Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
NATIONAL RESEARCH COUNCIL COMMISSION ON ENGINEERING AND TECHNICAL SYSTEMS 2101 Constitution Avenue Washington, D. C. 20418 COMMITTEE ON NASA SCIENTIFIC AND TECHNOLOGICAL PROGRAM REVIEWS Panel on Redesign of Space Shuttle Solid Rocket Booster June 22, 1987 The Honorable James C. Fletcher Administrator National Aeronautics and Space Administration 400 Maryland Avenue, S.W., Room 7137 Washington, DC 20546 Dear Jim: ~ am pleased to submit herewith the fourth interim report of the National Research Council's Panel for the Technical Evaluation of the Space Shuttle Solid Rocket Booster. The baseline design is essentially complete with only a few choices remaining to be made. The program of testing full scale implementation of the design is about to begin. This report addresses several of the more important questions remaining about both the design and the verification program. Background Since submitting our third report, the Panel has met formally four times. In addition, groups of our members have participated in four technical meetings with personnel from the Marshall Space Flight Center, Johnson Space Center, Langley Research Center, Morton Thiokol, Inc., and several of its subcontractors. Members have visited several subcon- tractors to review capabilities to cut and finish case hard- ware. The cooperation that we have received from personnel from NASA and Morton Thioko~ has been outstanding. The members of the redesign team have shared their technical insights and information about the program graciously and effectively. Current Status The direction of the redesign program has not changed significantly in the last several months. The focus of attention has been on refinements to the baseline design and preparation for testing the new design to verify that it meets specifications. 25 The National Research Council is the principal operating agency of the National Academy of Sciences and the National Academy of Engineering to serve government and other organizations
Letter to the Honorable James C. Fletcher 2 Schedule . The realities of this complex redesign, develop- ment, manufacturing, testing, and verification program, however, have forced substantial delays in the test program. For example, the f irst test of the new case field j oint hardware in a short-duration burn, original ly scheduled for January, now will not take place until ache end of July. We understand that the need] to prepare new mold tool ing for the insulation of the modified J-seal design contributed in part to your decision to change the launch schedule from February to June 19 8 8 . The target date f or del ivering the f irst f ~ ight set to Kennedy Space (enter has changed by about two months, to the middle of December 1987, shill Leaving considerable pressure on the booster testing program. Contingencies. More attention is now being paid to contin- gencies, but we remain concerned that alternatives for certain critical aspects of the design are not being pursued with sufficient vigor to minimize delays in the program should it become necessary to turn to alternatives. We understand the view of the program office that the delivery schedule will only be met by concentrating intensively on the baseline design--the principal need for this concentration being the shortage of manpower rather than budget. However, as we have stated before, paying greater attention to the development of alternatives could, on balance, reduce the risks to the schedule. The types and levels of risks to schedule and performance that are acceptable can only be decided by NASA management. We understand that you have reviewed the program's plans for pursuing design alternatives in critical areas in light of the schedule for delivery and launch and have decided that the risks to the program are acceptable. The Development and Verification Program The firing of the full-scaJe, full-duration Engineering Test Motor (ETM-1A) on May 27th raised the curtain on the most important part of the development and verification program. Although the ETM-1A could not incorporate much of the new hardware, the firing has given useful engineering information and exercised both the test facility and personnel. The essential features of the two most critical aspects of the new design, the case field joint and case-to-nozzle joint, will first be tested in the fulI-scaJe, fu11-duration firing of Development Motor No. 8, currently scheduler for August 23rd. 26
Letter to the Honorable James C. Fletcher 3 Reliability and Margins of Safety. Contemplation of the next year of scheduled full-scale, fu11-duration firings of demon- stration and qualification motors and the tests of full-scale hardware for short duration, reminds us once again of the unique feature of the solid propellant rocket: the flight article itself cannot be test-fired before launch. To compen- sate for this characteristic, the test program must establish reliability and margins of safety in special ways. Further- more, extraordinary care must be taken to assure that the materials and processes used to fabricate and assemble both test and flight articles are tightly controlled so that all are, as nearly as practical, identical. It is clear that conducting a statistically meaningful number of full-scale, full-duration ground tests of the rede- signed motor to establish reliability and margins of safety is not feasible. In circumstances such as these, several prac- tices can be undertaken to increase confidence in the design. One is to make detailed predictions of performance before major tests to demonstrate understanding by comparing the predictions with the results of equally detailed measurements. The redesign team intends to make explicit predictions in its test plans in the future and we will follow progress in this regard closely. Testing with Defects. A complementary method is to test the adequacy of margins of safety by introducing intentional flaws or defects into test articles. The redesign program relies on testing with inflicted flaws to verify the design where the flaws are of types not likely to be detectable by available methods of inspection. Tests with inflicted damage currently must demonstrate not only that the primary and redundant secondary seals are not violated under these conditions but also that they do not show signs of thermal distress. Since all types and sizes of flaws cannot be tested, a key issue in this regard is what constitutes the most severe flaw that has a reasonable probability of occurring. Testing with flaws that are not likely to occur in practice may yield mean- ing~ess or misleading results. Subscale tests have shown that a small direct path between mated parts of the insulation can result in a gas jet that impinges on an O-ring. Major damage can result from such a flaw. Not clear is whether these "engineered" flaws are representative of the most serious defects that might be expected to result during manufacturing and assembly. We conclude that determination of the "worst credible" flaw is essential for the success of the certifica- tion program. 27
Letter to the Honorable James C. Fletcher Another possible cause of thermal distress is circumferen- tial flow around the periphery of the joint. The JES test series, which has provided much understanding, has a burn duration of less than one second. There is concern that this duration may not be sufficiently long to reveal other possible modes of failure. Case-to-Nozzle Joint. The area of our greatest concern regarding testing with inflicted flaws, however, is the case-to-nozzle joint. Defects in the form of holes in the adhesive may be introduced as the nozzle housing slides past the case dome during assembly. Since the integrity of neither the adhesive nor the wiper O-ring will be assured by test after assembly, neither can be considered seals. Thus, the performance of the nozzle must be tested with defects in both. In our opinion, the probability is high that the primary O-ring will experience thermal distress when tested with a flaw that extends directly through both the adhesive and wiper O-ring. Therefore, Me conclude that the current baseline design of the case-to-nozzle joint is likely to fail to meet the requirement prohibiting the appearance of thermal distress when tested in the presence of this flaw. Consequently, we recommend that the redesign team address this problem with urgency. The worst credible flaw should be determined systematically using assembly demonstration tests and supporting analyses. Potentially modest modifications of the design which might preclude O-ring damage with such flaws should be reviewed. Aggressive pursuit of an alternative design, including acquisition of hardware and testing, also appears warranted to minimize consequences for the schedule if the current design or modifications of it cannot meet require- ments. Verification of Pressure Seals. Another matter of concern is the development of satisfactory procedures for verifying the O-ring pressure seals after assembly. The maximum allowable leak rate past pressure seals must be below the rate that would result in thermal distress, including erosion, of the O-rings but not so low that it cannot reliably be detected by the adopted method. The redesign team currently favors a maximum leak rate of 0.01 standard cubic centimeters per second. The current leak check protocol includes establishing the test pressure in a volume adjacent to the O-ring and watching for a decay of 2 psi within 10 minutes. Laboratory tests conducted by the redesign team indicate that thick coatings of the grease can interfere with the detection of leaks of this size. However, large amounts of grease are also 28
Letter to the Honorable James C. Fletcher . 5 thought to protect against leaks during firing. Removing as much grease as possible may assist in the detection of smaller leaks but may make the design less forgiving in operation. Testing at high pressure may also reveal small leaks, but the procedures are sensitive to very small changes in tem- perature. Under current leak check protocols, changes in temperature of the order of a few tenths of a degree Kelvin at a test pressure of 1,000 psi can falsely either indicate a leak or mask one. Also, the higher the test pressure the greater the potential for damaging the insulation, insulation bond lines, or O-rings. Work is underway to develop the most appropriate leak check procedures in light of the several trade-offs involved. The additional data from this program should be available in the next month. Further analysis and testing appears to be in order to determine if a higher maximum leak rate, hence lower test pressure, might be acceptable. Also, in some cases it might be appropriate to measure leak rates using alternative approaches, such as by measuring the buildup of test or tracer gases downstream of a seal rather than by measuring the decay of the test pressure upstream. Nozzle Ablatives The redesign of the internal parts of the nozzle incorpo- rates the carbon-phenolic materials used in the previous design, with changes in the orientation of the carbon fibers in the matrix and improved specification and qualify control of materials. The changes are intended to provide reduced, more consistent erosion and to eliminate fracture problems. Early evaluation of some features of the new design (fiber angle in two nozzle parts) has been provided by three full- sca~e static tests (two in the filament-wound motor program and the ETM-1A). Because the performance of ablative mate- rials is relatively unpredictable, a number of tests are needed to establish the adequacy of the design, materials, workmanship, and production. Thus far, we are encouraged by the work on redesign of the ablative parts of the nozzle. Nonetheless, as many tests of these parts should be scheduled as practical. Furthermore, the performance of the ablative parts in particular, should be closely monitored and evaluated by inspection of recovered flight motors on a continuing basis beyond the verification phase of the program. 29
Letter to the Honorable James C. Fletcher Materials and Process Control 6 As we previously observed, a successful solid rocket motor program depends on careful control of the materials and proc- esses used I manufacture both the test and flight articles. The manufacture of nonmetallic parts under carefully defined specifications is a continuing concern. O-r; He M~teri al . The effort of the past year to find an O-ring material with better low temperature resilience while meeting all other performance requirements has not been suc- cessful. Therefore, we agree that the fluoroelastomer used previously is an appropriate choice at this time since it will be used in combination with the modified metal parts and heaters to provide temperature control and hence satisfactory . . . resl fence. Specifications. An accurate and thorough definition of the fluoroelastomer material is urgently needed, however. The specifications supplied to us to date are incomplete. Once the final specifications are determined, it will be necessary to review the existing inventory of O-rings to verify that they were manufactured and tested in conformance with the specifications. Similar comments apply to the specification of composition, processing, performance, and acceptance of the adhesives that are to be used in the new design. While the adhesives and O-rings may contain proprietary products, adequate control of materials and processes requires their full and complete description. Only with this informa- tion can adequate engineering control be exercised over the manufacturing process. To maintain quality control for mate- rials used in critical applications, the suppliers must be restrained from making any changes to formulations or proc- esses without engineering approval. Means are available for sharing essential proprietary information to assure control while maintaining confidentiality. We recommend that agree- ments with suppliers be negotiated and implemented to permit stringent engineering control of the materials and processes used in the solid rocket booster consistent with the protec- tion of proprietary interests. The requirement for finishes of metal surfaces in contact with O-rings poses a challenge to manufacturing techniques. We are pleased to note both the reported improvements in machining techniques and laboratory studies of the relation- ships between surface finish and sealing effectiveness. Much work remains to be done. We will continue to fo1 low progress in these areas, including the measurement of surface finish. 30
Letter to the Honorable James C. Fletcher Activities at Kennedy Space Center - / As the design comes into final focus, our attention has also turned to the activities at Kennedy Space Center, where case segments are inspected, stored, and assembled into the flight articles. Inspection. Regarding nondestructive evaluation of the integ- rity of case segments at KSC, previous practice has been to inspect segments visually when they arrive. The inspections are intended to detect flaws in bonds between the propellant and inhibitors and between the insulation and case. The segments are then stored in a vertical position until use. During storage, the segments are subjected to varying condi- tions of temperature and humidity and the case-to-insulation bonds are subject to various stresses. Individual segments can be stored for periods from months to years. In ballistic missile programs, similar circumstances have lead to extensive programs to control environmental conditions under which the motors are stored and to determine the response of the com- ponents to aging. We suggest that similar considerations apply in the case of the Shuttle SRB. Further, we recommend that motor segments be inspected again using more than simple visual inspection when removed from Jong-term storage before assembly. Process Control. Environmental conditions may also affect the . assembly of the segments into boosters. The new design re- quires that a critical bonding process be performed during the assembly of field joints. While specific requirements are not yet defined, control of temperature, humidity, and cleanliness are essential for achieving reliable bonds. Other critical processes, such as lubrication of the field joint O-rings, are also performed in the Vehicle Assembly Building, emphasizing the need for a higher level of cleanliness in the facility than previously existed. We recommend that the contemplated design and installation of an environmental control system suitable for protecting case segments during assembly be vigorously pursued. The objective should be to insure the most complete environmental control feasible. Requirements for enhanced cleanliness and environmental control should be reviewed, established, and implemented. 31
Letter to the Honorable James C. Fletcher Plans for Flight Evaluation and Product Improvement - 8 While we, like NASA, have been preoccupied primarily with issues of redesign and testing, we continue to be concerned with the evaluation of performance and improvement of the product after flights resume. The 51-L accident demonstrated the importance of conducting an on-going program designed both to evaluate the reliability and performance of the booster and its components in flight and to provide a basis for near-term product improvement. A flight evaluation plan should provide for a comprehen- sive program of measurements in flight, inspection of recovered motors, and expeditious assessment of results of each flight before the next one. The verification plan currently envisions collecting extensive performance data on the first six flights of the new booster. But NASA must also remain in position to identify weaknesses in the design, manufacture, or assembly of the boosters in the long term so that corrective actions can be taken if necessary. Collection of performance data therefore should be continued until the data base is sufficient to demonstrate the required reiiabil- ity and answer outstanding questions about production control and component functioning. Thereafter, surveillance data should be gathered and the program should constantly be alert to the need for taking additional flight data. The flight evaluation program should also provide for design, production, evaluation, and introduction of limited modifications of the SRB in response to concerns of engineers regarding reliability of flight performance or problems in manufacturing, inspection, or assembly. The program should provide for preplanned opportunities for changes in order to introduce improvements required as a result of analysis of flight performance data. Finally, once again I would like to express our appreci- ation of the willingness of the redesign team to discuss technical details and plans with us. The team members are dedicated professionals working hard to achieve their goals and to keep us informed, too. On behalf of the Panel, I remain cc: Adm. Richard H. Truly Panel Members 32 Sincerely, H. Guyford Stever Chairman
