Control Architecture Modeling using Functional Energetic Method

Demonstration on a Hybrid Electric Vehicle

Mert M

ukc

1,2

, Philippe Fiani

, Sylvain Chavanne

, Lahsen Ait Taleb

Cristina Vlad

and Emmanuel Godoy

Sherpa Engineering, R&D Department, 92250, La Garenne Colombes, France

Laboratoire des Signaux et Syst

emes (L2S, UMR CNRS 8506), Centrale Sup

elec - CNRS - Univ. Paris-Sud,

Universit

e Paris Saclay, 3 rue Joliot Curie, F-91192, Gif-sur-Yvette cedex, France

Keywords:

Control Design, Energy Management, Complex Systems, Modeling, Integrated Design, Systems Modeling,

Interconnected Systems, System-level Design, System Veriﬁcation, Automotive Engineering, System

Architecture.

Abstract:

With the advances on component technology, communication and information, energy systems are becom-

ing more complex. In this context, energy optimization based on various criteria requires the development

of relevant and representative models that are able to characterize the system behaviour. Within this study,

functional modeling is used to represent a system at a higher level of abstraction, with simple equations, lo-

cal control loops and a decision manager (DM) for handling the energy ﬂow. The reduced complexity and

fast simulation of this model simplify the validation of system architecture and components sizing, as well as

the performances evaluation of energy management algorithms according to different criteria. Once this ﬁrst

validation is completed, the following step in the system design process is to test the same algorithms on a

more accurate model, represented at multi-physical level, that has its own local controllers and global resource

manager (GRM). One way to complete this second validation is to use the information computed using the

functional model, to design a high level controller of a more complex multi-physical model. To this purpose, a

solution is proposed to interconnect the two models, of the same system, that are represented at different level

of abstraction. First, it is shown how the GRM can be extracted from the functional model. Secondly, it is pre-

sented how this management system can be adapted in order to be used at multi-physical level. Both models

are developed for a plug-in parallel hybrid vehicle (PHEV), and the interconnection solution is illustrated for

the considered application.

1 INTRODUCTION

It is important to cite economic and ecological frame-

work that drags industry and researchers towards an

innovative energy management involving an associa-

tion of energy technologies, optimal control laws and

reﬁned components. Along with the advancements

on component technology, communication between

them and obtaining information make systems more

intelligent. This growth of intelligence also makes

the system more complex. For energy systems, this

complexity hardens the work on energy management

and control strategies which enforces the research on

system engineering. The most important difﬁculties

on this subject are:

• Choosing the system architecture;

• Setting and sizing the system components;

• Optimizing the ﬂow between multi-sources and

multi-consumers;

• Designing a control system architecture.

The goal of system engineering, as far as energy

systems are concerned, is to obtain optimal control

laws and energy optimisation. Increased complexity

and technology advancements on components drive

the research to evolve on modeling and simulation.

This research is made to design, identify and control

the system. All these challenges lead the system de-

velopers to research further on:

• A global and interactive approach to improve sys-

tematic innovation;

• A methodology for architecture evaluation and

system veriﬁcation from the early stages of the

system life cycle;

• A system representation from multiple points of

view, in order to deﬁne and analyse the main ob-

Mokukcu, M., Fiani, P., Chavanne, S., Taleb, L., Vlad, C. and Godoy, E.

Control Architecture Modeling using Functional Energetic Method - Demonstration on a Hybrid Electric Vehicle.

DOI: 10.5220/0006413300450053

In Proceedings of the 14th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2017) - Volume 1, pages 45-53

ISBN: 978-989-758-263-9

jectives;

• Design and identiﬁcation of a control system.

For these reasons, including the tendency to de-

crease time −to − market, the complex system has to

be represented at a higher level of abstraction that will

ease its global understanding within a structured envi-

ronment. In the literature, this type of representation

is associated with systemic theory (Le Moigne, 1977).

Usually, physical models of complex systems have

been represented and analysed using Bond Graph

modeling and multi-domain simulation (Brunet et al.,

2005). However, another interesting approach is the

functional modeling (Penalva, 1994), (Suh, 1998),

(Mokukcu et al., 2016), which is based on the fol-

lowing principle: a system can be deﬁned by basic

elements, modelled with an adequate level of com-

plexity, that faithfully describe the system behaviour.

By construction, Model Based System Engineer-

ing (MBSE) allows to specify and design systems at

different levels and to specify their elements and the

links between them. These links are: components and

information, requirements, architecture (functional,

multi-physical or otherwise), use cases and validation

tests (Fiani et al., 2016).

In (Fiani et al., 2016) and (Mokukcu et al., 2016),

three levels of modeling are introduced as:

• Teleological modeling: a system of missions ﬁ-

nalized with respect to system environment, gov-

erned by regulations and standards;

• Functional modeling: the system of missions can

be deﬁned by a set of key functions and the asso-

ciated architecture that realizes the missions;

• Multi-physical modeling: the key functions are

realized by a set of components and an equipment

usually provided by suppliers.

Figure 1: Multi level integrated design and simulation (Fi-

ani et al., 2016).

Fig. 1 illustrates the modeling steps for each sys-

tem representation at a different level of abstraction.

The development starts with requirements formula-

tion. Once the requirements are ﬁxed, the parame-

ters (P) and objectives (O) are deﬁned in order to ob-

tain a simulation model and its associated controller.

The evaluation of the resulting control system is per-

formed in simulation using validation criteria. If the

criteria are satisﬁed then the parameters of higher lev-

els of modeling will deﬁne the requirements of lower

levels of modeling. Otherwise, necessary modiﬁca-

tions are made in the design process. This mecha-

nism helps to transmit objectives or parameters be-

tween different representations of the system. It can

be also found out that higher modeling levels become

cascade controllers for lower modeling levels (Fiani

et al., 2016).

At functional level of abstraction, the system be-

haviour is represented from an energetic point of

view, using simple equations that allow reducing the

amount of time needed to complete a simulation. The

functional modeling methodology and its semantics

(Mokukcu et al., 2016), (Fauvel et al., 2014), (Fau-

vel, 2015) are based on a FU (Functional Unit), also

referred as OFS (Organico-Functional Set) approach.

In Fig.2, the representation of a functional model is

given. Each element

∑

represents a functional unit

while D elements are used for energy distribution.

Figure 2: Functional modeling representation.

However, at multi-physical level, the system can

be represented as a composition of controlled sub-

systems (Mokukcu et al., 2016). The block diagram

of a multi-physical representation is shown in Fig. 3,

where C, I, T and E denote the local controller,

input conditioning, trans f ormer and e f f ector, re-

spectively. The GRM block acts like an energy man-

agement system for the multi-physical model.

Figure 3: Multi-physical modeling representation.

Indeed, in the early stages of the design process,

a functional model is preferred to represent the com-

plex system in order to validate (by fast simulations)

the system architecture and components sizing, and

ICINCO 2017 - 14th International Conference on Informatics in Control, Automation and Robotics

also to evaluate the performances of local controllers

and energy management strategies for different mis-

sions using different criteria. Naturally, the following

stages in the system design process is to test the su-

pervision and control algorithms that have been de-

veloped using the functional model, on the multi-

physical model. In Fig. 4, the control architecture is

illustrated at functional and multi-physical levels of

abstraction. On the left side of the ﬁgure, the sys-

tem is represented at functional level and the energy

ﬂow within the system is managed by the supervi-

sion block DM using optimisation algorithms. On the

right side, C is a composition of local controllers of

the system, and P includes the physical subsystems.

Therefore, the main difﬁculty is to obtain the global

resource manager (GRM) of the multi-physical model

using information provided by local controllers and

decision manager (DM) of the functional model.

Figure 4: Control system representation.

The struggle of this extraction is given by: which

input/output of which physical subsystem should be

measured/estimated, how to use these signals to pro-

vide energy transfer information to functional model,

and after processing this information, how to trans-

form the computed power ﬂow reference into a phys-

ical reference signal and transfer it to all different

types of controllers of physical subsystems.

For a better understanding of these challenges in

the context of an energy system, a hybrid electric ve-

hicle (HEV) is considered as an example of a multi-

source/multi-consumer system.

In this work, the issue of interconnecting the func-

tional and multi-physical models is presented, and a

solution is proposed showing how the GRM can be

extracted from the functional representation and be

connected to local controllers of the multi-physical

model. In Section 2, the multi-physical and functional

modeling methods are brieﬂy introduced. In Sec-

tion 3, the interconnection procedure between the two

modeling levels is discussed. Section 4 presents both

models for a plug-in parallel hybrid vehicle, along

with the multi-physical model obtained as a result of

interconnection. Its performances are tested in simu-

lation for a speciﬁc mission. Conclusions and future

works are summarized in Section 5.

2 MODELING METHOD

REMINDERS

This section introduces brieﬂy the multi-physical and

functional modeling methods, which are further ap-

plied to model the behaviour of a gear motor group at

multi-physical and functional level, respectively.

2.1 Multi-physical Modeling

Multi-physical modeling aims to represent the archi-

tecture of technological equipment. Generally in in-

dustry the 0D-1D multi-physical modeling is used

for a complex system to optimise its sizing, for con-

trol laws design and validation. This multi-physical

model allows representing the complex system as a

whole and is used for simulations, analysis and pre-

diction of system performances.

The multi-physical model is composed by ana-

lytical models that provide an accurate description

of the multi-physical behaviour of the complex sys-

tem. The multi-physical model can be developed un-

der the simulation environment Matlab/Simulink us-

ing a component-based approach derived from the

Bond Graph methodology. It is a language that allows

the passage between physical and mathematical mod-

els using a block-diagram environment (Brunet et al.,

2005). On the other hand, the simulation tool is based

on a multi-port concept: a unique link is used to rep-

resent and simulate all the interactions between dif-

ferent components. In the multi-physical methodol-

ogy, this link is represented by energy transfer. More-

over, every link between physical model components

consists of a ﬂux variable and an effort variable that

depend on the physical domain. In Fig. 5, some ex-

amples are given for different domains. Despite the

advantages of multi-physical modeling (accuracy and

intermediate signals availability), the model design,

its simulation and validation are time consuming and

require expertise.

Figure 5: Multi-physical domains.

For these reasons, it is necessary to use a model

of a higher level of abstraction, which does not need

the deﬁnition of multi-physical elements, in order to

easily evaluate the system in the early stages of the

design process.

Control Architecture Modeling using Functional Energetic Method - Demonstration on a Hybrid Electric Vehicle

An example of multi-physical model for a gear

motor group is given in Fig. 6(b). The transformation

of electrical ﬂow into mechanical ﬂow is done using

a converter (1), an electrical motor (2), (3) and a gear

reducer (4), along with their local controller. For this

example, the physical behaviour of each component is

represented by a simple analytical model as follows:

= u

+ t

o f f

. (1)

= R

· i

+ L

+ E. (2)

= τ

− τ

ext

. (3)

ext

= α · τ

out

. (4)

where u

denotes the rotor voltage; u

is the con-

verter supply voltage; t

o f f

is the converter on/off

time; R

, i

, L

is the resistance, current and induc-

tance of the rotor, respectively; E is the electromotive

force; J

is the inertia; θ

is the motor angular posi-

tion; τ

is the electromagnetic torque, τ

is the loss

torque; τ

ext

is the motor output torque; α is the gear

constant and τ

out

is the gear output torque.

Figure 6: Multi-physical (a) & functional (b) model repre-

sentations.

2.2 Functional Modeling

The concept of functional modeling has been intro-

duced and detailed in our previous work (Mokukcu

et al., 2016). Unlike multi-physical modeling con-

cept, in this methodology, functional links are em-

ployed to represent ﬂow exchanges within the com-

plex system. In a functional model, the ﬂow can be ei-

ther energy, either matter or both, and it is exchanged

together with an information ﬂow. The exchange of

ﬂow is made between ﬁve types of elements (source,

storage, distribution, transmission and effector), each

of them having source and consumer ports. Source

ports receive expressed need from consumer ports and

they answer by supplying the requested need. In ad-

dition, consumer ports transmit demands of need to

source ports. This gives the method its modularity.

Brief information about the basic elements of func-

tional modeling is given in Table 1. The functional

model can be simulated using the Matlab/Simulink

simulation environment with a functional modeling li-

brary, which contains all this basic elements.

Table 1: Element types of functional energetic modeling

and associated functions (Mokukcu et al., 2016).

Source Storage Transformation Distribution Effector

Energy &

Matter

Source

Energy &

Matter

Storage

Energy & Matter

Transformation

in Different

Domains

Energy &

Matter

Distribution

Represents

Energetics

Services

In a functional model, the need computation starts

from the effector. For example, the need of energy for

a plus-in hybrid electric vehicle (PHEV) is calculated

by electrical auxiliary element or vehicle dynamics el-

ement (both of them effectors of the system). Then,

the energy need is transmitted to storages or sources

via distribution and transformation elements. Based

on information ﬂow, the storages and sources can de-

cide whether they are able to provide the requested

energy or not. Furthermore, distribution elements are

used to manage the energy ﬂow between sources and

storages, and to supply the requested energy to ef-

fectors as an answer to their need. If a hotel water

treatment system is considered, the need of water con-

sumption is calculated by hotel consumer element (ef-

fector) and the hotel logistics element must supply the

required amount of water, with suitable properties.

Fig. 6(a) illustrates an example of a functional

model, which also represents an energy transforma-

tion (electrical energy into mechanical and thermal

energy) but without considering the real physical be-

haviour. The model is described as below:

mech

= η · P

(5)

where P

mech

denotes the mechanical output power,

is the electrical input power and η is the efﬁ-

ciency. Besides (5), maximum and minimum power

limitations, P

min

and P

max

, are speciﬁed for each func-

tional element. Moreover, these elements include the

dynamic behaviour of the functional units. The dy-

namic behaviour is taken into account either by an in-

tegration for the energy-to-power transition, either by

adding 1

(or 2

) order transfer functions of the dif-

ferent elements such as transformations, storages or

effectors.

ICINCO 2017 - 14th International Conference on Informatics in Control, Automation and Robotics

The functional model allows fast simulations for

system evaluation (sizing, architecture, requirements

management) before choosing the technology, obtain-

ing the GRM of multi-physical model and simulating

the system as a whole. In the next section, the in-

terconnection between functional and multi-physical

modeling is presented, which comes to extract the

GRM from the functional model.

3 INTERCONNECTION

BETWEEN FUNCTIONAL AND

MULTI-PHYSICAL MODELING

Functional modeling deﬁnes key-functions (FUs), al-

locates and reﬁnes end-mission requirements to the

FUs and also deﬁnes the energy management system.

On the other hand, multi-physical modeling deﬁnes

the physical architecture or physical units, allocates

and reﬁnes the functional requirements to physical

units. To overcome the challenges associated with the

control laws design and energy management within

the entire system, functional and multi-physical mod-

els are interconnected. In this section, the difﬁculties

related to the interconnection are presented along with

the proposed solution. This solution is presented for

the gear motor group and the electromechanical en-

ergy transformation that are introduced as examples

in Section 2.

3.1 Problems of Interconnection

As presented in Fig. 4, the functional modeling level

includes a control strategy that will be used by the

control system of the multi-physical model. More-

over, this strategy is independent from technical com-

ponents, and is deﬁned according to the decision man-

ager allocated from end-missions model.

In Fig. 7, a representation of control systems and

Figure 7: Control system and ﬂow exchange representations

of functional and multi-physical modeling.

ﬂow exchanges of functional and multi-physical mod-

els is shown, for a battery electric vehicle (BEV).

Here, the challenge is to ﬁnd the adequate language

to connect both models.

As ﬂux exchanges are different between func-

tional and multi-physical models, the connection can-

not be done directly. Since multi-physical model

components need physical domain references and

functional model components require a power de-

mand reference for simulation, connecting the power

ﬂow to physical domain ﬂows can be a challenging

task. At this stage, the interesting features of the in-

terconnection can be expressed as follows:

• Functional modeling allows fast control archi-

tecture design and fast adaptation to eventual

changes in the system,

• Multi-physical representation is too complex and

time consuming when trials are accomplished.

3.2 Proposed Solution

A solution to the interconnection problem is to build

an interface between the multi-physical and func-

tional model. As illustrated in Fig. 8, this interface

contains passage equations between physical domain

and functional domain. It accomplishes the following

functions: determine the equivalent physical refer-

ences required for the multi-physical model based on

the power demand provided by the functional model;

measure/estimate the power supply that the system

is able to deliver using information from the multi-

physical model, and transfer the estimated power sup-

ply to the functional model. For each functional

model element, an interface is required in order to cal-

culate/adapt the necessary values.

If the electromechanical transformation element is

considered, the interface between this element and

Figure 8: Functional to multi-physical domain interconnec-

tion.

Control Architecture Modeling using Functional Energetic Method - Demonstration on a Hybrid Electric Vehicle

electrical propulsion group (drive and electric ma-

chine in this example) uses the following equations:

∗

f nc

= τ

∗

cns

. (6)

mech

f nc

. (7)

where P

∗

f nc

denotes the power demand; |

| is esti-

mated/ measured angular speed of rotor; τ

∗

cns

is torque

demand; P

mech

is calculated mechanical power of the

electrical machine and

f nc

is estimated/measured

output power of the motor.

In the next section, an example of PHEV is pre-

sented. First of all, model architectures of functional

and multi-physical models are given, and secondly,

simulation results using the functional model and the

multi-physical model with GRM are discussed.

4 APPLICATION EXAMPLE:

HYBRID VEHICLE ENERGY

MANAGEMENT SYSTEM

4.1 Motive

In a hybrid electric vehicle (HEV), and internal com-

bustion engine (ICE) and one or several additional

electric motors (EMs) are used for the vehicle pow-

ertrain. The ICE is supplied by fuel while the

EMs are supplied by a battery. These components,

usually allowing different possible interconnections,

form a complex and challenging multi-source/multi-

consumer system in terms of optimal control design

and energy management. Both objectives of the de-

sign process have to satisfy several vehicle services

like fuel consumption or comfort level. Although

there are optimization methods applied on HEVs,

they are implemented for a speciﬁc architecture of the

HEV and they usually require a priori knowledge of

the driving cycle. Thus, the problem is how to manage

the power split that globally satisﬁes the vehicle ser-

vices whenever the vehicle has a new task (Sciarretta

and Guzzella, 2007) and/or the system architecture is

reconﬁgured.

In this study, a parallel plug-in hybrid electric ve-

hicle (PHEV) is considered due to the resemblance to

a battery electric vehicle that has been highly investi-

gated over the last few years.

4.2 Control Architecture

The functional model of the parallel plug-in hybrid

vehicle has been developed in our previous work

(Mokukcu et al., 2016). The developed model, shown

in Fig. 9, is used to compute the power split bewteen

the multiple sources of the system, for different con-

ﬁgurations and missions of the vehicle. Moreover, it

allows to evaluate the fuel consumption, maximum

speed, maximum acceleration and regenerative brak-

ing power (Mokukcu et al., 2016).

Figure 9: Functional model of PHEV (Mokukcu et al.,

2016).

Thereafter, the next step in the design process is to

use the information provided by the functional model

(i.e. power signals for each source) for control de-

sign of a more complex multi-physical model. To this

purpose, the power signals are transformed into phys-

ical reference signals using an unique interconnection

element, which is added to each element of the func-

tional model to adjust the ﬂow nature, as shown in

Fig. 10. Using this link, the functional and multi-

physical models are able to exchange necessary val-

ues of power or physical references, as well as mea-

sured/estimated values.

Figure 10: Functional model of M2E transformation ele-

ment with its connection element.

The multi-physical model of the system is given

in Fig. 11, where each component is a system itself.

For example, Fig. 12 illustrates the representation of

the electric machine subsystem.

As it can be noticed, the system architectures are

similar in both functional and multi-physical models.

Thus, if there is any change at multi-physical level,

the functional model has to be adapted respectively.

ICINCO 2017 - 14th International Conference on Informatics in Control, Automation and Robotics

Figure 11: Multi-physical model of HEV.

Figure 12: Electric machine component group.

4.3 Simulation Results

The simulations are run under the following assump-

tion: the vehicle always moves forward. The model

parameters are consistent with those of a parallel

PHEV available on the market. Table 2 provides the

technical characteristics of the vehicle.

Table 2: Technical characteristics of PHEV.

Technical Characteristic Value

Fuel tank max. volume 45 l

ICE max. output power 70 kW @ 5000 rpm

ICE max. output torque 140 Nm @ 4500 rpm

Battery voltage 210 V

Battery capacity 50 Ah

EM max. output power 60 kW

EM max. output torque 200 Nm

Combined max. output power 100 kW

Vehicle curb mass 1500 kg

Vehicle SCx (Aerodynamic drag coeff.) 0.63

Vehicle wheel radius 0.635 m

To be able to compare the simulation results with

the manufacturers brochure, the vehicle performance

indicators are determined and their values are given in

Table 3.

Table 3: Performance indicators of PHEV.

Performance data Value

Combined consumption (WLTC cycle) 3.2 l/100 km

Electric drive range 25 km

Vehicle max. speed 180 km/h

Vehicle max.speed in e-drive mode 85 km/h

Vehicle max. acceleration (0-100km/h) 11.4 s

First of all, the functional model with its DM is

simulated using the WLTC (Worldwide harmonized

Light vehicles Test Cycle) that yields the vehicle

speed and the power demand illustrated in Fig. 13

(a), (b). In addition, the DM uses a ruled-based en-

ergy management strategy based on priorities, which

is implemented in the distributor elements of the func-

tional model. In this example, the functional model

has three main distributors that are detailed in Table 4

with their priorities.

Table 4: Distributor priorities.

Priority No Distributor 1 Distributor 2 Distributor 3

1 Drive Electrical Aux Electric Drive

2 Battery Charge Drive Fuel Drive

3 N/A N/A Brake System

Distribution 1 transmits mechanical energy sup-

ply from fuel to mechanical transformation element

to drive or mechanical to electrical transformation el-

ements. Distribution 2 transmits electrical storage en-

ergy supply to electrical auxiliary or drive. Distribu-

tion 3 transmits the energy need of vehicle dynamics

to electric drive supply element or to fuel drive supply

element or to the brake system.

The obtained results are also shown in Fig. 13.

The vehicle speed and power achieve the desired pro-

ﬁles and meet the requirements of the WLTC. The

regenerative braking can be observed between 1600s

and 1800s in the Fig. 13 (c), (d). We remark that

the vehicle speed on electric drive is limited to 85

km/h and the electrical storage/battery SOC (state of

charge) is limited to %20; beyond these values, the

electric drive is abandoned and electrical energy is

consumed just by electrical auxiliaries.

Figure 13: Parallel PHEV - simulation results using the

functional model (Mokukcu et al., 2016).

These results represent a ﬁrst validation of the

chosen architecture of the vehicle and of the energy

management strategy used to handle the power split

within the system. However, at this level of abstrac-

tion, the energy model cannot generate speciﬁc phys-

ical signals such as electric motor output torque or

Control Architecture Modeling using Functional Energetic Method - Demonstration on a Hybrid Electric Vehicle

battery output current.

Simulation results of the multi-physical model

and the proposed control arhitecture are illustrated in

Fig. 14.

Figure 14: Parallel PHEV - simulation results using the

multi-physical model with GRM.

According to Fig. 14, the following remarks can

be made:

• Vehicle power need pattern is compatible with al-

located source powers;

• When the vehicle surpasses 85 km/h (Fig. 14 (c))

the source power allocation moves to fuel source

power, but at the same time battery SOC decreases

(Fig. 14 (d)). The reason behind this is the con-

stant electrical auxiliary load;

• A slight increase in battery SOC is seen at the end

of the simulation showing the regenerative brak-

ing effect;

• Based on the parameters values shown in Table 2,

the results are consistent with the physical limits

of the components;

• From the acquired data, the fuel consumption can

be calculated from following equation:

4SOC

Fuel

· vol

Fuel

· 100 (8)

For this test scenario (WLTC), the obtained fuel

consumption is of 3.5l/100km. This result is

well approximated by the value given in Table 3

(3.2l/100km). Besides the fuel consumption, power

need and supply patterns have been compared. Slight

differences can be observed due to the system dy-

namic behaviour, especially at time instants with neg-

ative power supply.

With the proposed solution, the system can be ex-

amined globally but also locally. Each component of

the vehicle can be investigated separately if the simu-

lation model permits. Fig. 15 shows the electric ma-

chine results.

Figure 15: Hybrid vehicle electric machine results.

From the speciﬁc physical signals of the electric

motor, the following comments can be made with in-

formation that is given in Table 2:

• Output torque values of electric machine are

within its physical limits (maximum output torque

is 200 Nm);

• The angular speed of the electrical machine fol-

lows the vehicle speed with a certain gear ratio;

• The electric motor current is between the physical

limits with possibility of detailed analysis for re-

generative braking (for example motor K

(motor

torque constant) value is approximately 1, which

is acceptable);

• The electric motor mechanical power is illustrated

in order to calculate the motor and generator efﬁ-

ciencies.

These results highlight the advantages of a multi-

physical model with a GRM: detailed analysis of

components, better precision and, therefore, reliable

validation of simulations. Other physical components

(ICE, auxiliaries, battery etc.) can also be analysed

using the same simulation data. However, data ex-

ploitation depends on the multi-physical model com-

plexity.

4.4 Additional Comments on

Reconﬁguration

A major advantage of the functional model is the

ability to handle the system architecture reconﬁgura-

tion without reviewing the analytical modeling, which

cannot be avoided for a multi-physical representation

of the system. If a component is added or removed,

ICINCO 2017 - 14th International Conference on Informatics in Control, Automation and Robotics

the new conﬁguration can be validated in a fast and

efﬁcient way. An example of this interesting feature

is demonstrated in Fig. 16, highlighting the modular-

ity of functional modelling. Compared to Fig. 9, an

additional component is added to the system: a sec-

ond electric machine used for traction.

Figure 16: Reconﬁgurated HEV model.

Therefore, with faster simulations compared to

multi-physical modeling and ease of reconﬁgura-

tion, the functional modeling becomes a very useful

methodology for system modeling and simulation.

5 CONCLUSION AND FUTURE

WORK

In the presented work, a methodology of modeling a

control architecture was proposed using a functional

energy-based approach. The methodology was ap-

plied to a PHEV with WLTC use case, using Mat-

lab/Simulink simulation environment. The intercon-

nection between functional model and multi-physical

model serves to add a high level control to the multi-

physical system representation. At functional level of

abstraction, where the energy system exchanges need

and supply, the energy management algorithms are

easier and faster to adapt at this level of representa-

tion. Based on the functional model, the proposed

concept simpliﬁes the control design and computes

appropriate multi-physical reference signals for the

multi-physical model.A ﬁrst perspective of the pro-

posed modeling methodology is to establish a gener-

alized method to interconnect the two levels of mod-

eling.

Also, the simulated multi-physical model does not

contain gear box for the ICE. The components (ICE or

EM) models can be acknowledged as simple, which

means that the dynamic behaviour of the system and

its components cannot be thoroughly researched.

Future works intend to enhance the accuracy of

multi-physical model (by improving the components

models or by adding additional ones) in order to have

a more reliable validation. On the other hand, robust-

ness of the control architecture will be further investi-

gated.

Another perspective is to improve the energy man-

agement strategy used at functional level: more efﬁ-

cient algorithms for need/supply distributions should

be developed and integrated into the simulation mod-

els for further validation. In this work, distributors

use priorities for power need/supply distribution. For

an optimal power split, an otimization-based energy

management algorithm should be used.

Finally, the proposed method for control architec-

ture design will be tested on different types of appli-

cations that need an optimal energy management like

water treatment systems or building energy manage-

ment systems.

REFERENCES

Brunet, J., Flambard, L., and Yazman, A. (2005). A hard-

ware in the loop (HIL) model development and im-

plementation methodology and support tools for test-

ing and validation car engine electronic control unit.

Lecce, Italy. International Conference on Simulation

Based Engineering and Studies, TCN CAE.

Fauvel, C. (2015). Approche modulaire de l’optimisation

des ﬂux de puissance multi-sources et multi-clients,

vis

e temps reel. PhD thesis, Nantes, France.

Fauvel, C., Claveau, F., and Chevrel, P. (2014). Energy

management in multi-consumers multi-sources sys-

tem: a practical framework. pages pp.2260–2266,

Cape Town, South Africa. IFAC World Congress.

Fiani, P., Taleb, L. A., Chavanne, S., and Mokukcu,

M. (2016). Mod

elisation pour la conception et

l’

evaluation de syst

emes complexes. Revue Ing

enieurs

de l’automobile, vol.841.

Le Moigne, J. (1977). La th

eorie du syst

eme g

eral,

eorie de la mod

elisation. PUF, Paris, France.

Mokukcu, M., Fiani, P., Chavanne, S., Taleb, L. A., Vlad,

C., Godoy, E., and Fauvel, C. (2016). A new con-

cept of functional energetic modelling and simulation.

pages pp.536–541, Oulu, Finland. The 9th Eurosim

Congress on Modelling and Simulation.

Penalva, J. (1994). Sagace: la mod

elisation des system

dont la ma

ıtrise est complexe. volume vol.94, pages

pp.536–541, Montpellier. 2nd International Confer-

ence on Integrated Logistics and Concurrent Engi-

neering (ILCE).

Sciarretta, A. and Guzzella, L. (2007). Control of

hybrid electric vehicles. IEEE Control Systems,

vol.27:pp.60–70.

Suh, N. (1998). Axiomatic design theory for sys-

tems. Journal of Research in Engineering Design,

vol.10:pp.189–209.

Control Architecture Modeling using Functional Energetic Method - Demonstration on a Hybrid Electric Vehicle