Extraction and Separation Modeling of Orion Test Vehicles with ADAMS Simulation Usbaldo Fraire Jr.1 Jacobs Engineering, Houston, TX, 77058 Keith Anderson2 ATK Brigham City, Utah 84302 P.A. Cuthbert3 NASA-Johnson Space Center, Houston, TX, 77058 The Capsule Parachute Assembly System (CPAS) project has increased efforts to demonstrate the performance of fully integrated parachute systems at both higher dynamic pressures and in the presence of wake fields using a Parachute Compartment Drop Test Vehicle (PCDTV) and a Parachute Test Vehicle (PTV), respectively. Modeling the extraction and separation events has proven challenging and an understanding of the physics is required to reduce the risk of separation malfunctions. The need for extraction and separation modeling is critical to a successful CPAS test campaign. Current PTV-alone simulations, such as Decelerator System Simulation (DSS), require accurate initial conditions (ICs) drawn from a separation model. Automatic Dynamic Analysis of Mechanical Systems (ADAMS), a Commercial off the Shelf (COTS) tool, was employed to provide insight into the multi-body six degree of freedom (DOF) interaction between parachute test hardware and external and internal forces. Components of the model include a composite extraction parachute, primary vehicle (PTV or PCDTV), platform cradle, a release mechanism, aircraft ramp, and a programmer parachute with attach points. Independent aerodynamic forces were applied to the mated test vehicle/platform cradle and the separated test vehicle and platform cradle. The aero coefficients were determined from real time lookup tables which were functions of both angle of attack (α) and sideslip (β). The atmospheric properties were also determined from a real time lookup table characteristic of the Yuma Proving Grounds (YPG) atmosphere relative to the planned test month. Representative geometries were constructed in ADAMS with measured mass properties generated for each independent vehicle. Derived smart separation parameters were included in ADAMS as sensors with defined pitch and pitch rate criteria used to refine inputs to analogous avionics systems for optimal separation conditions. Key design variables were dispersed in a Monte Carlo analysis to provide the maximum expected range of the state variables at programmer deployment to be used as ICs in DSS. Extensive comparisons were made with Decelerator System Simulation Application (DSSA) to validate the mated portion of the ADAMS extraction trajectory. Results of the comparisons improved the fidelity of ADAMS with a ramp pitch profile update from DSSA. Post-test reconstructions resulted in improvements to extraction parachute drag area knock-down factors, extraction line modeling, and the inclusion of ball-to-socket attachments used as a release mechanism on the PTV. Modeling of two Extraction parachutes was based on United States Air Force (USAF) tow test data and integrated into ADAMS for nominal and Monte Carlo trajectory assessments. Video overlay of ADAMS animations and actual C-12 chase plane test videos supported analysis and observation efforts of extraction and separation events. The COTS ADAMS simulation has been integrated with NASA based simulations to provide complete end to end trajectories with a focus on the extraction, separation, and programmer deployment sequence. The flexibility of modifying ADAMS inputs has proven useful for sensitivity studies and extraction/separation modeling efforts. 1 Aerospace Engineer, Aeroscience and Flight Mechanics Section, 2224 Bay Area Blvd, Houston, TX, AIAA Member. 2 Aerospace Engineer, Advanced Engineering, P.O. Box 707, Brigham City, UT. 3 Aerospace Engineer, Aeroscience and Flight Mechanics Division, Houston, TX. 1 American Institute of Aeronautics and Astronautics Nomenclature A = Airborne A = Reference Area ref AESM = ADAMS Extraction and Separation Model AGL = Above Ground Level C = Force aerodynamic coefficient associated with the direction x C Moment aerodynamic coefficient associated with the direction xx = CDT = Cluster Development Test cm = Center of Mass cp = Center of Pressure Downdraft = Design variable with range of 2500±2000 lbs; dispersed during a Monte Carlo analysis. EDU = Engineering Development Unit Fa = Axial Drag Fn = Normal, up and down Fy = Side to side IMU = Inertial Measurement Unit L = Reference Length ref Mll = Moment about the roll axis Mm = Moment about the pitch axis Mln = Moment about the yaw axis (cid:1869)(cid:3364) = Dynamic Pressure Rho = Atmospheric density. (cid:1829) (cid:1845) (cid:4666)(cid:1872)(cid:4667) = Extraction parachute apparent drag area as a function of time. (cid:3005) (cid:3045)(cid:3032)(cid:3033) V = Velocity of the extraction parachute. ex Vm = Magnitude of the extraction parachute velocity V = Velocity of the mated CPSS/PTV cg. mated Vz = z-component of the extraction parachute cm velocity vector I. Introduction U NDERSTANDING the physics, dynamics and force interactions during extraction and separation events between an aircraft, test article, and platform are vital to the success of a Capsule Parachute Assembly System (CPAS) test campaign. Aerodynamics, atmosphere, vehicle configuration, mass properties, and the force interactions between these objects are key components that must be considered when designing for a drop test. Optimizing test vehicle mass properties to support end-to-end flight stability is critical, but a vehicle separation solution is necessary to initiate a favorable deployment sequence. Without a good separation solution, the CPAS drop test campaign would be at increased risk of loss-of-test-vehicle (LOTV) malfunctions during deployment. The first test of a fully integrated CPAS system, Cluster Development Test 2 (CDT-2), occurred during the Generation I testing phase. The configuration consisted of a capsule shaped vehicle called a Parachute Test Vehicle (PTV) mated to a Cradle and Platform Separation System (CPSS). The approximately 30,000 lbs mated vehicle was extracted using a C-17 aircraft at an altitude of 25,000 ft. The CDT-2 test resulted in a LOTV due to a programmer deployment malfunction that occurred during the initial stages of flight after the PTV/CPSS separation event1. Valuable test experience was gained from this effort that is implemented in subsequent flights. To provide insight to the force interactions of test articles during the intricate extraction and separation flight phases, the development of the Commercial off the Shelf (COTS) tool Automatic Dynamic Analysis of Mechanical Systems (ADAMS) was employed. Efforts were invested to model the extraction-separation event of CPAS test articles based on physics principles versus qualitative solutions or assumptions. Since the test execution of CDT-2, three successful Engineering Development Unit (EDU) capsule tests have been performed. EDU-A-CDT-3-3 was the first baseline capsule test after the CDT-2 mishap. The ADAMS Extraction and Separation Model (AESM) was used as the primary tool to provide nominal and Monte Carlo trajectories from aircraft extraction to PTV/CPSS separation through PTV programmer deployment. The state vector of the PTV in the AESM was delivered to Decelerator System Simulation (DSS) as initial conditions to provide predictions through vehicle touchdown. The preflight predictions were comparable to the flight test data. Post-test data reconstructions identified a delay in the cut command signal when the smart separation parameters 2 American Institute of Aeronautics and Astronautics were senseed by the avionnics system. NNew release meechanism geommetry was addeed to the AESMM to account ffor the ball and soocket interface interaction prioor to a full PTVV/CPSS separaation. The AEESM gained inncreased fideliity for the seccond capsule teest, EDU-A-CCDT-3-5. The development of the release meechanism geommetry simulateed the contact forces with higher fidelitty to give a bbetter match of the separation delay seen in fflight. An avioonics system deelay of 90 ms wwas also includded in the smaart separation loogic to provide a rrepresentative cut command signal as seenn in test. The ttest execution of EDU-A-CDDT-3-3 and EDDU-A- CDT-3-5 pproved the PTVV\CPSS contacct forces during separation coonsistently oriented the PTVV heat shield foorward at the releaase conditions ddespite the oppposing aerodynnamic momentss that force thee vehicle apex forward. The AAESM simulations predicted this behavior prioor to both testss. Post-Test daata reconstructtions introduceed a downdraft force, due to thee aircraft wakee, on the extraaction parachuutes that resultted in a betterr match to tesst results. Eacch test experienceed presumed ddowndraft forces and downddraft dispersionn were includeed in the Monnte Carlo analyysis to account forr the effects off this new parammeter. A concceptual three diimensional smaart separation wwindow was thhe answer to delivering preflight predictionns with repeatable PTV heat shieeld forward atttitudes. The tthree dimensioons used to boound favorable PTV attitudess were time, pitchh, and pitch rate. Each dimension wwas defined wwith an open and closed window criteeria or minimum/mmaximum valuues. The intrroduction of aan avionics tiime delay, rellease mechanism geometry and a downdraft force variable on the extraction parachutess lead to a famiily of solutionss triggering at different sides of the defined smmart separation window. Earrlier preflight ppredictions onlyy showed casees triggering att the maximumm pitch rate window. Test resultts showed the mmated vehicle ssensing smart sseparation paraameter at intennded rates, but due to the time ddelay separatedd at lower ratees than predictted. The incluusion of thesee parameters too the AESM pproved successful during the thiird capsule test, EDU-A-CDT-3-7, which eextracted fromm a C-17 and sseparated heat shield forward. II. Validating tthe AESM wwith DSSA Prior to integrating the ADAMS Exxtraction-Separration Model ((AESM), into the CPAS maainstream simuulation sequence, extensive trajeectory compariisons were madde using Decellerator Simulaation System AApplication (DSSSA)2. The pallet extraction simmulation, DSSAA, is a FORTRRAN legacy simmulation tool validated and used extensiveely for deelivering prefllight test prediictions for CPPAS in adddition to the earlier NASAA X-38 test proogram. DDSSA is a varriant of the DDSS tool with added paayload extractiion capabilitiess. DSSA modeels two siix degree-of--freedom (DOOF) masses with Figure 2: PPTV/CPSS att Extraction Dynamic Pressuree Chute load separate aeerodynamic chharacteristics, coupled by ann elastic 800 30 ATK Adams riser – onne body beingg the parachutte, and the oother the 700 DSSA 25 20 pduayrilnoga de.x tCCrraacrtriioenr aanidrc triapffot ffr apmitcph mcoonttiaaoocnts awt irtahm tph ec leepaar y(lRoCad) ( psf )560000 ( klb )15 10 is also mmodeled. Thee simulation ccannot track multiple 400 5 independennt bodies afteer the separation event, suuch as a 300 0 0 5 10 0 5 10 PTV/CPSSS would demonnstrate in flightt. Thus, it is only used ALPHA (ASA aerro) Y-Body rate 100 60 for test coonfigurations tthat include a single non-seeparating 00 40 vehicle coonfiguration suuch as a weighht tub on a pallet for -100 20 ChaPnAd-So, faf ncddo/onrd iat iConres ww Rereeeet ugrenn Vereahteicdl eu (ssCinRgV t)h efo mr Xat-e3dd88 .p oErtairolny ( deg )--320000 ( deg/s )--42000 of flight att selected times with favorabble pitch and ppitch rate -400 -60 -500 -80 dynamics. The crude separation haand-off conditiion was 0 5 10 0 5 10 time (s) time (s) assumed to be instantanneous. Due tto DSSAs proven test Figuure 1: AESM and DSSA Vaalidation Resuults heritage aand fidelity it was used too verify and validate preliminaryy trajectories pproduced by thee AESM. The anallysis in this secction was perfoormed prior to the EDU-A-CDT-3-3 test exxecution. The iintent of this annalysis was to commpare AESM results to the mated portionn of the DSSAA pallet extracction simulation3. The goal wwas to verify the AAESM by commparison to a siimilar yet indeppendent simulaation. The simuulation intervaal was taken froom the start of thee pallet (CPSS) extraction too the PTV test article separattion trigger. Thhe two simulattions compared very well and vverification waas successful. Figure 2 showws a diagram of the test arrticle and a snnapshot of an actual 3 Americcan Institute off Aeronautics aand Astronautiics extraction from CDT-2. Figure 1 is a sseries of plots oof simulation ooutput from AEESM and DSSAA. The green DSSA plot trace eends at the plannned PTV sepaaration time. Thhe AESM plott trace follows a hypotheticall trajectory wheere the system remmains mated past the sepaaration time. Visually, the two predictioons are very similar. Succcessful comparisonns with DSS oof the PTV in ffree flight afterr separation wwere also perforrmed. Additionnal comparisonns will be performmed as the test pprogram continnues. IIII. AESM MModeling Prinnciples ADAMS iis a multi-boddy dynamic annd motion anaalysis softwaree tool. ADAAMS helps enggineers to studdy the dynamics oof moving partts, how loads aand forces are distributed throoughout mechanical systemss, and to improove the performancce of their sysstems. The simmulation capabiility was first introduced on the ARES prooject. It was uused to successfullly simulate thee extraction evvents for the AARES Parachuutte Drop Test4.. The ARES mmodel simulatted the extraction, reorientation and release off the Jumbo DDrop Test Vehhicle (JDTV) ffrom the aircraaft and cradle. This modeling effort saved tthe program mmoney by elimminating the nneed to redessign the reorieentation and rrelease sequence. The next application of the simulation waas on CPAS. AADAMS was employed andd tailored to mmodel a Parachute Compartment Drop Test VVehicle (PCDTTV). Results from this effoort gave the CCPAS analysiss team confidencee to apply the cutting edge modeling to aa PTV configuuration. The PTV test configguration presented a challenge tthat was less ddynamically foorgiving compared to a PCDDTV. Anotherr challenge posed with testinng and modeling aa PTV is the faact that the cennter of mass (cmm) and center of pressure (cpp) are in close proximity commpared to a dart shhaped vehicle. The ARES JDTV and CPAS PCDTV are bboth aerodynammically stable wwhen deceleratting on its weighteed nose. The PPTV simulation objectives aare: 1) Determiine a point in time to releasse the PTV froom the CPSS, 2) DDetermine the interactive loads between bodies during exxtraction/separaation, 3) Providde state vector initial conditions to the DSS moodel, and 4) Ennsure a predictaable and repeattable separatioon between the PTV and CPSSS. A. Modeliing Componennts Thhe main compponents of thee AESM are ddefined by sixx rigid bodies as depicted inn Figure 3. Thhe interactionn and motion of the partss are governed by constrraints, enforcced motions, external forcces, contact forces betweeen bodies andd gravity. Eveery AESM partt has six degreees-of- freedoom (DOF) unttil the modelerr enforces connstraints betweeen the bodies. Initially the PTV, CCPSS, Programmmer, Prograammer Conflluence and rammp are all constrained relativee to each otherr. The motioon of the aircraaft ramp is connsidered an inpput parameter to this model. The motioon statement ddefines the mootion of the aaircraft ramp relative to theground for all 6 DOF, i.e., thhe airspeed andd ramp Figure 3: ADAMS Modeeling Componennts attitudee angles. The simmulation beginns as two indeppendent bodiess, the extractioon parachute annd aircraft rammp. At initializzation, the aircrafftt ramp encomppasses all the mmass propertiees associated wwith the matedd vehicle. Thiss occurs becauuse the componentts are constraiined relative too one another.. The CPSS iss fixed to the aircraft ramp and the PTV to the CPSS. AAs the time-varrying simulatioon executes, ppredetermined conditions aree met, constraiints are releaseed and more relatiive motion bettween the bodiies can take plaace. For exammple, the CPSSS is fixed in alll 6 DOF to thee ramp until a ½ gg load is appliedd to the CPSS from the extraaction line. Afftter the ½ g loadd is sensed, thee CPSS is allowwed to translate 3224 IN along the rail restraintss which are insstalled along thhe length of thee aircraft cargoo floor. As the mated vehicle reaaches the rampp, the pitch connstraint is lifteed and pitchingg motions can take place whhile gravity hollds the CPSS to tthe ramp. At the end of raamp, the CPSSS becomes an independent 6 DOF body free falling unnder a composite extraction parrachute. The PPTV remains ffiixed to the CPPSS until the coonditions of thhe “Smart Sepaaration Window” aare met and sepparation of thee test vehicles ttakes place. Affter separationn, the PTV becoomes an indepeendent 6 DOF boddy. The Programmer Parachhute remains fixxed to the PTVV until the disttance between the Programmmer and the CPSS aattach point is ggreater than the length of thee static line, 31 ft. After the length of the static line is exceeded, the Programmer is its owwn 6 DOF boddy. The Progrrammer Confluuence remains fixed to the PProgrammer unntil the distance beetween the Programmer Conffluence and thee PTV attach ppoints exceeds tthe length of thhe harness legss. 4 Americcan Institute off Aeronautics aand Astronautiics B. User DDefined Forcess There are three typees of user deffined forces aapplied in the AESM. Theey include a ddrag force froom the programmeer and extractiion parachute, an external foorce such as aeerodynamic forrces, and an innternal force suuch as contact forrces between boodies. The following forces have been defiined in the AESM. 1. Extracttion and Progrrammer Parachhute Drag Forcce Modeling Inflatioon modeling off dual extractioon parachutes inn the AESM iss represented bby using a compposite parachuute that has an equuivalent drag foorce. The AESSM starts whenn the extractionn parachute baag is pulled into the airstreamm from the aft end of the aircraft. The cluster of two extractioon parachutes aare assumed to inflate to full open and a draag area degradation factor is applied to accountt for any perforrmance loss duue to cluster intteractions and aircraft wake eeffects during the extraction phaase. The compposite drag forcce is applied too the extractionn parachute ceenter of pressurre (cp) and has thrree translationaal componentss that act at thee cp in the direection opposingg the velocity vector. The ggeneral equation of the three commponent extracttion parachute drag force vecctor model is: (cid:1832)(cid:1832) (cid:3404)(cid:1869)(cid:3364)(cid:1829) (cid:1845) (cid:4666)(cid:1872)(cid:1872)(cid:4667)(cid:3404)1(cid:3415) (cid:2025)(cid:1848)(cid:2870)(cid:1829)(cid:1829) (cid:1845) (cid:4666)(cid:1872)(cid:4667) (1) (cid:3005) (cid:3045)(cid:3032)(cid:3033) 2 (cid:3005) (cid:3045)(cid:3032)(cid:3033) The prrogrammer parrachute inflatiion model diffffers from the extraction paarachute drag force vector model because a time varying inflation currve based on empirical inflation characteristiics is applied.. Currently, the AESM iss only executeed and validated thhrough the proggrammer deplooyment line strretch event, Figgure 4. The programmmer phase oof flight is in tthe early deveelopment phasees and still needs tto be anchoredd to test data bbefore it is offficially introduuced in the CPAS mainstream siimulation seqquuence. The capability to model independentt extraction paarachutes usinng a time varyying inflation curve similar to tthe programmmer phase is beeing integratedd for future eefforts. This effortt would aid iin deriving innflation parammeters for extrraction Figure 4: Programmer Deployment parachutes Line Stretcch Event 2. Aerodynnamic Forces Aerodyynamic forces are applied thrroughout the dduration of thee simulation onn the mated PTTV/CPSS, the CPSS alone and PTV alone. TThere are a totaal of six aeroddynamic forcess and momentss acting on thee mated vehiclee from the end off ramp to the ttime of separaation of the PTTV from the CCPSS. After sseparation, the mated aerodyynamic forces are deactivated aand the CPSS alone and PTTV alone aeroodynamic forcees act on eachh independentt body respectivelly. The generaal equation of thhe aerodynamiic force is: (cid:1832) (cid:3404)(cid:1869)(cid:3364)(cid:1829) (cid:1827) (cid:3404)1(cid:3415) (cid:2025)(cid:1848)(cid:1848)(cid:2870)(cid:1829) (cid:4666)(cid:2009),(cid:2010)(cid:4667)(cid:1827) (2) (cid:3051)(cid:3051) (cid:3045)(cid:3032)(cid:3033) 2 (cid:3051) (cid:3045)(cid:3032)(cid:3032)(cid:3033) (cid:1839)(cid:3404)(cid:1869)(cid:3364)(cid:1829) (cid:1827) (cid:1838) (cid:3404)1(cid:3415) (cid:2025)(cid:1848)(cid:2870)(cid:1829)(cid:1829) (cid:4666)(cid:2009),(cid:2010)(cid:4667)(cid:1827) (cid:1838) (3) (cid:3051)(cid:3051) (cid:3045)(cid:3032)(cid:3033) (cid:3045)(cid:3045)(cid:3032)(cid:3033) 2 (cid:3051)(cid:3051) (cid:3045)(cid:3032)(cid:3033) (cid:3045)(cid:3032)(cid:3033) The maated aerodynammic forces lineearly increase from zero at raamp clear to ffull affect 0.755 s-RC later. TThis is intended too account for thhe wake effectss of the aircrafft. 3. Internaal Forces Internaal forces defineed in the AESMM fall into two main categories,, tension repressenting lines orr slings and coontact forces betwween solid boddies. The lines and slings include: 1) Extractiion Line. Thhis line runs frrom the Extraaction Parachute through the sinngle attach poiint on the CPSSS, 2) Four Harnness Sling Legss between the PTV attach ppoints Figuure 5: PTV/CCPSS Contact Interference aand Aeroodynamic Forrces and the Prrogrammer Coonfluence, 3) Programmer Static Line. This line runs betwween the singlee attach point on the CPSS to thhe Programmeer Bag, and 4)) Programmer Riser. This line rruns between the Programmmer Canopy annd the Programmeer Confluence. Internall contact forcees are defined bbetween: 1) AAircraft Ramp and the bottom off the CPSS platform, 2) PTV and CPSS, and 3) three ball and socket intterfaces betweeen the Figurre 6: PTV/CPPSS Contact Innterference PTV and the CPSS wwhich are part of the rrelease duringg Separation mechanismm, Figure 5. Thhe red ball is a rigid member of the 5 Americcan Institute off Aeronautics aand Astronautiics PTV geommetry and aquaa socket is a riigid member oof the CPSS. The most freqquent contact bbetween the veehicles occurs betwween the PTV heat-shield annd the front bummpers of the CCPSS right after separation as shownn in Figure 6. C. Atmosphere and Aero Coefficientts The attmospheric deensity is requiired in the caalculation useed for all aerodynammic and parachuute drag forcess. This is an innput into the AAESM via an importeed X-Y data taable with the ddensity (slugs/ft3) as a functiion of the altitude (ftt-MSL). For eeach drop test,, this table hass been updatedd with the local Yuma Proving Grounds (YPG) attmosphere for the month wheen the test is held. The sixx aerodynamic coefficients, FFigure 7, are taable look-ups hhandled in a similar mmanner except they are a funnction of two iindependent vaariables, α and β, wheere α is the anggle of attack annd β the side-sslip angle. Theere are 18 aerodynammic effects conssidered in the ssimulation, sixx forces and mooments on the mated PPTV/CPSS, CPPSS alone and PTV alone. D. Sensorrs and Logic SScripts AESMis run from a script, or a sequence of commmands that alllows the Fiigure 7: Mateed Aerodynammic program tto sense “triggger” conditioons, pause annd activate/deeactivate Definitions annd Coordinatee constraintss or external foorces and thenn continue withh the simulatioon. The Systtem5 script runss the simulation from the timme the extraction parachute iis pulled into the airrstream until thhe PTV has fallen under the PProgrammer paarachute 14 secconds. Sensors are set up to define the differennt triggers andd store a “state”” such as a timme when prograammer line streetch occurs so that this time ccan be referenced later as a consstant in the programmer inflattion curve equaations. E. Extracction Parachutte Center of PPressure The exxtraction parachhute is assumeed to be a rigidd body from thhe confluence aat the bottom oof all the suspension lines up around the surfaace of the canoppy. The center of mass (cm)) is at the centeer of the circlee forming the ccanopy skirt. The center of presssure (cp) is 5 inches aft of tthe center of mmass. By sepaarating the disttance between the cp and cm thee extraction parrachute would behave similarr to that observved from videoo. F. SSmart Separattion Conditionns TThe main objecctive for adoptiing the AESMM was to dettermine the optimal condditions of releaasing the PTVV from the CCPSS. The geeneral motion of the mated PTV/CPSS after it exits the aircraaft is similar to that of a ddouble Figure 8: Motion of thee PTV/CPSS aafter Exit frommthe Aircraft penddulum. The PTV/CPSS sswings abouut a point nnear the centter of pressure foor the extraction parachute annd at the attach point of the exxtraction line aand CPSS as deepicted on Figuure 8. The timming for the rellease of the PTTV from the CPPSS is determiined such that: 1) The PTV hheat shield willl rotate into the wiind. This will allow for a cleean deploymennt of the paracchutes and reduuce the risk off the PTV tanglling in the parachute lines, 2) thhe rotation ratee of the PTV Taable 1: Smartt Separation PParameters at release iis low enough so the PTV wiill not tumble Paarameter  Inputs  out of conttrol before the programmer pparachute has Time (sec)  1.0 < time < 4.00 time to depploy and inflaate, and 3) therre will be no unplanned impact betweeen the PTV aand CPSS to Pitch Angle (⁰)  -25o ≤ θ ≤ -3o enable the rest of the testt to flow smootthly. Pitch Rate (⁰/sec)  100o/s ≤ (cid:2016)(cid:4662) ≤ 25o//s An IMMU on board the CPSS mmeasures the acceleratioon, angular possitions and angular rotation rates. The coonditions referrenced to deteermine the timming of release aree known as thee, “Smart Separration Windoww (SSW).” Thhey included coonditions for thhe pitch attitudde, the pitch rate and time fromm ramp clear. The time vaariable was puut into the SSWW parameters to ensure suffficient clearance ffrom the aircraaft and, in the eevent of an IMUU failure, releaase at the end oof the time winndow to give thhe test 6 Americcan Institute off Aeronautics aand Astronautiics a chance tto proceed andd be successfuul. The AESMM simulation hhelped us deterrmine the conditions of the Smart Separationn Window as shhown in Error! Reference soource not founnd.. When thhe separation oof the PTV froom the CPSS takes place thhe following ssequence of evvents and interractive forces occuur. The aerodyynamic forces aacting on the teest articles channge from the mmated propertiees to the indepeendent vehicle prooperties. The once mated veehicle becomess independent 6 DOF objectts, Figure 9. AAs the mated vvehicle Figure 9: Interactive FForces Prior too Separation Figure 110: Interactivve Forces during Release separates, a level of conttact is maintainned between thhe surfaces of tthe release meechanism in thee form of a balll-and- socket joinnt. The PTV bball slides awayy from the soccket on the CPSSS and maintaains about the ssame pitch andd pitch rate attitudde as the CPSS,, Figure 10(a). Shortly aafter the ball aand socket cleaar each other, tthe front bumpper slide along the heat shieldd of the PTV, FFigure 10(b). Thee interaction annd aerodynamic effects induuce a moment aabout the PTVV cm to keep thhe PTV rotatinng in a heat shieldd first attitude. G. Prograammer Parachhute Implementation The Proggrammer Parachute is storedd on the side oof the PTV unttil the distancee between the attach points on the CPSS and Programmer ddeployment bagg exceeds the llength of the sstatic line connnecting the twoo markers togetther as seen in Figgure 11(a). Affter the 31 ft sttatic line lengthh has been excceeded, the Proogrammer becoomes its own 66 DOF body. Witthin the Prograammer deploymment bag are twwo masses, onne weighing 51.6 lbs represennting the mass of the canopy and risers, the seecond mass (222 lbs) is the Prograammer confluuence which joins the four harneess legs to the single riser. The four harness leegs are tensionn only springgs which connect thhe confluence tto the attach ppoints on the PTV. Two of the hharness legs aare 198.8 inches in length and two are 189.33 inches. When one or more of thhe harness leggs exceed its free length, the confluuence is pulled from the deploymennt bag and becomes an independent body, Figuure 11(b). Thhe static line attached between thhe CPSS and programmer ccontinues to pull on the Programmmer until the llength of the Prograammer riser ((48.8 ft) is ddeployed. This evennt is known aas line stretchh Figure Figure 111: Programmeer Inflation Seequence 11(c). Sennsors within the AESM store the time of line streetch and velociity of the progrrammer at line stretch to be uused for calcullations of the pprogrammer inflation curve in a similar mannerr as described iin the section ffor extraction pparachute inflaation. IVV. Data Recconstruction Results Post testt data reconstrructions of the extraction andd separation evvents were perrformed for EDU-A-CDT-3--3 and EDU-A-CDDT-3-5. Each test was uniquue and did not provide the saame separation solution. Therre were deviatiions in the extraction parachute pperformance, fforce interactioons, system eveent cut triggerss, vehicle attituude and day off flight winds whicch enabled varrying downdrafft forces. 7 Americcan Institute off Aeronautics aand Astronautiics A. Avioniics Time Delayy A timee delay or timee lag observed in measured fflight test dataa was evidentlyy missing commpared to the AAESM ppreflight prediictions. The on-board CPSSS avionics syystem includess IMU ssensors to meaasure time, pitcch, pitch rate pparameters andd are activated with a rrigged pin-pulll mechanism att the ramp cleaar event as seenn in Figure 12. Post- ttest reconstructtion comparisoons identified tthat the avioniccs system senssed the PPTV/CPSS sepparation criteriia on a delay ddue to processinng and executiing the ccut command tto the pyrotechhnic cutter sysstem. Anothher contributor to the oobserved avioonics time dellay was attribbuted to the release mechhanism cconfiguration shown in Figuure 13 that is used to securre the PTV onnto the Figure12: Avionics System CCPSS. Activatted by Pin-Pull at RC TThe release meechanism is coomprised of a bball and ssocket geomettry which is iinstalled on thhe PTV and CPSSS, respectively. The ball aand socket intterfaces requirre additional time to overcome internal interfference prior too complete sepparation. A rresult of the oobserved time delayys was the impplementation oof an AESM reelease mechannism geometry joining the PTV/CCPSS vehicles aand allowing bball/socket inteerferences to occcur prior to aa cleared separation.. To account ffor the known avionics time lag a conservaative 90 ms timme delay was includded in the smmart separationn logic to alllow for proceessing time onnce the Fiigure 13: Release conditions were satisfied and executionn of the cut commmand signal. Mechanismm B. Extracction Parachutte Inflation MModeling A clustter of two 28 ft Extraction PParachutes aree simulated as a composite pparachute. Thhe initial assummption madee for modelingg the drag forrce produced bby dual extracction parachutees was calcuulated withouut consideratioon of clusteer efficiency. However, data reconnstruction resuults confirmed that the extracction parachuttes cluster efficiency was reduced 15% in a cluster oof two. The magnitude of the drag forcce was reducced due to cluuster parachutee interactions dduring the inflation process. Each extraaction parachuute did not infflate to full oppen during thhe chaotic extrraction phasee due to interrference and ccrowding relattive to one annother. This ccluster interaaction can be sseen in Figure 14. Other facctors that contrribute to the reeduced Figure 144: Reduced CCluster clustter efficiency are the aircraft wwake, which iss currently not directly quantiified. Efficiency C. Extracction Force Line of Action The exxtraction forcce line of action (EFLA) is the tenssion that is exerted through the simulated risers and suspension linnes of the Extraction Parachutes about the attach poinnt on the mateed vehicle. The pre-ttest orientatioon of the EFLA waas set horizonntal to the Figure 155: PTV Attituude Dynamicss ground at extraction. Uppon further examinatiion of the model behavior, it was determineed there is aa direct correelation between tthe EFLA and the amplitude of the initial osccillation. The downward pitcch rate of the paallet at the rammps edge wass over predicted by a magnitude of 30% andd 60% for EDU--A-CDT-3-3 annd EDU-A-CDDT-3-5 respectiveely as shown inn Figure 15. Figure 16: Downdraft FForce Variatioon The post-test video vverified the EDDU-A- CDT-3-5 EFLA appearred oriented lowwer in 8 Americcan Institute off Aeronautics aand Astronautiics relation wwith the ramp plane compared to EDU-AA-CDT-3-3, Figgure 16. In order to vary the position of the extraction parachutes a nnew variable neeeded to be inttroduced in thee AESM and wwas called the ddowndraft forcce. As the extracttion parachute travels in the wwake of the aircraft, there is an airflow thaat streamlines over the aircraaft and converges behind the aiircraft in the rregion of the extraction paraachute, drivingg the extractioon parachute ffurther downward than the no wwake airflow. TThe downdraft force is used too control the pposition of the extraction parachute relative to the ramp planne. It is applied in the simulaation as a timee step functionn that is activatted 1.2 secondds after the first mmotion of the mmated vehicle oon the aircraft ramp. As thee PTV/CPSS iss extracted, the downdraft foorce is applied forr a one second duration either pushing the eextraction paraachutes below, in-line, or sligghtly above thee ramp plane. A nnominal peak vvalue for the coomposite drag force is about 39,000 lbs. Thhe equation forr the z-componnent of the extractiion parachute ddrag force is (cid:3398)(cid:1874) (cid:1832) (cid:3404)(cid:1869)(cid:3364)(cid:1829)(cid:3005)(cid:1845)(cid:3045)(cid:3032)(cid:3033)(cid:4666)(cid:1872)(cid:4667)(cid:3435) (cid:3053)(cid:3415)(cid:1874)(cid:3040)(cid:3439)(cid:3439)(cid:3397)(cid:1856)(cid:1867)(cid:1875)(cid:1866)(cid:1856)(cid:1870)(cid:1853)(cid:1858)(cid:1858)(cid:1872) (5) The doowndraft force seen in the twwo flight tests wwas 800 lbs annd 4100 lbs, reespectively witth the nominall input value set to 2500 lbs. EEach test had ddifferent extracction parachutee orientations. The orientation for each teest was positioned either over annd under or side by side as seeen in Figure 16. The 4100 llbs downdraft force used for EDU- A-CDT-3-5 produced a good match too the measuredd test data as sshown in Erroor! Referencee source not foound.. As the dowwndraft force inncreases, the mmagnitude of thhe pitch and pittch rate will deecrease. For EEDU-A-CDT-3-3, the magnitudeof the pitch raate was differennt during the innitial descent oof the payload, but the overaall timing of the PTV release hass proven to be eexceptional. D. Vehiclee Dynamics Predictinng vehicle dynnamics and thee time sequennce of the initiial phases of fflight has evolved from asssuming instantaneoous separation of bodies to mmodeling interrnal force interractions. The understandingg of how to iniitialize the AESMM with represenntative initial cconditions was the first lessoon acquired durring the post ttest reconstructtion of EDU-A-CDDT-3-3. Data comparisons showed that thhe preflight preediction reached the end of the ramp fasteer than the actual mmeasured true airspeeds. Thee slower flightt test extractionn of the PTV/CCPSS was a connsequence of ccontact forces expperienced betwween the CPSSS pallet and C--17 ramp. Thhe initial AESMM assumed noo frictional loss and resulted in a faster eextraction sequuence. The aaddition of fricctional forces between the palllet and ramp immproved the exxtraction timinng. V. Monte Carrlo Analysis A Moonte Carlo (MMC) capabilityy was develooped to analyzze the unknowwn trajectory tyypes that couldd be triggered with a unique set of parametters. To initiallize the MC aannalysis a repreesentative disppersion band for the aircraft rramp pitch annd pitch rate wwas required. These parametters were definned using a limmited test data sset acquired frrom an on-boardd C-130A andd C-17 aircrafft sensor tray. C-130A datta was includedd due to limitted availabilityy of C-17 onnly data. An initial Figuree 17: Vehicle Acceleration conservatiive dispersion dderived from tthe diverse dataa points resulteed in a [0˚-10˚] piitch and [0˚/s-10˚/s] pitch raate band. The dispersion bannd was intended too cover all thee C-130A andd C-17 data exxperience. Thhis proved to bbe too conservvative and a ssecond approach wwas implemennted to averagge the aircraft flight test pittch and pitch rate values aat ramp clear versus bounding extreme attituudes and rates. This approoach deliveredd a more repreesentative disppersion band of the dynamics tthat could be sseen when extracting from a CC-17. The folllowing criteriaa were used to initialize the aaircraft in the AESSM at ramp cllear. Post testt analysis demmonstrated that the derived ddispersion bandd captured whaat was experienceed in test by thee C-17. The acctual C-17 aircraft pitch and ppitch rate valuees seen during EDU-A-CDT--3-5 at ramp clearr were 3.5˚ annd 2.0˚/sec, resspectively. All other mass property inerttias were dispeersed ±10% annd the altitude waas set to ±500 fft. The MCC capability pprovided insigght into Table 22: Ramp Initiialization Paraameters trajectory variations andd introduced a unique solution thhat required saafeguarding to reduce C‐117 Aircraft Inittialization Paraameter  Minnimum  Maxiimum  malfunctioon risks during the Ramp Pittch Angle (⁰)  00.0  55.0  extraction//separation phhases of flightt. All Ramp Pitcch Rate (⁰/sec) 00.0  5..33  preflight MMC predictionns prior to EEDU-A- CDT-3-7 oonly entered thhe Smart Separration Windoww (SSW) from the maximumm pitch rate sidde. A MC cycle was consideredd a satisfactoryy solution if thhe PTV pitch rrate at separattion was abovee 0˚/sec and hhad a α ≥ 95˚. This 9 Americcan Institute off Aeronautics aand Astronautiics requiremennt helped oppoose any apex ffoorward attitudee at separationn and allowed adequate time for the Prograammer parachute tto recover conttrol authority oof the vehicle aand re-orient heeat shield forwward. The introoduction of thee down draft forcee variable traansformed howw AESM trajjectories enterred the SSWW and aided iin identifying three classificatiions of trajectoories. Each traajectory class iis defined by tthe SSW planee it enters: MMaximum Pitchh Rate (MaxPR), Minimum Pitch Rate (MinPR), annd Minimum Pitch Angle (MinPA). TThe AESM haas progressiveely been anchored to testt data and haas supported analysis effoorts to disclosse the physicaal conditions that producee unfavorablee attitudes annd identify the possibilityy of trajectories separatinng outside thee defined SSW using MC anaalysis. A new trend of Monte Carlo trrajectories waas observed wwith the implementation off the downdraaft force and aavionics time ddelay. Trajectoories in the righht conditions now could pottentially enter the non-optimaal SSW fromm the MinPR annd MinPA winddow. The noveel trajectory class that enteered the MinPPR window waas attributed with having downdraft foorces exceedinng 4300 lbs coupled with raamp pitch and ramp pitch ratte inputs beloow 2˚ and 1˚/ssec, respectiveely. One of thhe ramp initialization parammeters usuallyy had an inpuut FFigure 18: Trrajectory Classsifications below 1, ssuch as a 0.7˚/sec pitch ratte or 0.4˚ pitcch. Under thesse conditions, iif the MinPR wwindow was seet too low, succh as at 0˚/sec,, the PTV wouuld tend to acceelerate apex forwaard at separatioon with high riisks of snagginng lines and tuumbling withouut recovery. AAn example is shown by the oraange case showwn in Figure 19. Extreme ddowndraft forcees also producced cases withh flat or translaational separationss. This scenariio orients the PPTV as it sits oon the CPSS wwith an α=90˚ aand slides off wwith small ratess, as if apparently motionless, unntil the force oof the inflating programmer pparachute is appplied to the PTTV at the four attach points. The secoond novel trajeectory class whhich entered thhe MinPA winndow, like the MinPR class, had large dowwndraft forces assoociated with itss inputs. The rramp initializaation inputs weere contributorss to entering aa different side of the window. Thhe MinPA rampp inputs were bboth typically ggreater than 2˚ annd 2˚/sec, reespectively. The larger ramp initialization values affecteed the oscillatioon magnitudes of the pitch and piitch rate at exxtraction, but still producedd heat shield forwaard PTV attituudes with sattisfactory sepaaration results. Caases that enteered the MinPPA side weree also susceptible to triggering within the wwindow limitts and executing thhe separation ccut command outside the ddefined SSW. Thiss behavior wwas due to thhe systems poositive acceleration and implemenntation of the 990 ms avionic ssystem time lag. TThe SSW connditions were being met prrior to reaching thee systems peak acceleratioon and as a result separated annd acceleratedd outside thee defined wiindow. Conversely, trajectories enntering the MaaxPR window satisfy Figuree 19: Trajectoories Entering the Smart the SSW coonditions whenn the system hhas already beggun to Separatiion Window decelerate uunder the extraaction parachuutes on the neegative slope of thhe accelerationn curve. An uunderstanding of the systemms trigger and release cut coommand timinng was acquired aand implementted in the AEESM. Acquirinng an insight tto the systemms timing sequuence eliminateed the assumptionn of instantaneeous separationn and was repplaced with a sseparation obsserved in post--test reconstrucctions. The SSW event was noww split into twoo visible featurres, a trigger evvent in which tthe conditions are sensed andd a cut command signal event thhat is executed a delta time affter the systemm senses the defined separatioon conditions, FFigure 18. Earlier AEESM predictioons have modeeled higher piitch rates at seeparation thann actually experienced. Thiis was attributed tto the delays inn the separationn event. From when the condditions are sensed to when thhe actual strap ccutters are activated, a period off time has passsed resulting inn different pitcch angle and piitch rate statess. A typical raate loss for EDU-AA-CDT-3-5 waas approximatelly 10˚/sec. Thhe predicted sepparation pitch rate was 15˚/seec and the actuual test 1 0 Americcan Institute off Aeronautics aand Astronautiics