xmm fuel estimates

J. Martin, J. Martin, XMM-Newton XMM-Newton European Space Operations CenterEuropean Space Operations Center

ESOCESOCESOCESOC

XMM fuel estimates XMM fuel estimates

J. Martin J. Martin

European Space Agency (ESA)European Space Operations Centre (ESOC)

2323rdrd March, 2010 March, 2010


ESOCESOCESOCESOCMenuMenu

• BackgroundBackground• XMM Reaction Control SystemXMM Reaction Control System• Fuel estimationFuel estimation

– Book-keepingBook-keeping– PVT (PVT (Pressure-Volume-TemperaturePressure-Volume-Temperature: applying the Ideal Gas : applying the Ideal Gas

LawLaw))

– Thermal Propellant Gauging Technique (TPGT) Thermal Propellant Gauging Technique (TPGT) (thermal knocking)(thermal knocking)

• Review and Way AheadReview and Way Ahead

• Reference DocumentsReference Documents


ESOCESOCESOCESOCBackgroundBackground

• On July 15, 2009, DP TM TD057 (tank temps. range) went OOL HH On July 15, 2009, DP TM TD057 (tank temps. range) went OOL HH with a value of 10.41 degC. Spacecraft Anomaly XMM_SC-60 raisedwith a value of 10.41 degC. Spacecraft Anomaly XMM_SC-60 raised

• Tank1 Thermostat failedTank1 Thermostat failed• OOL triggered because Tank1 temperature was more than 10 degC OOL triggered because Tank1 temperature was more than 10 degC

lower than Tanks 2, 3 and 4lower than Tanks 2, 3 and 4

• Conclusion from Industry (Astrium) was that this excursion had no Conclusion from Industry (Astrium) was that this excursion had no impact on Reaction Control Systemimpact on Reaction Control System– No possibility of pressurant gas entering pipes downstream of Tank1 No possibility of pressurant gas entering pipes downstream of Tank1

unless Tank1 nearing empty (< ~6Kg)unless Tank1 nearing empty (< ~6Kg)– Total fuel then ~80Kg (or 20Kg per tank)Total fuel then ~80Kg (or 20Kg per tank)– Lower temp encourages more fuel into Tank 1Lower temp encourages more fuel into Tank 1

• RecommendationRecommendation– Maintain Tank1 temperature control using time-tagged commands Maintain Tank1 temperature control using time-tagged commands

keeping temperature range between tanks <5degCkeeping temperature range between tanks <5degC– Try to gauge fuel distribution between tanks to better determine when a Try to gauge fuel distribution between tanks to better determine when a

tank is near emptytank is near empty


ESOCESOCESOCESOCXMM Reaction Control SystemXMM Reaction Control System

T1 T2 T3 T4

PT-2

LV-1 LV-2

PT-1

PT-3

• 4 Tanks4 Tanks– Hydrazine FuelHydrazine Fuel– Helium PressurantHelium Pressurant–T1 Main TankT1 Main Tank–T2,T3,T4 AuxT2,T3,T4 Aux

• Latch ValvesLatch Valves– LV-1 feeds Branch ALV-1 feeds Branch A– LV-2 feeds Branch BLV-2 feeds Branch B

• Pressure TransducersPressure Transducers– PT-1 System PressurePT-1 System Pressure– PT-2 Branch A Press.PT-2 Branch A Press.– PT-3 Branch B Press.PT-3 Branch B Press.

• Reaction Control ThrustersReaction Control Thrusters– 4x Branch A4x Branch A– 4x Branch B4x Branch B


ESOCESOCESOCESOCBook-keeping: PrinciplesBook-keeping: Principles• Specific Impulse (Isp) of thrusters Specific Impulse (Isp) of thrusters

knownknownIsp ~ 2250Ns/kg but varies with inlet Isp ~ 2250Ns/kg but varies with inlet

pressurepressure

• Thrust knownThrust known– Nominally 25N (at 25bar)Nominally 25N (at 25bar)– Now <10N (<7bar)Now <10N (<7bar)– Can measure with flight calibrationCan measure with flight calibration

• Fuel mass flow through thrusters can Fuel mass flow through thrusters can be derivedbe derived ..Thrust = Isp * mThrust = Isp * m

..One firing: ∆m = m. ∆tOne firing: ∆m = m. ∆ton on

= Thrust. ∆t= Thrust. ∆ton on / Isp/ Isp

• Keep running count of thruster Keep running count of thruster activations (On-times)activations (On-times)mmfuelfuel = m = mloadedloaded - ∑∆m - ∑∆m

LV-1

PT-1

.m

mfuel


ESOCESOCESOCESOCBook-keeping: Running figuresBook-keeping: Running figures

0

100

200

300

400

500

600

time

fuel

[kg]

Remaining fuel (March 2010)

76.3 [kg]

Consumption last 12 month

5.29 [kg]

average fuel consumption

(since 2003-03-01)

0.48 [kg]

residual lifetime in month

116 [-]

extrapolated milageSep 2019


ESOCESOCESOCESOCBook-keeping: ErrorsBook-keeping: Errors

• BOL =5 % (<11Kg), EOL =20% (18Kg) [RD-2]BOL =5 % (<11Kg), EOL =20% (18Kg) [RD-2]• Actually EOL is when Inlet Pressure = 8 bar. PT-Actually EOL is when Inlet Pressure = 8 bar. PT-

01 < 7bar So we are already past EOL!!01 < 7bar So we are already past EOL!!• No calibration for Thrust and Isp below this No calibration for Thrust and Isp below this

pressure. Also calibration data is based on pressure. Also calibration data is based on constant thrust. Momentum dumping is via short constant thrust. Momentum dumping is via short pulse firings.pulse firings.

• PT-01 reading need to be corrected for PT-01 reading need to be corrected for Temperature variations and also converted to Temperature variations and also converted to correct inlet pressure to use these calibrations.correct inlet pressure to use these calibrations.


ESOCESOCESOCESOCPVT: PrinciplesPVT: Principles

LV-1

PT-1

.m

mfuel

PPHeHeVVHeHe = n = nHeHeRTRTHeHe

wherewhere

PPHeHe ~ PT-01 ~ PT-01

nnHeHe number of moles of He is number of moles of He is constant. Assume 165moles constant. Assume 165moles loaded (~660g)loaded (~660g)

R gas constantR gas constant

TTHeHe = average temp of He from = average temp of He from tank thermistorstank thermistors

VVHeHe= V= Vtanktank – m – mfuelfuel/ρ/ρfuelfuel

mmfuelfuel = = ρρfuelfuel*(V*(Vtanktank- n- nHeHeRTRTHeHe/P/PHeHe))

PHe VHe THe


ESOCESOCESOCESOCPVT: Spot ChecksPVT: Spot ChecksFuel Mass: Book-keeping vs PVT

0

20

40

60

80

100

120

140

160

180

2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Time

Fu

el M

ass

(Kg

)

Mfuel (Kg) Book-keeping

Mfuel (Kg) PVT

• PVT figures (first est. from FCT) follow the book-keeping (Flight Dynamics figures)PVT figures (first est. from FCT) follow the book-keeping (Flight Dynamics figures)• No evidence of leakageNo evidence of leakage

• Error bars for Book-keeping set at +/- 18KgError bars for Book-keeping set at +/- 18Kg• PVT errors estimated to be 12Kg [RD-2] PVT errors estimated to be 12Kg [RD-2]


ESOCESOCESOCESOCThermal Gauging: PrincipleThermal Gauging: Principle

• The thermal response of the propellant tank when heated is The thermal response of the propellant tank when heated is related to the propellant load.related to the propellant load.

• Apply the Energy equationApply the Energy equation

Q.∆t = m.Cp.∆TQ.∆t = m.Cp.∆T

– Where:Where:Q is applied heat rate minus heat loss to the environment (J/s)Q is applied heat rate minus heat loss to the environment (J/s)∆∆t is the change in time (s)t is the change in time (s)m is the mass of the tank system (kg)m is the mass of the tank system (kg)Cp is the specific heat of the tank system (J/kg.K)Cp is the specific heat of the tank system (J/kg.K)∆∆T is the change in temperature (K)T is the change in temperature (K)

• Ie. A full tank should have greater heat capacity, therefore requires Ie. A full tank should have greater heat capacity, therefore requires longer heating time to reach a given ∆T than an empty tank.longer heating time to reach a given ∆T than an empty tank.


ESOCESOCESOCESOCTank Duty CycleTank Duty CycleOver lifetimeOver lifetime

– Temperature gradient Temperature gradient increasesincreases

– Temperture cycle time Temperture cycle time (peak-to-peak) reduces(peak-to-peak) reduces

– Heater Off-time (or On-Heater Off-time (or On-time) reducestime) reduces


ESOCESOCESOCESOCFuel ratios for each tankFuel ratios for each tank

• For any tankFor any tank

m.Cp = [m.Cp]m.Cp = [m.Cp]tanktank + m + mfuelfuel.Cp.Cpfuelfuel + m + mHeHe.Cp.CpHeHe

If [m.Cp]If [m.Cp]tank tank andand mmHeHe.Cp.CpHeHe neglected neglectedIf Q considered constant (ca. 14W steady heat input)If Q considered constant (ca. 14W steady heat input)ThenThen

[m[mfuelfuel]]rev-1rev-1 / [m / [mfuelfuel]]rev-nrev-n = [∆t/∆T] = [∆t/∆T]rev-1rev-1/ [∆t/∆T]/ [∆t/∆T]rev-nrev-n

Or if ∆T is constant (thermostat set-points fixed)Or if ∆T is constant (thermostat set-points fixed)

[m[mfuelfuel]]rev-1rev-1 / [m / [mfuelfuel]]rev-nrev-n = [∆t] = [∆t]rev-1rev-1/ [∆t]/ [∆t]rev-nrev-n


ESOCESOCESOCESOCTry Tank 1Try Tank 1[∆t/∆T][∆t/∆T]rev-100rev-100/ [∆t/∆T]/ [∆t/∆T]rev-1000rev-1000

• prove of prove of concept: data concept: data from rev 100 from rev 100 compared to compared to rev 1000rev 1000

• In reality due to In reality due to the quantization the quantization of HK data the of HK data the method is method is highly highly dependent on dependent on the data the data selection for the selection for the determination determination of the slopeof the slope


ESOCESOCESOCESOCFuel Ratios: Fuel Ratios: Cycle Time MeasurementCycle Time Measurement

•Method 1 Method 1 –Spot checks of Tank cycles Spot checks of Tank cycles over 10 days at 1 year intervalsover 10 days at 1 year intervals


ESOCESOCESOCESOCFuel Ratios: Fuel Ratios: Cycle Time MeasurementCycle Time Measurement

•Method 2Method 2–Statistical analysis of all temperature data over 230 days Statistical analysis of all temperature data over 230 days in 2000 and in 2008in 2000 and in 2008

• Example of Cycle Example of Cycle times times

–Green -with angle Green -with angle correctioncorrection–Red - without angle Red - without angle correctioncorrection


ESOCESOCESOCESOCCompare to Book-keepingCompare to Book-keeping

• Methods Methods agree within agree within error error marginsmargins

•Tank 2 Tank 2 supplying supplying more Fuel?more Fuel?

• But 30Kg But 30Kg fuel missing?fuel missing?

Fuel content in 2008 with respect to 2000

Tank1 Tank 2 Tank 3 Tank 4

Method 1 Spot checks 38% ± 12 25% ± 13 43%* ± 8 54% ± 8

Method 2 Statistical Analysis 47% ± 6 33% ± 6 48% ± 6 52% ± 6

* Data taken from 2009 for Tank 3* Data taken from 2009 for Tank 3


ESOCESOCESOCESOCThermal Gauging: Method 3Thermal Gauging: Method 3

• In the simplest form a single tank system can In the simplest form a single tank system can be described as be described as – a Capacitance (m.Cp) a Capacitance (m.Cp) – with a Conductive coupling to the environment with a Conductive coupling to the environment

(the XMM base plate). (the XMM base plate).

• In the steady state condition with heater off the In the steady state condition with heater off the system will be at Tmin. system will be at Tmin.

• When heat is applied at a constant rate Qin the When heat is applied at a constant rate Qin the tank temperature will increase. tank temperature will increase.

• The higher the temperature with respect to The higher the temperature with respect to Tmin the higher the heat losses Qout until Tmin the higher the heat losses Qout until eventually the tank system reaches equilibrium eventually the tank system reaches equilibrium again at Tmax with Qin = Qout.again at Tmax with Qin = Qout.


ESOCESOCESOCESOCMethod 3: System ResponseMethod 3: System ResponseTANK 1 EXT Temperatures during Solar Flare Dec 2006

15

17

19

21

23

25

27

29

31

33

0 2 4 6 8 10 12 14 16 18 20

time (hours)

Tem

p (

deg

C)

Tank1 EXT Temp Measurements

fn dT.exp(-t/tau2)

fn dT.exp(-t/tau1)T = Tmax - dT(1-exp(-t/tau2)

Tmax

dT

T = Tmin + dT(1-exp(-t/tau1)

Tmin

dT/tau1

dT/tau2

When heating from Tmin to Tmax:When heating from Tmin to Tmax: T = Tmin + ∆T (1 - exp(-t/tau1))T = Tmin + ∆T (1 - exp(-t/tau1))

∆ ∆ T = Tmax – TminT = Tmax – Tmintau1 = Time constant for tau1 = Time constant for heatingheating

When cooling the function is:When cooling the function is: T = Tmax - ∆T (1 - exp(-t/tau2))T = Tmax - ∆T (1 - exp(-t/tau2))

∆ ∆ T = Tmax – TminT = Tmax – Tmintau2 = Time constant for tau2 = Time constant for coolingcooling


ESOCESOCESOCESOCMethod 3: AnalysisMethod 3: Analysis

• For the heating case:For the heating case:

Differentiating => d/dt(T) = ∆T/tau1 (exp(-t/tau1))Differentiating => d/dt(T) = ∆T/tau1 (exp(-t/tau1))

• In the initial condition when temperature T=Tmin and time t=0 the In the initial condition when temperature T=Tmin and time t=0 the gradient is therefore simply:gradient is therefore simply:

Initial gradient = ∆T/tau1Initial gradient = ∆T/tau1

• This initial gradient is shown as the blue line (previous slide)This initial gradient is shown as the blue line (previous slide)

• If it is assumed that at this instant the heat loss is zero (Qout = 0) If it is assumed that at this instant the heat loss is zero (Qout = 0) and all heater power is therefore applied to the tank (Qin = heater and all heater power is therefore applied to the tank (Qin = heater power), then if we apply the energy equation (power), then if we apply the energy equation (see Thermal Gauging see Thermal Gauging Principle slidePrinciple slide) we have:) we have:

Qin / m.Cp = ∆T / tau1Qin / m.Cp = ∆T / tau1


ESOCESOCESOCESOCMethod 3: AnalysisMethod 3: Analysis

• We now have to assume the mass and heat capacity of the Helium gas in We now have to assume the mass and heat capacity of the Helium gas in the tank can be neglected. If so then we can write:the tank can be neglected. If so then we can write:

m.Cp = [m.Cp]m.Cp = [m.Cp]tanktank + m + mfuelfuel.Cp.Cpfuelfuel = Qin.(tau1/∆T) = Qin.(tau1/∆T)

• or rearranging:or rearranging:

mmfuelfuel = ( Qin.(tau1/∆T) - [m.Cp] = ( Qin.(tau1/∆T) - [m.Cp]tank tank ) / Cp ) / Cpfuelfuel

• Simply by estimating tau1 and ∆T (Graphical best fit), it was possible to Simply by estimating tau1 and ∆T (Graphical best fit), it was possible to arrive at absolute values for the fuel in each tank.arrive at absolute values for the fuel in each tank.

• In practice it was difficult to find times where data was stable enough to In practice it was difficult to find times where data was stable enough to make a fit: Best data from periods of Solar Flares where S/C in same make a fit: Best data from periods of Solar Flares where S/C in same pointing for number of dayspointing for number of days


ESOCESOCESOCESOCMethod 3: ResultsMethod 3: Results

• Surprising ResultsSurprising ResultsTank 2 (as before) appears to have Tank 2 (as before) appears to have least fuelleast fuel

Tank 4 has 2 to 3 times more fuel Tank 4 has 2 to 3 times more fuel than Tank 1than Tank 1

Tank 1 (most important) looks OKTank 1 (most important) looks OK

2001 TANK1 TANK2 TANK3 TANK4

dT 11 5.2 6 8.5 DegC

tau1 18 3 4.3 32 hrs

tau2 13 2.5 4 27 hrs

Qin/m.Cp 0.61 1.73 1.40 0.27 degC/hr

Qout/m.Cp 0.85 2.08 1.50 0.31 degC/hr

m.Cp 20.95 7.38 9.17 48.19 W.hr/K

Qout 17.72 15.36 13.76 15.17 W

Mass fuel 22.49 6.79 8.88 54.52 Kg

Total 92.68 Kg

2009 TANK1 TANK2 TANK3 TANK4

dT 13 6.5 6 6.4 DegC

tau1 9.5 1.4 3 16 hrs

tau2 5.8 1.3 2.8 12 hrs

Qin/m.Cp 1.37 4.64 2.00 0.40 degC/hr

Qout/m.Cp 2.24 5.00 2.14 0.53 degC/hr

m.Cp 9.35 2.76 6.40 32.00 W.hr/K

Qout 20.97 13.78 13.71 17.07 W

Mass fuel 8.93 1.38 5.64 35.58 Kg

Total 51.53 Kg


ESOCESOCESOCESOCThermal Gauging vs Book-keepingThermal Gauging vs Book-keeping• Still 30 Kg missing…Still 30 Kg missing…


ESOCESOCESOCESOCAstrium reviewAstrium review

• In February 2010 FCT sent Fuel Mass Calculation document [RD-03] In February 2010 FCT sent Fuel Mass Calculation document [RD-03] to Industry (Astrium) for reviewto Industry (Astrium) for review

• March 2010 Detailed Analysis received [RD-04]March 2010 Detailed Analysis received [RD-04]• No need for alarm, the assymetry in fuel mass could not occurNo need for alarm, the assymetry in fuel mass could not occur

– Errors in initial loaded could account for ~2Kg difference between tanksErrors in initial loaded could account for ~2Kg difference between tanks– Fuel / pressurant gas migration between tanks during initial fuel Fuel / pressurant gas migration between tanks during initial fuel

orientation manoeuvres (LEOP) could also account for ~2Kg differenceorientation manoeuvres (LEOP) could also account for ~2Kg difference– Since LEOP the only way to account for more significantly fuel in Tank4 is Since LEOP the only way to account for more significantly fuel in Tank4 is

a leakage of pressurant gas (a leakage of pressurant gas (loss of 3bar!! Would be observable – loss of 3bar!! Would be observable – but is notbut is not))

– This is ruled out since a leakage is not observable in PVT analysis nor in This is ruled out since a leakage is not observable in PVT analysis nor in analysis of disturbance torques (Flight Dynamics)analysis of disturbance torques (Flight Dynamics)

• Two explanations for the assymetry calculatedTwo explanations for the assymetry calculated– Tank 4 is better insulated (this tank is on the cold side)Tank 4 is better insulated (this tank is on the cold side)– Temperature measurments are taken from thermistors placed near Temperature measurments are taken from thermistors placed near

middle of Tanks (in contact with He) rather than the top (in contact with middle of Tanks (in contact with He) rather than the top (in contact with Fuel)Fuel)


ESOCESOCESOCESOCWay AheadWay Ahead• Book-keeping Book-keeping

– – seeking better ground calibration data for thruster performance seeking better ground calibration data for thruster performance - Which temperatures to use for correcting pressure variations- Which temperatures to use for correcting pressure variations

• Tank gauging (Thermal knocking) requires a better tank model (thermal properties, Tank gauging (Thermal knocking) requires a better tank model (thermal properties, location of fuel, location of thermistors) and then calibration against flight telemetrylocation of fuel, location of thermistors) and then calibration against flight telemetry

Comparison of Relative Accuracies [RD-7]Comparison of Relative Accuracies [RD-7]


ESOCESOCESOCESOCWay Ahead – Way Ahead – Thermal KnockingThermal Knocking

• Flight Calibration?Flight Calibration?• Perhaps TANK 4 is best suitedPerhaps TANK 4 is best suited


ESOCESOCESOCESOCReferencesReferences

RD-1 Assessment of XMM Anomaly XMM_SC-60 and Recovery Action, Astrium, 21/07/2009RD-1 Assessment of XMM Anomaly XMM_SC-60 and Recovery Action, Astrium, 21/07/2009

RD-2 RD-2 RCS SSD - Annex B: Fuel Book-keeping Accuracy for XMM, RCS SSD - Annex B: Fuel Book-keeping Accuracy for XMM, XMM-MOC-SSD-0019-OAD, XMM-MOC-SSD-0019-OAD, March 2002March 2002

RD-3 XMM-Newton Fuel Mass Calculation, J.Martin, Draft, 01-02-2010RD-3 XMM-Newton Fuel Mass Calculation, J.Martin, Draft, 01-02-2010

RD-4 Assymetric Tank Depletion, Astrium, 11-03-2010RD-4 Assymetric Tank Depletion, Astrium, 11-03-2010

RD-5 XMM-Newton RCS Fluid Dynamics Analysis, XM-TN-BPD-0013RD-5 XMM-Newton RCS Fluid Dynamics Analysis, XM-TN-BPD-0013

RD-6 XMM-Newton RCS Flight Model Users Manual, XM-TN-BPD-0005RD-6 XMM-Newton RCS Flight Model Users Manual, XM-TN-BPD-0005

RD-7 RD-7 Comparative Assessment of Gauging Systems and Description of a Liquid Level Comparative Assessment of Gauging Systems and Description of a Liquid Level Gauging Concept for a Spin Stabilized Spacecraft, Hufenbach.B. 1997ESASP.398..561HGauging Concept for a Spin Stabilized Spacecraft, Hufenbach.B. 1997ESASP.398..561H

RD-8 RD-8 Flight Validation of the Thermal Propellant Gauging MethodFlight Validation of the Thermal Propellant Gauging MethodUsed at EADS Astrium, Used at EADS Astrium, L. Dandaleix, 2004ESASP.555E...9DL. Dandaleix, 2004ESASP.555E...9D