UTD Ray and Beam Methods for Analysis of Large EM Wave Problems
Prabhakar H. Pathak ([email protected])
The Ohio State University, ElectroScience Laboratory 1320 Kinnear Road, Columbus Ohio, 43212, USA
2
COPYRIGHT
©The use of this work is restricted solely for academic purposes. The author of this work owns the copyright and no reproduction in any form is permitted without written permission by the author.
3
ABSTRACTIt is well known from the independent works of Keller, Deschamps, and Felsen, respectively, that an electromagnetic (EM) point current source positioned in complex space produces a beam wave field which is highly localized about its forward propagation axis. It is noted that the EM field of a complex source beam (CSB) constitutes an exact solution of Maxwell's equations, and it is simply obtained by an analytic continuation of the exact closed form expression of the EM field for a point current in real space, where the real source coordinates are replaced by complex values. In its paraxial region, a CSB automatically reduces to a Gaussian Beam (GB). By controlling the values of the complex source coordinates, one can produce either a CSB with a very broad ( not well focused ) beam or a very narrow ( highly focused ) beam; consequently, such a CSB field can be made to pass smoothly from the field of a real point source to a plane wave field. It is clear that CSBs can serve as highly useful basis functions to represent EM fields, and indeed they have been used in this fashion by some researchers in this area. Here, additional useful methods for developing convergent CSB expansions to represent the fields of EM sources, via appropriate EM equivalence theorems, will be illustrated. Applications of such CSB expansions to analyze a class of electrically large practical antenna and scattering problems will be presented to demonstrate their utility.
Index Terms: Ray Optics, Diffraction, GTD, UTD, PTD, Beams, Hybrid Methods
4
BIOGRAPHYPrabhakar Pathak received his Ph.D (NVTP) from the Ohio State Univ (OSU). Currently he is Professor (Emeritus) at OSU. He is also a Courtesy Professor and an Adàunct Professor at Univ. of South Florida. Prof. Pathak is regarded as a co-developer of the uniform geometrical theory of diffraction (UTD). His interests continue to be in the development of new UTD solutions in both frequency and time domains, as well as in the development of fast Beam and Hybrid methods, for solving large antenna/scattering problems of engineering interest. Prof. Pathak has been actively presenting short courses and invited talks at conferences and workshops both in the US and abroad. He has authored/coauthored over a hundred journal and conference papers, as well as contributed chapters to seven books. Prior to NVVP, he served two consecutive terms as an Associate Editor of IEEE Trans. AP-S. He was appointed as an IEEE (AP-S) Distinguished lecturer (DL) from 1991-1993. Prof. Pathak was also appointed as the chair of the IEEE AP-S DL program during 1995 – 2005. He served as a member of the IEEE AP-S AdCom in 2010. He received the 1996 Schelkunoff best paper award from IEEE-AP-S; the ISAP 2009 best paper award; the George Sinclair award (1996) from OSU ElectroScience Laboratory; and, IEEE Third Millennium Medal from AP- S in 2000. Prof. Pathak received the Distinguished Achievement Award from IEEE AP-S in 2013. He is an IEEE Life Fellow, and a member of URSI-commission B.
• Two Basic Asymptotic High Frequency (HF)Methodologies can be Categorized as follows:
1. RAY OPTICAL METHODSa) Geometrical Optics (GO)b) Geometrical Theory of Diffraction (GTD)
[GTD = GO + Diffraction]c) Uniform Version of the GTD
2. WAVE OPTICAL METHODSa) Physical Optics (PO)b) Physical Theory of Diffraction (PTD)
[PTD = PO + Diffraction Correction]c) Incremental Theory of Diffraction (ITD) and
Equivalent Current Method (ECM)
UTD UAT
5
ON THE LOCALIZATION OF THE WAVE PROPAGATION AT HF
• The HF localization principle can be demonstrated viaasymptotic evaluation of the radiation integral as depictedbelow:
a) Radiation integral for the scattered field in the spatialdomain.
QE, QR and QS are critical points corresp. to end, stationaryand confluence of two stationary points, respectively ofthe integrand. For other values of Q’ the integral isnegligible due to destructive interference in the asymptoticHF regime.
2
from from from
( ) ( ) ( )
( ) ( | ) ( )
( | ) ,
( ) ( ) ( ) ( )R E S
i s
see S
S
jkR
ee
s r d sd
Q Q Q Q Q Q
E P E P E P
E P P Q J Q dS
er r j I R r rk R
E P E P E P E P
6
b) Radiation integral for the scattered field in the spectraldomain.
QE, QR and QS transform into critical within the spectrum aspoles, saddle points, branch points, etc. Only those planewaves reaching P from the nbhd of QE, QR and QScontribute significantly; all others interfere destructively.
( ) ( , )
( ) ( )
( ) ( ) ( ) ( )
s j k rx y x y
j k rS
S
s r d sd
E P dk dk f k k e
f j C J Q e dS
E P E P E P E P
2 2 2 2 2 2 ;
x y z
z x y x y
k k x k y k z
k k k k j k k k z z
as sumed 7
8
RAY METHODS & SOME APPLICATIONS• Unlike most other computational electromagnetic (CEM) techniques, asymptotic high frequency(HF) ray methods offer a simple picture for describing EM antenna/scattering phenomena.
Examples of EM antenna radiation and coupling problems of interest and some typical UTD rays.
• Rays wave effects are highly LOCALIZED at HF.• Primary focus here will be on the uniform geometrical theory ofdiffraction (UTD) type ray solutions.
• The need for UTD arises because classical geometrical optics (GO)ray method fails to predict diffraction !
ISB: Incident Shadow Boundary, RSB: Reflection Shadow Boundary, SSB: Surface Shadow Boundary
,0,1
,0,1~
~
r
i
rjksr
RRir
ii
jks
Oi
U
U
UesfQRQEPE
Us
ePCPE
r
i
Lit side of RSB
Shadow side of RSB
Lit side of ISBShadow side of ISB * depends only on
surface and wavefront geometry at & near
rsf
RQ
0rE
0&0 ri EE
rE
iE
iE
is
rs
9
• Keller and coworkers (1958; 1962) introduced a new class of rays, i.e. diffracted rays, todescribe diffraction in his geometrical theory of diffraction (GTD).• Diffracted rays exist in addition to geometrical optics (GO) rays.• Diffracted rays are produced at structural and material discontinuities, as well as atgrazing incidence on a smooth convex surface.
0rE
0&0 ri EE
rE
iE
iE
is
rs
GO
(GO + Diffraction)
RAY METHODS & SOME APPLICATIONS (cont.)
d
D
de
jksdDDSS
idS
jksdeEE
ide
esfQQTQEPE
esfDQEPPE
,~
~,
4
42
Examples of diffraction
0rE
0&0 ri EE
rE
iE
is
rs
des
dDs
dSE
dEE
des
DQ
10
RAY METHODS & SOME APPLICATIONS (cont.)• To find and , etc, in diffraction problems, one may:
(a) Solve appropriate, simpler canonical problems which model the LOCALgeometrical and electrical properties of the original surface in theneighborhood of diffraction points.
(b) An exact (or sometimes approximate) solution to a canonical problem isfirst expressed as an integral containing an exponent
(c) Canonical integral is then evaluated asymptotically, generally in closedform, as parameter becomes large (i.e. at HF).
(d) and are then typically found from (c) by inspection.(e) Canonical and generalized to arbitrary shapes by invoking
principle of locality of HF waves.
D
D T
dimensionsticcharacteri
2number wave
D
D TD
• Keller’s original GTD is not valid at and near ISB, RSB, SSB (i.e. in SB transitionregions).
• UTD developed to patch Keller’s original theory within the SB transition regions.
• GTD corrects GO, and GTD = GO + diffraction
• UTD corrects GTD, but usually UTD GTD outside SB transition regions.
D T
11
RAY METHODS & SOME APPLICATIONS (cont.)
• Additional Comments :
(a)Ufimtsev’s Physical Theory of Diffraction (PTD) (1950s) correctsPhysical Optics (PO). PO contains incomplete diffraction.
(b)PTD generally requires numerical integration on theradiating/scattering objects, hence, loses efficiency asfrequency increases.
(c) PTD does not describe creeping/surface wave diffraction onsmooth convex objects; hence, does not accurately predictpatterns in shadow zone of antennas on such complex objects.
(d)Conventional numerical CEM methods become rapidlyinefficient with increase in frequency.
(e) In contrast, UTD ray paths remain independent of frequency.(f) UTD offers an analytical (generally closed form) solution to
many complex problems that can not otherwise be solved in ananalytical fashion.
12
RAY METHODS & SOME APPLICATIONS (cont.)• In many practical applications of UTD, the following diffraction ray mechanisms dominate
[1] R. G. Kouyoumjian and P. H. Pathak,“A uniform geometrical theory ofdiffraction for an edge in a perfectlyconducting surface,” Proc. EEE, vol. 62,pp. 1448-1461, Nov. 1974.
Alternative ray solutions (UAT)[2] S. W. Lee and G. A. Deschamps, “AUniform Asymptotic theory of EMdiffraction by a curved wedge,”lEEETrans. Antennas Propagat., vol. AP-24,pp. 25-34, Jan. 1976.[3] Borovikov, V.A.and Kinber B.Ye,“Some problems in the asymptotic theoryof diffraction”, IEEE Proceeding, volume62, pp. 1416-1437, Nov. 1974.
[1] P.H Pathak, “An asymptotic analysis of the scatteringof plane waves by a smooth convex cylinder,” RadioScience, Vol 14 pp419-435, 1979[2] P.H Pathak et al, “A uniform GTD analysis of thediffraction of EM waves by a smooth convex surface,”IEEE Trans Ant and Propa. Vol 8 Sept 1980.
[1] P.H Pathak et al,”A uniform GTD solution for theradiation from sources on a convex surface,” IEEETrans Ant and Propa. Vol 29 July 1981.
[1] P.H Pathak and N. Wang,”Ray analysis of mutualcoupling between antennas on a convex surface,”IEEE Trans Ant and Propa. Vol 29 Nov 1981.
[1] K.C Hill and P.H Pathak, “A UTDsolution for EM diffraction by a corner ina plane angular sector,” IEEE Ant. Prop.Symp. June 1991.[2] K. C. Hill, “A UTD solution to the EMscattering by the vertex of a perfectlyconducting plane angular sector,” Ph.Ddissertation, The Ohio State University,1990.
(c) PEC Corner Diffraction
[1] G. Carluccio, “A UTD DiffractionCoefficient for a Corner Formed byTruncation of Edges in an OtherwiseSmooth Curved Surface,” IEEE Ant.Prop. Symp. June 2009.
(a) PEC Wedge Diffraction (b) PEC Convex Surface Diffraction
(Alt. Soln. by S.W. Lee in IEEE AP-S)
13
The Ohio State Univ. (OSU) ElectroScience Lab. (ESL) UTD based codes:
(a)OSU-ESL NEWAIR code(b)OSU-ESL BSC code
• Complex radiating and scattering objectsmodeled by simpler shapes consisting ofellipsoids, spheroids, cylinders, cone frustrums,flat plates, etc.
Some UTD code developments in USA during 1980’s – 1990’s
[1] R. G. Kouyoumjian and P. H. Pathak,“A uniform geometrical theory ofdiffraction for an edge in a perfectlyconducting surface,” Proc. EEE, vol. 62,pp. 1448-1461, Nov. 1974.
Alternative ray solutions (UAT)[2] S. W. Lee and G. A. Deschamps, “AUniform Asymptotic theory of EMdiffraction by a curved wedge,”lEEETrans. Antennas Propagat., vol. AP-24,pp. 25-34, Jan. 1976.[3] Borovikov, V.A.and Kinber B.Ye,“Some problems in the asymptotic theory of diffraction”, IEEE Proceeding, volume 62, pp. 1416-1437, Nov. 1974.
(a) PEC Wedge Diffraction
• In many practical applications of UTD, the following diffraction raymechanisms dominate
0
0
1
1 20
1 0 2 0
0 0
20 1
00
0 0 0 0
( ) ( )
lim ( ) ( ) ;
, , , ,
( ) ( )
d
Ed
d
d dd d jksd d d d
d
d dd id
E e d dP Q
e es eh
jksd dd i
E e d d d
E P E P es s
s P P
E P E Q Ds s
D D D
s eE P E Q Ds s
contains 4 terms which are a product of cot2
and
eDn
F kLa
2
UTD EDGE 2 TRANSITION
FUNCTION
jx j
x
F x j xe d e
14
where is theequivalent magnetic current in terms of the transmiting electric field in the slot aperture of area Sa; this replaces the aperture Sa which is now short circuited. Likewise, the radiation from a short thin monopole of height h and transmiting current
fed at the base Q’ on a convex surface can be found as
The UTD solution can predict complex surface dependent polarization effects resulting form surface ray torsion (see terms T1, T2, T3, T4, T5, T6).
( ) ( )S aM Q E Q n
[1] P.H Pathak et al,”A uniform GTD solution for theradiation from sources on a convex surface,” IEEE TransAnt and Propa. Vol 29 July 1981.
( )aE QSM
( )I l
15
The is obtained from uniform asymptotic solutions to problems of radiation by on conducting cylinders and spheres.
UTD Transition functions in A, B, C, D, H, S, n and N are the radiation Fock fcns.
,i m
p
0 at 0 at
S
L
PP
16
19
Geodesic surface ray
cylindercircular afor offunction a as )T(Q' a'
75 spheroid. prolate aon apart) 90 (phased antennaslot crossed Lindberg a of patternsRadiation
'
't
'b
'Q 'ˆ2
'ˆ1
't
'b
'Q
'n
t
b
n
curvature. of radii principal are R and R
Q'.at directionssurface principal denote 'ˆ and'ˆ
21
21
'at ;112
'2sin'
Torsion ; '''
1212
QRRRR
QT
TQQTQT gO
[1] P.H Pathak and N. Wang,”Ray analysis of mutualcoupling between antennas on a convex surface,”IEEE Trans Ant and Propa. Vol 29 Nov 1981.
(Alt. Soln. by S.W. Lee in IEEE AP-S)
contain the UTD transition functions corresponding to surface Fock fcns.
, , and ee he eh hh
20
21
CalculatedMeasured
Modeling of Boeing 737 aircraft
C. L. Yu, W. D. Burnside, and M. C. Gilreath, “Volumetric pattern analysis ofairborne antennas,” IEEE Trans. AP, Sep. 1978.
22
• J. J. Kim and W. D. Burnside, “Simulation and Analysis of Antennas Radiating in a ComplexEnvironment”, IEEE Trans. AP, April 1986.
23
-110
-100
-90
-80
-70
-60
0 4 8 12 16 20 24 28 32 36
Receive Antenna Distance from Nose (in.)
Cou
plin
g (d
B)
12 GHz (meas)12 GHz (calc)
wings not presentincludes double diffractionComputed by Dr. R. J. Marhefkaat OSU-ESL. This work was supported by Gary Roan at NRL
Comparison of Measured (NRL) and Calculated (NEC-BSC)Antenna Isolation with Receiver Moving above Center of Fuselage
24
Limitations of Existing UTD Codes• Existing UTD codes such as NEC-BSC and NEW-AIR have proven to
be successful over the past two decades.• However, these codes are based on the approximation of the
electrically large airborne platform in terms of canonical shapes, whichis a complicated task.
• Moreover, a canonical shape representation may lead toinaccuracies.
• Very limited capability to analyze material coatings
Canonical representation for NEC-BSC Canonical representation for NEW-AIR
25
New UTD code development
•Radiating object modeled with better fidelity viafacets based on CAD geometry data.
•UTD rays tracked in presence of facets. UTDray parameters obtained by mapping facets backto the original geometry (via bi-quadraticsurfaces or splines, etc).
•Does not require an expert user.
•Will eventually incorporate thin material coatingon metallic (or PEC) platform.
26
Quarter wavelength monopole on a KC-130 at 500 MHz(18.3λ0 height, 69.2λ0 wingspan, 49.7λ0 length at 500 MHz).
New code (Applied EM)
27
Quarter wavelength monopole on a KC-130 at 500 MHz(18.3λ0 height, 69.2λ0 wingspan, 49.7λ0 length at 500 MHz).
Mesh Geometry
Frequency (GHz)
# of facets for UTD
CPU time for UTD
# of facets for MLFMM
CPU time for MLFMM
KC-135 0.500 6,496 3.57 sec 420,466 1 h 28 min
Preliminary Validation
Near Field Far Field
32
New UTD analytical development
•UTD + slope for PEC planar faced wedges with thinmaterial coating in form useful and accurate forengineering applications.
•UTD + slope for curved PEC wedges with thin materialcoating.
•UTD for PEC corners in curved edges and surfaces withthin material coating.
•UTD for PEC edge excited surface (or creeping) rays(and its reciprocal problem) in a form useful forengineering applications.
• Material coatings are generally replaced by approximate boundary conditions (e.g. Impedance Boundary Condition [IBC])• Solutions to canonical problems with “approx.” boundary conditions formulated exactly via Wiener-Hopf (W-H) or
Maliuzhinets (MZ) methods for surfaces made up of planar structures. Ray solutions extracted analytically from them viaasymptotic procedures.
• W-H method:
– J. L. Volakis and T. B. A. Senior, “Diffraction by a Thin Dielectric Half-Plane”, IEEE Trans. AP, Dec.1987
– R. G. Rojas, “Wiener-Hopf Analysis of the EM Diffraction by an Impedance Discontinuity in a Planar Surface andby an Impedance Half-Plane”, IEEE Trans. AP, Jan. 1988.
– R. G. Rojas and P. H. Pathak, “Diffraction of EM Waves by a Dielectric/Ferrite Half-Plane and RelatedConfigurations”, IEEE Trans. AP, June 1989.
– J. L. Volakis and T. B. A. Senior, "Application of a Class of Generalized Boundary Conditions to Scattering by aMetal-Backed Dielectric Half Plane”, Proc. IEEE, May 1989.
– V. G. Daniele and G. Lombardi, “Wiener-Hopf Solution for Impedance Wedges at Skew Incidence”, IEEE Trans.AP, Sep. 2006 .
• MZ method:
– G. D. Maliuzhinets, “Excitation, Reflection and Emission of Surface Waves from a Wedge with Given FaceImpedance”, Sov. Phys.-Dokl., 1958.
– R. G. Rojas, “Electromagnetic Diffraction of an Obliquely Incident Plane Wave Field by a Wedge with ImpedanceFaces”, IEEE AP, July 1988.
– R. Tiberio and G. Pelosi and G. Manara and P. H. Pathak, “High-Frequency Scattering from a Wedge withImpedance Faces Illuminating by a Line Source, Part I: Diffraction”, IEEE Trans. AP, Feb. 1989, see also IEEETrans. AP, July 1993.
– M. A. Lyalinov and N.Y. Zhu, “Diffraction of a Skew Incident Plane Electromagnetic Wave by an ImpedanceWedge”, Wave Motion, 2006.
• Approx. skew incidence solution (MZ) for imp. wedges based on modifying the HP solution:– H. Syed and J. L. Volakis, “Skew incidence diffraction by an impedance wedge with arbitrary face impedances”,
Electromagnetics, Vol. 15, No.3, 1995.
Previous Work on Thin Material Coated Metallic Wedge Structures
33
An Approximate UTD Ray Solutionfor Skew Incidence Diffraction
by Material Coated Wedges of Arbitrary AngleT Lertwiriyaprapa, P. H. Pathak and J. L. Volakis
ElectroScience LaboratoryDepartment of Electrical and Computer Engineering
The Ohio State University
URSI, Chicago 2008
• Present solution based on spectral synthesis.• Solution useful and accurate for engineering applications.
3232
0
30
6090
120
150
180
-40
-20
0dB
0
30
6090
120
150
180
-40
-20
0dB
0
30
6090
120
150
180
210
240 300
330
-20
-10
0dB
Grounded Material Junction
UTD-MZUTDMZ
0
30
6090
120
150
180
210
240 300
330
-20
-10
0dB
Grounded Material Junction
UTD-MZUTDMZ
y
x
0
0
0
ˆ 0
ds
is
Semi infinite Material Slab
Diffracted Field
ro , ro
z
rn , rn
Scatt. Field
Total Field
E-pol H-pol
• MZ is R. G. Rojas, “Electromagnetic Diffraction ofan Obliquely Incident Plane Wave Field by a Wedgewith Impedance Faces”, IEEE AP, July 1988.
PW illumination
Numerical Results(3-D Junction Planar Material Slabs on a PEC Ground Plane)
•Comparison of UTD-MZ, UTD and MZ at r=5, =45, o=65, =/20,ro=2, ro=4, rn=5, and rn=1.
35
0
30
6090
120
150
180
210
240270
300
330
-40
-20
0dB
Coated PEC Wedge
PEC
UTD-MZMZ
0
30
6090
120
150
180
210
240270
300
330
-40
-20
0dB
Coated PEC Wedge
PEC
UTD-MZMZ
0
30
6090
120
150
180
210
240270
300
330
-40
-20
0dB
Coated PEC Wedge
UTD-MZMZ
0
30
6090
120
150
180
210
240270
300
330
-40
-20
0dB
Coated PEC Wedge
UTD-MZMZ
36
Scatt. Field
Total Field
r
• MZ is R. G. Rojas, “Electromagnetic Diffraction ofan Obliquely Incident Plane Wave Field by a Wedgewith Impedance Faces”, IEEE AP, July 1988.
E-pol H-pol
PW o-face illumination
Numerical Results(3-D Material Coated PEC Right-Angled Wedge)
•Comparison of UTD-MZ and MZ at r=10, =45, o=65, =/20, rn=2, and rn=5.
3734
Total Field
• MZ is R. G. Rojas, “Electromagnetic Diffraction ofan Obliquely Incident Plane Wave Field by a Wedgewith Impedance Faces”, IEEE AP, July 1988.
0 20 40 60 80 100 120 140 160 180-40
-20
0
20
HH
-pol
[dB
]
in Degrees
in Degrees0 20 40 60 80 100 120 140 160 180
-15
-14
-13
-12
-11
-10
HE
-pol
[dB
]
UTD-MZMZ
0 20 40 60 80 100 120 140 160 180-16
-14
-12
-10
EH
-pol
[dB
]
in Degrees
0 20 40 60 80 100 120 140 160 180-30
-20
-10
0
10
in Degrees
EE
-pol
[dB
]
UTD-MZMZ
0
0
0
ˆ 0
ds
is
Numerical Results(3-D Junction Planar Material Slabs on a PEC Ground Plane)
•Comparison of UTD-MZ and MZ at r=10, =45, o=65, =/20, ro=4, ro=2,rn=2, and rn=5.
3835
Total Field
• MZ is R. G. Rojas, “Electromagnetic Diffraction ofan Obliquely Incident Plane Wave Field by a Wedgewith Impedance Faces”, IEEE AP, July 1988.
r
0 50 100 150 200 250 300 350-100
-50
0
50
HH
-pol
[dB
]
in Degrees
0 50 100 150 200 250 300 350-80
-60
-40
-20
0
in Degrees
HE
-pol
[dB
]
UTD-MZMZ
0 50 100 150 200 250 300 350-80
-60
-40
-20
0
EH
-pol
[dB
]
in Degrees
0 50 100 150 200 250 300 350-80
-60
-40
-20
0
20
in Degrees
EE
-pol
[dB
]
UTD-MZMZ
Numerical Results(3-D Material Coated PEC Half Plane)
•Comparison of UTD-MZ and MZ at r=10, =45, o=65, =/20, ro=4, ro=2,rn=2, and rn=5.
3936
Total Fieldr
• MZ is R. G. Rojas, “Electromagnetic Diffraction ofan Obliquely Incident Plane Wave Field by a Wedgewith Impedance Faces”, IEEE AP, July 1988.
0 50 100 150 200 250-150
-100
-50
0
EH
-pol
[dB
]
in Degrees
0 50 100 150 200 250-80
-60
-40
-20
0
20
EE
-pol
[dB
]
UTD-MZMZ
0 50 100 150 200 250-60
-40
-20
0
20
HH
-pol
[dB
]
in Degrees
0 50 100 150 200 250-120
-100
-80
-60
-40
-20
0
HE
-pol
[dB
]
UTD-MZMZ
Numerical Results(3-D Material Coated PEC Right-Angled Wedge)
• Comparison of UTD-MZ and MZ at r=10, =45, o=65, =/20, ro=4, ro=2,rn=2, and rn=5.
0
30
6090
120
150
180
210
240270
300
330
-40
-20
0
Coated PEC Wedge
TETM
40
Scatt. Field Total Field
0
30
6090
120
150
180
210
240270
300
330
-40
-20
0
Coated PEC Wedge
TETM
Numerical Results(3-D Material Coated PEC Wedge, WA = 54o)
• r=5, =117, =66, r=5,=/20,rn=2.4, and rn=8.
Slope diffraction is included
z-directed currentmoment excitation
ANTENNAS ON CONVEX COATED STRUCTURESKittisak Phaebua and Prabhakar Pathak
• A Uniform Geometrical Theory of Diffraction (UTD) Ray Solution is developedto predict the radiation by antennas on smooth convex metallic surfaces withthin material coating.
• Metallic surface is assumed to be a perfect electric conductor (PEC).
• Thin coating ; ; = surface radii of curvatureAlso,
• For sufficiently thin material coating, one can approximate the actual boundary on the external surface by a surface impedance Zs
gkd k 2k
g
)( , rr
ˆ ˆ sn n E Z n H121 ( )1 (1 )
2o
r rr r
rs jZ kd kdZ
sZ
n
Arbitrarily oriented electric or magnetic point current at Q’ on external boundary
d (coating thickness)
PL
Q’
Q
PsPEC
Direct Ray
Surface diffracted Ray
Geodesic surface ray
Thin uniform material coating
1kd
41
MOTIVATION•UTD Ray Analysis can be applied to analyze radiation by conformal antennas
and antenna arrays in the presence of a smooth PEC convex surface with thinmaterial coating.
Single printed patchPrinted patch array
Printed cross dipole element
dPEC
Single slotSlot array
dPEC
Single monopole
42
SOME PREVIOUS RELATED WORK[1] P. Munk and P. H. Pathak, "A UTD Analysis of the Radiation and Mutual Coupling Associated with
Antennas on a Smooth Perfectly Conducting Arbitrary Convex Surface with a Uniform Material Coating,"Antennas and Propagation Society International Symposium, vol. 1, pp. 696 - 699, Jul. 1996.- UTD ray solution not in form convenient for applications. Also, not all UTD transition functions computed.
[2] N. A. Logan and K. S. Lee, "A Mathematical Model for Diffraction by Convex Surface," In Electromagneticwaves. R. ranger, Ed, Univ. Wisconsin Press, 1962.- No specific ray solution for radiation available.
[3] Wait, J. R., Electromagnetic Waves in Stratified Media, A Pergamon Press Book, McMillan Co., New York,1962.- Propagation of waves around the earth, spherical surface analyzed. No UTD ray solution presented- Similar to work by V. A. Fock, Electromagnetic Diffraction and Propagation Problems, New York, PergamonPress, 1965 (Original work in Russian was published in 1940s)
[4] L. W. Pearson, “A scheme for automatic computation of Fock-type integrals,” IEEE Trans. AntennasPropagat.,vol. AP-35, pp. 1111–1118, Oct. 1987.- Solution presented for only the scattering into shadow region of a coated circular cylinder.
[5] C. Tokgöz, P. H. Pathak and R. J. Marhefka," An Asymptotic Solution for the Surface Magnetic Field Withinthe Paraxial Region of a Circular Cylinder With an Impedance Boundary Condition", IEEE Trans. AntennasPropagat., vol. 53, no. 4, April 2005.- Mostly restricted to surface fields on cylinders due to point magnetic currents on the same surface.
[6] P. H. Pathak, N. Wang, W. D. Burnside and R. G. Kouyoumjian, “A uniform GTD solution for the radiationfrom sources on a convex surface”, IEEE Trans. Antennas Propagat., vol. AP-29, no. 4, pp. 609-622, July 1981.- UTD analysis restricted to smooth convex PEC surfaces.
[7] P. H. Pathak, R. J. Burkholder, Y. Kim and J. Lee, "A Hybrid Numerical-Ray Based Analysis of Large ConvexConformal Antenna Array on Large Platforms," Presented at ACES conference in Finland, April, 2010.- Hybrid numerical UTD solution restricted to complex antennas on locally smooth convex PEC surfaces.
43
ANALYTICAL FORMULATION• A UTD solution for radiation by an arbitrarily oriented or on an arbitrary
smooth convex surface with a uniform surface impedance boundary condition (IBC)is developed from canonical solutions.
(Prabhakar Pathak & Kittisak Phaebua)
• Canonical problems to be solved pertain to (or ) with arbitrary orientation on circular cylinders and spheres with IBC.
• Generalization of canonical solutions to arbitrary convex surface performedheuristically based on the principal of locality of HF wave phenomena
ed p md p
( )edp Q ( )mdp Q
(a) Canonical circular cylinder problem geometry (b) Canonical spherical problem geometry
44
10 20 30 40
30
210
60
240
90270
120
300
150
330
180
0
Ray SolutionCST-Microwave Studio
NUMERICAL RESULTS (CYL)Radius of cylinder 4Thickness of dielectric coating 0.02Length of cylinder 50
2 .1 (T eflo n )r 1r
a t 8 0 oE a t 6 0 oE
a t 9 0 oE
Frequency of operation = 10 GHz
10 20 30 40
30
210
60
240
90270
120
300
150
330
180
0
Ray SolutionCST-Microwave Studio
10 20 30 40
30
210
60
240
90270
120
300
150
330
180
0
Ray SolutionCST-Microwave Studio
ˆNormal electric current source, ( . )J J n
45
NUMERICAL RESULTS (CYL)
a t 9 0 oE
ˆTangential magnetic current source, M ( . )t M b
Radius of cylinder 4Thickness of dielectric coating 0.02Length of cylinder 50
2 .1 (T eflo n )r 1r
Frequency of operation = 10 GHz
10 20 30 40
30
210
60
240
90270
120
300
150
330
180
0
Ray SolutionCST-Microwave Studio
46
NUMERICAL RESULTS (SPH)
nE E
ˆNormal electric current source, ( . )rJ J n ˆTangential magnetic current source, M ( . )t M b
Radius of sphere 4Thickness of dielectric coating 0.02 2 .1 (T eflon ) ; r 1r
Frequency of operation = 10 GHz
10 20 30 40 50
30
210
60
240
90270
120
300
150
330
180
0
Ray SolutionCST-Microwave Studio
10 20 30 40
30
210
60
240
90270
120
300
150
330
180
0
Ray SolutionCST-Microwave Studio
nE E
47
48
A UTD Diffraction Coefficient for a Corner Formed by Truncation of Edges
in an Otherwise Smooth Curved Surface
Giorgio Carluccio(1), Matteo Albani(1), and Prabhakar H. Pathak(2)
(1) Department of Information Engineering, University of SienaVia Roma 56, 53100 Siena, Italy, http://www.dii.unisi.it
(2) ElectroScience Laboratory, The Ohio State University1320 Kinnear Road, 43212 Columbus – OH, USA,
http://electroscience.osu.edu
IEEE International Symposium on Antennas and Propagation and USNC/URSI NationalRadio Science Meeting
June 01-05, 2009
49
UTD Vertex Diffraction CoefficientShadow Boundary Cones (SBCs) and Shadow Boundary Planes (SBPs):
50 -180 -150 -120 -90 -60 -30 05
10
15
20
25
30
[dB
]
Tot UTDTot Uniform Asym PO
-180 -150 -120 -90 -60 -30 0-10
-5
0
5
10
15
20
25
30
[dB
]
GOD-UTD-ABD-UTD-DAV-UTD-ATot
-4 -2 0 2-4-202
-3
-2
-1
0
1
2
3
4
5
BC
x
A
D
z
y
III Example: Vertex Double Transition
Scan Center on the Vertex A
FieldE
3 , 45 , 180 0r
We consider a smooth convex parabolic surface illuminated by an electric point source
FieldE
51
Remarks• A UTD diffraction coefficient for a corner formed by truncation of edges in asmooth curved surface was presented.
• A PO diffraction coefficient is derived by asymptotical evaluation of the POintegral, to understand how the surface curvature affects the diffracted fieldtransitional behavior.
• The UTD diffraction coefficient was obtained by heuristically modifying theUTD diffraction coefficient for a corner in a flat surface, on the basis of theprevious PO result.
• Numerical examples show how the proposed diffracted coefficient smoothlycompensates for the abrupt discontinuity occurring when the GO field or thesingly diffracted at edges abruptly vanish.
• Valid for astigmatic ray tube illumination.
• Can be extended to include thin material coating.
53
i to
'Pie
Edge excited surface rays
• Presently UTD solution has been obtained for ISB and SSB farapart.
• Work is in progress to obtain an asymptotic solution useful forengineering applications when ISB and SSB regions overlap.
54
Comments• Keller’s original GTD is not valid at and near ISB, RSB, SSB (i.e. in
SB transition regions).
• UTD developed to patch Keller’s original theory within the SBtransition regions.
• GTD corrects GO, and GTD = GO + diffraction
• UTD corrects GTD, but usually UTD GTD outside SB transitionregions.
• UTD ray paths remain independent of frequency.
• UTD offers an analytical (generally closed form) solution to manycomplex problems that can not otherwise be solved in an analyticalfashion.
• UTD in some cases must be augmented by PO/PTD or ECM
• PO/PTD can give rise to spurious contributions from theshadow boundary line in a smooth body.
• PO/PTD does not incorporate creeping wave effects.
•PO can correct for UTD transport singularities at and nearthe confluence of edge induced GO ray shadow boundariesand GO/diffracted ray caustics (e.g. forward radiation fromparabolic reflector antennas).
RSB + diffracted caustic
offset paraboloidal reflector
reflected ray caustic
feed
55
SPECTRAL THEORY OF DIFFRACTION (STD)
• UTD/GTD requires a RAY OPTICAL incident field
• If the incident field is NON-RAY OPTICAL, then it must berepresented by:a) continuous set of PLANE WAVES (e.g. Plane Wave
Spectrum)b) discrete set of RAY OPTICAL fields
• Each constituent RAY OPTICAL field in the SPECTRALDECOMPOSITION of a NON-RAY OPTICAL incident wavecan be treated by UTD.
• The total UTD solution is then a summation of the UTDresponse to each constituent RAY OPTICAL incident field
R. Tiberio, G. Manara, G. Pelosi, R. KouyoumJian: ‘’HF EM scattering of plane waves fromdouble wedges,” IEEE Trans AP-37, pp. 1172-1180, Sept. 1989
Y. Rahmat-Samii, R. Mittra ‘’A spectral domain interpretation of HF phenomena,” IEEETrans AP-25 pp. 676-687, Sept. 1977
56
ADDITIONAL COMMENTS ON ASYMPTOTIC HF METHODS
• Asymptotic HF methods are powerful for analyzing a wide varietyof electrically large EM problems.
• Conventional CEM numerical solution methods based on selfconsistent wave formulations become rapidly inefficient, or evenintractable, for solving large ‘’EM problems’’.
• UTD is more developed, especially for handling smooth convexboundary diffraction.
• HF wave optical methods (PO, PTD, ITD, SBR) have not directlyincorporated creeping waves.
• Ray optical methods require ray tracing. More efficient but lessrobust. Ray paths independent of frequency.
• Wave optical methods require numerical integration on the largeobject. Less efficient but more robust. Does not scale withfrequency.
60
62
Beam Methods & Some Applications• Beams provide useful basis functions for representing EM fields.• Ray methods fail at caustics (focii) of ray systems. Caustics are formed byintersection or envelopes associated with the same class of rays. Beamsremain valid in regions of real ray caustics
• Beams can be used to treat large reflector systems and radome problemsefficiently.• Beams can be used to improve the speed of conventional Moment Method(MoM) solution of governing EM integral equation (IE) for the radiating object.• Beams can also be used for Near Field Far Field transformations requiredin near field measurements.
62
• UTD for real source excitation of wedge developedvia first order Pauli-Clemmow method (PCM) [1-4] forasymptotic solution of canonical wedge diffractionintegral along a steepest descent path (SDP)•First order PCM not strictly valid (for poles crossingthe SDP away from saddle point); hence analyticcontinuation of UTD for complex source locationwithout further study is questionable!•First order Van der Waerden method (VWM) [1-4] isvalid even where PCM fails.
y
x
EQ~Oz
)2(
nW
A
,,r
'~,'~,'~ r
Complex Source Beam (CSB) Diffraction by a Wedge
• However, one can show that the first order VWM method, upon using a key rearrangement(and combination) of terms, yields:
PCM) (Extended EPCMΔPCMVWM Key Step• Next, for the special wedge case of interest, it is shown analytically (and verified numerically)that .0
)negligible is (since UTD UTD EUTD
for wedge EUTD EPCM VWM for wedge UTD PCM :Note
For a wedge
Therefore, analytic continuation of UTD for a wedge is OK for
complex waves
1. T. B. A. Senior and J. L. Volakis, “Approximate Boundary Conditions in Electromagnetics,” The Institute of Electrical Engineerings, London, 1995.2. L. B. Felsen and N. Marcuvitz, “Radiation and Scattering of Waves,” Englewood Cliffs, NJ: Prentice-Hall, 1973.3. C. Gennarelli and L. Palumbo, “A uniform asymptotic expansion of typical diffraction integrals with many coalescing simple pole singularities and a
first-order saddle point,” IEEE Trans. Antennas and Propagat., vol. AP-32, pp. 1122-1124, Oct. 1984.4. R. G. Rojas, “Comparison between two asymptotic methods,” IEEE Trans. Antenn- as and Propagation, vol. 35, no. 12, pp 1492-1493, Dec 1987.
61
Numerical Result 1
bbb ˆ
y
x
eE zQ ~,0,0~ Oz
'~,'~,'~ r
)2(
nW
A
bb
,,rP
r 'r
0 50 RSB(135.86) ISB(224.14) 3000
1
2
3
4
5
6
7x 10-3
,degM
ag
Einc Eref EUTDdiff EUTD
total EEUTDdiff EEUTD
total
0 50 100 150 ISB(224.14) 300
0.5
1
1.5
2
, deg
0 50 RSB(135.86) 200 250 300
0.5
1
1.5
2
, deg
V2a V2
b V2=V2a*V2
b
V2a V2
b V2=V2a*V2
b
222.5 223 223.5 224 224.5 225 225.5 226-8
0
8x 10
-3
, Deg
Im[pi ]
134 134.5 135 135.5 136 136.5 137 137.5-8
0
8x 10
-3
, Deg
Im[pr ]
• Additional term in EUTD solution• Trajectories of
0Im rp
0Im ip
nss
nV
p
pp
a
2
~cot'~~
2~cos
~~
4,2
~~1~4,2
kFV pb
rip,Im
30WA
10b
6.60b 4.217b
2r 45zp ri ˆˆ ,
5'r 140' 50'
Complex Source Beam (CSB) Diffraction by a Wedge(cont.)
62
0 50 RSB(112.37) 150 200 ISB(247.67) 3000
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
x 10-3
, degM
ag
Einc Eref EUTDdiff EUTD
total EEUTDdiff EEUTD
total
Numerical Result 2
246.5 247 247.5 248 248.5 249-0.015
0
0.015
, Deg
Im[pi ]
111 111.5 112 112.5 113 113.5-0.015
0
0.015
, Deg
Im[pr ]
0 50 100 150 200 ISB(247.67) 300
0.5
1
1.5
2
, deg
0 50 RSB(112.37)150 200 250 300
0.5
11.5
22.5
, deg
V2a V2
b V2=V2a*V2
b
V2a V2
a V2=V2a*V2
b
bbb ˆ
y
x
eE zQ ~,0,0~Oz
'~,'~,'~ r
)2(
nW
A
bb
,,rP
r 'r
• Additional term in EUTD solution• Trajectories of
0Im rp
0Im ip
nss
nV
p
pp
a
2
~cot'~~
2~cos
~~
4,2
~~1~4,2
kFV pb
rip,Im
45WA
10b
57.78b 57.258b
12r 40
zp ri ˆˆ ,
20'r 100' 70'
Complex Source Beam (CSB) Diffraction by a Wedge(cont.)
66 SB1(5.3) 40 60 80 100 120 140 SB2(174.7)0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
, deg
Mag
HzPCM Hz
EPCM
Surface Wave 58.35jZS
0 20 SB1(54.7) 80 100 SB2(125.3) 160 1800
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
, deg
Mag
HzPCM Hz
EPCM
Leaky Wave 79.1794.327 jZS
2sin2 2,14
2,1
ppj
pp e
M
y
x
jXRZ s Shadow boundaries are determined when
0Im 2,1 pp
Other Complex Waves
68
Sequence of CSP beams
Feed Antenna
Reflected & Diffracted Fields of each CSP beam
Feed Radiation Pattern
Fast analysis of the reflector antennas.Radome
AntennaCSP beams
Antenna radiation in the presence of radomes.
• A GB -UTD (PO based) method was previouslyreported in [1].
• With the CSP method, feed pattern is expandedinto a set of CSBs.
• Each CSP beam field is scattered from thereflector by using complex extension of UTD.
• A 2-D case for a single beam illumination wasreported in [2].
• This new fully 3-D CSP-UTD approach (UTD forbeams) is expected to be more accurate then [1].
• The field of the antenna is first expanded intoa set of CSP beams.
• Each beam is next tracked through theradome.
• The transmitted beams are summed up atthe observer location.
• Complex ray tracing can be employed forbeam tracking through the radome [3,4].
[1] H. T. Chou, P. H. Pathak and R. J. Burkholder, “Novel Gaussian Beam Method for the Rapid Analysis of Large Reflector Antennas”, IEEE Trans. AntennasPropagat., 2001[2] G.A.Suedan and E.V. Jull, ”Beam diffraction by planar and parabolic reflectors,” IEEE Trans. Antennas Propagat., 1991.[3] X. J. Gao and L. B. Felsen, “Complex ray analysis of beam transmission through two-dimensional radomes”, IEEE Trans. Antennas Propagat., 1985[4] J. J. Maciel and L. B. Felsen, “Gaussian beam analysis of propagation from an extended plane aperture distribution through dielectric layers, part 1 - planelayer,” IEEE Trans. Antennas Propagat., 1990.
Large Antenna Applications
69
'~r
b
p • The CSB-UTD solution is valid for analyzing CSBexcited focus-fed parabolic reflector antennas since the caustics are now in complex space for the CSB excitation case.
• The PO analysis for a CSB excited parabolic reflector(a) loses its accuracy in the region of the main beam
when a CSB axis hits near the edge. can beimproved by adding the additional edge diffractionterm based on Physical Theory of Diffraction (PTD).
(b) becomes more accurate when a CSB axis hits theactual reflector surface away from the edge.
• The present CSB-UTD & CSB-EUTD solution for aCSB excited PEC curved wedge is obtained by analyticallycontinuing the UTD solution for a PEC curved wedgeexcited by a real point source (or even real astigmatic ray)to deal with a CSB (or even more generally a complexastigmatic beam, i.e. CAB) illumination of a curved edgein a curved surface.
• The CSB reflected from doubly curved surface become anastigmatic Gaussian beam in paraxial region.
mfyxz
4
22
CSB-UTD Diffraction by a Curved Wedge
70 -50 0 50 100 150 200 250
0.2
0.4
0.6
0.8
1
1.2
, deg
MA
G
EPO
ECSB-EUTD
ECSB-UTD
Numerical Result : Finite parabolic reflector illuminated by a CSB
-50 0 50 100 150 200 250
0.2
0.4
0.6
0.8
1
1.2
1.4
, deg
MA
G
EPO
ECSB-EUTD
ECSB-UTD Mfyxz
4
22
b
5000r
S S
)270,100,40(),50,10(,ˆˆ,10,35,5,60 SQbpbfhd BMM
Symmetry Plane
Transverse Plane
CSB-UTD Diffraction by a Curved Wedge (cont.)
71
Numerical Result : Finite parabolic reflector illuminated by a CSB
-50 0 50 100 150 200 250
0.1
0.2
0.3
0.4
0.5
, deg
MA
G
EPO
ECSB-EUTD
ECSB-UTD
-50 0 50 100 150 200 250
0.1
0.2
0.3
0.4
0.5
0.6
0.7
, deg
MA
G
EPO
ECSB-EUTD
ECSB-UTD Mfyxz
4
22
y
z
b
5000r
QB
SfM
y
xOO
QB
S
dM
h
)270,100,40(),60,20(,ˆˆ,10,35,5,60 SQbpbfhd BMM
Symmetry Plane
Transverse Plane
CSB-UTD Diffraction by a Curved Wedge (cont.)
72
Numerical Result : Finite parabolic reflector illuminated by a CSB
-50 0 50 100 150 200 250
0.05
0.1
0.15
0.2
0.25
, deg
MA
G
EPO
ECSB-EUTD
ECSB-UTD
-50 0 50 100 150 200 250
0.05
0.1
0.15
0.2
, deg
MA
G
EPO
ECSB-EUTD
ECSB-UTDMfyxz
4
22
b
5000r
S
S
)90,0,40(),10,17(ˆˆ,10,30,50 SQxpbfd BMM
Symmetry Plane
Transverse Plane
CSB-UTD Diffraction by a Curved Wedge (cont.)
73
Numerical Result : Finite parabolic reflector illuminated by a CSB
-50 0 50 100 150 200 250
0.5
1
1.5
2
, deg
MA
G
EPO
ECSB-EUTD
ECSB-UTD
-50 0 50 100 150 200 250
5
10
15
20
, deg
MA
G
EPO
ECSB-EUTD
ECSB-UTDMfyxz
4
22
y
z
b
200rQB
SfM
y
xOO
QB
S
dM
)0,10,40(),35,20(ˆˆ,15,30,50 SQbpbfd BMM
Symmetry Plane
Transverse Plane
CSB-UTD Diffraction by a Curved Wedge (cont.)
74
Feed Radiation
GB Expansion
n = launching points in feed plane
m = number of GBs from each n
CSBs/GBs Illuminating a reflector
75
< 200 GBs
NUM-PO
Time < 5 min/iter
Time = 5 or 6 hrs/iter
Approx. 30 iter’s*
Offset Shaped Reflector for CONUS Contour Beam Using GBs
85D
* Rahmat-Samii’s paper
Normalized co-polarized contours based on the GB approach CONUS coverage by a shaped concave reflector with a feed pattern at 12 GHz with l =18.51. Approximately 200GBs were used.
Normalized co-polarized gain contours based on the numerical PO integration approach for the same shaped reflector case as above. ?
76
Conclusions
• CSB expansion methods for EM radiation arepresented employing three different variants of thesurface equivalence theorem.
• The analytical properties (validity region, truncation,etc.) of the approach are investigated.
• It is shown that accurate and efficient fieldrepresentations can be obtained by convenientlytruncating the beam expansion.
• It is demonstrated that the expansion idea is applicableto a class of EM radiation/scattering problems.
77
Band Width/Complexity Materials
Elec
tric
al S
ize
UTD
PO/GO
MOM
FEM
MLFMM
Hybrid Methods are required for in-situ
analysis
Source: http://www.feko.info/feko-product-info/technical
Computational Methods
78
Hybrid Method & Some Applications• In many applications, large antennas (arrays) and largeantenna platforms contain both large and small featuresin terms of wavelength.
•For electrically large parts on radiating object, UTD raymethod is useful but not valid for electrically smallportions.
•For highly inhomogeneous and electrically small region(e.g. complex antenna elements/arrays) the FE-BI ornumerical methods are useful, and UTD is not applicablehere.
•A hybrid combination of FE-BI (or other suitablenumerical methods) and UTD could handle the entireproblem not otherwise tractable by each single approachby itself.
79
Simple slot phased array in a convex PEC surface
Material Treatment
Tapered Absorber
Antenna Array Elements
Radome Cover Flush with Platform
PEC Platform
PEC Platform
Complex phased antenna array slightly recessed in a convex platform and covered by a radome
Conformal Array Configurations
80
Local Array Parttreated by full wave numerical methods.
Present Collective UTD Solution
converts numerical array solution into rays launched from
array aperture.
External Platform PartCollective UTD rays launched
from aperture efficiently excite external platform which
is analyzed by UTD.
Proposed Hybrid Numerical-UTD Approach
81
Actual problem
A local array modeling for FEM
PEC
Local UTD RayeqsM
Structure outside the local aperture region is ignored
A local array modeling for FE-BI
ABC
Local Array Part Treated by FEM, FE-BI
82
• Collective UTD rays launched by arrayaperture distribution obtained via numericalsolution to local array part.
• Collective UTD rays launched from a few flashpoints in the array aperture.
Observation point
Collective UTD RaysPanuwat Janpugdee and Prabhakar H. Pathak
83
• Collective UTD rays efficiently launched from arrayaperture interact with external platform.
• Rays from platform could interact back with the arrayand so on. These effects can be included if desired.
• Collective UTD surface rays provide direct couplingbetween two arrays located on the same platform.
Convex Surface
Observation point
External Platform Interaction Part
84
Less efficient for large arrays – need to trace a large number of rays (with all the constructive & destructive interferenceeffects).
Lacks physical picture for describing collective array radiation and surface field excitation mechanisms.
Integration to existing UTD codes for predicting the platform interaction effects via UTD is straightforward –but less efficient.
Comparison with Conventional Element-by-element UTD Field Summation Approach
85
• Asymptotic UTD ray solutions for the radiation andsurface fields produced by a single point currentsource on a smooth convex surface have beendeveloped previously by Pathak et al.
Pathak et al., IEEE Trans. AP, vol. AP-29, pp. 609-622, Jul. 1981.Pathak et al., IEEE Trans. AP, vol. AP-29, pp. 911-922, Nov. 1981.
nn
lt
lb
s
Q
LP
i
n
n
t
b
t
Q
SP
Qb
Conventional EBE Sum Utilizes UTD Solution for a Single Current Source
86
• Antenna array operated at 25 GHz (K-band); = 1.2 cm• Aperture size = 1 ft. × 1 ft. = 25.4 × 25.4 • If sampling at every /4, one needs 102×102 (10,404)
sample points and needs to trace such large number ofrays.
• Conventional UTD approach becomes less efficient thancollective UTD approach.
Hypothetical Example
Material Treatment
Tapered Absorber
Antenna Array Elements
Radome Cover Flush with Platform
PEC Platform PECPlatform
Example of a LARGE Array
87
Describes fieldsproduced by the wholearray aperture at once interms of only a few UTDrays arising from specificpoints in the interior, andon the boundary of thearray aperture – highlyefficient.
Provides physical picture for describing collectivearray radiation and surface field excitation mechanisms.
Integration to most existing UTD codes for predictingthe platform interaction effects via UTD is not directbecause it requires some code modifications to allowfor a new input description.
Observation point
Present Collective UTD Ray Approach
88
• A scanning phased array on a slowly varyingconvex platform can be modeled by a parametricsurface patch, such as a bi-quadratic surface, etc.
uv
x
yz
11r
12r13r
21r31r
32r33r
22r
23r
u
v
0 1
1
1/2
1/2
uv-plane
Curved Surface Modeling
89
( )
1
( ( , ))i iu v
Kj u v
ii
A r u v C e
* Fields produced by each TW can be represented in terms of a set of UTD rays.
Found numerically by matching to the actual distribution
• A traveling wave (TW) expansion for realistic aperturedistributions obtained from FEM, FE-BI, etc.
uv
uv-space
Traveling Wave Expansion
90
• DFT (Discrete Fourier Transform)
• CLEAN (or “Extract and Subtract”)
• Prony’s Method
• Other Available Methods
TW Extraction Methods
91
Observation point
uv
Each TW current radiates a small set of collective UTD rays.
1
( )K
lUTD
lE P E
4 4 4
1 1 1
GO ed cd GO edUTD GO i ei j i
i j iE E U E U E E E
Transition fields
Collective UTD Ray Fields
92
A
ssQie
je
x
yz
ˆssn
ssQ
P ˆsss
ˆ ( )u sst Q
ˆ ( )v sst Q
ssu
ssv
1 2
1 2( )( )ss
ss ssjksGO
ss ss ssss ss
E A es s
Astigmatic Ray Field
ˆ( | )ss L ss ssA T P Q L
( ' )( 2 ) ( )ssjk Qss
ssss
j J QL ek
Geometrical Optics orLocal Floquet Wave
93
A
ieje
,s eiQ,
ˆ ( )u s eit Q
,ˆs ein
,s eiQ
ˆeisP
,s eiu
ie
( )eijksd ei
ei eiei ei ei
E A es s
Conical WaveLit Region
, ˆ( | ) dei L s ei eiA T P Q D
,/ 4 ( ),
2
2 ( )ˆ[ ] | |
s eij jk Qei s eid
eiei v v ei u
j e J Q eDk s r k E
Edge Diffracted Fields
94
A
ieje
,s eiQ
( )dei
djksds s ei
ei ei d d dei ei ei
E A es s
Shadow Region
eiQ
SP
eiQ
deis
ie
1/ 6
,,,
, ,
( )( )ˆ( | )
( ) ( )ei g s eis eijkts ds
ei S s ei eis ei g s ei
Qd QA T P Q D e
d Q Q
,/ 4 ( )
,
2
2 ( )ˆ[ ] | |
s eij jk Qei s eids
eiei v v ei u
j e J Q eDk t r k E
Edge-Excited Surface Diffracted Fields
95
ˆcin
ciQ
P ˆcisSpherical Wave
ijjkscij ij
ij
eE As
Lit Region
A
ieje
ijQ
ˆ( | ) cij L ij ijA T P Q D
( )
2
( )ˆ ˆ[ ][ ]
ijjk Qei ej ijc
ijij u u ij v v
J Q eD
k s r s r
Corner Diffracted Fields
96
( )dij
djksijcs s
ij ij d d dij ij ij
E A es s
Shadow Region
ciQ
SP
ciQdcis
A
ieje
ijQ
1/ 6
0 ( )ˆ( | )
( ) ( )ijjkt g ijs cs
ij S s ij ijij g ij
QdA T P Q D ed Q Q
( )
2
( )ˆ ˆ[ ][ ]
ijjk Qei ej ijcs
ijij u u ij v v
J Q eD
k t r t r
Corner-Excited Surface Diffracted Fields
97
2( )2 4 4
, ,21 3
( , , , )4
ijj kGO GO
ci j ei cj i ej ei eji j ij
eE E W k k k kk
2( )2 42
1
1 ( )2
eij k
ei ei eii ei
e F k Uk
2( )4 4
2
3
1 ( )2
ejj k
ej ej ejj ej
e F k Uk
2,( )4 4
2,
3 ,
1 ( )2
ci jj kd dei ei ci j ej
j ci j
eE E F kk
2 2 2 21 1 2 22 21 2
1 1 2 2 1 21 2 1 1 2 2
( , , , ) ( ) ( )x y x yx xW x y x y F x F x
y y x y x y
1 21 1 2 2
1 2
( , , , )y y T x y x yx x
Fields in Transition Regions
98
Taylor distribution
vL
uW
0n
0u
0v
O
0k
f = 24 GHza = 150.0 cm (120.0 )Wu = 37.5 cm (30.0 )Lv = 50.0 cm (40.0 )Broadside scanCurrent polarization: 45 w.r.t. axial direction
105 TWs were used !!
Aperture on a PEC Circular Cylinder
99
vL
uW
0n
0u
0v
r
O
Circumferential plane cutr = 40 (near zone)REF: 25.36 sec.TW-UTD: 3.78 sec.
Aperture on a PEC Circular Cylinder (cont.)
100
Oblique plane cut: = 40r = 40 (near zone)REF: 24.24 sec.TW-UTD: 3.76 sec.
vL
uW
0n
0u
0vr
O
Aperture on a PEC Circular Cylinder (cont.)
101
Circumferential plane cutr = 9600 (far zone)REF: 31.25 sec.TW-UTD: 4.18 sec.
0v
0u
0n
vLuW
a O
r
Aperture on a PEC Circular Cylinder (cont.)
102
0v
0u
0n
a
0k
0
0O
0 0Scan direction: 30 , 90
0.55 , 0.244s sL W
9.0 GHzf 100.0a
0.65u vd d 101 101 elements
93 TWs (0.9%) were used !!
sW
vd
ud
sL
Slot array on a PEC Circular Cylinder
103
Axial plane cut (scan plane)r = 100 (near zone)REF: 34.57 sec.TW-UTD: 5.47 sec.
0n
0u
0vrO
Slot array on a PEC Circular Cylinder (cont.)
104
Oblique plane cut: = 60r = 100 (near zone)REF: 38.40 sec.TW-UTD: 6.12 sec.
0n
0u
0vr
O
Slot array on a PEC Circular Cylinder (cont.)
105
0v
0u
0nr
a
O
Axial plane cut (scan plane)far zoneREF: 33.48 sec.TW-UTD: 4.39 sec.
Slot array on a PEC Circular Cylinder (cont.)
106
Oblique plane cut: = 45far zoneREF: 51.39 sec.TW-UTD: 6.08 sec.
0v
0u
0n
a
O
r
Slot array on a PEC Circular Cylinder (cont.)
112
Conclusions• An asymptotic UTD ray solution has been developed fordescribing, in a collective fashion, the fields radiated bylarge conformal antenna arrays on a doubly curved,smooth convex surface.
• The present solution will provide an efficient linkbetween the local array part to be analyzed numericallyand the full external platform part to be analyzed by UTD,in a hybrid method for analyzing large complexantenna phased arrays integrated into a realistic complexplatform.
• The present collective UTD ray solution shows a goodagreement with the conventional element-by-elementUTD field summation solution.
113
The Ohio State Univ. (OSU), ElectroScience Lab.(ESL), Columbus, Ohio, USA
(1)Prof. Robert J.Burkholder(2)Dr. Youngchel Kim(3)Dr. His-Tseng Chou (now Prof. at Yuan-Ze Univ. Taiwan ROC)(4)Dr. Titipong Lertwriyaprapa (now at King Mongkut university)(5)Dr. Pawuwat Janpugdee (Was member of research staff at
TEMASEK Labs. NUS, Singapore; Now at Chulalongkorn univ.)(6)Dr. Koray Tap (Now at ASELSAN, Ankara, Turkey)(7)Prof. W.D. Burnside(8)Dr. R.J. Marhefka(9)Dr. Nan Wang(10)Prof. Jin-Fa Lee
The work presented is done in conjunction with the following researchers
Univ. of Siena, Italy(1)Giorgio Carluccio(2)Prof. Matteo Albani(3)Prof. Stefano Maci
Applied EM(1)Dr. Cagatay Tokgoz(2)Dr. C. J. Reddy