Using Tapered Channels to Improve LAD Performance for Cryogenic Fluids: Suborbital Testing Results

Many innovations will be required in critical spacecraft systems, including propulsion technology, to enable long duration space missions. Advancing cryogenic storage and transfer technology is a continuing area of research among spacecraft designers. The contents of spacecraft propellant tanks comprise both liquid and vapor, and in a low gravity environment, the liquid propellant will tend to drift and adhere to tank walls due to the lack of buoyancy forces. One of the main challenges in tank design is to ensure the delivery of vapor-free liquid to an engine during an engine restart or to a receiving tank during a fuel transfer process.

When the propellants are storable and have a low vapor pressure (e.g., hydrazine), a flow of vapor-free propellant is accomplished using propellant management devices (PMDs) that utilize capillary or surface tension forces to capture the liquid in the tank. A common PMD is a capillary liquid acquisition device (LAD) that enables the safe transfer of liquid propellants from anywhere inside of the tank to the outflow system, as shown in Figure 1, using screen channels. Such a device was used in the Space Shuttle Orbital Maneuvering System propellant tanks. When liquid outflow commences, surface tension forces within the screen mesh will act as a barrier to vapor flow so that only liquid flows through the pores of the screen. The screen's resistance to vapor passage, known as the bubble point pressure, is dependent on the liquid surface tension and the screen's pore size. Vapor flow is blocked as long as the pressure differential across the screen is less than the bubble point pressure.

Figure 1

Using LADs with cryogenic propellants has yet to be demonstrated because heat leaking into the saturated fluid will vaporize the liquid within the channels, creating vapor bubbles. Since these vapor bubbles are within the screened channel, the screen bubble point pressure is no longer relevant. After enough vaporization occurs, the screens will dry and the capillary barrier to the vapor inflow from the tank will no longer be present. Therefore, designing a LAD that can passively remove the vapor generated internally is ideal.

It is proposed that a simple change to the existing LAD geometry can be incorporated to reduce the risk of vapor bubbles being present during propellant transfers or in engines for long duration space missions. The modified design utilizes a tapered tube such that the channel's cross-sectional area increases with distance away from the tank outlet, as shown in Figure 2. The principle of the design is that capillary or surface tension forces will drive the vapor bubble along the channel to the wider end of the channel, where it can vent back into the tank through a window screen. The bubble clearing prevents the accumulation of vapor bubbles in the channel, keeps the screens wetted, and maintains the desired capillary barrier from vapor flow entering the channel interior with a simple design that does not add additional mass to the system. The development of this technology for long duration space missions would reduce costs associated with complex propellant tank designs, increase the efficiency of tank filling operations, and reduce or eliminate propellant loss from thrusting to orient the ullage in the storage tank.

Figure 2

Several research groups have previously studied capillary flows in tapered channels. A numerical model of capillary motion inside of tapered channels has been developed (Weislogel et al., 2011; Metz et al., 2009) as well as fundamental experiments involving vapor droplets that were driven by bubble injection and by surface tension forces were accomplished in a 1-g environment (Hartwig et al., 2014). In addition, the capillary flow experiments aboard the International Space Station (CFE-1, CFE-2, IFCI, VGI) studied critical wetting phenomena in different capillary geometries, including tapered channels (Blackmore et al., 2011). However, these studies were accomplished either by bubble shake tests, channel rotation, or by actively pumping fluid and vapor into the channels, generating initial momentum-driven bubble migration. These experiments were also conducted in channels with large tapered angles, where vapor bubbles did not span the entire channel width. Past results of two-phase flow in channels have shown that fluid motion and interaction are vastly different in microchannels compared with channels or tubes that are larger (Colin et al., 1996). The work on tapered channels accomplished at Southwest Research Institute (SwRI) also indicated that the test channel cross-sectional dimensions, including the taper angle, must be small, so that the bubble spans the entire channel width for surface tension forces to move the bubble along the channel in full-scale applications (Dodge et al., 2002). Excluding the research accomplished by SwRI on bubble motion in screen channels for cryogenic liquid management, experiments that incorporate tapered screen LADs that demonstrate bubble motion through surface tension forces only are not reported in the literature.

An analytical model derived by SwRI from the conservation of mass and momentum predicts the bubble velocity within tapered LADs. A ground experiment was designed to simulate bubble movement within a tapered channel and to provide data for model validation. The ground experiment utilized isopropanol and an air bubble as this fluid combination provided favorable scaling parameters based on the Ohnesorge number. In addition, the ground test unit was also designed to minimize gravity effects by limiting the channel depth, thus maximizing the aspect ratio of the bubble. Figure 3 shows a time lapse of the bubble movement through the LAD from the ground test. As shown in Figure 3, the bubble is elongated when it is first injected into the LAD and then becomes spherical at the wider end due to the surface tension difference between the leading and trailing edges of the bubble. These ground experiments investigated different taper angles (2.5°, 5°) and bubble sizes (25 μL, 50 μL) to compare with the model. A subset of the results with a 5° taper angle is shown in Figure 4. In all tests, the bubble moved toward the wider end of the channel until the bubble was nearly spherical. The initial and final bubble velocities for the tests with a 5° taper angle were on the order of 2.5 mm/s and 0.6 mm/s, respectively. The model was able to predict bubble velocities within 0.2 mm/s of the rates recorded in the 5° ground tests.

Figure 3

Figure 4

The next step in advancing tapered LAD technology was to test the concept in a relevant microgravity environment. Utilizing funding from NASA's Flight Opportunities Program, several tapered channels were tested aboard Blue Origin's New Shepard vehicle in December 2019. This vehicle provided more than 3 min of high-quality microgravity. This duration of microgravity is beneficial for this test program because it takes at least 1 min for the bubble to migrate to the wider end of the channel. A photograph of the tapered LADs flown in this experiment is shown in Figure 5. LADs 3, 4, and 5 were identical to the 5° ground test geometry and LADs 1 and 2 were scaled-up versions that utilized an 8° and 5° taper angle, respectively. Utilizing LADs in the flight test that have the same geometry as the ground test allows for gravity effects to be isolated and characterized. It is also important to recognize that larger scale tapered LADs cannot be tested with gravity present because larger bubbles will not be in contact with all surfaces in the channel. The fluids used in this test included isopropanol, HFE-7000, and water to study the effect that fluid properties have on bubble movement. Video cameras recorded bubble movement during all phases of the flight.

Figure 5

The suborbital test showed that the LADs that contained HFE-7000 (LAD 4) and isopropanol (LAD 5) showed bubble movement toward the wider end of the channel during the microgravity phase of the flight. The LADs that contained water (LADs 1, 2, and 3) did not show any bubble movement. The leading and trailing edges of the bubbles were tracked over time with edge detection and cross-correlation image processing algorithms in MATLAB^® and are displayed in Figure 6 alongside the ground test data. The results presented in Figure 6 show that the bubble velocity in the ground test is much higher and nonlinear when compared with the flight test. The bubble velocity in the ground test averaged about 1.5 mm/s in the first 5 s of the experiment and then reduced as the bubble geometry became more spherical and the surface tension force was reduced between the bubble leading and trailing edges. The bubble velocities in the HFE-7000 and isopropanol from the flight tests were nearly linear and averaged about 0.1 mm/s and 0.05 mm/s, respectively. Another clear difference is that the bubble in the HFE-7000 LAD from the flight test traveled twice the distance as compared with the bubble in the isopropanol LAD from the ground and flight tests.

Figure 6

There were two main variations between the ground test and the flight test that can explain the differences observed in the bubble movement. First, the bubbles in the isopropanol and the HFE-7000 LAD were two and three times the volume, respectively, as compared with the bubble utilized in the ground test. Efforts were made to maintain similar bubble volumes, but due to the volatility of the fluids and the relatively small liquid volumes within the LADs (about 1.5 mL), evaporation of the liquid resulted in larger bubbles. These larger bubbles will travel further in the same size LAD because the difference in surface curvature between the leading and trailing edges, the capillary driving force, can be maintained over a longer distance. In addition, larger bubbles should generate higher initial velocities; this was, however, not observed in the flight data. This reduced velocity is a result of the second main difference between the ground tests and the flight tests: the bubble was manually injected into the LADs used during the ground tests, and the flight test hardware was designed to operate unattended. As shown in Figure 2, when surface tension drives a bubble out of a tapered channel, the liquid in front of the bubble is expelled from the channel, and liquid from the tank is drawn into the channel behind the bubble. The ground test simulated this behavior by utilizing an open-loop design with an upstream and downstream reservoir, while the flight test was designed with a closed passage that required the displaced liquid to flow around the bubble to the narrow end of the channel. The closed passage resulted in increased viscous drag in the liquid film between the bubble and the LAD channel, which linearized the velocity profile of the bubble. Despite these two differences that affect bubble movement physics, it is important to note that significant bubble movement was still observed in the flight test, thus demonstrating that tapered LAD channels could be utilized to passively remove internally generated vapor bubbles.

Tapered LAD research will be further explored in two additional funded suborbital tests that will expand on several lessons learned from this first flight. First, water will not be utilized as the working fluid. Despite water having a favorable scaling relationship to cryogens, the contact line drag (also known as “stiction”) between the bubble and LAD channel surface could not be overcome by surface tension forces only. Second, a liquid passage will be incorporated into the closed-loop design to allow the displaced liquid in front of the bubble to return to the narrow end of the channel behind the bubble. Finally, future designs will incorporate better seals in the fill and drain ports to minimize the liquid boil-off prior to the launch in an effort to better control the bubble size. The next suborbital test, currently planned for fall 2020, will refly several LADs used in the previous test, but they will be modified to incorporate the liquid passage as shown in Figure 7. In addition, the next test will include four larger scale versions of the LAD to study bubble movement as the bubble nears the end of the channel. These LADs will also utilize isopropanol and HFE-7000 as the working fluids.

Figure 7

The basic principles of tapering LAD channels to passively manage internally generated cryogenic vapor bubbles have been established. The concept has been successfully tested in ground experiments as well as in a relevant microgravity environment aboard a suborbital flight. There were several experimental improvements realized from the first tapered LAD suborbital test that will be incorporated into the next two suborbital flight tests. These funded experiments will provide model validation data at different scales and will utilize fluids that simulate cryogenic properties to advance tapered LAD design strategies. However, it is important to recognize that tapered LAD tests will eventually need to incorporate cryogenic test fluids and long duration microgravity in field-scale hardware to validate the design for future long duration space missions that utilize cryogenic storage tanks.