You are on page 1of 18

INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

Intelligent Control Of An Electric Vehicle


(ICEV)
Taoufik Chaouachi, Kamel Jema

Abstract: The electric vehicle allows fast, gentle, quiet and environmentally friendly movements in industrial and urban environments. The automotive
industry has seen the opportunity to revive its production by replacing existing vehicles due to the reluctance of oil reserves around the world. In order to
greatly reduce countries' dependence on oil, strategic sectors such as transport must increasingly integrate technologies based primarily on clean and
renewable energy. Governments must implement large-scale measures to equip themselves with electric vehicles and build large recharge networks.
The traditional system for conversions of conventional vehicles into electric vehicles consists of replacing the internal combustion engine and the
gearbox with electrical components (engine and gearbox, or engine and gearbox), retaining the rest of the elements Transmission (transmission shafts,
etc.).

Index Terms: Strategy of control, random perturbation method , fuzzy logic,power train vehicle,electric vehicle,Permanent magnetic synchronus motor.

1 INTRODUCTION instantaneous variation of the error xMi/ t , to synthesize


Reducing CO emissions represents a challenge for the a fuzzy vector KxMi [6], [7],[8].
transport sector. Transportation produces approximately 23
percent of the global CO emissions from fuel combustion.
With rapid urbanization in developing countries, energy
consumption and CO emissions by urban transport are
increasing quickly[18]. Over the last few decades, the
environmental impact of oil transport infrastructure,
combined with the reluctance of energy resources, has led
to renewed interest in an electric transport infrastructure.
Electric vehicles differ from fossil fuel vehicles in that the
electricity they consume can be produced from a wide Fig.1.Model of a fuzzy controller
range of sources, including fossil fuels, nuclear power and
We defined the degree of membership of the error xi to a
renewable Tidal energy, solar and wind energy or any
combination of these. The energy consumption of these x
x

i
and
electric vehicles varies according to the fuel and the i t
membership function mfunc r by mfuncr mfuncr
technologies used to produce electricity. Electricity can then
be stored in the vehicle using an on-board battery[1], [5]. the degree of membership of the error variation to another
The electric motors are mechanically simple and often membership function [18].
achieve high energy conversion efficiencies over the whole
range of speed and power developed and can be controlled Let then:
with high accuracy. They can also be combined with x
x
i
regenerative braking systems that have the ability to
convert motion energy into stored electricity. This can be R mfunc .
k
i
r
t
mfuncr
used to reduce brake system wear and reduce total energy (1)
consumption. Electric motors can be finely controlled and
provide high starting torque, unlike internal combustion
engines, and do not require multiple gears to match power
curves. This eliminates the need for gearboxes and torque
converters. Electric vehicles offer silent and smooth
operation and therefore have less noise and vibration than
internal combustion engines. In this work, we have
incorporated new strategies for controlling the state
variables of an electric vehicle based on fuzzy controllers.
We attached particular importance to the mechanical
transmission system by means of an appropriate modeling
of the symbiosis between the electrical management
system and the mechanical management system.

2 FUZZY STRATEGY Fig.2. Rulers treatment

2.1 Fuzzy Studied Model


The topology of each fuzzy controller to integrate, was A surface S Rk swept by the control vector for a rule Rk
based on an interaction between two input variables, is given by: K
characterized successively by an error xMi and an x i

128
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

S Rk Rk .mfuncr where :
(2) NTG : Negative Very Big,
NG : Negative Big,
N : Negative,
NM : Negative Medium ,
NP : Negative Small,
ZE : Zero,
PP : Positive Small ,
PM : Positif Medium,
P : Positif,
PG : Positif Big,
PTG : Positif Very Big.
Fig.3.Surface related to rule Rk
The overall area swept by the fuzzy vector K xi after use of
Indeed, we have implemented a matrix of fuzzy inferences all the rules is formulated as follows :
through a number of membership functions MFi , k N Rk

x ( N 11 ), That for its variation S Rk


Both for the error x SK k

x N
xi
N Rk
( x 11 ) The
(4)
same for the control vector Uf (
t t
With : N R k number of rules that are used. Thus, the
NU f 11 ), [15]. fuzzy vector K is none other than the abscissa of the
x i

center of gravity of the overall surface S K swept by the


Thus, the total number of fuzzy inferences is given by : xi

control vector K and deduced in accordance with the


N rgles N x .N x 121
x i

relationship.
t (3)

These membership functions are specified in the following


table:

Table 1 :
Membership functions
1 MF1 MFNTG
2 MF2 MFNG
3 MF3 MFN
4 MF4 MFNM Fig.4. Total surface swept by the fuzzy control vector K xi
5 MF5 MFNP 1

Rk .mfuncr ( K xi ).K xi .K xi
6 MF6 MFZE K xnfi 1
1

7 MF7 MFPP Rk .mfuncr ( K xi ).K xi


1 (5)
8 MF8 MFPM nf
Where : K xi is an un-fuzzy vector.
9 MF9 MFP
10 MF10 MFPG 2.2New Rescaling Technique
A new technique of scaling will be of great importance for
11 MF11 MFPTG the magnitudes of the state variables that may possibly
exceed the extreme limits quoted. In other words, all sizes
to be treated xi giving rise to an error x and a variation of
i

129
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

error x . Must undergo a transfer at the base of fuzzy


i 3ELECTRIC VEHICLE MODEL
t The model of the electric vehicle studied is illustrated in
variables evidenced by the interval [-1,1] , to generate the figure 2. We focus mainly on the components of the power
fuzzy input required for processing by the designated system that provide electric traction namely[11], [14]:
controller in accordance with the following system of The photovoltaic panel,
equations [9]: The buck converter,
The battery,
xf B . x [1 B ] The DC-Link,
i i
The DC/AC converter,
xf x
i B . i [1 B ] The filter,
t t
t t The permanent magnet synchronous machine (PMSM).
2
B ss
Gmax [mfuncr ( x )] Gmin ss
[mfuncr ( x )]

i i

B 2

t Gmax ss
[mfuncr ( xi )] Gmin
ss
[mfuncr ( xi )]
t t (6)

Where:

ss
1 ( x )2
Gmax ( mfuncr ( xi )) max .exp i 2

. 2 . 2.


ss
1 ( xi )
2

Gmin ( mfuncr ( xi )) min .exp
2. 2

. 2.


xi

( )2
G ss ( mfunc ( xi )) max 1
.exp t
max
r
t
. 2. 2
. 2




Fig.5. Model of the studied electric vehicle

)2
xi
(
G ss ( mfunc ( xi )) min 1
t
min . exp
r
t . 2. 2. 2
4 PHOTOVOLTAIC GENERATOR




We start with a reminder of the electrical equations of the
(7)
integrated photovoltaic generator. We highlight the model of
the basic component of this generator, namely the
Where i i denote respectively the mean and standard photovoltaic cell with a real model.
deviation of the error and variation of error x and x . i
i
t 4.1 Real Model of a Photovoltaic Cell
Similarly, the control vector K must undergo a transfer to In order to highlight the physical phenomena that occur in
xi
the photovoltaic cell, we must consider:
the basic quantities studied, rescaled, to assign the value The leaks resulting from the edge effects of the junction
K xnfi and this means the system of equations: PN and this by the integration of a shunt resistor Rsh in
the equivalent scheme of the cell,
K xnfi BK n f .K xi [1 BK n f ] The losses due to the contacts and the connections and
xi xi
which will be modeled, in the equivalent diagram, by a
G [mfuncr ( K xi )] Gmin
ss ss
[mfuncr ( K xi )]
BK n f max series resistance Rs .
xi 2 (8)

Where:

( K xi )2
ss 1 K

Gmax [ mfuncr ( K xi )] max . 2. . exp
xi

2
. 2
K
i x K
x i

( K xi )2
G ss [ mfunc ( K )] min 1
. exp
K
xi

min r xi
K . 2. 2. 2

xi K
x i
(9) Fig.6. Real modelof a photovoltaic cell

Through this new technique of rescaling, we assigned a


dynamic behavior to the fuzzy controller in order to ensure
better tracking of the variable to control [9].
130
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

The current supplied by the photovoltaic cell is then:

V p VRs .I p V p Rs .I p
I p I ph I s .e T 1 (10)
Rsh

The current I cc supplied by the photovoltaic cell when the


latter has a zero voltage to these terminals ( V p 0V ):

RsV.TI cc Rs .I cc
I cc I ph I s .e 1 (11)
Rsh

The voltage Vco delivered by the photovoltaic cell when the


latter does not discharge any current ( I p 0 A ) is: Fig.8. Scheme of a photovoltaic generator

The current supplied by the photovoltaic generator is


I V expressed as follows:
Vco VT .Ln ph co 1 (12)
Is Rsh .I s V p Rs .I p V R .I
I p n p .I ph n p .I s .e ns .VT 1 p s p
(14)
The optimum power Pop delivered by the photovoltaic cell is Rsh

defined when it delivers an optimum current I op under an


Taking into account that:
optimum voltage Vop :

Rs 0
Pop Vop .I op (15)
Rsh
(13)

then:

V p R s .I p

I p n p .I ph n p .I s . e ns .VT
1 (16)


The photovoltaic generator then delivers energy to the rest
of the power circuit of the electric vehicle as shown in figure
9.

Fig.7. Current voltage characteristic of a photovoltaic cell

4.2 Model Of A Photovoltaic Generator


We present in the following the developed scheme of the
model of a photovoltaic generator, figure 8.

Fig.9. Electric vehicle battery charging system

5 BUCK CONVERTER
In order to ensure a conditioned transfer of the energy
supplied by the photovoltaic generator ( Vp , Ip ) to the rest
131
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

of the power system of the electric vehicle, we have soon as the terminal ends are closed in an electrical
integrated a buck converter. In fact, by checking the duty circuit,chemical reactions start to take place in the cell,
cycle buc of this converter, according to the requirement of causing the flow of a current. However,due to the charge
the PMSM motor, we identify the level of the requested transport in the electrolyte, and the chemical reactions at
power of the photovoltaic (PV) generator. Let Tbuc be the the surfaceof the plates and the current in the plates, there
is a resistance to the current, which iscalled the battery's
commutation cycle of the buck converter. From 0 to internal resistance, R. A simple circuit model of a battery is
buc.Tbuc we have: depicted in Figure 10, where the OCV isdepending on the
SOC, and the resistance is constant. There are, however, a
number offactors that are not included in this simple model,
dI p such as the charge accumulation atthe plates, which gives
Vp Lbu . Vsoc (17) capacitive contributions to the resistance, SOC and
dt temperaturedependence of all parameters (OCV, R and C),
and finally a self-discharge of the battery,that can be
buc.Tbuc to Tbuc we have:
modeled as a shunt resistance to the OCV.
From

dI p
0 Lbu . Vsoc (18)
dt
And thus:

Vsoc
Ip .( t Tbuc ) I pmin (19)
Lbuc
On the other hand :

Vsoc
I pmax .( buc 1 ).Tbuc I pmin (20)
Lbuc Fig.10. Electric vehicle battery

7 BOOST CONVERTER MODEL


The current I p ripple noted I p is evaluated through the The energy delivered by the storage battery ( Vb , I b )must
relationship: meet the requirements imposed instantly, namely:
Maintain a DC link voltage level Vdc required by the
1 buc
I p I pmax I pmin .Vsoc (21)
operating point of the PMSM motor,
Determine the level of desired reference voltages (
Lbuc . fbuc * *
vdc , vqc ) that the converter C PMSM must supply.
Where :
1
f buc
Tbuc
Thus, we conclude that the average value of the voltage at
the terminals of the storage battery Vsoc is expressed as a
function of the voltage applied by the photovoltaic generator
V p in accordance with the following relation:

Vsoc buc .Vp (22)

6 BATTERY MODEL
The charge separation that takes place in each battery cell
give rise to a cell voltage, orOpen Circuit Voltage, OCV. As Fig.11. Boost converter
132
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

For these and other reasons, we have integrated a boost


(step-up) converter with a duty cycle boc and a 2 t pas
Vdc . ( Pboc Pm ).dt
commutation cycle Tboc . Cdc t
(26)

Referring to the figure 11, we write: Where pas is the step of calculation.

1
Vdc .Vb (23)
8 PERMANENT MAGNET SYNCHRONOUS MACHINE
1 boc (PMSM) MODEL
The topology of the permanent magnet synchronous
machine (PMSM) is illustrated in figure 14.By performing
As already2 reported, Any operating point of the PMSM
the transform Park to stator windings with respect to a
motor automatically imposes a DC-Link voltage level Vdc reference associated with the rotor of the machine, we
Therefore, we relied on the control of the duty cycle boc . obtain a bipolar fictitious machine with axes ( d, q ),[14],
For this purpose, a fuzzy controller evaluates an [17].
instantaneous reference of this duty cycle boc
*
as a function
*
of a desired level of the voltage reference level Vdc , figure
12.

Fig.14.Permanent magnet synchronous machine (PMSM)


relating to (d, q) axes.

Fig.12. Fuzzy control of the DC link voltage The magnetic flux f created by the permanent magnet at
the rotor is oriented along the d-axis. Thus, the components
Referring to the above figure, the reference voltage Vb* That
of the voltage at the terminals of the stator of the PMSM
the battery must supply is given by : machine are given by the following system:

Vb* ( 1 boc
*
).Vdc (24) ddm
vdm Rs .idcm .qm
dt
Under these conditions, the level of the charge voltage Vsoc (27)
v R .i dqm
of the battery is given by:
.dm

qm s qcm
dt
Vsoc Rb .I b Vb*
(25) The components of the fluxes along the ( d, q ) axes are
Vsoc buc .V p given by the following system:

We thus go back in order to estimate the duty cycle buc of dm Ld .idcm f


the buck converter as shown in figure 13, [15]. (28)
qm Lq .iqcm
Taking into account the fluxes equations dm and qm , we
write :
didcm
vdm Rs .idcm Ld . .Lq .iqcm
dt
Fig.13.Model of duty cycle buc control (29)
v R .i L . diqcm .( L .i )
In figure 13, the DC-Link voltage level Vdc is evaluated by qm s qcm q dt d dcm f

taking into account the transit powers Pboc and Pm :


133
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

On the other hand, the dynamics of the machine is


We then obtain two equivalent schemes along the ( d, q )
described by the following model:
axes, figures 15 and 16.

d Tem Tr . f r
dt Jm
(34)
d
dt

Where : Tr Is the resistive torque, f r coefficient of friction,


J m moment of inertia of the rotating mass and is the
Figure 16 Equivalent scheme of the PMSM machine
according to the q-axis angular position. We have adopted the method of Runge
Kutta of order 4 to evaluate the components of the current
(idcm , iqcm ) Provided by the converter CPMSM .

We write then:

vdm Rs .idcm .Lq .iqcm


idcm Fdcm( idcm ,iqcm ,t )
Ld
(35)
i vqm Rs .iqcm .( Ld .idcm f ) F ( i ,i ,t )


qcm
Lq
qcm dcm qcm

Fig.16.Equivalent scheme of the PMSM machine according


to the d-axis Therefore:

The active and reactive powers exchanged by the machine 1


i (n 1) idcm (n) . pas.(rkdcm1 2.rkdcm 2 2.rkdcm 3 rkdcm 4 )
are: dcm 6
(36)
i (n 1) i (n) 1 . pas.(rk
qcm1 2.rkqcm 2 2.rkqcm 3 rkqcm 4 )

qcm qcm
6

Pm vdm .idcm vqm .iqcm



(30) With:

Qm vqm .idcm vdm .iqcm


rkdcm1 Fdcm (idcm , iqcm , t )

rk 1 1 1
Fdcm (idcm . pas.rkdcm1 , iqcm . pas.rkqcm1 , t . pas )
dcm2 2 2 2
We also write : (37)
rk 1 1 1
Fdcm (idcm . pas.rkdcm2 , iqcm . pas.rkqcm2 , t . pas )
dcm3

2 2 2
rk
dcm4 Fdcm (idcm pas.rkdcm2 , iqcm pas.rkqcm2 , t pas )
Pm .( dm .iqcm qm .idcm ) (31)
p And:

Consequently, the electromagnetic torque


Tem is then: rkqcm1 Fqcm( idcm ,iqcm ,t )

rk 1 1 1
Fqcm( idcm . pas.rkdcm1 ,iqcm . pas.rkqcm1 ,t . pas )
qcm2
dm .iqcm qm .idcm
2 2 2
(38)
1 1 1
Tem (32)
rk
qcm3
Fqcm( idcm . pas.rkdcm2 ,iqcm . pas.rkqcm2 ,t . pas )
2 2 2
rk
p qcm 4
Fqcm
( idcm
pas .rkdcm 2
,iqcm
pas .rkqcm 2
,t pas )

In order to plan the process of numerical integration of the 9 QUADRATIC CONTROL OF THE DC LINK VOLTAGE
equations of state of the PMSM machine and by referring to LEVEL
the system of equations (33), we write: The level of the DC bus voltage is conditioned by the power
Pdc transited from the boost converter Pboc
didcm vdm Rs .idcm .Lq .iqcm
balance

dt to the CPMSM converter Pm ,we write then:


Ld
(33)
diqcm vqm Rs .iqcm .( Ld .idcm f )

dt Lq
134
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

The notion of quadratic control comes from the fact that we


treat the square of the DC bus voltage Vdc . Indeed, the
Pdc Pboc Pm (39) control strategy is divided into two blocks namely a
calculation block and a control block. In order to attribute to
this strategy a better dynamic of pursuit of the desired
reference Vdc*2 , we have opted for a fuzzy control, figure 17.
knowing that :

Pdc Vdc .I dc

1 (40)

boc 1 .Vb .I boc
P
boc

The fluctuations of the power Pdc are evaluated according


to the following relation:

Pdc P
Pdc .Vdc dc .( Iboc I m ) (41)
Fig.17. Calculation and control models of the DC-link

Vdc I dc
voltage

10 SVPWM STRATEGY APPLIED TO THE CPMSM


We then find:
CONVERTER
In this section, we are interested in developing the space
vector pulse width modulation strategy (SVPWM) applied to
the CPMSM converter, figure 18.
Pdc P
Pdc .Vdc dc .( Iboc mcd .idcm mcq .iqcm ) (42)
Vdc I dc

Where mcd , mcq are the index modulations of the

CPMSM converter. The power involved in the DC bus is:

1 d ( Vdc2 )
Pdc .Cdc . (43)
2 dt
then : Fig.18. model of the CPMSM converter
The output voltages of the C PMSM converter are
t pas
2 expressed as follow:
Vdc2 . Pdc .dt (44)
Cdc t di
vdcm R f .idcm L f . dcm .L f .iqcm vdm
dt
(46)
Which leads to:
v R .i L . diqcm .L .i v


qcm f qcm f
dt
f dcm qm

2 t pas
Vdc . ( Pboc Pm ).dt (45) R f , L f are respectively the resistance and the inductance
Cdc t of the filter.

We write likewise:

135
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

vdcm mcd .Vdc x , x


dp qp (d,
(47) q)-axiscomponents of the positive
vqcm mcq .Vdc sequence in the PSRF frame,

Also : x , x

dn qn (d, q)-axiscomponents of the negative

I m mcd .idcm mcq .iqcm



sequence in the NSRF frame,

(48) Thus, the fluctuations of the components of the current

I m f cm ( idcm ,iqcm )
supplied to the machine are given by:

i
Current fluctuations I m at the input of the CPMSM dcm 2..i .sin(2. ) 2..i . cos(2. )
t dcmn qcmn
converter are written as follows:
i (51)
vdcm vqcm qcm 2..i .sin(2. ) 2..i . cos(2. )
I m .idcm .iqcm (49) t qcmn dcmn
Vdc Vdc *
The components of the reference voltage ( vdm *
, vqm )

Taking into account the Fortescues theory based on applied to the terminals of the PMSM machine are
decomposition into a positive synchronous reference frame evaluated in accordance with the following block diagram:
(PSRF) and a negative synchronous reference frame
(NSRF), figure 19, We implement control loops of currents
and voltages.

Fig.20. Voltage control loops at the terminals of the PMSM


machine

We define the variations of the deviations idcm , i , and


qcm
Fig.19.Positive and a negative synchronous references
frame (PSRF, NSRF) V dc
by the following relations:

The current components (idcm , iqcm ) provided by the


id cm .sin( 2. ) 2..i . cos(2. )
converter CPMSM will be decomposed according to the 2..idcmn

qcmn
t
following model.
iq cm .sin( 2. ) 2..i . cos(2. )
2..iqcmn
t
dcmn
i i i .cos( 2. ) i
dcm dcmp dcmn
. sin( 2. ) (52)
Tem Tr . f r
qcmn

i i i .cos( 2. ) i . sin( 2. )
(50)
t Jm
qcm qcmp qcmn dcmn
Vd c Pdc
Where:
t Cdc.Vdc
x , x

d q (d, q)-axiscomponents in the PSRF frame,

136
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

The vectors Sf dcm and Sf qcm generated by the fuzzy


controllers integrated in the (d, q) axis current control loops
are defined as follows:

di
Sf dcm Rs .idcm Ld . dcm
dt
(53)
Sf R .i L . diqcm


qcm s qcm q
dt
* *
The references of the components vdm and vqm of the
voltage at the terminals of the PMSM machine being
obtained and taking into account the existence of the (
R f , L f ) filter of the low-pass type,which completely
eliminates all frequencies above the cutoff frequency while
passing those below unchanged. We then constitute the
* *
components of the reference voltage v dc and v qc that the
converter must generate:
Fig.21. SVPWM technic

vdc R f .idc .L f .iqc vdm


* * * * With referring to Figure 21, we write:

* (54)

vqc R f .iqc .L f .idc vqm


sin .( R f .idc* .L f .iqc* v*dm ) cos .( R f .iqc* .L f .idc* v*qm )

* * *
tan g
1

cos .( R f .idc .L f .iqc vdm ) sin .( R f .iqc .L f .idc vqm )
* * * * * *
(57)

angle( vc1 )
* *
The components of the reference voltage vdc and vqc are
The distribution of the states of the keys K1 K 2 K 3 as well as
subsequently transferred to a fixed reference frame ( , )
in order to apply the SVPWM technique. their application times T , over a modulation period TH are
illustrated in Figure 22.
*
vc
e j . .t
.v *
dqc
(55)

From where:

v* c cos .( R f .idc* .L f .iqc* v*dm ) sin .( R f .iqc* .L f .idc* v*qm )

* (56)
vc sin .( R f .idc .L f .iqc vdm ) cos .( R f .iqc .L f .idc vqm )

* * * * * *

Fig.22.Distribution of key control times


It is then necessary to locate the sector of belonging of the
desired reference voltage vector v c which the
*
CPMSM The keys K1 K 2 K 3 application times in a sector i are
defined as follows:
converter must supply, figure 21.

*
v
Ti 2 .
c
.TH . sin[ i . ]
Vdc 3
(58)
T 2 . vc .T . sin[ ( i 1 ). ]
*



i 1
Vdc
H
3

Therefore, the average value of the voltage vc


supplied by the CPMSM converter is:

137
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

1 1 Vdcm ( n ) .L f .I qcm ( n ) Vdm ( n )


I dcm ( n ) 2. . f .T 1 .I dcm ( n 1 ) Lf
.
1
c e 2. . f c
1 j.

Te
2 j .( i 1 ). 2 (65)
vc .T0 .vc 0 Ti . .Vdc .e 3 Ti 1 . .Vdc .e 3 (59) I ( n )
qcm
1
.I qcm ( n 1 )
1 Vqcm ( n ) .L f .I dcm ( n ) Vqm ( n )
.
TH 3 3 2. . f c .Te 1 1

Lf 2. . f c

Te

Where Te is the sampling period.


Where :
12 PHASE LOCKED LOOP (PLL)
T Ti Ti 1 The transformations of Park, by means of the use of the
matrix P , Of axes (d, q) that we have established, Were
T0 H (60)
4 based on arguments perfectly identical to those of the
systems of voltages (currents) studied, ( Vic , I icm ).The aim of
This voltage will be transferred to the (d, q) reference frame this approach was to reconstitute reference variables,
according to the following relation: assigned to the control loops, which were extremely
precise, and on the other hand to aim at a perfect
synchronization of the procedures for digital processing of
1 2 j .( i 1 ).

2 j.

the various state variables.
vdqc e j . . .T0 .vc 0 Ti . .Vdc .e 3
Ti 1 . .Vdc .e 3 (61)
TH 3 3
On the basis that:
11 TECHNIQUE DE FILTRAGE

.Vc max .cos(argP arg( vic ( t ))
3
vdc
In order to improve the quality of the energy exchanged
between the CPMSM converter and the PMSM machine, we 2
have integrated a low pass filter ( R f , L f ). The filter cut-off
(66)
v 3 .V . sin(argP arg( v ( t ))
frequency ( f c ) is given by:

qc
2
c max ic

In these circumstances, we must ensure that:


Rf
fc
argP arg( vic (t ))
(62)
2. .L f (67)

By application of the Laplace transform, the currents at the It then follows:


output of the ( R f , Lf )filter are expressed as follows:

.Vc max .(argP arg( vic ( t ))


3
Vdcm Vdm .L f .I qcm vqc (68)
I 2
R f p.L f
dcm

(63)
I Vqcm Vqm .L f .I dcm
Therefore, the phase locked loop (PLL) must guarantee the
digital procedure established in figure 23.
qcm R f p .L f

Numerically one writes:
Lf Vdcm ( n ) .L f .I qcm ( n ) Vdm ( n )
I dcm ( n ) R .T L .I dcm ( n 1 ) L
f e f
Rf f
Te
(64)
I ( n ) Lf Vqcm ( n ) .L f .I dcm ( n ) Vqm ( n )
.I qcm ( n 1 )
qcm R f .Te L f L
Rf f
Te
Fig.23. Phase locked loop (PLL)
Or :
13 MODEL OF ELECTRIC VEHICLE TRACTION (EVT)
The model of the electric vehicle traction that emphasizes
the different resistive torques Tsrm , Tsrw , Tw opposed to the

138
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

electromagnetic torque Tem developed by the PMSM motor This leads to:
d m
is illustrated in figure 24. Tem ( Dm r 2 .Dw ). m ( J m r 2 .J w ). r .Tw
dt (72)

Finally:

d m
Tem Dtm . m J tm . r .Tw (73)
dt
where:

Dtm Dm r 2 .Dw
(74)
J tm J m r .J w
2
Fig.24.Chain of traction

Where:
The mechanical torque T w applied to the wheel of the
Tsrm is the speed reducer torque of the PMSM side, electric vehicle is given by the relation:
Tsrw is the speed reducer torque of the wheel side,
Tw is the wheel torque. Tw Fr .Rw (75)

In order to clarify the distribution of these different torques


on the electric vehicle's shaft, we have used an analogy
with the single-phase electric transformer, figure 25. Rw The radius of the EV wheel.
Fr The resistive force (N),
On the other hand, the road and the quality of the coating
have resistances external to the advance of the vehicle on
a longitudinal plane. All the forces at the advancement can
be represented by a single force of the second order: Under
these conditions, the dynamics of the PMSM machine will
be governed by the following mathematical model:
Fig.25. Equivalent model of mechanical torque transfer

d m Tem r .Tw Dtm . m


By referring to the equivalent model of figure 25, we write:


d m dt J tm
T
em D . J . Tsrm (76)
d m
m m m
dt
(69)
T D . J . d w T
dt
m

srw w w w
dt
w

With:
With:
m , p is the number of pair poles.
w r . m p
(70)
Tsrm r .Tsrw 14 THE FORCES RESISTING MOVEMENT OF THE
r Is the speed reduction ratio. ELECTRIC VEHICLE
In order to evaluate with great precision the electromagnetic
torque which the motor of the electric vehicle (EV) must
Then:
develop, to overcome the total resisting force Fr opposed
d m d ( r . m )
Tem Dm . m J m . r .( Dw .r . m J w . Tw ) to the movement of the vehicle, we must establish the
dt dt (71)
139
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

balance of all the forces Fr F ri


likely to act during the Where :
i
M EV The mass of the vehicle,
movement of this vehicle. In effect, a moving vehicle is
subjected overall to a force opposed to the penetration into g The acceleration of gravity,
the air Frap , to a force of opposition to the rolling Frrol and k1rrol , k2 rrol Coefficients of rolling resistance.
possibly to a force of opposition to the rise in a slope Frst .
14.3 Slope force
14.1 Force of penetration into air This force Frst is created at times when the vehicle attacks
This force Frap is created by the mass of air that the vehicle a slope. In fact, it depends on the weight of the vehicle PEV
in motion wishes to pass through. and the angle of inclination of the road p used by the
vehicle.

Fig.26. Effect of the force of penetration into the air

The force required to penetrate the air Frap , it is calculated


by applying the relation:
Fig.28. Effect of the sloping terrain force

1
Frap . EV .S .Cx .VEV2 (77) The force to be defeated on a sloping terrain Frst is
2 calculated by mean of the following relation:

Where :
Frst PEV . sin( p ) (79)
EV The density of the vehicle,
S The surface taken to the wind,
Cx Where PEV M EV .g .
Coefficient of penetration into air,
VEV 14.4Centrifugal Force
Electric vehicle speed.
In Newtonian mechanics centrifugal force Fc is a force of
14.2 Rolling resistance force inertia directed outside the axis of rotation which seems to
act on all objects when they are observed in a rotating
This force Frrol is applied to the vehicle as a result of the
reference frame. The concept of centrifugal force can be
contact of these wheels with the road followed. applied to vehicles in rotational motion according to a
certain curvature.

Fig.27. Effect of the rolling resistance force

The rolling resistance force Frrol calculated by applying


the relation:

Frrol M EV .g .( k1rrol k2 rrol .VEV2 )


Fig.29. Centrifugal force applied to anelectric vehicle
(78)

140
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

Its intensity is given by the formula:


F P .k
M EV .VEV2 r 0 VE 1rrol

Fc (80) r1 p ) PVE . sin( p )



F ( (86)
Rc Fr 2 .( VVE2 ) ( 1 .VE .S .Cx PVE .k2 rrol M EV ).VVE2
2 Rc
Rc Is the distance from the axis of rotation to the center of
gravity of the electric vehicle (EV), i.e. the radius of Therefore, we write the total resistant force opposed to the
curvature of the path in meters (m).This reaction force Fc is movement of the vehicle is three-dimensional:
sometimes described as a centrifugal inertial reaction [37]

Fr fVE ( PVE , p ,VVE2 )


[38] that is a force which is centrifuged, which is an equal
and opposite reactive force to the centripetal force which is (87)
the curvature of the path.
The dynamics of the electric vehicle brought on the motor
shaft then becomes:

d m Tem r .Rw . fVE ( PVE , p ,VVE2 ) Dtm . m



dt J tm
(88)
d m

dt
m

Then:

d m Tem r .Rw . fVE ( PVE , p ,VVE2 ) Dtm . m



dt J tm
Fig.30. Forces applied in a curved path to anelectric vehicle
d m 1 .V
(89)
By referring to the previous figure, we write:
dt r .Rw
VE

VEV2
tg ( )
Taking into account the link between the linear speed of the
(81) electric vehicle VVE and that of the wheels w :
Rc .g
Thus, the total resultant force Fr applied to a moving vehicle VVE Rw . w
in the different types of roads is given by the relation:
(90)
w r . m
Fr Frap Frrol Frst Fc (82)
We obtain:
Then:
1 M
Fr 2 .( m2 ) r 2 .Rw2 .( .VE .S .Cx PVE .k2 rrol EV ). m2 (91)
1 M .V 2 2 Rc
Fr .VE .S .Cx .VVE2 M VE .g .( k1rrol k2rrol .VVE2 ) PVE .sin( p ) EV EV (83)
2 Rc
15 SIMULATIONS AND RESULTS
Finally: We simulated the behavior of an electric vehicle with the
following parameters:
1 M
Fr PVE .k1rrol PVE . sin( p ) ( .VE .S .Cx PVE .k2 rrol EV ).VVE2
2 Rc
(84) VE 1.3Kg.m 3 The density of the vehicle,
This total resistive force can be expressed as follows:
S 1.2m 2 The surface taken to the wind,
C x 0.29 Coefficient of penetration into air,
Fr Fr 0 Fr1( p ) Fr 2 .( VVE2 ) (85) VVE Electric vehicle speed.
M VE 1400Kg The mass of the vehicle,
Where: g 9.81m.s 2 The acceleration of gravity,

141
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

k1rrol 12,5.10 3 , k 2rrol 32,4.10 6 Coefficients of


rolling resistance.
r 0.92 the speed reduction ratio.
Rw 0.3m The radius of the EV wheel.

The parameters of the PMSM motor are:


stator resistance: Rs 0.4578 , d-axis inductance:
Ld 0.00334 H , q-axis inductance: Lq 0.00334 H
,flux: f 0.171weber ,inertia: J m 0.01469Kg.m 2 ,
Fig.33. Time evolution of the electromagnetique and
pair poles: p 4, nominal torque: Temn 20 N .m , resistive torques
friction: Dm 0.00334 N .m We simulated the case
We simulated the case where from the instant t 1.2s up
where from the instant t 1.2s up to t 5s the electric to t 5s the electric vehicle attacks a slope of 8% .
vehicle attacks a slope of p 300 .

Fig.31. rotational speed of the PMSM motor m Fig.34.Time evolution of m

During the rise in a slope the variations in the speed of the


PMSM are similarly insignificant.

Fig.32. Time evolution of electric vehicle speed

Once the electric vehicle EV starts the rise of the slope, the
associated regulators act so that the rotational speed of the Fig.35.Torque evolution during the rise in a slope of 8%
PMSM motor m catches up to its reference value m* in an
extremely short time. Therefore, the vehicle speed VEV The electromagnetic torque Tem perfectly compensates for
remains constant during the climb. This demonstrates the the resistive torque Tw resulting essentially from the sloping
robustness of integrated fuzzy control loops. rise p .

142
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

We notice that when the vehicle picks up speed V EV at the


beginning of the descent, the regulators engage a braking
strategy and we notice that the speed regains its reference
value. On the other hand, we observe the engagement of a
regenerative braking strategy because the torque becomes
negative and the motor returns electrical energy to the
battery.

Fig.36. Time evolution of the DC link voltage Vdc


The level of the DC bus voltage Vdc remains insensitive to
the torque rise of the electric vehicle, which proves that the
control strategies adopted have perfectly isolated the
engine power system from any abrupt change in the EV
vehicle operator mode.
Fig.39. Time evolution of Tem and Tw during a negative
slope 8%

Fig.37. Time evolution of current

The q-axis component of the current iqcm supplied by the Fig.40.Time evolution of idcm and iqcm during a negative
converter CPMSM acts in a notable manner to compensate slope 8%
for the resistive torque.This results from the fact that this
component is the most weighted dm.iqcm in the expression
of the electromagnetic torque Tem . We simulated the descent
of the electric vehicle with a negative slope of 8% .

Fig.41. Time evolution of VEV during a negative slope 8%

Similarly, during the descent into a ramp, the fuzzy control


strategies act in such a way as to attenuate the excesses of
certain state variables. We simulated the behavior of the
Fig.38.Time evolution of m during a negative slope 8% electric vehicle the moment it takes a turn, that is to say
when it is subjected to the centrifugal force Fc which pushes
it towards the outside of the curve.

143
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

Guillermo, MOURE Christophe, BOLTSHAUSER


Sandro
Volume 6 - Issue 1 - Pages 88-94.

[2] DESIGN CONSIDERATIONS FOR FAST AC


BATTERY CHARGERS
BERTOLUZZO Manuele, BUJA Giuseppe, PEDE
GIOVANNI
Volume 6 - Issue 1 - Pages 147-154, World Electric
Vehicle Journal Vol. 6 - ISSN 2032-6653 - 2013
WEVA.
Fig.42. Time evolution of VEV during a curved path with a
[3] M. ISMAIL, M. Mustafa HASSAN Load Frequency
curvative radius Rc 100m Control Adaptation Using Artificial Intelligent
Techniques for One and Two Different Areas Power
System International Journal of Control, Automation
and Systems IJCAS, Vol. 1, NO. 1, january 2012, p12-
p23.

[4] A. M.A.Haidar,A. Mohamed and A.


Hussain,Vulnerability control of large scale
interconnected power system using neuro-fuzzy load
shedding approach,Expert Systems With Applicatons,
Elsevier 37.pp. 3171-3176, 2010.

[5] MOTOR-GENERATOR CONTROL TO IMPROVE


SHIFT QUALITY FOR A DUAL MODE POWER SPLIT
TRANSMISSION
Figure.43.Time evolution of VEV during a curved path with HONG Sungwha, CHOI Woulsun, AHN Sunghyun,
a curvative radius Rc 10m LEE Geontae, SON Hanho, KIM Yongjoo, KIM
Hyunsoo
Volume 6 - Issue 2 - Pages 245-250, World Electric
During the turning, the control strategies fuzz act to reduce
Vehicle Journal Vol. 6 - ISSN 2032-6653 - 2013
the excess speed and this by an evaluation of new
WEVA .
references ensuring an extreme degree of safety.
[6] Alireza Alfi, "Hybrid state-feedback sliding-mode
16 CONCLUSION controller design using fuzzy logic for four-wheel-
In this work, we are interested in the implementation of steering vehicles ", Vehicle System Dynamics, vol. 47,
intelligent strategies based on fuzzy controllers to ensure Issue. 3, pp. 265-284, 2009.
optimal operation of an electric vehicle.we can say that the
design and optimization of the traction chain of an electric [7] Alireza Alfi, "Swarm optimization tuned Mamdani fuzzy
vehicle is a multidisciplinary problem that must take into controller for diabetes delayed model ", Turkish
consideration, at a minimum, batteries, mechanical Journal of Electrical Engineering and Computer
transmission and electromagnetic powered electronics. The Sciences, vol. 21, pp. 2110-2126, 2013.
search for a minimal cost is added to the scientific and
technological difficulties. The future electric vehicle must be [8] Alireza Alfi, "Optimal design of type-2 fuzzy controller
able, with appropriate control, to use the energy stored in using particle swarm optimization for HVAC systems",
the batteries in order to sustain the electricity network Automatika-Journal for Control, Measurement,
during periods of peak consumption or in case of Electronics, Computing and Communications, vol. 55,
emergency (shutdown of a production plant) . The energy vol. 1, pp. 69-78, 2014.
stored in the battery of the vehicle could also supply the
electrical requirements of the dwelling. This technology [9] D. Nazari, M. Abadi, M. H. Khooban, A. Alfi, M. Siahi
requires that the charger embedded in the vehicle as well "Design of optimal self-regulation Mamdani-type fuzzy
as the interface between the vehicle and the electrical inference controller for type 1 diabetes mellitus",
network are bidirectional. The results obtained are Arabian Journal for Science and Engineering, vol. 39,
extremely encouraging and prove that the more we refine pp. 977-986, 2014.
the electromagnetic model of the electric vehicle the more
we obtain a favorable ground for a better implementation of [10] M. H. Khooban, A. Alfi and Abadi, "Control of a class
the strategies of control especially the fuzzy strategies. of nonlinear uncertain chaotic systems via an optimal
type-2 fuzzy PID controller", IET Science,
REFERENCES Measurement & Tecnology, vol. 7, no. 1, pp. 50-58,
[1] CONTROL SYSTEMS FOR HIGH PERFORMANCE 2013.
ELECTRIC CARS
PEDRET Paula, WEBB Jonathan, , BAYONA
144
IJSTR2017
www.ijstr.org
INTERNATIONAL JOURNAL OF SCIENTIFIC & TECHNOLOGY RESEARCH VOLUME 6, ISSUE 01, JANUARY 2017 ISSN 2277-8616

[11] M. Hassan Khooban, Alireza Alfi, Davood Nazari and


Maryam Abadi, "Teaching learning based optimal
interval type-2 fuzzy PID controller design: A
nonholonomic wheeled mobile robots", Robotica, vol.
31, no. 7, pp. 1059-1071, 2013.

[12] A. Alfi, A. Akbarzadeh Kalat and M. Hassan Khooban ,


Adaptive fuzzy sliding mode control for
synchronization of uncertain non-identical chaotic
systems using bacterial foraging optimization Journal
of Intelligent & Fuzzy Systems 26(2014) 2567-2576,
DOI: 10.3233/IFS-130928 IOS Press.

[13] Oscar Castillo, Patricia Melin, Jausz Kacprzyk, Wiltold


Pedrycz, Type 2 Fuzzy Logic : Theory and
Applications, IEEE International Conference on
Granular Computing, 2007.

[14] IMPACT OF ELECTROMOBILITY ON AUTOMOTIVE


ARCHITECTURES
LUCCARELLI Martin, RUSSO SPENA Pasquale,
MATT Dominik
Volume 6 - Issue 1 - Pages 1-8, World Electric Vehicle
Journal Vol. 6 - ISSN 2032-6653 - 2013 WEVA .

[15] H. Jouini, K. Jema, and S. Chebbi,Voltage Stability


Control of Electrical Network using Intelligent Load
Shedding Strategy based on Fuzzy Logic, Hindawi
Publishing Corporation, Mathematical Problems in
Engineering, Volume 2010, Article ID 341257, 17
pages.

[16] K. Jema, New Fuzzy Approach for Modeling


Random Disturbances, Journal of Control Science
and Engineering, Volume 2, Number 1, P 16-27,
February 2014 (Serial Number 2).

[17] A. Ouederni, K. Jema, H. Trabelsi, "Modelling and


Control of a PMSG-Based Variable Speed Wind
Turbine under Unbalanced Conditions", International
Journal of Research and Reviews in Electrical and
Computer Engineering (IJRRECE), Vol. 2, No. 1,
March 2012, ISSN: 2046-5149.

[18] IMPLEMENTING ELECTRIC VEHICLES IN URBAN


DISTRIBUTION: A DISCRETE EVENT SIMULATION
LEBEAU Philippe, MACHARIS Cathy, VAN MIERLO
Joeri, MAES Guillaume
Volume 6 - Issue 1 - Pages 38-47, World Electric
Vehicle Journal Vol. 6 - ISSN 2032-6653 - 2013
WEVA.

145
IJSTR2017
www.ijstr.org