DYNAMIC SIMULATION OF MARS-03 ENTRY, DESCENT AND LANDING SYSTEM

Size: px

Start display at page:

Download "DYNAMIC SIMULATION OF MARS-03 ENTRY, DESCENT AND LANDING SYSTEM"

Godwin McBride
5 years ago
Views:

DYNAMIC SIMULATION OF MARS-03 ENTRY, DESCENT AND LANDING SYSTEM Chia-Yen Peng and Walter Tsuha Jet Propulsion Laboratory California Institute of Technology 4800 Oak Grove Drive Pasadena, CA 91109

1 DYNAMIC SIMULATION OF MARS-03 ENTRY, DESCENT AND LANDING SYSTEM Chia-Yen Peng and Walter Tsuha Jet Propulsion Laboratory California Institute of Technology 4800 Oak Grove Drive Pasadena, CA ABSTRACT. NASA s Mars Exploration Rover mission will place two landers on the surface of Mars using an innovative approach to entry, descent, and landing. Since it is difficult, expensive, and in some cases impossible to test the entry, descent, and landing system in Mars-like environment, an extensive simulation effort has been undertaken to validate the concept and to predict the performance of the system. A critical concern is the dynamical behavior of the system under winds as it descends through the Martian atmosphere. This paper summarizes multi-body dynamic simulations of the terminal phase of the entry and descent. 1. INTRODUCTION After the loss of Mars Polar lander in 1999, NASA dramatically modified its plans for Mars exploration. The 2001 lander mission was cancelled and will be replaced by a more ambitious attempt to land two large rovers in With far greater mobility than the 1997 Mars Pathfinder rover, the two powerful new Mars rovers will each be able to trek up to 100 meters (about 100 yards) a day across the Martian surface. Each rover will carry a sophisticated set of instruments - the Athena Science Payload - that will allow it to search for evidence of liquid water in the planet's past. On June 4, 2003, the first Mars Exploration Rover (MER) spacecraft, Figure 1, is scheduled for launch on a Delta II launch vehicle from Cape Canaveral, Florida. After a seven months flight, it will enter the Martian atmosphere in January A second lander and rover will follow a short time later. In a spectacular landing similar to that of the Pathfinder spacecraft, a parachute will deploy to slow the MER spacecraft. Then a cocoon of airbags will inflate around the vehicle to cushion the shock of impact. The first bounce on the Martian surface will reach more than one hundred feet into the air. The airbags will bounce about a dozen times, and could roll as far as one kilometer (0.6 miles). When the spacecraft finally comes to a stop, it will turn itself to an upright position so that the rover can deploy properly. The airbags will deflate and retract. Three "petals," like the petals of a flower, will open to reveal a tightly folded rover. Piece by piece the rover will take shape, deploying its camera mast, antennas, wheels, and solar arrays. The landed portion of each mission features a design that is dramatically different from that of Mars Pathfinder. Where Pathfinder had a lander and the small Sojourner rover, each MER spacecraft will carry just a large, long-range rover. The rover has a mass of nearly 180 kilograms (about 380 pounds) and has a range of up to 100 meters (about 100 yards) per Martian day, or sol. Immediately after landing, each rover will begin reconnaissance of its landing site by taking a 360-degree visible color and infrared image panorama. Athena scientists will choose rock and soil targets and command the rovers to begin their exploration, leaving the lander structure behind. When a rover reaches a target, its multi-jointed arm will deploy and the target will be examined with a microscope and two spectrometers. The "RAT" (Rock Abrasion Tool) will be used to expose fresh rock surfaces for study. Images and spectra of interesting rocks and soils will be taken daily. Initial targets may be close to the landing site, but later targets might be far afield. These rovers will be able to travel almost as far in one sol as the Sojourner rover did over its entire lifetime. Surface operations will last for at least 90 sols, extending into the early summer of Cruise Stage Backshell Lander Heatshield Figure 1. Mars Exploration Rover spacecraft. 1123

2 Entry Turn & HRS Freon Venting: E - 1:30:00 Cruise Stage Separation: E - 0:30:00 Entry: E - 0 s, 129 km, 5.4 km/s wrt atmos., γ = inertial, -12 relative Peak Heating / Deceleration: 34 W/cm 2,6.5earthg Parachute Deployment: E s, 10.8 km, 449 m/s wrt atmos. Heatshield Separation: E s Lander Separation: E s Bridle Deployed: E s Radar Ground Acquisition: L - 34 s, 2.4 km above ground Airbag Inflation: ~450 m, L - 10 s Rocket Firing: L - 6 s, 125 m, 76 m/s Bridle Cut: L - 3 s, 16 m Landing: E s Bounces: >15, Rolls Up to 1 km Roll to a Stop: Base Petal Down, Landing + 2 min Deflation / Petal Latch Firing: Landing + 90 min Figure 2. Sequence of events during Mars Exploration Rover entry, descent, and landing. The mission seeks to determine the history of climate and water at a site on Mars where conditions may once have been favorable to life. The landing sites have not yet been chosen, and will be selected on the basis of intensive study of orbital data collected by the Mars Global Surveyor spacecraft and other missions. The sites will be ones at which there is clear evidence that liquid water was once present; possibilities might include former lakebeds or hydrothermal deposits. The rovers' scientific instruments will be used to read the geologic record at each site, to investigate what role water played there, and to determine how suitable the conditions would have been for life. 2. ENTRY, DESCENT, AND LANDING To meet the MER mission requirements, a passive Entry, Descent, and Landing (EDL) approach has, Figure 2, been developed based on proven technologies and the mission experience obtained from the Viking and Mars Pathfinder. The EDL approach for the Mars Exploration Rover mission consists of a series of entry, descent, and landing events, which culminate in the deployment of a large, long-range rover on the surface of Mars. The Mars Exploration Rover spacecraft is composed of a cruise stage and an entry vehicle. The 825 kg entry vehicle consists of a tetrahedral lander, encased by a heatshield and backshell. After the interplanetary cruise, the entry vehicle will separate from the cruise stage. During the 36 minutes from cruise stage separation to the landing, the spacecraft has no active control over its trajectory or its orientation. The onboard computer can only choose the timing of the series of events culminating in surface impact. Spinning at 2 RPM, the computer will detect the beginning of atmospheric entry by monitoring onboard accelerometers. Following peak deceleration, a disk-gap-band parachute will be deployed. Shortly thereafter the heatshield is discarded, exposing the lander and the inside of the backshell. The lander will then be lowered from the backshell on a 20-meter tether. At altitudes below 2.4 km, a radar altimeter in the lander will be used to estimate both altitude and decent rate. Based on this stream of information, the flight computer will determine the critical timing of the final actions. Approximately ten seconds before surface impact, airbags surrounding the lander will be inflated, and then three solid rocket motors in the backshell will be fired to bring the lander to a stop. While the rockets are still firing, the lander will cut itself free from its bridle, and then fall to the surface. The flight computer in the lander decides (before airbag inflation) when it will initiate airbag inflation, rocket firing and bridle cut, but otherwise has no control over its attitude or trajectory. 1124

3 3. CHALLENGES OF TERMINAL DESCENT Early on in the system development, it became clear that the MER upper-bound terminal velocity of 85 m/sec is 32% higher than that of Mars Pathfinder. The major reasons for the terminal velocity increase are: Heavier descent mass; Less dense atmosphere; Higher landing site elevation. As a result, the MER EDL system has to increase the diameter of parachute to reduce its terminal velocity and use a Retrorocket-Assisted Deceleration (RAD) system with more total impulse to decelerate the system vertically. The RAD system consists of three retrorockets mounted inside the backshell. Wind induces three-body pendulum oscillations in the parachute-backshell-lander system that can result in large bridle angles between the vertical and the bridle line at retrorocket firing, as shown in Figure 3. The bridle line is defined as the straight line connecting the center of mass of the backshell with that of the lander. The bridle angle (θ ) is defined as the angle between the vertical and the bridle line. A large bridle angle degrades the vertical deceleration performance. In addition, it induces large horizontal velocity (V h ) during RAD firing. For example, a 20 degrees bridle angle can result in a horizontal velocity of 29 m/sec. Adding the large RAD-induced horizontal velocity to other velocity components can exceed the impact velocity capability of airbag. Lander Parachute Backshell θ V h Bridle Line Figure 3. Horizontal velocity induced by RAD firing. To reduce the RAD-induced horizontal velocity, a new Transverse Impulse Rocket System (TIRS) was incorporated into the RAD system. The combined TIRS/RAD system is called system. TIRS consists of three small rockets (2,000 N thrust each) aimed at the backshell center of mass and spaced 120 o apart. As illustrated in Figure 4, impulsively impart a transverse V to the backshell in order to reduce the average off-nadir angle during RAD firing. To simplify the firing algorithm, a constrained system was developed that constrains both the magnitude and direction of the horizontal velocity correction by firing, at RAD ignition, either a single or two TIRS rockets simultaneously. Bridle Cut V Start of RAD firing V Figure 4. Concept of system. 4. SIMULATIONS OF TERMINAL DESCENT As described above, the dynamics of the multi-body descent system plays a crucial role in the effectiveness of the MER EDL approach. The parachute-backshell-lander system has no control over its pendulum motion, and any deviation of the rockets from vertical will induce lateral velocity and degrade vertical deceleration. Swinging of the lander can also affect the accuracy of the radar altimeter during terminal descent. Multi-body system dynamics during the time just before rocket firing are influenced by a number of factors. Most important are the parachute aerodynamics, lower atmosphere winds, and residual oscillations inherited from the earlier separation events. The heatshield and lander separations are themselves dependent on initial conditions inherited from atmospheric entry and parachute deployment. These unique aspects of the MER EDL approach, and the difficulties associated with testing in Mars-like conditions, made it necessary to conduct an end-to-end simulation, from parachute deployment till ground impact, for verification of the system design concept and prediction of the system performance. The specific objectives of the end-to-end terminal descent simulations are: Verifying the EDL system performance with ; Studying key parameters affecting the effectiveness of the system; Providing key data for designing and testing of the EDL sub-systems. 1125

Figure 5. System configuration at start of simulation. 5. MULTI-BODY DYNAMIC MODEL In order to assess the range of performance of the EDL system, a three dimensional multi-body dynamic model was developed in ADAMS [1].

4 Figure 5. System configuration at start of simulation. 5. MULTI-BODY DYNAMIC MODEL In order to assess the range of performance of the EDL system, a three dimensional multi-body dynamic model was developed in ADAMS [1]. This model is intended to simulate the behavior of the system over the entire terminal descent phase, from parachute deployment until surface impact. The simulation must cover this range because the initial conditions of each event are determined by the final state of earlier events. Terminal descent lasts approximately two minutes, and with system frequencies as low as 0.07 Hz, there is not enough time for large-scale motions to damp out between events. The parachute (including shroud lines) was treated as a rigid body, as were the backshell, heatshield, and lander. Additional rigid bodies (or parts ) were added at the confluence points of the triple bridles above and below the backshell. Multiple rigid bodies were connected by joints and force relationships. Figure 5 shows the essentially two-body system (parachute and entry vehicle) at the start of terminal descent. During final descent, the system takes the form shown in Figure 6, and is essentially a three-body system. The heatshield has been jettisoned, and the lander has been deployed at the end of a 20-meter bridle system. Note that all rigid bodies were specified in ADAMS by their mass properties, initial position and orientation. The graphical representations shown in the figures are for reducing geometry errors and for visualization. Figure 6. System configuration during final descent. The initial conditions of the bodies (altitude, velocity, spin rate and orientation) were taken from the results of atmospheric entry simulations performed at NASA Langley Research Center. In addition, an initial disturbance was imparted to represent the effect of parachute deployment. The various rigid bodies were connected by various joints and forces during the course of the simulation. Joints represented constraints, such as fixed joints (6 degree of freedom) and ball joints (3 degree of freedom). Connections such as bridles were modeled by explicitly defining force relationships between bodies. Damping was also included. In addition to the forces used to connect bodies, there were three external forces included in the simulation. The simplest was Mars gravity (3.72 m/sec 2 ), which was easily entered in ADAMS. Another force was the thrust of the solid rocket motors. The thrust profile of each rocket was specified as an explicit function of time. The aerodynamic forces on the parachute and on each body were formulated assuming an axisymmetric flow [2]: 1126

5 Fx Fy VV A C N = 1 2 ρ ( α) x, sinα (1) VV A C N = 1 2 ρ ( α) y, sinα (2) Fz = 1 2 ρv AC 2 T( α), (3) where ρ is the atmospheric density, V=(V x,v y,v z ) is the velocity of the body relative to the wind, A is the reference area, α is the total angle of attack, and C N and C T are aerodynamic coefficients for each of the bodies. In these equations, the velocity vector is resolved in the reference frame of the body, and the forces are applied at the center of pressure. The forces in equations (1) through (3) were defined in ADAMS using interpolation tables for C N and C T as a function of α, and other formulas for the velocity components. In order to avoid numerical ill-conditioning, the lookup table for the normal force coefficient actually contained values of C N /sinα, which asymptotes to a constant value at zero angle of attack. The atmospheric density ρ was defined by using an interpolation table for ρ as a function of altitude. The most important aerodynamic forces are those acting on the parachute, but forces were also applied to the backshell and lander (as well as the backshell/lander/heatshield combination before heatshield separation and backshell/lander combination after heatshield separation. The axisymmetric formulation was used as an approximation to the actual forces acting on the lander. Forebody wake effects, as well as vortex shedding by upstream bodies, were ignored for this simulation. The actual C N and C T of each of the bodies, including the disk-gap-band parachute in Mars conditions were selected based on wind tunnel tests performed at NASA Langley Research Center and parachute drop tests conducted by Pioneer Aerospace Corporation. Stability of the EDL system is governed largely by the parachute s aerodynamic stability. But even a stable parachute is subject to dynamic oscillations due to wind shear as it descends. Wind speeds of up to 20 m/sec are anticipated at the landing site [3], but almost nothing is known about the amount of wind shear over length scales of interest for this lander. For the ADAMS simulation, the wind velocity was treated as a vector function of altitude. The wind profile was entered as a lookup table function. 6. DISPERSIONS ON SIMULATION PARAMETERS A large number of uncertain parameters affect the motion of the system. The initial condition is determined by trajectory analysis, based on errors in entry angle and attitude as well as uncertainties in the model of the upper atmosphere. An important input to the simulation is the wind model, particularly wind shear. Other critical drivers include the performance of separation systems and rocket motors, and errors in estimating altitude and descent rate from the radar altimeter. The complicated interaction of system dynamics and multiple uncertain parameters made it impractical to determine worst-case conditions. Instead, a Monte Carlo simulation approach was undertaken to estimate the likely range of surface impact conditions and other engineering data by dispersing all critical simulation parameters. Typical Monte Carlo simulation studies included 200 to 1,000 runs. 7. RESULTS Figure 7 animates an ideal terminal descent simulation from /RAD ignition till bridle cut. Corrected by the constrained firing, the average backshell angle, which is approximately equal to the bridle angle, from vertical is near zero, as shown in Figure 8. Comparing to the simulation withoutinfigure9,itisseenthatthefiringcan effectively reduce the RAD-induced horizontal velocity. In this ideal case, the RAD-induced horizontal velocity is reduced to near zero and the horizontal velocity at bridle cut till ground impact is almost identical to that before RAD firing. y z x Bridle Cut Middle of Firing /RAD Ign. Figure 7. Simulation of terminal descent with firing. Hit by Wind T=1.65 sec 1.04 degrees RAD/TIRS Ign. T=6.48 sec 13.1 degrees Bridle Cut T=9.43 sec 13.6 degrees Figure 8. Backshell angle during terminal descent with firing. 1127

6 Horizontal Velocity (m/sec) Vh before RAD firing Time (sec) No With Figure 9. Horizontal velocity during terminal descent. Impact Velocity Magnitude (m/sec) airbag-driven req. (Vm<24 m/sec) Cross Wind Velocity (m/sec) with Figure 10. Impact velocity magnitude vs. cross wind speed. When subjected to cross winds of 10, 15, 20, 25 m/sec 500 m above ground and no dispersions on simulation parameters, the simulation results of ideal firing runs are compared to that without in Figure 10. Without, the impact velocity magnitudes will exceed the airbag capacity of 24 m/sec when cross wind velocities are larger than 15 m/sec. With ideal firing, impact velocity magnitudes are reduced to be well within the airbag capacity. When run with dispersions on all simulation parameters, the impact velocity vs. impact angle from vertical are plotted with current airbag capacity in Figures 11 and 12 for the cases of and with, respectively. It is clearly observed from the figures that reduces the number of runs exceeding current airbag capacity, especially in the region of high velocities associated with large impact angles. When run with dispersions on all simulation parameters and variation on terrain slope (+/- 15 degrees, normal distribution), the impact velocity results of vs. are compared in seven bins of horizontal impact velocities without (Figures 13 and 14). It is observed that, with current constrained, the airbag horizontal (tangential) requirement are met and insurance is provided against wind shears for un-mitigated horizontal velocities up to 30 m/sec. The dispersions of vertical (normal) impact velocity are also reduced, and will allow lowering of bridle cut height altitude to meet airbag requirement of 12 m/sec. Impact Velocity Magnitude (m/sec) X airbag capacity O simulation data Impact Angle (deg) Figure 11. Impact velocity vs. airbag capacity (). Impact Velocity Magnitude (m/sec) with X airbag capacity O simulation data Impact Angle (deg) Figure 12. Impact velocity vs. airbag capacity (with ). Tangential Impact Velocity (m/sec) Horizontal Impact Velocity Bins w/o (m/sec) Horizontal Impact Velocity Bins w/o (m/sec) >30 airbag-driven req. (Vt<20 m/sec) Figure 13. Tangential impact velocity ( vs. ) 1128

7 Normal Impact Velocity (m/sec) Horizontal Impact Velocity Bins w/o (m/sec) >30 target velocity with 15 m Hcut = 10.6 m/sec EDL system to achieve a successful landing under a wide range of circumstances. ACKNOWLEDGEMENT The work described herein was conducted by the Jet Propulsion Laboratory, California Institute of Technology, under contract with National Aeronautics and Space Administration. REFERENCES [1] ADAMS/Solver Reference Manual, Version 11.0, Figure 14. Normal impact velocity ( vs. ) These above-described results proved to be very important in predicting/evaluating the terminal descent performance with various versions of firing algorithms. The results were also useful to develop design criteria and qualification tests for the airbags. Considerable effort was required to provide sufficient abrasion resistance for the worst-case landing approach angles and orientations. Besides the firing algorithm development and airbag design, many other spacecraft sub-system designs also benefited from these simulation results. Worst-case loads in bridles were extracted from the runs. The design of the descent rate limiter depended greatly on the simulations. Also, the swinging of the lander affects communications with Earth during the descent. [2] Kenneth S. Smith, Chia-Yen Peng, and Ali Behboud, Multi-body Dynamic Simulation of Mars Pathfinder Entry, Descent and Landing, JPL Document D-13298, April 1, 1995, Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California. [3] A. Seiff, Mars Atmospheric Winds Indicated by Motion of the Viking Landers During Parachute Descent, Journal of Geophysical Research, Vol. 98, No. E4, pages , April 25, CONCLUSION An extensive effort has been successfully conducted to simulate the multi-body dynamics of the MER EDL system during terminal descent. This effort is critical for verification of the EDL concept and prediction of its performance under Mars-like conditions. The simulation results illustrate the effectiveness of system to reduce the horizontal impact velocity induced by RAD firing. Without, terminal descent performance will exceed current airbag capacity with step wind velocities larger than 15 m/sec. With ideal, performance is greatly enhanced to stay within current airbag capacity for step wind velocities at least up to 25 m/sec. Under dispersed simulation conditions, current design of enables the reduction of horizontal (tangential) impact velocities from 30 m/sec to less than 20 m/sec. The simulation results also provide critical information for the design and test of various EDL sub-systems (parachute, bridle, descent rate limiter, lander, airbag, and etc.). The simulations are also used to evaluate the clearance loss during various separation events. When combined with a comprehensive test program, these simulations will give confidence in the ability of the MER 1129

Long-Range Rovers for Mars Exploration and Sample Return

2001-01-2138 Long-Range Rovers for Mars Exploration and Sample Return Joe C. Parrish NASA Headquarters ABSTRACT This paper discusses long-range rovers to be flown as part of NASA s newly reformulated Mars