Motivation: the current power grid is . . .

Kron Reduction of Graphs

with Applications to Electrical Networks

Florian Dörfler and Francesco Bullo

Center for Control,

Dynamical Systems & Computation
University of California at Santa Barbara
http://motion.me.ucsb.edu

“. . . the greatest engineering achievement of the 20th century.”

Center for Nonlinear Studies [National Academy of Engineering ’10]
Los Alamos National Labs, New Mexico, June 8, 2011
Article available online at: http://arxiv.org/abs/1102.2950 “. . . the largest and most complex machine engineered by humankind.”
[P. Kundur ’94, V. Vittal ’03, . . . ]
Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 1 / 18 Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 2 / 18

Motivation: the envisioned power grid Motivation: the envisioned power grid

Energy is one of the top three national priorities Energy is one of the top three national priorities

Expected developments in “smart grid”: Expected developments in “smart grid”:

1 large number of distributed power sources 1 large number of distributed power sources
2 increasing adoption of renewables 2 increasing adoption of renewables
3 sophisticated cyber-coordination layer 3 sophisticated cyber-coordination layer

/ challenges: increasingly complex networks & stochastic disturbances / challenges: increasingly complex networks & stochastic disturbances
, opportunity: some smart grid keywords: , opportunity: some smart grid keywords:
control/sensing/optimization distributed/coordinated/decentralized control/sensing/optimization distributed/coordinated/decentralized

Today: “reducing the complexity by means of circuit and graph theory” Today: “reducing the complexity by means of circuit and graph theory”
Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 3 / 18 Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 3 / 18
 Kron reduction of a resistive circuit Kron reduction of a resistive circuit

1 Nodal analysis by Kirchho↵’s and Ohm’s laws:

2 Partition circuit equations via boundary nodes & interior nodes :
I =Y ·V k   
Yik Yjk Iboundary Yboundary Ybound-int Vboundary
= T
I 2 Cn nodal current injections Iinterior Ybound-int Yinterior Vinterior
V 2 Cn nodal voltages/potentials Ik Yk,shunt
1 1 -1
Y 2 Cn⇥n nodal conductance matrix 8 30 30

1 1 1 1
1 1 1
2 3 1 1
.. .. .. . .. 1 1
. ..
6 . . . 7 1
30 30 30 30

1 1
6 P
n 7 1 1 1 1
Y = YT =6
6 Yi1 . . . Yik + Yk,shunt . . . Yin 7
7 1 1
1
1 1
4 k=1,k6=i 5 8 30 30

.. . .. .. ..
. .. . . .
1 -1 -1

= { weighted Laplacian matrix } + diag Yk,shunt = “loopy Laplacian” Boundary nodes ⇤

⌅ arise as natural terminals in applications.

Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 4 / 18 Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 5 / 18

Kron reduction of a resistive circuit Kron reduction of graphs

3 Kron reduction: eliminate interior nodes • via Schur complement:

Consider either of the following three equivalent setups:
Yred = Y /Yinterior = Yboundary Ybound-int · Yinterior 1 T
· Ybound-int 1 a connected electrical network with conductance matrix Y , terminals
⇤
⌅ , interior nodes• , & possibly shunt conductances
1 1 -1 0.97 -0.66
8 30 30

8
0.39 0.08 1.92 2 a symmetric and irreducible loopy Laplacian matrix Y with partition
1 1 1 1
⌅ ,• ), & possibly diagonally dominance
8

1
1
1 1 (⇤
1 1 1
0.15 0.05 0.21
1 0.11 an undirected, connected, & weighted graph with boundary nodes ⇤
⌅,
30 30 30 30

1 1 3
1 1 1 1
8
1 30
1
1
30
1 8
1
0.98 8
0.06 8
1.73 interior nodes • , & possibly self-loops
1 -1 -1 0.72 -1

G
8 30 30

conductance 1 1
8

original circuit “equivalent” reduced circuit 1 1 1 1

8 27 27 8

1
1 1 1 matrix Y 1
1 1 1 1 1
1 1
I =Y ·V Ired = Yred · Vboundary 1 1 1 1
30 30 30 30

1
27 27 27 27

1 1
1 1 1 1 1 1 1 1 1
1 loopy Laplacian 1
1 1 1 1 1 1
8 30 30

matrix Y 8 27

1 27 8

G. Kron, “Tensor Analysis of Networks,” Wiley, 1939.

Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 6 / 18 Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 7 / 18
Kron reduction of graphs Kron reduction of graphs

Kron reduction via Schur complement: Yred = Y /Yinterior

Kron reduction via Schur complement: Yred = Y /Yinterior
G
8 30 30

1 1 1 1 conductance 1 1
8 27 27 8

1
1 1 1 matrix Y 1
1 1 1 1 1
1 1
1 1 1 1 Relation of spectrum and algebraic properties of Q and Qred ?
30 30 30 30

1
27 27 27 27

1 1
1 1 1 1 1 1 1 1 1
1 loopy Laplacian 1
1 1 1 1 1 1
8 30 30

matrix Y 8 27

1 27 8

1 How about the graph topologies and the e↵ective resistances?

Kron reduction Kron reduction Kron reduction

What is the e↵ect of a perturbation in the original graph on the
reduced graph, its spectrum, and its e↵ective resistance?
0.39 0.08 1.92 transfer conductance Gred
8

matrix Yred
8

0.08
8
Finally, why is this graph reduction process of practical importance
0.15 0.11 0.05 0.21
0.39
0.15 0.21
1.92 and in which application areas?
0.11 0.05
0.98 0.06 1.73 Kron-reduced 0.98
0.06
1.73
loopy Laplacian Yred
8

Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 8 / 18 Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 9 / 18

Kron reduction of graphs: applications Kron reduction of graphs: applications

Purpose: construct low-dimensional equivalent circuits / graphs / models
Electrical impedance tomography Smart grid monitoring

Simplest non-trivial case: star- transformation Yred Ycut

[A. E. Kennelly 1899, A. Rosen 1924]
8 8

1.0
Kron reduction to reconstruct spatial conductivity through cut-set variables
1/3 1/3
30
[E. Curtis and J. Morrow ’94 & ’00] [I. Dobson ’11]
1.0 1.0
8

8
8
1/3 8

Representation of integration chips

32 32 IEEE TRANSACTIONS
IEEE TRANSACTIONS ON COMPUTER-AIDED
ON COMPUTER-AIDED DESIGN OFCIRCUITS
DESIGN OF INTEGRATED INTEGRATED
ANDCIRCUITS
SYSTEMS,AND
Reduced power network modeling
VOL.SYSTEMS,
29, NO. 1,VOL. 29, NO.
JANUARY 1, JANUARY 2010
2010
!"#$%&'''%()(*%(+,-.,*%/012-3*%)0-4%5677*%899: !"#$%&'
8
8

8 8 8 8 8

8 15
37 9 10 10
37 9
10
10 37 38 02 38
29 10 30 25 30 25
03 26
30 25 26 28 26

δi / rad
2 38 2 29 04 2 29

18 27 05
1 24 9 5 1 3 27
8 8

1 3 27

6
6
28

1
28
17

35

35
1

39
39
18 18 22
22
F 16 0
4 4

Engineering applications:
9 9
17 21
6 17 21
3

smart grid monitoring, circuit theory, model reduction for power and water 1 15 35 -5 8 8
5
8 8
21 0 2 45 6 24
8 10 24
7
39 22 7
12
4 14 12 14 15 23 14 15 23
6 6

36
36
31

31
5 15 16

7
16
6 12 19 23 13

2
13

2
06 11

networks, power electronics, large-scale integration chips, electrical impedance tomography, data-mining, . . .
11

13 20 36 07
10
7 31 11 10 10
19 08 19
7
8 8

10 34 33

δi / rad
8 2
32
20
33
09 32 20
33
32 5 4 5 3
3
34 34 4
9 4
3
5
0

Mathematics applications:
5

8 8 8 8 8

Fig. 9. The New England test system [10], [11]. The system includes
sparse matrix algorithms, finite-element methods, sparse multi-grid solvers, Markov 10 synchronous generators and 39 buses. Most of the buses have constant
active and reactive power loads. Coupled swing dynamics of 10 generators
are studied in the case that a line-to-ground fault occurs at point F near bus
-5
0 2 4
TIME / s
6 8 10

16.

chain reduction, stochastic complementation, applied linear algebra & matrix analysis, Dirichlet-to-Neumann map, . . . Fig. 5. conductance
Two reducedmatrices
conductance
of thematrices of the same network with 59
Fig. 10. Coupled swing of phase angle i in New England test system.
Fig. 5. Two reduced same network with 59
for stability analysis and control
The fault duration is 20 cycles of a 60-Hz sine wave. The result is obtained
by numerical integration of eqs. (11).

Physics applications: knot theory, Yang-Baxter equations and applications, high-energy physics, statistical
for sparse computation terminals. The terminals.
matrix on The
onbut
on the right, all the five
right has fewerright
right,
has fewer
nonzeros,
matrix
the left on the
results
all butnodes
internal
and nonzeros,
when lefteliminating
five internal
results when
nodes are
are eliminated.
only five andmoreonlyrows,fivethan
all eliminating

more therows,
all internal nodes;
internal nodes;
Theeliminated.
matrix on Thethe matrix on the
than the first matrix,
first matrix,
test system can be represented by

Hi
˙i =

˙i =
i,

Di + Pmi Gii Ei2

10
Ei E j · (11)
are provided to discuss whether the instability in Fig. 10
and hence the and hence the network
corresponding network721isversus
smaller: 721resistors.
versus 1711 resistors. i occurs in the corresponding real power system. First, the
corresponding is smaller: 1711 fs

mechanics, vortices in fluids, entanglement of polymers & DNA, . . . [F. Dörfler & F. Bullo ’11, J.H.H. Perk & H. Au-Yang ’06]
By selecting
By selecting specific
[J. Rommes and W. H. A. Schilders ’09] internalspecific
the reduced
the reduced network,
nodes internal
fill-in network,
(five in nodes
fill-in during
during elimination
(five toin be
this case)
elimination
of the
thispreserved
case) to inbe preserved in
of the nodes
other internal other internal
is nodes is
[F. Dörfler and F. Bullo ’09]
· {Gij cos( i j)

where i = 2, . . . , 10. i is the rotor angle of generator i with

j=1,j=i
+ Bij sin( i j )},
classical model with constant voltage behind impedance is
used for first swing criterion of transient stability [1]. This is
because second and multi swings may be affected by voltage
fluctuations, damping effects, controllers such as AVR, PSS,
limitedThe
limited considerably. considerably.
difference The difference
becomes even becomes
bigger as even bigger as
the number of the number of respect to bus 1, and i the rotor speed deviation of generator
i relative to system angular frequency (2 fs = 2 60 Hz). and governor. Second, the fault durations, which we fixed at
terminals increases.
terminals increases. 1 is constant for the above assumption. The parameters
20 cycles, are normally less than 10 cycles. Last, the load
fs , Hi , Pmi , Di , Ei , Gii , Gij , and Bij are in per unit condition used above is different from the original one in
system except for Hi and Di in second, and for fs in Helz. [11]. We cannot hence argue that global instability occurs in
Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 10 / 18 Florian Dörfler (UCSB)
the reduced the reducedand
network), network), andcan
since this since this can bebefore
be computed
Kron Reduction
computed before
The mechanical input power Pmi to generator i and the
magnitude Ei of internal voltage in generator i are assumed
Center for Nonlinear Studies
the real system. Analysis, however, does show a possibility
of global instability in real power systems. 11 / 18
to be constant for transient stability studies [1], [2]. Hi is IV. T OWARDS A C ONTROL FOR G LOBAL S WING
doing the actualdoing elimination
the actual elimination we always
we always include include
a check on a check on the inertia constant of generator i, Di its damping coefficient,
and they are constant. Gii is the internal conductance, and
I NSTABILITY

this. If ittoiseliminate
this. If it is decided
conductanceconductance
matrix Gk can
Cholesky factorization
Cholesky factorization
decided to
alleliminate all internal
internal nodes, nodes, the reduced
the reduced Gij + jBij the transfer impedance between generators i
and j; They are the parameters which change with network
Global instability is related to the undesirable phenomenon
that should be avoided by control. We introduce a key
matrix Gk can beefficiently
be computed computedusingefficiently
the using the topology changes. Note that electrical loads in the test system
are modeled as passive impedance [11].
mechanism for the control problem and discuss control
strategies for preventing or avoiding the instability.
G22 = LLT G22 = LLT B. Numerical Experiment A. Internal Resonance as Another Mechanism
Coupled swing dynamics of 10 generators in the Inspired by [12], we here describe the global instability
test system are simulated. Ei and the initial condition with dynamical systems theory close to internal resonance
k = G11 − QQ
T
Gk = G11 − G
QQ T ( i (0), i (0) = 0) for generator i are fixed through power
flow calculation. Hi is fixed at the original values in [11].
[23], [24]. Consider collective dynamics in the system (5).
For the system (5) with small parameters pm and b, the set
1
 Kron reduction of graphs: properties Kron reduction of graphs: properties

Kron reduction of a graph with 8

2 Iterative 1-dim Kron reduction: k+1
Yred k /
= Yred •
⌅ , interior • , non-neg self-loops
boundary ⇤
30

) topological evolution of the corresponding graph

27 28 25

loopy Laplacian matrix Y 28

1)
8 8

2)
8 8

Schur complement: Yred = Y /Yinterior

30 30 30 30

30

30

X
27 25 27 25

27 25 27 25

28 28

28

8
30

X 28

30

8
8
30

30

8
8
30

30
8

8
30

30

3) 7)
8 8

Properties of Kron reduction:

8

30 30

X
30

X 27
25
25

...
1 Well-posedness: set of loopy Laplacian matrices is closed 8
30

30

8
8
30

30

8
8

!"#$%&'''%()(*%(+,-.,*%/012-3*%)0-4%5677*%899: !"#$%&'
8
8

) Equivalence: the following diagram commutes:

8 15
37 9 10 10
37 9
10
10 37 38 02 38
29 10 30 25 30 25
03 26
30 25 26 28 26
δi / rad

2 38 2 29 04 2 29

18 27 05
5
1 24 9 1 3 27 1 3 27

Kron reduction of

6
6
28
30

1
28
17

35
35
1

39
39

18 18 22
22
F 16 0
4 4
9 9
17 21
6 17 21
3
1 15 35 -5 8 8
5
21 0 2 45 6 24
8 10 24

n | |
7
39
0
7
22

Y = Yred
12
4 14

Yred = Yred
12 14 15 23 14 15 23 8

6 6

36
36
31

31
5 15 16

7
16
6 12 19 23 13

2
13
2

11
06 11

13 20 36 07
10
7 31 11 10 10
19 08 19
10 34 33 7
1
δi / rad

32
09 32

2 n | |+1
20
2 20

Yred
8 33 33

9
32

3
5 4 5

0
3
34

5
4
3
34

5
4

Yred ... Yred

8

Fig. 9. The New England test system [10], [11]. The system includes
10 synchronous generators and 39 buses. Most of the buses have constant
active and reactive power loads. Coupled swing dynamics of 10 generators
-5
0 2 4 6 8 10
iterative Kron reduction
are studied in the case that a line-to-ground fault occurs at point F near bus TIME / s
16.
Florian Dörfler (UCSB) Fig. 10. Kronswing
Coupled Reduction Center for Nonlinear Studies
of phase angle i in New England test system. 12 / 18 Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 13 / 18
The fault duration is 20 cycles of a 60-Hz sine wave. The result is obtained
by numerical integration of eqs. (11).
test system can be represented by
˙i = i,
10
Hi are provided to discuss whether the instability in Fig. 10
˙i = Di i + Pmi Gii Ei2 Ei E j · (11) occurs in the corresponding real power system. First, the
fs

Kron reduction of graphs: properties

· {Gij cos( i

where i = 2, . . . , 10. i is the rotor angle of generator i with

j ) + Bij sin(
j=1,j=i
i j )},
classical model with constant voltage behind impedance is
used for first swing criterion of transient stability [1]. This is
because second and multi swings may be affected by voltage
fluctuations, damping effects, controllers such as AVR, PSS,
Kron reduction of graphs: properties
respect to bus 1, and i the rotor speed deviation of generator
i relative to system angular frequency (2 fs = 2 60 Hz). and governor. Second, the fault durations, which we fixed at
20 cycles, are normally less than 10 cycles. Last, the load
1 is constant for the above assumption. The parameters
fs , Hi , Pmi , Di , Ei , Gii , Gij , and Bij are in per unit condition used above is different from the original one in
[11]. We cannot hence argue that global instability occurs in

by edge to grounded node ⌥

⌃
system except for Hi and Di in second, and for fs in Helz. 1) 2)
8 8

8 8

3 Augmentation: replace self-loops

The mechanical input power Pmi to generator i and the
magnitude Ei of internal voltage in generator i are assumed
the real system. Analysis, however, does show a possibility
of global instability in real power systems. 30 30 30 30

to be constant for transient stability studies [1], [2]. Hi is IV. T OWARDS A C ONTROL FOR G LOBAL S WING
the inertia constant of generator i, Di its damping coefficient,
X
27 25 27 25

I NSTABILITY
27 25 27 25

and they are constant. Gii is the internal conductance, and 28 28

28

Global instability is related to the undesirable phenomenon

Gij + jBij the transfer impedance between generators i
G
and j; They are the parameters which change with network
8

topology changes. Note that electrical loads in the test system

27 27 8

that should be avoided by control. We introduce a key

mechanism for the control problem and discuss control
strategies for preventing or avoiding the instability.
G 8 27 27 8

8
30

X 28

30

8
8
30

30

8
8
30

30
8

8
30

30

are modeled as passive impedance [11].

B. Numerical Experiment A. Internal Resonance as Another Mechanism
Coupled swing dynamics of 10 generators in the Inspired by [12], we here describe the global instability 3) 7)
8 8

28
8

27 27 27 27 27 27 27 27

test system are simulated. Ei and the initial condition with dynamical systems theory close to internal resonance
( i (0), i (0) = 0) for generator i are fixed through power [23], [24]. Consider collective dynamics in the system (5).
For the system (5) with small parameters pm and b, the set
30 30

X 30

X ...
flow calculation. Hi is fixed at the original values in [11].
Pmi and constant power loads are assumed to be 50% at their {( , ) 2 S 1 R | = 0} of states in the phase plane is 27
25
25

ratings [22]. The damping Di is 0.005 s for all generators.

8 27 27 8

called resonant surface [23], and its neighborhood resonant 8 27 27 8

Gii , Gij , and Bij are also based on the original line data band. The phase plane is decomposed into the two parts:
in [11] and the power flow calculation. It is assumed that resonant band and high-energy zone outside of it. Here the 8
30

30

8
8
30

30

8
8

the test system is in a steady operating condition at t = 0 s, initial conditions of local and mode disturbances in Sec. II
that a line-to-ground fault occurs at point F near bus 16 at indeed exist inside the resonant band. The collective motion
t = 1 s 20/(60 Hz), and that line 16–17 trips at t = 1 s. The before the onset of coherent growing is trapped near the
fault duration is 20 cycles of a 60-Hz sine wave. The fault resonant band. On the other hand, after the coherent growing,
) Equivalence: the following diagram commutes:
is simulated by adding a small impedance (10 7 j) between
bus 16 and ground. Fig. 10 shows coupled swings of rotor
it escapes from the resonant band as shown in Figs. 3(b),
4(b), 5, and 8(b) and (c). The trapped motion is almost
integrable and is regarded as a captured state in resonance
angle i in the test system. The figure indicates that all rotor
angles start to grow coherently at about 8 s. The coherent
growing is global instability.
[23]. At a moment, the integrable motion may be interrupted
by small kicks that happen during the resonant band. That is,
4 Topological properties:
C. Remarks augment
the so-called release from resonance [23] happens, and the

It was confirmed that the system (11) in the New Eng-

land test system shows global instability. A few comments
Ycollective motion crosses the homoclinic orbit in Figs. 3(b),
4(b), 5, and 8(b) and (c), and hence it goes away from
the resonant band. It is therefore said that global instability Y interior network connected ) reduced network complete
(')$ at least one node in interior network features a self-loop
Kron reduction
Authorized licensed use limited to: Univ of Calif Santa Barbara. Downloaded on June 10, 2009 at 14:48 from IEEE Xplore. Restrictions apply.
Kron reduction ) all nodes in reduced network feature self-loops
of 27

of 27

5 Algebraic properties: self-loops in interior network

Yred augment Yred decrease mutual coupling in reduced network
increase self-loops in reduced network
Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 14 / 18 Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 15 / 18
 We start by considering “regular” electrical networks with scalar valued currents and p
electrical network consists of an interconnection of purely resistive elements. Such
Kron reduction of graphs: properties generally described by graphs whose
Kron nodes
reduction of represent connection points between resis
graphs: properties
correspond to the resistors themselves. The effective resistance between two nodes in
6Spectral properties:
is defined as the potential drop between
E↵ective the Rtwo
resistance ij : nodes, when a current source with 7

interlacing property: i (Y )  i (Yred )  i+n |⌅| (Y )

Ampere is connected across the two nodes (cf. Figure 1). For a general network, the com
) algebraic connectivity 2 is non-decreasing

e↵ect of self-loops on loop-less Laplacian matrices: u

2 (Lred ) + max{ } 2 (L) + min{ }

) self-loops weaken the algebraic connectivity 2

PSfrag replacements
Example: all mutual edges have unit weight

Ampere o
8 27 27 8

Kron reduction
27 27 27 27

8 27 27 8
8
8

1 Amp

without self-loops: 2 (L) = 0.39  0.69 = 2 (Lred )

with unit self-loops: 2 (L) = 0.39 0.29 = 2 (Lred )
Florian Dörfler (UCSB)
Figure 1: A resistive electrical network for which a 1 Ampere current was is injected at
Kron Reduction Center for Nonlinear Studies 16 / 18 Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 17 / 18

at the reference node o. The resulting potential difference Vu Vo is the effective resis
o.
Kron reduction of graphs: properties Conclusions
7 E↵ective resistance Rij : resistances relies on the usual Kirchoff’s and Ohm’s laws.
G
8 30 30

1 1 1 1 conductance 1 1
8 27 27 8

1
1 1 1 matrix Y 1
1 1 1

⌅To establish the connection between graph effective resistances defined in the prev 1 1 1 1
Equivalence and invariance of Rij among ⇤ nodes: 1 1 1 1
30 30 30 30

1
27 27 27 27

1 1
1 1 1 1 1 1 1 1 1
loopy Laplacian
trical networks we need to consider an abstract generalized matrix Y 8

electrical network in which

1 30
1
1
30
1 8
1 8
1 27

1
1
27
1 8

augment
Y and resistors
Y are k k matrices. For such networks,KronKirchoff’s
Kron reduction reduction
current law can be defi
Kron reduction
except that currents are added as matrices. Kirchoff’s voltage law can also be defined in
Kron reduction Rij potentials drops
Kron reduction
transfer conductance
across edges are added as matrices.matrix Kirchoff’s
Y
G 0.39
voltage laws show the
8
0.08 8
1.92
red
red
8

0.08
8

0.39 1.92
of 27

i, j of
valued node potential function. Ohm’s law takes loopy
27

the following matrix form

Kron-reduced
0.15
0.98
0.11 0.05 0.21
0.98
0.15
0.11
0.21
0.05
1.73
0.06 1.73 0.06
Laplacian Y
8

red

Yred Yred
augment Ve = Re ie ,
Kron reduction is important in various applications

no self-loops: Rij among ⇤

⌅ uniform ,
where
1 ⌅ ie is a generalized k k matrix
Analysis ofcurrent flowing
Kron reduction through
via algebraic graphthe edge e of the electric
theory
R = 2 |Yred (i, j)|
generalized resistance of that edge,
ij
and Ve is a &generalized k k matrix potential dr
self-loops: Rij among ⇤
⌅ &⌃
⌥ uniform , R1 = ⌅2 |Yred (i, j)| + max{ } Open problem: directed complex-weighted graphs
Generalized resistances are always symmetric positive definite matrices.
ij

Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 17 / 18 Florian Dörfler (UCSB) Kron Reduction Center for Nonlinear Studies 18 / 18

The generalized electrical networks so defined share many of the properties of “re
works. In particular, Kirchoff’s and Ohm’s laws uniquely define all edge currents an