1
Crushing Martian Regolith for the Preliminary Development of Water Extraction Hardware
MAE 435 – Project Design & Management II
Final Report
Team Members:
Nicholas Sestito
Christopher Graham
Interdisciplinary Member:
Ian Kaplan (CE)
Project Advisor:
Dr. Robert Ash
Table of Contents
Table of Contents
List of Figures
Abstract
Introduction
Methods
Input Sizing Justification
Output Sizing Justification
The lumped System Calculation
Preliminary Regolith Crushing Method Research
Crusher selection
Stress Analysis of an Ice Block
Crushing Power Estimate
Discussion
Conclusion
Reference Page
Appendices
Appendix 1: Gantt Chart
Appendix 2: Thermal Model Sample Calculation
List of Figures
Figure 1: Water Extraction Schematic
Figure 2: Drag Coefficient as a function of Reynolds number [1]
Figure 3a: Cd Vs Log10 (CdRe^2)
Figure 4: pipe length vs grain diameter
Figure 5: pipe diameter vs grain diameter
Figure 6: Basic Jaw Crusher Diagram
Figure 7: Ice Block Wireframe
Figure 8: von Mises stress distribution
Figure 9: Displacement in an ice block
Figure 10: Scaled Chipmunk Jaw Crusher
Abstract
With increasing interest and research being done towards the exploration and colonization of Mars, one important driver will be determination of the feasibility of extracting water from Martian surface material as a local feedstock for the production of consumables. Orbiters andMars surface Rovers have generated sufficient data confirming the presence ofsignificant quantities of water within the Martian regolith to justify the development of preliminary hardware designs to demonstrate the feasibility of water excavation. Furthermore, research and testing by an earlier ODU design team has been performed to determine the mechanical behavior of permafrost made from a Martian Regolith Simulant, in orderto guide the design of harvesting equipment. These tests demonstrated that the excavation and removal of one litersize blocks of frozen Martian regolith/ice mix can be accomplished. Extraction of water from these cryogenic permafrost blocks will be too slow and inefficient, making it necessary to crush these 1000 cc blocks into small-diameter particles that can be heated to liberate water more easily. This work is an attempt to develop a crushing process to produce optimized permafrost grains that are suitablefor efficient water extraction. Through a Lumped system thermal analysis, a practical heating processhas been developed to guide the output regolith grain diameters, over a range from one to five millimeters.
1
Introduction
As a consequence of the 27-month, low energy Mars cargo delivery interval, it will not be possible to provide the consumables required for a crewed base at Mars without making use of local materials. Furthermore, the propellant mass that would be required for crewed return flights from Mars to Earth exceed planned launch vehicle capabilities. Most consumable materials and chemical rocket propellants can be produced from water and carbon dioxide, and more than 95 % of the Martian atmosphere is carbon dioxide. Consequently, if a reliable source for water can be found at Mars, it should be possible to use that water and the Martian atmosphere to produce a variety of consumables and rocket propellant. One method for producing methane and oxygen for rocket propellant is based on the Sabatier reaction, which can use hydrogen produced in the electrolysis of water, along with CO2 from Mars atmosphere[2].
Since CO2 is plentiful in the Martian atmosphere, the only molecule needed to start producing fuel for a return trip can be hydrogen. Hydrogen is the smallest molecule, and is the lightest to carry into space, however, storing large quantities over long periods of time is problematic, because it must be maintained as a liquid at temperatures on the order of 20 K. The surface of Mars is covered in a fine soil like material known as regolith, which shields icy material from the sun, reducing its sublimation rate[3]. If icy regolith is harvested, electrolysis can convert the segregated and distilled water extracted from the regolith into hydrogen and oxygen, which can eliminate the need to haul hydrogen from earth[2].
The Mars surface rovers have conducted scientific research on the surface of Mars. The Phoenix Lander in particular gathered data in regard to regolith properties, including soil water ice concentration and density. The Phoenix lander also observed surface ice, which when exposed, sublimated within a few days[4].
With chemical and particle measurements obtained from the Mars landers and rovers, NASA was able to develop a technique for simulating Mars regolith from volcanic material found in Hawaii. That simulant was called JSC Mars-1, which was believed to accurately mimic the properties of Mars regolith. The company that NASA worked with (Orbitec) later made a second batch of simulant from the same source, and called it JSC Mars-1A. JSC Mars-1A has similar chemical composition to Mars regolith, based on Martian rover testing, but has a different density, which is believed to be a result of the presence of organic material and water in the terrestrial simulant. To make it more accurate, the ODU JSC Mars-1A simulant was baked at high temperature to remove water and burn off organic compounds[3].
Since the ambient pressure on the Martian surface is below the triple point for water, any mined icy regolith must be processed in a pressurized environment, so that heating the regolith will cause the ice to melt, boil and/or evaporate, enablingits collection as water distillateon colder surfaces (above freezing) rather than attempting to collect sublimating water vapor. By crushing the icy regolith into smaller sized particles than one liter blocks, the surface area of the regolith feedstock per unit mass increases dramatically, which reduces the time needed to extract a specified mass of water.
The purpose of this project is to design a rock (Martian permafrost) crushingmachinethat will operate in Martian temperatures and can crush the required amount of regolith needed to produce approximately 100 kg of water each sol. A prototype of the crushing process will be 3D printed, built and tested with similar required production rates to prove its functionality and feasibility.
Methods
Input Sizing Justification
The geometry of the regolith/permafrost input material is one of the initial criteria in developing a rock crushing system. The ODU spring 2015 RASC-AL team had determined a practical regolith excavation method that involves excavation by means of a regolith/ice saw. The ice saw would be employed to cut 10 cm deep grooves in the frozen regolith Martian surface, resulting in a grid pattern consisting of 10cm by 10cm squares, 10 cm deep. This procedure would thus leave one-liter-size cantilever beams on the surface, ready to be broken off and extracted [5, 6]. On that basis, an excavation area requirement was estimated to be 2m by 2m per sol. This geometry of liter sized cubes was assumed as a basic block size for designing a crushing process for enabling the efficient extraction of water from Martian regolith.
Output Sizing Justification
In order to determine an optimal regolith grain size, produced by a one or two-stage crushing machine, the efficiency with which water can be extracted from an individual regolith grain over a range of characteristic sizes was examined. Furthermore, a conservative regolith-water contenthad to be specified in order to determine roughly howmuch excavated regolith would need to be collected and processed in order to produce a sufficient quantity of water each sol. Considering estimates done in the Northern Martian hemisphere, the subsurface water content (by mass) ranged between 44% and 11% [7]. From the Spring 2015 RASC-AL group’s report, it was determined that by extracting 15% water by mass, from the excavated regolith, less than 0.5 m3/sol of regolithwould be needed to yield the desired 100 kg/sol of water[6]. Thus the required “permafrost” volume corresponds to 500 harvested one-liter regolith blocks per sol. With a designwater content and an established excavation rate, a thermal model was developed to determine a range of acceptable output sizes and the heating requirements for water extraction could be estimated subsequently.
The ODU spring 2015 RASC-AL team provided a recommendation for a Martian colony power source in their final report[6]. The recommended power source was a Japanese prototype reactor, designed initially for lunar base operations, known as the RAPID-L [8]. The regolith/permafrost crushing system and water extraction process are expected to operate within the constraints of the RAPID-L’s 200 kW electrical output, and its associated heat rejection rate of 5 MW, at a temperature of 800 K. Furthermore, the water extraction process was assumed to utilize heated gaseous carbon dioxide, supplied from the RAPID-L at a temperature of 600 K, enabling the system to take advantage of the reactor’s rejected waste heat. The estimatedavailable output wasmore than 4500 kW at a heat rejection temperature of 800 K.
The heating process assumed in order to determine theoptimal grain size produced by the rock crusher has been explored. After several iterations, thenominal water extraction heating approach represented in Figure 1 was the basis for the grain-sizingheat transfer analysis.
Figure 1: Water Extraction Schematic
Initially, the system wasonly assumed to be pressurized to just below 10 kPa, in order to raise the phase changetemperature of water above the triple point. However, the density of the carbon dioxide heating fluid was low, and the associated heat transfer estimates produced excessive feedstock residence times. By increasingthe density of the hot carbon dioxide, employing apressure of 100 kPa, heat transfer rates could be improved and the resulting overall system dimensions appeared to be reasonable. The regolith grains are processed through the crushing system, at a specified temperature and pressure where any exposed water on the grain’s surface cannot rapidly vaporize or sublimate. The subfreezing grain output would then becollected in a grain hopper, as shown in the Figure 1. At the bottom of the grain hopper is a screw-type auger of sufficient length to prevent the temperatures within the hopper from warming to the melting point of ice. The main purpose of the auger is to maintain carbon dioxide pressures near 100 kPa, during the time when the lid to the rock crusher (supplying the liter-size permafrost block feedstock) is open to receive an additional batch of blocks prior to crushing. The interior pressure of the system, must simultaneously control the rate at which regolith grains entering the system are heated to extract water. The grains are droppeddown the hot CO2 pipe where forced convection controlled both by the terminal velocity of the grains (in Mars gravity) and the upward-directed velocity of hot CO2, supplied by a blower drives the phase change of water. Finally, the mixture of steam and humid carbon dioxide is routed to a heat exchanger where it is condensed and the water is stored for use. Some of the main components of the system that are either to be determined or required include the length of pipe along which the grains fall, the optimal speed of the blower, the fluid properties of hot CO2, and both the inlet and outlet temperatures of the regolith grains. The method that has been selected to estimate these parameterswas a lumped system heat transfer approach. The main assumption of this approach entails that once the center of the regolith grain has reached a specified temperature above the boiling point, in the set environment, the water content in the grain will be vaporized and separated as the latent heat of sublimation (combining the latent heat of fusion with the latent heat of vaporization) is supplied.
The lumped System Calculation
A series of trial calculation were performed, varying theregolith permafrost grain diameter from 1 millimeter to 5 millimeters. A heat transfer lumped system calculation is capable of determining the time in which a spherical particle will uniformly change from an initial to a final temperature, provided the system meets certain criteria, such as a minimal Biot number. In this case, the initial and final temperature of the regolith grain can be specified. A conservative, below average Martian surface temperature of 200 K was assumed for the liter-size blocks and a temperature 375 K, above the boiling point of water, were assumed as the initial and final temperatures of a processed grain. Then, the inlet temperature of the hot CO2 was specified as 600 K, as a result of a heat exchanger from the RAPID-L rejected heat output, to the source of the system’s carbon dioxide.
In order to estimate a forced convection heat transfer coefficient, the Nusselt number was needed. Therefore, it was necessary to estimate the fluid properties of the hot CO2.The dynamic viscosity of the hot carbon dioxide and the grainsurface were estimated using Sutherland’s formula for pure carbon dioxide:
The remaining fluid properties of CO2, including thermal conductivity, specific heat and density were determined using the National Institute of Standards and Technology (NIST) - thermophysical properties of fluids calculator. Furthermore, the Prandtl was determined using the dynamic viscosity, thermal conductivity and specific heat of CO2.
It was necessary to estimate the characteristic Reynolds number of a spherical particle (assumed smooth) falling through the hot CO2 pipe, and then select an appropriate correlation to estimate the Nusselt number. In order to ignore the influence of auger speed and the initial acceleration (or deceleration) of the injected grain particles, the model grain particle was assumed to initially fall at its terminal velocity, corresponding to the gravityon Mars.Therefore, the coefficient of drag on each grain, of possibly varying diameter,wasrequired. Since the Reynolds number depended on the terminalvelocityofthe grain, the terminal velocity depended on the coefficient of drag, and the coefficient of drag depended on the Reynolds number, a straight-forward terminal velocity calculation could be used.
However, an alternative method was employed using a modification of the terminal velocity equation, as
well as a coefficient of drag vs Reynolds number plot for a smooth sphere.
Five points were selected on a drag coefficient vs Reynolds number curve where the Reynolds number ranged from 1 to 1000, using Figure 2 above. The values of CdRe2 were calculated and their logarithm of base ten was calculated, employing a logarithm base ten scale. From that curve fit, two individual plots of drag coefficient and Reynolds number versus log10(CdRe2) were created. The most similar curve-fit function was employed using the Microsoft Excel plotting tool and is shown in Figure 3a and Figure 3b. An equation for Reynolds number as a function of log10(CdRe2) and an equation for Drag coefficient as a function of log10(CdRe2) are shown in the lower right portion of each plot. These equations were employed to provide Reynolds number, drag and Nusselt number estimatesfor grain diameter-based Reynolds numbers below 1000. However, if the Reynolds number of the grain was found to be greater than 1000, meaning the log10(CdRe2) > 5.6, then the drag coefficient was set equal to a constant at 0.4 (by simplification of the Cd vs Re plot in Figure 2). This method was used assuming that the Reynolds of the desired grains did not exceed 5x105.
Figure 3a: Cd Vs Log10 (CdRe^2)
Figure 3b: Re Vs Log10 (CdRe^2)
The curve-fit functions allowed for calculations of the Reynolds number and the coefficient of drag, then lead to the terminal velocity for each grain trial between 1mm and 5mm. The Nusselt number was then determined and, in turn, the Heat transfer coefficient assuming forced convection for each grain diameter trial was estimated.
Finally the lumped system relationship could be utilized to find the time required for each grain to be individually and uniformly heated to the final temperature of 375 K.
Once the heating time required for each grain diameter was determined, the length of the heated CO2 oven pipe could be estimated. Multiplying the terminal velocity by the required heating time, corresponding to the correct grain size, a pipe length estimation was found. However, for grain diameters of 3mm and above the length of pipe ranged between 8 and 25 meters, producing unreasonable pipe length estimates. Incorporating a blower to provide a vertical hot carbon dioxide flow through the pipe allowed for the larger grain sizes to remain a viable option to for water extraction. The blower extended the time available to heat the falling grain, since the overall velocity over the grain remained equal to the terminal velocity; however, the overall pipe length became significantly shorter and more realistic, as depicted in Figure 5. Furthermore, an estimate for pipe diameter as a function of the regolith grain diameter was determined and that variation is shown in Figure 6. The method used to estimate each pipe diameter is provided in Appendix 2.
Figure 4: pipe length vs grain diameter
Figure 5: pipe diameter vs grain diameter
Preliminary Regolith Crushing Method Research
In researching the crusher design schemes and standards for rock crusher design, several factors had to be considered. The most basic factor is the reduction ratio, which is the comparison of the longest dimension of the starting rock compared with the product of the crusher. Jaw crushers and cone crushers had ratios of 6 to 1and 8 to 1, respectively, while some other crushers’ utilized 2.5 to 1 ratios (a double roller crusher), or a 20 to1 ratio (the hammer mill model or the double impeller impact breaker crusher). The greater the reduction ratio, the smaller the output could be, and the more reasonable the ideal size for the thermal modeling could be [9].
A Concern that came into Play later in the Project was how large or heavy the crusher would be. Most crusher types are not designed to be portable or small, with weight in the tons as well as large sizes. Limiting crusher selection to lab crushers, it is possible to find a few existing models that can reduce the cubes down inside the 1-5 millimeter range, however, modifications will need to be made to optimize the process. The only type of crusher that is commonly found in lab size is the Jaw Crusher type, due to their simple design. Bico’s WD Chipmunk Jaw Crusher (3 hp, 3 phase or 3hp 1 phase) is able to crush regolith down to 1.6 mm, but its opening is only 7.5cm by 10.1cm, so either the cubes would need to be cut in 7.5 by 10 by 10cm blocks, or the crusher would have to be redesigned to get the jaw plate opening wider [10].