SPIS multi time scale and multi physics capabilities:SPIS multi time scale and multi physics capabilities: development and
application to GEO charging and flashover modeling
ROUSSEL J.-F., DUFOUR G., MATEO-VELEZ J.-C. ONERATHIEBAULT B. ArtenumANDERSSON B SSCANDERSSON B. SSCRODGERS D., HILGERS A. ESAPAYAN D. CNES
Outline
Introduction
New modelling capabilities Multi time scale Multi time scale Multi physics
Si l ti
Sep
. 201
0
Simulation cases Charging in GEO Flashover expansion
C –
Alb
uque
rque
–20
-24
Conclusions and perspectives
11th
SC
TC
2
SPIS context and project overview
SPINE (Spacecraft Plasma Interaction Network in Europe) community setup around year 2000 (A. Hilgers, J. Forest...): An idea was born: gather European efforts for SC-plasma interactions Exchange: knowledge, data, codes, results... Boost the development of a common simulation toolkit: ESA ITT in 2002 => SPIS
SPIS D l SPIS Development (Spacecraft Plasma Interaction Software) : Initial development: 2002 – 2005
ONERA-Artenum consortium ESA/ESTEC TRP contract
Sep
. 201
0
ESA/ESTEC TRP contract Major solver enhancement: 2006 – 2009
Mostly ONERA ESTEC ARTES contract, French funding
this presentation
SPIS releases (open souce): - v4.0 July 2009 - v4.3 soon
C –
Alb
uque
rque
–20
-24
Others: Some community developments Some CNES-funded enhancements (EP, ESD) ESD t i i d lli l t l t d (ESA TRP)
next presentation (Pierre Sarrailh)
11th
SC
TC
3
ESD triggering modelling almost completed (ESA TRP) Next steps: EP integration (Astrium), SPIS-GEO (Artenum), SPIS-Science...
( )
Overall status of SPIS code SPIS-UI:
Real framework: task monitor, data management, script console (jython)… Interfacing with modeler/mesh-generator, postprocessing tools…
SPIS-Num: Plasma:
Matter models: PIC (leapfrog/exact (potential P1)), Boltzmann distribution, multi physics E field solver: Poisson, non linear Poisson, singularities (wires, plates) Volume interaction: CEX (MCC)
Spacecraft: Material properties: secondary emission (under electron/proton/UV), conductivities
Sep
. 201
0
Material properties: secondary emission (under electron/proton/UV), conductivities (surface/volume, intrinsic/RIC), field effect, sputtering (recession rate, products generation and transport)
Equivalent circuit: coatings (RLC) + user-defined discrete components (RCV), implicit solver Sources: Maxwellian Axisymmetric two axes
C –
Alb
uque
rque
–20
-24 Sources: Maxwellian, Axisymmetric, two axes
Specific features: Time integration: control at each level (population, plasma, simulation) Numerical times: integrate fast processes over a smaller duration (electrons/ions, plasma/SC…)
11th
SC
TC
4
Multiscale capabilities: cell = box / 100,000 Modularity: OO (Java), “plug-in” classes (Java introspection)
Outline
Introduction
New modelling capabilities Multi time scale Multi time scale Multi physics
Si l ti
Sep
. 201
0
Simulation cases Charging in GEO Flashover expansion
C –
Alb
uque
rque
–20
-24
Conclusions and perspectives
11th
SC
TC
5
Multi time scale: requirements
Examples of modelling requirements: Charging in GEO: fast absolute charging (ms) / slow relative charging (mn) ESD triggering: slow precharge (mn) / very fast electron avalanche (ns!) + gg g p g ( ) y ( )
steep field emission
Surface potentials => SC circuit: Plasma
New SW requirements:spacecraft ground (node 0) Electric super node 1 Electric super node 2
C0 = Csat A0/Atot C1 = Csat A1/Atot C2 = Csat A2/Atot
Sep
. 201
0
New SW requirements: Circuit:
Inductances " C " ( h i h h G h ) i d f d fi d
0 sat 0 tot 1 sat 1 tot 2 sat 2 totIf CSat >0
C –
Alb
uque
rque
–20
-24 "Exact Csat" (charge conservation through Gauss theorem) instead of user defined
Csat
Circuit solver: I li i
Newton type solver with: - dI/dV predictor (a matrix) with validity control
11th
SC
TC
6
ImplicitVariable, automatic time step
dI/dV predictor (a matrix) with validity control - automatic time steps (saturating validity)
The circuit solver: a test case0
Implicit solver / automatic time step: An example (GEO charging with -4000
-3000
-2000
-10000 20 40 60 80 100
elecNode0ground elecNode1surface elecNode3surface
very large electron flux) Quite large range of time scales
0
-7000
-6000
-5000
-4000
-3000
-2000
-10000 2 4 6 8 10
elecNode0ground elecNode1surface elecNode3surface
4000
-3000
-2000
-1000
00 0,2 0,4 0,6 0,8 1
elecNode0ground elecNode1surface elecNode3surface
Sep
. 201
0 -7000
-6000
-5000
-7000
-6000
-5000
-4000
0
C –
Alb
uque
rque
–20
-24
-4000
-3000
-2000
-1000
00 0,02 0,04 0,06 0,08 0,1 elecNode0ground
elecNode1surface elecNode3surface
-3000
-2000
-1000
00 0,002 0,004 0,006 0,008 0,01
11th
SC
TC
7-7000
-6000
-5000
4000
-7000
-6000
-5000
-4000 elecNode0ground elecNode1surface elecNode3surface
Outline
Introduction
New modelling capabilities Multi time scale Multi time scale Multi physics
Si l ti
Sep
. 201
0
Simulation cases Charging in GEO Flashover expansion
C –
Alb
uque
rque
–20
-24
Conclusions and perspectives
11th
SC
TC
8
Multi physics modelling requirement
Typically simulate in a single simulation: Dense quasi-neutral regions Low density, space charge regions
low density, space charge
Low density, space charge regions
Examples: Ambient plasma
positively biased surface
space charge region (sheath)at rest / sheath:
Expanding plasma /
Sep
. 201
0
low density, space charge region
p g pfast electrons ahead of the plasma front (ESD, EP i iti )
C –
Alb
uque
rque
–20
-24
cathode spot, plasma source
EP ignition…):
Method:
11th
SC
TC
9
multi-zone, interface handling at sheath edge or plasma bubble edge
The physics at the boundary
Child-Langmuir theory: The current emitted at the boundary is the maximum allowed by space charge
(space charge limitation): E = 0 in the emission plane (Te = 0) Case 1D analytical: jCL ~ V 3/2 / d 2
ideal Child Langmuir situation (Te=0, no ions)ideal Child Langmuir situation (Te=0, no ions)ideal Child Langmuir situation (Te=0, no ions)V,
collecting
Sep
. 201
0
xelectron source (dense
plasma in
collecting electrode
C –
Alb
uque
rque
–20
-24
potential (x 4̂/3)potential (x 4̂/3)charge density (~x^ 2/3)
potential (x 4̂/3)charge density (~x -̂2/3)potential due to larger charge
plasma in pour case)
11th
SC
TC
10
potential (x 4/3)charge density (~x -̂2/3)
charge density (~x -̂2/3)potential due to larger chargelarger charge (larger current)
potential due to larger chargelarger charge (larger current)potential due to smaller chargesmaller charge (smaller current)
Algorithm
Multi-physics solver design:
Loop 1: l
3. In case of "artificial source", electron balance for the expanding plasma bubble
plasma dynamics
2. Coupling of quasi-neutral region and space charge region (Child-Langmuir 3D)
Define n0 and Te in Boltzmann distribution n0exp(e/Te) for electrons in dense quasi-neutral region
Loop 2: CL
Modify local parameter defining jCL-3D = jCL-1D , itself defining boundary where jth = jCL-3D
2. Coupling of quasi neutral region and space charge region (Child Langmuir 3D)
no
electron b l
condition
Loop 3: "floating
Sep
. 201
0
Particle dynamics (PIC electrons)
1. Vlasov-Poisson solving
no
noyes
CL condition:En = 0 on
yes
balance ok?potential" of
the plasma bubble (i l di
C –
Alb
uque
rque
–20
-24
yes Poisson equationStationary?
or given loop
number
(including cathode spot), still lacks
11th
SC
TC
11The end
lacks stability
Test case 1 – bubble LD T t l b bbl
total
Test case: plasma bubble expansion
Electron density:y composed of Boltzmann
electrons in dense ion zone (quasi neutral)(q )
and PIC electrons in low density zone (non neutral)
Sep
. 201
0
Boltzmann PIC
C –
Alb
uque
rque
–20
-24 Boltzmann PIC
11th
SC
TC
12
Test case 1 – bubble LD S
ep. 2
010
C –
Alb
uque
rque
–20
-24
11th
SC
TC
13
Test case 2 – bubble HD
Test case2: plasma bubble expansion
Higher electron density (x 100)
Sep
. 201
0
Boltzmann PIC
C –
Alb
uque
rque
–20
-24 Boltzmann PIC
11th
SC
TC
14
Test case 2 – bubble HD S
ep. 2
010
C –
Alb
uque
rque
–20
-24
11th
SC
TC
15
Test case 3 – Child-Langmuir test case
1D test case, close to ideal CL 1 case
Sep
. 201
0C
–A
lbuq
uerq
ue –
20-2
4
SPIS current density = 0.35 A/m2
CL theory at Te = 0:jCL = 0 233 A/m2
11th
SC
TC
16
- jCL = 0.233 A/m2
- fine, as much as can be checked
Specific difficulties encountered
Some extra instabilities were discovered and had to be handled (with specific algorithms):
CL condition feed back loop can be unstable (Bohm type instability) if too small ion gradient: extracting electric field can be due to a positive space charge in the space charge
zone, and increasing the electron emission is not necessarily an improvement! stability condition (li / d) (es / kTe) < 1, with: li the typical scale of ion density
variation (fixed at electron time scale), d the sheath size, s the sheath potential drop and Te the electron temperature
=> need to consider a possible positive space charge in the space charge zone to
Sep
. 201
0
=> need to consider a possible positive space charge in the space charge zone to improve the algorithm
Bi-stable behaviour: A region can be consistently considered as:
C –
Alb
uque
rque
–20
-24 A region can be consistently considered as:
either in quasi-neutral zone (high density) or in the space charge zone (smaller density) (appeared when considering Ne not strictly = Ni in quasi-neutral)
11th
SC
TC
17
(appeared when considering Ne not strictly Ni in quasi neutral)
=> control of non neutrality w.r.t. cell size / Debye length ratio
…
Outline
Introduction
New modelling capabilities Multi time scale Multi time scale Multi physics
Si l ti
Sep
. 201
0
Simulation cases Charging in GEO Flashover expansion
C –
Alb
uque
rque
–20
-24
Conclusions and perspectives
11th
SC
TC
18
Modelling charging in GEO
Plasma dynamics : Blocking of photo/secondary Blocking of photo/secondary
emission by the barrier (small barrier height compared to potentials involved)
Sun side Equipotential at-3100 V
Saddle point at– 3010 V
Accuracy of (collected) currents: small object in a large computation box (noisy) => backtracking
d d ( b f l f d
-4500 V (S/C ground)
-3000 V S/C solar array
Sep
. 201
0
needed (can be useful for detector also e.g.) Shade side
-3500 V
C –
Alb
uque
rque
–20
-24
Multi-time scale modelling Implicit SC circuit solver
11th
SC
TC
19
Comparison with NASCAP modelling
Published model (Davis et al) Published model (Davis et al)
"Validation of NASCAP-2K spacecraft-environment interactions calculations", V. A. Davis, M. J. Mandell, B. M. Gardner, I. G. Mikellides, L. F. Neergaard, D. L. Cooke and J. Minor, 8th Spacecraft Charging Technology Conference, Huntsville, Alabama, USA, 20-24 oct. 2003
Similar model with SPIS (B. Andersson, SSC)
Sep
. 201
0
Similar model with SPIS (B. Andersson, SSC)
Comparison of potential maps and time variation
C –
Alb
uque
rque
–20
-24
11th
SC
TC
20
Potential maps (t=1000s)
Comparison with NASCAP Globally good agreement Small local differences (OSR e.g.) Small local differences (OSR e.g.) Often smaller gradient
Sep
. 201
0C
–A
lbuq
uerq
ue –
20-2
4
Without SEE
11th
SC
TC
21
Time evolution
Comparison with NASCAP: falls within NASCAP-series code
results
Node potentials vs Time
-2000
00 100 200 300 400 500 600 700 800 900 1000
Sep
. 201
0
-6000
-4000
entia
l [V]
elecNode0ground elecNode0surface elecNode1surface elecNode2surface elecNode3surface
C –
Alb
uque
rque
–20
-24
-12000
-10000
-8000
Pote elecNode4surface
elecNode5surface elecNode6surface elecNode7surface
SC ground
11th
SC
TC
22
-14000
Time [s]
SC ground (the one to be compared)
-Part of the
/Chas
i
PVSA (shadow
id ) OSR
PVSA (solar
id )Main SC
t tTop
A t
Circular anten
- -
s/c si side) OSR side) structure Antenna nae
Material
Black kapton Kapton OSR Solar Cells Teflon
Non-conducting paint
Graphite
Absolute Charging (kV)
NASCAP/GEO -10.0 -8.2 to -13.1-8.23 to -
10.7-5.2 to -
7.68 -7.5 to -12.7-8.3 to -
10.3 N/A
SEE None in -7.5 to -SEE Handbook -8.6
None in model -7.3 to -9.6 -3.6 to -5.7 -6.8 to -11.3
7.5 to 11.3 N/A
Nascap-2k -12.0 -11.5 to -14.4-10.0 to -
13.7-7.2 to -
10.8 -7.9 to -14.0-7.9 to -
14.0 N/A
SPIS 10 9 12 9 11 7 6 1 9 8 9 7 10 9
Sep
. 201
0
SPIS -10.9 -12.9 (-10.9 to -
13.9)
-11.7 -6.1(-5.8 to -
6.4)
-9.8(-7.9 to -11.6)
-9.7(-9.6 to -
9.8)
-10.9
Differential charging (kV)
C –
Alb
uque
rque
–20
-24 NASCAP/GEO 1.8 to -3.1 1.77 to -0.7 4.8 to 2.3 2.5 to -2.7 1.7 to -0.3 N/A
SEE Handbook
None in model 1.3 to -1.0 5 to 2.9 1.8 to -2.7 1.1 to -0.3 N/A
Nascap-2k 0 5 to -2 4 2 to -1 7 4 8 to 1 2 4 1 to -2 2 to -0 2 N/A
11th
SC
TC
23
Nascap 2k 0.5 to 2.4 2 to 1.7 4.8 to 1.2 4.1 to 2 2 to 0.2 N/A
SPIS -2.0(0 to -3.0)
-0.8 4.8(5.1 to 4.5)
1.1(3 to -0.7)
1.2(1.1 to 1.3)
0
Outline
Introduction
New modelling capabilities Multi time scale Multi time scale Multi physics
Si l ti
Sep
. 201
0
Simulation cases Charging in GEO Flashover expansion
C –
Alb
uque
rque
–20
-24
Conclusions and perspectives
11th
SC
TC
24
Flashover experiment: the setup Inverted Potential Gradient charging
of a large coupon (CNES R&T) In JONAS plasma tank
(ONERA/DESP) Monitoring of flashover current
individually on some of the strings
canon à électrons
Large PVSA coupon: 1.33 x 0.6 m, 19 strings
Sep
. 201
0
source à plasma
feuille de diffusion (V=+10 kV)
caméra reliée à un magnétoscope
distanceréglable
C –
Alb
uque
rque
–20
-24
Arcoupon de GS
réglable
trajectoire de la sonde de potentielsonde de potentiel
plateau support
More positive potential (SEE or plasma neutralisation)
Initially positive (global potential rise, BO), then progressive neutralisation (FO)
11th
SC
TC
25
cage de Faraday cellules du GS
negative potential (bias)~ zero potential (e- emission, BO)
Flashover experiment: the measurements
Individual currents on intermediate string (9 to 13) + others together
This ESD (No 16) on string 15
ESD n°16, initiated on string 15
0,2
0,4
rrent
s (A
) Delay of flashover on successive strings consistent with plasma
Sep
. 201
0
-0,2
0,00,0E+00 5,0E-05 1,0E-04 1,5E-04Time (s)
Cu
Blow Off
consistent with plasma expansion at 104 m/s
Ahead of the plasma
C –
Alb
uque
rque
–20
-24
-0,6
-0,4
Blow-Off
string 9
string 10
string 11
t i 12
front, smaller current thought to be carried by electrons only (limited by space
11th
SC
TC
26-1,0
-0,8string 12
string 13
other strings
(limited by space charge)
FO modelling: the challenge
Multi physics: dense zone (plasma expansion) low density with positive potential low density with positive potential
Multi time scale Large currents in dense plasma
S ll t t f th l ( l t ) Small currents out of the plasma (electrons)
Difficult of coupling of both: Propagation of the plasma edge on the PVSA
Sep
. 201
0
=> brutal variation of current expected on the cells Moving singular point
C –
Alb
uque
rque
–20
-24 plasma bubble
in expansion
~ +1000 V ~ 0 V
11th
SC
TC
27
PVSA
~ 0 V
FO modelling
Initial conditions After the blow off PVSA ground already g y
back to ~ 0 Reason: loop 3 of
multiphysical model is needed to model that ("floating potential of the plasma bubble)
ESD model
Sep
. 201
0
ESD model Ion source on a 5x5 cm
patch (Maxwellian, 10 mA – 1A range, 10-100 eV)
C –
Alb
uque
rque
–20
-24 eV)
Electron: Boltzmann distribution (10 eV, n0 = 1016 m-3) in the dense one + PIC in the space
11th
SC
TC
28
zone + PIC in the space charge zone
FO modelling: plasma expansion / ion density
Sep
. 201
0C
–A
lbuq
uerq
ue –
20-2
4 11
th S
CTC
29
FO modelling: plasma expansion / ion density + zone boundary
Sep
. 201
0C
–A
lbuq
uerq
ue –
20-2
4 11
th S
CTC
30
FO modelling: plasma expansion / electron density
Sep
. 201
0C
–A
lbuq
uerq
ue –
20-2
4 11
th S
CTC
31
FO modelling: plasma expansion / electron density + isocontours
Sep
. 201
0C
–A
lbuq
uerq
ue –
20-2
4 11
th S
CTC
32
FO modelling: potential
Sep
. 201
0C
–A
lbuq
uerq
ue –
20-2
4 11
th S
CTC
33
FO modelling: current collectionESD n°16, initiated on string 15
0,2
0,4
urre
nts
(A)
Reaches all strings successively, but not the whole panel Expansion speed somewhat too fast Electron current ahead of the plasma correct
0 6
-0,4
-0,2
0,00,0E+00 5,0E-05 1,0E-04 Time (s)
Cu
Blow-Off
string 9
string 10
In presence of plasma, the plotted current is too large due to faster neutralisation and potprocessing artefact (current x 3 here due to an absence of current update in th i li it l I > I + dI/dV V)
-1,0
-0,8
-0,6 string 11
string 12
string 13
other string
the implicit solver I => I + dI/dV V)
0
0,E+00 1,E‐05 2,E‐05 3,E‐05time [s]
0 05
0,00
0,E+00 1,E‐05 2,E‐0
zoom on current y-axis
Sep
. 201
0 ‐10
‐5
t [A]
other strings (ESD side) ‐0,20
‐0,15
‐0,10
‐0,05
t [A]
C –
Alb
uque
rque
–20
-24
25
‐20
‐15
curren string 13
string 12string 11string 10string 9 ‐0,40
‐0,35
‐0,30
‐0,25
current
other strings (ESD side)string 13string 12string 11string 10
11th
SC
TC
34
‐30
‐25 string 9other strings (other side)
‐0,50
‐0,45
, string 10string 9other strings (other side)
Conclusions on FO simulation capabilities
Plasma expansion in volume with dynamical adjustment of the zone boundary: Operational Robustness could be improved: Robustness could be improved:
Several instabilities were identified (with respect to the ideal CL approach) Stabilisation schemes were proposed and tested => ok, sometimes needing manual tuning (UI-
accessible parameters) Improved schemes could be developed:
Improved feed back function (non linear…) Theoretically proven
Plasma expansion on S/C surfaces: interaction of zone boundary with surfaces
Sep
. 201
0
Plasma expansion on S/C surfaces: interaction of zone boundary with surfaces Expansion certainly less favourable than in reality:
Total neutralisation of PVSA not demonstrated in SPIS, although exists in experiments Stopped expansions exist in experiments but are thought to be more related to cathode spot
C –
Alb
uque
rque
–20
-24 Stopped expansions exist in experiments, but are thought to be more related to cathode spot
phenomena
Difficulty = presence of a singular point (line) where zone boundary crosses panel No specific handling => steps/peaks in current when boundary reaches a new node
11th
SC
TC
35
p g p p y => detailed study of this singular region needed (in particular centring of all quantities)
Conclusions and perspectives
Major solver improvements for SPIS: multi-physics, multi time scale Very ambitious Achievements:
Most features operational, not bad given the difficulties Still a lack of robustness of some points
Many applications: ESD triggering (electron avalanche), cf. next presentation: mn scale (charging) to ns scale
(electron avalanche)
Sep
. 201
0
Positive biasing in a dense plasma (probes on SC, connectors on PVSA…)…
Perspectives
C –
Alb
uque
rque
–20
-24
Revisit some of the control features: to improve robustness To simplify the control parameters (more sophisticated automation)
11th
SC
TC
36
Complete (stabilise indeed) the external loop for multi physics ("floating potential" of the plasma bubble)
Top Related