drag free and attitude control g. sechi tas-i (buoos)eotvos.dm.unipi.it/tasi gg presentation...

All rights reserved © 2009, Thales Alenia Space

BU OOS1

Drag Free and Attitude Control

G. SechiTAS-I (BUOOS)


BU OOS2

DFAC – Functional requirements

The Drag-Free and Attitude Control (DFAC) sub-system has in charge:

the spacecraft attitude control after launcher separation, in order to guarantee safe power and communication conditions;spacecraft spin-up to achieve the spin rate required by scientific observation (360deg/s) (spin axis aligned with the normal to the orbital plane);stabilization of PGB-satellite relative position (whirl control);drag compensation with very high rejection (< 1/50000) to permit the detection of the EP violation (if any);to keep PGB-spacecraft relative spin angle and attitude safe enough for PGB suspension, PGB sensors and actuators integrity.

Above objectives have been achieved considering active and passive attitude control solutions.


BU OOS3

DFAC mode architecture

The AOCS mode architecture has been designed to be compatible with both the separation modes provided by the VEGA launcher:

Spin stabilized mode (up to 30 deg/s, nutation angle half cone < 5deg)Three axis stabilized mode (de-pointing < 1.5deg, angular rate < 1.0 deg/s)

This choice allows us to share the launch with another companion satellite.

In spin-stabilized mode, after separation the attitude control system recovers the spin axis pointing, and it maintains a significant gyroscopic stability (the spacecraft is not stable due to gravity gradient; fuel saving).

In three-axis stabilized mode, after separation the attitude control system damps the angular rate, and points the satellite spin axis as close as possible to the Earth magnetic field (close to the Earth Mean Rotation axis at the epoch). A small spin-rate around this axis is introduced (fuel saving).


BU OOS4

DFAC mode organization

Rate Damping reduces the residual angular rate after separation;Coarse Pointing Mode points the satellite Z-axis aligned with Earth magnetic field (SAA<35deg), with a reduced spin-rate around Z-axis.Fine Pointing Mode permits improvement in spin-axis pointing tanks to the use of Sun sensor, orbit propagator, Earth magnetic field model, Sun propagator.In Spin-Up Mode, the satellite Z-axis pointing and the Z-axis angular rate are controlled (< 2deg, < 1deg/s). In Satellite Spinning Mode, the spacecraft is not controlled at all (spin stabilized, magnitude of the angular momentum equals to 950Nms (1Hz spin rate)).Drag-Free Mode (organized in sub-modes) provides all the functionality required by the scientific observation.

SBM Stand-By Mode SUM Spin-Up Mode

RDM Rate Damping Mode SSM Satellite Spinning Mode

CPM Coarse Pointing Mode DFM Drag-Free Mode

FPM Fine Pointing Mode

With PRIMA heritage only the grey box must be developed


BU OOS5

DFAC equipment

The equipment used for satellite attitude control are:RDM : GYRO, Cold-gas on/off thruster (MGM only for FDIR)CPM : MGM, GYRO, Cold-gas on/off thrusterFPM : MGM, SS, GYRO, Cold-gas on/off thrusterSUM : MGM, SS, GYRO, Cold-gas on/off thrusterSSM : MGM, SS, GYRO (for attitude determination, no-control)DFM : specific equipment described in next viewgraphs.

MGM Magnetometer

SS Sun Sensor

GYRO Gyroscope


BU OOS6

DFAC – Architecture

RDM, CPM and FPM are “classical” AOCS control modesIn FPM the calibration of the three-axis magnetometers shall be done, to achieve the required Z-axis pointing accuracy. The calibration permits to recover bias error, scale factor error, non-orthogonality error.

Spin-Up Mode covers a “critical” phase for the mission. It has in charge of:the accurate pointing of the spin-axis (<1deg half-cone with respect to the normal to the orbital plane);spin rate up to 360deg/s (1 Hz).

It is considerate “critical” because it is very expensive in terms of fuel consumption (950Nms is the final satellite angular momentum magnitude). So, it is necessary to implement fine attitude determination scheme, effective FDI for monitoring.

The attitude determination is based on MGM, SS, Earth magnetic field model and orbit propagator, Sun propagator. MGM shall be sampling up to 20Hz. Thruster: 0.3-0.5 N, MIB < 0.005 Ns.


BU OOS7


After spin-up phase is concluded, the satellite spin axis drifts as result of the following perturbation sources:

orbital plane precession;solar pressure;atmospheric drag;gravity gradient;residual magnetic dipole;eddy current;residual torque left by thrusters assembly.

Residual magnetic and residual torque left by thrusters assembly (thrust direction de-pointing, FEEP failure) are potentially the major responsible of the drift. In particular, the major impact is due to the 1Hz components of the residual magnetic and residual torque.Requirement has been provided on residual magnetic, distance between COP/COA and satellite COM in order to reduce spin-axis drift.Mechanical stops shall cope with tilt between PGB spin axis and satellite spin-axis up to 15÷20deg.


BU OOS8


Spin axis evolution for 1.5 year mission (residual perturbing torque zero at the start of the mission).

COP/COA – COM distance < 0.035m

Satellite de-spin negligible


BU OOS9


Spin axis evolution for 1.5 year mission after one FEEP fault (residual perturbing torque not zero at the start of the mission).

COP/COA – COM distance < 0.035m

Satellite de-spin negligible


BU OOS10


No specific maneuver for re-pointing has been foreseen. This starts from:in 1 year mission, the de-pointing is still acceptable with scientific objectives. Special care will be put in COM positioning (<0.035m), residual dipole (requirements on magnitude of first harmonic).mission extension will be conditioned to de-pointing, fuel availability.

The foreseen cold-gas fuel mass is still compatible with one re-pointing maneuver .


BU OOS11


Eddy current (1/2)

Since the satellite moves with respect to the magnetic field vector, torques caused by induced currents (named eddy currents but also known as Foucault current) must be considered. They produce dissipative torques (damping torques).

Accurate assessment of these torques is extremely difficult and simplifying assumptions regarding the shape of the spinning section, its electromagnetic properties (electrical conductivity), and the nature of the interaction with the ambient field are required to approximate the magnetic field effects.The instantaneous value of the eddy current damping torque on a body whose angular velocity vector is in a magnetic field is

where is a constant that depends on the geometry and electrical conductivity of the rotating object.

( )BωBT ××−= ek

ek


BU OOS12


Eddy current (2/2)

One day-average magnetic field

Preliminary assessment shows that de-spin is not significant for satellite. In any case, foreseen spin-rate control recovers any rate variation with respect to PGB (satellite follows the PGB).

De-spin effect in not negligible on PGB since its external cover is constituted by aluminium honeycomb and carbon fibers, with mu-metal shield inside. Since the PGB inertia moment around spin axis is in the order of 1kgm2, in order to limit the PGB de-spin to 10% in two years mission then .

Preliminarily and considering the PGB external cover like a cylinder with length 0.5m, radius 0.3m, the mu-metal layer shall not be larger than 0.1÷0.2 mm (improvement from multi-layer solution (cascade) is expected).

⎥⎥⎥

⎦

⎤

⎢⎢⎢

⎣

⎡

+−=

22

00

YX

SeO

BBk ωT

2XB 2

YB

4.8623e-011 4.9997e-011

7.15<ek


BU OOS13

Drag Free Mode

Drag-free Mode has the following functional requirements:the stabilization of the relative displacement in the plane XY, limiting the magnitude;the rejection of any disturbances (drag is the expected major one, but others are solar pressure, thrusters noise, etc.) at spinning rate (1Hz) on overall axes;the stabilization of the spin rate versus variation of the spacecraft inertia moment due to thermo-elastic deformations. It is necessary to preserve the integrity of the suspension and sensors.

The rejection of drag at 1Hz in XY plane shall be greater than 1/50000.

In order to meet the above functional requirements, a quite complex control architecture has been designed.


BU OOS14

Drag Free Mode

Drag-free block diagrams


BU OOS15

Drag Free Mode

The more challenging controllers is the drag-free in the XY plane.The complexity starts from the high rejection at high frequency (1Hz), considering the response time of the available thrusters (> 30ms).

100

102

104

106

10-6

10-5

10-4

10-3

10-2

10-1

100

101

Frequency - [Hz]

Rej

ectio

n

Rejection comparison for different controllers

PIDPI2DPI3DRequirement

In order to provide an overview of the control complexity, it is possible to show what it is possible to obtain using “simple”control structure:

PID : 3rd order controllerPI2D : 4th order controllerPI3D : 5th order controller

The sample frequency impacts on selected hardware (thrusters, sensors, computation capability). The objective is to reduce the sampling frequency, to leave room as much as possible to available technologies.


BU OOS16

Drag Free Mode

More complex algorithms have been designed in order to achieve the required performances with still acceptable hardware requirements. The considered control solution is based on the integration of a sort of notch-filter (actually a harmonic oscillator in the state observer to model periodic drag).The limited ratio between sampling frequency (10Hz) and the frequency at which the disturbance occurs (1Hz) has required a special care in the derivation of the plant discrete model embedded in the observers.The controller have been designed according to state-space approach based on the state observer and gain feedback functions.

All other controllers have been designed according to the same methodology, used in TAS-I since several years (GOCE is the first relevant application ofsuch a method for control design).


BU OOS17

Drag Free Mode

Notch filter (harmonic oscillator in the state observer) introduces new relevant requirement.

10-1 100 10110-5

10-4

10-3

10-2

10-1

100

101Sensitivity - 1/Rejection

Frequency - [Hz]

Mag

nitu

de

1.998 1.9985 1.999 1.9995 2 2.0005 2.001 2.0015 2.00210-8

10-7

10-6

10-5

10-4

10-3

10-2Sensitivity - 1/Rejection

Frequency - [Hz]

Mag

nitu

de

0.999990 0.999995 1.000000 1.000005 1.000010 1.000015 1.000020 10-9

10-8

10-7

10-6

10-5

10-4Sensitivity - 1/Rejection

Normalized frequency - []

Mag

nitu

de

The knowledge about the angular rate shall bein the order of 10-5 |ω|.


BU OOS18

Drag Free Mode

Drag-free controller in XY plane has been designed considering the model in the inertial reference frame (significant reduction of numerical problems):

The controller commands the required force in an inertial reference frame;The actual thruster commands (in body reference frame) are computed by modulation starting from above commanded force ;The acquired measurements (in body reference frame) are reported in inertial reference frame by de-modulation.

Modulation and demodulation functions are neededRate information is used to tune the observers in Z drag-free and whirl controllers, and for modulation/demodulation functions (XY drag-free controller).


BU OOS19

Whirl and other controllers

To stabilize the whirl movement it is enough to introduce a simple control action proportional to the inertial XY linear velocity. It means that whirl control needs the knowledge of satellite angular rate too.Drag-free on Z axis does not provide additional difficulties.In order to maintain the relative rotation between PGB and satellite limited in time (Sun-eclipse transition), it is necessary to add the spin-rate control.

Spacecraft & Environment

Spin-rate controller

Thruster Commanded Torque

Relative rotation between spacecraft and PGB around Z axis (∆θZ)


BU OOS20

DFM sensors

DFM requires specific additional sensors for PGB – satellite relative linear position and angular rotation, and satellite rate measurement.They are capacitor sensors (sampling frequency 10Hz):

X-Y position:bias (not relevant)noise 0.01÷0.05 µm/√Hz (*)

Z position:< 10-3m (not critical)noise 0.01÷0.05 µm/√Hz (*)

Z rotation:< 10-3radnoise 0.1 µrad/√Hz (*)

(*) lower value permits fuel saving.

Rate sensorIn the frame of the study, a specific optical rate sensor has been designed.The angular rate shall be known with relative accuracy in the order of 10-4÷10-5, with FOV > ±40 deg.Details will be provided during tomorrow presentation.


BU OOS21

DFM actuators

Thruster assemblyMain forces and torque to be compensated are:

force on XY plane and Z axis for drag-free;torque on Z axis for spin rate control.

Two assemblies have been considered (6 and 8 (baseline) thrusters)


BU OOS22

DFM actuators

Thruster requirements (1Hz)

No Parameter Unit Value Comments

1 Maximum thrust µN >=150 50% margin

2 Max thruster response time[1] ms 40 @ commanded step (up and down) >= 60 µN

3 Resolution (quantization) µN 24 TBC, not critical

4 Max noise µN/√Hz 18 Around 1Hz

5 Scale factor error % 12 Peak

6 Update com rate Hz 10 TBC

7 Total impulse Ns 4500 20 % margin

8 Minimum thrust µN <=10 TBC

9 Vector stability rad 0.17 Peak, at 60 µN

10 Centrifugal acceleration g <4.4 20 % margin, 0.75m spacecraft radius

[1] Thrust response time is defined as the time required to achieve the 90% of the commanded step, and to remain definitively over this threshold.


BU OOS23

DFM actuators

Two technologies have been considered, both leaded by Italian industries:Field Emission Electrical Propulsion (FEEP) from ALTA S.p.A.;Cold gas propulsion system (CGPS) from TAS-I S.p.A.

FEEP Thrusters are being developed for the ESA Lisa Pathfinder (LPF) mission and the CNES Microscope mission. Thruster development is nearly completed, and the preparation of the Lisa Pathfinder FEEP Cluster Assembly (FCA) Qualification Model is ongoing. Manufacturing of FM parts for LPF was also released.

GAIA Cold-gas Micro Propulsion system (GCPS), currently under qualification at TAS-I, represents the reference design and technology starting points for configuring/realizing both Microscope and Galileo Galilei.


BU OOS24

FEEP status

No Parameter Unit Value FEEP status

1 Maximum thrust µN >=150 Thruster is designed and currently being qualified for a maximum thrust of 150 µN.Command capability is, at present, greater than 204.8 µN, and thrust up to 540 µN was recorded during one test

2 Max thruster response time ms 40 Current response time (for 60 µN step from 0 to 60 µN) is about 80 to 150 ms, (depending on thrust and up or down command), with command frequency at 10 Hz.Step response can be improved up to 30-40ms reducing internal delay, fall time, by biasing minimum thrust (e.g. working with thrust higher than 70 µN) and/or adding some internal dissipation.

3 Resolution (quantization) µN 24 Thruster/PCU are designed and currently being qualified for a thrust resolution of 0.1 µN

4 Max noise µN/√Hz 18 The thruster is being qualified for 0.03µN/√Hz (range 0.006 to 5 Hz)


BU OOS25

FEEP status

5 Scale factor error % 12 PCU allows scale factor correction and re-calibration with a 12 bit resolution (individual command correction).Requirement is not deemed critical.

6 Update command rate Hz 10 Already available for Lisa Pathfinder

7 Total impulse Ns 4500 Thruster is designed vs. a requirement of 2900 Ns (Lisa Pathfinder). Life test (on QM) will be performed up to 1100 Ns (with possible extension to higher total impulse). Analysis will be performed to predict EOL performance. At present, > 1000 Ns were verified at EM level.

8 Minimum thrust µN <=10 Thruster is designed and currently being qualified for a minimum thrust of 0.3 µN.

9 Vector stability rad 0.17 For thrust greater than 10 µN is always met.

10 Centrifugal acceleration g <4.4 Not met by current design. Modification of thruster design, and, in particular, of tank position and shape, to minimize hydrostatic head will permit to achieve the requirement.


BU OOS26

Cold-gas status

No Parameter Unit Value CGPS Status

1 Maximum thrust µN >=150 Thrust levels up to 500 mN achievable

2 Max thruster response time ms 40 about 100 ms: commanded thrust level below 50 mN: 100 to 200 ms: commanded thrust level in the 50 to 500 mNrange

3 Resolution (quantization) µN 24 1 µN achievable with the current GAIA Design

4 Max noise µN/√Hz 18 1 µN/√Hz from 0.01 Hz to 1 Hz0.045 µN/√Hz from 1 Hz to 150 Hz achievable with GAIA design

5 Scale factor error % 12 1 for GAIA

6 Update command rate Hz 10 1 Hz for GAIA


BU OOS27

Cold-gas status

7 Total impulse Ns 4500 Same Total impulse figure required for GAIA700 million cycles at 10 Hz, in open loop, performed on the TV EM

8 Minimum thrust µN <=10 1 µN achievable with the current GAIA Design

9 Vector stability rad 0.17 No data available at the moment, not critical

10 Centrifugal acceleration g <4.4 No risk of valve opening induced by the centrifugal force has been recognized. In fact, the centrifugal force (0.174 kg) is lower than the spring strength (1 kg).


BU OOS28

DFM actuators

ESA Lisa Pathfinder FEEP are almost in line with required response time, but need modifications of tank positioning and shape to meet the requirement on maximum centrifugal acceleration. Additional activities during Phase B are needed to extend the performances of already available LPF FEEP to GG FEEP. GAIA CGPS are compliant with required maximum centrifugal acceleration, but need modifications on electronic box and control algorithms to meet the requirement on response time. Additional activities during Phase B are needed to extend the performances of already available GAIA CGPS to GG CGPS.Both manufacturers do not recognize problems to meet the requirements pending the additional activities above indicated.

FEEP has been considered as baseline since they are closer to our requirements. Nevertheless, cold-gas solution is attractive for synergy (same technology for coarse mode and drag-free mode).


BU OOS29

DFAC simulator

In the frame of the study, a reduced numerical simulator (DFAC-sim) for drag and attitude control design issues and performance analysis have been developed. It implements simplified models but still representative for control design. The non-linear model is constituted by 25 state variables:

XY PGB-satellite linear relative movement (4 state variables);Z PGB-satellite linear relative movement (2 state variables),Z PGB-satellite angular relative displacement (2 state variables);inertia around Z axis (to take into account Sun-eclipse transition);thrusters dynamics (8 thrusters, 2 state variables for each thrusters).

The attitude dynamics for X and Y axes have not been taken into account due to their vary slow variation (very high gyroscopic stability around Z-axis) with respect to dynamics required by drag-free, whirl and spin-rate controllers (several orders of magnitude). No coupling exists.


BU OOS30

DFAC simulator

Thruster model takes into account:Slew-rateMinimum and max thrust valuesNoise profileResponse timeQuantizationScale factorsMounting and thrust direction uncertainties

Sensor model takes into account:Noise profileBiasScale factor

10-1 100 10110-5

10-4

10-3

10-2

10-1

100

101Sensitivity - 1/Rejection

Frequency - [Hz]

Mag

nitu

de


BU OOS31

Simulation results

Thrusters:noise = 6 µN/√Hzresolution = 6 µNscale factor error < 5%, uncertainty on thrust direction < 5deg

PGB-spacecraft relative position:bias= 1 µmnoise= 0.01 µm/√HzScale factor=1e-4

PGB-spacecraft relative rotation:bias= 1 µradnoise= 0.1 µrad/√HzScale factor=1e-4

Spacecraft mass = 500kgPGB mass = 45 kgPGB suspension

oscillation period = 150sQ= 90


BU OOS32

Simulation results

Simulated perturbing force in XY plane (peak acceleration = 0.2 10-6 m/s2)

Time series Spectrum


BU OOS33

Simulation results

Whirl controller not activated, drag-free controller not activated (1/2)

Perturbing force (XY plane, body reference frame)


BU OOS34

Simulation results

Whirl controller not activated, drag-free controller not activated (2/2)

PGB-spacecraft relative position


BU OOS35

Simulation results

Whirl and spin-rate controllers activated, drag-free controller not activated(1/2)

PGB-spacecraft relative position (angular rate relative accuracy = 10-4)


BU OOS36

Simulation results

Whirl and spin-rate controllers activated, drag-free controller not activated(2/2) (Zoom in)

One-side spectral density of the PGB-spacecraft relative position.


BU OOS37

Simulation results

Whirl, spin-rate and drag-free controllers activated– No thruster and measurement noises (1/2)

PGB-spacecraft relative position (angular rate relative accuracy = 10-4)


BU OOS38

Simulation results

Whirl, spin-rate and drag-free controllers activated – No thruster and measurement noises (2/2)

PGB-spacecraft relative position (angular rate accuracy = 10-4)

Rejection better than 1/100000


BU OOS39

Simulation results

Whirl, spin-rate and drag-free controllers activated (all noises present)

PGB-spacecraft relative position (angular rate accuracy = 10-4)


BU OOS40

Conclusions

GG DFAC architecture (functional organization, operating modes, equipment selection) has been defined.

The design of the control algorithms for fine drag compensation, whirl control and spin rate has been completed. Results from simulation clearly show that:

the proposed solutions permit to meet requirements with margins considering available technologies. during Phase A2 of the study, improvements in the control performances have been achieved.

Minor open points are still present on thrusters’ performances (particularly for response time (cold-gas), and maximum centrifugal acceleration (FEEP)). According to thrusters manufacturers, the still open points may be solved pending additional activities to be done in early Phase B.No show stopper have been envisaged from DFACS side.

drag free and attitude control g. sechi tas-i (buoos)eotvos.dm.unipi.it/tasi gg presentation...

Documents