AIAA 2003-5108 RSTANDIMG OF SPACE SHUTTLE RSRM SULATIONJOINTS B. B. McWhorter, D. E. Bolton, S. V. Hicken, L. D. Allred, D. J. Cook ATK Thiokol Propulsion Corp., Brigham City, UT ABSTRACT The Space Shuttle Reusable Solid Rocket Motor by exposure to hot gases. After early operation, a J- (RSFUvl) uses an internal insulation “J-joint” design joint that has been pulled apart will come back for the mated insulation interface between two together as the 1-joht deformation decreases. This J- assembled RSRM segments. In this assembled joint heating event is relatively short and occurs only (mated) segment Configuration, this J-joint design during the first part of motor operation. Internal serves as a thermal barrier to prevent hot gases from instrumentation was developed for another N1-scale affecting the case field joint metal surfaces and 0- static test in February 2000. The static test rings. A pressure sensitive adhesive (PSA) provides instrumentation did indeed prove this theory to be some adhesion between the two mated insulation correct. Post-test inspection revealed very similar surfaces. In 1995, after extensive testing, a new charring characteristics as observed on RSRM-55. ODC-free PSA (free of ozone depleting chemicals) This experience of the development of a new PSA, was selected for flight on RSRM-55 (STS-78). Post- its testing, the RSR1M-55 flight, followed by the J- flight inspection revealed that the J-joint, equipped joint investigation led to good “lessons leamed” and with the new ODC-free PSA, did not perform well. to an additional fundamental understanding of the Hot gas seeped inside the J-joint interface. Although RSRM J-joint function. not a flight safety threat, the J-joint hot gas intrusion on RSRM-55 was a mystery to the investigators since INTRODUCTIAONDN B ACKGROUND the PSA had previously worked well on a full-scale static test. A team was assembled to study the 1-joint This paper presents an investigation into an internal and PSA further. All J-joint design parameters, rubber joint of the Space Shuttle Reusable Solid measured data, and historical performance data were Rocket Motor (RSRM). Before the explanation of re-reviewed and evaluated by subscale testing and this rubber joint is given, some basic background analysis. Although both the ODC-free and old PSA information regarding the RSRM assembly will be were weakened by humidity, the ODC-free PSA given here. The RSRM is composed of four strength was lower to start with. Another RSRM full- segments. These four segments are assembled, or scale static test was conducted in 1998 and mated, to form a full length RSRZM. The assembly of intentionally duplicated the gas intrusion. This test, the four segments makes three “field joints” between along with many concurring tests, showed that if a J- the segments. Each field joint is composed of a joint was 1) mated with the new ODC-free PSA, 2) “tang” and “clevis”, which are the mated parts of exposed to a history of high humidity (Kennedy steel case ends. Each case field joint is sealed by Space Center levels), and 3) also a joint which two O-rings. This design is a redundant sealing experienced significant but normal joint motion (J- system joint deformation resulting from motor pressurization dynamics) then that J-joint would open (allow gas There are two other features in the joint design that intrusion) during motor operation. When all of the serve as thermal barriers to this redundant sealing data from the analyses, subscale tests, and full-scale system. These features are a “capture feature” 0- tests were considered together, a t h qe merged. ring and an internal insulation “J-joint” Most of the joint motion on the RSRM occurs early in configuration. This internal insulation J-joint design motor operation at which point the J-joints are pulled is the configuration for the mated rubber insulation into tension. If the new PSA has been weakened due interface between two assembled RSRM segments to humidity, then the &joint will partially pull apart (see Figure 1). Since there are three field joints, (inboard side), and the J-joint surfaces will be charred the= are also three J-joints - forward, center, and aft. Copyright Q 2003 by ATK Thiokol Propulsion, A Division of ATK Aerospace Company, Published by the American Institute of Aeronautics and Astronautics, Inc. with permission. 1 This J-joint design is the first t h e db arrier that subscale testing of several PSA types, a new ODC- prevents hot gases from affecting the case field joint free PSA was selected. Its strength was less than the metal surfaces and O-rings. old PSA. However, it was thought that the interference fit and pressure actuation of the J-joint design were the controlling parameters that made the J-joint an effective hot gas barrier. The ODC-free PSA was successfully tested on an RSRM full-scale static test (referred to as FSM-5). Armed with this success, the new ODC-free PSA was selected for flight on STS-78.T he flight set of solid rocket motors for STS-78 was “RSRM-55”. (The remainder of this report will refer to this solid rocket motor flight set as “RSRM-55”. The new ODC-free PSA will be referred to as the “new PSA” hereafter in this report.) THE PROBLEM The flight occurred as planned. Post-flight inspection revealed that the J-joints, assembled with the new PSA, did not perform well. Hot gas had Figure 1. Assembled RSRM Showing seeped into the J-joint interfaces. The inboard Internal Indatlon Jjoint Conflguratlon portions of the J-joint interface surfaces were heavily for the Three Fbld Joints heat affected and charred on four of the six joints (there are two motors for each flight set - hence six In this assembled (mated) segment configuration, this total field joints). The four severely affected J- rubber J-joint design serves well as a barrier to hot joints were the aft and center J-joints for each motor. gases. It works so well, in fact, that the J-joint almost The forward J-joints were only slightly heat affected acts as a seal, but it is not officially considered a seal. or sooted in four small areas. Figure 2 illustrates Since the J-joint is composed of internal rubber with a cross-section schematic where the severe insulation, this component can be visualized as two charring was located. Figures 3 and 4 show example rubber parts joined together, or pressed against each photos of the heat affected J-joint mating surfaces. other, in the assembled RSRM configuration. On one As shown in Figure 2, the heat effect and charring of side of the interface is the rubber insulation part the rubber interface surfaces were limited to only the referred to as the “J-leg” (see Figure 1). On the other J-joint surfaces inboard of the radius start. Although side is the rubber insulation part referred to as the the J-joints did not perform as intended, their “clevis”. (Hereafter in this paper, the term “J-joint” outboard (outer diameter) interfaces appeared to applies to the mated configuration of these rubber have remained closed and did work as a thermal components.) barrier (this region is also identified in Figure 2). No joint metal or O-rings were affected by hot gas. Three aspects of the J-joint design enable it to be an Some soot did penetrate past the radm start on at effective gas barrier: 1) the mated joint makes an least one of the joints, but this joint still served well interference fit, 2) the configuration of the J-joint as a thermal barrier. enables the internal motor gas pressure, at least partially, to produce a pressure actuated seal between the two mated insulation surfaces, and 3) a pressure sensitive adhesive (PSA) provides some structural adhesion between the two mated insulation surfaces. SELECTION OF A NEW PSA In 1995, selection of a new environmentally friendly PSA was initiated. The desired new PSA was to be free of ozone depleting chemicals (ODC-free). After 2 Although no heat reached the O-rings or metal Heat Effect Depths surfaces, this J-joint gas inmion was a first time At &leg TP ever occurrence in the RSRM program. RSRM motors had been flown over 100 times before with no PSA failure. The J-joint gas intrusion on RSRM- 55 was a mystery to the investigators since the new PSA had previously worked well on the FSM-5 full- scale static test. Compounding the problem further, the forward J-joints of RSRM-55 experienced only minor heating in a few small areas, whereas the center and aft J-joints were heat affected extensively. INITIAL INVESTIGATION Figure 2. Heat Effect and Char on the J-joint Intmrdaca Surfacer of RSRM-6S A team was quickly assembled to study the J-joint, PSA, and all other materials and processes. The evidence confronting the team was perplexing. A first impression was that the new FSA somehow caused the J-joint to have gas intrusions. However, the testing that was done for the new PSA would suggest that the new PSA could not be the cause for gas intrusion. It was true that the only obvious change to the J-joint mating process was the adoption of the new PSA and ODC-free cleaning process. But, the following were also true: 1) p e" new" PSA had been tested on the aft J- joint of the FSM-5m otor successfully. During no& operation, this aft joint insulation experiences the most severe heating effects and thermal ablation compared to the other joints. So logically, it was Figum 5. Photograph of Heat-Affected and thought that this joint would provide the most Charred &joint Surface From FISRM-66 conservative test bed for the J-joint and PSA system. 2) Lab tests had shown that the strength of the new PSA was lower than the old PSA. However, the design of the J-joint was such that it was thought to be pressure actuated. In other words, the J-leg should remain seated to the opposite clevis side of the joint by motor chamber gas pressure. Under this assumption, the J-leg should remain seated to the clevis even while the clevis is experiencing deformation due to the chamber pressure. Figure 5 illustrates this deformation under motor chamber pressure and how the J-leg will move With the clevis and thereby remain seated to the clevis in this condition, Therefore, the reduction in PSA strength was not considered an important factor. 3) When the segments are mated, an interference fit is produced between the J-leg and clevis. This Figure 4. Photograph of Heat-Affected and interference fit should allow the J-leg to remain in Charred J-joint Surface from RSRM-55 contact with the clevis during the prelaunch time frame. The J-leg-to-clevis interface should never go into tension during prelaunch. 3 4) Over 100 RSRM motors had either flown or had (other than the PSA) were installed prior to the joint been static tested with the “old” PSA, and no heat assembly; and 4) the J-joint assembly process, effects were ever observed in a normal rubber including use of the new PSA and ODC-free process, interface of the J-joint (this statement does not was inadequate. include static tests in which intentional flaws were created in the I-joint to study the various aspects of Material and fabrication records were reviewed to joint performance) determine if any manufacturing material or process changes could have triggered the J-joint problem. Processes at Kennedy Space Center (KSC) were reviewed. All procedures were in agreement with Thiokol and KSC requirements. No changes were - found that could have affected the J-joint except Inboard 1 for, the incorporation of the new PSA. =J.I. J is srnent The RSRM-55J -joints were studied to insure that their configuration was within family. Preflight inspection data was studied. There was normal axial engagement during assembly to produce the desired interference fit between each J-leg and corresponding clevis. Photos and samples of the post-flight J-joints were studied. The condition of the mated surfaces were observed and measured. Figure 5. JJoint Configuration The heat affected regions of the J-joints were During Motor Operation mapped and recorded. Other than the anomalous heat affected rubber, the J-joints were normal. Since the initial investigation revealed contradictory evidence that the new PSA caused the gas intrusion, Lab analysis of the samples revealed that the heat nothing was assumed. Broader questions were effect was severe and resulted from material addressed. Was there another material, joint dynamic, temperatures exceeding 1800’F. The char (complete assembly process, or motor performance parameter material thermal decomposition) depth on the J-leg that caused the J-joint gas intrusion? Was an and clevis surfaces varied from 0 to 40 mils (Figures undetected bad lot of a material used in 3 and 4). J-joint insulation char was only observed manufacturing? Did the motor internal operational along the J-joint surfaces inboard of the J-joint radius heating environment change? Did something in (Figure 2). processing change - such as transportation in an unusually rough or cold environment? When did the In conjunction with the hardware study, the J-joints opening and the heating occur? Were the J- environments to which RSRM-55 hardware was joints opened at ignition or prior to ignition? Or, exposed were studied. No data was found that could an unusual post-operation event during re-entry indicated that the RSRM experienced anything other have caused the J-joints to open and then become than a normal flight environment. Vibration and heat affected? temperature records during transportation were normal. Processing environments were within An extensive fault tree was developed. The fault tree family. did not assume that the J-joint gas intxusion was due to the application of the new PSA. Instead, all An ODC-free cleaning process for the J-joint had possible root causes for the event were considered. also been added to the assembly. However, lab tests The fault tree branched into four mjor categories of showed that the ODC cleaning process had no causes, and each of those causes were subdivided adverse effects on the J-joint or PSA. further and those cause were divided until a possible root cause could be postdated. The four major areas A re-examination of post-flight history data was for defining causes were 1) a motor performance made so that the condition of RSRM-55c ould be parameter caused these J-joints to open; 2) some compared and contrasted to earlier post-flight data. different joint dynamics were encountered which Much of this data are thickness measurements of the caused the J-joints to open; 3) inadequate materials 4 remaining insulation components. Internal insulation new PSA, prior to its incorporation into the RSRM- is ablated and eroded during flight. Thickness 55, had been thought to be conservative and measurements of the remaining insulation thorough. Since the root causes for J-joint gas components are made for every flight. The intrusion were not understood, the team concluded remaining inboard thickness of the J-joint rubber that the prior PSA testing had been missing components, including the J-leg and clevis, were something. As RSRM production resumed using the within family. The thickness measurements of all old, proven PSA, the investigation team now other internal insulation components, such as concentrated on a longer and more in-depth study of propellant grain inhibitors and stress relief flaps, were the J-joint and PSA physics. The other variables within family. The case insulation thermal causing the gas intrusion needed to be understood decomposition depths were within family. This before another new ODGfree PSA or modified PSA review of post-flight data concluded that all of the could be considered for the RSRM program. measurements and observations of RSRM-55 were within family, except for the anomalous J-joint heat In addition to the fault tree, an event tree was effects already under investigation. No other similar developed. The event tree laid out all RSRM events, heat effects had been observed prior to RSRM-55. non-operational and operational, from pre-ignition through splashdown. The purpose of this event tree Reviews were conducted of earlier certification static was to define the timing for the opening and tests in which the J-joints were flawed to demonstrate subsequent heating of the J-joint. As studies of all the robust RSRM field joint design. One full-scale possible root causes of failures were conducted, the test, QM-6,in corporated a ‘tvave defect” which did scenarios that could provide the thermal and result in heat effect and sooting patterns similar to structural boundary conditions to create the J-joint those measured on RSRM-55. In QM-6, the wave heating effects were also studied. All events were - defect was made so that the J-leg and clevis did not investigated including motor bending due to contact for a circumference of several inches. The ignition of the shuttle main engines, RSRM ignition, remainder of the joint was normal. Post-test and all other operating phases of the RSRM. Events inspection revealed heat affected rubber in the wave prior to motor operation as well as post-motor defect region. In this region, soot patterns were very operation, such as re-entry events, were considered. similar to those patterns on RSRM-55. Obviously, As the investigation proceeded into the many this early certification test, performed with a flawed possible root causes defined by the fault tree, J-joint and with the older and stronger PSA, produced evidence that either supported or refuted an event heating characteristics that were similar to RSRM-55. time for the gas intrusion was formed. With this However, RSRM-55 had no wave defect flaws, and information, the event tree table was filled, and all data indicated that the shape of the J-joints were likely scenarios that would create a J-joint gas normal. intrusion emerged. The design background of the J-joint was reviewed. This continued in-depth study reviewed many A structural analysis done earlier showed that the hypotheses. With each new hypothesis, other high pressure gas during motor operation would force possible variables of the physics were considered. the J-leg against the clevis - the J-joint was pressure Subscale testing, lab testing, and analysis were actuated. Under this loading, the strength reduction conducted to determine the merit of these of the PSA should not be a cause for J-joint opening. hypotheses. In the end, two other variables, one for It was concluded that the prelaunch and flight a storage condition and the other for an operation environment must have introduced another, and as of condition, were determined to be important. As with yet, unknown variable into the physics of the J-joint most unexpected events, a combination of root and PSA operation. causes would explain this PSA failure and J-joint gas intrusion. - DETAILED bWESTIGATXON ROOTC AUSE AND TIMING HUMIDITY Obviously, there were some other combinations of PSA strength tests had already been completed prior physics at work that prevented the new PSA from to RSRM-55. The new PSA strength had already functioning properly in the J-joint. Testing of the been determined to be less than the old PSA strength. 5 But, the old assumptions of the operating physics Due to loads, dynamics, and the structure of the allowed, in theory, a weak PSA to be acceptable. For assembled motor, each field joint experiences the reasons stated above (interference fit and pressure somewhat diffemt motion. Post-flight inspection of actuation) it was thought that this strength parameter RSRM-55r evealed that the amount of heat effect in was not important. each J-joint strongly correlated to joint position - forward, center, or aft. There was little to In early tests, humidity had been included. insignificant heating in the forward J-joints on both Unfortunately, the exposures were for a short period motors. The center J-joints, however, were heavily of time and were only for the PSA application heat affected and had the most char and sooting. The process. These tests showed no significant effect of aft J-joints did experience severe heating and humidity. Extended humidity exposure was not done charring, but not as bad as the center &joints. for the J-joint rubber prior to the PSA application. During the investigation, conclusions of old During the investigation, the strength tests were structural analyses were revisited, and new analyses performed again, but this time, these tests were were conducted. The case joint motion was performed under different conditions. Samples were examined closely. During motor operation, the exposed to periods of humidity that would simulate internal propellant grain deforms due to the gas the KSC conditions prior to joint assembly. PSA pressure distribution across the surface of the grain. strength tests were conducted for a matrix of aged Also, the case membrane strains more than the case and humidified samples. The results showed that joint regions since there is extra thickness in the case humidity exposure time was an important governing joint. This elastic straining causes the case joints to variable that adversely affected PSA strength. rotate - that is, as the membrane on either side of the joint expands, the internal joint surface becomes Both old and new PSAs were adversely affected by slightly convex. This joint rotation, combined with aging in high humidity environments. However, the motion of the internal propellant grain since the strength of the old PSA started out higher, deformation, results in a relative displacement its strength remained higher in humid environments between the two sides of the J-joint. The J-joint compared to the new PSA. deforms and tends to go into tension as illustrated in Figure 5. Extended humidity exposure effects were considered important since the new PSA had been certified and The amount of resulting axial displacement at the J- tested in Utah where the climate is dry. For flight leg-to-clevis interface could not be precisely motors, the J-joint rubber is exposed to long periods calculated. The combination of the visco-elastic of high humidity prior to joint assembly as the properties of the propellant grain and rubber segments are stored at KSC. After assembly, the PSA insulation components in response to the case is weakened by this moisture. Although, believing deformation and internal pressurization is an the J-joint was pressure actuated during operation, the extremely complicated problem. However, the investigation team was confident that the weakening available results showed that the forward field joint effects of humidity was an important variable in should experience the least amount of J-joint motion determining the root cause for the PSA failure. during operation. The center field joint experiences the most J-joint motion. The aft field joint JOINT MOTION experiences more motion than the forward, but not as much as the center J-joint. This variation in J-joint Another operating condition of the joint that had been motion corresponded exactly with the varying studied during the design of the RSRM had to do with amounts of heat effect observed in the J-joints of joint motion that occurs during motor operation. RSRM-55. Joint motion had been studied thoroughly for case joint structural integrity and for proper O-ring This correlation was too much of a coincidence not operation. Exactly how joint motion might fit into to to have something to do with the physics that the physics of the J-joint and PSA failure, however, resulted in hot gas intrusion into the J-joints of had not been as thoroughly studied. RSRM-55. But again, the J-joint was thought to be completely pressure actuated during operation. In other words, internal gas pressure would force the J- leg against the moving clevis insulation. With 6 pressure actuation, the PSA did not have to be a strong adhesive. It was apparent that this joint 1 physics needed to be reexamined. LNcomm PREssuRlE ACTUATION The re-analysis of this theory proved that the J-leg is not fully pressure actuated. Along the inboard portion of the. J-leg, from the J-leg radius inboard to the J-leg tip (Figure 6),t here is little to no pressure m actuation. The reasons can be explained as follows: Consider a J-joint assembled with an ineffective PSA. Upon J-joint assembly, there exists contact between Figure 6. J-joint Auembled Configuration the insulation clevis and J-leg (see Figure 6). Prior to Motor Operation Showing Between the J-leg tip and radius, there exists a certain Uncompread Fme Volume amount of free volume created by the shapes of the mating surfaces. This geometry was created intentionally to assure J-leg tip contact with the clevis and to assure an interference fit at assembly. When 1 the J-leg is pressed into the clevis by motor gas b impressed pressure, the free volume is allowed to compress, as C ing L 3rd illustrated in Figure 7. The pressure of the free I' volume increases with this compression until static equilibrium is achieved at or near the pressure of the motor chamber. This situation now results in a J-leg with essentially the same pressure inside the joint as there is in the motor chamber. This is essentially equivalent to the condition just after assembly, before motor operation, when there is atmospheric pressure (14.7 psi) on both sides of the J-leg. Only now, Figure 7. Jjoint Deformation Configuration during motor operation, there is roughly motor During Motor Operation Showing pressure on both sides of the J-leg. So contact Compresssd Free Volume pressure at the tip is approximately the same as it is just after assembly - hence no pressure actuation. In A key point here is the observation that the free fact, due to clevis &formation during motor volume geometry is such that it can be compressed operation, as shown in Figure 7, the J-leg must be as the J-leg rubber presses it. If the volume was pulled away from its assembled position, and so there fixed so that it would not be compressed, then the may be even less pressure where the J-leg tip contacts pressure in the free volume would remain at 1 the clevis. atmosphere during motor operation. In that case pressure actuation may work. For the region outboard of the radius, the condition of having a fixed volume is almost that situation. In Figure 6, the volume adjacent to the capture feature O-ring is surrounded by rigid case metal hardware. During motor pressurization, the J-leg rubber presses into this outboard volume somewhat. But mostly, the volume is controlled by the rigid surfaces of the case. Therefore, this region outboard of the J-leg radius is pressure actuated as illustrated in Figure 7. It remains pressure actuated whether the J-leg 7 inboard region remains in contact with the clevis side edge of a suction cup away), then the pressure or not (Figures 8a and b). actuation of a rubber-to-rubber interface, with PSA, will not be effective. I- In summary, the length of the J-leg from its radius inboard to its tip is not pressure actuated (refer back to Figure 7). The length of the J-leg outboard its radius, and including its radius region, is pressure actuated. During motor pressurization, the insulation clevis deforms due to joint rotation and propellant grain deformation. This motion tends to pull the J- P leg towards the clevis. If the PSA is strong enough to hold it (Figure 8a), then the J-leg will maintain contact following pressure equilibrium. If, however, this clevis movement is significant and if the PSA (a) Pmper PSA Operation has weak adhesive strength, the J-leg tip will peel away from the clevis (Figure 8b). The J-leg length outboard of the radius region, however, will be sucked into the clevis since motor chamber pressure will exceed the pressure in the volume in the capture feature region. THE RSRM-55 PSA FAILURE THEORY When all of the data from the analyses, sub-scale tests, FSM-5,a nd RSRM-55 were considered together, a theory emerged that combined the effects of 1) long term humidity exposure weakened PSA and 2) J-joint motion during operation with the (b) Failed PSA correct physics of pressure actuation. It was postulated that pressure actuation by itself will not hold the J-joint together. The adhesive strength of Figure 8. &joint Defomation During the PSA is at least needed during a part of motor Operation With Proper PSA Operation operation. It was postulated that the new PSA would and With Failed PSA be strong enough to work if the J-joint and PSA was affected by one of the two variables - humidity or A second advance in the understanding of the - significant joint motion but not both. pressure actuation effects on the PSA was made in the laboratory. These tests involved tensile and peel PROOFFO R THE PSA FAILURTEH EORY testing rubber/PSA/rubber samples under high gas pressure and under only one atmosphere of pressure. Because of the complicated nature of the J-joint It was noted that tensile buttons tended to have higher physics, the theory stated above could only be strengths under high gas pressure environments than proven by use of a full-scale RSRM static test. The under ambient pressure. This indicates that the investigation team understood that one data point rubber-to-rubber interface, with PSA but no volume, had already been gathed - the effects of joint does tend to have some pressure actuation effect in motion alone. FSM-5 had demonstrated that the new tensile mode. This difference in strength was not PSA will work on an aft J-joint, which has observed for peel tests - there was no pressure significant motion. In that test, however, the J- actuation effect for peel modes. The important point joint/PSA system had not been exposed to higher to recognize is that if the failure mode is from the humidity for extended times. This was the main data inside out (like a suction cup), then the pressure point that indicated that both humidity and joint actuation effect of a rubber-to-rubber interface, with motion were needed to fail the new PSA. PSA, can be important. If, however, the failure mode peels from the edge to the inside (like peeling the 8 The investigation team decided that the next full- TIMING AND DURATIOONF T HE mATING EVENT scale test (FSM-7) needed to demonstrate what would happen with both variables affecting the aft J-joint. Now, the variables that caused new PSA to fail in the So, the FSM-7 aft J-joint, using the new PSA and the J-joint were understood. However, the exact timing ODC-free cleaning process, was also processed with a for the failure was not known. The duration for simulated KSC humidity (a process that targeted a which the J-joint was exposed to hot gas was not simulated exposure of 60% to 80% relative known. Thermal and struchlral analyses were used humidity). This would make the aft J-joint experience to approximate the answers to those questions. the combined effects of humidity and joint motion. The investigation team felt that this would simulate THERMALMODJ%IRNEGSU LTS as close as possible the aft J-joint on RSRM-55. Thermal models were needed to reproduce the char To further test that joint motion alone would not configurations measured on RSRM-55. These cause the PSA to fail, the center J-joint on FSM-7 models, if they accurately reproduced the char was processed with the same variables but without configurations, could then possibly address the humidity. timing and the duration of the J-joint hot gas intrusion. The forward J-joint, which experiences little J-joint motion, was assembled with long exposure to Charred J-leg specimens were taken from RSRM-55. humidity. The new PSA was applied to one half of Cross-sections of the J-leg were made, and the char the J-joint surface, and the old PSA was applied to configurations were studied. Figure 9 shows a the other half. Both halves had partial flaws designed typical cross-section of an RSRM-55J -leg. The J- to allow hot gas to penetrate the J-joint well after leg mating surface is identified on the figure. In motor ignition. The gas pressure would reach to the normal operation, this mating surface stays in contact start of the radius region - but not beyond that point. with the clevis, so this surface would not be charred. In RSRM-55, the PSA failure allowed the J-joint to The FSM-7 results were conclusive. The post-test open theEby exposing this mating surface to heat. observations of the aft J-joint revealed strong heat The char depth on this mating surface varied from 30 effects, including charring, on the J-joint interface to 0 mils. surfaces. The charring did not extend pass the radius, which tended to prove that the outboard portion of the Under a charred surface is the pyrolysis zone. This J-joint was pressure actuated (similar to RSRM-55 aft is the material that has undergone heat J-joints). But the inboard portions of the aft J-joint decomposition but has not been fully converted to were definitely exposed to heat during motor char (Char is material that is fully thermally operation. decomposed and carbonized). In Figure 9, the depth of the pyrolysis zone at the mating surface is thin The center J-joint, meanwhile, was not heat affected relative to the charred material indicating that the at all. The new PSA held this joint together - in the heating event was intense but short in duration. joint that experiences the most J-joint motion. The forward J-joints did not open or experience any - heat effect even with the partial intentional flaws. The conclusions from this test were significant. The new PSA was failing during motor operation. This indicated that this heating event did not occur during re-entry, which of course, a static test motor does not experience. The new PSA would fail if the J-joint experienced both extended humidity exposure and significant joint motion. 9 Heating a d It was understood that the char configuration in Chalng of J-hp Figure 9 was indicating the timing of the heating Inboard TP: event for the mating surface. Thermal models were Normal created that could simulate the char configurations Perknnaar 6ent Chs shown in Figures 3,4, and 9. All scenarios were L a mo f considered for the modeling effort. These models J-teg M dng assumed a variety of boundary conditions for this surface- attempt. With this effort, possibilities were examined as follows: 1) J-joint gas intrusion after motor operation and during re-entry: In this case, the models showed that there was not enough heat remaining in the motor after separation to reproduce the char thickness and configuration of the char observed in Figure 9. 2) J-joint gas intrusion early in operation and remaining open throughout motor 0-peration: Thermal models were set up to simulate char Figum 9. RSRM-66 JJeg Cross Section formation in this scenario. In his case, the heating Showing Char Thlokness time would be extensive. Thermal models could not reproduce the char and pyrolysis depths observed in Also noted on this figure, is the typical J-leg tip char. Figure 9 with this case. If 20 to 40 mils of char were This char results fr0mnorma.l motor operation to be produced in this long duration environment, heating of the J-leg inbod diameter surface. This then the pyrolysis depths would have been much char and heat affected depth varies from 0.2 inches to greater than those observed on RSRM-55. The 0.4 inches from the original p-fired inboard surface pyrolysis depths on RSRM-55J -joints were very thin depending on J-joint location. When this J-leg tip compared to the char depths. This kind of char and material is heat affected and charred, the pyrolysis pyrolysis configuration can only be reproduced with gasses build pore pressure in the rubber material, and a short duration but intense heating environment. the material will try to expand. This expansion So, this thermal model did not reproduce the effects normally results in cracking of the charred material. seen. If there is space for it to do so, the J-leg char will 3) J-ioint gas intrusion late in operation: A variety expand and bend other material around it. of two dimensional models and conjugate flow models were created. This scenario would reproduce In Figure 9, there is bent material that is the char of the relative char and pyrolysis depths observed. the mating surface. This char on the mating surface However, these models would not repmduce the was bent by the heat effect that advanced through the sharp corner of non-heat affected material and the J-leg tip. Remember that there were two surfaces that - bent char layer on the mating surface. If the heating were heated the J-leg tip surface and the mating event happened late in motor operation, this material surface. These two surfaces are positioned mghly would have had heat approach it from two directions 90' to each other. at the same time. In this case, the material would not have sharp comers. Instead, the material would be In Figure 9, there is a fairly sharp comer of non-heat rounded as the two char fronts would increase the affected material between the two charred surfaces. heat flux into this region by two dimensional If the heating events for both surfaces occurred at the - conduction. So again, this thermal model did not same time, then this corner would be rounded due reproduce the effects seen in Figure 9. to the effects of two dimensional heating. However, this sharp comer indicates that the heating events for 4) J-ioint gas intrusion early in motor owration the two surfaces occurred at different times. Normal causing early heating with a short duration: In this I-leg tip heating occurs during all of motor operation. scenario, the J-joint opens and the mating smfaces The observable post-fire char of the J-leg tip would char severely early in motor operation. However, be the result of normal motor performance. due to some condition (such as a decrease in J-joint deformation causing the J-leg to re-contact the clevis), the heating event stops shortly after it begins. 10