Program
Background - History PEEC Introduction
EMC Europe 2014 tutorial on Partial Element Equivalent Circuit (PEEC) based analysis of EMC problems
Jonas Ekman
EISLAB
Lule˚ a University of Technology Jonas.Ekman@ltu.se
Giulio Antonini
UAq EMC Lab University of L’Aquila Giulio.Antonini@univaq.it
Sep 1, 2014
Program
Background - History PEEC Introduction
1. Program
Monday Sep 1st
14:00-14:30 Introduction and the
history of PEEC
Jonas Ekman (Lulea University of Technology)
14:35-15:05 Recent progress Giulio Antonini &
- Acceleration techniques Daniele Romano (University of L’Aquila)
15:10-15:40 Parameterized PEEC
models
Francesco Ferranti &
Giulio Antonini (Ghent University &
University of L’Aquila) COFFEE
16:10-16:40 Circuit and
electromagnetic co-simulation for active, non-linear circuit systems
Sohrab Safavi & Jonas Ekman (Lulea University of Technology)
16:45-17:15 Industry view on PEEC
for EMC simulations
Didier Cottet (ABB)
17:20-17:50 Application of PEEC
formulation to analysis of electrical networks in aircraft
Mauro Bandinelli &
Alessandro Mori (Ingegneria dei Sistemi (IDS))
Program
Background - History PEEC Introduction
2. Background - History
2.1. Edward Bennett Rosa
1861-1921, Chief Physicist at the National Bureau of Standards. 1 2
”The actual self-inductance of any closed circuit will be the sum of the self- inductances of all the parts, plus the sum of the mutual inductances of each one of the component parts on all the other parts”
2.2. Frederick Warren Grover
1876-1973, Associate Physicist at the National Bureau of Standards, Prof.
Union College. 3
1E. B. Rosa & L. Cohen, ”Formulae and tables for the calculation of mutual and self- inductance”, Bulletin of the Bureau of Standards, National Institute of Standards & Tech- nology, vol. 5, no. 1. 1907. https://archive.org/stream/formulae5113219089393rosa
2E. B. Rosa & F. W. Grover, ”Formulas and tables for the calculation of mutual and self-inductance”, Bulletin of the Bureau of Standards, National Institute of Standards &
Technology, vol. 8, no. 1. 1907. http://www.g3ynh.info/zdocs/refs/NBS/Sci169noerr.pdf
3F. W. Grover, ”Inductance calculations”, Courier Dover Publications, 2004.
Program
Background - History PEEC Introduction
2.3. Albert E. Ruehli
1937-, IBM T.J. Watson Research Center, NY, USA.
“A new comprehensive theory of inductance is offered, which is called the theory of partial inductance since the calculations are based on the inductance of loop segments.” 4
“Specifically, the use of partial capacitances (defined later) in conjunction with partial inductances leads to a three-dimensional technique, and partial-element equivalent circuits result in a complete characterization of three-dimensional in- terconnection structures.” 5
“So-called partial element equivalent circuits are proposed, which are based on an integral equation description of the geometry rather than a differential equation description.” 6
”It was incomprehensible to Grover that we wanted to compute inductances for such small objects as integrated circuits (ICs).”
4A. E. Ruehli, ”Inductance Calculations in a Complex Integrated Circuit Environment”, IBM Journal of Research and Development, vol. 16, no. 5, 1972.
5A. E. Ruehli & P. A. Brennan, ”Efficient Capacitance Calculations for Three-Dimensional Multiconductor Systems”, IEEE Trans. on Microwave Theory and Techniques, vol. 21, no.
2, 1973.
6A. E. Ruehli, ”Electrical analysis of interconnections in a solid-state circuit environment”, IEEE Int. Solid-State Circuit Conf., pp. 64-65, 1972.
Program
Background - History PEEC Introduction
2.4. Giulio Antonini
1969-, Prof. at University of L’Aquila.
• Acceleration using wavelet theory, fast multipole method, waveform re- laxation, reluctance modeling
• Material modeling including dielectrics and magnetics
• Stability and passivity enhancements and studies
• Skin-effect models
Program
Background - History PEEC Introduction
3. PEEC Introduction
3.1. Basic theory
The theoretical derivation starts from the expression of the total electric field in free space as
E~T(~r, t) = ~Ei(~r, t) −∂ ~A(~r, t)
∂t − ∇φ(~r, t). (1)
If the observation point, ~r, is on the surface of a conductor, the total electric field can be written as ~ET = J (~~r,t)σ , in which ~J (~r, t) is the current density on a conductor. To transform into the electric field integral equation (EFIE), the definitions of the electromagnetic potentials, ~A and φ are used.
A(~~ r, t) =
K
X
k=1
µ Z
vk
G(~r, ~r0) ~J (~r0, td)dvk (2)
φ(~r, t) =
K
X
k=1
1
0
Z
vk
G(~r, ~r0)q(~r0, td)dvk (3)
Combining results in the well known electric field integral equation.
Program
Background - History PEEC Introduction
The EFIE to be solved E~i(~r, t) =
J (~~ r, t)
σ + µ
Z
v0
G(~r, ~r0)∂ ~J (~r0, td)
∂t dv0 +∇
0
Z
v0
G(~r, ~r0)q(~r0, td)dv0 (4)
Interpreting the EFIE as KVL
It is possible to interpret each term in the above equation as KVL since
V = RI + sLpI + Q/C (5)
This results in interpreting:
• The first RHS term as Resistance,
• The second RHS term as Partial inductance
• The third RHS term as Coefficient of potential
Program
Background - History PEEC Introduction
The resulting equivalent circuit (QS)
C
aaaaLpmm Rmm
a b
C
bbFigure 1: One way of visualizing the interpretation of the EFIE as a partial element equivalent circuit (quasi-static approximation).
Program
Background - History PEEC Introduction
The resulting equivalent circuit (FW)
P 1
iiI
PiP 1
jjI
PjLpmm Rmm
I
CiI
Cja
I
bVmmL
Figure 2: One way of visualizing the interpretation of the EFIE as a partial element equivalent circuit (full-wave).
Program
Background - History PEEC Introduction
3.2. Simple PEEC model for two wires in free space
IPi
(a)
P1ii IPi P1jj IPj
Lpmm Rmm
ICi ICj
a I b
VmmL
P1ii IPi P1jj IPj
Lpmm Rmm
ICi ICj
a I b
VmmL
IPi
(b)
Figure 3: PEEC model for two conducting wires (a) using controlled voltage- and current- sources to account for the electromagnetic couplings (b).
Program
Background - History PEEC Introduction
3.3. Non-orthogonal formulation
1. Developed in 20037.
2. Enables the modeling of geometrically complex structures.
3. Consistent with the orthogonal formulation.
4. More complex partial elements. Compare:
Lpnonort = µ Z
a
Z
b
Z
c
Z
a0
Z
b0
Z
c0
(ˆa · ˆa0) ha0 (6) G[ ~rg(a, b, c); ~rg(a0, b0, c0)] da0db0dc0 ha da db dc with
Lport= µ
2π[ln 2l/w
1 + t/w + 0.5 + 2
9l/w(1 + t/w)] (7)
7A. E. Ruehli, G. Antonini, J. Esch, J. Ekman, A. Mayo, A. Orlandi, ”Nonorthogonal PEEC Formulation for Time- and Frequency-Domain EM and Circuit Modeling,” IEEE Transactions on EMC, 45, pp. 167-176, 2, 2003.
Program
Background - History PEEC Introduction
a
b c
(a)
Lp22
Lp33
Lp44 P33
1 Ip
P44
1 Ip
P11
1 Ip
P22
1 Ip
-+
Lp11
-+
-+ -+
VL
VL
VL
VL
1
4
3 2
Ip1
Ip4
Ip3 Ip2
3
2
1
4 I3
I1 I2
I4
f2
f1
f4
f3
(b)
Figure 4: Nonorthogonal metal patch (a) and PEEC model (b).
Program
Background - History PEEC Introduction
a
b c
(a)
Lp22
Lp33
Lp44
C33
C44 C11
C22
Lp11
C14
C23
C34
C12 C13
C24
(b)
Figure 5: Nonorthogonal metal patch (a) and QS-PEEC model (b).
Program
Background - History PEEC Introduction
a
b c
(a)
Lp11 1 4 {Lp11value} C11 1 0 {C11value} Lp22 1 2 {Lp22value} C22 2 0 {C22value} Lp33 2 3 {Lp33value} C33 3 0 {C33value} Lp44 4 3 {Lp44value} C44 4 0 {C44value} K13 Lp11 Lp33q {Lp13
value}2
{Lp11value}{Lp33value} C12 1 2 {C12value} K24 Lp22 Lp44q
{Lp24value}2
{Lp22value}{Lp44value} C13 1 3 {C13value} C14 1 4 {C14value} C23 2 3 {C23value} C24 2 4 {C24value} C34 3 4 {C34value} (b)
Figure 6: Nonorthogonal metal patch (a) and SPICE .cir-file in (b).
Program
Background - History PEEC Introduction
3.4. Steps performed in a PEEC solver
Graphical User
Interface (GUI) Meshing Partial Element Post-processor
Calculation Matrix
Formulation Matrix Solution
3.4.1. GUI
Creating, editing, archiving simulation projects.
3.4.2. Meshing
Orthogonal vs. nonorthogonal. Electrically large or small. Model simplification.
3.4.3. Partial element calculation
Orthogonal vs. nonorthogonal. Closed formulas vs. numerical integration.
Program
Background - History PEEC Introduction Graphical User
Interface (GUI) Meshing Partial Element Post-processor
Calculation Matrix
Formulation Matrix Solution
3.4.4. Matrix formulation
Nodal analysis vs. Modified nodal analysis.
3.4.5. Matrix solution
Direct, iterative, and/or accelerated.
3.4.6. Post-processing
Visualize current and potential distribution, port parameter calculations.
Program
Background - History PEEC Introduction
3.5. Summary of approach
• Pros:
+ Same time- and frequency- domain model.
+ Direct inclusion of other SPICE circuit elements.
+ Solvable with SPICE-like solvers.
+ Engineering-type of approach.
• Cons:
- Large number of partial elements (speed-up, approximations).
- Complex materials and geometries?
• The PEEC method is equivalent to a Method of Moments solution (if the Moment Method is formulated in a special way).