Smart LiFePO
4
battery modules in a fast charge
application for local public transportation
F. Baronti, R. Roncella, R. Saletti
Dipartimento di Ingegneria dell’Informazione
University of Pisa
Via Caruso 16, 56122 Pisa, Italy
r.roncella@iet.unipi.it
G Pede, F. Vellucci
Laboratorio Veicoli a Basso Impatto Ambientale
ENEA
Santa Maria di Galeria (Roma), Italy
giovanni.pede@enea.it
Abstract—This paper describes the research effort jointly
carried out by the University of Pisa and ENEA on
electrochemical energy storage systems based on Lithium-ion
batteries, particularly the Lithium-Iron-Phosphate cells. In more
detail, the paper first illustrates the design and experimental
characterization of a family of 12 V modules, each of them
provided with an electronic management system, to be used for
electric traction. Then, the sizing of the energy storage system for
an electric bus providing a service with “fast and frequent”
charge phases is described.
Keywords—Lithium-ion batteries; electric propulsion; battery
management systems; Life cycle assessment
I. INTRODUCTION
The market of the electric energy storage systems based on
traditional lead-acid batteries provides a large choice of
options and configurations for both the battery sizes available
and the manufacturers of cells and systems. Instead, there are
many manufacturers of cells based on the Lithium-ion
chemistry but only a few producers of modules or entire
systems. If the largest players of the automobile market (first
in Japan and then in Germany) have tackled the problem by
means of direct agreements with the manufacturers, the small
and medium enterprise suffers the lack of a mature market for
the Lithium-ion storage systems. This is particularly true for
the Italian scenario [1].
The availability of standard modules provided with
electronic monitoring and managing capabilities would allow
the manufacturers of market-niche vehicles, such as the
electric minibuses for the local public transportation, but also
many other off-road vehicles (e.g. agricultural machines, fork-
lifts and other industrial electrical vehicles), to use the last
generation of battery technology without acquiring the know-
how necessary for the in-house design of the battery system.
These manufacturers can benefit of the modular approach, by
building batteries of various sizes for the different applications
by using the same modules, with obvious advantages in terms
of production volumes and cost reductions [2]-[4].
ENEA and the University of Pisa are carrying out a joint
research effort in the framework of the Program Agreement
“Ricerca di Sistema Elettrico”, funded by the Italian Ministry
of Economic Development. During the last years they have
developed 12 V standard modules with capacity values
spanning from 30 Ah to 100 Ah, which are provided with an
advanced battery management system [5]. The seamless
connection of these modules may allow the producer of
market-niche electric vehicles to implement the required
energy storage in an effective way.
This paper first describes the design and the experimental
characterization of the above mentioned modules. Then, it
shows an interesting use of the modules to implement the
battery suitable for the energy storage of a public
transportation electric bus characterized by a service profile
where the charge phases are frequent and fast. This application
is rather demanding for the energy storage system, as it
requires frequent fast charges for the entire duration of the
daily service, i.e., around 16 h.
II. T
HE STANDARD MODULES
As mentioned above, the basic idea underlying this work is
the development of a standard 12 V battery module to be used
in many applications for electric traction and more in general
for automotive applications [4]. The initial investigation has
led to the identification of the most suited type of chemistry.
The Lithium-Iron-Phosphate (LFP) cells seem to be the best
alternative to lead-acid ones for electric traction, because of
the voltage range and the good intrinsic safety features that
this chemistry provides. TABLE I. shows the voltage ranges
of several Lithium-ion cells as a function of the specific
chemistry, together with the possible voltage ranges of a
series-connected 4-cell module. It stands that the module built
with LFP cells provides the voltage range most similar to that
found in lead-acid batteries.
There are many different kinds of electric traction vehicles,
starting from the “purely electric” to the “hybrid” ones, where
the traction is due to the joint action of an electric motor and
TABLE I. CHARACTERISTICS OF VARIOUS LITHIUM-ION CHEMISTRIES [6]
Chemistry
Cell voltage
(min, nom, max) (V)
# of
cells
Battery voltage
(min, nom, max) (V)
LCO
2.7
−
3.7
−
4.2
4
10.8
−
14.8
−
16.8
LMO
2.75
−
3.7
−
4.2
4
11
−
14.8
−
16.8
NMC
2.7
−
3.6
−
4.2
4
10.8
−
14.4
−
16.8
NCA
2.7
−
3.6
−
4.2
4
10.8
−
14.4
−
16.8
LFP
2.5
−
3.2
−
3.65
4
10
−
12.8
−
14.6
an internal combustion engine. As far as the hybrid vehicles
are concerned, the large spreading of different topologies, e.g.
parallel, series, plug-in rechargeable, charge sustaining, charge
depleting, etc., does not allow the rise of a widely accepted
standard, capable of representing all the different alternatives.
In addition, the development complexity of a hybrid vehicle
tends to narrow its application only to the large and medium
size enterprises, which have the power to establish direct
agreements with the battery producers. Besides the usual
meaning of vehicles as road and off-road transportation
means, the electric energy storage is gaining more importance
also in the marine applications. New battery generations are
expected to be used in boats and vessels for the electric
propulsion in protected areas, for maneuvering in the harbors
and for powering the auxiliary electric appliances, particularly
when anchored.
Given the above considerations, we believe useful the
development of a standard battery module for the electric
traction of vehicle. Keeping in mind the possible applications,
the minimum and maximum capacities are set to 30 Ah and
100 Ah, respectively, with an intermediate size of 60 Ah.
The first investigation step has been the experimental
characterization of the various cell types, coming from
different manufacturers, e.g. Thundersky, Kokam, Valence,
HiPower, etc. The choice of the chemistry and the producer
allowed us to define the preliminary specification of the
control electronic needed to monitor and manage the battery
module, the Battery Management System (BMS). Besides the
monitoring and passive equalization functions, common to any
BMS for vehicular applications, we also have investigated the
key issues of the battery State-of-Charge (SoC) estimation, the
active charge equalization of the series connected cells and the
possible influence on the battery behavior of hysteresis
phenomena, often found in LFP cells [7]-[9]. A comparative
evaluation of commercial BMSs has also been carried out, in
order to extract the information useful for the design and
implementation of an advanced BMS for the module and for
the entire battery (intended as a seamless connection of
modules).
III. M
ODULE DESIGN AND CHARACTERIZATION
Three different modules have been realized [10]. Their
electrical and physical characteristics are listed in TABLE II.
and TABLE III. respectively. One of the aim of the research is
to allow the realization of batteries of any size by simply
interconnecting the elemental modules, without worrying of
the thermal design of the battery or installing a dedicated
cooling system. The solution adopted is based on a forced-air
active cooling system consisting of very cheap fans like those
used in consumer electronic applications.
In fact, each module is provided with three fans located on
the cell container top cover, as it is shown in Fig. 1. The fans
are 50 x 50 x 20 mm
3
in size and are supplied by the 12 V
module voltage with a current of 500 mA. The fans are
activated by the BMS when the temperature sensors attached
to the cells detect a value above +45°C. The fan load can be
considered negligible if compared to the electric motor load,
so that the benefit of the active battery temperature control is
enjoyed with minimal penalty.
The temperature control of the battery is very important as
the thermography of the LFP cell subjected to a discharge of
3C (three times the battery capacity expressed in ampere)
shows an average temperature of 40°C, with hot spots where
the temperature locally reaches the quite high value of 70°C.
The thermal imaging of a 100 Ah LFP cell taken during the
experiment is shown in Fig. 2 [11]. It is worth noting that the
temperature is not uniformly distributed and shows one hot
spot below the positive terminal of the cell.
IV. BMS
ARCHITECTURE AND TESTING
Every battery composed of Lithium-ion cells must be
accompanied by the BMS, an electronic control circuit that
provides the monitoring/management functions that allow the
cells to remain in their safety operation area [12]. The main
innovative feature of the BMS developed for the modules
described in this paper is the active equalization of the cell
charges, realized according to the pack-to-cell topology [9],
[13]. The peculiar feature is the presence of a circular
balancing bus shared between the cells and the modules that
allows the intra-module but also the inter-module equalization
[15]. Charge balancing is achieved by transferring energy from
the more charged cells/modules to the less charged cell,
avoiding the energy dissipation that occurs in passive
Fig. 1. Battery management system board with the cooling fans.
TABLE II. E
LECTRICAL PARAMETERS OF THE MODULES
Size Small Medium Large
Series connected cells 4 4 4
Nominal capacity (Ah) 30 60 100
Temperature (discharge) (°C)
−20÷+65 −20÷+65
−
20÷+65
Temperature (charge) (°C) 0÷+45 0÷+45 0÷+45
Discharge
@ +23°C
Max continuous current
[A]
90 180 300
Peak current @ 60 s (A) 150 300 500
Cut-off voltage (V) 2.5 2.5 2.5
Charge
@ +23°C
Charge method CC/CV CC/CV CC/CV
Max charge current (A) 30 60 100
Max suggested cell
voltage (V)
3.65 3.65 3.65
Cell cut-off voltage (V) 3.85 3.85 3.85
TABLE III. MODULES PHYSICAL PARAMETERS
Module size Length (mm) Depth (mm) Height (mm) Volume (l)
30 Ah 251 160 220 8.83
60 Ah 277 166 256 11.77
100 Ah 279 198 337 18.62
balancing, where the charge in excess is dissipated and not
transferred.
Each module is provided with a BMS board that manages
4 cells. The BMS functions of the entire battery are jointly
carried out by the electronic boards belonging to each module
that are connected in daisy chain. The boards are identical to
each other to standardize their production, but each one may
be configured as “master” or “slave” in the daisy chain
connection. Configuring a board means properly configuring
the firmware on the inner microcontroller. Since the firmware
is designed with a modular and parameterized structure, the
cell capacity is one of the parameters that are configured.
Thus, the BMS board is always the same, regardless of both
the module function inside the chain and the module size.
The BMS boards communicate to each other by means of
an isolated CAN bus. The CAN bus is available outside the
battery to allow the connection of an external PC for
supervision, diagnostic and logging functions. Each BMS
board is attached to the module cells by means of another
printed circuit board, by which the power connections from
cell to cell and the sense connections to each cell terminal are
realized. This solution dramatically simplifies the module
assembly, as any connection cable to be manually wired is
missing. Should a mass production start, the assembly would
completely be automated.
As stated above, the cell balancing is realized according to
the pack to cell topology, in which a DC/DC converter fed by
the module voltage provides some recharging current to a
particular cell selected by means of a switch matrix consisting
of MOS switches. Fig. 3shows the active balancing circuit, in
which the balancing current to the less charged cell may reach
the value of 2 A.
The upper switch reported in Fig. 3 allows the charge
balancing between different modules, when activated. As the
balancing bus structure is circular, the DC/DC current from
one module can reach a cell selected in another module
realizing the inter-module balancing. Furthermore, the
balancing currents coming from different modules may join
into a single cell, to increase the current value up to 6 A and
shorten the balancing time. Each of the module cells is
provided with a temperature sensor for independent
temperature readings. An isolated Hall-based current sensor
interfaced to the BMS board acting as master of the battery
measures the battery current.
V. C
ELL AND MODULE PRELIMINARY LIFE TESTS
The proposed approach for the implementation of a
generic battery needs an evaluation that also takes the
economical aspects into account. To this end, one of the most
important factors is the expected lifetime of the cells when
exposed to deep and intense charge/discharge cycles that may
overcome the rated values. As preliminary experimental test,
we have subjected a 60 Ah cell to the cycle shown in Fig. 4.
Fig. 2. Thermal imaging of a 100 Ah LFP cell showing a hot spot at 70°C.
Dimensions are expressed in mm.
Fig. 3. BMS active balancing circuit. The switch matrix to select a
particular cell and the DC/DC converter to recharge it are shown.
Fig. 4. Experimental fast charge/discharge cycle as life test of the battery.
The application in which the cell is used is an urban
transportation minibus and the cycle represents a scenario in
which the battery is subjected to fast, brief and repeated
recharges.
The cycle consists of a balanced sequence corresponding
to an expected path of 5 km followed by 5 min of fast recharge
at 3C, corresponding to the stop at the end of the route. It is
worth noting that 3C is 3 times (fast charge) the value of the
maximum recharge current for a complete recharge and could
be dangerous for the battery life. At the same time, the
recharge lasts only 5 min (brief charge). Up to now, the cell
has been subjected to tests lasting more than 1,000 h, for more
than 2,000 charge/discharge cycles, without any observable
degradation of the performance. This corresponds to a mission
of more than 13,000 km.
VI. C
ASE STUDY: THE “GULLIVER” REPOWERING
The minibus Gulliver produced by Tecnobus is a small bus
for urban public transportation rather spread in small/medium
cities of the country. The range autonomy requested in the
application described above is achieved by substituting the
whole battery pack at midday, with a new fully charged one. A
traditional lead-acid battery with the capacity sufficient to
complete the daily mission without intermediate recharging
stops would result in excessive weight and size of the battery
itself. However, the solution doubles the costs, as it requires
the availability of two batteries per each minibus: one of them
is under recharge at the deposit while the other is operational
on the bus. Fig. 5 shows the Gulliver minibus during the
battery replacement. The operation is not straightforward but it
is quicker that a full battery recharge.
As proposed above, a fast and partial recharge (5-10 min)
any time the bus stops at the first stop of the daily route would
allow an alternative way of managing the battery energy
during the day. In such a framework, the design of the energy
storage is based on a route of a few kilometers and leads to a
battery of “minimum size”, provided that the policy of
frequent and fast recharges is applied to the battery.
A battery as small as that of a hybrid vehicle needs in any
case to fulfill the requirements in term of available power,
both during the discharge and recharge phases. The Lithium-
ion batteries are the perfect replacement for the lead-acid ones,
because of their high specific power and the almost symmetric
behavior.
A. Battery minimum sizing
The first activity has thus been carried out to verify the
possibility of utilizing the previous described Lithium-ion
modules to re-power the Gulliver bus. We chose the 100 Ah
module as the elemental building block of the new battery.
The modules may be discharged at 3C (the rated value) and
recharged at 3C (for brief periods), as in the lifetime test
described in Section V, without appreciable degradations. We
chose to limit the recharge current to 2C, to reduce the
possible stress of the cell. The characteristics of the Gulliver
minibus and the expected daily route are reported in TABLE
IV. These parameters are the basis for a correct sizing of the
battery. All the other parts and characteristics of the bus are
left unchanged, in order to end up with a transformation kit
consisting only of the new battery and the relevant charger. A
simple wired battery/charger connection through a Combo
unified connector is the most straightforward way to
implement the recharge. An appealing alternative solution
consists of the Schunk pantograph depicted in Fig. 6. An
inductive charger seems at the moment too expensive for the
application and it will be left to further development of the
research.
The main constraint in the battery design is the maximum
recharge current that is limited to 2C. When this limit is
fulfilled, the maximum rated discharge current of 3C seems to
Fig. 5. Battery replacement at midday stop for the Gulliver minibus.
Fig. 6. Schunk recharge station (from www.schunk-sbi.com).
TABLE IV. G
ULLIVER MINIBUS AND COURSE CHARACTERISTICS
Bus mass (1,200 kg of lead-acid cells) 6,000 kg
Average speed 20 km/h
Course length (circular) 5 km
Specific energy consumption (per km) 700 Wh/km
Energy for traction 3.5 kWh
Max available recharge power (380 V/64 A) 43 kW
Max energy recharged in 5 min 3.6 kWh
TABLE V. PARAMETERS OF THE BATTERY
Battery voltage 72 V
Module type 12 V/100 Ah
Max string recharge current (2C) 200 A
Number of strings in parallel 3
Total battery capacity 300 Ah
Max. available battery recharge current 600 A
Max. battery power during discharge 64.8 kW
be sufficient for the bus drive. Therefore, the parameters of the
battery, the size of which is sufficient for the application, are
reported in TABLE V. It is worth noting that the battery
voltage value of 72 V allows the utilization of a large kind of
commercial chargers of sufficient power, which are rather
cheap because of the series production.
Fig. 7 shows the energy stored in the battery during the
simulation of a duty cycle of about 8 h, consisting of
consecutive repetitions of the course, interleaved with fast
(2C), brief (6 min) and frequent (every 21 min) recharges. The
course is repeated several times up to a total length of 115 km.
It is worth noting that more than 50% of the battery energy has
been utilized after 8 h of duty, half of the expected daily
operational time. Some additional recharge is thus needed to
bring the battery back to the full state of charge, to cover the
entire operational period. If the full recharge was not executed,
the vehicle could operate for 200 km before the collapse of the
battery state of charge below 10%. In any case, the full battery
charge would guarantee 14 h of continuous operation.
B. Battery Life Cycle Assessment and proposed sizing
A proper design of the battery must consider some safety
margins that take into account the natural degradation of the
cells parameters due to aging [14]. The battery design is thus
devoted to achieve a long operational period of the vehicle. To
replicate the same properties of the original lead-acid battery,
the Lithium-ion battery should provide a capacity of 585 Ah.
The idea is to design a battery with a size larger than the
minimum value calculated in the above subsection. If we
connect 6 strings of 72 V in parallel according to the scheme
shown in Fig. 8, the volume of the new battery is almost the
same of the lead-acid one to be replaced, but the vehicle mass
decreases from 6,000 kg to 5,520 kg. An important
consequence deriving from having doubled the minimum size
battery is that the battery current is halved in term of C-rate, as
well as the SoC interval of operation. The battery is
significantly less stressed and it is thus foreseen that its life
could increase of at least a factor 2, because the life of a
battery not only depends on the capacity and power supplied
to the load. In fact, the way and the depth of the
discharge/recharge cycles affect the battery life.
Manufacturers indicate that the life of a battery improves by
reducing the depth of discharge. Life tests carried out in our
laboratories on lead-acid batteries demonstrated that halving
the current triplicates the battery life and that the relationship
between aging and current is not linear, being the losses in the
battery proportional to the current squared [16], [17]. A further
advantage of the current reduction is the reduction of the
battery temperature that in its turn leads to less stress.
As the life-tests described in Section V are still running,
the end-of-life cost of the new battery is not available yet, and
so the comparison between the costs of the “electric” and the
“conventional” public transportation solutions is not possible.
However, life tests carried out on other cells of similar
chemistry show that a useful life expectation (calculated when
the residual capacity reaches 80% of the initial value) of
100,000 km is a reasonable value. In addition, a vehicular
battery might be used for a second life application in
stationary storage systems after the above limit is reached.
As far as the module cost is concerned, only half of it is
due to the cells. Therefore, a used module may be regenerated
many times by substituting the cells only, with costs well
below that of a new one. If we take the 100 Ah module as
example, the cost of the cells (presently around 220 €/kWh),
the two electronic boards, the fans and the container, we
foresee a cost of 300÷400 €/kWh, with a production of 10,000
modules, corresponding to about 300 buses.
VII. C
ONCLUSIONS
The standard modules described in this paper are suitable
for the public transportation application considered, even if
consisting of LFP cells, cheaper and safer than the NMC ones.
Standardizing the module in three sizes further allows cost
reduction, as any battery size can be built with series/parallel
combinations of modules.
The preliminary life-test experiments on the considered
LFP cells show that the cells withstand fast charges without
performance degradation and are thus suited to a public
transportation application where a minibus is subjected to
continuous brief cycles of fast charges, with the valuable
advantage of halving the mass of the storage system.
The completion of the currently ongoing life tests is
necessary to assess the economic advantage derived from the
Fig. 7. Battery energy as a fucntion of time with fast recharges cycles.
Fig. 8. Final battery layout with 6 paralled strings of modules.
![TABLE I. CHARACTERISTICS OF VARIOUS LITHIUM-ION CHEMISTRIES [6]](/figures/table-i-characteristics-of-various-lithium-ion-chemistries-6-1zakw5d7.webp)
