10th Conference on Manoeuvring and Control of Marine Craft

10th Conference on Manoeuvring and Control of Marine Craft

August
10th
10th 24-26, 2015.
Conference
Conference on Copenhagen, and
on Manoeuvring
Manoeuvring Denmark
Control
Control of Marine Craft
August 24-26, 2015. Copenhagen, and
Denmark
Availableofonline
MarineatCraft
www.sciencedirect.com
August 24-26,
August 24-26, 2015.
2015. Copenhagen,
Copenhagen, Denmark
Denmark

ScienceDirect
IFAC-PapersOnLine 48-16 (2015) 273278
Modeling
Modeling of
of a
a Large
Large Marine
Marine Two-Stroke
Two-Stroke
Modeling
Diesel of
Engine a Large
with Marine
Cylinder Two-Stroke
Bypass Valve
Diesel Engine
Diesel Engine with
with Cylinder
Cylinder Bypass
Bypass Valve
Valve
and
and EGR
EGR System
System
and EGR System
Guillem Alegret Xavier Llamas
MortenGuillem Alegret Xavier
Vejlgaard-Laursen Llamas
Guillem
Guillem Alegret
Alegret Xavier Lars
Xavier Eriksson
Llamas
Llamas
Morten Vejlgaard-Laursen
Morten Vejlgaard-Laursen
Vejlgaard-Laursen Lars

Lars Eriksson
Lars Eriksson
Eriksson
Morten

MAN Diesel & Turbo, Copenhagen, Denmark
MAN Diesel & Turbo, Copenhagen, Denmark
Vehicular
MAN
MANSystems,
Diesel Dept.
Diesel & of Electrical
& Turbo,
Turbo, Engineering
Copenhagen,
Copenhagen, Denmark
Denmark Link
oping


Vehicular Systems,
University,
Vehicular Dept.
Sweden,
Systems, of Electrical Engineering
Engineering Link
Dept.xavier.llamas.comellas@liu.se
of Electrical
Electrical Engineering o
oping
Link ping
Vehicular Systems,
University, Dept.
Sweden, of Link
xavier.llamas.comellas@liu.se o ping
University,
University, Sweden,
Sweden, xavier.llamas.comellas@liu.se
xavier.llamas.comellas@liu.se
Abstract: A nonlinear mean value engine model (MVEM) of a two-stroke turbocharged marine
diesel engineA
Abstract:
Abstract:
Abstract: A isnonlinear
developed,
nonlinear
A is
nonlinear
mean
mean
mean
value
parameterized
value engineand
value engine
engineand
model
model
model
(MVEM)
validated
(MVEM)
(MVEM)
of
against
of a two-stroke
measurement
of aa two-stroke
two-stroke
turbocharged
data. The marine
turbocharged
turbocharged goal is
marine
marine
diesel
to
dieselhaveengine
a
engine is developed,
computationally
developed, parameterized
fast and
parameterized accurate
and validated
engine
validated modelagainst
against that measurement
captures
measurement the data.
data. mainThe dynamics
The goal
goal is
is
diesel
to haveengine
a is developed,
computationally parameterized
fast and and
accurate validated
engine modelagainst that measurement
captures the data. main The
dynamicsgoal is
and
to
to have can
have abe used in
abecomputationallythe
computationally development
fast and
fast and of of control
accurate
accurate systems
engine
engine model
modelfor thethat
that newly
captures
capturesintroduced
the
the main EGR
main system.
dynamics
dynamics
and
The can
and can
tuning be used in
procedure the development
used is explained, control systems
and thesystems
result isfor for the
a six-state newly introduced
MVEM with EGR EGR system.
seven system.control
and
The can
tuning be used
used in
in the
procedure theuseddevelopment
development
is explained,
of
of control
control
and thesystems
result isfor the
the newly
aa six-state newlyMVEM introduced
introduced with EGR
seven system.
control
inputs
The
The that
tuning
tuning capture
procedure
procedure the main
used
used is system
explained,
is explained, dynamics.
and the result is six-state MVEM
and the result is a six-state MVEM with seven control with seven control
inputs
inputs that
that capture
capture the the main system dynamics.
inputs
2015,that IFACcapture the main
(International main system
systemofdynamics.
Federation dynamics.
Automatic Control) Hosting by Elsevier Ltd. All rights reserved.
Keywords: Engine modeling, diesel engines, parametrization, validation, nonlinear systems
Keywords:
Keywords: Engine Engine modeling,
modeling, dieseldiesel engines,
engines, parametrization,
parametrization, validation, validation, nonlinear
nonlinear systems systems
Keywords: Engine modeling, diesel engines, parametrization, validation, nonlinear systems
1. INTRODUCTION Exhaust
1.
1. INTRODUCTION
INTRODUCTION Exhaust
Manifold
Turbine
Turbine p
1. INTRODUCTION Tscav
Exhaust
Exhaust
Manifold Turbine
Turbine t,out
pt,out
Manifold
The upcoming Tier III regulation (International Maritime Tscav
scav
Tscav Manifold t pt,out
pt,out
The
The upcoming
upcoming Tier
Tier III
III regulation
regulation (International
(International Maritime
Maritime T scav p exh X O,exh tt t,out
Organization,
The upcoming2013) 2013) is the
Tier III next milestone
regulation for EGRMaritime
(International technol- p exh XO,exh
pexh tt

Organization,
Organization, 2013) is is the
is the
the next milestone
next milestone
milestone for EGR
forsystem technol-
EGR technol-
technol- p exh X O,exh
exh XO,exh
O,exh
ogy in large two-stroke
Organization, 2013) engines.
next The EGR for EGR is used blow cyl
ogy
ogy in
in large
large two-stroke
two-stroke engines.
engines. The
The EGR
EGR system
system is
is used
used CBV
to
ogy reduce
in large N O emissions
two-stroke
x by recirculating a fraction
engines. The EGR system is used of the Blower blow
blow
blow
cyl
cyl
Engine
inj CBV
to reduce
to reduce
reducegas N
NO O x
Ointo emissions
emissions by
by recirculating
recirculating a
a fraction
fraction of
of the
the
Blower blow cyl Engine
Engine
t inj
CBV
CBV
CBV
exhaust
to N x the
emissions scavenging
by manifold.
recirculating a This
fractionresults
of in
the Blower
Blower
cyl
Engine
inj
inj
tinj
inj
exhaust
exhaust gas
gas into
x
into the
the scavenging
scavenging manifold.
manifold. This
This results
results in
in t inj
inj
EVC u tc
a lower combustion
exhaust gas into the peak temperature
scavenging manifold. andThis consequently
results in leak tinjEVC

inj cbv
ucbv tc
a lower
lower combustion
combustion peak
peak temperature
temperature and
and consequently
consequently
leak
leak blow fuel

EVC
eng
EVC u cbv
tc
tc
a reduction
lower in N Ox formation.
combustion peak Due to the
temperature and high financial
consequently leak
leak blow del EVC
eng
eng ucbv
cbv tc
a
a reduction
reduction in
in N
N O
O x formation.
formation. Due
Due to
to the
the high
high financial
financial
blow
blow del fuel
fuel eng
eng
costs of performing tests on a real
a reduction in N Ox formation. Due to the high financial
x engine, a reliable and blow del
del
del fuel
fuel
costs of ucov
fast
costs of performing
costsdynamic
of performing
performingenginetests
tests
model
tests
on
on isa
on a an
a
real
real
engine,
engine, a
realimportant
engine, a
reliable
a tool for and
reliable
reliable and
the
and ucov cool c
fast dynamic engine model is an important tool for the u cov
ucov
cov p scav XO,scav cool p cc
fast dynamic
development ofengine
new EGR model is
controlan important
systems. tool for the egr pscav cool pc,out
fast dynamic engine model is an important tool for the X cool c,out c pc,in
pc,out c
pscav
scav O,scav cool pc,out
development
development of
of new
new EGR
EGR control
control systems.
systems.

egr
egr
egr p XO,scav Tscav
O,scav
scav XO,scav
c,out
pc,in
T
p c,in
development of new EGR control
A lot of research can be found in literature about Meansystems. egr
Scavenging Tscav
T scav pc,in
T c,in
c,in
scav
Tscav Compressor Tc,in
A
A lot
lot of
of research
research can
can be
be found
found in
in literature
literature about
about Mean
Mean Scavenging
manifold
Scavenging Compressor c,inT c,in
Value
A lot Engine
of research Models
can (MVEM)
be found inwith EGR about
literature systems Meanfor Scavenging
manifold Compressor
Value
Value Engine
Engine Models
Models (MVEM)
(MVEM) with
with EGR
EGR systems
systems for
for manifold
manifold
Compressor
automotive
Value Engine engines,
Models e.g., Wahlstr
(MVEM) o m
with andEGR Eriksson
systems (2011)
for Fig. 1. Structure of system with state variables (blue) and
automotive engines, e.g., Wahlstr o m and Eriksson (2011)
automotive
and
and
Nieuwstadt
automotive
Nieuwstadt
engines,
engines,et al.e.g.,
et al.
Wahlstr
(2000).
e.g., Wahlstr
(2000).
o
om
However,
However,m and
andmuch
much
Eriksson
Eriksson
less (2011) Fig.
(2011)
less research
research
1.
Fig. control
Fig. 1. Structure
1. Structure
Structure inputs of system
of (red)
of system
system with
with state
with state
variables (blue)
state variables
variables (blue) (blue) and
and
and
and been
has
and Nieuwstadt
done in
Nieuwstadt et the
et al. (2000).
al. (2000).
same area However, much
with much
However, large lessless
marineresearch
two-
research control
control inputs
inputs (red)
(red)
has been done in the same area with large marine two- mass control
fraction, inputs
p (red)
and X , compressor outlet pres-
has
stroke
has been
been done
diesel in
doneengines. the
in the same same area
A fewarea with
examples large
with large marine
are marine
Blanke two-two-
and mass fraction, pscav and XO,scav , compressor outlet pres-
stroke diesel engines. A few examples are Blanke and mass fraction, pscav and XO,scav
stroke
stroke diesel
Anderson diesel engines.
(1985), A
A few
Theotokatos
engines. few examples
(2010) where
examples are
are Blanke
an MVEM
Blanke and
and masssure, fraction,
sure,
pc,out , exhaust
p p scav and
scav manifold
X O,scav ,, compressor
compressor
pressure and oxygen outlet mass
outlet pres-
pres-
c,out ,, exhaust manifold pressure and oxygen mass
O,scav
Anderson
Anderson
of a marine (1985),
(1985),
engine Theotokatos
Theotokatos
was (2010)
(2010)
developed, where
where
and Hansenan
an MVEM
MVEM
et al. sure,
fraction,
sure, p
p p
c,out , exhaust
and
exhaust X manifold
manifold and pressure
turbocharger
pressure and
and oxygen
speed,
oxygen mass
tc .
mass
Anderson
of (1985), Theotokatos (2010)and where an MVEM fraction, p
c,out exh
exh and X O,exh
O,exh and turbocharger speed, tc ..
of a
of a marine
(2013)
a marine
marine
engine
engine
where aengine similarwas
was
wasmodel
developed,
developed, and Hansen
of the engine
developed, and Hansen
used here
Hansen
et
et
al.
al. The
et was
al. fraction,
control
fraction,
The control
p
p exh
exh
and
inputs
and
inputs
X
are
X
are
fuel mass
O,exh
fuel
O,exh
and
and
mass
turbocharger
flow, m
turbocharger
flow, m
f uel , EGR
,
speed,
speed,
EGR

blower
tc
blower tc .
(2013)
(2013) where
where a
a similar
similar model
model of
of the
the engine
engine used
used here
here was
was The
speed, control
blow ,inputs are
fuel injection fuel mass
time, flow, tinj , m f uel
ff uel
fuel ,, EGR
injection blower
angle
proposed.
(2013) where a similar model of the engine used here was The
speed, control
inputs
,, fuel are fuel
injection mass
time, flow, m
ttinj ,, fuel uel EGR
injection blower
angle
proposed. speed,
blow fuel injection time, fuel injection angle
proposed.
proposed. inj , exhaust
speed, blow , fuel
blow valve closingtime,
injection angle, inj
tinj , EV
fuel C , injection
cut-out anglevalve
In this study the proposed MVEM is based on the inj ,, exhaust
(COV) exhaust
position,
valve
valve closing
closing
ucovclosing
, and CBV
angle,
angle,
EV C ,, cut-out
position, cut-out
C , ucut-out
cbv . Figure
valve
valve
In
In this
this study
study the
the proposed
proposed MVEM
MVEM is
is based
based on
on the
the inj , exhaust
(COV)
inj position, valve
u , and CBV angle, EV C
position,
EV u . valve1
Figure 1
4T50ME-X
In this studytest theengine from MAN
proposed MVEM Dieselis&basedTurbo,onwhich the (COV)
gives an position,
overview uofcov ,
the andmodel.CBV The position,
engine u cbv
cbv . Figure 1
model . Figure
consists
4T50ME-X
4T50ME-X test
test engine
engine from
from MAN
MAN Diesel
Diesel &
& Turbo,
Turbo, which
which (COV) position, u cov , and CBV position, ucbv
cov 1
is a
4T50ME-X turbocharged
test engine two-stroke
from MAN diesel engine
Diesel with
& Turbo, direct
which gives
gives
of an
an
several overview
overview
interconnected of
of the
the model.
model.
submodels The
The engine
engine
which are model
model consists
consists
introduced
is
is aa turbocharged
a turbocharged
turbocharged two-stroke
two-stroke and diesel
diesel engine
enginevalve with direct
withtiming.
direct in gives
of an
several overview
interconnected of the model.
submodels The engine
which model
are consists
introduced
injection,
is uniflow scavenging
two-stroke dieselvariable
engine with direct of several
several
the following interconnected
subsections. submodels which which are are introduced
introduced
injection, uniflow scavenging and
and variable valve timing. of interconnected submodels
injection,
It can provide
injection, uniflow
uniflow scavengingrated
a maximum
scavenging and variable
power ofvalve
variable valve kW at in
timing.
7080 timing. in the
the following
following subsections.
subsections.
It
It can provide a maximum rated power of 7080 kW at in the following subsections.
123can
It RPprovide
can M . It isa
provide maximum
a equipped
maximumwith rated
ratedanpower EGR of
power 7080
7080 kW
ofsystem and at
kW a 2.1 Turbocharger
at
123
123 RP M .. It is equipped with an EGR system and a
123 RP M . It is equipped with an EGR system and aa 2.1
RP
Cylinder M It
Bypass is equipped
Valve (CBV).with an
The EGR
purpose system
of the and
valve 2.1 Turbocharger
Turbocharger
Cylinder
Cylinder
is Bypass
Bypass
to keep Bypass
Cylinder the desired Valve
Valve
Valve
(CBV).
(CBV). The
turbocharger
(CBV).
The
Thespeedpurpose
purpose
whenof
purpose
of the
ofthe valve 2.1 Turbocharger
valve
theengine
the valve
is
is to
to keep the desired turbocharger speed when the engine The turbocharger model includes submodels for the com-
is to keep
operates the
the desired
keepunder high EGR
desired turbocharger
rates. Inspeed
turbocharger thosewhen
speed when the
the engine
situations less The
engine The turbocharger model includes submodels shaft.for the com-
operates
operates
energy
under
is under
transferred
high
high EGR
EGR
through
rates.
rates. In
In
the turbine,
those
those situations
situations
thus
less
part of less The turbocharger
pressor,
the pressor,
the turbine
turbocharger
the turbine
model
and the
model
and
includes
the
connecting
includes
connecting
submodels
submodels shaft.
for
for the
the com-
com-
operates
energy is under
transferred high EGR
through rates.
the In those
turbine, situations
thus part of less
the pressor, the the turbine
turbine and and thethe connecting
connecting shaft. shaft.
energy
compressor is transferred
air mass through
flow is the
bypassed turbine,
to thus
boost part
the of the
turbine. pressor,
Compressor
energy
compressor is transferred through the turbine, thus part of the
compressor air
compressor air mass
air mass flow
mass flow is
flow is bypassed
is bypassed to
bypassed to boost
to boost the
boost the turbine.
the turbine. Compressor
turbine. Compressor
The mass flow and efficiency models of the compressor are
Compressor
2. MODELING The
The massmass flow and
and efficiency
efficiency models
flowparameterization models of
of the
the compressor
compressor are
are
based
The on the
mass flow and efficiency modelsof the of performance
the compressor mapsare in
2.
2. MODELING based on the parameterization of the performance maps in
2. MODELING
MODELING based
SAE format.
based
SAE
on
on
format.
the
the parameterization
The turbochargerof
parameterization
The turbocharger of the
speed
the
speed
performance
and the compressor
performance
and the
maps
maps
compressor
in
in
The MVEM consists of six states and seven control inputs. mass SAE
SAE format.
flow
format. in The
the
The turbocharger
performance
turbocharger map speed
speed are and
corrected
and the
the compressor
in order
compressor to
The
The MVEM
MVEM consists
consists of
of six
six states
states and seven
seven control
and pressure control inputs. mass
inputs. massinto flow in
in the
flowaccount the performance
performance map
map are corrected
corrected in
areconditions. inTheorder
order to
to
The states are
MVEM scavenging
consists of six manifold
states and seven and oxygen
control inputs. take
mass flow in the changes in ambient
performance map are corrected in com-
order to
The states
The states
states are are scavenging
are scavenging
scavenging manifoldmanifold pressure
manifold pressure
pressure and and oxygen
and oxygen
oxygen take take into
take into account
into account
account changes changes
changes in in ambient
in ambient conditions.
ambient conditions.
conditions. The The com-
The com-
com-
The
2405-8963
Copyright 2015, IFAC (International Federation of Automatic Control)
IFAC 2015 273 Hosting by Elsevier Ltd. All rights reserved.
Copyright
Peer review IFAC responsibility
2015 273Control.
Copyright
Copyright under
IFAC 2015
IFAC 2015 of International Federation of Automatic
273
273
10.1016/j.ifacol.2015.10.292
IFAC MCMC 2015
274 Guillem Alegret et al. / IFAC-PapersOnLine 48-16 (2015) 273278
August 24-26, 2015. Copenhagen, Denmark

pressor mass flow is modeled using super-ellipses centred The power generated by the turbine and the power con-
at the origin. A similar approach is found in Leufven and sumed by the compressor are defined as in Dixon (1998)
Eriksson (2013). The explicit expression of a super-ellipse  e 1

is P t = t m t cp,e Tt,in 1 (t ) e (6)
  n  n1  
c m c cp,a Tc,in a 1
m c,corr = a 1 (1) Pc = (c ) a 1 (7)
b c
where c is the pressure ratio over the compressor,
pc,out /pc,in . The variables a, b and n are described by third 2.2 Control Volumes
order polynomials of the corrected turbocharger speed, so
the model has 12 tuning parameters. The model consists of three control volumes. The com-
pressor outlet and the two manifolds, they are all modeled
The compressor efficiency is modeled by parameteriz- with standard isothermal models as proposed in Heywood
ing the manufacturer performance map with rotated and (1988) and Eriksson and Nielsen (2014).
translated ellipses. The implicit expression of an ellipse
rotated and translated from the origin to (a0 , b0 ) is as The pressure at the compressor outlet is described by
follows d Ra Tc,out
 2 pc,out = cm
(m cool m
cbv ) (8)
(x a0 ) cos (y b0 ) sin dt Vc,out
+
a where Vc,out is the control volume size, and it has to be
 2 (2)
(x a0 ) sin + (y b0 ) cos estimated and Tc,out is described in (14).
=1
b At the scavenging manifold, the temperature is assumed to
where in this case x corresponds to m c and y corresponds be constant since the cooler is considered to be ideal and
to c . The coefficients a0 , b0 , a, b, and are described using capable of maintaining a constant scavenging temperature.
second order polynomials of c so the model consists of Two states are needed to fully characterize the manifold,
15 parameters to estimate. the pressure and the oxygen mass fraction. The pressure
is governed by the following differential equation
Turbine
The turbine corrected mass flow is described as in Eriksson d Ra Tscav
pscav = (m egr m
cool + m del ) (9)
and Nielsen (2014) dt Vscav
 where Vscav is the volume of the manifold and has to be
m t,corr = Ct 1 kt t (3) estimated. The oxygen mass fraction is described as in
where turb is the pressure ratio over the turbine, Wahlstrom and Eriksson (2011)
pt,out /pexh . Moreover, kt and Ct are parameters to be d Ra Tscav
estimated. XO,scav = (XO,exh XO,scav ) m egr +
dt pscav Vscav
(10)
The turbine efficiency is commonly modeled using the Ra Tscav
(XO,a XO,scav ) m cool
Blade Speed Ratio (BSR), e.g. Wahlstr om and Eriksson pscav Vscav
(2011) and Eriksson and Nielsen (2014) where XO,a is the mass fraction of oxygen in dry air.
Rt t As in the previous manifold, two states characterize the ex-
BSR =    (4)
1 1e haust manifold, the pressure and the oxygen mass fraction.
2 cp,e Tt,in 1 t The exhaust pressure is driven by the following differential
equation
where Rt is the turbine blade radius. The turbine efficiency d Re Texh
is again modeled with rotated and translated ellipses pexh = (m cyl m egr m exh,out ) (11)
using (2). In this case x corresponds to the BSR and dt Vexh
y corresponds to the t . The coefficients a0 , b0 , a, b, with
and are described as second order polynomials of the m
exh,out = m tm cbv (12)
corrected turbocharger speed, thus 15 parameters need to and where Vexh is the exhaust manifold volume and a
be determined. tuning parameter, and m cyl = m del + m f uel . The oxygen
mass fraction is defined in a similar manner as in the
Connecting Shaft scavenging manifold
The turbocharger shaft speed is described by Newtons
second law using the power recovered from the exhaust d Re Texh
XO,exh = (XO,cyl XO,exh ) m cyl (13)
gas by the turbine and transferred to the compressor dt pexh Vexh
d Pt Pc where XO,cyl is the oxygen mass fraction coming out
tc = (5) from the cylinders. Since the injected fuel combustion is
dt Jt tc
assumed to be ideal and complete, XO,cyl is calculated as
where the parameter Jt corresponds to the overall tur- equation (16) in Wahlstrom and Eriksson (2011).
bocharger inertia. Pt and Pc are the turbine and com-
pressor powers, respectively. Note that the mechanical 2.3 CBV
efficiency is not included in (5), it is already included in
the turbine efficiency of the SAE map. The CBV model consists of a submodel for the compressor
outlet temperature, a submodel for the flow through the
CBV valve and a submodel for the flow through the cooler.

274
 IFAC MCMC 2015
August 24-26, 2015. Copenhagen, Denmark
Guillem Alegret et al. / IFAC-PapersOnLine 48-16 (2015) 273278 275

The temperature at the compressor outlet is calculated flow through all cylinders. The same approach is found
using the definition of the adiabatic efficiency of the in Hansen et al. (2013) and Theotokatos (2010). The same
compressor from Dixon (1998) generic equation (15) is used, and in this case the m is
 a 1  the delivered mass flow m del through the cylinders, is
(c ) a 1 the pressure ratio over the cylinders pexh /pscav , i and Ri
Tc,out = Tc,in 1 + (14)
c correspond to the heat capacity ratio and the specific gas
constant of air. Aef f is the effective area of the restriction,
The mass flow through the CBV is modeled as a com- and has to be estimated.
pressible turbulent restriction. A generic formulation of
It is common to characterize the scavenging process in two-
the model is presented as follows
 stroke engines with the scavenging efficiency scav and the
Aef f pin 2 i  2 i +1 
trapping efficiency trap . Their definitions can be found in
m = i i (15)
Ri Tin i 1 Heywood (1988). The delivery ratio (DR) is defined as the
for this case, m
is the mass flow through the CBV, is the ratio between the delivered flow and the ideal flow at the
pressure ratio pexh /pc,out , i and Ri are the heat capacity scavenging manifold density
 
ratio and the specific gas constant of air, respectively. Aef f 2 m del Ra Tscav
corresponds to the CBV effective area Acbv , which in this DR = (20)
ncyl eng V1 pscav
case is variable depending on the control input ucbv , and
it is defined as follows The model proposed here is a combination of the two
limited ideal models introduced in Heywood (1988), the
Acbv = Amax (1 cos(ucbv )) (16)
2 perfect displacement and the complete mixing. The perfect
where Amax is a tuning parameter that corresponds to the displacement assumes that the burned gases are displaced
maximum area of the restriction. by the fresh gases without mixing, on the other hand,
the complete mixing model assumes instantaneous mixing
The mass flow through the cooler is described by an of the gases when fresh mixture enters the combustion
incompressible turbulent restriction, described in Eriksson chamber. By introducing the tuning parameters Kse1 and
and Nielsen (2014) Kse2 in the complete mixing model (21) and (22), an

pc,out (pc,out pscav ) intermediate formulation is obtained, with the purpose of
m cool = kcool (17) taking into account the late exhaust valve closing.
Tc,out
scav = 1 eKse1 DR (21)
where kcool is a parameter to be estimated.
1 eKse2 DR
In situations where the CBV is open, the turbine inlet trap = (22)
DR
temperature cannot be assumed to be equal to the exhaust
temperature. To consider the temperature drop caused Limited pressure diesel cycle
by the CBV flow, the perfect mixing model described in As an overview, six changes to the cycle presented in
Eriksson and Nielsen (2014) is used Wahlstrom and Eriksson (2011) have been incorporated.
Texh cp,e m exh,out + Tc,out cp,a m
cbv (i) The constant volume burned ratio xcv is considered
Tt,in = (18)
cp,e m exh,out + cp,a m
cbv variable. The maximum pressure rise in the cylinders is
With this formulation, an algebraic loop is encountered regulated by the control system as a safety measure. The
between the Tt,in and the m t calculations. In order to regulation is accomplished by delaying the injection. To
break the algebraic loop, it is assumed that m exh,out be able to model late injection, the xcv is considered a
in (18) can be approximated by its steady state value linear function of the start crank angle and duration of
m exh,out = m cyl m egr . the injection. The model is shown in (23). A similar model
for xcv is shown in Lee et al. (2010).
The exhaust oxygen measurement equipment is installed xcv = c1 + c2 inj + c3 tinj (23)
downstream of the turbine. When the CBV is open, it
affects the measurement. Therefore, a new oxygen mass where the three parameters ci have to be estimated.
fraction is calculated in (19) for validation purposes. (ii) The compression process is considered to start when
the exhaust valve closes. In that instant the crank angle
XO,exh m
exh,out + XO,a m
cbv is given by EV C . The volume of the combustion chamber
XO,t = (19) based on the crank angle is used in the limited pressure
mt
In this expression, the m
exh,out used is described by cycle calculations, and it is defined as equation (4.3) from
equation (12). Eriksson and Nielsen (2014). Also, the expansion process
is assumed to last until the bottom dead center.
2.4 Cylinders (iii) Both the compression and the expansion processes
are considered polytropic Jiang et al. (2009) in order
The mass flow through four-stroke engines is commonly to consider heat exchange with the cylinder walls, both
modeled with the volumetric efficiency as in Wahlstrom polytropic exponents of the compression and expansion
and Eriksson (2011) and Heywood (1988). For two-stroke are tuning parameters.
engines, the mass flow through all cylinders can be approx- (iv) The delivered mass flow is assumed to be heated by
imated with the flow through a compressible turbulent a tuning factor dTcyl before the cycle starts. The heating
restriction. The continuous flow represents the average

275
IFAC MCMC 2015
276 Guillem Alegret et al. / IFAC-PapersOnLine 48-16 (2015) 273278
August 24-26, 2015. Copenhagen, Denmark

affects both the trapped and the short-circuited flows (25). The existence of a leak in the recirculation valve is known,
The pressure of the trapped gas when the combustion however, its magnitude is unknown. The leak mass flow
chamber is sealed is assumed to be the scavenging pressure, m leak is modeled as a compressible turbulent restriction,
while the temperature is described by like in (15). But in this case, the is the pressure ratio over
T1 = Tcyl (1 scav ) + scav (Tscav + dTcyl ) (24) the recirculation valve pexh /pscav , i , and Ri corresponds
to the heat capacity ratio and the specific gas constant
The algebraic loop between the initial cycle temperature, of exhaust gas respectively. Aef f corresponds to the leak
T1 and the cylinder out temperature, Tcyl , is solved using effective area. The resulting mass flow through the EGR
the previous sample value for Tcyl similar to what is done system is
in Wahlstrom and Eriksson (2011).
egr = (m
m blow m
leak ) f (ucov ) (29)
(v) The cv,a before the constant volume combustion starts where f (ucov ) describes the valve dynamics as
and the cp,a at the beginning of the constant pressure com-  1

bustion are calculated based on the temperatures at the f (ucov ) = 1 e cov ucov (30)
respective crank angles. To perform such calculation, the
NASA polynomials are used to describe these parameters ucov only regulates the flow during start-up of the system,
in terms of temperature. The same polynomials found in then the flow is controlled using the blower speed.
Goodwin et al. (2014) are used here.
3. EXPERIMENTAL DATA AND TUNING
(vi) To determine the exhaust temperature Texh , charac-
PROCEDURE
terized by the mixture of the short-circuited flow and the
trapped flow in the cylinder at their respective tempera-
tures, the perfect mixing model is used again in the same The parameters in the submodels are estimated using
manner as (18). The short-circuited flow is defined as engine measurements. The measured signals: pc,in , pt,out ,
Tc,in and eng are used in the estimation and in the
m del m
sh = m trap (25) validation of the model in the same manner as if they were
inputs to the model.
Using the pressures and the volumes of each process in the
thermodynamic cycle, the indicated power of the cycle is Unfortunately, the oxygen sensors were not properly cali-
computed using equations (2.14) and (2.15) in Heywood brated before the measurements. Thus the stationary val-
(1988). To sum up, the limited pressure cycle has eight ues cannot be trusted and will not be used for estimation
parameters to determine. purposes. More information (types, starts and stops and
number of steps) about each of the dynamic datasets can
be found in the top part of Table 2.
2.5 EGR loop
The following relative error is used to quantify the differ-
The EGR loop model consists of a blower to overcome ence between the modeled signals, ymod , and the measured
the pressure difference between exhaust and scavenge signals, ymeas
manifolds, a recirculation valve and a cut-out valve (COV) ymod [k] ymeas [k]
to manage the start-up of the EGR system. The flow is erel [k] = N (31)
considered ideally cooled to scavenging temperature. 1/N j=1 ymeas [j]
the euclidean norm of this relative error is used as the
EGR Blower
objective function to minimize in the tuning procedure.
The performance map is expressed in a non-dimensional
space described by the Head Coefficient () and the Flow
Coefficient (), their definitions are shown in (26) and (28) 3.1 Submodels initialization
respectively
 1  Some of the submodels are initialized using the maps

2 Tscav cp,e blow 1 provided by the component manufacturer. Table 1 presents
= (26) the stationary errors of the submodels that are initialized.
2
(blow Rblow )
where blow is the blower angular speed, Rblow is the Table 1. Relative errors of the initialized submodels
blower blade radius and blow is the pressure ratio over
the blower pscav /pexh . Model mc c mt t m
blow
Mean rel. error [%] 3.11 0.69 0.19 0.31 0.47
The non-dimensional performance map is modeled with Max rel. error [%] 14.5 3.24 0.45 1.03 0.87
the same approach as the compressor mass flow, but
here only one speed line is parameterized. Therefore, the
parameters a, b, and n are constants and need to be To get an initial guess and to avoid overparametrization in
estimated. the pressure limited cycle submodels, e.g. (23), a few extra
  n  n1
stationary measurements are used, which are not used
=a 1 (27) later in the model simulation, e.g., the maximum cylinder
b
pressure and ordered cylinder compression pressure. This
Rearranging the definition of , the mass flow through the initialization is based on a least-squares optimization with
blower m
blow is obtained Texh and the indicated power of the cycle as objectives.
pexh  3
 Since some of the submodel inputs are not measured, e.g.,
m blow = blow Rblow (28) m del , those submodels have to be used in this initialization.
Re Tscav

276
 IFAC MCMC 2015
August 24-26, 2015. Copenhagen, Denmark
Guillem Alegret et al. / IFAC-PapersOnLine 48-16 (2015) 273278 277

3.2 Overall stationary estimation

5 5
x 10 x 10
4
4

pscav [Pa]
3.5

pexh [Pa]
3.5

Since there are no mass flow measurements apart from 3

3

Meas
m egr , some submodels cannot be properly initialized, e.g., 2.5
Mod
2.5

the m eng or the m

cbv . Therefore all the parameters have 2
0 500 1000 1500 2000 2500 3000
2
0x 105 500 1000 1500 2000 2500 3000

to be estimated together, since the optimization problem 4

tc [rad/s]
1800

pc,out [Pa]
cannot be separated. Another reason for estimating all pa- 3.5

1600
rameters at the same time is that it is difficult to attain the 2.5
3

same stationary levels for the modeled and the measured 1400
2
0 500 1000 1500 2000 2500 3000 0 500 1000 1500 2000 2500 3000
signals by fixing the previously estimated parameters. The 1
1

overall estimation is performed with 27 different stationary 0.8

X O,scav [-]
0.8

X O,t [-]
0.6
points extracted from the estimation datasets. A point is 0.6

0.4 0.4

considered stationary when the pressure and temperature 0.2 0.2

signals are stabilized. 0

0 500 1000 1500 2000 2500 3000
0
0 500 1000 1500 2000 2500 3000
6000 800

The measured states are used as inputs in the optimization

P eng,i [kW]
5500
750

Texh [K ]
5000
since they cannot be integrated for isolated stationary 4500
700

points. To ensure that the model outputs are stationary at 4000

650

the stationary points, the derivative terms of (5), (8), (9) 3500
0 500 1000 1500 2000 2500 3000
600
0 500 1000 1500 2000 2500 3000
12
and (11) are added into the objective function weighted by 1

eng [rad/s]
COV

Valves [-]
11.5 0.8
the mean of the measured state to provide fair comparison. 0.6 CBV
11
The objective function is defined as 10.5
0.4

0.2
M 
 N i 2
1 (x [n]) 10
0 500 1000 1500 2000 2500 3000
0
0 500 1000 1500 2000 2500 3000
Vstat () = N (32) 1000 5

NM i=1 n=1 1/N

i
j=1 xmeas [j]
blow [rad/s]

egr [kg/s]
800 4

600 3
S N
1  i 400 2

(e [n])2

m
+ 200 1

N S i=1 n=1 rel 0

0 500 1000 1500 2000 2500 3000
0
0 500 1000 1500 2000 2500 3000
time [s] time [s]

where the first row minimizes the residuals of the dynamic
models. With x1 = pscav , x2 = pexh , x3 = pc,out and
x4 = t , the second row minimizes the relative error of the
EGR mass flow, the exhaust temperature and the engine
indicated power. N is the number of stationary points
available. The vector represents the parameters to be
estimated, which in this case are all the static parameters,
except for the compressor parameters and the turbine
efficiency parameters. This selection has proven to be a
good trade-off between objective function complexity and
model accuracy.
3.3 Dynamic estimation
Keeping the static parameters already estimated fixed, the
next step is to tune the parameters of the dynamic models
(5), (8), (9), (11) and (30). From the available datasets, 13
step responses were extracted and used in the estimation.
These steps consist of EGR blower speed steps, fuel flow
steps and CBV steps. In the same manner as it is done
in Wahlstr om and Eriksson (2011), the measurements and
the model outputs are normalized so the stationary errors
have no effect on this estimation. The objective function
used is
J  D
1 
Lz
Vdyn () = (ximeas,n [l] ximod,n [l])2 (33)
i=1 z=1
L z
l=1
where xi are the control volume pressures and the tur-
bocharger speed, J, is the number of states (excluding
oxygen mass fractions), D is the number of steps used,
and Lz is the length of each step. The parameter vector is
thus = [Jt , Vscav , Vexh , Vc,out , cov ]T .
Fig. 2. Model simulation vs measurements of dataset 11.
4. MODEL VALIDATION
Table 2 presents the mean relative errors in percentage for
all dynamic datasets. The 27 extracted stationary points
are used to compute the mean required in the denominator
of (31). This is done to provide a fair comparison between
them, so all errors for different datasets are weighted with
the same mean value. Excluding the EGR mass flow, the
model errors are below 6.28% and in general below 3%. A
higher error is observed for the EGR mass flow, where the
mean for all datasets is 7.34%.
Figure 2 shows the states of the model compared to the
measurements for dataset 11. Since the oxygen measure-
ments are not calibrated, the modeled and the measured
signals are normalized to compare only the dynamic be-
havior. This dataset has a load step, several EGR blower
speed steps and a start and stop of the EGR system which
is coupled to the CBV operation. It can be observed that
the model captures the dynamics of the system.
5. CONCLUSION
An MVEM for a large marine two-stroke engine is pro-
posed and validated. The estimation is done with part of
the datasets available while the validation against mea-
surements is done for another set of datasets. The overall
agreement of the states is good, and the model is able to
capture the general state dynamics.
Nevertheless this model is the first step towards a more
general model to be used for development of control
strategies. The next step is low load modeling, where new
components need to be introduced.

277
IFAC MCMC 2015
278 Guillem Alegret et al. / IFAC-PapersOnLine 48-16 (2015) 273278
August 24-26, 2015. Copenhagen, Denmark

Table 2. Top: number and type of steps contained in each dataset. Bottom: mean relative errors
in % of the tuned model for the absolute measured signals in the estimation and validation
datasets.
Estimation Datasets Validation Datasets
DS1 DS2 DS3 DS4 DS5 DS6 DS7 DS8 DS9 DS10 DS11 DS12 DS13 DS14 DS15
m f uel steps 1 1 5 1 1 0 1 4 3 6 1 0 0 0 7
blow steps 9 1 f ixed 5 7 1 8 9 11 f ixed 9 9 5 7 f ixed
uCBV steps 0 0 0 0 0 0 2 1 3 2 3 4 3 4 5
EGR start/stop 1 1 0 0 0 1 3 0 0 0 3 4 3 5 0
CBV start/stop 0 0 0 0 0 0 2 1 3 0 3 4 3 4 0
pscav 1.34 3.03 1.71 2.25 2.71 1.27 2.36 2.32 1.92 2.47 2.67 3.12 3.22 6.23 2.94
pc,out 1.32 2.94 1.71 2.24 2.70 1.24 2.31 2.39 1.96 2.51 2.49 2.88 2.98 6.28 2.96
pexh 1.72 3.07 2.36 1.80 3.10 1.60 2.16 2.78 2.36 3.11 2.18 2.52 2.69 6.13 3.68
tc 0.92 4.21 0.91 0.96 2.03 0.83 1.44 1.13 1.04 1.27 1.14 1.23 1.23 4.61 1.96
Texh 1.36 4.22 1.39 1.48 1.10 0.77 2.66 1.39 1.96 1.91 1.93 1.60 1.61 2.37 2.83
Peng,i 1.57 1.41 1.76 2.11 2.52 2.70 1.66 1.44 2.01 1.73 1.85 1.95 2.24 1.58 2.29
m egr 7.23 8.35 10.15 6.54 4.77 6.07 5.62 6.40 8.88 7.82 6.58 7.87 9.34 7.31 7.14

REFERENCES Wahlstrom, J. and Eriksson, L. (2011). Modelling diesel

Blanke, M. and Anderson, J.A. (1985). On modelling large engines with a variable-geometry turbocharger and ex-
two stroke diesel engines: new results from identification. haust gas recirculation by optimization of model param-
IFAC Proceedings Series, 20152020. eters for capturing non-linear system dynamics. Proceed-
Dixon, S. (1998). Fluid Mechanics and Thermodynamics ings of the Institution of Mechanical Engineers, Part D:
of Turbomachinery. Butterworth-Heinemann. Journal of Automobile Engineering, 225, 960986.
Eriksson, L. and Nielsen, L. (2014). Modeling and Control
of Engines and Drivelines. John Wiley & Sons. Appendix A. NOMENCLATURE
Goodwin, D.G., Moffat, H.K., and Speth, R.L. (2014).
Cantera: An object-oriented software toolkit for chemi- Table A.1. List of symbols
cal kinetics, thermodynamics, and transport processes. A Area [m2 ]
http://www.cantera.org. Version 2.1.2. B Bore [m]
Hansen, J.M., Zander, C.G., Pedersen, N., Blanke, M., and c Connecting rod length [m]
Vejlgaard-Laursen, M. (2013). Modelling for control cp Specific heat at constant pressure [J/(kgK)]
of exhaust gas recirculation on large diesel engines. cv Specific heat at constant volume [J/(kgK)]
Proceedings of the 9th IFAC conference on Control J Inertia [kg m2 ]
Applications in Marine Systems. m Mass flow [kg/s]
Heywood, J.B. (1988). Internal Combustion Engine Fun- ncyl number of cylinders [ ]
damentals. McGraw-Hill. p Pressure [P a]
International Maritime Organization (2013). MARPOL: P Power [kW ]
Annex VI and NTC 2008 with Guidelines for Imple- R Gas constant [J/(kgK)]
mentation. IMO. s Stroke [m]
Jiang, L., Vanier, J., Yilmaz, H., and Stefanopoulou, A. T Temperature [K]
(2009). Parameterization and simulation for a tur- V Volume [m3 ]
bocharged spark ignition direct injection engine with XO Oxygen mass fraction [ ]
variable valve timing. SAE Technical Paper 2009-01- angle [rad]
0680. Specific heat capacity ratio [ ]
Lee, B., Jung, D., Kim, Y.W., and Nieuwstadt, M. (2010). Efficiency [ ]
Thermodynamics-based mean value model for diesel Pressure ratio [ ]
combustion. Engineering for Gas Turbines and Power, Flow Coefficient [ ]
135(9), 193206. Head Coefficient [ ]
Leufven, O. and Eriksson, L. (2013). A surge and choke Rotational speed [rad/s]
capable compressor flow model - validation and extrap-
olation capability. Control Engineering Practice, 21(12), Table A.2. Subscripts
18711883.
Nieuwstadt, M., Kolmanovsky, I., Moraal, P., Ste- a air inj injection
fanopoulou, A., and Jankovic, M. (2000). EGR-VGT blow blower meas measured
control schemes: experimental comparison for a high- c compressor mod modeled
speed diesel engine. IEEE Control Systems Mag., 20(3), cool cooler scav scavenging manifold
6379. cyl cylinder t turbine
del delivered trap trapped
Theotokatos, G. (2010). On the cycle mean value mod-
e exhaust gas x, corr corrected quantity
elling of a large two-stroke marine diesel engine. Proceed-
egr EGR gas x, in inlet of x
ings of the Institution of Mechanical Engineers, Part M:
eng engine x, out outlet of x
Journal of engineering for the maritime environment,
exh exhaust manifold x, n normalized x
224(3), 193206.

278