Comprehensive computational modeling of hypergolic propellant ignition

Swanand Sardeshmukh, Purdue University

Abstract

Ignition and combustion of hypergolic propellants mono-methyl hydrazine (MMH) and red fuming nitric acid (RFNA) is investigated computationally. A hierarchical approach is chosen to study parametric behavior of isolated processes and complex interactions thereof, in this transient phenomenon. Starting with a homogeneous reactor, performance of three reduced kinetic mechanisms is assessed first, followed by the study of auto-ignition delay as a function of initial composition and thermal state of the mixture. Macroscopic features as well as the structure of opposed diffusion flame are studied next, followed by the study of opposed liquid jets and the gas layer at the interface. Lastly, effects of transport properties on gas phase kinetics are studied using impinging vapor sheets. In order to mitigate the effects of non-linearities in the transient solution, an adaptive time stepping method is proposed for the homogeneous reactor. In this method, physical bounds are imposed on the explicit guess of a solution for assessing linear behavior, based on which, level of time step adaption is determined. This method is ascertained to provide accuracy necessary with marginal increase in the cost of computations. It is anticipated that such a method, upon adoption in multi-dimensional computations of reacting flows, can capture accurate flame location and structure due to enhanced resolution of the non-linear processes. Temporal history of various species during the auto-ignition of premixed vapors is utilized to understand the well known behavior of hypergolic propellants MMH and RFNA, namely dependence on thermal state of the mixture and relatively smaller ignition delay variation for oxidizer rich mixtures. In particular, initial assimilation of \ce{NO2} followed by subsequent release is found to be the key mechanism in ignition of the premixed vapors. Mapping the ignition delay as a function of initial state of the mixture, vapor temperatures above 650 K showed realistic ignition delays. Rich mixtures were seen to have up to two orders of magnitude variation in ignition delay compared to less than an order of magnitude variation for fuel lean cases. Further relaxing the assumption of superheated vapors, premixed liquids showed competing vaporization and liquid reaction, affecting the final temperature and ignition delay in the gas phase. Investigation of the diffusion flame is carried out under varying strain rates, operating pressures and inlet temperatures with consideration of macroscopic features: peak temperature and flame width, along with the detailed flame structure. Primary effect of increasing the strain rate is found to reduce the flame residence time, providing a cut-off for the post-flame reactions, in turn, reducing the peak temperature and width of the diffusion flame. Reduced diffusion at higher pressures limits the pre-flame reactions while variations in concentrations of the reactants at higher and lower pressures lead to different reaction paths. Change in the inlet temperatures is found to be responsible for reducing the concentrations of oxidizer derived species, affecting the width of the flame. Focusing on the gas layer between the liquid jets, it is seen that the liquid reaction is a surface phenomenon, leading to instabilities due to rapid volumetric expansion, which is characteristic of phase change. Resulting instabilities are anticipated to promote liquid mixing, reactions and additional instabilities. However, intermittent contact and limited liquid interaction is seen to be the cause of low average temperatures that are significantly lower than the vapor phase ignition threshold. Lastly, behavior of impinging vapor sheets is studied to discern the effects of pressure and transport through bath gases helium, neon and argon, on the ignition process. Injection and location of ignition are found to be affected by pressure due to altered diffusivities, while the variation in transport properties is found to affect the location of ignition as well as the shape and spread of subsequent flame. While greater diffusion through helium is found to allow smooth interface, batch mixing due to roll-up in argon, is seen to allow isolated reaction regions, that can lead to an earlier ignition due to retention of intermediate species within the reaction zone. The mode of upstream flame spread from ignition location in helium is therefore found to be smooth, while in argon, ignition is seen to occur at multiple locations that subsequently connect, leading to earlier formation of a continuous flame front. The hierarchical exploration of physics involved in ignition and combustion of hypergolic propellants MMH and RFNA has established a structured framework for modeling, while quantifying various influencing factors. Macroscopic features and detailed structure of an opposed diffusion flame is studied for the first time, noting the effects of strain rate and operating conditions for the vapor phase propellants MMH and RFNA. Two phase reacting flow simulations showed evidence of instabilities at very small time and length scales, indicating feasibility of modeling phenomena like popping or liquid stream separation, observed in practice. Alteration in the transport properties and corresponding change in vapor phase mixing, ignition and flame propagation revealed that higher diffusivity in helium leads to smooth ignition and flame propagation while argon shows isolated mixing, earlier ignition and faster flame spread due to several spot ignitions.

Degree

Ph.D.

Advisors

Heister, Purdue University.

Subject Area

Engineering|Chemical engineering

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS