Finite Element Analysis of Antennas and Phased Arrays in the Time Domain Jian-Ming Jin Center for Computational Electromagnetics and Electromagnetics Laboratory Department of Electrical and Computer Engineering University of Illinois at Urbana-Champaign Urbana, Illinois 61801 July 30, 2012 Antennas Mini-Symposium
56
Embed
Finite Element Analysis of Antennas and Phased … Element Analysis of Antennas and Phased Arrays in the Time Domain Jian-Ming Jin Center for Computational Electromagnetics and Electromagnetics
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Finite Element Analysis of Antennas and Phased Arrays
in the Time Domain
Jian-Ming Jin
Center for Computational Electromagnetics
and Electromagnetics Laboratory
Department of Electrical and Computer Engineering
University of Illinois at Urbana-Champaign
Urbana, Illinois 61801
July 30, 2012
Antennas Mini-Symposium
Time vs. Frequency Domain
Multiple excitation / angular sweep
Dispersive material modeling
Steady state phenomena
Frequency domain:
Broadband simulation / frequency sweep
Nonlinear material/device modeling
Transient phenomena (coupling paths,
resonances, multiple bounces)
Physical insight useful to design engineers
Time domain:
Basic TDFEM formulations and unconditionally stable algorithm
Perfectly matched layers (PML) for mesh truncation
Hybrid FDTD-TDFEM for efficient mesh truncation
Time-domain FE-BI for accurate mesh truncation
Time-domain waveguide port boundary condition (WPBC)
Highly efficient domain decomposition methods
Periodic TDFEM with Floquet ABC for mesh truncation
Modeling of general dispersive, lossy, and anisotropic materials
Explicit TDFEM (Discontinuous Galerkin and Huygens’ methods)
Tree-cotree splitting to fix low-frequency breakdown problems
Hybrid field-circuit-network simulation based on TDFEM
Applications to scattering, antennas, antenna arrays, microwave
devices, RF circuits, FSS, photonic crystals, etc.
Progress in TDFEM
Maxwell’s equations:
TDFEM Basics
imp
( )( ) ( )
tt t
t
HE M
imp
( )( ) ( ) ( )
tt t t
t
EH E J
Curl-curl wave equation:
0 o
0
1ˆ ˆ ˆ 0 on n Y n n S
t
E E
2imp imp
2
1 ( ) ( )( )
t tt
t t t
J ME EE
Radiation boundary condition:
Weak-form representation:
TDFEM Basics
Spatial discretization:
2nd-order ordinary differential equation:
o
2
02
imp imp
1ˆ ˆ( ) ( ) ( )
V S
V
dV Y n n dSt t t
dVt
E E ET E T T T
J MT
edge
1
( , ) ( ) ( )
N
i i
i
t E t
E r N r
2
2
{ } { }[ ] [ ] [ ]{ } { }
d E d ET R S E f
dt dt
Unconditionally Stable TDFEM
( 1) ( 1){ } { } { }
2
n nd E E E
dt t
2 ( 1) ( ) ( 1)
2 2
{ } { } 2{ } { }
( )
n n nd E E E E
dt t
( 1) ( ) ( 1){ } { } (1 2 ){ } { }n n nE E E E
Unconditionally stable when ! 1 4
Newmark-beta method:
1910 – 1981
Civil Engineering
University of Illinois
Temporal discretization via Newmark- method:
TDFEM Basics
Unconditionally stable time-marching equation:
( 1) ( )
2 2
( 1) ( 1) ( ) ( 1)
2
1 1 1 2 1[ ] [ ] [ ] { } [ ] [ ] { }
( ) 2 4 ( ) 2
1 1 1 1 1 1[ ] [ ] [ ] { } { } { } { }
( ) 2 4 4 2 4
n n
n n n n
T R S E T S Et t t
T R S E f f ft t
1
( 1) ( )
2 2
( 1) ( 1) ( ) ( 1)
2
1 1 1 2 1{ } [ ] [ ] [ ] [ ] [ ] { }
( ) 2 4 ( ) 2
1 1 1 1 1 1[ ] [ ] [ ] { } { } { } { }
( ) 2 4 4 2 4
n n
n n n n
E T R S T S Et t t
T R S E f f ft t
Major Challenges:
TDFEM Challenges
Modeling of Large Computational
Domains and Finite Arrays
Modeling of Infinite Periodic
Structures
Truncation of Open Free Space
Modeling of Waveguide Ports
Modeling of Dispersive Material
J. M. Jin and D. Riley, Finite Element Analysis of
Antennas and Arrays. Wiley, 2009.
TDFEM Mesh Truncation
Absorbing boundary conditions (ABC)
Easy and highly efficient, yet approximate
Perfectly matched layers (PML)
Direct implementation in TDFEM
Complicated formulation, uses a single mesh
Implementation via FEM-FDTD hybrid
Highly robust, uses a hybrid mesh
Boundary integral equations (BIE)
Most accurate and most expensive
Total-Field / Scattered-Field Interface
Near-Field to Far-Field Transformation Boundary
Perfectly Matched Layer
E inc
Finite-Difference Region
(Structured Hexahedra, Explicit)
Gri
d T
erm
inati
on
Finite-Element Region
(Unstructured, Implicit)
FEM-FDTD Hybrid
Implicit – Explicit Solution Scheme
Time-Domain Finite Element Approach (Hybridized with FDTD)
RCS Validation: Standard Benchmark
incEincE
Elevation (degrees)
Mo
no
sta
tic
RC
S(d
Bsm
)
-90 -60 -30 0 30 60 90-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
Measurement (VV)
FETD-FDTD (VV)
Elevation (degrees)
Mo
no
sta
tic
RC
S(d
Bsm
)
-90 -60 -30 0 30 60 90-60
-55
-50
-45
-40
-35
-30
-25
-20
-15
-10
Measurement (HH)
FETD-FDTD (HH)
VV-Polarization HH-Polarization
(a) Double Ogive Geometry
Surface Current with 30 GHz
RF Illumination (b) Monostatic RCS at 9 GHz.
FEM-FDTD Hybrid Example
J. M. Jin and D. Riley, Finite Element Analysis of Antennas and Arrays. Wiley, 2009.
Input impedance of a cavity-backed
microstrip patch antenna
Detailed parameters: See J. Jin, The Finite
Element Method in Electromagnetics (2nd edition),
New York: Wiley,2002.
TDFEM-PML Example
Waveguide Port Modeling
Simplified feed model: electric probe feed
Simplified feed model: voltage gap
Waveguide port boundary condition (WPBC)
Coaxial Cable Rectangular Waveguide Microstrip Line
inc)(ˆ UEE Pn
1
TMTM
TE
1
TETEM
0
TEM
0
)(
)()()(
m S
mm
S
m
m
m
S
dS
dSdSP
Eee
EeeEeeE
inc inc TEM TEM inc
0 0
TE TE inc
1
TM TM inc
1
ˆ ( )
( )
( )
S
m m
m S
m m
m S
n dS
dS
dS
U E e e E
e e E
e e E
Time-Domain Formulation:
Assume dominant mode
incidence:
incidence TMdominant )(2
incidence TEdominant )(2
incidence TEM)(2
incTM
1
incTE
1
incTEM
0
inc
f
f
f
e
e
e
U
Waveguide Port Model
Z. Lou and J. M. Jin, “Modeling and
simulation of broadband antennas
using the time-domain finite element
method,” IEEE Trans. Antennas
Propagat., vol. 53, no. 12, pp. 4099-
4110, Dec. 2005.
Monopole Antennas
mm 1.0a
mm 2.3b
mm 32.8h
mm 1.0a
mm 2.3b
23.1 mmh
' 2.0 mmh
o30
Measured data: J. Maloney, G. Smith, and W. Scott, “Accurate computation of the radiation from simple
antennas using the finite difference time-domain method,” IEEE Trans. A.P., vol. 38, July 1990.
Microwave Resonator Filter
Measured data:
J. R. Montejo-Garai and J. Zapata,
“Full-wave design and realization of
multicoupled dual-mode circular
waveguide filters,” IEEE T-MTT,
vol. 43, pp. 1290-1297, June 1995
Dispersive Modeling
Constitutive relations: 0 0( ) ( ) ( ) ( )et t t t D E E
0 0( ) ( ) ( ) ( )mt t t t B H H
Weak-form solution:
2
1
0 2
0
2imp
0 2
1 ( ) ( )( ) ( ) ( ) ( )
( )( ) ( )ˆ( )
e
V
e
S V
t tt t
t t
tt tt dV n dS dV
t t t
E ET E T Q T T
JE HT T T
1
0
( ) 1( ) ( )
tt t
t
HQ E
0 0 imp
( ) ( )( ) ( ) ( ) ( )m m
t tt t t t
t t
H HE H M
Dispersion Modeling Example
Frequency (GHz)
r,
r
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20
2
4
6
8
10
12
-r''
r'
r'
-r''
r() = 1 + 9/(1+ j 510
-11)
r() = 2 + 4/(1+ j 210
-10)
Frequency (GHz)
|
|,|T
|
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 20
0.2
0.4
0.6
0.8
1
Exact - Reflection
Exact - Transmission
FETD - Reflection
FETD - Transmission
Dispersive Electric & Magnetic Slab
Transmission Coefficient
Normal Incidence
Reflection Coefficient
( ), ( )
incE
10 cm
Dispersion Modeling Example
( ), ( )
incE
10 cm
Frequency (GHz)
r
0.6 0.8 1 1.2 1.4-40
-20
0
20
40
60
80
100
-r''
r() = 1 +
r
2/ (
r
2+ j 2
r-
2)
r= 2 10
9
r= 5 10
7
r'
Frequency (GHz)
|
|,|T
|(d
B)
0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4-140
-120
-100
-80
-60
-40
-20
0
20
FETD (Reflection)
FETD (Transmission)
Exact (Reflection)
Exact (Transmission)
Normal Incidence
Reflection Coefficient
TransmissionCoefficient
Four-Arm Sinuous Antenna
20-mil Dielectric Substrate
Frequency (GHz)
Inp
ut
Active
Imp
ed
an
ce
(Oh
ms)
4 5 6 7 8-20
-10
0
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
Predicted Resistance (Volmax)
Predicted Reactance (Volmax)
Measured Average Resistance (80 Ohms 4-12 GHz)
Measured Average Reactance (0 Ohms 4-12 GHz)
Predicted Average Resistance (84.4 Ohms 4-8 GHz)
Predicted Average Reactance (4.4 Ohms 4-8 GHz)
Measured (average)
Predicted (average)
4-Arm Sinuous
20-mil Substrate
Thin Substrate Drops Theoretical Free-Space Resistance of 133.3
Ohms Down To Approximately 80 Ohms
Sinuous Antenna Etched on 20-mil Substrate*
*N. Montgomery and D. Riley, “Broadband antenna predictions using state-of-the-art hybrid FETD-FDTD,”
URSI Symposium Digest, July 2001.
Coaxial
Feeds
D. J. Riley and C. D. Tuner,
“Volmax: A solid-model-based,
transient, volumetric Maxwell
solver using hybrid grids,”
IEEE AP. Mag., vol. 39, Feb. 1997.
The entire antenna structure
is partitioned into three
sections
Adjacent sections are
connected by WPBC
TDFEM result is obtained by
cascading S-parameters of
three sections
Vlasov Antenna (Return Loss)
Measured
TDFEM
q
Vlasov Antenna (Gain pattern)
Five-Monopole Array (Geometry)
unit: inch
Finite Ground Plane:
• 12’’ X 12’’
• Thickness: 0.125’’
SMA Connector:
• Inner radius: 0.025’’
• Outer Radius: 0.081’’
• Permittivity: 2.0
Monopole Array (Impedance Matrix)
1 2 3 4 5
5
4
3
2
1
Domain Decomposition
Time-Domain Dual-Field Domain
Decomposition (DFDD):
Decomposes the computational
domain into small subdomains
Computes both the electric and
magnetic fields
Employs a leapfrog time-marching
scheme similar to the FDTD
Couples subdomains by exchanging
surface fields at the interfaces
Physical interpretation as application
of Huygens’ principle or equivalence
principle
1V
2V
3V
4V
Interfaces
E, H
No global interface problems to solve!!!
2
111 02 2
0
1 ir
r c t t
JEE
subdomain 1
1 1
2 21 1 1
ˆn n
s n
J H
subdomain 2
2
222 02 2
0
1 ir
r c t t
JEE
interface 1n̂
2n̂
2
11 12 2
0
1 1ri
r rc t
HH J
1 1 1ˆn n
s n M E
2 2 2ˆn n
s n M E
2
22 22 2
0
1 1ri
r rc t
HH J
Time step n Time step n+1/2 Time step n+1
1 1
2 22 2 2
ˆn n
s n
J H
subdomain 1
subdomain 2
interface 1n̂
2n̂
2
111 02 2
0
1 ir
r c t t
JEE
subdomain 1
1 1
2 21 1 1
ˆn n
s n
J H
subdomain 2
2
222 02 2
0
1 ir
r c t t
JEE
interface 1n̂
2n̂
1 1
2 22 2 2
ˆn n
s n
J H
Dual-field domain decomposition method (DFDD):
DFDD Flowchart
• Break down the entire FEM system into many smaller subsystems to
significantly speed up FEM analysis and reduce memory requirements
Physical Interpretation
E, H
E, H
E, H E, H
PMC
Js
E, H
PEC
Ms
E, H
E, H
E, H
Equivalence principle #1
Original problem Equivalent problem
Equivalence principle #2
E-field calculation
at time step n
H-field calculation
at time step n+1/2
Stability Analysis
Eigenvalue Analysis
TMct
10
21
2
1
0
0
h
e
M
MM
0
0
Q
PT
dSnnnjiP j
S
i
B
NN ˆˆˆ,
dSnnnjiQ i
S
j
B
NN ˆˆˆ,
TDFEM
(central difference)
FDTD 222
0 111
11
zyx
ct
SMc
t1
0
21
TDFEM
(Newmark-Beta)
Method
t
TMct
10
21
DFDD-TDFEM
Stability Condition
Only depend on local mesh
Example: Vlasov Antenna
Number of Subdomains
1 8
Total Number of Unknowns
964,826 8 X
135,414
Peak Memory (MB)
15,000 990
Factorization Time (s)
3,582 105
CPU Time per Step (s)
10.9 0.88
Volmax data: Courtesy of D. Riley.
Computational Performance (Parallel)
Tested on SGI-Altix 350 system
with multiple Intel Itanium II
1.5GHz processors
Each subdomain is assigned to a
different processor
Speedup
10-by-10 Vivaldi Array
2.8 million unknowns
Distributed on 72
processors
Solving time per step: 0.3 s
X-band Phased-Array Antenna Phased-Array Antenna with
Distributed Feed Network
MMIC phase-shifters & T/R modules
256 elements
Applied Radar, Inc.
New applications:
• Wireless communications
• Video games
Phased-Array Antennas
8× 8 Vivaldi Phased Array
E-plane
H-plane
o o45 , 90s sq
8× 1 Vivaldi Phased Array
Single-Stage Divider
8× 1 Vivaldi Phased Array
Wilkinson Divider
Multi-Stage Divider
Explicit TDFEM
m
For element m:
tttc
immmrm
r
JEEE 002
2
2
0
11
im
r
m
r
mrm
r ttcJ
HHH
1112
2
2
0
Weak-Form Representation:
Vector Wave Equation:
2
02 2
0
0 0
1
ˆ
m
m m
m mri m i i
rV
im mi i
V S
dVc t t
dV n dSt t
E EN E N N
J HN N
2
2 2
0
0
1
1ˆ
m
m B
m mri m i i
r rV
i im i m
r rV S
dVc t t
dV n dSt
H HN H N N
N J N EDFDD-ELD
Explicit TDFEM
m
1 12 2
11 1n e e n e n
m m m m m m
n ne e n
m l m l m
e A B e C e
D h E h f
3 1 12 2 2
1
1
n n nh h h
m m m m m m
h n h n n
m l m l m
h A B h C h
D e E e g
For element m:
Pass to the
adjacent elements
Receive from
the adjacent elements
1n
me 1n
le
Pass to the
adjacent elements
Receive from
the adjacent elements
32
n
mh
32
n
lh
2n
me DFDD-ELD
Numerical Example
Mono-conical antenna
DGTD Methods
• J. S. Hesthaven and T. Warburton, “Nodal high-order methods on unstructured
grids,” J. Comput. Phys., vol. 181, pp. 186–211, 2002.
• T. Lu, P. Zhang, and W. Cai, “Discontinuous Galerkin method for dispersive and
lossy Maxwell’s equations and PML boundary conditions,” J. Comput. Phys.,
vol. 200, pp. 549–580, 2004.
• L. Fezoui, S. Lanteri, S. Lohrengel, and S. Piperno, “Convergence and stability
of a discontinuous Galerkin time-domain method for the 3D heterogeneous
Maxwell equations on unstructured meshes,” ESAIM: M2AN, vol. 39, no. 6, pp.
1149–1176, 2005.
• T. Xiao and Q. H. Liu, “Three-dimensional unstructured-grid discontinuous
Galerkin method for Maxwell’s equations with well-posed perfectly matched