Available at: http://digitalcommons.calpoly.edu/theses/284
Date of Award
MS in Mechanical Engineering
The Mechanical Engineering Department at California Polytechnic State University in San Luis Obispo currently maintains a lab-scale hybrid rocket motor for which nitrous oxide is utilized as the oxidizer in the combustion system. Because of its availability, the same two-phase (gas and liquid) nitrous oxide that is used in the combustion system is also routed around the throat of the hybrid rocket’s converging-diverging nozzle as a coolant. While this coolant system has proven effective empirically in previous tests, the physics behind the flow of the two-phase mixture is largely unexplained. This thesis provides a method for predicting some of its behavior by modeling it using the classic gas dynamics scenarios of Rayleigh and Fanno flows which refer to one-dimensional, compressible, inviscid flow in a constant area duct with heat addition and friction. The two-phase model produced utilizes a separated phase with interface exchange model for predicting whether or not dryout occurs. The Shah correlation is used to predict heat transfer coefficients in the nucleate boiling regime. The homogeneous flow model is utilized to predict pressure drop. It is proposed that a Dittus-Boelter based correlation much like that of Groeneveld be developed for modeling heat transfer coefficients upon the collection of sufficient data.
Data was collected from a series of tests on the hybrid rocket nozzle to validate this model. The tests were first run for the simplified case of an ideal gas (helium) coolant to verify the experimental setup and promote confidence in subsequent two-phase experimental results. The results of these tests showed good agreement with a combined Rayleigh-Fanno model with a few exceptions including: (1) reduced experimental gas pressure and temperature in the annulus entrance and exit regions compared to the model and (2) reduced experimentally measured copper temperatures uniformly through the annulus. These discrepancies are likely explained by the geometry of the flowpath and location of the copper thermocouples respectively. Next, a series of two-phase cooled experiments were run. Similar trends were seen to the helium experiment with regards to entrance and exit regions. The two-phase Rayleigh homogeneous flow model underpredicted pressure drop presumably due to the inviscid assumption. Ambiguity was observed in the fluid temperature measurements but the trend seemed to suggest that mild thermal non-equilibrium existed. In both cases, the dryout model predicted that mist flow (a post-CHF regime) occurred over most of the annulus.
Several modifications should be implemented in future endeavors. These include: (1) collecting more data to produce a heat transfer coefficient correlation specific to the nitrous oxide system of interest, (2) accounting for thermal non-equilibrium, (3) accounting for entrance and exit effects, and (4) developing a two-phase Fanno model.