-
Lithium Ion Battery Management Strategies for
European Space Operations Centre Missions
Thomas Ormston, Viet Duc Tran, Michel Denis and Nic Mardle
European Space Agency, Darmstadt, Germany
Luke Lucas and Laurent Maleville
LSE Space GmbH, Darmstadt, Germany
Kees Van Der Pols
Telespazio VEGA Deutschland GmbH, Darmstadt, Germany
Eective battery management on a space mission is one of the key
factors in ensuringmission success and longevity. Given the
reliability of modern spacecraft, the unavoidableageing of
batteries can become a critical life-limiting factor. To improve
this, it is necessaryto have a strategy for management and
monitoring of spacecraft batteries that is tailoredto both the
mission prole and the battery technology in use. This paper will
focus onseveral missions own from the European Space Agency's (ESA)
European Space Opera-tions Centre (ESOC) in Darmstadt, Germany. The
main case studies in this paper focuson missions that regularly use
their Lithium Ion batteries, although a summary of othermissions
that contain Lithium Ion batteries will also be presented. Lithium
Ion batteriesare currently the prevailing battery technology in use
on current and future EuropeanSpace Agency missions.
The paper will begin with an overview of the Lithium Ion battery
technology thathas largely replaced all others for modern space
batteries. Their proper managementrequires dierent techniques
compared to previous space battery technologies; for
instancecompared to the previous Nickel-Cadmium technology, Lithium
Ion battery deep dischargesshould be avoided where possible - which
increases the risk of using deep discharges tomeasure degradation.
The paper will describe the characteristics and inuencing factors
ofLithium Ion battery degradation, along with an overview of
research aimed at prolonginglifetime of the batteries. The paper
will also summarise methods available in order tomeasure the
absolute or relative degradation of Lithium Ion batteries and the
limitationsof these methods based upon the capabilities of each
spacecraft and the mission prole.
The paper will then detail the actual operational implementation
of this information ontwo representative ESA missions. The rst case
study will be Mars Express, which hasbeen ying three Lithium Ion
batteries for ten years and using them for prolonged eclipseseasons
2-3 times per year. The power demand of the spacecraft is high and
the availablemargin in the power system is low, therefore modelling
and management of the batteriesis critical to the mission. The
second case study will be ESA's CryoSat-2 mission, whichhas been
ying a single Lithium Ion battery for 4 years. The battery is
younger, and the
Spacecraft Operations Engineer, Earth Observation Missions -
EarthCARE (HSO-OER), European Space Operations Cen-tre, 64293
Darmstadt Germany.
Spacecraft Operations Engineer, Earth Observation Missions -
Aeolus (HSO-OEA), European Space Operations Centre,64293 Darmstadt
Germany.
Spacecraft Operations Manager, Planetary Missions - Mars Express
(HSO-OPM), European Space Operations Centre,64293 Darmstadt
Germany.
Spacecraft Operations Manager, Earth Observation Missions -
CryoSat-2 (HSO-OEE), European Space Operations Centre,64293
Darmstadt Germany.
Spacecraft Operations Engineer, Planetary Missions - Mars
Express (HSO-OPM), European Space Operations Centre,64293 Darmstadt
Germany.
Spacecraft Operations Engineer, Earth Observation Missions -
SWARM (HSO-OEW), European Space Operations Centre,64293 Darmstadt
Germany.
1 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
SpaceOps 2014 Conference 5-9 May 2014, Pasadena, CA
AIAA 2014-1883
Copyright 2014 by European Space Agency. Published by the
American Institute of Aeronautics and Astronautics, Inc., with
permission.
SpaceOps Conferences
-
power system has more margin but eclipse seasons are an almost
constant feature of theroutine mission (albeit with varying
duration eclipses). In addition, the satellite ies in
anon-sun-synchronous orbit, which makes the assessment of the
expected state of batterycharge more dicult. An overview of the
techniques used on other ying ESOC missionswill also be presented
(Herschel, Planck, GOCE, Venus Express and Rosetta).
The paper will describe new operations that have been introduced
to manage the degra-dation of the batteries, including specially
designed settings that, while respecting the al-lowed usage prole
of the battery, modify the charge and discharge management
strategiesand other ight operations to almost halve the rate of
degradation compared with theworst-case design assumption. In
addition, the methods used by each mission to assessabsolute and/or
relative battery degradation in ight will be discussed.
The paper will conclude with an overview of the lessons that
have been learnt so far atESOC from missions ying Lithium Ion
batteries. These lessons could be used as a modelfor current and
future operators of spacecraft with Lithium Ion batteries on how to
bestmanage their batteries for longevity, mission reliability and
success.
I. Introduction
The majority of ESA missions now use Lithium Ion batteries as
their primary method of power storage.This battery technology has
many advantages for spacecraft and mission design and is foreseen
for manyof the upcoming missions to be own by ESOC too. However,
although they are an accepted and welcomeadvancement for spacecraft
design, the operational use of Lithium Ion batteries in ight is
still a relativelynew area.
It is becoming increasingly common that spacecraft will survive
long past their design life and as suchthe proper operational
management of Lithium Ion batteries from day one of the mission,
including prior tolaunch, is critical. This eectively requires two
components - determining the level of battery degradationand
reducing the rate of battery degradation.
At present the strategies for this management have been
developed largely as a benet of ight experience,and often
independently from one another due to the dierent generations of
missions and the dierentdemands and constraints of various
missions. This paper aims to summarise Lithium Ion battery
technologyas a primer to operators working with the technology and
to further go on to highlight some diering butrepresentative
examples of Lithium Ion battery management at ESOC. It will
conclude with a summary ofsome lessons learned during the time ESOC
has been operating spacecraft with Lithium Ion batteries.
II. Space Battery Technology
For the majority of spacecraft, whether scientic, communication
or other purpose, in LEO, MEO, GEOor interplanetary, power and its
supply at all times is a matter of careful consideration.
During sunlight illumination power is normally provided by solar
arrays, however spacecraft in all ofthe orbits mentioned will
experience solar eclipses and need a secondary power source for
these periods.Rechargeable batteries have been used for this
purpose as they are particularly applicable to long
durationmissions. Non-rechargeable energy sources e.g. primary
batteries for short missions or radioisotope ther-moelectric
generator power sources for deep space missions are not within the
scope of this paper. Whilerechargeable batteries have been a common
part of spacecraft systems since early in the space age, thebattery
chemistry has undergone signicant changes.
A battery is made up of a number of cells in series and
parallel. The arrangement of the cells providesthe required current
and voltage. The voltage of any cell or battery depends upon the
electrochemistry ofthe cell. It is here that changes in the cell
electrochemical make-up have brought about great changes inspace
batteries. This electrochemistry impacts the key parameters of a
battery, namely:
Capacity - Dened as the number of Ampere Hours (Ah) a battery
can deliver at room temperatureuntil it reaches a cut-o voltage
where it can no longer deliver power routinely (commonly two
thirdsof the fully charged voltage); it is dependent upon the size
of the battery.
Specic Energy - The energy stored per unit mass and measured in
Watt Hours per Kilogram(Wh/kg). Reduced mass of the battery for the
same power storage allows a heavier and/or morecomplex payload.
2 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
Energy Density - The energy stored per unit volume and measured
in Watt Hours per Litre (Wh/l).Reduced volume of the battery for
the same power storage has a similar positive impact on
spacecraftdesign.
Cycle Life - The number of charge/discharge cycles a battery can
undergo and still provide theminimum required voltage.
Although not a parameter of the battery itself, but rather a
parameter of the state of a battery, theDepth of Discharge (DoD) is
dened as the Ah capacity drained from battery, divided by the real
Ahcapacity of the battery at that point (i.e. including any
measured or assumed degradation of the battery).This measure is
also sometimes referred to as its inverse, the State of Charge
(SoC), equal to 1 minusthe DoD. The SoC refers to the level of
charge remaining in the battery. Both of these measures are
oftenreferred to as percentages, achieved by multiplying the value
by 100.
A. Space Battery Technology Evolution
The rst common space batteries in the 1960s were Nickel Cadmium
(Ni-Cd). They became the coretechnology of many missions and were
the most common battery in use up to the mid-1980s. They are
wellcharacterised and known. The space industry rides on the back
of the tried and tested Ni-Cd battery and itis still powering
spacecraft like XMM-Newton today.
A Ni-Cd battery is temperature sensitive and has a typical
specic energy of approximately 25 Wh/kg.Ni-Cd has a high cycle life
which is important for long duration missions; however, signicantly
the Ni-Cdbattery is subject to the `Memory Eect'. The Ni-Cd battery
`remembers' its most frequently used DoDand does not work well
beyond that. That is to say, a Ni-Cd battery frequently discharged
to 25% DoD willnot be able to discharge lower than 25% DoD,
therefore eectively losing capacity and rendering the batteryunable
to provide the required power to the spacecraft. The potential loss
of capacity can be circumventedby regularly `re-conditioning' the
battery by discharging it nearly completely. While this operation
restoresthe battery it requires additional hardware on-board the
spacecraft and additional periodic operations ofthe spacecraft in
orbit. More recently, cadmium has also evoked environmental
concerns and is subject toregulatory scrutiny.
In the years since the rst space batteries, such as the early
ESA mission ESRO-2 in 1968, improvementshave been made in
electrochemistry and newer cell types have been developed. In the
1980s, Nickel Hydrogen(Ni-H2) batteries, a successor to Ni-Cd, came
into use. Ni-H2 cells combine and make the best of two
dierentchemistries - that of the Nickel Oxide electrode of the
Ni-Cd cell and that of the Hydrogen catalyst electrodeof the fuel
cell.
Their chemistry allows a deeper DoD for a comparable cycle than
Ni-Cd, resulting in a lower requiredAh capacity which in turn leads
to lower mass. As the cost of space ight depends upon mass, any
savingis highly advantageous.
Ni-H2 cells have been widely used in both LEO and GEO
spacecraft. The evolved chemistry lead to ahigher reliability and
longer lifetime in orbit compared to Ni-Cd, while avoiding the
penalty of regulatedcadmium.
In 2001 a move away from nickel-based batteries and to a new
chemistry began with the ight of therst Lithium Ion (Li-ion)
battery aboard Proba-1, an ESA technology demonstrator. The Li-ion
cell oers asignicant leap in performance. They oer a high specic
energy of 85-130 Wh/kg, a 3-5 fold improvementin specic energy
compared to Ni-Cd cells. A comparison of mass and volume of Ni-Cd,
Ni-H2 and Li-ionbatteries can be seen in gure 1. The huge gains in
mass and volume oer new design possibilities forincreasing payload
mass, volume and/or power demand. In addition the modular concept
of Li-ion batteriesgives benets of simplicity while also allowing
exibility in accommodation which was exploited in ESAmissions such
as Mars Express, CryoSat-2 and Philae.
3 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
... ..Ni-Cd
.Ni-H2
.Li-ion
.0 .
50
.
100
.
150
.
200
.
250
.
300
.
350
.
400
.
450
.
Mass(kg)
.
. ..Cells . ..Battery Packaging
(a) Battery Mass Comparison
... ..Ni-Cd
.Ni-H2
.Li-ion
.0 .
50
.
100
.
150
.
200
.
250
.
300
.
350
.
400
.
450
.
Volume(litres)
.
. ..Cells . ..Battery Packaging
(b) Battery Volume Comparison
Figure 1. Mass and volume comparison of Lithium Ion (Li-ion,
carbon cells) against heritage battery tech-nologies. Values given
are for a 10 kWh battery with a maximum DoD of 75%.1
The Li-ion cell is also magnetically `cleaner' than the nickel
batteries, which can be signicant in sensitiveinstrumentation.
Aside from the spacecraft itself, the handling of Li-ion is also
considerably easier. Priorto launch the Li-ion battery is very easy
to store and has a long storage life. As many programs
experiencedelays and launch slips, this pre-launch factor is
non-negligible. After launch, once in operation the Li-ionbattery
does not suer from any `memory eect'. This makes the battery easier
to operate and does awaywith the need for additional hardware for
the `re-conditioning' process and the added complexity of
theoperation itself.
B. Space Battery Technology Usage
Li-ion batteries are now ying on or have own on LEO, GEO and
Interplanetary missions. As their spaceheritage becomes
established, their performance has proven to be even better in ight
than predicted. Mis-sions are able to y longer and experience more
cycles at higher DoD while continuing to achieve substantialpayload
results.
Examples of space batteries in ESA spacecraft and the evolution
of battery chemistry is shown in table1 below.
4 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
Table 1. Summary of battery type and mission type for a
selection of ESA missions.
MissionName
MissionPurpose
LaunchDate
FlightDuration
Orbit BatteryType
NumberofBatteries
BatterySize
ESRO-2 Science 17.05.68 3 years LEO Ni-Cd 1 3 Ah
COS-B Science 09.08.75 7 years HEO Ni-Cd 1 6 Ah
Marecs-A Telecom 20.12.81 14.75 years GEO Ni-Cd 2 21 Ah
Giotto Science 02.07.85 7 years DeepSpace
Ag-Cd 4 16 Ah
Olympus Telecom 12.07.89 4 years GEO* Ni-CdNi-H2
11
24 Ah35 Ah
ERS-1 EarthObservation
17.07.91 5 years LEO Ni-Cd 4 24 Ah
Eureca Microgravity 31.07.92 1 year LEOy Ni-Cd 4 40 AhXMM-Newton
Science 10.12.99 Ongoing HEO Ni-Cd 2 24 Ah
Proba-1 TechnologyDemonstration
22.10.01 Ongoing LEO Li-ion 1 9Ah
Envisat EarthObservation
01.03.02 10 years LEO Ni-Cd 8 40 Ah
MSG-1 Weather 28.08.02 Ongoing GEO Ni-Cd 2 29 Ah
Integral Science 17.10.02 Ongoing HEO Ni-Cd 2 24 Ah
Mars Express Science 02.06.03 Ongoing Planetary Li-ion 3 22.5
Ah
SMART-1 Science 28.09.03 3 years Lunar Li-ion 5
Rosetta Science 02.03.04 Ongoing DeepSpace
Li-ion 3 16.5 Ah
MSG-2 Weather 21.12.05 Ongoing GEO Ni-Cd 2 29 Ah
Venus Express Science 09.11.05 Ongoing Planetary Li-ion 3 24
Ah
GOCE EarthObservation
17.03.09 4.5 years LEO Li-ion 1 78 Ah
Herschel Science 14.05.09 4 years L2 Li-ion 1 39 Ah
Planck Science 14.05.09 4.5 years L2 Li-ion 1 39 Ah
CryoSat-2 EarthObservation
08.04.10 Ongoing LEO Li-ion 1 78 Ah
*Batteries were used for LEOP and for 2 major recoveries. Both
batteries were frozen due to a spacecraft anomaly, thensuccessfully
recovered to operation.
yRetrieved by Space Shuttle and returned to Earth at end of
mission.
This evolution of ESA missions to the use of Li-ion batteries is
thanks to the benets of the technology.Figure 2 summarises the
battery types of the table above and graphically shows how space
battery technologyhas evolved over time as greater energy density
technologies have become available.
5 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
1950 1970 1990 2010
Energy Density
Silver-Zinc
Nickel-Cadmium/ Argon-Zinc
Nickel- Hydrogen
Lithium-Ion
Sputnik 1 1956
PROBA 2001
Mars Express 2003
CryoSat-2 2010
Sentinel Family 2014
Bepi Colombo 2016
Figure 2. Evolution of the energy density and use in the space
eld of battery technologies over time.
C. Lithium Ion Lifetime Management
Despite the demonstrated advantages of Li-ion batteries for
space use, they do still degrade over time and assuch this must be
managed to ensure that the optimal possible lifetime of the
batteries is achieved. Lifetimemanagement can only be addressed
after dening the health descriptors of the batteries. Battery
healthindicators at the highest level are:
Remaining Capacity - The same as the Capacity mentioned earlier,
but adjusted to give the realamount of Ah that the battery can
deliver. This is governed by two types of degradation;
degradationby discharge cycles and degradation due to ageing.
Internal Resistance - The internal resistance of the cells in
the battery. This is key to the performanceof the battery - higher
currents in or out cause a loss of resulting energy that can be
delivered. Theincrease of internal resistance is predominantly
governed by temperature, but also by ageing.
Battery Health
Internal ResistanceRemaining Capacity
Capacity fadedue to ageing
Capacity fadedue to cycling
Figure 3. The constituent parts that make up the health of a
battery.
Lifetime management for Lithium Ion batteries aims to minimise
the possible impacts from the types ofdegradation discussed above.
The key to minimising degradation is to manage and mitigate where
possiblethe degradation drivers given below.
6 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
Capacity Fade Due To Cycling
The capacity fade due to cycling is primarily impacted by the
DoD reached on a given cycle and thetotal number of cycles
experienced by the battery. Deeper and/or more frequent discharges
will leadto a higher capacity fade rate.
Capacity Fade Due To Ageing
The capacity fade due to ageing cannot be avoided, however
storing the batteries colder and not fullycharged signicantly helps
keeping the ageing rate low. The capacity fade due to ageing is
quicker athigher temperatures. Li-ion batteries should always be
kept below 20 degC and if possible close to 0degC. 0 to 5 degC
seems to be the optimum range for Li-ion battery longevity. A
reduction of thisageing of the batteries can be realised by
abstaining from continuously trickle charging the batteriesonce
they have reached 100% SoC.
Stress factors due to (dis)charging at high rates are also good
to avoid, as this keeps internal resistanceduring cycling inside a
low range.
All degradation types described above are more pronounced in the
rst phases of the lifecycle due to theexponential character of most
eects.
III. Case Study 1: Mars Express
A. Power System Design Description
The Power subsystem on Mars Express consists of two solar array
wings and three batteries, both connectedto a Power Conditioning
Unit by means of Array Power Regulators (APRs) and Battery
Charge/DischargeRegulators (BCDRs), respectively. The Power
Conditioning Unit (PCU) takes the input of these powersources and
regulates the power to a 28V bus. The Power Distribution Unit (PDU)
then distributes this28V regulated bus to the various users on the
spacecraft. The PCU has three operating modes, dependingon the
availability of the power sources versus the power demand of the
spacecraft:
APR Mode
Spacecraft Power Demand + Battery Charge Demand Maximum Possible
Solar Array Supply
In this case the maximum solar array supply still exceeds the
amount required by the spacecraft, butis not enough to reach the
maximum battery charge rate. In this mode the solar array is
operatedat maximum power output, supplying the whole requirements
of the spacecraft power bus, and anyremaining power being routed to
charge the battery, meaning that charging rate varies depending
onthe instantaneous level of unused solar array power.
BDR Mode
Spacecraft Power Demand (alone) >Maximum Possible Solar Array
Supply
In this case the solar arrays are operated at maximum power
output but this is still not enough tosupply the spacecraft demand.
In this case the batteries are discharged to make up the shortfall
inavailable power. A typical example of this case would be during a
spacecraft eclipse.
In the specic case of Mars Express there is a design anomaly
which means that the harness connectingthe solar cells to the APRs
is incomplete, causing a lower power performance than was designed.
Thewiring harness between the Solar Arrays and the APRs is such
that the Maximum Power Point Trackingvoting system cannot function
as was foreseen. At best 72% of the design power is available from
the solararrays. This reduced power was and is a restrictive issue
for Mars Express. Nonetheless, the mission has
7 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
outperformed the dened scientic goals in terms of lifetime
extension and scientic return due to carefulmanagement of the power
system and accurate planning tools.
In terms of the batteries used on Mars Express, they consist of
three 24 Ah (at Beginning of Life) Li-ionbatteries. Each battery is
composed of multiple Sony Cells manufactured into space qualied
hardware byABSL Space Products.
B. Battery Usage Prole
On Mars Express, the battery usage is dened by two factors - the
eclipse seasons and augmenting the poweravailable from the arrays
during special operations or low-power seasons. The key driver for
the degradationof the battery are the large and frequent discharges
caused by eclipse seasons. These seasons can last anumber of months
with typically a short gap in between. In the rst years of the
mission, the eclipses werelonger than in recent years, with maximum
eclipse duration reaching 90 minutes in 2004, reducing to a
peakeclipse duration during a season of 40 minutes in 2014. This
aligns well with the state of the power system,as the demands
placed on the batteries are reduced as they age. The degradation
rates of the battery havenot been as high as anticipated, and as
such there is power margin to work with and extended discharges
(i.e.on top of eclipse demand for science pointings as well as
outside eclipse seasons) are allowed and relativelycommon. This
trade-o allows extended science objectives to be realised by
allowing certain pointings thatimpact the DoD following an eclipse.
An example of such a \double discharge" case is given in this
graph,going to 30% DoD twice in one orbit.
.....
07:00
.
14:00
.0 .
5
.
10
.
15
.
20
.
25
.
30
.
35
.
08:00
.
09:00
.
10:00
.
11:00
.
12:00
.
13:00
.
Time
.
Depth
ofDischarge
(%)
Figure 4. Mars Express battery DoD over approximately one orbit
on DoY 2014-074. First peak is due tosolar eclipse by Mars, the
second due to science pointing that required suboptimal array
pointing.
The eclipse seasons on Mars Express cover approximately 60% of
the mission life, and during this timethere will be at least one
discharge/charge cycle every orbital revolution. Throughout the
years, being ableto more accurately plan and monitor discharges and
their degradation impact has been critical to ensure
8 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
proper management of the batteries as a limited resource
available to the mission.
C. Lifetime Preservation Measures
As the Mars Express batteries are indeed a limited resource, it
is important to take any necessary measuresto reduce the rate of
their degradation. Following the information from the previous
chapter, 5 key principlesfor reduction in degradation rate were
identied:
Minimise depth of discharge
In the case of eclipses, there is no way to avoid the discharge,
but we try to make sure the DoD isonly as high as strictly
required. In the orbit, the most interesting part for most science
observationsis around closest point to Mars (pericentre). The
eclipses take place just before pericentre or overlapwith it, from
which two power-related problems arise. During eclipse, if payload
is switched on, theDoD will be higher as a result of more power
demand. After eclipse, a special pointing might berequired to
enable the science observation and the sun aspect angle to the
arrays cannot be optimised,leading to extra discharge outside of
eclipse. Minimising DoD comes down to a trade-o between thevalue of
performing a science observation which eectively enlarges the DoD
and on the other handsafeguarding the longevity of the Mars Express
mission by preserving the batteries. Mars Expresshas mission
planning rules (max routine DoD = 45%) and resource allocation
processes in place toaddress requests for higher DoDs in a
structured manner. Special power optimised pointings have
beendeveloped to assist in this process. The transmitter is o by
default in eclipse and not switched onuntil well after the eclipse
to create a solid recharge margin. Also, in some mission seasons,
the largestheater groups are phased to use most power outside of
eclipse by pre-eclipse boost heating.
Minimise number of cycles
Minimising the number of cycles is achieved by making sure that,
outside of eclipses, no excess power isdemanded from the solar
arrays requiring a battery discharge, unless strictly planned for
and allowedduring the mission planning process. This is also the
reason for not performing too many BatteryCapacity Measurements in
ight, since they have a signicant impact based on high DoD and
extradischarge/charge cycle.
Store at correct temperature
The temperature at which Mars Express batteries are nominally
operated is between -5 and 0 degC.While the batteries do have
heaters to ensure they do not get too cold, they are largely above
thistemperature and the driver for their temperature is the steady
state temperature of the Mars Expressspacecraft platform, rather
than specic battery thermal control. During discharges,
temperatureexcursions of +10-15 degC are seen, depending on the DoD
and discharge rates. This can be seenbelow in gure 5.
9 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
20100303 20110407 20120511 20130615
6
4
2
0
2
4
6
8
10
12
Date
Tem
pera
ture
(deg
C)
Battery 1Battery 2Battery 3
Figure 5. Mars Express battery temperature from 2010 until 2014.
Raised periods correspond with batteryin use during eclipse
seasons.
Store at lower state of charge
It is known that a lower charge level will result in a slower
degradation rate due to ageing. On top ofthat, it has been proven
in ight on Mars Express that reducing the nominal state of charge
(wheneverpossible outside of eclipse seasons) from 100% down to
80-90% had a positive eect on the availablecapacity by the time the
new eclipse season was due to start. In gure 6 we can see that
during eclipseseasons, the discharge started from 100% SoC, and
outside eclipse seasons the SoC was lowered (therewere still some
discharges in this period but these were unavoidable and analysed
to ensure they weresmall and safe).
10 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
20100303 20110407 20120511 2013061530
40
50
60
70
80
90
100
Date
Stat
e of
Cha
rge
(%)
Figure 6. Mars Express battery SoC from 2010 until 2014.
Inter-eclipse seasons show extended periods wherebattery end of
charge level was lowered to 80-90% for lifetime preservation
purposes.
Keep charge and discharge rates low
The discharge rate is controllable based on which platform and
payload equipment is on and by designthe charge rate is set to the
maximum that can be delivered by the arrays after subtracting
thespacecraft bus requirements. This is up to a charge rate limit
of 9 Amps (but the BCRs can also beset to 3 Amps). Since the solar
array harness to the APRs is incomplete, the output from the arrays
isless, but from the point of view that slower charge is better for
battery preservation, this has a positiveeect here. The regular
charge rates lie between 1 and 4 Amps depending on the mission
season.
D. Capacity Measurement
The analysis of the power budget of Mars Express only covered
the originally planned mission duration of 1Martian year plus a
second Martian year as an extension (totalling approximately 4
Earth years). Estimateson battery capacity degradation therefore
only cover this period and take a very conservative predictionof
the degradation rate. To be able to assess the health state of the
batteries more accurately, and togather trends for predictive
models, the Flight Control Team had to develop its own empirical
methodsbased on data available from telemetry. A test under
laboratory conditions is not possible due to severalfactors: no
calibration of the sensing equipment is possible in-ight, no
possibility to remove the batteryfrom the circuit, no possibility
to control the discharge rate and no possibility to ne-control the
thermalenvironment. The method used here was initiated by the Mars
Express ight control team with assistancefrom the ESA battery
technology experts and further developed and ne-tuned by the Venus
Express ightcontrol team. The model used consists of tting a
battery voltage curve (based on the Beginning of
Life`State-of-Charge vs. EMF' curve) to the voltage measurements as
obtained from telemetry. The modeloptimises for initial energy
capacity degradation factor and the losses due to internal
resistance dissipation.
11 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
The internal resistance modelling is variable to account for
(dis)charge rates as well as thermal eects (heat-up during
discharge and cool-down during charge). Of particular interest is
the successful modelling of thetransition between discharge and
charge. The model is covered in more detail in Ref. 2.
.....21 .
21:5
.
22
.
22:5
.
23
.
23:5
.
24
.
24:5
.
25
.
14:00
.
15:00
.
16:00
.
Time
.
Battery
Voltage
(V)
.
. ..Model
. ..Telemetry
Figure 7. Battery voltage for Venus Express Deep Discharge Test
9, performed on DoY 2013-285. The batteryvoltage can be seen to
drop past the critical level where a faster rate of discharge
begins. The model tting,used to assess degradation, can be seen to
t well with the telemetry data.
To obtain a complete dataset for this analysis the operators had
to plan for dedicated deep discharges,as data available from
routine eclipses was not sucient to obtain a signicant excursion of
the EMF curveoutside its linear initial part. The discharge down to
60-65% DoD showed the portion of the curve wherethe steeper second
part started, the exact position of which is indicative of the
battery degradation level.Example telemetry and model tting of such
a test is shown in gure 7. These dedicated Deep DischargeTests were
performed in visibility (for safety reasons) and were triggered by
rotating the Solar Arrays edge-on to the Sun. The visibility
requirement causes the discharge rate to be relatively high as the
X-bandtransmitter is on. On-board protection was added to ensure
that at a given battery voltage the arrays wouldrotate back to the
normal position even in case of loss of ground contact.
Both Mars Express and Venus Express ight control teams have
determined long term trends of thecapacity evolution (which is
useful for long term mission planning and approval of mission
extensions) andused it for short term prediction of the battery
voltage behaviour for activities where the power marginavailable is
critical. Establishing a better relation between SoC and voltage
allowed the ight control teamto be able to use the margin, when
needed, more safely. Any allowed DoD gure in the mission
planningcycle can be coupled to the corresponding voltage used and
remaining capacity.
12 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
IV. Case Study 2: CryoSat-2
A. Power System Design Description
The power system of CryoSat-2 is driven by the constraints of
its non-sun-synchronous orbit, which featuresa wide variability in
sun incidence angle on the spacecraft and periods of eclipse where
no solar array poweris available.
The CryoSat-2 Power Subsystem (EPS) provides the following
functions:
Generation of electrical power by means of a solar array
Control, storage and distribution of electrical power to/via a
main bus
Battery management (charge/discharge/protection)
Provision of unregulated main bus power to the units attached to
this bus in the range of 22 to 34 V
Provision of status monitoring and telecommand interfaces for
subsystem operation and performanceevaluation
Provision of adequate redundancy and protection circuitry to
avoid failure propagation and to ensurerecovery from any
malfunction within the subsystem and/or load failure.
The CryoSat-2 EPS comprises the following units:
A two panel xed (non-rotating) GaAs Solar Array with 11
electrical sections per panel, regulated bymeans of a shunt
system
A single 78 Ah battery, consisting of 52 strings with 8 Li-ion
cells each, manufactured by ABSL SpaceProducts.
Combined Power Control and Distribution Unit
Battery charge is controlled by the PCDU in order to prevent the
battery from experiencing thermalstress. During insucient solar
array power, either due to eclipse or high demand, for example in a
dawn-dusk orbit with sub-optimal array alignment, the energy stored
in the battery will be used to satisfy thebus power demand. The
Battery is charged by applying an IV-method with charge current as
the controlparameter besides battery voltage. Eight commandable end
of charge (EoC) voltage limits (steps of 200mV) allow compensation
of any potential cell parameter variations of the battery.
B. Battery Usage Prole
The CryoSat-2 orbit is a polar orbit but it is not
sun-synchronous and consequently the orbital plane rotateswith
respect to the sun direction. The nodal plane regresses at a rate
of about 0.25 per day. It thereforemakes half a revolution,
sampling all local solar times, in just over 8 months. This means
that the satellitefaces great variations in solar illumination and
there are periods when it ies along the dawn-dusk line andis in
constant sunlight, but with only one of the solar arrays
illuminated. At other periods it ies in thenoon-midnight plane with
both solar arrays illuminated and undergoes eclipses. At maximum
extent theeclipse duration is around 36 minutes.
This leads to seasons of battery use, as with Mars Express, at
the present state of the spacecraft. However,as the arrays age, the
poor illumination in the dawn-dusk seasons could lead to more
regular battery use tocompensate for array output.
C. Lifetime Preservation Measures
Although the CryoSat-2 mission is younger and has more margin in
its power system, it is still consideredimportant to take measures
to reduce the rate of battery degradation in order to preserve the
resource ofbattery life for future use. As for Mars Express,
CryoSat-2 obeys the 5 key principles of Li-ion battery
lifepreservation:
13 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
Minimise depth of discharge
This is performed \de facto" on CryoSat-2, rather than actively
managed as with Mars Express. This isthanks to the greater design
margins in the CryoSat-2 power system and the more stable and
repetitivenature of its power demand. This allows the operators to
ensure that no excessive discharges will occuras long as routine
operations are conducted.
Minimise number of cycles
As with Mars Express, the cycles from eclipse periods are
unavoidable. However, outside of eclipseperiods, extraneous cycles
are largely avoided by the design margins, as in the previous
point.
Store at correct temperature
The batteries on CryoSat-2 are maintained by thermal control to
a region above 10 degrees. This is stillmostly low enough to
prevent excessive capacity fade degradation. The increased
temperature (withrespect to Mars Express) prevents the increase in
internal resistance at lower temperatures, which hastwo
drawbacks:
{ Lower charge/discharge eciency, as the internal losses are
higher
{ Lower charge capability, as the end of charge voltage limit is
reached already at a lower SoC, dueto the higher voltage drop
Store at lower state of charge
During the early days of a LEO type mission, batteries are often
being charged to 100% SoC evenwhen the missions DoD is around 30%.
This unnecessary margin in the power system has been usedon
CryoSat-2 to reduce the degradation due to ageing by minimising the
SoC of the battery during thewhole mission. For example, there
could be no energy demand from the battery for several weeks
everyfew months when the spacecraft is in a dawn-dusk orbit. During
these battery non-operational periodsthe SoC of the battery is
reduced in order to reduce the ageing degradation. However, it is
important toensure that the reduction in SoC during mission does
not interfere with the energy demand during fulloperation. During
the transition between no eclipse to longest eclipse the battery
voltage at the endof the discharge is monitored and when it crosses
a predened threshold the end of charge is modiedsuch as to always
have enough power available to ensure the safety of the satellite
even in a worst caseanomaly, while maintaining a low (around 80%)
SoC at the end of the charge whenever possible.
Keep charge and discharge rates low
This is another area where CryoSat-2 has passive control, with
low charge/discharge rates being ensuredby the spacecrafts
conservative design margins rather than by active operator
control.
D. Capacity Measurement
CryoSat-2 faces similar issues to Mars Express in that measuring
internal battery characteristics such as theinternal resistance and
the battery open circuit voltage is not achievable in ight. In fact
the CryoSat-2housekeeping telemetry provides us only with the
battery charge/discharge current and the battery terminalvoltage,
along with a summary of these in the on-board software. In
addition, the solar arrays have a xedposition and so cannot be
turned away to provoke a deep discharge of the battery, as with
Mars Express.Therefore estimations of the battery degradation
parameters have to be performed with routine ight data.
1. Battery internal resistance estimation
One of the best ways in ight to estimate the battery internal
resistance is to induce a high impulsivecurrent demand and measure
the corresponding terminal voltage drop, a method detailed in the
CryoSat-2User Manual.3 This method works well if the current demand
is large but the current demand on CryoSat-2is typically quite low,
with the largest peaks being around 3 amps, for example when a
heater is switchedon. However, even with this low current the
method can still be applied to see if trends on the evolution ofthe
battery internal resistance appear over satellite lifetime.
14 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
2. Battery Capacity Estimation
To estimate capacity of the battery without deep discharges, the
CryoSat-2 User Manual3 proposed a methodbased on comparing the
voltage drop and integrated current produced by the battery. Using
the theoreticalcurve for a new battery which converts voltage into
state of charge provided by ABSL, one can get thestate of charge
change between these two points and then estimate the capacity by
computing what a 100%discharge will produce. This method is similar
to the deep discharge method used on Mars Express but doesnot
provide reliable results for CryoSat-2 due to the relatively small
discharges.
Instead it has been preferred to monitor the degradation of the
battery by comparing the expected Ahconverted from the voltage drop
seen in TM using the ABSL curve and the eective Ah measured by
telemetryfor each longest eclipse. The ABSL curve provides a
conversion between voltage level of the battery andstate of charge
and can be used therefore to produce an estimate of the Ah that a
given voltage drop shouldproduce for a battery with a given
capacity. Periodically we compare the power that should be produced
by abattery with a capacity as estimated at the previous longest
eclipse with the power produced now computedby integrating the
current for the latest longest eclipse. This method is similar to
the previous method but,while not capable of providing an absolute
value of remaining capacity, it does provide a relatively
accuraterelative value of capacity lost between tests.
Using the percentage of reduction between the power produced at
the previous longest eclipse and theone at the last eclipse
measurement gives us a factor to apply to the last estimated
capacity of the batteryto estimate the current capacity of the
battery. This iteration was started using the capacity given by
ABSLfor the battery at beginning of life.
The results of this method are shown in gure 8 below and show a
reasonable match with the predictionmade by ABSL.
.....
2008
.
2016
.60 .
65
.
70
.
75
.
80
.
85
.
90
.
95
.
100
.
2009
.
2010
.
2011
.
2012
.
2013
.
2014
.
2015
.
67% Capacity Specied EoL Performance
.CryoSat-2 Launch
.
Time
.
Rem
ainingBattery
Capacity(%
)
.
. ..Evaluated Capacity from TM
. ..ABSL Battery Capacity Prediction
Figure 8. Evaluated vs. predicted battery capacity degradation
for the CryoSat-2 battery. The manufacturer(ABSL) predicted
degradation of capacity matches well with the battery capacity
degradation that has beenevaluated using telemetry.
15 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
V. Summary of Li-ion Battery Management on other ESOC
Missions
As mentioned previously, ESA's Proba-1 mission, launched in
2001, was the rst spacecraft to orbit Earthwith commercial Li-ion
battery technology on board. Since then almost all ESA missions
have been equippedwith Li-ion batteries, gradually bringing to an
end the era dominated by conventional nickel-cadmium
andnickel-hydrogen batteries for spacecraft.
The management and usage of Li-ion batteries diers from mission
to mission depending on their orbitprole and power system design.
As has already been mentioned, what all missions using Li-ion
batterieshave in common are the important operational constraints
for maximum charge and discharge voltagesand currents, correct
settings for end of charge levels and the avoidance of thermal
stress. All this iscommonly taken into account and managed by the
PCDU in combination with respective thermal regulationwhich by
design optimises the battery lifetime and reduces the operational
overhead. Furthermore the levelof operational experience has
increased with time and missions, providing a set of applied
best-practicescontributing to the battery longevity, and extending
missions considerably as highlighted in the above casestudies. In
the process of justifying the technical and operational feasibility
to extend an ESOC mission,Li-ion batteries have therefore proven to
be an essential factor.
To complement the above case studies, the following table gives
an overview of other ESOC missions andtheir respective battery
management characteristics. While the above case studies apply to
missions withpower systems designed to cope with regular eclipse
encounters or, in case of Mars Express, operations waybeyond their
nominal lifetime, other missions exist where lithium-ion batteries
are not used in routine ightat all or have shown to require no
additional operational measures. For example, Herschel and Planck,
eachequipped with a 36 Ah Li-ion battery, were ying in an orbit
without eclipses where together with the solararray design and
spacecraft orientation sunlight was permanently available to supply
the bus load. Hence,the battery was only used during launch when no
solar array power was available. Given that Lithium-ionbatteries do
not require reconditioning and are free of memory eects, no
activities for capacity preservationor capacity measurements were
needed throughout the mission.
GOCE had a 78 Ah battery at the front of the spacecraft which
operated awlessly throughout the mis-sion, encountering seasons of
16 eclipses per day with peak durations up to 32 minutes. Except
for staying inthe dened boundaries in which the battery had to be
operated no particular battery management strategieshad to be
applied due to the mission life-limiting factor being propellant,
not battery life. Battery perfor-mance assessment using in-ight
telemetry to compute the battery capacity spent during discharge
cyclesshowed good performance of the batteries. A battery
simulation on ageing and number of charge/dischargecycles based on
in-ight telemetry conducted by industry in preparation for the low
orbit operations in 2012concluded a battery degradation due to
ageing and cycling of about 6%, whereas 19.5% was expected atthat
time. Shortly before the re-entry the battery was successfully
operated at temperatures above 80 degCfor a short time.
The Rosetta mission is a similar case to Herschel/Planck where
the batteries are not used in routine.Although not foreseen, the
three 16.5 Ah batteries were used for a Mars y-by in February 2007
when thespacecraft entered solar-eclipse and for solar array
performance tests in 2010 and 2014 (before and after deepspace
hibernation) where a step-wise sun o-pointing of the solar arrays
made use of the batteries to supplythe required power and assess
how much power the solar arrays could provide. To preserve the
batterycapacity the end-of-charge level was lowered to 87%.
Venus Express on the other hand is similar to Mars Express where
the extension of the mission required adetailed understanding of
the degradation level and solar eclipse seasons could only be
survived with the on-board battery. To determine the battery
capacity Deep Discharge tests through solar array o-pointing
wereperformed, as with Mars Express. The collected data was tted to
dedicated models from which predictionson the capacity degradation
and internal resistance trend could be made. Furthermore, the
end-of-chargelevel was lowered outside eclipse seasons to 80%.
Having kept the batteries in a good shape will be crucialin the
2014 period where the batteries will be used to allow extended
periods of low aspect angles on thearrays and high DoDs during
aerobraking in the Venusian atmosphere.
16 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
Table 2. Summary of Li-ion battery monitoring and management
strategy for ESOC missions.
Mission Battery Usage Prole Lifetime PreservationMeasures
Capacity MeasurementIn-Flight
GOCE Launch, eclipses and peakpower demands
None Performed In-ight telemetry usedfor battery
simulationtests
Venus Express Launch, eclipses and peakpower demands
Battery end-of-chargelevel lowered to 80% SoC
Deep discharge tests andMonte-Carlo tting
Rosetta Launch, Mars swing-by inFeb. 2007 (eclipse) andSolar
Array performancetests in 2010 and 2014
Battery end-of-chargelevel lowered to 80% SoC
None performed
Herschel/Planck Launch only None Performed None Performed
VI. Lessons Learned
Throughout this process lessons have been learned that could
contribute to future or current operatorsworking with Lithium Ion
batteries.
1. All operational lifetime preservation measures presented in
this paper are useful, even though indi-vidually they may seem
minor. Each measure, when applicable, allows to slow down the
capacitydegradation by up to a few percent over mission lifetime
with respect to the spacecraft manufacturer'sprediction.
2. The mission, spacecraft and power system design heavily
inuence the ability to perform lifetimepreservation measures.
Operators should work with the spacecraft designers to ensure as
far as possiblethat the design allows the exibility to perform
these measures.
3. Operators should also work with the spacecraft designers to
ensure that sucient telemetry is availablein ight to accurately
monitor battery health. This could include increase of sensitivity
of sensors,number of sensors or frequency of samples.
4. Ground and space system modelling of the SoC and DoD, either
in prediction, planning or telemetry,should take into account a
variable degradation factor and allow this factor to be changed
easily inorder to match the real degradation.
5. The most eective (and only absolute) way to reliably measure
true degradation of a Lithium Ion spacebattery in ight is through a
Deep Discharge Test.
6. The ability to perform a Deep Discharge Test should be
included in the design of spacecraft powersystems, even if they
normally would not have that ability (i.e. CryoSat-2). The
mechanism used todo this would have to be carefully considered to
ensure it could be performed at negligible risk to themission.
7. Measures to preserve and monitor Lithium Ion batteries should
be implemented by operators as part ofthe operations concept of a
mission, planned for before launch, to maximise their eectiveness.
Whileit is never too late to implement such measures, they are most
benecial if started from day one.
8. A close working relationship between operations teams,
spacecraft and power system designers andbattery experts is key to
minimising the degradation and maximising the usage and lifetime of
a spacebattery.
VII. Conclusion
This paper has reviewed some background on Lithium Ion space
batteries and focussed on their use andmanagement on missions own
from ESOC. Even in this small selection it is clear that the
strategies being
17 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
-
employed are variable. There are a number of reasons for this,
as highlighted in the paper. These includemission design factors,
such as the usage prole of the battery, the length of the mission
and the hardwareavailable. However, the operational factors of how
the battery is monitored and operated are also variedacross
missions.
Ultimately though, it is clear that there are a number of common
threads that can be followed to bestmonitor Lithium Ion batteries
and to reduce the rate of degradation. The lessons learned point to
theneed to implement such strategies from the moment spacecraft
operations start and to consider them in thespacecraft design and
operations concept. It is also clear that there is still
harmonisation to be done betweendierent operators of Lithium Ion
batteries, spacecraft designers and battery experts. This will
result in anoptimal realisation of the potential of this battery
technology both in the spacecraft design phase and formany years of
operations to come - maximising the potential and longevity of our
space missions.
Acknowledgments
T. Ormston thanks all of the co-authors of this paper for their
hard work and for producing an excellentoverview of Lithium Ion
battery usage from across the spectrum of ESOC missions.
The authors thank the battery technology experts from within ESA
and from our industrial partnersthat have assisted us in building
up our operational experience and knowledge of working with Lithium
Ionbatteries.
References
1Dudley, G. and Verniolle, J., \Secondary Lithium Batteries for
Spacecraft," ESA Bulletin, , No. 90, May 1997, pp. 50{54.2Sousa, B.
and Van Der Pols, C. L., \Breath in, breath out, how healthy are
the batteries on Mars and Venus Express,"
SpaceOps, Stockholm, Sweden, June 2012.3Astrium GmbH, .,
\CryoSat-2 User Manual," CS-MA-DOR-SY-0001.
18 of 18
American Institute of Aeronautics and Astronautics
Dow
nloa
ded
by 1
89.1
37.1
47.1
47 o
n Ju
ly 7
, 201
5 | ht
tp://a
rc.aia
a.org
| DOI
: 10.2
514/6.
2014-1
883
IntroductionSpace Battery TechnologySpace Battery Technology
EvolutionSpace Battery Technology UsageLithium Ion Lifetime
Management
Case Study 1: Mars ExpressPower System Design DescriptionBattery
Usage ProfileLifetime Preservation MeasuresCapacity Measurement
Case Study 2: CryoSat-2Power System Design DescriptionBattery
Usage ProfileLifetime Preservation MeasuresCapacity
MeasurementBattery internal resistance estimationBattery Capacity
Estimation
Summary of Li-ion Battery Management on other ESOC
MissionsLessons LearnedConclusion