SPECIAL SECTION ON ADVANCED MODELING AND CONTROL OF COMPLEX MECHATRONIC

SYSTEMS WITH NONLINEARITY AND UNCERTAINTY

Received April 2, 2018, accepted April 24, 2018, date of publication May 15, 2018, date of current version June 19, 2018.
Digital Object Identifier 10.1109/ACCESS.2018.2835395

Energy Management Strategy for Atkinson Cycle

Engine Based Parallel Hybrid Electric Vehicle
MUHAMMAD ASGHAR 1 , AAMER IQBAL BHATTI1 , (Senior Member, IEEE),
QADEER AHMED 2 , (Member, IEEE), AND GHULAM MURTAZA2
1 Department of Electrical Engineering, Capital University of Science and Technology, Islamabad 44000, Pakistan
2 Center for Automotive Research, The Ohio State University, Columbus, OH 43212, USA
Corresponding author: Muhammad Asghar ([email protected])

ABSTRACT Energy management strategies are widely used for the fuel efficiency of hybrid electric
vehicle (HEV). Most of the on road hybrid cars use Atkinson cycle engine instead of Otto cycle engine,
on account of better efficiency. The authors have already proposed a control oriented extended mean
value engine model (EMVEM) of the Atkinson cycle engine and also developed and evaluated its control
framework. The prior research was limited to evaluate the fuel efficiency of the engine only, while HEV
employs both the engine and the motor. The presented work is an attempt to devise and demonstrate
energy management strategy (EMS) by giving full consideration to the powertrain using Atkinson cycle
engine. A novel energy management strategy based on the vehicle speed for Atkinson cycle engine for
HEV is proposed. A parallel hybrid electric vehicle is modeled in Matlab/Simulink through a forward
facing simulator. The proposed EMS with Atkinson cycle engine control framework exhibits the significant
improvement in the fuel economy around 12.30% for standard Manhattan driving cycle and 7.22% for the
modified federal urban driving schedule (FUDS) driving cycle in comparison with the Otto cycle engine.

INDEX TERMS Energy management strategy, Atkinson cycle engine, fuel economy, drivability.

I. INTRODUCTION HEV powertrains is the designing of energy management

A. MOTIVATION control (EMC) methods to enhance fuel economy benefits.
Environmental challenges and reduction of global crude oil
reserves gained the attention of researchers and automo- B. LITERATURE SURVEY
bile manufacturers for exploration of novel vehicle tech- The control of power flows between ICE and the electric
nologies. Hybrid Electric Vehicles (HEVs) established a motor is referred as the energy management strategy. The
thought for minimizing the fuel consumption and green- drivetrain topology includes the power flow routes and the
house gases (GHG) emissions. Transportation sector con- transmission ratios between the two power sources (Engine
sumes about 66% of total oil consumption in the world and Motor) and the wheels of the vehicle shown in Fig. 1.
and 50% of that is utilized by small passenger cars and The energy management strategies for HEVs are catego-
trucks [1]. In this unique circumstance, automakers must rized keeping in view multiple criteria. On the basis of
lessen the oil consumption and GHG emissions by present- the amount of information and the optimization method,
ing advanced fuel-effective innovations and furthermore by the following three groups can be found in the literature.
utilizing alternative fuels. Fuel economy improvement and Global optimization strategies, local optimization strategies,
reduction in emissions are achieved through engine start- and heuristic methods. Global optimization methods seek
stop operation, brake energy recuperation [2] and advanced the optimal result, that is the optimal power split sequence.
technology engine such as Atkinson cycle engine [3]. In this This optimal power split sequence is guaranteed by execut-
manner, incorporating fuel-efficient Atkinson cycle engine in ing a global minimization over the complete driving cycle,
HEVs can enhance the vehicle efficiency and lessen the GHG supposed to be decided in advance. The Dynamic program-
outflows. This paper investigates the execution of a parallel ming (DP) technique ensured the globally optimal solution
HEV with a fuel-efficient innovation (e.g.Atkinson cycle and had been discussed in several publications [4]–[8]. This
engine). Another vital factor for improving the efficiency of technique is not practicable online, on account of a priori

2169-3536
2018 IEEE. Translations and content mining are permitted for academic research only.
28008 Personal use is also permitted, but republication/redistribution requires IEEE permission. VOLUME 6, 2018
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
M. Asghar et al.: EMS for Atkinson Cycle Engine Based Parallel HEV

characteristic of Atkinson cycle engine, the longer expansion

ratio than compression ratio can be realized.
The potential benefits, in relation to fuel economy
enhancement and decline in exhaust emissions of the
VVT incorporated Atkinson cycle engine over the usual Otto
cycle engine, come from the reality that the typical spark
ignition engines are designed to achieve optimal efficiency
by taking into account the full load. While, the automotive
engines operate at part loads most of the times and the load
is controlled with the conventional throttle that results in
more engine part load losses. Correspondingly, the thermal
FIGURE 1. Parallel architecture of the vehicle. efficiency of the conventional SI gasoline degrades consider-
ably at engine light loads, on account of fixed compression
ratios varying from 8 to 12 as well as limitations of fuel
information about the driving pattern and high computational quality and engine knocking. Moreover, the authors [30], [31]
demands. Local optimization strategies do not require the have derived and explored the thermal efficiencies of several
information of future driving pattern, rather depending upon engine cycles and have figured out that the optimum ther-
onboard data. The solution is provided through a sequence mal efficiency contributing thermodynamic cycle is the over-
of local (instantaneous) minima, estimated at every instant. expanded cycle (Atkinson cycle).
Equivalent consumption minimization strategy (ECMS) is So, to cater for the SI engine performance degrading
one of these methods [9]–[11]. This strategy is relying on behavior, hybrid electric vehicles are using the over-expanded
the transformation of battery power into fuel consumption Atkinson cycle engine appreciated through VVT and incor-
(equivalent). Unlike the global optimization methods, these porated with Variable Compression Ratio (VCR) mecha-
methods are implementable online. Heuristic methods do not nism [31]. About 50% fuel consumption is reduced by HEVs
accomplish any minimization, instead, the control action is through the Atkinson cycle engine as one of its propulsion
discovered at every instant utilizing empirical rules [12]–[14]. system.
The main useful aspect of heuristic methods lies with the In this connection, the availability of simple and accurate
reduction of computational time and with the disadvantage physically motivated control-oriented engine model incor-
of a non-optimal solution. porated with advanced technologies and the engine robust
Earlier HEV research is categorized into three groups as controllers have been the key contributors in the accom-
per the Internal Combustion Engine (ICE) type. In the first plishment of the SI engine performance [32]. In this per-
category, Compression Ignition (CI) engines have been uti- spective, the authors contribution towards a control-oriented
lized as a part of various HEV models [5], [15]–[19]. The EMVEM model of the Atkinson cycle engine with variable
CI engines have been utilized as a part of sport utility intake valve actuation [33] exist in literature. Besides,
vehicles (SUVs), trucks, and buses. In the second category, the authors [3] developed and evaluated a control framework
Spark Ignition (SI) engines have been integrated as a primary for the Atkinson cycle engine based on its EMVE model
source of energy of HEV [20]–[22]. In the third category, Low in fuel economy perspective for the driving cycles like
Temperature Combustion (LTC) engines have been utilized NEDC, FUDS and FHDS. This study introduced on the
as the main energy source of HEV [23]–[27]. Recently, the average about 7% fuel saving over Otto cycle engine. The
SI engines with Atkinson cycle have turned out to be more research [34] has figured out that there is around 6% decrease
well known in the auto market (e.g., Toyota Prius, Lexus in fuel consumption with VCR Over-expanded engine cycle,
RX 450h, Honda Accord PHEV, and Ford C-Max). using Cylinder by Cylinder Engine Model (CCEM) model
Few studies are found in the literature that explored incor- in comparison with the Otto cycle engine during the typical
porating Atkinson cycle engine with HEVs. In [28], an new European driving cycle (NEDC). About 19% reduction
Atkinson cycle engine was utilized by Toyota R&D group in specific fuel consumption (SFC) of the VVT along with
to build the mileage advantage of HEVs. In another inves- VCR mechanism built-in Atkinson cycle engine with regards
tigation [29], a Honda Accord Plug-in Hybrid Electric to the usual Otto cycle spark ignition engine [31], [35].
Vehicle (PHEV) was upgraded in light of the Atkinson cycle
engine and around 10% better fuel economy was accounted C. CONTRIBUTION
compared to the SI engine. Most of the prior research has been conducted to examine
An Atkinson cycle engine is attained with Variable Valve either the benefits of Atkinson cycle engine incorporated with
Timing (VVT) by using the similar crankshaft scheme as advanced mechanisms alone or on the optimization of fuel
for the Otto cycle engine owing to the recent technological economy of HEV using conventional Otto cycle SI engine
developments, which eliminates the requirement of conven- as one of its propulsion system. The fuel economy investi-
tional throttle at some engine operating range [30]. With gation of the parallel HEV using Atkinson cycle engine in
optimal engine intake and exhaust valves timing, the primary comparison with conventional Otto cycle is not present in

VOLUME 6, 2018 28009

 M. Asghar et al.: EMS for Atkinson Cycle Engine Based Parallel HEV

the literature so far, to the best of authors knowledge. The TABLE 1. Physical parameters of vehicle.
main challenge for the designing of Hybrid Electric Vehicles
is the coordination of onboard energy sources and optimal
power flow control for both the electrical and the mechanical
paths. This requires the utilization of an appropriate con-
trol strategy or energy management strategy. The proposed
energy management strategy utilizes the Boolean rules and
threshold values for the charge-sustainability and the imple-
mentation of torque distribution between the energy sources.
The torque distribution between the energy sources is decided
by selecting the threshold of speed of the vehicle. At the
starting of the vehicle, the whole torque demand is provided
by the electric motor and when the speed crosses over the
selected threshold, the whole torque demand is accomplished
by the engine. The kinetic energy recuperated is utilized to TABLE 2. VVT engine characteristics.
run the motor, contributing towards fuel economy and fur-
thermore, the engine running at part loads improves the fuel
efficiency of the vehicle. The paper provides a comprehen-
sive framework to understand the benefits of Atkinson cycle
engine based electrified powertrains and presents an EMS,
that takes into account the efficient use of Atkinson cycle
engine and delivers better fuel economy as compared to Otto
cycle engine based HEV.
The significant contribution of this research is two-fold.
First, it explores the fuel economy of an Atkinson cycle
engine based parallel HEV drivetrain at part load operat-
ing conditions. Secondly, it finds out the effect of proposed
EMS on the fuel economy for parallel HEV using an Atkinson
cycle engine.

D. STRUCTURE OF THE PAPER before the manual transmission. In the series configuration,
This research paper is structured as follows. In Sec II, Parallel usually two electric motors are present: one acts as a motor,
Hybrid Power-train modeling of the vehicle is presented. and the other operates as a generator. The driving power
Problem formulation is given in Sec III. Vehicle Modeling has to the wheels is received from the motor, which in turns
been described in Sec IV. Development of proposed energy receives the power from the generator connected to the
management strategy along with supervisory control is dis- engine or from the battery bank. Thus, there is an electrical
cussed in Sec V. Simulation results along with the discussion summation of engine and battery power. The parallel config-
of results are narrated in Sec VI. The conclusion with future uration of Hybrid Electric Vehicle is selected for simulation
work is discussed finally in Sec VII. purposes (shown in Fig. 1). Because it is a small vehicle,
series architecture cannot be implemented due to less space
II. PARALLEL HYBRID POWER-TRAIN MODELING availability.
OF VEHICLE For a driving cycle, the selection is made of Manhattan and
A parallel hybrid electric vehicle is modeled with physical modified FUDS drive cycles with the vehicle velocity Vveh .
specifications as shown in Table 1. This is small and light The power request Ptotal is found as in Equation (1)
weighted vehicle, allowing maneuverability on narrow roads. [5], [36], [37] (neglecting the grade force) [38].
Allowable maximum speed is 50 Km/h, having the capacity 1
of two passengers and their luggage. A 12-volt lead acid Ftotal = ρa Af Cd Vveh 2 + MgCr + Ma
2
battery is employed for ignition and lighting power. Char- Vveh
acteristics of VVT engine along with gear specifications are ωveh =
Rw
described in Table 2. 1
Depending on the power-train architecture (the way of Ttotal = ρa Af Cd Rw 3 ωveh 2 + MgCr Rw + MaRw
2
arrangement of various power-train components), two main Ptotal = Ttotal ωveh (1)
categories of Hybrid Electric Vehicles are present: parallel
and series configuration. In the parallel HEV, mechanical In the above relations, ρa indicates the Air density, Af repre-
summation exists between the ICE and motor needed for the sents the Frontal area of the vehicle, Cd is the coefficient of
driving of wheels. Two power sources torques are combined Drag, Vveh indicates the velocity of the vehicle, M represents

28010 VOLUME 6, 2018

M. Asghar et al.: EMS for Atkinson Cycle Engine Based Parallel HEV

FIGURE 2. Simulation model of parallel HEV.

the total mass of the vehicle, g is the Gravitational accelera- III. PROBLEM FORMULATION
tion, Cr is the coefficient of Rolling resistance, a represents The purpose of the Energy management problem is to attain
the acceleration, Rw is the radius of the wheel, ωveh represents the minimum fuel consumption, while keeping in view
speed of the vehicle in radians per second, Ttotal is the total the design limitations of every component and drivabil-
torque and Ptotal represents the total power. ity/performance constraints. The optimization objective taken
for this research activity is the fuel consumption during a
A. SIMULATION MODEL OF PARALLEL HEV journey and mathematically is represented as follows
Simulation Model for the parallel HEV is presented in Fig.2. Z tf
This model comprised of the drive cycle input to the driver J (x) = ṁ(x(k))dt (2)
control module, the energy management strategy module t0
and the plant module (vehicle module). First, the reference where J (x) represents the cost function, ṁ is the rate of Fuel
vehicle speed (desired) Vd for time t is given to the driver mass, x(k) is the state of the system, while t0 and tf represents
module. On the basis of this speed Vd , the acceleration and the initial and final time of driving cycle. The constraints
brake signals are calculated in the driver model through PID considered are as;
controller. These signals are given to the HEV controller, 1. Charge-sustainability: For the charge-sustaining mode
in which the selection of Electric machine and Engine mode of the accumulator, the following integral constraint will lead
is accomplished on the basis of speed and torque demand. as follows
In the Atkinson cycle engine model, late Intake Valve closing Z tf
t
Timing (IVT) load control scheme is employed, whereas Ebatt (t)|tf0 = Pbatt (t)dt = 0 (3)
conventional throttle is kept wide open. However, in the t0
parallel HEV model, the Atkinson engine block calculates the The equation (3) shows that battery energy Ebatt (t) from
torque produced by the engine and this torque is forwarded the initial time t0 to final time tf remains zero by integrating
according to HEV controller scheme. the battery power Pbatt (t), that indicates zero utilization of the

VOLUME 6, 2018 28011

 M. Asghar et al.: EMS for Atkinson Cycle Engine Based Parallel HEV

battery power. The fuel consumption can only be calculated, A. ATKINSON CYCLE ENGINE MODEL DESCRIPTION
when the charge sustaining criteria is fulfilled, otherwise, one The precision of the spark ignition engine model with vari-
should have to take care of the charge used. able innovative technology as well as model-based robust
The SOC dynamics is therefore given as follows control strategies play a vital role in the improvement
p
2 − 4.(P of the engine performance [40]–[43]. In this perspective,
˙ =− 1 V oc − Voc batt ).(Rint )
x(t) (4) the authors [33] have developed a physically provoked
Qmax 2 ∗ (Rint ) control-oriented EMVEM Atkinson cycle engine model,
where Rint is the internal resistance of the battery, in which the modern technologies as flexible valve timing,
Voc indicates the voltage at the terminals of open circuit, over-expansion, VCR and Atkinson cycle’s realization have
Qmax represents the maximum battery charge and Pbatt indi- been incorporated, and is described in the ensuing subsection.
cates the battery power.
2. Driver’s demand: The driver’s demand should be satis- 1) CONTROL-ORIENTED ATKINSON CYCLE
fied, the total output torque of power-train must be same as ENGINE EMVEM MODEL
of driver’s demand; A physically motivated Atkinson cycle engine’s control-
3. Physical limits of actuators: the output of the compo- oriented EMVEM dynamics consisting of the modeling
nents in the power-train should be limited to their operational processes like air dynamics which includes intake manifold
range and are given as follows, pressure dynamics with Variable Valve Actuation (VVA)
incorporated for the realization of the Atkinson cycle, throttle
ωice min ≤ ωice ≤ ωice max
body, air mass suction system, the rotational dynamics with
ωem min ≤ ωem ≤ ωem max an alternate Late Intake Valve Closing (LIVC) control strat-
0 ≤ Tice ≤ Tice max egy [33], [44] and then the fuel dynamics [45], based on
0 ≤ Pice ≤ Pice max • Fluid Dynamics and Thermodynamics Principles
Tem min ≤ Tem ≤ Tem max • Atkinson cycle for the in-cylinder dynamics analysis
• Inertial laws
Pem min ≤ Pem ≤ Pem max
is given as
SOCmin ≤ SOC ≤ SOCmax
ωice = ωem (5) Ṗm = 91 χ(p) − (2 − λ)92 Pm ωe α(Pm , ωe )
1
where ωice min and ωice max represents the minimum and the ω̇e =


Tind (λ) − Tpump − Tfric − Tload (6)
Je
maximum speed of the engine, Tice is the torque of ICE,
Tem indicates the torque of the electric motor, Pice represents where,
the power of ICE, Pem indicates the power of the electric
Tm R
motor, ωem min and ωem max represents the minimum and max- 91 = CD Ae Pa γc
imum speed of the motor. The SOCmin and SOCmax are taken Vm
Vivc
as 0.4 and 0.8 indicating the minimum and the maximum state 92 =
of charge of the battery respectively. 4π Vm
 
Pm
χ(p) = 1 − exp 9 −9
IV. VEHICLE MODEL Pa
For the proposed EMS procedure, a simple but the func- D 2 φ + φcl
tional mathematical model (Quasi-static model) is preferred Ae = π (1 − cos( ))
4 φcl
rather a complex nonlinear model. These models are exten- s s
1 2 γ +1
sively used to design and evaluate the energy management γc = γ( ) γ −1
strategies (EMS) of HEVs. Obviously, the transient behavior RTa γ +1
cannot be exhibited by such models [39], but if, we are
A theoretical Atkinson cycle engine torque termed as indi-
focusing our attention on the fuel economy evaluation, they
cated torque, generated as a process of air to fuel mixture
are still reasonably serving for the comparison of the global
burning is described as
performance of different energy management strategies. For
comparison of fuel efficiency, two different ICEs are studied. Vd
These ICEs include SI engine and Atkinson cycle engine. Tind = ηatk mep
4π
As, focus of the research is on exploring the behavior of
while
Atkinson cycle engine, so modeling of Atkinson cycle engine
γ −1
is described here. rc ζ (rc − λ)
In this work, an Atkinson engine model is a dynamic mep = [ γ −1
+ (λ − 1)(γ − 2)]Pm
(γ − 1)(re − 1) re
model, while the electric machine and battery model are
quasi-static models. In the following section, modeling of the 1 [(γ − 1)λγ − γ λγ −1 + 1]
ηatk = 1 − γ −1
−
major HEV components is presented. re ζ λγ −1

28012 VOLUME 6, 2018

M. Asghar et al.: EMS for Atkinson Cycle Engine Based Parallel HEV

where
QLHV ηc
ζ =
Tm Cv (AFR + 1)
 
ηc = ηcmax −1.6082 + 4.6509σ − 2.0764σ 2
where, ηc is the combustion efficiency with 0.75 < σ <
1.2 [46] and ηcmax is the maximum ηc typically considered as
90 to 95 % for a SI engines [40], [42]. The Pumping torque
essential to carry out pumping action and engine frictional
torque mathematically is given as [40]
Vd
Tpump = (Pa − Pm)
4π
Vd N N2
Tfric = [(0.97 + 0.15 3 + 0.05 6 )105 ]
4π 10 10
Furthermore, to explore the fuel economy of an Atkinson
cycle engine, the stable fuel dynamics subsystem utilized [45]
in this framework is expressed as:
1

m̈ff = −ṁff + Xf ṁfi (7)
τf
FIGURE 3. Theoretical PV representation of an ideal Atkinson cycle
where, engine [33], [42].

ṁfv = (1 − Xf )ṁfi and the engine displaced volume Vd is described as

ṁfo = ṁfv + ṁff π
ṁfi = (2 − λ)92 Pm ωe α(Pm , ωe ) Vd = B2 S (10)
4
whereas, ṁff is the fuel film mass flow, τf is the fuel evap- The EMVEM model parameters, their comprehensive
oration time constant and Xf is the fraction of the injected description and values are as depicted in Table 3.
fuel deposited on the intake manifold or close to the intake The fuel consumption is calculated using the relations as
valve as a liquid. The input to the fuel dynamics model is follows.
the injected fuel flow mfi , whereas output mfo is the engine
intake valve fuel mass flow. The description and values of Pice = Tice ωice
the parameters are as illustrated in Table 3. Pice .BSFC
ṁ =
3600.1000
2) INTAKE VALVE TIMING PARAMETER (λ) Pfuel = ṁ.1H
To incorporate VVA phenomenon in the conventional Pice
η(Tice , ωice ) =
SI engine for the Atkinson cycle realization, a novel intake Pfuel
valve timing (IVT) parameter λ is introduced [33]. It has 3600.1000
= (11)
prodigious importance to summarize the physical dynamics BSFC.1H
of the Atkinson cycle engine, besides the accomplishment of where Pice represents the engine power, Tice indicates the
the advantages of the overexpansion and VCR characteristics. engine torque, ωice represents the engine speed, ṁ represents
Mathematically, it can be expressed as mass fuel rate, BSFC expresses the Brake Specific Fuel
re Consumption, Pfuel is the fuel power, η represents engine
λ = , 1 ≤ λ ≤ 1.60 (8)
rc efficiency and 1H exhibits the fuel heating value.
where,
Vc + Vd B. ELECTRIC MACHINE
re = The efficiency of an electric machine is modeled as a quasi-
Vc
Vc + Vivc static map of its power and operating speed. It is assumed
rc = that the efficiency η is constant for bi-directional operation
Vc
of an electric machine. Therefore, the power flowing in and
where Vc is the clearance volume and the combustion cylin-
out of the electric machine in motoring and generating modes
der’s displaced volume in accordance with the intake valve
is represented as follows.
closing time (IVC) is specified by Vivc = V1 is depicted
in Fig. 3 and is described in the following. Pm (k) = ηPelec (k) motoring (12)
π B2 h n p oi 1
Vivc = r + a − acosθ + r 2 − a2 sin2 θ (9) Pm (k) = Pelec (k) generating (13)
4 η
VOLUME 6, 2018 28013
 M. Asghar et al.: EMS for Atkinson Cycle Engine Based Parallel HEV

TABLE 3. EMVEM parameters description and nominal values. and from the battery are given in the equation (14).
Ps (t) = Voc (t, SOC).I (t)
Pess (t) = Ps (t) − Ploss (t)
= Voc (t, SOC).I (t) − I (t)2 .Rint (t, SOC) (14)
Due to the dependence of an electric system on the battery,
the energy storage capacity of the battery should be excessive
than that required to meet the driving cycle. How the state of
charge SOC(k) is modified for the next state of charge is given
in the following equation (15).
SOC(k + 1)
q
2 −4(R ).T .ω .η−sgn(Tem )
Voc − Voc int em em em
= SOC(k)− (15)
2 ∗ (Rint ).Qb
Where SOC(k + 1) is the next state of charge of a battery,
Rint (internal resistance) and Voc (open circuit voltage) are the
functions of the battery SOC, Qb represents the maximum
battery charge. The overall performance of HEV depends on
the role of the battery. The battery model is shown in Fig. 4.
The battery used is Li-ion type of 4.8 KWh rating indicated
in Table 4.

FIGURE 4. Battery model for hybrid electric vehicle.

where Pm (k) represents the motor power; Pelec (k) is
the battery electric power; and η represents the motor V. PROPOSED ENERGY MANAGEMENT STRATEGY
efficiency. The power-train controller consists of a supervisory con-
Permanent Magnet motor of 07 kW rating is used as a troller (decision of the appropriate operating mode) and the
secondary power source as indicated in Table 4. The motor is energy management controller (which decides the torque
capable of providing the total torque demanded, so the vehicle sharing among the energy sources), while satisfying the
can be run in pure electric mode also. overall demand of driver. The thresholds and rules for each
of these two controller levels are narrated in the following
TABLE 4. Physical specifications of the motor and the battery. subsections.

A. SUPERVISORY CONTROL
As far as the supervisory control is concerned, the following
rules are implemented.
1. At low speed, positive torque: the power-train works in
pure electric vehicle (EV) mode, keeping engine in off state.
This is presented by region A as shown in Fig.5.
C. BATTERY 2. At high speed, positive torque: the power-train works
The battery pack comprises of multiple modules with a com- in thermal (engine) mode. This is presented by region B as
bination of parallel and in series which as a whole taken as shown in Fig.5.
a voltage source with a series resistance. The total battery 3. While operation with a negative torque, the supervisory
power is the total power delivered to or from the electric controller shut off the engine for fuel saving. Since the vehicle
machine. The internal power of the battery Ps (t) and the is in decelerating mode, the power-train works in regenerative
terminal voltage power Pess (t) with current I (t) flowing to mode. This is presented by region C as shown in Fig.5.

28014 VOLUME 6, 2018

M. Asghar et al.: EMS for Atkinson Cycle Engine Based Parallel HEV

2) ENGINE MODE
When SOC ≤ SOCmax , the vehicle system is in engine mode
only, in this mode battery does not need charging and speed
of the vehicle is more than the threshold selected Wsel , while
the torque demand is met by the engine only.

3) REGENERATIVE MODE
When SOC ≤ SOCmin , the vehicle system is in regen-
erative mode only, in this mode battery need charging
and it is charged either through recuperation of kinetic
energy or through engine mode. These three modes are well
depicted in Fig.5. The threshold of speed selected Wsel is the
boundary line between regions A and B.
FIGURE 5. Operating points of engine and motor in different modes.

VI. SIMULATION RESULTS AND DISCUSSION

For all above operating modes, the supervisory control This section narrates the simulation results for implementing
rules implemented are summarized in Table 5; the proposed EMS with VCR Atkinson cycle engine con-
trol framework over the conventional Otto cycle engine at
TABLE 5. Supervisory control strategy parameters. variable load conditions during the standard Manhattan and
modified FUDS driving cycles.
To optimize fuel economy, the initial and final constraints
chosen for SOC is 0.55, which is met in the SOC profile of
Otto and Atkinson cycle for Manhattan drive cycle shown
in Fig 7. The proposed EMS regulates the engine ON/OFF
status according to the SOC values.

B. ENERGY MANAGEMENT
After the decision of mode selection by the supervisory
controller, the torque sharing among the ICE and the elec-
tric machine is decided as follows in the Fig. 6. Three
modes of operation of the vehicle have been discussed,
namely Pure Electric mode, Engine mode, and Regenerative
mode.

FIGURE 7. SOC profile of Otto cycle engine and Atkinson cycle engine for
Manhattan drive cycle.

To evaluate the substantial benefits of the proposed

strategy, the Atkinson cycle engine along with variable
LIVC load handling strategy instead of the conventional
throttle is utilized. The control strategy proposed for the
assessment of fuel economy is explained for Manhattan drive
FIGURE 6. Flowchart of energy management strategy. cycle, while the same approach is also adapted for modified
FUDS drive cycle. For the simulation purposes, Manhattan
and modified FUDS drive cycles are considered due to their
1) PURE ELECTRIC MODE frequent stop/start driving pattern, which in turns represents
When SOC ≤ SOCmax , the vehicle system is in electric the driving pattern of congested cities. The motor drive
vehicle mode only, in this mode battery does not need is facilitated by kinetic energy recuperation related to the
charging and speed of the vehicle is less than the threshold frequent stop/start behavior experienced within these drive
selected Wsel , while the torque demand is provided by the cycles. The supervisory control (which decides the mode
motor only. selection) is well replicated in the Fig. 5, showing each

VOLUME 6, 2018 28015

 M. Asghar et al.: EMS for Atkinson Cycle Engine Based Parallel HEV

mode of operation. The comparison of torque provided by

Atkinson cycle engine and the electric motor is exhibited
through Fig. 8. For the proposed EMS, engine is turned
off during the regenerative mode and motor behaves like a
generator, providing kinetic energy for charging purposes.
The mode choice of proposed energy management strategy
is depicted by Fig. 9, showing Electric launch, Engine only
and Regenerative modes. The vehicle always starts in Electric
mode and when the speed crosses over the selected threshold,
it goes in the Engine mode and while decelerating, receives
the kinetic energy. The kinetic energy recuperated is used to
charge the battery, thus saving fuel energy. Fig. 10 depicts
the tracking velocity profile of the proposed strategy for
FIGURE 10. Actual and desired velocity profile of Manhattan drive cycle.
Manhattan drive cycle. The result also shows that the
HEV components’ sizing fulfills the performance require-
ments during the driving cycle. The pre-transmission par-
allel HEV is modeled in the Matlab/Simulink environment the standard Manhattan and modified FUDS drive cycles,
by using the forward simulator for the comparison of the is depicted in Fig. 11. The standard FUDS drive cycle is
performance of proposed EMS. The flexible LIVC load con- modified due to its high-speed profile. The improvement in
trol strategy has been employed in the EMVEM modeling fuel consumption is achieved about 12.30% for the variable
approach of Atkinson cycle engine instead of the conven- loading conditions during the Manhattan drive cycle, whereas
tional throttle. about 7.22% during the modified FUDS driving cycles for
the proposed EMS. Simulation results of the fuel economy
for a parallel hybrid electric vehicle for the proposed EMS is
shown in Table 6. The torque load is dependent on the speed
and the mass of the vehicle of the vehicle as shown in
equation (1). As the mass of the vehicle is constant, so by
lowering the threshold of speed, the torque load on the engine
can be reduced, resulting in more fuel economy. The improve-
ment in fuel consumption during the Manhattan drive cycle is
more than that of modified FUDS driving cycle as Manhattan

FIGURE 8. Torque profile of motor and engine for Manhattan drive cycle.

FIGURE 11. Comparison of fuel economy for Manhattan and modified

FUDS drive cycles.

TABLE 6. Simulation results of fuel economy of Atkinson cycle engine in

comparison with Otto cycle engine.

FIGURE 9. Mode choice for Manhattan drive cycle exhibited for energy
management strategy.

The fuel consumption comparison attained with proposed

EMS along with Atkinson cycle engine over the conventional
Otto cycle engine with the variable loading condition during

28016 VOLUME 6, 2018

M. Asghar et al.: EMS for Atkinson Cycle Engine Based Parallel HEV

drive cycle is inclined more towards the engine part loads [7] J. Wang, Q. Wang, P. Wang, J. Wang, and N. Zou, ‘‘Hybrid electric vehi-
as compared to modified FUDS driving cycle. Even though cle modeling accuracy verification and global optimal control algorithm
research,’’ Int. J. Autom. Technol., vol. 16, no. 3, pp. 513–524, 2015.
the proposed EMS shows very promising results, there is [8] Z. Chen, W. Liu, Y. Yang, and W. Chen, ‘‘Online energy management of
room for further improvement in the fuel economy and the plug-in hybrid electric vehicles for prolongation of all-electric range based
performance of a parallel HEV. Further improvement can on dynamic programming,’’ Math. Problems Eng., vol. 16, Nov. 2015,
Art. no. 368769.
be achieved through some optimal control strategy, but the [9] S. Onori, L. Serrao, and G. Rizzoni, ‘‘Adaptive equivalent consumption
algorithm of proposed EMS in this research is computation- minimization strategy for hybrid electric vehicles,’’ in Proc. ASME Dyn.
ally efficient due to the simple rules based on the Heuristic Syst. Control Conf., 2010, Jan. 499-505.
[10] L. Serrao, S. Onori, and G. Rizzoni, ‘‘A comparative analysis of energy
methods. management strategies for hybrid electric vehicles,’’ J. Dynamic Syst.,
Meas., Control, vol. 133, no. 3, pp. 12–31, 2011.
VII. CONCLUSIONS AND FUTURE DIRECTIONS [11] C. H. Zheng, G. Q. Xu, S. W. Cha, and Q. Liang, ‘‘Numerical comparison
of ECMS and PMP-based optimal control strategy in hybrid vehicles,’’ Int.
In this research, a novel energy management strategy (EMS) J. Automotive Technol., vol. 15, no. 7, pp. 1189–1196, 2014.
has been proposed for an Atkinson engine based parallel [12] B. M. Baumann, G. Washington, B. C. Glenn, and G. Rizzoni,
hybrid electric vehicle and evaluated through a forward facing ‘‘Mechatronic design and control of hybrid electric vehicles,’’ IEEE/ASME
Trans. Mechatronics, vol. 5, no. 1, pp. 58–72, Mar. 2000.
simulation for standard Manhattan and modified FUDS drive [13] B. K. Powell, K. E. Bailey, and S. R. Cikanek, ‘‘Dynamic modeling and
cycles. The simulation results depict that the proposed EMS control of hybrid electric vehicle powertrain systems,’’ IEEE Control Syst.,
provide substantial improvements to a parallel HEV in terms vol. 18, no. 5, pp. 17–33, Oct. 1998.
[14] Z. Chen, Y.-J. Pan, and J. Gu, ‘‘Integrated adaptive robust control for mul-
of fuel economy exhibited for the standard Manhattan and tilateral teleoperation systems under arbitrary time delays,’’ Int. J. Robust
modified FUDS driving cycle for variable loading condition Nonlinear Control, vol. 26, no. 12, pp. 2708–2728, Aug. 2016.
along with the benefits of Atkinson cycle engine. On the [15] N. Jalil, N. A. Kheir, and M. Salman, ‘‘A rule-based energy management
strategy for a series hybrid vehicle,’’ in Proc. IEEE Amer. Control Conf.,
average around 12.30% and 7.22% improvement in fuel con- vol. 1. Jun. 1997, pp. 689–693.
sumption is achieved with proposed control strategy along [16] A. Brahma, Y. Guezennec, and G. Rizzoni, ‘‘Optimal energy management
with Atkinson cycle engine over the Otto cycle engine during in series hybrid electric vehicles,’’ in Proc. Amer. Control Conf., vol. 1.
2000, pp. 60–64.
the standard Manhattan and modified FUDS driving cycles. [17] C.-C. Lin, H. Peng, and J. W. Grizzle, ‘‘A stochastic control strategy for
The Atkinson cycle engine is proved significantly efficient hybrid electric vehicles,’’ in Proc. Amer. Control Conf., vol. 5. Jun. 2004,
than conventional SI engine, especially for the engine part pp. 4710–4715.
[18] S. Kermani, S. Delprat, T.-M. Guerra, R. Trigui, and B. Jeanneret, ‘‘Predic-
load operating range. tive energy management for hybrid vehicle,’’ Control Eng. Pract., vol. 20,
We plan to evaluate the developed control framework at the no. 4, pp. 408–420, 2012.
vehicle level with a different topology of drive-train, such as [19] D. F. Opila, X. Wang, R. McGee, R. B. Gillespie, J. A. Cook, and
J. W. Grizzle, ‘‘An energy management controller to optimally trade off
series Hybrid Electric vehicle. fuel economy and drivability for hybrid vehicles,’’ IEEE Trans. Control
Syst. Technol., vol. 20, no. 6, pp. 1490–1505, Nov. 2012.
ACKNOWLEDGEMENT [20] A. Sciarretta et al., ‘‘A control benchmark on the energy management
of a plug-in hybrid electric vehicle,’’ Control Eng. Pract., vol. 29,
The authors would like to acknowledge Center for Energy pp. 287–298, Aug. 2014.
Research and Development Lahore (Punjab) for furnishing [21] L. Tribioli, M. Barbieri, R. Capata, E. Sciubba, E. Jannelli, and G. Bella,
technical support for this research project and also control and ‘‘A real time energy management strategy for plug-in hybrid electric
vehicles based on optimal control theory,’’ Energy Procedia, vol. 45,
signal processing research group (CASPR) of Capital Univer- pp. 949–958, Jan. 2014.
sity of Science & Technology Islamabad, 44000, Pakistan. [22] M. Shams-Zahraei, A. Z. Kouzani, S. Kutter, and B. Bäker, ‘‘Integrated
Special thanks to the Center for Automotive Research (CAR) thermal and energy management of plug-in hybrid electric vehicles,’’
J. Power Sour., vol. 216, pp. 237–248, Oct. 2012.
at the Ohio State University, Columbus, USA; for providing [23] A. Delorme, A. Rouseau, T. Wallner, E. Ortiz-Soto, A. Babajimopolous,
the technical guidance for this work. and D. Assanis, ‘‘Evaluation of homogeneous charge compression ignition
(HCCI) engine fuel savings for various electric drive powertrains,’’ in Proc.
EVS, Shenzhen, China, 2010, pp. 1–8.
REFERENCES [24] B. Lawler, E. Ortiz-Soto, R. Gupta, H. Peng, and Z. Filipi, ‘‘Hybrid
[1] A. A. Malikopoulos, ‘‘Supervisory power management control algorithms electric vehicle powertrain and control strategy optimization to maximize
for hybrid electric vehicles: A survey,’’ IEEE Trans. Intell. Transp. Syst., the synergy with a gasoline HCCI engine,’’ SAE Tech. Paper 2011-01-
vol. 15, no. 5, pp. 1869–1885, Oct. 2014. 0888, 2011, pp. 1115–1126.
[2] S. A. Oleksowicz et al., ‘‘Regenerative braking strategies, vehicle safety [25] A. Solouk and M. Shahbakhti, ‘‘Modeling and energy management of
and stability control systems: Critical use-case proposals,’’ Vehicle Syst. an HCCI based powertrain for series hybrid and extended range electric
Dyn., vol. 51, no. 5, pp. 684–699, 2013. vehicles,’’ Int. J. Powertrains, vol. 6, no. 2, p. 1, 2017.
[3] G. Murtaza, A. I. Bhatti, and Q. Ahmed, ‘‘Design, development, and [26] R. Hanson et al., ‘‘Highway fuel economy testing of an RCCI series hybrid
evaluation of a control framework for an Atkinson cycle engine,’’ J. Dyn. vehicle,’’ SAE Tech. Paper 2015-01-0837, 2015.
Syst., Meas., Control, vol. 140, no. 5, p. 051005, 2017. [27] A. Solouk and M. Shahbakhti, ‘‘Energy optimization and fuel
[4] C.-C. Lin, H. Peng, J. W. Grizzle, and J.-M. Kang, ‘‘Power management economy investigation of a series hybrid electric vehicle integrated
strategy for a parallel hybrid electric truck,’’ IEEE Trans. Control Syst. with diesel/RCCI engines,’’ Energies, vol. 9, no. 12, p. 1020,
Technol., vol. 11, no. 6, pp. 839–849, Nov. 2003. 2016.
[5] A. Sciarretta, M. Back, and L. Guzzella, ‘‘Optimal control of parallel [28] N. Kawamoto, K. Naiki, T. Kawai, T. Shikida, and M. Tomatsuri,
hybrid electric vehicles,’’ IEEE Trans. Control Syst. Technol., vol. 12, no. 3, ‘‘Development of new 1.8-liter engine for hybrid vehicles,’’ SAE
pp. 352–363, May 2004. Tech. Paper 2009-01-1061, 2009.
[6] A. Yazdani and M. Bidarvatan, ‘‘A comparative analysis for optimal control [29] A. Yonekawa, M. Ueno, O. Watanabe, and N. Ishikawa, ‘‘Develop-
of power split in a fuel cell hybrid electric vehicle,’’ SAE Tech. Paper ment of new gasoline engine for accord plug-in hybrid,’’ SAE Tech.
2016-01-1189, 2016. Paper 2013-01-1738, 2013.

VOLUME 6, 2018 28017

 M. Asghar et al.: EMS for Atkinson Cycle Engine Based Parallel HEV

[30] D. Akihisa and S. Daisaku, ‘‘Research on improving thermal efficiency AAMER IQBAL BHATTI (SM’07) received the
through variable super-high expansion ratio cycle,’’ SAE Tech. Paper 2010- M.S. degree in control systems from Imperial
01-0174, 2010. College London, London, U.K., and the Ph.D.
[31] J. Martins, K. Uzuneanu, B. S. Ribeiro, and O. Jasansky, ‘‘Thermodynamic degree in control engineering from the University
analysis of an over-expanded engine,’’ SAE Tech. Paper 2004-01-0617, of Leicester, Leicester, U.K., in 1998.
2004. He is currently a Control Engineer in
[32] A. Stefanopoulou, ‘‘Modeling and control of advanced technology dynamic modeling, state/parameter estimation,
engines,’’ Ph.D. dissertation, Dept. Electron. Elect., The Univ. Michigan,
controller design, and controller/observer imple-
Ann Arbor, MI, USA, 1996.
mentation/testing experience in automotive and
[33] G. Murtaza, A. I. Bhatti, and Q. Ahmed, ‘‘Control-oriented model of
atkinson cycle engine with variable intake valve actuation,’’ J. Dyn. Syst., aerospace industries with 22 years of experience.
Meas., Control, vol. 138, no. 6, p. 061001, 2016, paper No. DS-15-1344. He has applied his expertise mainly in automotives, unmanned aerial vehi-
[34] B. Ribeiro, J. Martins, and N. Kothari, ‘‘Otto and VCR Miller engine cles, radar system design, fuel cells, and underground coal gasification. He
performance during the European driving cycle,’’ SAE Techn. Paper 2006- has involved in control of real life engineering systems and has produced
01-0440, 2006. four prototypes so far. He is currently the Head of the Controls and Signal
[35] B. Ribeiro and J. Martins, ‘‘Direct comparison of an engine working under Processing Research Group, Capital University of Science and Technology,
Otto, Miller and diesel cycles: Thermodynamic analysis and real engine having over four Ph.D. students. He has already supervised 19 Ph.D. thesis
performance,’’ SAE Tech. Paper 2007-01-0261, 2007. in his areas of expertise. He has authored or co-authored 130 journal and
[36] G. Rizzoni, L. Guzzella, and B. M. Baumann, ‘‘Unified modeling of hybrid conference publications. He has also won four prestigious national research
electric vehicle drivetrains,’’ IEEE/ASME Trans. Mechatronics, vol. 4, awards.
no. 3, pp. 246–257, Sep. 1999.
[37] G. Rizzoni and V. I. Utkin, ‘‘Advances in automotive control 1998,’’
in Proc. 2nd IFAC Workshop, vol. 7. Loudonville, OH, USA,
Feb./Mar. 1998, p. 262.
[38] J. Rauh and D. Ammon, ‘‘System dynamics of electrified vehicles: Some
facts, thoughts, and challenges,’’ Vehicle Syst. Dyn., vol. 49, no. 7,
pp. 1005–1020, 2011.
[39] W. Sun, Y. Zhang, Y. Huang, H. Gao, and O. Kaynak, ‘‘Transient-
performance-guaranteed robust adaptive control and its application to QADEER AHMED (S’08–M’11) received the B.S.
precision motion control systems,’’ IEEE Trans. Ind. Electron., vol. 63, degree in mechatronics and control engineering
no. 10, pp. 6510–6518, Mar. 2016. from the University of Engineering and Technol-
[40] J. B. Heywood, Internal combustion engine fundamentals. New York, NY, ogy, Lahore, Pakistan, in 2007, and the M.S. and
USA: McGraw-Hill, 1988. Ph.D. degrees in control systems from Moham-
[41] L. Guzzella and C. Onder, Introduction to Modeling and Control of Internal mad Ali Jinnah University, Islamabad, Pakistan,
Combustion Engine Systems. Zurich, Switzerland: Springer, 2004. in 2009 and 2011, respectively. He is currently a
[42] W. W. Pulkrabek, Engineering Fundamentals of the Internal Combustion Research Scientist with the Center for Automotive
Engine. Englewood Cliffs, NJ, USA: Prentice-Hall, 2003. Research, The Ohio State University, Columbus,
[43] Z. Chen, B. Yao, and Q. Wang, ‘‘µ-synthesis-based adaptive robust con-
OH, USA. He is involved in next generation of
trol of linear motor driven stages with high-frequency dynamics: A case
energy efficient vehicle propulsion systems with a focus on modeling, esti-
study,’’ IEEE/ASME Trans. Mechatronics, vol. 20, no. 3, pp. 1482–1490,
Jun. 2015. mation, control, and diagnostics of automotive systems. He has authored over
[44] J. Yao, Z. Jiao, and D. Ma, ‘‘A practical nonlinear adaptive control of 71 international peer reviewed publications. He was a recipient of the Lumley
hydraulic servomechanisms with periodic-like disturbances,’’ IEEE/ASME Research Award in 2018.
Trans. Mechatronics, vol. 20, no. 6, pp. 2752–2760, Dec. 2015.
[45] E. Hendricks and S. C. Sorenson, ‘‘Mean value modelling of spark ignition
engines,’’ SAE Tech. Paper 900616, 1990.
[46] G. H. A. Alla, ‘‘Computer simulation of a four stroke spark ignition
engine,’’ Energy Convers. Manage., vol. 43, no. 8, pp. 1043–1061, 2002.

GHULAM MURTAZA received the M.Sc. degree

MUHAMMAD ASGHAR received the B.E. in physics from the University of the Punjab,
degree in electrical engineering from the Uni- Lahore, Pakistan, in 1994, the M.S. degree in com-
versity of Engineering and Technology, Lahore, puter engineering with specialization in control
Pakistan, in 1993, and the M.S. degree in power systems from the Center for Advanced Studies in
networks from the Institut Nationale Polytechnic Engineering, Islamabad, in 2010, and the Ph.D.
de Grenoble, France, in 2003. He is currently pur- degree in control engineering from the Capital
suing the Ph.D. degree in hybrid electric vehicles University of Science and Technology Islamabad,
from the Capital University of Science and Tech- Pakistan, in 2017. Since 1998, he has been with
nology, Islamabad. He is currently an Operational industry in the field of electronics, instrumenta-
Engineer with the Department of Power Engineer- tion, and control design for electromechanical systems. His research interests
ing, Islamabad. are Atkinson cycle SI engine modeling and control applications emphasizing
His research interests include power engineering, and optimization and on hybrid vehicles propulsion system and inertial navigation sensors.
parameters estimation.

28018 VOLUME 6, 2018