Available online at www.sciencedirect.

com

ScienceDirect
IFAC PapersOnLine 54-14 (2021) 221–226
Stability
Stability Analysis
Analysis of of DC
DC Microgrids
Microgrids With
With
Stability Analysis
Switched of DC Microgrids
Events With
Stability Switched
Analysis of DCEvents
Microgrids With
Switched Events
VladimirSwitched Events
Toro ∗ Eduardo Mojica-Nava ∗∗
Vladimir Toro ∗∗ Eduardo Mojica-Nava ∗∗
VladimirNaly Toro
∗∗
Eduardo Mojica-Nava
Rakoto-Ravalontsalama
∗
∗∗∗
∗∗
Vladimir
VladimirNaly Toro
Toro ∗ Eduardo Mojica-Nava
Eduardo
Rakoto-Ravalontsalama Mojica-Nava ∗∗∗
∗∗∗ ∗∗
VladimirNaly Rakoto-Ravalontsalama
Toro ∗
Eduardo Mojica-Nava ∗∗∗ ∗∗
∗
Naly Rakoto-Ravalontsalama
Naly Rakoto-Ravalontsalama ∗∗∗ ∗∗∗
∗ Universidad Nacional de Colombia,
Naly Rakoto-Ravalontsalama Bogota, 111321, (e-mail:
∗
∗
Universidad
Universidad Nacional Nacional de
de Colombia,
Colombia, Bogota,
[email protected]). Bogota, 111321,111321, (e-mail:
(e-mail:
∗ Universidad Nacional de Colombia,
Colombia, Bogota,Bogota, 111321,111321, (e-mail:
(e-mail:
∗∗ Universidad Nacional
∗ Universidad Nacional
de
[email protected]).
[email protected]).
de Colombia,
Colombia, Bogota,Bogota, 111321, 111321, (e-mail:
(e-mail:
∗∗Universidad
∗∗
Universidad Nacional
Nacional de
[email protected]).
[email protected]).
de
∗∗ Universidad Nacional
∗∗ Universidad Nacional
de Colombia, Bogota, 111321, (e-mail:
Colombia,
[email protected])
[email protected]).
Bogota, 111321, (e-mail:
Universidad Nacional de Colombia,
deNantes,
Colombia,
[email protected]) Bogota,
Bogota, 111321,
111321, (e-mail:
(e-mail:
∗∗ ∗∗∗
Universidad IMT [email protected])
Atlantique, de Colombia, France,
Bogota, 44307, (e-mail:
111321, (e-mail:
∗∗∗
∗∗∗ IMT [email protected])
[email protected])
Atlantique, Nantes, France, 44307, (e-mail:
∗∗∗ IMT Atlantique,
∗∗∗ IMT Atlantique,
Nantes, France,
[email protected])
[email protected]) 44307, (e-mail:
IMT Atlantique, Nantes, France,
Nantes, France, 44307,
[email protected]) 44307, (e-mail:
(e-mail:
∗∗∗ [email protected])
IMT Atlantique, Nantes, France, 44307, (e-mail:
[email protected])
[email protected])
[email protected])
AbstractThis paper analyzes the stability of a direct current microgrid with a decentralized
AbstractThis
AbstractThis
switched control paper
paper using analyzes
analyzes the stability
the stability of
differential-algebraic of aa direct
direct and
equations current
current microgrid
microgrid
Lyapunov with aaThe
with
functions. decentralized
decentralized
decentral-
AbstractThis
AbstractThis
switched control paper
paper using analyzes
analyzes the
the stability
stability
differential-algebraic of
of aa direct
direct
equations current
current
and microgrid
microgrid
Lyapunov with
with
functions. aaThedecentralized
decentralized
decentral-
switched
ized
AbstractThis control
controllers paper using
regulate differential-algebraic
analyzes the voltage,
the stability achieve equations
of the
athe and
power-sharing
direct current Lyapunov functions.
condition,
microgrid with and The
aThe decentral-
guaranty
decentralized the
switched
switched
ized control
control
controllers using
using
regulate differential-algebraic
differential-algebraic
the voltage, achieve equations
equations and
and
power-sharing Lyapunov
Lyapunov functions.
functions.
condition, and The decentral-
decentral-
guaranty the
ized controllers
non-Zeno
switched condition.
control regulate
using They thealsovoltage,
fulfill
differential-algebraic achieve
the the
droop-control
equations power-sharing
and condition
Lyapunov condition,
for optimal
functions. andpower
Theguaranty
dispatch.
decentral- the
ized
ized controllers
controllers
non-Zeno condition. regulate
regulateThey the
thealsovoltage,
voltage,
fulfill achieve
achieve
the the
the
droop-control power-sharing
power-sharing
condition condition,
condition,
for optimal and
andpower guaranty
guaranty
dispatch. the
the
non-Zeno
The event
ized condition.
of either
controllers regulateThey
connecting thealso fulfill
orfulfill
voltage, the
disconnecting droop-control
achieve a converter
the condition
power-sharing is analyzed for optimal
as a switched
condition, andpower dispatch.
event.
guaranty Itthe
is
non-Zeno
non-Zeno condition. They also the droop-control condition for
for optimal power dispatch.
The
The
shown
non-Zeno
The thatcondition.
event
event
event
ofthe
of either
either
condition.
of system
either
They
connecting
connecting
They
also orfulfill
or
is asymptotically
connecting also orfulfill
the
disconnecting
disconnecting
the
droop-control
stable
disconnecting
aaunder
converter
converter
droop-control
a this
converter
condition
is analyzed
is analyzed
class
condition
is
optimal
as aa switched
as
of switching
analyzedfor optimal
as a
power
switched
events
power
switched
dispatch.
event.
event.
after It is
It
either
dispatch.
event. It
is
is
The
shown
shown event
that of
thatofor either
the
the system
system connecting
is or
asymptotically
is asymptotically disconnecting
stable
stable a converter
under this is analyzed
class of as
switching a switched
events event.
after It
either is
connecting
The
shown
shown event
connectingthat
that or the
the
or
disconnecting
either
system
system connecting
disconnecting is a distributed
or disconnecting
is asymptotically
asymptotically
a distributed
distributed stable
stable aunder
generator.
under
under
generator.
this class
converter
this
this classclass
of switching
is analyzed
of
of switching
switching
events after
as a switched
events
events after event.
after
either
It is
either
either
connecting disconnecting a generator. BY-NC-NDevents
CCswitching
the of licenseafter either
Copyright
shown
connecting
connectingthat©or or2021
the The Authors.
system
disconnecting
disconnecting aThis
is asymptotically
a is an open
distributed
distributed stable
generator.
generator. under
articlethis
accessunder class
(https://creativecommons.org
Keywords:
connecting Microgrid,
or disconnecting /licenses/by-nc-nd/4.0/)
differential-algebraic
a distributed generator.equation, decentralized control, switched system.
Keywords: Microgrid,
Keywords: Microgrid, differential-algebraic
differential-algebraic equation, equation, decentralized
decentralized control, control, switched
switched system. system.
Keywords:
Keywords: Microgrid,
Microgrid, differential-algebraic
differential-algebraic equation, equation, decentralized
decentralized control, control, switched
switched system. system.
Keywords: Microgrid, differential-algebraic equation,
1. INTRODUCTION decentralized
der switching control,Inswitched
are shown. system.
Domı́nguez-Garcı́a and Trenn
1. INTRODUCTION
1. INTRODUCTION der
der switching
switching
(2010), are shown.
are shown.
an analytical In Domı́nguez-Garcı́a
In Domı́nguez-Garcı́a
framework for jumps and impulse and Trenn
and Trenn
de-
1. INTRODUCTION
1. INTRODUCTION der
der switching
switching
(2010), an are
are
analytical shown.
shown. In
In
framework Domı́nguez-Garcı́a
Domı́nguez-Garcı́a
for jumps and and
and
impulse Trenn
Trenn
de-
1. INTRODUCTION (2010),
tection
der an
in
switching analytical
switched
are shown. framework
systems with
Inwith for jumps
applications
Domı́nguez-Garcı́a and toimpulse
power
and de-
sys-
Trenn
Distributed energy resources (DERs) have increasingly (2010), (2010),
tection
tection
an
an
in
in
analytical
analytical
switched
switched
tems is presented.
framework
framework
systems
systems
Passivity with
for
for jumps
jumps
applications
applications
conditions
andtoimpulse
and
to analyze to
impulse
power
power
de-
de-
sys-
sys-
switched
(2010), answitched
analytical framework for jumps andto
toimpulse de-
Distributed
Distributed
become attractive
Distributed
energyalternatives
energy
energy
resources for
resources
resources
(DERs)
(DERs)
energyhave
(DERs)
have
have espe- tection
increasingly
increasingly
generation,
increasingly
tection
tems is
tems
systems
in
in switched
is presented.
presented.
are presented
systems
systems
Passivity
Passivity
by
with
with applications
applications
conditions
conditions
Ñañez et al.
to analyze
to analyze
(2017)
power
power sys-
sys-
switched
switched
including an
Distributed
become
become energy
attractive
attractive resources
alternatives
alternatives (DERs)
for
for energy
energy have increasingly
generation,
generation, espe-
espe- tection
tems
tems
systemsis
is in switched
presented.
presented.
are presented systems
Passivity
Passivity
by with applications
conditions
conditions
Ñañez et al. to
to to
analyze
analyze
(2017) power sys-
switched
switched
including an
cially renewable
Distributed
become attractive sources.
energy resources
alternativesA direct current
(DERs)
for energy (DC)
have microgrid
increasingly
generation, espe- systems
example are
of presented
an electrical by system.
Ñañez etAal. DAE (2017) modelincluding
with an
the
become
cially
cially attractive
renewable
renewable alternatives
sources.
sources. A
A for
direct
direct energy
current
current generation,
(DC)
(DC) espe-
microgrid
microgrid tems
systems
systems
exampleis presented.
are
are
of presented
presented
an Passivity
electrical by
by conditions
Ñañez
Ñañez
system. et
et A al.
al.
DAE to analyze
(2017)
(2017) model switched
including
including
with an
an
the
(MG) renewable
become interconnects
attractivesources.DERs
alternatives thatfor operate
energy in DCmicrogrid
generation, through
espe- example example
Lyapunov ofapproach
an electrical system.
tobycontrol Aal.DAE model with the
cially
cially
(MG)
(MG)
a commonrenewable
interconnects
interconnects sources.
bus, each DERs
DERs
of them
A
A direct
direct
that
that
might
current
current
operate
operate have in
(DC)
(DC)
in
in DCmicrogrid
DC
different power systems
through
through example
Lyapunov are
of
ofby presented
an
an electrical
electrical
approach to et Ñañez alternate
system.
system.
control etAA DAE
DAE
alternate (2017) current
model
model
current
MGsthe
including
with
with
MGs
is
an
the
is
cially
(MG)
(MG)
a renewable
commoninterconnects
interconnects
bus, sources.
DERs
each DERs A
of response
them direct
that
that
might current
operate
operate have (DC)
in DC
differentmicrogrid
through
DC throughpower Lyapunov
presented
example approach
ofby De Persis
anDeelectricalto control
al.
system. alternate
(2016), A DAE where current
model MGs
algebraic
with is
con-
the
a common
capacity and bus, each
dynamical of them might Justo have
et different
al. (2013). power
Some Lyapunov
Lyapunov
presented approach
approach Persisto
to control
control
et al. alternate
alternate
(2016), where current
current MGs
MGs
algebraic is
is
con-
(MG)
aacapacity
common
commoninterconnects
bus,
bus, each DERs
each of them
of response
them that
might
might operate have
have in DC through
different
different power
power presented
ditions correspondby De Persis
totoan et al. (2016),
equilibrium where
between algebraic
demanded con-
acapacity
of common
the main
capacity
and
and
and
dynamical
dynamical
objectives
bus, each ofof
dynamical
response
an isolated
them
response might
Justo
Justo
Justo MG etare
et
have
et
al.
al. (2013).
(2013).
voltagepower
different
al. (2013). regu- Lyapunov
Some
Some
Some
presented
presented
ditions
ditions
power
approach
by
by De
De Persis
correspond
correspond
and power Persis
to point.
to
set an
an
control
et
et al. alternate
(2016),
al.This
(2016),
equilibrium
equilibrium
where
where
between
between
approach
current
does
MGs
algebraic
algebraic
demanded
demanded
not con-
is
con-
con-
con-
capacity
of
of the
the
lation, main
mainand dynamical
objectives
objectives
power-sharing orof
ofresponse
an
an
the isolated
isolated
conditionJusto MG
MG et
that al.
are
are (2013).
voltage
voltage
each Some
regu-
regu-
converter presented
ditions
ditions
power by De
correspond
correspond
and power Persis
to
to
set an
an et
point. al. (2016),
equilibrium
equilibrium
This where
between
between
approach algebraic
does demanded
demanded
not con-
capacity
of
of the
the main
lation, mainandobjectives
dynamical
objectives
power-sharing orofresponse
oftoan
anitsisolated
the isolated
conditionJusto MG
MG etare
that al.each
are (2013).
voltage
voltage Some
regu-
regu-
converter power
ditions and
sider switched power
correspond set
events. point.
Gross
to point. This
an equilibrium approach
et al. approach
(2018) between doesdemanded
consider not con-
failures
lation,
ofshares
the
lation,
power-sharing
power
main according
objectives
power-sharing
or
orof the
an
the
condition
maximum
isolated
condition MG that each
capacity,
thatare converter
voltage
each and the power
regu-
converter
power
sider
sider
or
and
and
switched
switched
variations
power
power
in
set
set
events.
events.
the point.
Gross
Gross
network’s
This
This
et al. approach
ettopology
al. (2018)
(2018) as
does
does not
consider
consider
switched not con-
con-
failures
failures
events
lation,
shares
shares power-sharing
power
power
plug-and-play according
according
capacity or totheits
to itscondition
which maximum
maximum
is the that each
capacity,
capacity,
capacity of converter
and
connect-the power
and the sider
sider
or and power
switched
switched
variations in set
events.
events.
the point.
Gross
Gross
network’s This
et
et al.
al.
topology approach
(2018)
(2018) as does not
consider
consider
switched con-
failures
failures
events
lation,
shares
shares power-sharing
power
power according
according or totheits
to itscondition
maximum
maximum that each
capacity,
capacity, of converter
and
and the
the sider or variations
applied switched in the
to electrical network’s
systems
Gross etmade topology as
of rotatoryswitched events
machines.
plug-and-play
plug-and-play
ing or power
shares
capacity
capacity
disconnecting
according
which
which
a to
converter is the
is
its maximum
theatcapacity
capacity
ancapacity, of
arbitrary connect-
connect-
andtime.
the or
or
or variations
variations
applied
applied
Mojica-Nava inevents.
in the
to electrical
to electrical
et al.
network’s
the(2019)
network’s
systems
systems present
al. (2018)
topology
topology
made
made a as consider
as
of rotatory
of
switched
switched
rotatory
complete
failures
events
events
machines.
machines.
decentralized
plug-and-play
plug-and-play
ing or capacity
capacity
disconnecting a which
which
converter is
is the
theat capacity
capacity
an of
of
arbitrary connect-
connect-
time. variations
applied to in the
electrical network’s
systems topology
made of as switched
rotatory events
machines.
ing or
Approachesdisconnecting
plug-and-play for MG a
controlconverter
can beat an arbitrary
classified as time.
central- applied
Mojica-Nava
Mojica-Nava to electrical
et al. systems
(2019)
et al. (2019) present
presentmade a of rotatory
complete
a complete machines.
decentralized
decentralized
ing
ing or
Approaches
ized,or forcapacity
disconnecting
ordecentralized,
Approachesdisconnecting
for MGand
MG
aa which
control
control
converter
converter
can
can
distributed
is the
beat
beat capacity
an
an arbitrary
classified
Guerrero
of
arbitrary
classified etasas
as
connect-
time.
time. applied
central-
central-
al.central-
(2011).
switchedtocontroller
Mojica-Nava
Mojica-Nava
switched electrical
et
et al.
controlleral.
for
(2019)
(2019)
for
a DC
systems
a DC
MG
present
present
DC MGmade whose
aa complete
complete
whose
stability
of rotatory is proved
machines.
decentralized
decentralized
stability is proved
proved
ing
Approaches
Approaches
ized, disconnecting
for
decentralized, MG
for MGand a
control
control
and converter
can
can
distributed beat an arbitrary
classified
beGuerrero
classified
Guerrero etas time.
al.central-
(2011). switched
using a
Mojica-Nava controller
passivityet al. for
approach,
(2019) a MG
the
present whose
switched
a complete stability
control is
commutes
decentralized
ized, decentralized,
Centralized controllers distributed
need a robust et
communication al. (2011).
sys- switched
switched
using a controller
controller
passivity for
for
approach, aa DC
DC MG
MG
the whose
whose
switched stability
stability
control is
is proved
proved
commutes
Approaches
ized,
ized, for MGand
decentralized,
decentralized,
Centralized control
and
controllers can
distributed
distributed
need beGuerrero
classified
Guerrero etas
et al.central-
al. (2011).
(2011). using
between
sys- switched a passivity
two
controller approach,
subsystems
for a DC the
depending
MG switched
whose control
onstability
the error commutes
of the
is proved
Centralized
tem and
ized, controllers
thecontrollers
changesand
decentralized,
Centralized theiraaastructure
need
in need
distributed
robust
robust
robust
communication
communication
Guerrerorequire etrecalculate
communication
sys-
al. (2011).
sys-
using
using
between
between
aa passivity
generated passivity
two
two
approach,
approach,
subsystems
subsystems
voltage compared
the
the
depending
depending
switched
switched
to switched on control
on
a reference. control
theAterror
the
commutes
commutes
error
this of the
of the
article
Centralized
tem
tem and
and
control the
the controllers
changes
changes
parameters in
in
such need
their
their
as a robust
structure
structure
gains. A communication
require
require
distributed sys-
recalculate
recalculate
controller using
between
between a
generated passivity
two
two approach,
subsystems
subsystems
voltage compared the
depending
depending
to a on
on
reference. control
the
the At commutes
error
error
this of
of the
the
article
Centralized
tem
tem and
and
control thecontrollers
the changes
changes
parameters in need
in
such their
their
as an astructure
robust
structure
gains. communication
require
require
A distributed
distributed sys- between
recalculate
recalculate
controller generated
we extendtwo voltage compared
thesubsystems
lastcompared
approach totoaareference.
depending switched
on theAt At thiscreated
event
error article
of the
control
allows parameters
flexibility; such
however, as gains. A
appropriate controller
communication generated
generated
we extend voltage
voltagelastcompared
the last approach totoaaaareference.
toto reference.
switched At this article
thiscreated
event article
created
tem
control
control
allowsand the changes
parameters
parameters
flexibility; in
such
such
however, their
as
as structure
gains.
gains.
an A
A
appropriate require
distributed
distributed recalculate
controller
controller
communication we
when extend
generated a the
converter
voltage approach
is connected
compared to a orswitched
disconnected
reference. event
At this to the
article
allows flexibility; however, an appropriate communication we
we
when extend
extend the
the last
last approach
approach to
to a
a switched
switched event
event created
created
network
control
allows
allows
network
is also necessary.
parameters
flexibility;
flexibility;
is also such as
however,
however,
necessary.
Inan
In an
a decentralized
gains.
a A distributed
appropriate
appropriate
decentralized
approach,
communication
communication
approach,
sev- when
controller
sev- MG.
we extendaa converter
Also, converter
an
the analysis
last
is connected
is connected
of
approach the Zeno
to a
orbehavior
or disconnected
disconnected
switched is
event
to the
to
performed the
created
network
eral
allows is also
changes can
flexibility;necessary.
be performed
however, Inan a decentralized
over the
appropriate system approach,
parameters
communication sev- when when
MG.
MG.
aa converter
Also,
Also, converter
anfast
an analysis
analysis
is
is connected
connected
of the
of Zenoor
thebetween
Zeno
disconnected
orbehavior
disconnected
behavior to
to the
is performed
is performed the
network
network is
is also
cannecessary.
also necessary. In
In aa decentralized
decentralized approach,
approach, sev-
sev- MG. because
when this
a thisconverter transitions
is connected subsystems endanger
eral
eral changes
changes
and changes
network the is can be performed
be
plug-and-play
also
performed
necessary. capacity over
over the
the
is the
better
In a decentralized
system
system parameters
parameters
compared
approach, with
sev- MG. MG.
because
because
Also,
Also,
the components
an
thisan
analysis
anfast
analysis
fast transitions
transitions
of the of
of the
ofconverterZenoor
Zeno
thebetween
between disconnected
behavior
behavior
and subsystems
subsystems
degrade
is to the
is performed
performed
endanger
endanger
the power
eral
eral
and changes
the can
can be
be
plug-and-play performed
performed
capacity over
overis the
better system
system parameters
parameters
compared with because Also, this analysis
fast transitions the Zeno
between behavior
subsystems is performed
endanger
and
eral the plug-and-play
the other
changes approaches
can be capacity
because
performed is the
better
a complete
over compared
communication
system parameterswith because the
the this fast transitions
components
components of for
of the the
the between
converter
converter andsubsystems
and degrade
degrade endanger
the
the power
power
and
and
the the
other plug-and-play
theisplug-and-play
approaches capacity
capacity
because is
a al. better
is (2017).
complete compared
better communication with
compared with because
communication quality.
the Conditions
this fast transitions
components of the existence
between
converter and of Zeno
subsystems
degrade behavior
endanger
the are
power
the other
system
and approaches
not essential
theisplug-and-play because
Meng
capacity et a complete
is (2017).
better communication
compared with the the components
quality.
quality. Conditions
Conditions of the
for
for converter
the
the existence
existence and of
of degrade
Zeno
Zeno the
behavior
behavior power
are
are
the
the other
other
system approaches
approaches
not essential because
because
Meng etaa al.
complete
complete communication presented
components in Ames of et
the al. (2005)
converter for
and a class
degrade of
the system
power
system
the otheris not essentialbecause
approaches Meng eta al. (2017). communication quality.
complete quality.
presented
presented
Conditions
Conditions
in Ames
in Amesfirstfor
et the
for
et the
al.
al.
existence
existence
(2005) (DFQ)
(2005) forof
of aaZeno
for Zeno
class
class
behavior
behavior
ofsystems.
of
are
are
system
system
system is
is not
The connection
system not essential Meng
Meng et
or disconnection
essential et al. of (2017).
al. a DER implies abrupt quality.
(2017). known as diagonal
Conditions for quadrant hybrid
The connection or disconnection of (2017).
presented
presented
known as in
in etAmes
Ames
diagonal et the
et
first al.
al. existence
(2005)
(2005)
quadrant forof aaZeno
for
(DFQ) class
classbehavior
hybrid of
ofsystems.are
system
system
systems.
system
The
changes isinnot
connectionthe essential
or
MG’s Meng etand
disconnection
power, al.
of aa DER
DER implies
voltages implies
can abrupt presented
abrupt
generate known
In Heymann
known
as
as
diagonal
in Ames
diagonal
first
al. (2002)
et
first
quadrant
are
al. presented
(2005)
quadrant
(DFQ)
for
(DFQ) athehybrid
conditions
class
hybrid of system
systems. for
The
The connection
connection
changes in the or
or disconnection
disconnection
MG’s power, and of
of aa DER
DER
voltages implies
implies
can abrupt
abrupt
generate known
In
In
the Heymann
Heymann as diagonal
existence et
et
of al.
al.
the first
(2002)
(2002)
Zeno quadrant
are and
are
set presented
presented(DFQ)
Zeno the
thehybrid systems.
conditions
conditions
executions. for
for
changes
jumps in
and the MG’s
impulse power,
variations, and voltages
endangering can generate
the generate
compo-
abrupt known as diagonal first
The
changes
changes
jumps connection
in
in
and the
the or disconnection
MG’s
MG’s
impulse power,
power,
variations, and
and of voltages
a DER implies
voltages
endangering can
canthe generate
compo-
In
theHeymann
In
the Heymann
existence
existence
et
of al.
et
of al.
the Zenoquadrant
the(2002)
(2002)
Zeno
are
are and
set
set
presented
presented
and (DFQ)
Zeno
Zeno
the
thehybrid
executions.
executions.
systems.
conditions
conditions for
for
jumps
nents and and impulse
inloads variations,
and variations,
degrading endangering
the power quality the generate
compo-
Kumar In
changes
jumps
jumps
nents and
and
and the MG’s
impulse
impulse
loads power, and
and variations,
degrading voltages
endangering
endangering
the power canthe
the compo-
quality compo-
Kumar TheHeymann
the
the principalet
existence
existence of
of al.
the(2002)
the Zeno
Zeno set
contribution are
set presented
of and
this Zeno
and Zeno the
paperexecutions. conditions
is the
executions. analysisfor of
nents
et
jumps al. and
and loads
(2017). and
Switched
impulse degrading
events
variations, the
can power
be
endangering quality
represented
the Kumar
as
compo- a The
the
The principal
existence
principal contribution
of the Zeno
contribution setof
of this
and
this paper
Zeno
paper is the
executions.
is the analysis
analysis of
of
nents
nents
et al. and
and loads
(2017). and
loadsSwitched degrading
and degrading
Switched events the
the power
can power quality
quality
be represented
represented Kumar
Kumar
as aa The a DC MG
principal when DERs
contribution are connected
of this or
paper disconnected
is the analysis fromof
et
set
nents al.
of (2017).
differential-algebraic
and loads and events
degrading equations can
the be
(DAEs)
power to
quality as
determine
Kumar The
a DC
a DC principal
MG
MG when
when contribution
DERs
DERs are
are of this
connected
connected paper
or
or is the analysis
disconnected
disconnected from
fromof
et
et
set al.
al. (2017).
(2017). Switched
Switched
of differential-algebraic
differential-algebraic events
events
equations can
can be
be represented
(DAEs) represented
to determine as
as aa The
determine a DCDAE’s
principal
MG perspective,
when contribution
DERs considering
of this paper the space
is the constraint,
analysis of
set
et of
if there
al. are impulses
(2017). Switched equations
andevents
jumps in the
can (DAEs) states.to
be represented Examplesas a aaajump-free,
DC
DAE’s
DAE’sMG perspective,
when DERs are
perspective,
and impulse-free are connected
connected
considering
considering
conditions.
or
or disconnected
the
the disconnected
space
space
from
from
constraint,
Also, constraint,
the stabil-
set
set
if of
of
theredifferential-algebraic
differential-algebraic
are impulses and equations
equations
jumps in (DAEs)
(DAEs)
the states.to
to determine
determine
Examples DC
DAE’sMG when DERs
perspective, are connected
considering or
the disconnected
space from
constraint,
if there aresystems
impulses and jumps in the states. Examples
Ñañez ajump-free,
DAE’s perspective, considering
ifforthere
set
if
for
power
of
there
power are impulses
aresystems
impulses
systems
areand
differential-algebraic
areand
presented
equations
jumps
jumps
presented
in the
in
in Trenn
(DAEs)
the
Trenn
(2009),
states.
states.to determine
(2009),Examples
Examples
Ñañez
jump-free,
ityDAE’s
ajump-free,
and impulse-free
and
of the perspective,
and
impulse-free
controller
impulse-free usingthe
conditions.
conditions.
is considering
shownconditions. the
space
Also, constraint,
Also,
Lyapunov
space
Also,
the
the stabil-
stabil-
functions.
constraint,
the stabil-
for
ifet power
al.
there (2017),
aresystems are presented
Domı́nguez-Garcı́a
impulses inand Trenn Trenn (2009),
(2010).Ñañez
In jump-free,
ity
ity of
of the
the and impulse-free
controller
controller is
is shown
shown conditions.
using
using Also,
Lyapunov
Lyapunov the stabil-
functions.
functions.
for
for
et
et
power
power
al.
al.
Trenn systems
(2017),
(2017),
(2012), areand
are jumps in
presented
presented
Domı́nguez-Garcı́a
Domı́nguez-Garcı́a
conditions for stability
in
inand
and the
Trenn
Trenn
and
states.
Trenn
Trenn
(2009),
(2009),Examples
(2010).
(2010).
solvability
Ñañez
In
un-
Jumps,
In jump-free,
Ñañez ity
ity of
of the
Jumps, the
impulses,
controller
impulses,
and
and impulse-free
controller and is Zeno conditions.
is shown
shown
behavior
using
using Lyapunov
Zeno behavior
behavior
are
Also,
Lyapunov
are
studied
the using
stabil-
functions.
functions.
studied using
for
et power
al. systems
(2017), are presented
Domı́nguez-Garcı́a inand Trenn Trenn (2009),
(2010).Ñañez
In Jumps,
ity of the impulses,
controller and is Zeno
shown using are
Lyapunov studied using
functions.
et al. (2012),
Trenn
Trenn (2017), conditions
(2012), Domı́nguez-Garcı́a
conditions for stability
for stability andand andTrenn (2010). un-
solvability
solvability In Jumps,
un- Jumps, impulses,
impulses, and and Zeno Zeno behavior
behavior are are studied
studied using
using
et
Trenn
Trennal. (2012),
(2017), conditions
(2012), Domı́nguez-Garcı́a
conditions for
for stability
stability andand andTrenn (2010). un-
solvability
solvability In Jumps, impulses, and Zeno behavior are studied using
un-
2405-8963 Copyright © 2021 The Authors. This is an open access article under the CC BY-NC-ND license.
Trenn (2012), conditions for stability and solvability un-
Peer review under responsibility of International Federation of Automatic Control.
10.1016/j.ifacol.2021.10.356
222 Vladimir Toro et al. / IFAC PapersOnLine 54-14 (2021) 221–226

the proposed switched control. Finally, the conditions for The output of the system yi is the voltage at each converter
power-sharing and optimal power dispatch are presented. Vi , so expression (2) is a state-dependent switching control
law with yi = Vi . The feedback control law determines
The rest of this paper is organized as follows. Section
the current supplied by each source (Mojica-Nava et al.
2 presents the electrical model for the microgrid with
(2019))
decentralized control and related Lyapunov functions. In
ui = −yi (3)
Section 3, the connection of a converter to the MG is
studied using tools for DAEs, stability is analyzed based To simplify (1) , we work with the admittance values given
on consistency space and projectors, and conditions for by Yij = R1ij . Connections are represented by a weighted
optimal dispatch are considered too. Simulation results are Laplacian matrix Lw whose elements are defined by
presented in Section 4. Finally, conclusions are presented n
in Section 5. Lij = Yij if i = j Lij = −Yij if i = j
j∈Nij
2. DC MICROGRID MODEL AND SWITCH with Yij = 0 and YN i = R1N i = 0. The load terms
CONTROL
YLi = R1Li = are included in the Lw matrix.
  n 
For modeling the microgrid the next DC converter model
based on current is presented  Y ij −Y 12 · · · −Y 1n 
 j∈Nij 
 n 
  
 −Y Yij · · · −Y2n 
 21 
Lw =  j∈Nij


 .. .. .. .. 
 . . . . 
 
 n
 
 
−Yn1 −Yn2 · · · Yij
j∈Nij
Figure 1. Two DC sources connected through a transmis- Defining the set of inputs Isi as the column vector given
sion line 
by Is = [Is1 Is2 · · · Isn ] . The system (1) without
control input can be written in a matrix form as
Two converters are connected in parallel through a trans-
C V̇ = Is − Lw V
mission line Rij as shown in Figure 1. Isi is the current (4)
ILn = −Lw V
source, usually controlled by PWM (Pulse Width Modu-
lation), RN i is the internal resistance of the source (which where C ∈ Rn×n is a diagonal matrix whose elements are
will be omitted assuming a very high value), and RLi is the the capacitance values Ci , and ILn is the line current Iij .
load for converter i. This model is extended to n parallel
Mojica-Nava et al. (2019) demonstrated the stability for
converters as follows
n system (4) based on a passivity approach. System (4) is
dVi Vi  1
Ci = Isi − + (Vj − Vi ) + ui (1) asymptotically stable for closed-loop under switched con-
dt RLi Rij trol law (2) with Lyapunov function Hp (v) = Hq (v) where
j∈Ni
n v denotes the vector of voltages V , p and q denote the
 1 function before and after the switching event, respectively.
Iij = (Vj − Vi )
Rij Defining the Lyapunov function Hp (v) = Hq (v) = 21 v 2 for
j∈Ni
the close-loop system in both cases:
where Ci is the parallel capacitance for filtering, Vi is the
voltage for DGi , Ni represents the set of neighbors of node ∆V ≥ 0, replacing the value of Ḣ(v) = v dv
dt
i, Iij is the current between converter i and j, and ui is  
Is Lw V
the control term. Ḣ = V − ≤0
C C
The next state-depending switching law is proposed by  
−PDG Lw V
Mojica-Nava et al. (2019) Ḣ = V − ≤0
 VC C
 PDGi
 ∆Vi ≥ 0 −PDG Lw V 2
ui = P Vi (2) Ḣ = − ≤0
2 C C
 DGi − k∆Vi

∆Vi < 0 −PDG − Lw V ≤ 02
Vi
where ∆Vi = Vref − Vi is the difference between the ∆V < 0
 
voltage reference Vref and the voltage at the i-th converter. −PDG − k∆V 2 Lw V
Ḣ = V − ≤0
PDGi = Vi Isi is the power at the ith converter, and k is VC C
a constant value for setting the effect of voltage variation, PDG ∆V 2 Lw V 2
usually 0 < k ≤ 1. Ḣ = − − − ≤0
C C C
2 2
Microgrid model assumptions: The voltage in the ca- −PDG − ∆V − Lw V ≤ 0
pacitors is strictly positive Vi > 0, and current sources In both cases, Lw represents the values of the positive-
are assumed to be stable and with a large enough internal definite Laplacian matrix Lw , PDGi are the positive power
resistance capable of supplying any load RN i −→ ∞. values of a diagonal matrix PDG which is positive definite,
 Vladimir Toro et al. / IFAC PapersOnLine 54-14 (2021) 221–226 223

and ki is a positive value. Then, Ḣ remains negative along where σ = {1, 2, · · · , p} is the variable that represents the
the trajectories, Hp (x) = Hq (x) are Lyapunov functions switch state, Ep , and Ap ∈ Rn×n , Bp ∈ Rn×m , x ∈ Rn is
for the system, and the system is asymptotically stable. the state vector, and u ∈ Rm is the input vector (Trenn
(2012)).
Switched systems might present some behaviors exclusive
Definition 2. Admittance matrix The connection among
for these systems. Transitions among subsystems depend
converters at t < 0s is defined by the admittance matrix
on the states’ conditions,the number of transitions is
Lw1 , at t = 0s an additional converter is (dis)connected
limited in a real model. However, when modeling it might
with admittance matrix defined by Lw2 .
occur an infinite number of transitions in a finite time, this
is known as Zeno behavior or Zenoness (Heymann et al. Assumption 1: We assume the total number of convert-
(2005)). ers is previously known, so n − 1 converters are intercon-
nected at t < 0s with ILn = 0, then at t = 0s converter n
is connected.
The initial associated weight value changes from zero
to a new value for that converter. System equations
are presented before and after the switching event in a
matrix form. The DAEs equations with pairs (E1 , A1 ) and
(E2 , A2 ) are written in matrix form.
t < 0       
C 0 V̇ −Lw1 0 V 1
= + U1 (5)
Figure 2. Converter switched control diagram 0 0 I˙Ls Lw1 Is ILs 0
t≥0
Definition 1. The existence of a nonempty Zeno set indi-        
cates the existence of a possible Zeno equilibrium (Zheng C 0 V̇ −Lw2 0 V 1
= + U2 (6)
(2006)). 0 0 I˙Ls Lw2 Is ILs 0
where U1 is the vector of inputs U1 = [u1 u2 · · · un−1 ],
The Zeno set corresponds to the intersection of the do- and U2 = [u1 u2 · · · un ].
mains of each subsystem. The sets are given by
S1 : Vi ≥ Vref S2 : Vi < Vref The DAEs description already represents jumps and im-
pulses. These are allowed in the system’s solution by
The Zeno set is given by
selecting an appropriate distributional framework Trenn
Zc = S1 ∩ S2 = {∅} (2012). Next, some important definitions for the switched
Then, the set Zc is empty so there are not Zeno points in system are given.
system (1) under switched control law (2). Definition 3. Regularity (Trenn (2012)). Matrix pairs (E1 ,
A1 ), (E2 , A2 ), · · · , (En , An ) are regular if the pencil
3. DC MICROGRID UNDER SWITCH EVENT (sEσ −A) is regular, it means if det(sE−A) is a polynomial
different from zero. This condition guarantees that the
Suppose a microgrid with n converters, then the switch Si switched system always has a solution, no matter the
with i = {1, 2, · · · , n} is closed at t = 0s connecting the initial value.
nth converter to the MG through a transmission line Yij
as shown in Figure 3. The weight values for converter n This condition is fulfilled by systems (5) and (6) whose
change depending on if it is connected or not. At t < 0s determinants are given by ∆1 = −sC − Lw1 and ∆2 =
values Yn,n are set as zero, so all values for column n and −sC − Lw2 .
row n are zero. Definition 4. Consistency space and projectors (Liberzon
and Trenn (2012)). Solutions for the switched system are
given inside a space known as the consistency space. Ma-
trices pairs (Ei , Ai ) are regular if they can be represented
in the quasi-Weierstrass form.
The Wong sequences give an algorithmic way to determine
the quasi-Weierstrass form. The consistency projectors
Π(E,A) , the differential projectors Πdiff
(E,A) , and the impulse
projectors Πimp
(E,A) are defined by Trenn (2012). Those are
calculated for systems (5) and (6) as follows
Figure 3. General diagram of a DC MG. The fifth converter
is connected when the switch s5 is closed Π(E1 ,A1 ) = I2n×2n Π(E2 ,A2 ) = I2n×2n

After the switch event, each converter updates its state Then, the space is consistent.
variables according to Kirchhoff laws. The system is rep- Definition 5. Impulse-free (Trenn (2012)). A switched sys-
resented as a differential-algebraic system in the canonical tem defined by a set of DAEs, with regular matrix pairs
form (Ep , Ap ) cannot produce an impulse for all solutions if the
Eσ(t) ẋ(t) = Aσ(t) x(t) + Bσ(t) u(t), following impulse-free condition holds:
224 Vladimir Toro et al. / IFAC PapersOnLine 54-14 (2021) 221–226

Eq (I − Πq )Πp = 0 converter to supply the demand according the maximum

Definition 6. Sufficient condition for jump free (Trenn power capacity at each converter.
(2009)):
Proof. In steady-state, the equilibrium equation is given
(I − Πq )Πp = 0.
by
Definition 7. Condition for inputs Ñañez et al. (2017). For U − Lw V = 0, (8)
each σ ∈ Σ,
Πimp
σ Bσ = 0
Assuming a quadratic cost function given by uc =
 1 2
Definition 8. Stability condition for switched systems i∈N 2 ri ui , subject to (8). Applying the Lagrange mul-
Trenn (2012). For a switched DAE with regular pairs tipliers method, we have the next equation

(Ep , Ap ) ∈ {1, · · · , p̄}, and consistency projectors ΠEp ,AP , L(u, v, λ) = ri u2i + λ(u − Lw V ), (9)
the system is asymptotically stable for all switching signals i∈N
if 
with ∇Lu = i∈N ri ui + λ = 0, ∇Lv = −λ(−Lw ) = 0,
1) Ep ẋ = Ap x is asymptotically stable with Lyapunov and ∇Lλ = u − V (−Lw ) = 0. Matrix −Lw can be written
function Hp , as an admittance matrix that includes the values of the
2) Fulfills the impulse-free condition Eq (I − Πq )Πp = 0 for local loads. Then, the solution of ∇Lλ = u − V (−Lw ) = 0
all q ∈ (1, ..., p), and are the set of values λ = c1n , with c ∈ R. Then, the
3) Hq (Πq x) ≤ Hp (x) for all x ∈ Hp∗ =imΠp . condition for the optimal control input is given by
Theorem 9. Consider the switched system represented by c
DAEs (5) and (6) with decentralized control (2) with ui = − , (10)
ri
regular matrix pairs (E1 , A1 ), (E2 , A2 ), and Lyapunov
functions Hp and Hq . If the system satisfies the conditions Then, the optimal input value for the ith converter is
of definition (8), then the system is asymptotically stable. given as a function of constant c over the cost value
 1  
Proof. The impulse-free condition is satisfied for both ri . Where c i∈N = i∈N Isi + i∈N IL (Zhao and
ri
subsystems Dörfler (2015)). As control (2) acts as a droop control in
q=1 q=2 two states, being the same as ∆V = 0, optimal condition
p = 1 Eq (I − Πp )Πp = 0 Eq (I − Πp )Πp = 0 (10) is fulfilled, and the theorem is demonstrated.
p = 2 Eq (I − Πp )Πp = 0 Eq (I − Πp )Πp = 0
4. SIMULATION RESULTS
And the jump condition (8) Hq (Πq x) ≤ Hp (x) is fulfilled
for the Lyapunov function, with 1/2(Πq x)2 < Hp (x). Then A five converter DC microgrid is simulated in Simulink
the system is asymptotically stable. considering two cases as follows
3.1 Optimal Control Criteria and Power Sharing 4.1 Case 1
Different resistance values at each transmission line cause The DC MG is simulated using the model of Figure 3,
that voltages differ from the load side compared with the four interconnected converters share power, a linear load
output side at each converter. That results in poor voltage is connected at t = 0.2s, then a fifth converter (n = 5) is
regulation and insufficient power-sharing which is fulfilled connected at t = 0.4s. The MG parameters are shown in
achieving the next condition Table 1.
Table 1. MG parameters
Pi Pj
= ··· = . (7)
Pmax,i Pmax,j Converter 1 2 3 4 5
Maximum Active
360 330 300 270 240
If the reference voltage is achieved as t → ∞, the last Power (W)
Ii Ij Line Resistor Ri (Ω) 0.64 0.50 0.90 0.70 0.82
expression can be rewritten as Imax,i = · · · = Imax,j , which
Output Capacitor 2.2mF
is an equilibrium among each source. Defining value ri = Vref 80V
1
Imax,i , power sharing condition is equal to ri Ii = · · · = rj Ij k 10Ω−1
or ri Pi = · · · = rj Pj . Local load 5.0 Ω

Optimal control for MGs solves the problem of dispatching The Weighted Laplacian matrix is given by Lw
DERs based on its marginal cost and availability. It is not
always desirable that a source supplies power according  
to its maximum capacity. Sometimes a converter can −13.77 3.56 2.67 2.99 2.78 1.76
have major priority over others to reduce the cost of  3.56 −15.52 3.11 3.42 3.21 2.20 
 
power generation. This requirement corresponds to a cost  2.67 3.11 −11.96 2.53 2.33 1.31 
Lw = 
minimization problem with restrictions.  2.99 3.42 2.53 −13.23 2.64 1.62 
 2.78 3.21 2.33 2.64 −12.39 1.41 
Theorem 10. Assume that system defined by (1) is asymp-
totically stable with control law (2), where each source has 1.76 2.20 1.31 1.62 1.41 −8.32
a dispatch factor ri . Then, the power-sharing condition is
also the optimal power dispatch condition for the system, Figure 4 shows the voltage at each converter. A load is
this means the optimal power to be delivered for each connected at t = 0.2s, and converter five is connected at
 Vladimir Toro et al. / IFAC PapersOnLine 54-14 (2021) 221–226 225

t = 0s. The interrupter is closed at t = 0.3s connecting

the fourth and fifth converters. The parameters of the
MG are shown in Table 2, and the physical connections
are represented by matrices Lw1 and Lw2 for the instants
before and after closing sn , respectively.
Table 2. MG parameters
Figure 4. Voltage output at each converter for Case 1 Converter 1 2 3 4 5
Maximum Active
360 330 300 270 240
Power (W)
Line Resistor Ri (Ω) 0.64 0.50 0.90 0.70 0.82
Output Capacitor 2.2mF
Vref 80V
k 10Ω−1
Local load 5.0 Ω
R12 = R23 = R34 = 0.1 Ω
R45 = R15

 
−1.8333 0.8333 0 0 1.0000
 0.8333 −1.2878 0.4545 0 0 
Figure 5. Current measured at each converter for Case 1  
Lw1 = 0 0.4545 −1.7045 1.2500 0 
 0 0 1.2500 −2.5833 1.3333 
1.0000 0 0 1.3333 −2.3333
 
−3.8333 0.8333 0 2.0000 1.0000
 0.8333 −1.2878 0.4545 0 0 
 
Lw2 = 0 0.4545 −1.7045 1.2500 0 
 2.0000 0 1.2500 −.5833 1.3333 
Figure 6. Power Sharing Among Converters for Case 1 1.0000 0 0 1.3333 −2.3333
t = 0.4s. Notice the difference in the voltage regulation
before and after connecting the load. Also, voltage reg-
ulation improves when the fifth converter is connected.
The generated voltage does not present peaks or jumps.
Figure 5 and Figure 6 show the current and power at each
converter, respectively. The current and power-sharing
condition is achieved. There are current peaks because
of the load connections, but mostly because of the fifth
converter’s connection. Figure 8. Voltage output at each converter for Case 2

4.2 Case 2

Figure 9. Current output at each converter for Case 2

Figure 10. Power measured at each converter for Case 2

The voltage at each converter is shown in Figure 8. The
load is connected at t = 0s, and the fifth converter is
connected at t = 0.3s. After the system starts, voltage
Figure 7. Second case MG model outputs stabilize in t < 0.1s, as expected voltage regulation
improves when the fifth converter is connected. Voltage
Here, the MG has the configuration presented in Figure 7. output presents neither peaks nor oscillations. The power-
A linear load is connected to the second converter at sharing condition is kept, as shown in Figure 9 and
226 Vladimir Toro et al. / IFAC PapersOnLine 54-14 (2021) 221–226

Heymann, M., Lin, F., Meyer, G., and Resmerita, S.

(2005). Analysis of Zeno behaviors in a class of hybrid
systems. IEEE Transactions on Automatic Control,
50(3), 376–383. doi:10.1109/TAC.2005.843874.
Johansson, K.H., Egerstedt, M., Lygeros, J., and Sastry, S.
(1999). On the regularization of Zeno hybrid automata.
Systems and Control Letters, 38(3), 141–150. doi:
Figure 11. Control signal applied at the first converter for 10.1016/S0167-6911(99)00059-6.
Case 2 Justo, J.J., Mwasilu, F., Lee, J., and Jung, J.W. (2013).
AC-microgrids versus DC-microgrids with distributed
Figure 10. However, as in the first case, voltage regulation energy resources: A review. Renewable and Sustainable
is not as good as expected. There is a trade between power Energy Reviews, 24, 387–405.
sharing and voltage regulation as shown in Mojica-Nava Kumar, D., Zare, F., and Ghosh, A. (2017). DC Microgrid
et al. (2019), and depends directly on selecting the factor Technology: System Architectures, AC Grid Interfaces,
k. Figure 11 shows the control signal at the first converter Grounding Schemes, Power Quality, Communication
it changes at t = 0.3s when the additional converter Networks, Applications, and Standardizations Aspects.
is connected and presents a small ripple because of the IEEE Access, 5, 12230–12256.
switching. Liberzon, D. and Trenn, S. (2012). Switched nonlinear
differential algebraic equations : Solution theory , Lya-
5. CONCLUSION
punov functions , and stability. Automatica, 48(5), 954–
963.
A DC MG with a decentralized controller was modeled
Meng, L., Shafiee, Q., Trecate, G.F., Karimi, H., Fulwani,
as a DAEs problem when an additional converter was
D., Lu, X., and Guerrero, J.M. (2017). Review on Con-
connected or disconnected. Stability, impulse-free, and
trol of DC Microgrids and Multiple Microgrid Clusters.
jump-free conditions were demonstrated. Stability for the
IEEE Journal of Emerging and Selected Topics in Power
controller was proved using a Lyapunov functions ap-
Electronics, 5(3), 928–948.
proach, and also non-Zeno behavior was verified. It was
Mojica-Nava, E., Rey, J.M., Torres-Martinez, J., and
checked that variations over the network’s topology when
Castilla, M. (2019). Decentralized Switched Current
converters are connected could also be analyzed using the
Control for DC Microgrids. IEEE Transactions on
DAEs approach. It was demonstrated that the decentral-
Industrial Electronics, 66(2), 1182–1191.
ized control fulfills the power-sharing condition. Also, the
Trenn, S. (2009). Regularity of distributional differential
conditions for optimal power dispatch were presented.
algebraic equations. Mathematics of control, signals,
REFERENCES and systems, 21(3), 229.
Trenn, S. (2012). Switched Differential Algebraic Equa-
Ñañez, P., Sanfelice, R.G., and Quijano, N. (2017). No- tions, 189–216. Springer London, London.
tions and a passivity tool for switched dae systems. In Zhao, J. and Dörfler, F. (2015). Distributed control and
2017 IEEE 56th Annual Conference on Decision and optimization in DC microgrids. Automatica, 61, 18–26.
Control (CDC), 3612–3617. Zheng, H. (2006). Simulating zeno hybrid systems beyond
Ames, A.D., Abate, A., and Sastry, S. (2005). Sufficient their zeno points. Technical report, Technical Report
conditions for the existence of zeno behavior. In Pro- No. UCB/EECS-2006-114.
ceedings of the 44th IEEE Conference on Decision and
Control, 696–701.
De Persis, C., Monshizadeh, N., Schiffer, J., and Dörfler, F.
(2016). A lyapunov approach to control of microgrids
with a network-preserved differential-algebraic model.
In 2016 IEEE 55th Conference on Decision and Control
(CDC), 2595–2600.
Domı́nguez-Garcı́a, A.D. and Trenn, S. (2010). Detection
of impulsive effects in switched daes with applications
to power electronics reliability analysis. In 49th IEEE
Conference on Decision and Control (CDC), 5662–5667.
Gross, T., Trenn, S., and Wirsen, A. (2018). Switch in-
duced instabilities for stable power system DAE models.
IFAC-PapersOnLine, 51(16), 127–132.
Guerrero, J.M., Vasquez, J.C., Matas, J., De Vicuña,
L.G., and Castilla, M. (2011). Hierarchical control of
droop-controlled AC and DC microgrids - A general
approach toward standardization. IEEE Transactions
on Industrial Electronics, 58(1), 158–172.
Heymann, M., Lin, F., Meyer, G., and Resmerita, S.
(2002). Analysis of zeno behaviors in hybrid sys-
tems. Proceedings of the IEEE Conference on De-
cision and Control, 3(December), 2379–2384. doi:
10.1109/cdc.2002.1184191.