- Email: [email protected]
Designing planetary protection into the Mars Observer mission
Adv. Space Res. Vol. 15. No. 3, pp. (3)257_(3)260. 1995
copyright Q 1994 COSPAR Printed in Great Britain. All rights re.served. 0273-l 177/95 $7.00 + 0.00
DESIGNING PLANETARY PROTECTION INTO THE MARS OBSERVER MISSION T. H. Sweetser, C. A. Halsell and R. J. Cesarone Jet Propulsion Luboratory, California Institute of Technology, 4800 Oak Grove Drive, Pasadena, CA 91109, U.S.A.
ABSTRACT Planetary protection has been an important consideration during the process of designing the Mars Observer mission. It affected trajectory design of both the interplanetary transfer and the orbits at Mars; these in turn affected the observation strategies developed for the mission. The Project relied mainly on the strategy of collision avoidance to prevent contamination of Mars. Conservative estimates of spacecraft reliability and Martian atmosphere density were used to evaluate decisions concerning the interplanetary trajectory, the orbit insertion phase at Mars, and operations in orbit at Mars and afterwards. Changes in the trajectory design, especially in the orbit insertion phase, required a refinement of those estimates. INTRODUCTION The planet Mars occupies a special place among the objects of solar system exploration. Mars is the most likely place besides Earth for life to have originated, and it offers us the best hope of finding an independent evolution to compare to our own. In order to protect that hope it is important that we prevent any evidence of Martian life from being contaminated by Earth-based life. For this reason planetary protection rules for Mars exploration were established by NASA to comply with international agreements reached through COSPAR. The Mars Observer Project will use an orbiting spacecraft at Mars to make global measurements of elemental and mineralogical surface composition, topography, and climate /1,2/. As an orbiting mission it is classified as a Category III mission with respect to planetary protection This means that the Project must meet certain standards of cleanliness in the manufacture and launch of the spacecraft and must meet certain criteria regarding the probability of impact at Mars. The cleanliness standards were satisfied but will not be discussed in this paper. The impact criteria which Mars Observer was designed to meet were the following: -
the probability of launch vehicle impact at Mars must be less than lo-’ (for Mars Observer this refers to the Transfer Orbit Stage, or TOS); the probability of spacecraft impact at Mars must be less than 10m4before 2009; the probability of spacecraft impact at Mars must be less than 5 x 10’ before 2039.
The Mars Observer Project defined impact at Mars to be any occurrence of an altitude less than 100 km. INITIAL TRAIFCTORY DESIGN The overall trajectory design for Mars Observer is straightforward. In outline it involves a direct transfer to Mars, insertion initially into an ellipse at Mars which acts as a holding orbit until the Sun-Mars-orbit plane geometry is what is desired for science, and then a series of maneuvers resulting in a near-circular mapping orbit. The details of the trajectory design, however, have changed significantly over the years. Mars Observer was originally designed for a 1990 launch on a Type II transfer to Mars in which the spacecraft would reach Mars after the aphelion of its heliocentric transfer orbit in order to arrive with the minimum Mars-relative velocity. This would be the Cruise Phase of the mission. The Orbit Insertion Phase of the mission would be initiated when the spacecraft executes the Mars Orbit Insertion (MOD maneuver, which would place the vehicle into an initial l-day capture orbit. From the perspective of the trajectory engineer, an approach to the planet over either of its poles was equally feasible. The North
(3)257
(3)258
T.H.Sweetser etal
polar approach would result in a mapping orbit characterized by descending equatorial crossings on the lighted side of the planet, with the South polar approach producing the reverse geometry. There were no strong arguments put forth for either geometry although there was some slight preference among the Project science community for the sunlit descending nodes. Thus the North polar approach became the baseline target for the MO1 and its resulting l-day orbit. The original mapping orbit was chosen to have an altitude of 361 km with a 3-sol repeat of its ground track (a “sol” is a Mars day). In addition it is Sun-synchronous with its descending node at 2 PM mean solar time. The orbit is near-polar, near-circular, and frozen (i.e., fixed in eccentricity and periapsis location, which is over Mars’s South polar region). In this discussion the period from the beginning of the mapping orbit to the end of planetary protection requirements in 2039 will be referred to as the Orbit Phase of the mission. INITIAL PLANETARY PROTECTION ANALYSIS To begin the planetary protection analysis, the probability of impact before 2009 was treated as a consumable quantity and apportioned equally to each of the three mission phases described above. Each phase could then be examined independently, where for each phase the probability of impact satisfies
P, = C PiQi with Pi being the probability that a given orbit will impact before 2009 and Qi being the probability that the maneuver changing that orbit will fail. For this initial analysis (documented in /3/) the Project used the conservative rule of thumb that any maneuver has a 1% chance of failing. Launch and Cruise Phase Aimpoint Biasing In order to reduce the probability of impact after launch and during the Cruise Phase, it was necessary to aim the TOS and spacecraft away from Mars. This technique is known as aimpoint biasing. The target for injection and the aimpoints for each Trajectory Correction Maneuver (TCM) were chosen to minimize the sum of the deterministic velocity changes at each maneuver while still meeting planetary protection requirements. To do this, Monte-Carlo simulations of the cruise included the effects of injection, orbit determination, and maneuver execution errors on a candidate trajectory in order to calculate contours of impact probability in the target plane at Mars. Maneuver and injection aimpoints were then adjusted to reduce the total velocity change needed while staying on the desired probability contours, and the simulations were redone with this new trajectory. After a few iterations this process converged on a final set of aimpoints for each launch day. In all cases the injection was biased to the TOS requirement of 1.0 x 10.‘. TCM-2, -3, and -4 were biased to Pi = 1.0 x 10m4which, when combined with the assumed Qi value of 0.01, would result in a negligible contribution to the impact probability. TCM-1, however, required some additional work. The aimpoint for TCM-1 was indirectly defined to be the aimpoint which minimized TCM-1 and TCM-2 velocity changes given the injection and TCM-2 aimpoints. These latter aimpoints were adjusted until the optimized TCM-1 aimpoint had a 0.3 x 10m2TCM-1 impact probability. which accounted for most of the cruise allocation of impact probability once maneuver reliability was taken into account ( (0.3 x 10T2)x (1 x 10e2)= 0.3 x 10e4). Orbit Insertion Phase Compliance Analysis of the Orbit Insertion Phase indicated that the cumulative value of the probability of impact for this mission phase was well within prescribed limits. The orbits that characterized the Orbit Insertion Phase were shown to be sufficiently stable as to be in accord with all planetary protection requirements, Mapping Orbit Lifetime and Quarantine Orbit A much more detailed analysis of the Orbit Phase was undertaken by Chen-wan Yen at IPL. Some of the factors that influenced this analysis were the anticipated timings and magnitudes of the peaks of upcoming Solar cycles, the area coefficient (reciprocal of the ballistic coefficient) of the Mars Observer spacecraft, and the Martian atmosphere (a new model of which was developed for this study /4/). Because of uncertainties in key variables, the analysis was of necessity parametric. An important conclusion, however, was that if the spacecraft area coefficient were greater than 0.020 m’/kg then the 361 km mapping orbit would be too likely to decay before the year 2009. Although the actual Mars Observer spacecraft configuration was not known at that time, its area coefficient was expected to be between 0.015 and 0.033 m2/kg and thus likely to result in a planetary protection requirement violation. The solution was to plan to maneuver the spacecraft into a higher orbit (called a quarantine orbit) at the end of the mission.
MarsObserver Planetary
Protection
CT)259
CHANGES TO THE TRAJECTORY DESIGN In the late 1980s a number of modifications to the original mission plan were initiated. The first of these was a 25 month delay in the launch date to September 1992, with a change in the primary launch vehicle from the Space Shuttle Orbiter to a Titan III. In order to make the launch period as long as possible and to maximize the spacecraft mass that could be delivered at Mars, days were included at the beginning and end which required significant spacecraft maneuvers during the Cruise Phase. Another major change was to have the spacecraft go into a 3-day capture ellipse at MO1 rather than into the previously planned l-day orbit, This would decrease the duration of the MO1 bum by about 15% and result in a significant reduction in the maneuver’s gravity losses (which increase with the square of the bum duration), thus reducing the propellant requirements for the mission. A new maneuver, Ellipse Change Maneuver - 2 (ECM-2), was included in the Orbit Insertion Phase 21 days after MO1 to reduce the orbit period from 3 days to the original 1 day. Also the Transfer to Low Orbit maneuver (TLO) was split into two maneuvers to reduce its gravity losses. A final major change was in the altitude of the mapping orbit. Again this was motivated by a desire to reduce the propellant requirements of the mission. By raising the mapping orbit to 378 km mean altitude with a 7-sol repeat of the ground track, this savings was achieved in several ways. The sum of the maneuvers in the Orbit Insertion Phase was reduced slightly since the final orbit was higher. Also the increased altitude meant that less drag would affect the spacecraft, so orbit maintenance requirements would be reduced. Finally there was the possibility that the higher altitude orbit would eliminate the need for a maneuver into quarantine orbit. PLANETARY PROTECTION RE-EXAMINED The change in the launch period did not require any change in the planetary protection strategy for the Cruise Phase. The only real difference in the transfer trajectories was that some now included large deterministic maneuvers during the Cruise Phase; for these trajectories the probability that the TOS would hit Mars was vanishingly small. All that was needed was to find aimpoints for the new interplanetary transfer trajectories. Since the launch would be two years later, the spacecraft would not be required to stay in orbit quite as long for planetary protection purposes. That, combined with the increase in the mapping orbit altitude, made mission designers even more hopeful that a quarantine orbit would not be needed. Since the spacecraft configuration was fixed by then, they initiated a study to determine its area coefficient more precisely, but an allocation of propellant for performing a quarantine raise maneuver was retained as a precaution. Most problematic was the addition of the 3-day orbit in the Orbit Insertion Phase. Numerical integration of the orbit, reinforced by theoretical analysis of the perturbations, showed that solar gravity causes a long-period cyclic variation in the eccentricity of the orbit. This would cause impact with Mars about five years after arrival if a spacecraft failure left it stranded in that orbit. This eccentricity variation is also present in the smaller intermediate ellipses of the Orbit Insertion Phase, but is too small to result in impact. Ironically, an asymmetry in the orbits now appeared: if the Project had decided to use an approach over the South Pole into MO1 then a phase difference in the perturbations would have made the 3-day orbit safe. The Project had a number of options to consider at this point. Mars Observer could forgo the use of the 3-day orbit altogether, although this option would become more difficult to implement as spacecraft sequence development proceeded. As an alternative the Project could change the arrival to come over the South pole of Mars, but this would involve finding out if the spacecraft design would allow rotation in the opposite direction; this option carried the risk of requiring a change to the spacecraft design at too late a date to be supported. Finally, since the probability of impact depended on the probability of spacecraft failure and the 1% maneuver failure estimate was very conservative, the Project could derive a more realistic model of spacecraft reliability in the hope that the reliability would be high enough to allow the use of the 3-day orbit for the short period of time needed. The Project chose to pursue this last option, with the first as a fallback option. An extensive analysis of spacecraft failure histories done by Milena Krasich at JPL led to the conclusion that the best model for spacecraft reliability is the Weibull distribution:
where R(t) is the spacecraft reliability at time t hours, h is the spacecraft failure rate per hour after one hour of operations, and fi (the shape parameter) has to do with how the failure rate of a system changes with time. When a system failure rate is constant, b equals 1 and the Weibull distribution reduces to the familiar exponential distribution. Interplanetary spacecraft, however, have failure rates that decrease with time (as is true of complex systems generally). Models of Voyager, Magellan, and Galileo resulted in shape parameters ranging between 0.47 and 0.52 so a p value of 0.55 was conservatively chosen for Mars Observer. An
T.
(3)260
H.Sweetser ef al.
TABLE. Probabilities of Impact at Mars Launch vehicle (TOS) Spacecraft before 2009 Cruise Phase Orbit Insertion Phase Orbit Phase Spacecraft before 2039
< 1o-5 0.4 x lQ4 0.1 x 10-4 0.1 x lo4 0.2 x lo4 0.8 x 1O-2
additional analysis was done to identify systems within the Mars Observer spacecraft which are necessary to perform successful maneuvers; a failure analysis of these systems yielded a value of 1.419 x 10M6 for h. The probability of a failure between two maneuvers preventing the second maneuver from being performed is
Q = WJ - NtJ where t, and t2 are the times of the first and second maneuvers respectively. Throughout the Mars Observer mission the above model results in a much lower probability of maneuver failure than the conservative rule of thumb value of 1% and the planetary protection plan for the mission was revised to reflect this. At the same time, Duane Roth at JPL reanalyzed the orbit determination errors at arrival. An improvement was made in the way that Mars position knowledge errors were accounted for which significantly reduced the expected error in the spacecraft state relative to Mars. FINAL PLANETARY PROTECTION PLAN The planetary protection plan for launch was not changed, although launch targets continued to be adjusted up to the last possible minute to account for improvements in knowledge of the behavior of the launch system. For the Cruise Phase, navigation analysis showed that biasing is not needed for TCM aimpoints because of the improved spacecraft reliability estimate and the reduction in the expected orbit determination error. Removal of the TCM biasing saved several meters per second of cruise AV. The result of the improved spacecraft reliability model for the Orbit Insertion Phase is that the 3-day initial ellipse can be used. For the Orbit Phase, spacecraft configuration analysis done by Angus McRonald at JPL found that the area coefficient is higher than expected and ranges between 0.30 m2/kg and 0.38 m2/kg, depending on the orientation of the spacecraft. Thus &spite the later arrival into the mapping orbit and its higher altitude a quarantine orbit is necessary. An increase to a quarantine altitude of 405 km is sufficient. One advantage of the improved reliability estimate is that this raise to quarantine orbit can be done as late as the turn of the century, thus allowing the possibility of an extended mission within the nominal mapping orbit. The results of all this analysis (documented in /5/) are shown in the Table. ACKNOWLEDGMENTS The research described in this paper was performed by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. REFERENCES 1.
D. Ainsworth, Mars Observer: Return to the Red Planet, Astronomy, 20, # 10, p. 28 (1992).
2.
A.L. Albee, R.E. Arvidson, and F.D. Palluconi, Mars Observer mission, J. Geophys. Rex # E5, 76657680 (1992).
3.
J.B. Barengoltz, Mars Observer Planetary Protection Plan, JPL Internal Document D-4481, Jet Propulsion Laboratory, Pasadena, California, (October 1985).
4.
A.I.F. Stewart, private communication (1987).
5.
J.B. Barengoltz, Mars Observer Planetary Protection Plan Supplement 1: Prelaunch Report, JPL Internal Document D-4481 Supplement 1, Jet Propulsion Laboratory, Pasadena, California, (Js 1992).
97,
copyright Q 1994 COSPAR Printed in Great Britain. All rights re.served. 0273-l 177/95 $7.00 + 0.00
DESIGNING PLANETARY PROTECTION INTO THE MARS OBSERVER MISSION T. H. Sweetser, C. A. Halsell and R. J. Cesarone Jet Propulsion Luboratory, California Institute of Technology, 4800 Oak Grove Drive, Pasadena, CA 91109, U.S.A.
ABSTRACT Planetary protection has been an important consideration during the process of designing the Mars Observer mission. It affected trajectory design of both the interplanetary transfer and the orbits at Mars; these in turn affected the observation strategies developed for the mission. The Project relied mainly on the strategy of collision avoidance to prevent contamination of Mars. Conservative estimates of spacecraft reliability and Martian atmosphere density were used to evaluate decisions concerning the interplanetary trajectory, the orbit insertion phase at Mars, and operations in orbit at Mars and afterwards. Changes in the trajectory design, especially in the orbit insertion phase, required a refinement of those estimates. INTRODUCTION The planet Mars occupies a special place among the objects of solar system exploration. Mars is the most likely place besides Earth for life to have originated, and it offers us the best hope of finding an independent evolution to compare to our own. In order to protect that hope it is important that we prevent any evidence of Martian life from being contaminated by Earth-based life. For this reason planetary protection rules for Mars exploration were established by NASA to comply with international agreements reached through COSPAR. The Mars Observer Project will use an orbiting spacecraft at Mars to make global measurements of elemental and mineralogical surface composition, topography, and climate /1,2/. As an orbiting mission it is classified as a Category III mission with respect to planetary protection This means that the Project must meet certain standards of cleanliness in the manufacture and launch of the spacecraft and must meet certain criteria regarding the probability of impact at Mars. The cleanliness standards were satisfied but will not be discussed in this paper. The impact criteria which Mars Observer was designed to meet were the following: -
the probability of launch vehicle impact at Mars must be less than lo-’ (for Mars Observer this refers to the Transfer Orbit Stage, or TOS); the probability of spacecraft impact at Mars must be less than 10m4before 2009; the probability of spacecraft impact at Mars must be less than 5 x 10’ before 2039.
The Mars Observer Project defined impact at Mars to be any occurrence of an altitude less than 100 km. INITIAL TRAIFCTORY DESIGN The overall trajectory design for Mars Observer is straightforward. In outline it involves a direct transfer to Mars, insertion initially into an ellipse at Mars which acts as a holding orbit until the Sun-Mars-orbit plane geometry is what is desired for science, and then a series of maneuvers resulting in a near-circular mapping orbit. The details of the trajectory design, however, have changed significantly over the years. Mars Observer was originally designed for a 1990 launch on a Type II transfer to Mars in which the spacecraft would reach Mars after the aphelion of its heliocentric transfer orbit in order to arrive with the minimum Mars-relative velocity. This would be the Cruise Phase of the mission. The Orbit Insertion Phase of the mission would be initiated when the spacecraft executes the Mars Orbit Insertion (MOD maneuver, which would place the vehicle into an initial l-day capture orbit. From the perspective of the trajectory engineer, an approach to the planet over either of its poles was equally feasible. The North
(3)257
(3)258
T.H.Sweetser etal
polar approach would result in a mapping orbit characterized by descending equatorial crossings on the lighted side of the planet, with the South polar approach producing the reverse geometry. There were no strong arguments put forth for either geometry although there was some slight preference among the Project science community for the sunlit descending nodes. Thus the North polar approach became the baseline target for the MO1 and its resulting l-day orbit. The original mapping orbit was chosen to have an altitude of 361 km with a 3-sol repeat of its ground track (a “sol” is a Mars day). In addition it is Sun-synchronous with its descending node at 2 PM mean solar time. The orbit is near-polar, near-circular, and frozen (i.e., fixed in eccentricity and periapsis location, which is over Mars’s South polar region). In this discussion the period from the beginning of the mapping orbit to the end of planetary protection requirements in 2039 will be referred to as the Orbit Phase of the mission. INITIAL PLANETARY PROTECTION ANALYSIS To begin the planetary protection analysis, the probability of impact before 2009 was treated as a consumable quantity and apportioned equally to each of the three mission phases described above. Each phase could then be examined independently, where for each phase the probability of impact satisfies
P, = C PiQi with Pi being the probability that a given orbit will impact before 2009 and Qi being the probability that the maneuver changing that orbit will fail. For this initial analysis (documented in /3/) the Project used the conservative rule of thumb that any maneuver has a 1% chance of failing. Launch and Cruise Phase Aimpoint Biasing In order to reduce the probability of impact after launch and during the Cruise Phase, it was necessary to aim the TOS and spacecraft away from Mars. This technique is known as aimpoint biasing. The target for injection and the aimpoints for each Trajectory Correction Maneuver (TCM) were chosen to minimize the sum of the deterministic velocity changes at each maneuver while still meeting planetary protection requirements. To do this, Monte-Carlo simulations of the cruise included the effects of injection, orbit determination, and maneuver execution errors on a candidate trajectory in order to calculate contours of impact probability in the target plane at Mars. Maneuver and injection aimpoints were then adjusted to reduce the total velocity change needed while staying on the desired probability contours, and the simulations were redone with this new trajectory. After a few iterations this process converged on a final set of aimpoints for each launch day. In all cases the injection was biased to the TOS requirement of 1.0 x 10.‘. TCM-2, -3, and -4 were biased to Pi = 1.0 x 10m4which, when combined with the assumed Qi value of 0.01, would result in a negligible contribution to the impact probability. TCM-1, however, required some additional work. The aimpoint for TCM-1 was indirectly defined to be the aimpoint which minimized TCM-1 and TCM-2 velocity changes given the injection and TCM-2 aimpoints. These latter aimpoints were adjusted until the optimized TCM-1 aimpoint had a 0.3 x 10m2TCM-1 impact probability. which accounted for most of the cruise allocation of impact probability once maneuver reliability was taken into account ( (0.3 x 10T2)x (1 x 10e2)= 0.3 x 10e4). Orbit Insertion Phase Compliance Analysis of the Orbit Insertion Phase indicated that the cumulative value of the probability of impact for this mission phase was well within prescribed limits. The orbits that characterized the Orbit Insertion Phase were shown to be sufficiently stable as to be in accord with all planetary protection requirements, Mapping Orbit Lifetime and Quarantine Orbit A much more detailed analysis of the Orbit Phase was undertaken by Chen-wan Yen at IPL. Some of the factors that influenced this analysis were the anticipated timings and magnitudes of the peaks of upcoming Solar cycles, the area coefficient (reciprocal of the ballistic coefficient) of the Mars Observer spacecraft, and the Martian atmosphere (a new model of which was developed for this study /4/). Because of uncertainties in key variables, the analysis was of necessity parametric. An important conclusion, however, was that if the spacecraft area coefficient were greater than 0.020 m’/kg then the 361 km mapping orbit would be too likely to decay before the year 2009. Although the actual Mars Observer spacecraft configuration was not known at that time, its area coefficient was expected to be between 0.015 and 0.033 m2/kg and thus likely to result in a planetary protection requirement violation. The solution was to plan to maneuver the spacecraft into a higher orbit (called a quarantine orbit) at the end of the mission.
MarsObserver Planetary
Protection
CT)259
CHANGES TO THE TRAJECTORY DESIGN In the late 1980s a number of modifications to the original mission plan were initiated. The first of these was a 25 month delay in the launch date to September 1992, with a change in the primary launch vehicle from the Space Shuttle Orbiter to a Titan III. In order to make the launch period as long as possible and to maximize the spacecraft mass that could be delivered at Mars, days were included at the beginning and end which required significant spacecraft maneuvers during the Cruise Phase. Another major change was to have the spacecraft go into a 3-day capture ellipse at MO1 rather than into the previously planned l-day orbit, This would decrease the duration of the MO1 bum by about 15% and result in a significant reduction in the maneuver’s gravity losses (which increase with the square of the bum duration), thus reducing the propellant requirements for the mission. A new maneuver, Ellipse Change Maneuver - 2 (ECM-2), was included in the Orbit Insertion Phase 21 days after MO1 to reduce the orbit period from 3 days to the original 1 day. Also the Transfer to Low Orbit maneuver (TLO) was split into two maneuvers to reduce its gravity losses. A final major change was in the altitude of the mapping orbit. Again this was motivated by a desire to reduce the propellant requirements of the mission. By raising the mapping orbit to 378 km mean altitude with a 7-sol repeat of the ground track, this savings was achieved in several ways. The sum of the maneuvers in the Orbit Insertion Phase was reduced slightly since the final orbit was higher. Also the increased altitude meant that less drag would affect the spacecraft, so orbit maintenance requirements would be reduced. Finally there was the possibility that the higher altitude orbit would eliminate the need for a maneuver into quarantine orbit. PLANETARY PROTECTION RE-EXAMINED The change in the launch period did not require any change in the planetary protection strategy for the Cruise Phase. The only real difference in the transfer trajectories was that some now included large deterministic maneuvers during the Cruise Phase; for these trajectories the probability that the TOS would hit Mars was vanishingly small. All that was needed was to find aimpoints for the new interplanetary transfer trajectories. Since the launch would be two years later, the spacecraft would not be required to stay in orbit quite as long for planetary protection purposes. That, combined with the increase in the mapping orbit altitude, made mission designers even more hopeful that a quarantine orbit would not be needed. Since the spacecraft configuration was fixed by then, they initiated a study to determine its area coefficient more precisely, but an allocation of propellant for performing a quarantine raise maneuver was retained as a precaution. Most problematic was the addition of the 3-day orbit in the Orbit Insertion Phase. Numerical integration of the orbit, reinforced by theoretical analysis of the perturbations, showed that solar gravity causes a long-period cyclic variation in the eccentricity of the orbit. This would cause impact with Mars about five years after arrival if a spacecraft failure left it stranded in that orbit. This eccentricity variation is also present in the smaller intermediate ellipses of the Orbit Insertion Phase, but is too small to result in impact. Ironically, an asymmetry in the orbits now appeared: if the Project had decided to use an approach over the South Pole into MO1 then a phase difference in the perturbations would have made the 3-day orbit safe. The Project had a number of options to consider at this point. Mars Observer could forgo the use of the 3-day orbit altogether, although this option would become more difficult to implement as spacecraft sequence development proceeded. As an alternative the Project could change the arrival to come over the South pole of Mars, but this would involve finding out if the spacecraft design would allow rotation in the opposite direction; this option carried the risk of requiring a change to the spacecraft design at too late a date to be supported. Finally, since the probability of impact depended on the probability of spacecraft failure and the 1% maneuver failure estimate was very conservative, the Project could derive a more realistic model of spacecraft reliability in the hope that the reliability would be high enough to allow the use of the 3-day orbit for the short period of time needed. The Project chose to pursue this last option, with the first as a fallback option. An extensive analysis of spacecraft failure histories done by Milena Krasich at JPL led to the conclusion that the best model for spacecraft reliability is the Weibull distribution:
where R(t) is the spacecraft reliability at time t hours, h is the spacecraft failure rate per hour after one hour of operations, and fi (the shape parameter) has to do with how the failure rate of a system changes with time. When a system failure rate is constant, b equals 1 and the Weibull distribution reduces to the familiar exponential distribution. Interplanetary spacecraft, however, have failure rates that decrease with time (as is true of complex systems generally). Models of Voyager, Magellan, and Galileo resulted in shape parameters ranging between 0.47 and 0.52 so a p value of 0.55 was conservatively chosen for Mars Observer. An
T.
(3)260
H.Sweetser ef al.
TABLE. Probabilities of Impact at Mars Launch vehicle (TOS) Spacecraft before 2009 Cruise Phase Orbit Insertion Phase Orbit Phase Spacecraft before 2039
< 1o-5 0.4 x lQ4 0.1 x 10-4 0.1 x lo4 0.2 x lo4 0.8 x 1O-2
additional analysis was done to identify systems within the Mars Observer spacecraft which are necessary to perform successful maneuvers; a failure analysis of these systems yielded a value of 1.419 x 10M6 for h. The probability of a failure between two maneuvers preventing the second maneuver from being performed is
Q = WJ - NtJ where t, and t2 are the times of the first and second maneuvers respectively. Throughout the Mars Observer mission the above model results in a much lower probability of maneuver failure than the conservative rule of thumb value of 1% and the planetary protection plan for the mission was revised to reflect this. At the same time, Duane Roth at JPL reanalyzed the orbit determination errors at arrival. An improvement was made in the way that Mars position knowledge errors were accounted for which significantly reduced the expected error in the spacecraft state relative to Mars. FINAL PLANETARY PROTECTION PLAN The planetary protection plan for launch was not changed, although launch targets continued to be adjusted up to the last possible minute to account for improvements in knowledge of the behavior of the launch system. For the Cruise Phase, navigation analysis showed that biasing is not needed for TCM aimpoints because of the improved spacecraft reliability estimate and the reduction in the expected orbit determination error. Removal of the TCM biasing saved several meters per second of cruise AV. The result of the improved spacecraft reliability model for the Orbit Insertion Phase is that the 3-day initial ellipse can be used. For the Orbit Phase, spacecraft configuration analysis done by Angus McRonald at JPL found that the area coefficient is higher than expected and ranges between 0.30 m2/kg and 0.38 m2/kg, depending on the orientation of the spacecraft. Thus &spite the later arrival into the mapping orbit and its higher altitude a quarantine orbit is necessary. An increase to a quarantine altitude of 405 km is sufficient. One advantage of the improved reliability estimate is that this raise to quarantine orbit can be done as late as the turn of the century, thus allowing the possibility of an extended mission within the nominal mapping orbit. The results of all this analysis (documented in /5/) are shown in the Table. ACKNOWLEDGMENTS The research described in this paper was performed by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration. REFERENCES 1.
D. Ainsworth, Mars Observer: Return to the Red Planet, Astronomy, 20, # 10, p. 28 (1992).
2.
A.L. Albee, R.E. Arvidson, and F.D. Palluconi, Mars Observer mission, J. Geophys. Rex # E5, 76657680 (1992).
3.
J.B. Barengoltz, Mars Observer Planetary Protection Plan, JPL Internal Document D-4481, Jet Propulsion Laboratory, Pasadena, California, (October 1985).
4.
A.I.F. Stewart, private communication (1987).
5.
J.B. Barengoltz, Mars Observer Planetary Protection Plan Supplement 1: Prelaunch Report, JPL Internal Document D-4481 Supplement 1, Jet Propulsion Laboratory, Pasadena, California, (Js 1992).
97,