Energetic Macroscopic Representation based modeling

and control for battery/supercapacitor hybrid energy

storage system in HEV

I. INTRODUCTION
Electric vehicles are powered by renewable and more efficient energy sources, such
as: Fuel Cell, Supercapacitor and Li-ion Battery. Each source has advantages and
disadvantage. FC has the highest energy density and the best efficiency, but it is not safeful
and isn’t be used in fact. Battery has high energy density, so electric vehicle can move a long
distance with the energy be provided by battery. However, the power density is very low not
enough to response the requests for acceleration. In the other hand, the energy density of
supercapacitor is low but the power density is very high. Therefore, supercapacitor can satisfy
the acceleration process. To take advantage of strong points and overcome weak points,
batteries are combined with supercapacitors into a hybrid energy storage system. This
combination helps energy consumption more effect.

To verify result of the combination between battery and supercapacitor, the modeling
and simulation are conducted. In control theory, we usually model an object by the transfer
function. Although it is easy to build the control structure for the object, it is very difficult to
help engineer having an overview of all object or system. In this report, Energetic
Macroscopic Representation (EMR) is suggested to model hybrid energy storage system.
EMR has introduced in 2000 for research development in complex electromechanical drives
especially drive systems. EMR is used widely in electric vehicles, hybrid electric vehicles,
trains and renewable energy generation system. Battery/Supercapacitor hybrid energy storage
system will be modeled and simulated based on EMR in this report. For an emphasis on the
modeling and control of hybrid energy storage system, a simple low-pass filter control
strategy is adopted. Then the supercapacitor state-of-charge (SOC) regulation is calculated to
have the supercapacitor system ready for the future events. The result of simulation will show
effect of the hybrid energy storage system.

1
II. EMR FUNDAMENTALS
EMR is not a controller design method. EMR is a modeling method, is built on
breaking a complex system into power-linked interconnected subsystems. EMR
supercapacitor schematic is based on four elements and two principles.

 Elements

There are four basic elements: source element, accumulator element, conversion
element and coupling element. The source elements (ovals) is terminal element which
generate or receive energy (e.g. a battery). The accumulator elements (rectangles) are
dynamic elements storing energy, such as: inductor store magnetic energy, capacitor store
electric energy. The conversion element is described by one square for mono-physical
conversion and one circle for multi-physical element. It is considered as energy conversion
without energy accumulator. Example, buck converters convert the voltage through one duty
cycle D: Uout=D ×Uin. The coupling element is introduced to distribute energy. It also
includes of two kind coupling elements: mono-physical conversion and mutil-physical
element. All elements are showed in Table 1.

Source element Accumulator element

Conversion element
Mono-physical element Mutil-physical element

y2

Coupling element
Mono-physical element Mutil-physical element

y2

y1
x3

y3
Table 1. elements of EMR.

2
  Principles

EMR is based on two principles: interaction principle and causality principle.

The interaction principle is power interaction between two sources S1 and S2 as Fig.1.
action
S1 S2
reaction

Power = action x reaction

Figure 1. interaction principle.

The causality principle relates to dynamic elements. Physical causality is integral, or

the output is integral of the input. For example to capacitors, the voltage is integral of
current:

simulation 1
Uc= ∫ ic dt
C
∫
Figure 2. causality principle.

III. STRUCTURE AND MODEL OF THE BATTERY/SUPERCAPACITOR

HYBRID ENERGY SYSTEM
In this report, the battery/supercapacitor hybrid energy system is adopted to improve
battery
the level of battery Itract
ibat working condition. The configuration of battery/supercapacitor hybrid
energy storage system is showed in Fig.3.
Ubattery Ubus load

supercapacitor inductor converter ich

isc

Usc Uconv

coupling

Figure 3. scheme
of battery/supercapacitor system.

3
 A DC/DC converter locates between supercapacitor pack and DC link whereas battery
pack is connected with DC link directly. Due to this report focuses on the management and
control the energy flow between supercapacitors and batteries, the load is simplified by a
current source. The current demands are satisfied by controlling charge and discharge states
of supercapacitors through controlling DC/DC converter.

A. The modeling of battery pack and supercapacitor pack

The battery is the main source in hybrid energy system. The electrical equivalent
circuit is showed in Fig.4. Two resistances (Rc and Rd) represent the charge and discharge.
They are seen as a function of temperature and current.
Rd
ibat
Rc

Ub Ubat

Figure 4. Battery equivalent circuit.

The battery SOC is calculated by the equation:

1
SOC=SOC o + i dt
3600 A bat ∫ bat
(1)

Where, Abat, SOCo, ibat are the battery capacity, the initial value of the battery SOC and current.

The supercapacitor is considered as an auxiliary source in hybrid energy storage

system. The electrical equivalent circuit is showed in Fig.5. there are three parameters: a
capacitance (C), a resistance Ri representing the charging and discharging resistance and a
resistance Rp representing the self-discharging losses.
Ri
+ isc
+
Usc Uc C Rp
-
-

Figure 5. supercapacitor equivalent circuit.

The voltage Uc depend on the capacitance C and the electrical charge Q. the electrical
charge Q will change according to the supercapacitor current Icell, as expressed in (2) and (3).

Q
U c= (2)
C

Q t +1=Q t −I cell . ∆t (3)

To ensure the energy storage capacity and required voltage Uc, the supercapacitor
system need many cells in series ns and in parallel np. this results in the total voltage Usc and

4
current Isc, seen (4) and (5). The internal resistance Ri is affected by the current and
temperature. Additionally, Rp represents the self-discharging losses and affects only the long-
term energy storage performance of supercapacitors. To simplify the model, its influence is
ignored.

U sc=ns .(U c −Ri . I cell ) (4)

I sc =n p . I cell (5)

B. Power electrics modeling

 Filter inductor L

The filter inductor has two important functions: store kinetic energy and filter output
voltage. Therefore, the inductor is represented by one accumulator element with tow inputs
(Usc and Uconv) and one output (IL). The mathematical model of inductor is showed below:

diL
U sc−U conv =L + R .i L (6)
dt

Where, L and R are inductance and internal resistance of filter inductor, Uconv is voltage of
electric converter, iL is current going through the inductor and it is equal isc.

 Electric converter

Because the functions of supercapacitor are to provide energy (discharge state) and
receive energy of generative branking, the converter is a bi-directional DC/DC converter. The
converter only convert voltage through a duty cycle m, so it is expressed by a converter
element.

(7)

Where, Ubus is DC bus voltage, Iconv represent the output current of converter, IL = Isc is current
of supercapacitor and m is duty cycle.

 Electric coupling

The electric coupling element represents the connection between battery pack,
supercapacitor bank with traction load part.
Ubatter Utract
y Itract U tract =U bus =U bat (8)
Ibatte
ry
{ I tract =I bat + I conv

Iconv Ubus

5
 C. Traction load

This report is only focused on the energy storage system, so the traction load is
assumed to be an equivalent current source.

After having EMR model of each component, the EMR modeling of

battery/supercapacitor hybrid energy system is established by linking blocks according to the
interaction principle. Finally, the EMR model is shown in the top half part in Fig.6. the rest
part below is the control structure built on inversion rule.

Ubat Utract
BAT
TER Ibat Itract load
Y

SUP inductor chopper Iconv system

ERC
APA Ubus
CIT
OR mref

IL_ref Uconv_ref

current
Control controller control
trategy

Figure 6. EMR model of battery/supercapacitor hybrid energy system

IV. CONTROL STRATEGIES AND METHODS

In this report, a control objective is to reduce the peak current and the Root Mean
Square (RMS) current of battery for several specific driving cycles. In the control structure
shown in Fig.6:
Usupercapac
itor_mes

Figure 7. PI control of
6 the inductor current
The current controller is a PI controller which is considered to control filter inductor current
following the setpoint IL_ref. The components of this control law are shown in Fig.7. IL_ref is
calculated based on power distribution control law between battery and supercapacitor.

To attain the reference input current of supercapacitor, this report uses the low-pass
filtering method. The ideal of method is to maintain the battery current following low-pass
filtered current and divert the supercapacitor through adjusting the filter time constant τ . The
reference input current of supercapacitor is calculated by formula (9).

I L =I tract ¿) (9)
ref

To ensure the vehicle runs effectively and prolong the battery life, there are two
restriction conditions. The first is current restriction, the current provided by hybrid energy
storage system should be equal to current which is required by the load:

I tract ( t )=I b ( t ) +m. I c ( t ) (10)

The second is SOC restriction. Both SOC of battery and supercapacitor should be within a
allowed limit, namely:

SOC bmin ( t ) ≤ SOC b ( t ) ≤ SOC bmax ( t ) (11)

SOC cmin (t) ≤ SOC c ( t ) ≤ SOC cmax (t) (12)

The reference input current of supercapacitor not only depends not only on the load
current, but also on supercapacitor SOC and the vehicle velocity. Both of them should be
taken into account. The supercapacitorheme of control strategy is shown in Fig.8. a PI
controller is used to determine a current request based on the error between the real and
reference SOC. The real SOC is estimated similar the fomula (1). The reference SOC is
assumed to be 80% when the vehicle speed is zero, while 30% at maximum speed.

Figure 8. scheme of control strategy

7
V. SIMULATION RESULTS
After building successfully, the EMR of hybrid energy storage system is given into
Matlab/Simulink environment to simulate. There are three urban cycles which are chosen to
validate the model: a modified ECE15-L Urban Cycle, NEDC and BeijingBus driving
cycles. The hybrid erergy storage system uses Ni-MH batteryteries and Maxwell
PC2500 ultra-capacitors. Parameters of batteries and supercapacitors are listed in Tabel 2.

Energy Sources
Parameters
Ni-MH Battery Supercapacitor
Capacity 27 Ah 2700 F
Voltage 1.2 V 2.5 V
Resistance 5 mΩ 1 mΩ
Weight 0.75 kg 0.725 kg
Cell number 250 120
Table 2. parameters of batteries and supercapacitors
Fig.9 shows the peak current delivered from battery with different filter time constants
over modified ECE-15-L urban cycle. When the filter time constant increases, the battery
charge and discharge currents decrease. If the time constant is low, the effect of sharing
battery peak current by supercapacitor will be weaken. Otherwise, high time constant will
increase cost and weight of cells. So τ = 10 is chosen in the following simulation.

Figure 9. the battery peak current with different τ value

In Fig.10 and Fig.11, the simulation results of hybrid energy storage system
with control strategy are illustrated both in NEDC and BeijingBus drive cycles. The
top graph is demanded vehicle speed; the mid graph is current characteristics of

8
battery, supercapacitor and load when the vehicle speed chances; the bottom graph is
SOC curves of battery and supercapacitor. Both in Fig.10 and Fig.11, the current graph
demonstrates the supercapacitors ability in separating the current segments changing
rapidly, which undertakes the burben of batteries. About the SOC profile, the
supercapacitor’s real SOC follows the target through maintaining supercapacitor
voltages and vehicle speed inversely.

Figure 10. Simulation results under NEDC drive cycle.

9
 Figure 11. Simulation results under BeijingBus drive cycle.

Finally, Table 3 shows the comparison of energy storage system with and
without supercapacitors. As indicate in this table, if the battery and supercapacitor is
combined, the peak and RMS current of battery are reduced largely.

Drive Energy Battery current (A)

cycles sources Peak value RMS value

B 74 -73.95 17
NEDC
B+UC 32.64 - 52.18 12.5

B 259.7 -172 30.6

BeijingBus
B+UC 79.08 -89.6 19.2

Table 3. The battery current before and after combining supercapacitor.

VI.
CONCLUSIONS
The combination between batteries and supercapacitors takes advantage of positive
points and overcomes negative points of each source. An EMR modeling is used to
investigated. A low-pass filter method and supercapacitor SOC control algorithm are
introduced. The hybrid energy system’s EMR modeling and control structure are simulated in
Matlab/Simulink enviroment. The simulation results show that the supercapacitors have
ability to reduce batteriespeak and RMS current on regenerative braking and satisfy the
acceleration by supply power suddenly. So, the battery lifetime is prolonged and the
efficiency of energy system is advanced.

However, many researchs remain to be done. After combining batteries and

supercapacitor, the issue about cost, size and control strategy optimization is considered.
Moreover, what happen if SOC b ≤ SOC bmin? In the future, the fuel cells will be considered to
replace batteries or all three sources (fuel cell, battery and supercapacitor) will be combined in
a hybrid energy storage system.

10