Reusable Launch Vehicle: Technology Development and Test Program (1995)

A key component in the success of the reusable launch vehicle program is the development of a reusable cryogenic tank system (RCTS) that can withstand the environments of launch and reentry and can meet the weight and reusability goals of the RLV. In the past, fuel tanks, such as the space shuttle external tank, have been jettisoned before the vehicle entered orbit. The semi-conformal and integral reusable tanks proposed for the RLV, however, not only store propellants but are also part of the primary structure of the vehicle. Because the tanks are used to store propellants, which comprise most of the vehicle volume, they contribute significantly to the mass of the primary vehicle structure. To date no reusable cryogenic tanks of this scale have been used in flight, but two subscale models have been built and tested at LH₂ temperature under the auspices of the National Aerospace Plane (NASP) technology program. The NASA/industry program is building on the experience of the NASP program.¹

Technology development during Phase I is designed to demonstrate the relative merits of both composite and metallic materials for the RCTS in the X-33 and in potential RLV configurations. The RCTS program includes development of metallic Al-Li alloys (primarily for LOX) and composite tanks (primarily for LH₂). At this point, however, data are incomplete for evaluating the material properties, life cycle, manufacture, inspectability, and repairability of some tank materials being considered for reusable cryogenic tanks. Therefore, the objective of RCTS technology development is to determine whether these tanks can be functionally produced and whether weight, reuse, cost, and operations requirements for X-33 and RLV configurations can be met.

During Phase I a series of developmental tests are planned to provide data to determine whether reusable cryogenic tanks can be integrated into an X-33 flight test vehicle to support a demonstration of the SSTO by the end of the decade. The following

tasks must be accomplished for the metallic and/or composite tanks to meet the Phase II go-ahead criteria:

1. At least one metallic (Al-Li) tank will be constructed and integrated with the required TPS, health monitoring, and attachment subsystems and will be under test. Current plans call for two such tanks to be manufactured and integrated for test. Appropriate coupon and other element testing (e.g. LOX compatibility, reusability) required to achieve this goal will be completed and documented. All applicable subscale testing will have been conducted to scaled (to full-scale RLV) pressures and loads.

2. At least one graphite composite tank will be constructed and integrated with the required TPS, health monitoring, and attachment subsystems and will be under test. Current plans call for two such tanks to be manufactured and integrated for test. Appropriate coupon and other subscale testing (e.g., LOX compatibility) to achieve this goal will be completed and documented.

3. The material selection for both fuel and oxidizer tank subsystems will be completed and documented. The selection must consider performance (e.g., weight, strength) producibility, inspectability, and operability characteristics.

4. A documented analysis will have been completed demonstrating that the selected materials and tank subsystems are scaleable to a full-scale RLV and will adequately be demonstrated by an X-33 vehicle. This analysis will contain the correlations between analytical predictions and experimental test results. These correlations will be at a level of confidence sufficient to ensure that analytical tools are valid for purposes of full-scale vehicle design. Estimated requirements for the RLV, which will be supported by this analysis are a minimum of 100 lifetime missions including depot maintenance not more than every 20 missions, volumetric weight targets (which will be updated for selected X-33 configuration) of 0.7 lb/ft³ or less for an oxidizer tank and 0.5 lb/ft³ or less for a liquid hydrogen tank, with leakage rates within the limits set for the space shuttle.

The NASA/industry programs and the findings and recommendations for Al-Li alloys and composites are discussed separately in the following sections.

The objective of the NASA/industry cryogenic tank technology program is to test structurally a "flight like" Al-Li cryogenic tank to validate the analysis and manufacturing methods used in its design. Advanced technologies, such as near-net–shape extrusion, near-net forging, and spin forming of Al-Li, are expected to be demonstrated at the NASA technology readiness level (TRL) of 6. (On the TRL scale, TRL 6 refers to demonstration of a system/subsystem model or prototype in a relevant ground or space environment.)

The cryogenic tank technology program tank will be 14-ft in diameter and will include near-net–shape extruded, integrally blade-stiffened panels; net-shape, spin-formed bulkheads; and near-net–shape, roll-forged stub adapters. The tank, which is being built now, is an all Al-Li alloy 2195 tank. External cryogenic insulation will be installed later. On one tank, the TPS panels will be installed, and a life cycle test will be conducted on the ground.

One industry partner has contracted with the Russians to build a LOX tank system from the Russian Al-Li alloy 1460. Russia is also building an Al-Li LOX tank for the DC-XA from alloy 1460, with some components fabricated from the Russian equivalent of alloy 2219. This tank is designed to SSTO loads and environment; however, the Al-Li alloy 1460 currently being used for this tank is not optimized. The industry partners will also be focusing on developing a LOX tank constructed from Reynolds alloy 2195 and building, friction-stir–welded tank 3 ft in diameter.

The Super Lightweight Tank (SLWT) for the space shuttle at the NASA Michoud Assembly Facility in New Orleans, Louisiana, is being constructed using alloy 2195 thick plate stock with Al alloy 2219 ring frames. There are a number of key SLWT milestones that are important to the technology for the X-33 LOX program. Fabrication of the advanced launch test article began in February 1995, and the advanced launch test article proof test was scheduled for October 1995. The welding techniques and procedures developed in this program apply directly to the X-33 program, and the results of tests on the welded article should be particularly informative. The advanced launch test article tests are scheduled to be completed by May 1996 and will be available for use in the X-33 program. The size of the LOX and LH₂ tanks for the RLV will be similar to those of the space shuttle SLWT; therefore, demonstrating producibility of a large tank like the SLWT will be reassuring.

In addition to alloys 2195 and 1460, an isotropic Al-Li alloy is being developed under Air Force Contract F33615-92-C-5914 with the University of Dayton and its subcontractor ALCOA. The alloy, designated AF(UDRI), contains more than 2-percent–weight lithium, which is the minimum for achieving approximately 10 percent savings in weight (compared with conventional aluminum alloys of equivalent mechanical properties) if only density and modulus are considered. The mechanical properties of alloy AF(UDRI) are far more isotropic than the mechanical properties ever observed for a similar density Al-Li alloy and are extremely encouraging. For instance, in-plane anisotropy has been reduced to less than 4 percent for 2 percent cold-worked recrystallized plate and to less than 8 percent for 6 percent cold-work unrecrystallized

plate; fracture toughness was nearly doubled in the short, transverse direction and was improved significantly in- plane; and fatigue-crack growth improved significantly compared with 7050-T7451, a leading non-lithium bearing plate alloy. A comparison of alloy AF(UDRI) with alloy 2195-T8, based on similar thickness and degree of recrystallization, is shown below:

Because of the higher lithium content, alloy AF(UDRI) has a 3 percent lower density and a 6 percent higher modulus than alloy 2195. Although AF(UDRI) is not as mature as alloys 2195 or 1460, it may offer some advantages, and its development should be monitored for possible future use.

MSFC will conduct mechanical tests on the Russian Al-Li alloy 1460 and Reynolds alloy 2195 and will specifically evaluate reusability. Reusability testing is defined as testing time-dependent properties, such as fatigue-crack–growth rate and susceptibility to stress-corrosion cracking. Alloy 1460 will be tested for baseline tensile, LOX compatibility, fracture, fatigue, weld development, and stress-corrosion. Tests will be conducted at both cryogenic and high temperatures.

A critical step in the manufacture of the LOX cryogenic tank is welding the chosen Al-Li alloy. Although alloy 2195 can be welded, weld repair and/or second-pass welding is a major problem. MSFC experiments with an Al-Si alloy filler have resulted in some improvement in weld repair. In cooperation with MSFC, the Edison Welding Institute, Boeing, and Reynolds Aluminum, one industry partner will be examining a new method of "friction-stir welding." The friction-stir–welding process shows promise and could, if successful, solve the welding problems associated with Al-Li alloys. Data for alloy 2195 on the space shuttle SLWT will also be incorporated into the database.

Al-Li alloys are primarily used for the LOX cryotanks. One industry partner is considering the use of a titanium (Ti) honeycomb tank as backup for the composite LH₂ tank. The Ti alloy (Ti-6Al-4V ELI) is being considered for use in the face sheet.

In general, the criteria for Phase II goals appear to be well conceived and reasonable, with the exception of the requirement of 0.7 lb/ft ³ or less for the oxidizer tank, which may not be appropriate as a universal, absolute target. Appropriate weight targets for all major components should be based on the system design-engineering

process. Target weights depend on specific design concepts and parametric trade studies to optimize the overall design rather than on individual components of the design.

The development program for the Al-Li tank primarily for LOX application is robust because more than one approach is being developed in almost all of the critical categories. Among NASA and the industry partners, two alloys of Al-Li with somewhat different properties are being used (Reynolds alloy 2195 and the Russian alloy 1460). Three fabrication techniques (machining to net-shape extrusion, net-shape spin forming, and net-roll forging) and two welding techniques (standard and friction-stir welding, which is under development) will be used. Finally, with respect to the scale of tanks to be fabricated and tested, there are one 14-ft-diameter, one 3-ft-diameter, and two 8-ft-diameter tanks being developed in the RLV program; a 28.5-ft-diameter by 154-ft-long tank made of Al-Li alloy 2195 with Al alloy 2219 rings is being developed as a replacement for the space shuttle's external tank. The Al-Li tank for the SLWT will clarify the issue of scaling the smaller test tank data to the full-scale RLV tank, which is approximately 40 ft in diameter. As is required by the decision criteria for Phase II approval, more than one organization is conducting tests to characterize material properties, including reusability (e.g., fatigue-crack–growth rate) and LOX compatibility for both the 2195 and 1460 alloys.

Despite the robustness of the Phase I development program, there are several technical areas of concern affecting producibility, operability, and reusability that are detailed in the following sections.

Using Al-Li alloys for the LOX cryogenic tank is an excellent choice. Current tests are focused on Reynolds alloy 2195, which has been selected for the SLWT, and the Russian alloy 1460.

Alloy 1460 is similar, although slightly lower in alloy content, to ALCOA alloy 2090, except a small amount of scandium (Sc) has been added. The Russians claim Sc improves weldability of the alloy by refining the grain size in the pool of molten metal formed during welding.² Alloy 1460 was developed specifically for use with LH₂ and LOX. With properties similar to alloy 2090, alloy 1460 has a higher modulus and a lower density than alloy 2195. Alloy 1460 may not be as strong as alloy 2195; however, their specific properties are similar.

Except for data in the proceedings of recent conferences on Al-Li and aluminum alloys, little has been published about alloy 1460.^2,3 Recent preliminary data provided to one industry partner indicate that the properties of alloy 1460 are at least 25 percent above specifications. Because alloy 1460 is relatively new (the effect of Sc on the properties of alloy 2090 is not well documented), there will have to be extensive tests to characterize all product forms (including welded products) to establish a reliable database. Tests should include wide panel testing, not just coupon testing, of the weld zone because cutting out coupons may relieve the residual stresses associated with the welding process that can affect fracture toughness and fatigue properties. Fracture

toughness and fatigue-crack–growth studies should be conducted over the total temperature range, including cryogenic temperatures.

Alloy 2195 exhibits better cryogenic ductility and significantly greater strength than the conventional alloy for cryogenic tanks, alloy 2219. Alloy 2195 also exhibits a positive fracture toughness ratio when subjected to a range of temperature (from room temperature to cryogenic temperatures), which is an important consideration for cryotanks. Greater strength, coupled with higher modulus and lower density, can lead to significant weight savings. The alloy also has good corrosion resistance, excellent fatigue properties,⁴ can be near-net–shape formed, and, with proper precautions, can be adequately welded.⁵ However, there has been some concern about consistent producibility. Specifically, some thick plate has failed to meet minimum properties at T/8, where T is the thickness of the plate. Consequently, there may be problems associated with through-thickness anisotropy. Recent microstructural and texture analyses have shown that an excellent product from the microstructural point of view can be produced with current methods.⁶ However, each casting and product form should be extensively characterized to ensure that it meets the required microstructure/properties criteria.

Weldability and weld repair are major areas of investigation. MSFC has been experimenting with an Al-Si alloy filler, which has resulted in some improvements in weld repair. But the committee believes MSFC should exercise caution in using an Al-Si filler material because silicon combines with Al-Li to form an AlLiSi phase that attracts and absorbs moisture, which increases susceptibility to stress-corrosion cracking. Alloy 8090 was found to be very susceptible to stress-corrosion when the alloy contained an excess of only 0.08 percent weight silicon.⁷ NASA should include stress-corrosion tests of weld zones of alloy 2190 to ensure that the filler material does not produce a weld zone susceptible to stress-corrosion cracking. This test is not included in the current program. If the weld repair problem is not resolved, an alternative would be to cut out the defective area and replace it with virgin metal because no problems have been identified if the first weld is sound.

The Russians claim that alloy 1460 is weldable and have reported weld zone strengths of more than 40 ksi. Russian alloy 1217, which is used for weld wire, contains an Sc addition and appears to be superior to conventional weld wire for Al-Li applications. Boeing has obtained weld zone strengths approaching 40 ksi for alloy 2090. Weld zone strengths should be verified and stress-corrosion tests should be conducted on the weld zone of alloy 1460.

Friction-stir welding seems to be a promising solution to the problem of welding Al-Li alloys. However, the process is in an early stage of development, and extensive tests are needed to determine the feasibility of using this technology on the thin sheet (less than 1/8-inch-thick) that will be used for the LOX tank. Results obtained from

welding 1/4-inch-thick alloy 2195 for the proposed demonstrator LOX tank may not be scaleable to the proposed X-33 tank.

Near-net-shape forming is another technology that is critical for the LOX cryogenic tank and will require extensive microstructural and property characterization of the finished products. Properties obtained on the machined isogrid panels for the SLWT may be different from the properties of the near-net-shaped formed panels that are planned for the X-33.

Russia is building an Al-Li LOX tank from alloy 1460 for the DC-XA using some components fabricated from Al alloy 2219. The Russians have extensive experience in building welded aircraft structures from Al-Li alloys 1420 and 1421, which should be helpful in constructing the LOX tank. The tank is designed to SSTO loads, and the environment and fabrication methods should be scaleable to the requirements for SSTO. Because of the limited database on alloy 1460 as well as possible lot-to-lot variations in Al-Li alloys, extensive coupon testing for durability—fatigue, fracture, stress-corrosion cracking (time-dependent properties), thermal stability, and similar tests will be needed as well as full-scale tests of the tank. The fatigue-crack growth and stress-corrosion behavior of welded and weld repaired alloy 1460 should also be determined. Al-Li alloys have been shown to be more anisotropic than conventional aluminum alloys; therefore, the texture of these alloys should be characterized and properties in the L, T, and 45° orientations should be measured. If thin structure is to be produced by chemical milling of a product thicker than 10 mm, properties of the milled product should be determined. In addition, all properties should be determined for the entire range of temperatures relevant to the X-33. Some of these tests are planned as part of the McDonnell Douglas Aerospace/Langley Research Center (LaRC) plan, which is funded separately. If the funding is withdrawn, other funding will be necessary to ensure that the database for the X-33 program is adequate.

The Al-Li LOX tanks constructed during Phase I must be scaleable to the tank size required for Phase II. Problems with the through-thickness properties of thick plate, the weldability of thin sheet, and other properties of the candidate Al-Li alloys must be identified and resolved prior to tank construction in Phase II. Tests of the time-dependent properties of alloy 2195, similar to the tests suggested for alloy 1460, should also be conducted to establish the durability of this alloy system. If near-net–shape manufacturing methods will be used for the X-33, the properties of these product forms must be established. By the end of Phase I, structural models should be in place to validate the concepts being considered for construction of the Phase II tank.

Composite materials are the prime candidates for construction of the LH₂ tank for the X-33. However, a metallic Al-Li option should be considered as a backup. A titanium (Ti) honeycomb tank is being considered as a backup for the composite LH₂ tank, and the Ti alloy (Ti-6Al-4V ELI) is being considered for the face sheet. Alpha and alpha/beta titanium (e.g., Ti-6Al-4V) are susceptible to embrittlement in the presence of very small levels (less than 250 ppm) of gaseous hydrogen.⁸ Nelson and Williams at NASA-Ames have published a series of papers describing the susceptibility of alpha/beta alloys to gaseous hydrogen alloys; including Ti-6Al-4V.^9,10,11 Nelson and Williams have shown that the apparent susceptibility of these alloys depends, in a complex way, on microstructure, test temperature, crack tip strain rate, and hydrogen pressure. However, Nelson has shown that all microstructures of Ti-6Al-4V are susceptible to hydrogen embrittlement.¹² Alloy Ti-6Al-4V ELI has a very low interstitial content, and it is known that interstitials increase the susceptibility of alpha and alpha/beta titanium to hydrogen embrittlement. Therefore, in spite of its very low interstitial content, Ti-6Al-4V ELI requires extensive characterization studies (similar to the studies conducted by Nelson on conventional alloy Ti-6Al-4V) before it can be used in a reusable LH₂ tank.

The following recommendations are intended to ensure that special attention is paid to several areas of concern. The committee realizes that NASA and the industry partners are aware of most of these concerns and are attempting to resolve them.

• Recent microstructural and texture analyses of alloy 2195 have shown that the current processing methods produce an excellent product. However, each casting and product form should be characterized extensively to ensure that the required microstructure/properties are obtained. Other product forms requiring characterization include all welded and weld repaired products, as well as extruded near-net-shape formed products.

• Because of the limited database for alloy 1460 and possible lot-to-lot variations in Al-Li alloys, extensive coupon testing of all alloy 1460 product forms should be conducted.

• Because Al-Li alloys have been shown to be more anisotropic than conventional aluminum alloys, their texture, strengths, and elastic moduli should be characterized, and properties in the L, T, and 45° orientations should be measured. If a thin structure is produced by chemical milling of a product thicker than 10 mm, properties of the chem-milled product should be determined for the entire range of temperatures relevant to the X-33.

• Weldability and weld repair strength are major issues. Although alloy 2195 can be welded, weld repair or second-pass welding is a major

problem. MSFC has been experimenting with an Al-Si filler material. The committee recommends caution when using an Al-Si filler because it forms an AlLiSi phase that attracts and absorbs moisture, which increases the potential for stress-corrosion cracking. Alloy 2195 weld zones should also be tested for stress-corrosion.

• Friction-stir welding should be vigorously pursued and demonstrated using the thin sheet Al-Li (less than 1/8-inch) that will be used for the LOX tank.

• The scaleability criterion must be rigorously satisfied because of issues related to the through-thickness properties of thick plate, the weldability of thin sheet, and other factors. Validated structural models must be in place for designing the tanks in succeeding phases.

• In addition to the 2195 and 1460 alloys, the progress of an isotropic Al-Li alloy (alloy AF[UDRI]) being developed by the University of Dayton and its subcontractor, ALCOA, should be followed. This alloy shows very good material properties when compared with other conventional aluminum alloys, and it may be a candidate for future use.

• Al-Li should be considered as a backup composite material for the LH₂ tank if system studies show that the performance penalty acceptable. Caution is recommended regarding use of Ti honeycomb, which is being studied by one industry partner, because the Ti alloy Ti-6Al-4V is susceptible to hydrogen embrittlement and not enough hydrogen embrittlement studies have been conducted on the ELI variant of this alloy.

• Thermal/load cycle testing on the alloy 2195 tank is strongly recommended to ensure that the integrated system will satisfy reusability requirements and to provide a database comparable to the database designed for the alloy 1460 tank-flight and ground test articles in the DC-XA program.

• Weight predictions (not only the 0.7 lb/ft³ for an oxidizer tank or 0.5 lb/ft³ for a hydrogen tank given in the decision criteria) must be defined for each vehicle concept; achievement of these (scaled) weights must be verified with properly designed test articles.

Several issues are critical to the use of organic-matrix composites for cryotanks: weight, producibility, permeability, and cycle life. All the contractors and NASA centers are aware of these concerns and are working to address them. The activities range from testing small-scale coupons to testing tanks of substantial, but subscale, dimensions. Three approaches to fabrication of a cryotank are being investigated. Any or all of them may produce a satisfactory product; however, given the somewhat speculative state of

this technology, the probability of success is greatly enhanced by the pursuit of diverse construction concepts.

One approach to the fabrication of the organic-matrix composite tank is to use carbon cloth layups impregnated with epoxy formed to shape. The layups are then cured in an autoclave. Historically, the layup approach has been used, for example for fabricating the heat shield on reentry vehicles. The second approach uses filament-placement machines to apply graphite filaments coated with epoxy to a mandrel in the desired shape, followed by curing in an autoclave. In the third approach, a honeycomb or foam core is sandwiched between sheets of graphite epoxy, which are used to fabricate the tank. At least two sizeable tanks will be fabricated using each method, one 8 ft in diameter and 16 ft long, the other, 8 ft in diameter and 9 ft long. Fabrication of a third, slightly larger, tank is in the initial planning stage. Several contractors will test material properties, subscale tank/bottles, and panels to determine the basic strength (therefore weight) and reusability of the tanks.

Permeability testing has a long history. Contractors in the NASP program tested numerous 2-inch samples that were 0.030 inches thick (much thinner than the proposed RLV or X-33 tanks) in LH₂ at pressures on the order of 50 psi. When toughened epoxy was used in these tests, there was no evidence of significant microcracking or permeability. The samples were cycled extensively under load prior to being cut into coupons for the permeability tests. Also a 10-inch diameter composite tank was subjected to hundreds of cycles in LH₂ at pressures up to 300 psi without leaking.

Two of the NASP contractors built and tested subscale tanks. One of the tanks was tested at 7 psi. This tank was subjected to 10 full LH₂ temperature cycles. During each cycle, the tank was pressure-cycled 10 times for a total of 100 pressure cycles. The tank was then filled with liquid nitrogen, taken to full pressure, and loaded simultaneously with the maximum predicted positive and negative structural loads. Four hydrogen detectors were placed in various locations on the tank for each test, and air samples were taken at various locations near the tank skin during the low-temperature tests. No significant amount of hydrogen was detected. The tank was thoroughly tested for leaks at ambient temperature using helium both before and after testing. The only leakage was at the flange joints. It should be noted that this tank was mounted in a titanium, metal-matrix fuselage section to simulate a flight-like environment. A second tank leaked severely on the first test but not through the composite. A bonded invar ring joining the bulkhead to the cylinder leaked at the invar ring joint. This tank was later tested successfully.¹³

Indications to date are that if strain caused by combined thermally induced and mechanical loads is kept below certain levels, microcracking is minimized and permeability is not a serious problem. Refining this limit will satisfy a major design criterion.

Various tests are underway on cryostats in support of the X-33 and RLV. One industry partner has developed a composite hydrogen line and ball valve for the DC-XA auxiliary power unit. A 50+ cycle temperature and pressure test in LH₂ was recently completed on this valve at MSFC. The valve is also being tested in LH₂ to account for the characteristics of composites, and the results to date are encouraging. MSFC is also

producing a number of 32-inch-diameter, half-scale tanks being used to evaluate structural and integration-related issues, such as stiffener attachment, Y-joint design, and others.

In addition to planned tests on subscale tanks, all of the contractors will constantly monitor the tanks for leaks of all types, including permeability, particularly as the number of pressure and temperature cycles increases and microcracking may become more significant. One method of dealing with permeability problems is to line the tank. This would create another set of problems, however, and probably increase production costs. Thus, this option is to be avoided, if possible. All evidence to date indicates that when proper design and material are used permeability is not a major concern. Because it will not be easy to develop leak-free tanks, one of the industry partners is investigating using an aluminized mylar liner. This liner has been tested for a limited number of cycles, but the test results are promising. The developer is less concerned with the graphite-epoxy membrane than with joints and penetrations (based on the NASP experience).

In an attempt to resolve the permeability question, an industry partner plans to test a large number of 6-inch-diameter, organic-matrix composite tanks to evaluate various materials and construction techniques, as well as liners. These small tanks are relatively inexpensive and should be useful for evaluating the factors discussed above through a number of cycles. Unfortunately, because of their small size, the scaling issues are significant. Four 3-ft-diameter tanks, one of them of sandwich construction, will also be built. These tanks will be subjected to several thermal and stress cycles (e.g., five). One tank will be used only for evaluating vehicle health-management instrumentation. Funding recently has been added to this program to allow testing one of the tanks for 100 cycles. The tank construction to be chosen for the 100 cycle tests and whether the tank will have a liner depend upon results of the 6-inch-tank tests. Performing 100-cycle tests on the 3-ft tank may verify life expectancy.

Integration of TPS and insulation with the cryogenic tank is an area being investigated by all of the contractors. This problem is being addressed by 4-ft by 5-ft panel tests and tests on scaled tanks. One industry partner has recently recognized that larger scale testing is desirable and has approved production of a 10-ft-diameter tank. The current plan is to build a two-lobed tank with a mechanical interlobe joint.

In general, the criteria for Phase II goals appear to be well conceived and reasonable, with one exception. It is not clear that the requirement of 0.5 lb/ft³ or less for the hydrogen tank is appropriate as a universal, absolute target. Appropriate weight targets for all major components should come from the design-engineering process.

Target weights depend on the specific design concepts and on the parametric trade studies to optimize the overall design rather than on individual components of the design.

Development of the organic-matrix composite propellant tank appears to be well focused toward achieving the ultimate goals of the project. The development program in this area is robust because the participants are pursuing a multiplicity of technical approaches. There are, however, several areas of concern, which are discussed in the following sections.

Cycle life of the tanks is a far more significant problem than it was with expendable launch vehicles. Previous tanks were certified for as few as five temperature and pressure cycles; however, a reusability target of 100 missions could easily require certifying the tanks to as many as 200 or more cycles. There is no precedent for this level of reusability, particularly for composite tanks. In addition to the temperature and pressure cycling, requirements for reusability and multiple orbital flights mean the tanks must withstand particle impact damage and be repairable. The results of the 10-inch-tank tests are most encouraging, but more data are needed for larger tanks and correct load levels. The large-tank test programs now being planned will not test nearly enough cycles to give real comfort. Also, it is difficult to accumulate the required number of cycles when testing large tanks because of the length of time required for the chill down, test, and warm-up cycles. Small sample tests will continue and will certainly add some degree of comfort, but they cannot fully address the problem.

Fabrication of the tanks is also a major area of concern. Although the three proposed techniques are conventional and well understood, application of the techniques to tanks of the required size and weight per enclosed volume is not. The tanks needed for the RLV will range from 25 ft to 40 ft in diameter and will vary in length. As noted previously, autoclaves of the correct dimensions may not be available. Thus, the tanks may require out-of-autoclave assembly of segments that have been autoclaved; or the tanks may have to be cured by other methods. Tooling for the fabrication process may also be expensive. The hand layup process eliminates some of the tooling and equipment costs, but this process may not be practical for a tank of the required size. A broad goods-dispensing technique has been used experimentally as an alternative to the hand layup method. This technique seems to retain the advantages of using cloth.

The committee believes the sizes selected for the test tanks are reasonable; nonetheless, the committee is concerned about producing full-scale tanks that maintain the required material properties some five times larger than the test tanks.

The issue of whether autoclaving is required is very important. At this time, it is not clear that there will be an autoclave large enough to accommodate the full-scale

tanks or primary structures (discussed in a separate chapter). The cost and schedule impact of building an autoclave of this size should be evaluated. Tests have been scheduled to resolve the issue of autoclaving versus non-autoclaving.

The critical issue of joining the tank to the intertank structure is receiving the same multiplicity of evaluations both in design and tests, which reduces the risk in this important area.

The primary focus of development of the composite tank is on the LH₂ tank; however, there is a possibility that composites will be used for the LOX tank as well. Compatibility is, of course, the key. Results of early testing at ambient pressures by one of the industry partners are encouraging. Tests at required pressures representative of the vehicle application should be expedited to determine if using composites for the LOX tank is a viable option. Sometimes there is an immediate emotional reaction about the risks of using composites with oxygen. This line of thought ignores the fact that, once ignited with oxygen, both aluminum and stainless steel, which are commonly used tank materials, burn furiously. It must be clearly demonstrated that the risk of ignition for organic-matrix composites is not significantly greater than for conventional materials. The possibility of using oxygen-compatible liners is also being explored.

The following recommendations are presented to ensure that special attention is paid to areas of concern. The committee recognizes that the program participants are aware of most of these concerns and are in the process of addressing them.

• A detailed plan addressing the producibility of full-scale composite tanks should be developed, and the advisability of some convincing fabrication demonstration should be evaluated.

• The autoclave issue should be resolved as soon as possible.

• Evaluation of the use of composite tanks for LOX should be continued to resolution.

• Thermal/load cycle testing should be conducted on all the 8-ft-diameter or larger tanks with cryoinsulation, interfaces with neighboring components and TPSs affixed to ensure that the integrated system satisfies reusability requirements. This testing will provide a database comparable to the database for the tank that will be tested and flown on DC-XA.

• Weight predictions (not only the 0.7 lb/ft³ for an oxidizer tank or 0.5 lb/ft³ for a hydrogen tank as given in the decision criteria) must be defined for each vehicle concept, and (scaled) achievement of these weights, with properly designed test articles, should be verified.

1. Private communications with the managers of the NASP program.

2. Fridlyander, I.N., A.M. Dritz, and T.V. Krymova. 1992. High-strength weldable 1460 alloy for cryogenic application. Pp. 1245-1250 in Aluminum-lithium: 6th International Conference on Aluminum-Lithium. Oberunsel, Germany: Deutsche Gesellschaft fuer Materialkunde Informationsgesellschaft mbH.

3. Fridlyander, J. 1994. Advanced Russian Aluminum Alloys. Pp. 80-87 in the 4th International Conference on Aluminum Alloys: Their Physical and Mechanical Properties, vol. 2. Atlanta, Georgia: The Georgia Institute of Technology.

4. Slavik, D.C., C.P. Blankenship, Jr., E.A. Starke, Jr., and R.P. Gangloff. 1993. Intrinsic fatigue-crack growth rates for Al-Li-Cu-Mg alloys in vacuum. Metallurgical Transactions A 24A(8):1807-1817.

5. Bhat, B.N., T.T. Bales and E.J. Vesely, Jr., eds. 1994. Aluminum-Lithium Alloys for Aerospace Applications Workshop. NASA Conference Publication 3287. Washington, D.C.: Government Printing Office.

6. Skrotzki, B., G.J. Shiflet and E.A. Starke, Jr. 1995. Characterization of Thick and Thin Plate of 2195: Final Report. Contract A110041. New Orleans, Louisiana: Martin Marietta Manned Space Systems.

7. Lewis, R. E. E.A. Starke, Jr., W.C. Coons, G.J. Shiflet, E. Willner, J.G. Bjeletich, C.H. Mills, R.M. Harrington, and D.N. Petrakis. 1987. Microstructure and properties of Al-Li-Cu-Mg-Zr (8090) heavy section forgings. Journal de Physique Colloque 48(C-3): 643-652.

8. Paton, N.E. and J.C. Williams. 1974. Pp. 409-431 in Hydrogen in Metals; Proceedings of the International Conference on the Effects of Hydrogen on Materials Properties and Selection and Structural Design, Champion, Pennsylvania, September 23-27, 1973. Metals Park, Ohio: American Society for Metals.

9. Nelson, H.G., D.P. Williams, and J.E. Stein. 1972. Environmental hydrogen embrittlement of an α-β titanium alloy: Effect of microstructure. Metallurgical Transactions 3(2): 469-475.

10. Williams, D.P. and H.G. Nelson. 1972. Gaseous hydrogen-induced cracking of Ti-5A1-2.5Sn. Metallurgical Transactions A 3(8): 2107-2113.

11. Nelson, H.G. 1973. Environmental hydrogen embrittlement of an α-β titanium alloy: Effect of hydrogen pressure. Metallurgical Transactions A 4(1): 364-367.

12. Nelson, H.G. 1974. Aqueous chloride stress-orrosion cracking of titanium—a comparison with environmental hydrogen embrittlement. National Aeronautics and Space Administration Report Number: NASA-TM-X-62314.

13. Private communications with the managers of the NASP program.