pump shafts and bearings. The ability to estimate these loads is important because (1) it is difficult to measure such loads in experiments and (2) a high load on a shaft can cause an inducer to rub against the shroud that houses the pump and, in the worst case, can result in seizing or in failure of the shaft.

This work was done by Ashvin Hosangadi, Vineet Ahuja, Ronald Ungewitter, and Sanford M. Dash of Combustion Research and Flow Technology, Inc. for Marshall Space Flight Center. For further information, contact Sammy Nabors, MSFC Commercialization Assistance Lead, at This email address is being protected from spambots. You need JavaScript enabled to view it.. Refer to MFS-32586-1.

NASA Tech Briefs Magazine

This article first appeared in the March, 2009 issue of NASA Tech Briefs Magazine.

Read more articles from this issue here.

Read more articles from the archives here.