Attitude Control on TET-1 - Experiences from the First Year of
Contents
1. 7 SpaceOps 2014 Pasadena California May 05 09 B IMU outages The IMU stopped delivering valid data for more than five hours six times during the past year Both IMU 1 and IMU 2 had three outages each Two times a drop to Safe Mode followed because of low battery power due to the ensuing attitude deviations The problem could always be mitigated by commanding a power cycle of the hardware An IMU outage has severe consequences on the attitude control There are two IMUs on board but they are operated in cold redundancy This means that a switch or a restart has to be commanded unless the data flow resumes by itself Until then the ACS will use a rotation rate derived from in sequence of preference and or accuracy 1 the star trackers 2 Sun sensor and magnetometer data combined 3 the magnetic field measurements 4 Sun sensor data The rates as determined from the star tracker quaternions are just as accurate as the ones derived from the IMU In this case TET 1 fulfills the OOV requirements without any constraint A combination of data from Sun and magnetometer is the second best solution The accuracy however is so much lower that experiments such as N15 are influenced negatively In eclipse only the rotation rate derived from the magnetic field measurements is available The time derivative of all three components of the magnetic field vector can be determined However the accuracy of each of the three axes is very different2. May 05 09
3. can thus easily control the attitude of a 100 to 150 kg satellite Satellite rotation rates during N7 experiment g Ou gt aame Q w fao om ao Som 07 35 Time hh ss 07 40 Roll Pitch Yaw Figure 5 N7 experiment on DoY172 2013 The experiment started at 07 29 TET 1 was commanded into SPM Around 07 40 N7 was heated up enough to start autonomous pulsing After 4 minutes N7 was switched off and TET was commanded to nominal attitude mode SPFM It needed around 5 minutes to slew back and to compensate this deviation C N15 Infrared camera This experiment comprises three cameras to detect high temperature events on the Earth e g forest fires or volcanic activity There are two IR cameras near infrared and one optical camera Processing is done directly within the NVS German abbreviation for payload supply system This is possible because the NVS also gets attitude information from the ACS It might be necessary to reconfigure the mass memory for data storage depending on the scenery that has to be observed The experiment was designed to match the TET orbit Deviations from this Sun synchronous orbit SSO reduce the amount of interesting objects a higher altitude changes the ground resolution and therewith reduces data quality in comparison to BIRD 3 axis stabilization with nadir pointing is necessary for imaging It is possible to rotate the satellite by 30 degree to enlarge the area obs
4. the control loop but telemetry will still be available It 1s not possible any more for ACS to overwrite the internal status of the wheel as was the case in the previous software The monitoring of the RW status after the switch on is important Failure in monitoring it closely was one of the causes that led to the loss of RWI e A safe mode using the magnetic torque rods as the sole actuators was designed which will be used automatically when one of the remaining wheels starts having problems During the Sun Pointing Rotate Mode the Star Trackers are pointing towards the Earth and thus are unusable for most of the orbit The Threshold of the fuse is temperature depended An increasing temperature lowers the threshold of the fuse Current is not interrupted immediately when the threshold is reached but after a persistency The default temperature limit stored in the firmware is 100 C instead of the 80 C written in the EEPROM K There was no explicit monitoring of the status included in the nominal flight procedure for restarting of the RWs With this procedure the wheels were set into use already shortly after a power cycle Thus the ACS software could not recognize that there possibly was an EEPROM read error and a failure mode It just sent the first control commands to the wheels overruling the failure message 9 SpaceOps 2014 Pasadena California May 05 09 e If the rotation rates of the satellite are derived from magnetic field m
5. Attitude Control on TET 1 Experiences of the First Year of Operations Markus Hobsch Fabiana Cossavella Sebastian L w Jaap Herman Deutsches Zentrum fiir Luft und Raumfahrt e V German Aerospace Center M nchner Strafe 20 82234 Wessling Germany The micro satellite TET 1 carries several technology experiments It is the first in a series offering the possibility of in orbit verification of new equipment made in Germany by the industrial and scientific aerospace community TET 1 was launched 22nd July 2012 and is operated by the German Space Operations Center Attitude and attitude control is influenced by several of the experiments Special attitude control modes are required for a number of experiments in order to point the satellite in a prescribed direction or to a specific location on Earth These comprise an experiment with three infra red cameras a pico thruster and finally a new type S band transponder The Li Polymer battery was not expected to have any effect on attitude control However it was discovered that charging and discharging the battery disturbs the magnetic field sensors thus a different approach to attitude control is required when it is in use Implementation of and special demands on the attitude control system for these experiments will be presented The mission is experimental with high demands on the attitude control system The envisaged duration was one year only with a possible prolongation of a further
6. EPROM read error occurred in the internal processor of RW1 It continued working with firmware default values It continued using these obsolete values for the field orientation after it was taken back into the ACS loopk which results in a very high idle current This led to increasing temperatures of the motor and caused damage to the isolation of the winding The consequence was a short circuit in the winding which is irreversible Thus reaction wheel is definitely lost D Mitigation measures A software upgrade was installed on 3 May 2013 the prime incentive being an improved FDIR functionality The main points were e The internal error counter of the RWs will be set to zero if no error occurred over the last two hours Resets of the counter are logged This new feature was implemented in order to mitigate the operational effects of the events described in chapter III A e The ACS is now capable to restart autonomously units by a power cycle if the communication to these is interrupted The new FDIR was designed an autonomous power cycle can be defined by a new command for the units described in Table 2 It also limits the overall maximum number of reboots The new concept safeguards against RW errors as described in chapter III A Unit Trigger Persistence In ACS loop Isolationby ACS No Table 2 Units that are allowed reboot autonomously e If there is a non nominal RW mode after switch on this wheel will be excluded from
7. aporates and then escapes through a nozzle when this is opened The torque from a 1 sec pulse is lt 50 Nm Measurements with the star cameras are used to determine the effect on the attitude The Suspend Mode is needed for experiments with the pico propulsion system All ACS sensors are used but no actuators will be commanded There were two operational scenarios one in which the pico thrusters were switched on for about two minutes only in order to perform a health check and a second where the thruster were also allowed to fire It is not possible to command the start of the pulsing by command because this is governed by temperature Scenario two was tested 13 times Pico propulsion was activated either for 14 or 22 minutes In order to meet the requirements no torques are commanded to the reaction wheels and wheel desaturation is also disabled ascertaining that only the pico thrusters influence the attitude The accumulated thrust duration was approximately 4 minutes in the 14 min and 12 minutes in the 22 min experiment A maximum of 0 725 rpm rotation rate around the S C z axis was observed in the first experiment which is caused by a mean acceleration of 0 018 s and an attitude change of 5 SpaceOps 2014 Pasadena California May 05 09 123 56 again around the z axis of the S C see Figure 5 At the end of the 22 min experiment TET was accelerated to a mean rotation rate of 1 6 rpm around the z axis of the S C The pico thruster
8. due to the orientation of the satellite with respect to the magnetic field As a consequence the attitude errors can increase from 30 during phases with sunlight to 180 in eclipse see Figure 8 The IMU outages could have two possible causes 1 A freeze of the IMU 2 A link problem between IMU and ACS Tests and analyses did not provide any exhaustive explanation for the outages Introducing an additional FDIR Fault Detection Isolation and Recovery mechanism reduced the consequences and operational effort necessary after an IMU outage It simply switches the relay autonomously off and on again when communication between ACS and IMU is interrupted see chapter I D IMU outage on DOY12276 off 0 9 on 1 on e999 0 NN Aa e N 3 fe Py N 3 gt Q A 18 00 Time hh ss e IMUNOCOMP e SENDPUOK Fclips Ro Pith Yaw Status 0 Figure 8 Attitude deviations during one of the outages of IMU 1 Attitude deviations increased from 5 to 180 due to the outage In this case the rates were determined from Sun sensor and magnetometer data The noise is seen to increase by a factor of 100 during eclipses where only magnetometer data can be used In this instance there was no information from the star trackers due to blinding and too high rotation rates of the satellite C Loss of one RW The increased operational load due to outages of reaction wheels and or IMUs resulted in the loss of RW1
9. e disturbances on the MFS have no impact on the attitude determination During LEOP it was tested whether effects measured on ground can be measured in space The effect of the magnetic dipole generated by the payload battery on MFS1 can be seen in Figure 3 the effect on MFS2 in Figure 4 The analysis shows that MFS1 is not noticeably affected A small jump is seen when charging starts and stops A different behavior is observed for MFS2 which is mounted closer to the payload battery The induced moment disturbs the measurements of the magnetic field of the Earth during the entire period of battery operations An autonomous switch from the measured data to the model of the magnetic field is made after about five minutes The magnetic field vector determined by the MFS is fudged during battery charge or discharge processes Nevertheless the ACS is still compliant with the requirements The ACS usually uses the model based magnetic field vector instead of the vector derived from the measured data for attitude control The latter is only used for safety checks If the vectors differ more than 5 the algorithm will allocate the error to the model and will deactivate the use of the model The magnetic field vector derived from the data of the magnetometers will be used 4 SpaceOps 2014 Pasadena California May 05 09 Effect of the payload battery on the magnetic field sensor aB Som Q 11 40 11 50 12 00 Time
10. easurements only see Section II the ACS control will be suspended completely This FDIR was developed to reduce the load on the RWs if the rates are poorly determined IV Conclusions The TET 1 OOV mission is a success despite some severe challenges to the ACS The definite loss of one RW was caused by a fatal chain of events The thermal problems of the bus and the ensuing non nominal operations resulted in a poor determination of the attitude which in turn increased the load and temperature of the actuators leading to outages and errors New FDIRS were implemented to safely continue the mission with three reaction wheels only Board autonomous responses improved the robustness of operations and eliminated payload interruptions caused by outages of ACS units TET 1 was an experimental mission with high demands on the attitude control system The envisaged mission duration was one year only with a possible prolongation of a further year Components were therefore not chosen for longevity However the actual amount of disruptions due to IMU and or RW outages was higher than expected Experience is useful and transferable for following missions such as BIROS that will be launched in 2015 Problems caused by the design cannot be repaired in flight It was shown that more information on the status of the ACS and its units is useful to diagnose contingencies earlier and to handle them better This info will be available indeed in the upcoming BIROS mis
11. erved Pointing stability of EPM 18 offset around roll axis fAYAES a LN LA Ww 2 13 34 30 13 34 31 13 34 32 13 34 33 13 34 34 13 34 35 13 34 36 13 34 37 13 34 38 13 34 39 13 34 40 Time hh ss exe Roll exe Pitch YAW wet farcmin sec Figure 6 Stability of the spacecraft during Nadir Pointing One example of a N15 data take the rate errors in EPM is shown Once TET I reaches nadir pointing the stability remains below 1 7 arcmin s during the 10 s of the data take This experiment is essential for the upcoming scientific DLR Earth observation mission FireBird in which TET 1 and BIROS launch planned 2015 will form the space segment The attitude stability of the commanded pointing f Satellite launched in 2001 6 SpaceOps 2014 Pasadena California May 05 09 during an N15 data take is displayed in Figure 6 The pointing meets the required stability Thus the Firebird mission can be supported without any restrictions D N18 KERAMIS KERAMIS comprises three experiments probing new techniques and modules for data transmission in the Ka band Among others high frequency modules for transmission and reception in the Ka band have been qualified for use in space An S band transponder was used for a bi directional transmission of measurement data The experiment serves as a transponder providing up and down link Target pointing Mode TPM was needed for the transmission A constant signal from the ground stati
12. h space weather was found It is possible to restart the wheels not only by a power cycle relay switching but also by performing a soft reboot The latter would have solved the contingency described above and would have been better for the operational lifetime than the actually commanded power cycles It also was decided to reset the error counter weekly The second cause could be a freeze of the RW s core processing unit CPU Again this behavior had not been observed in pre launch tests and it was tried unsuccessfully to reproduce it It is most likely caused by radiation as well although no direct correlation with tests and space weather could be found A soft reset does not solve the problem and only a power cycle brings back the wheel into control loop The two different failure cases cannot be distinguished in the telemetry Thus a power cycle was commanded in all fourteen cases that occurred so far RW speed in the four and three wheel configurations O AN KORONO TARP IN AN A a gt 4 b C AF AYI DESIR MA XN PAM O i oO Q v a N oT Q 17 00 18 00 19 00 20 00 21 00 22 00 23 00 Time hh ss Figure 7 Wheel speed in the four and three wheel configurations RW3 was isolated around 20 00 commanded torque dropped to zero and its speed remained constant green line Attitude control was taken over by RWI RW2 and RW4 in a three wheel configuration leading to higher load and higher amplitude in the wheel speeds
13. hh ss PAYLOAD CURRENT o d B_meas dt d B_est dt Figure 3 Effect of payload battery on magnetometer 1 DoY207 2012 The charge current blue line axis on the left hand side of the payload battery indicates the NI switch on and switch off The time derivatives of the measured magnetic field vector and of the estimated magnetic field vector are displayed in red and green axis on the right hand side The latter is calculated by the estimator module and is used by the ACS software The magnetic field vector is also used as input for reaction wheels desaturation using the magnetorquers The MCS is taken out of the control loop during this experiment in order to avoid erroneous commanding Effect of the payload battery on the magnetic field sensor Current A 13 30 Time hh mm 13 40 PAYLOAD CURRENT e e 1 B_meas dt d B_est dt Figure 4 Effect of payload battery on magnetometer 2 DoY208 2012 The charge current blue line axis on the left hand side of the payload battery indicates the NI switch on and switch off The time derivatives of the measured and estimated magnetic field vectors are shown as red and green lines respectively scale on the right hand side B N7 Pico propulsion system The pico propulsion system is based upon the Resisto jet actuation principle It contains a propellant system from which the working fluid is led into an evaporation chamber via a propulsion feed system It is heated there ev
14. left hand side still shows one experiment which was eliminated during the mission preparation N4 Multi frequency GPS The three black cylinders on the back side of the satellite are the baffles of the IR camera N15 The sensor bus N6 and the GPS experiment N16 are mounted on top of the spacecraft On the side the GPS antenna N16 and the Li Polymer battery N1 are visible The KERAMIS payload N18 is mounted above the IR camera On the right hand side the radiator of the battery N1 on top of the GPS payload the KERAMIS antenna and the Microjet propulsion system N7 can be seen Not visible in this picture are the three solar cell experiments N2 N8 N9 which are located on the front side the HW BOSS N17 and the mass memory N19 A short description of the experiments is given in Table 1 II Observations of interactions between ACS and experiments A N1 Li Polymer battery The battery was charged or discharged and parameters such as charge current voltage and temperature were recorded Charging is of course only possible during phases with sunlight The measurements of the magnetometer especially those of MFS2 are disturbed by this experiment This is because a dipole moment is generated within the payload during charging or discharging Battery operations have a high load on the spacecrafts power system Therefore the Sun pointing mode SPFM 1s prescribed In this mode the star cameras provide the attitude measurements so th
15. of 96 minutes The bus is based upon the one used for BIRD which has already been verified in orbit ACS Subsystem Engineer and Flight Director DLR GSOC 82234 Wessling Germany Markus Hobsch dlr de gt ACS Subsystem Engineer DLR GSOC 82234 Wessling Germany Fabiana Cossavella dlr de 3 ACS Subsystem Engineer DLR GSOC 82234 Wessling Germany Sebastian Loew dlr de ACS Team lead DLR GSOC 82234 Wessling Germany Jaap Herman dlr de 1 SpaceOps 2014 Pasadena California May 05 09 antenna Payloads Low Gain Backward direction gt Forward direction Service segment High Gain antenna antenna Figure 1 TET 1 design View of the TET 1 satellite Shown are the different segments of the spacecraft the arrangement of payloads and antennae as well as the flight direction The solar arrays are seen from the back The attitude control system ACS itself is very specific because it has to meet several requirements that are coming from the different experiments For example TET 1 has no standard operational mode but uses a distinct attitude mode for each experiment The preferred ACS mode during the first year of operations was Sun Pointing Fix Mode SPFM The Sun Pointing Rotate Mode SPRM had to be used due to thermal issues for a longer test period The Earth pointing mode EPM must be commanded for ground station contacts when the high gain antenna is required EPM is no
16. ol during experiments in order to measure effects on ACS Suspend mode N8 Next generation solar cells Improve calibrations against Sun pointing specific atmospheric conditions and derive solar cell degradation due to UV and particle radiation N9 Next generation solar cells Test performance solar cells Measurements must be in by ASTRIUM are integrated in the solar Sun pointing fix mode generator and deliver power to SPFM the bus when running N15 IR camera Set of three cameras to detect Earth or target pointing high temperature events like mode EPM TPM forest fires N16 Secondary frequency GPS Precise orbit determination and Forward or backward Earth limb sounding flight direction measurements Earth pointing mode EPM system satellite to ground TPM 3 SpaceOps 2014 Pasadena California May 05 09 communication modules receiver transmitter N19 Memory orbit radiation Investigate impact of radiation none Experiment on memory units Table 1 List of experiments A short description of the purpose of each experiment is given in column two the requirements for ACS in column three N16 N4 GPS antenna N1 Radiator Secondary Frequency GPS N6 Sensor Bus N7 Microjet GPS antenna N18 KERAMIS N18 Antenna N4 Multi frequency GPS N15 IR Camera Figure 2 Payload arrangement View of the payload segment and the location of different experiments The image on the
17. on in Weilheim WHM was transmitted to the satellite Some TET 1 contacts over WHM were used for this experiment and normal up and down link had to be suspended during this experiment No real time monitoring was possible III Challenges for ACS A Data from reaction wheels flagged unhealthy On TET 1 four RW 90 2 smart wheels of the German manufacturer Astro Feinwerktechnik Adlershof GmbH are used They are mounted in a triangular pyramid configuration Several issues with the reaction wheels were a major point of concern to ACS during the first year The communication between the ACS software thread and the RWs was interrupted several times because wheels were flagged unhealthy The first response was a power cycle by switching the relay of the corresponding wheel off and on again No particular wheel could be identified as the root cause of the problem because each of the four wheels was delivering invalid data for roughly the same amount of time Detailed analysis identified two possible causes for the unhealthy data The smart wheels are designed such that a reset occurs autonomously if an inconsistency is recognized In this case an internal counter increments by one step for each occurence Data will be flagged unhealthy if the counter exceeds 20 It was not possible to reproduce this behavior and the large amount of reboots on ground so the most likely explanation is an effect of the radiation in space although no correlation wit
18. on the 222 day of the mission Since then TET 1 is flying with a three wheel configuration Commanded attitude mode in this instance was SPFM with a bias of 30 in pitch 8 SpaceOps 2014 Pasadena California May 05 09 An IMU outage occurred at a moment that the satellite was in Sun Pointing Rotate Mode SPRM n One hour and ten minutes later all four RWs were isolated by the ACS The RWs were rebooted all at once by the nominal procedure power cycle five times However at each next contact they were again not used by the ACS Recovery was only achieved by rebooting the reaction wheels 2 3 and 4 This strongly suggests that the root cause of the problem lies with RW1 Following analysis showed that it was irretrievably lost The high attitude deviations due to the IMU outage caused not only high loads on the actuators but also temperatures increasing up to 95 C on the reaction wheel bus There are thermal cutoffs embedded into the wheels Because the temperature of the whole system heated up the threshold of the fuse of RW1 decreased that much that the amount of current flow needed to operate the wheel triggered the fuse This led to a collapse of the communication interface of the whole wheel bus and an outage of all four wheels Lowering the temperature of RWI would have increased the threshold of the fuse again But the high temperatures were not visible in the TM available to the operations team In this critical phase an E
19. sion Acknowledgments The authors would like to thank Julian Gude for his careful reading of the manuscript Figures 1 and 2 were provided by Kayser Threde GmbH References Proceedings i K M ller A Kumar Balan Battery Operations for the TET 1 Spacecraft these proceedings Reports Theses and Individual Papers S Low J Herman D Schulze and C Raschke Modes and More Finding the Right attitude for TET 1 SpaceOps Stockholm 2012 K BrieB W Baerwald T Gerlich H Jahn F Lura H Studemund The DLR small satellite mission BIRD Acta Astronautica 2000 M A Othman A E Makled Evaluation of Resisto Jet thrusters Engineering model for Space Application Aerospace Sciences amp Aviation Technology 2009 S Stoltz Ereignisanalyse der Reaktionsrader im Zeitraum DOY 13060 13061 TET AFW TN 0101 Astro und Feinwerkstechnik GmbH 2013 C Raschke ACS Beantwortung ARI TET AFW TN 0507 Astro und Feinwerkstechnik GmbH 2012 T Terzibaschian DDF Lageregelung und Navigation TET AFW TN 1005 Astro und Feinwerkstechnik GmbH C Raschke Technical Note zum N1 Problem TET_AFW_TN_0097 Astro und Feinwerkstechnik GmbH 2012 i A H nle et al TET 1 Space Segment User Manual TET KTH UM 0002 Kayser Threde GmbH 2011 19 E Lorentz TET1 N15 ICD IR Kamera TET1 N15 ICD Deutsches Zentrum f r Luft und Raumfahrt 2013 10 SpaceOps 2014 Pasadena California
20. t necessary if downlink with low rate is enough B Attitude information Because TET 1 has no orbit control the determination and control of the attitude came into main focus There are five types of sensors with high redundancy on board for attitude determination on TET 1 e Star Trackers with 2 Camera Head Units CHU e Inertial Measurement Units IMU e Coarse Sun Sensors CSS e Magnetic Field Sensors MFS e Global Positioning System GPS receivers The following actuators are available for attitude control e four Reaction Wheels RW e Magnetic Coil System MCS The Star Trackers can determine the attitude most accurately An accuracy of 1 arcmin can be achieved with one camera 30 arcsec if both cameras deliver valid data The aberration correction derived from the GPS data improves these values by up to 29 arcsec The attitude can also be estimated from the last valid measurement and the rotation Xsa axis points in flight direction Zsa towards the Sun Zsa points towards the Sun rotation around the Sun vector Nadir Ponting mode Zsa pointing towards Nadir Ysa axis perpendicular to orbital plane 1 The MCS was originally foreseen for wheel desaturation only The S W update see chapter HI D made the MCS available as actuator for attitude control during ACS Safe Mode 2 SpaceOps 2014 Pasadena California May 05 09 rate derived from IMU if both cameras deliver invalid or no data This is only done for lt 200 min beca
21. use the drift of the IMU is 1 h The attitude can still be determined if only data from CSS and or MFS are available for attitude determination The Sun and magnetic field vector are determined by the On board Navigation system ONS model and compared with the measured values The model has a variance of 0 3 and a maximum error of 0 8 compared to measured values C Payload All experiments and their requirements on the ACS are summarily described in Table 1 Figure 2 shows the location of the experiments in the spacecraft A more detailed description of three interesting experiments has already been presented in Ref 2 Nevertheless some influence ACS routine and contingency operations These are the Li Polymer battery N1 the pico propulsion system N7 the infrared camera N15 and KERAMIS N18 These examples were chosen because their requirements and operational impacts on the ACS are diverse and most Stringent Payload Name Description Requirements on the ACS N1 Li Polymer Battery Charge and discharge cycles to Measurements must be in investigate behavior Sun pointing fix mode SPFM No magnetorquer commanding cells N6 Sensor bus system and Verification of application None current drain capability of a sensor bus system The current drain works as a variable electrical load for the experimental solar cells N2 N8 and N9 N7 Pico propulsion system New kind of propellant system No active attitude contr
22. year Components were therefore not chosen for longevity However the actual amount of disruptions due to sensor and or actuator outages and due to idiosyncrasies in the spacecrafts thermal budget was higher than expected The ensuing challenges for the attitude control system will be discussed A software upload in May 2013 mitigated several of the issues addressed above The improvements in the software and autonomous on board reactions will be presented in the final Section together with some recommendations for the follow on missions TET 2 and BIROS I Introduction T first Section gives a short overview of the TET 1 mission the payload experiments which were performed in the first year and their impact on attitude control The German Technology Experiment Carrier TET 1 was successfully launched on 22 July 2012 and marks the first mission within the OOV on orbit verification program The eleven payloads on board were tested in a space environment over a one year period until 31 October 2013 The mission now continues under the name Firebird primarily making IR images of the Earth A The TET 1 mission The primary goal of the mission is to offer a platform for the verification of eleven experiments ten German and one Dutch in low Earth orbit The payload comprises among others a new kind of solar cells new batteries and a new GPS receiver type The orbit is Sun synchronous with a height of 523 km and a resulting orbital period
Download Pdf Manuals
Related Search