The present patent application claims the priority of Singapore patent application n° 10202010561W filed on October 26,2020.

TECHNICAL FIELD

The present invention relates to a method for fast charging a battery cell with an extended life and to a fast-charging system implementing such method.

BACKGROUND OF THE INVENTION

As compared to other rechargeable batteries operating at the ambient temperatures such alkaline- electrolyte and acid-electrolyte based batteries, lithium-ion batteries (LIB) show the best combined performances in terms of energy density (Ed), power density (Pd), life span, operation temperature range, lack of memory effect, lower and lower costs and recyclability.

The LIB market is expanding exponentially to cover the three main applications: a) mobile electronics (ME) (cellphones, handhold devices, laptop PCs . . . ), b) electromobility (EM) (e-bikes, e-cars, e-buses, drones, aerospace, boats,...), and c) stationary energy storage systems (ESS) (power plants, buildings/houses, clean energy (solar, wind, . . . ), industry, telecom . . .

The fastest growing market segment of LIB is the electromobility market.

In electromobility energy density goes with the operation time and driving range of any electric vehicle (EV). Higher Ed provides longer driving range when using a battery pack of a fixed weight (kg) and volume (1).

The energy density of LIB has been steadily improved since their commercialization. However, recent years showed a slowdown in Ed increase with a plateau around 250 Wh/kg and 700 Wh/1 at the cell level.

Because of Ed and Pd limitations current EV, which are mostly LIB powered, have a driving range of about 250 km to 650 km per full charge and a full charging time above 60 min.

Current internal combustion cars can fill a tank in 5-10 min and provide a driving range up to 900 km.

To ensure success public acceptance of EV for the coming energy transition the most serious option today is fast charging. Current fast charging stations for EV provide a limited amount of charge below 60 min because of: 1) overheating (reaching a safety temperature limit), and/or 2) overcharging (reaching a safety voltage limit). Common charging methods for Lithium-Ion Batteries are disclosed in the Journal of Energy Storage 6 (2016) 125-141, as shown by Prior Art Figure 1.

Except for the “voltage trajectory” method, all other LIB charging methods apply a constant current and/or a constant voltage in at least a step of the charging process.

There is no indication of cell’ cycle life nor of the cell’ temperature profde when these methods are used for 0-100% full charging of a LIB in less than 60 minutes (fast charging). There is no indication the methods stated apply to all battery’ chemistry

With reference to Prior Art Figure 2 the typical CCCV (Constant Current- Constant voltage) charging and Constant Current discharge profde, during the Constant Current step, the voltage increases from its initial value to a set voltage value (up to 4.4V). During the Constant Voltage step, up to 4.4V, the current drops to a set value (here 0.05C or C/20).

During the rest time, current is nil, and voltage drops to reach an open-circuit voltage (OCV)

During the CC discharge, the current is fixed, and voltage drops to a limit (here 2.5V)

During the following rest time, current is nil, and voltage increases to a new OCV value.

With reference to Prior Art Figure 3 that features Multistage constant current charge profile (MSCC), two charge currents have been applied successively to the cell, Il and 12, (where in general Il >12).

Il is applied until voltage reaches a first value VI Then 12 is applied until voltage reaches a value of V2 and so on.

Other currents Ij can be applied until a voltage Vj is reached, where Vl>V2>V3>...Vj>Vj+l.

The MSCC charge process ends when either the target capacity is reached, or a voltage high limit is reached or a temperature limit is reached.

CCCV and MSCC are the most popular charging methods used in lithium-ion batteries today. CCCV and MSCC are simple and convenient methods if the full charging time is above 2 hours.

Both CCCV and MSCC are based on applying one or several charging constant current(s) (CC) up to preset voltage limit(s), then for CCCV by applying a constant voltage (CV).

Both CCCV and MSCC cannot realistically be used to charge a battery in less than one hour because of: 1) excess heat generation, 2) lithium metal plating on the anode side, which may create an internal short circuit and thermal runaway event, 3) the reduction of the battery life due to accelerate ageing. Moreover, when used for charging batery cells connected in series, CCCV requires cell balancing, as discussed, for example, in the paper “Implementation of a LiFePO4 batery charger for cell balancing application”, by Amin et al. / Journal of Mechatronics, Electrical Power, and Vehicular Technology 9 (2018) 81-88.

Active cell balancing, required for high power applications, has the disadvantage of slow balancing speed and thus time-consuming, complex switching structures, it also needs advanced control techniques for switch operation.

Fast charging (FC) protocols are reviewed in the paper “Lithium-ion batery fast charging: a review” published in eTransportation 1 (2019) 100011. Issues of fast-charging are identified for fast-charging with charging time<lh: heat generation, lithium plating, materials degradation, limited charge uptake within tch (ASOC<100%), reduced cycle life, safety, and thermal runaway.

The paper in Journal of Energy Storage 29 (2020) 101342 recites CCCV limitations in fast charging and discloses that cycle life decreases when the Total Charge Time (TCT)= CCCT+CVCT decreases.

As recited in eTransportation 1 (2019) 100011, to date, no reliable onboard methods exist to detect the occurrence of crucial degradation phenomena such as lithium plating or mechanical cracking. Techniques for detecting lithium plating based on the characteristic voltage plateau are promising for online application, but fully reliable methods to distinguish lithium stripping from other plateauinducing phenomena, or to detect plating where no plateau is observed, have not yet been reported.

Many studies on fast charging protocols have been of empirical or experimental nature, and therefore their performance has only been assessed for a limited range of cell chemistries, form factors, and operating conditions. Such results cannot be easily extended to other cell types or ambient temperatures, as supported by the often-conflicting findings reported by different authors.

Moreover, a batery cell is considered to be at the end of its life when its discharge capacity represents only a percentage of its initial capacity after a predetermined number of charge cycles. Typically, a percentage of 80% for the capacity after 2000 cycles is an indicator of a batery cell at the end of its life. Usually, end-of-life bateries used in demanding applications such as electric mobility are then withdrawn and finally assigned to a second life. This has important economic consequences as well as in terms of life cycle.

A main objective of the invention is to overcome these issues by proposing a new method for fast charging batery cells, which allows an extension of life for batery cells, whatever the charging time. MAIN SYMBOLS AND DEFINITIONS i, I = Electric current intensity (A, mA. . . ) v, V= Cell voltage (in Volt, V)

Qch, q _C h= charge capacity (Ah, mAh. . . )

Qdis, qdis= discharge capacity (Ah, mAh...)

Qnom= cell’ nominal capacity (Ah, mAh...)

C-rate= current intensity relative to the charge time in hour.

IC-rate is the current intensity needed to achieve Qnom in Ih

2C-rate is the current intensity needed to achieve Qnom in 0.5h

0.5C-rate is the current intensity needed to achieve Qnom in 2h

SOC= state of charge relative to Qnom (in %)

SOH=state of health is the actual full capacity of the cell relative to the initial Qnom

SOS=state of safety estimated risk of thermal runaway

A= The time derivative of voltage (A=^ in V.s ¹ ) t _s = step time (in s) tch = charge time (in min)

SUMMARY OF THE INVENTION

This goal of the battery cell’ life extension is achieved with a method for charging a battery cell provided with charge/discharge terminals to which a charging voltage can be applied with a flowing charging current, characterized in that it comprises the steps of: applying to terminals of said battery cell a plurality of constant voltage stages Vj, where Vj+i> Vj , j=l, 2. . . , k, each voltage stage comprising intermittent nj voltage plateaus, between two successive voltage plateaus within a voltage stage, letting said charging current going to rest (1=0 A) for a rest period R , l<p<nj. until either one of the following conditions is reached: a pre-set charge capacity or state of charge (SOC) is reached, the cell temperature exceeds a pre-set limit value T _[im and the cell voltage has exceeded a pre-set limit value Vi _im .

- collecting data on at least two previous discharge capacities measured or during previous charge cycles for said battery cell,

- calculating a relative variation (AQ/Q) of said discharge capacity, from said collected data, - comparing said calculated relative capacity variation (AQ/Q) to a predetermined threshold (s),

- if said calculated relative capacity variation (AQ/Q) exceeds said predetermined threshold (s), modifying at least one charge parameter among a selection of charge parameters including the duration of said voltage plateau, the voltage stage shift, and the rest time, so as to bring back said relative capacity variation (AQ/Q) below said predetermined threshold (s).

A transition from a voltage stage Vj to the following stage Vj+i is advantageously initiated when if™ , p=nj reaches a threshold value .

The life extension method of the invention can further comprise a step for calculating the following stage Vj+i as = Vj + AV(j), with AV(j) relating to the current change AI(j)= I ™- if™, p=nj.

The method of the invention can further comprise the steps of measuring the intensity (Io) of current in the battery cell during a voltage stage Vj, calculating an intensity variation (AI(j)) as AI(j) = Io - I limit, with Limit defined a predetermined limit current calculating a voltage variation (AV(j)) as AV(j) = K _n . AI(j ), with K _n defined as an adjustable coefficient,

-applying a new voltage stage Vj+i =Vj+AV(j) to the terminals of said battery cell.

The successive K-values K _n -i to K _n can be determined by using a machine-learning technique, so as to maintain a sufficient charge of the battery cell.

The passage from a voltage plateau to the other is initiated either by detecting a current variation Al greater than a predetermined value, or by detecting a current smaller than a limit C-rate.

A limit C-rate which allows to move from a voltage plateau to another can be determined as C- Rate. (1+a), with a defined as a coefficient provided for compensating the rest time between two voltage plateaus.

The life extension method of the invention can further comprise the steps of: between two successive current rest times /? _; ^p-1 and R within a voltage stage Vj, and a pending voltage plateau, detecting the flowing pulse-like current dropping from an initial value I 'p reaches a final value if™ where l<p<nj , ending said pending voltage plateau, so that said flowing pulse-like current drops to zero for a rest time R , with said voltage departing from Vj., after the rest time R is elapsed, applying back said voltage to Vj. The life extension method of the invention can further comprise an initial step for determining an initial K-value and a charge step from inputs including charging instructions for C-rate, voltage and charge time.

The life extension method of the invention can further comprise a step for detecting a C _S hift threshold, leading to a step for determining a shift voltage, by applying a non-linear voltage equation and using K-value and AC -rate.

The life extension method of the invention can be applied to a combination of battery cells arranged in series and/or un parallel.

According to another aspect of the invention, there is proposed a system for fast-charging a battery cell provided with charge/discharge terminals to which a charging voltage can be applied with a flowing charging current, said system comprising an electronic converter connected to a power source and designed for applying a charging voltage to the terminals of a battery cell, said electronic converter being controlled by a charging controller designed to process battery cell flowing current and cell voltage measurement data and charging instruction data, characterized in that it further comprises: means for collecting data on at least two previous discharge capacities measured or estimated during previous charge cycles for said battery cell, means for calculating a relative variation (AQ/Q) of said discharge capacity, from said collected data, means for comparing said calculated relative capacity variation (AQ/Q) to a predetermined threshold (s) and for delivering an information on exceeding said threshold (s), and in that said charging controller is programed to modify at least one charge parameter among a selection of charge parameters including the duration of said voltage plateau, the voltage stage shift, and the rest time, so as to bring back said relative capacity variation (AQ/Q) below said predetermined threshold (s).

The electronic converter can advantageously include a microcontroller with processing capabilities enabling (i) implementation of artificial methods and (ii) online storage and computation of VSIP data.

This invention discloses a Voltage Staged Intermittent Pulse battery charging method and charging systems (VSIP) consisting of:

The total full (100% ASOC) charging time is below 180 min, below 90 min and below 30 min

Applying a plurality of constant voltage stages Vj , where Vj+i> Vj , j=l, 2. . . , k.

Each voltage stage consists of intermittent nj voltage plateaus Between two successive voltage plateaus with a voltage stage the current goes to rest (1=0 A) for a period 7??, l<p<nj .During the current rest period R? the voltage departs from Vj.

Between two successive current rest times 7? ’ ^-1 and R within a voltage stage Vj the flowing pulselike current drops from an initial value 7™ to a final value if™ where l<p<nj

When if™ is reached, the current goes to rest (drops to zero) for a rest time R

After the rest time R is elapsed the voltage goes back to Vj

The transition between voltage stage Vj to the following stage Vj+i takes place when if™ , p=nj reaches a threshold value lf„^

The voltage step AV(j)= Vj+i - Vj relates to the current change AI(j)= I ™- if™, p=nj

The VSIP charge process proceeds until either one of the following conditions is reached: 1) a preset charge capacity or state of charge (SOC) is reached, 2) the cell temperature exceeds a pre-set limit value T _[im and, 3) the cell voltage has exceeded a pre-set limit value Vi _im .

The main characteristics of the VSIP method are:

VSIP fully charges a battery (ASOC=100%) in a time lower than 30 min.

The charging time is even lower if ASOC<100% (partial charge such as for example from 20 to 100%, ASOC=80%)

The cell voltage during VSIP may exceed 4.5 V in LIB, 2V in of alkaline cells and 3 V in lead acid batteries

During VSIP none of the voltage and current is constant for a period higher than 3 min.

The temperature difference between the cell temperature Tcell and the ambient temperature Tamb remains below 25 °C (Tcell - Tamb <35 °C) during VSIP

The VSIP operating parameters are adjustable according to the cell’ chemistry, SOC, SOH and SOS

VSIP parameters adjustment can be performed using artificial intelligence (Al, such as machine learning, deep learning...)

VSIP applies to individual battery cells as well as to cells arranged in series and in parallel (battery modules, battery packs, power wall, . . . )

VSIP applies to a variety of battery cell chemistries including and not limited to LIB, solid-state lithium batteries, sodium-based anode cells, zinc-based anode cells, alkaline, acid, and high temperature cells (i.e. molten metal cells), ....

Two successive VSIP current and voltage profiles can be different from each other.

The advantages provided by the fast-charging VSIP method according to the invention are: VSIP is a universal charging technology that applies to all types of rechargeable batteries, including lead acid, alkaline, lithium ion, lithium polymer and solid-state lithium cells and for any application, including but not limited to ME, EM and ESS.

VSIP fully charges batteries (from 0 to 100% SOC) below 60 min and below 30 minutes, while keeping the cell’ temperature below 50 °C (safety) and providing long life span. VSIP can apply for quality control (QC) of batteries for specific applications (stress test). Because VSIP is an adapted charging method it extends the life span of batteries under any operation conditions (power profile, temperature, . . . )

VSIP increases the energy density of battery cells versus their rated energy density. Although VSIP is designed for fast charging it also applies to longer charging times tch> 60 min

A fast charge cycle performance index is also provided as: with = normalized cycle performance index i= cycle number ti=charge time @ 1 ^th cycle (hr)

Qdisch ⁼ discharge capacity @ i ^th cycle (Ah)

Qnom = ⁿ ominal capacity (Ah) n= cycle number when Qaisch/Qnom falls below ~ 80%

A new technology for safely fast charging LIB based on Voltage Step Intermittent Pulse (VSIP) has been demonstrated.

VSIP is an adapted charging technology with adjustable parameters either manually or using artificial intelligence methods and techniques

VSIP 100% SOC charge below 20 min is possible while keeping low temperatures (<45 °C) and long cycle life (>1300#).

Partial charge (ASOC<100%) can be performed below 10 min

Voltages above 4.5V can be safely reached under VSIP charge.

There is no sign of lithium plating during VSIP charge.

Over 1000 charge-discharge cycles can be achieved with ASOC<100% with VSIP charge.

VSIP can be used for: 1) cell’s quality control. 2) single cells and for cells arranged in series and in parallel (battery module and battery pack), 3) storage capacity enhancement,

Fast charging performance index can be used as a metrics to compare fast charge protocols. Furthermore, with the NLV based charge method according to the invention, it is no longer necessary to provide cell balancing for the charging of battery cells connected in series, since it is the charging voltage that is now controlled. Thus the fast-charging method of the invention provides intrinsic balancing between the battery cells.

DESCRIPTION OF THE FIGURES

Figures showing Prior Art:

Figure 1 is a schematic description of prior art charging methods;

Figure 2 shows Typical CCCV charging and CC discharge profde;

Figure 3 shows Multistage constant current charge profde (MSCC) ;

Figure 4 and Figure 5 show The CCCV limitations in fast charging;

Figures showing the invention:

Figure 6 shows typical voltage and current profdes during VSIP charge and CC discharge cycles;

Figure 7 shows typical voltage and current profdes during VSIP charge and CC discharge (here full charge time is 26 min);

Figure 8 shows typical voltage and current profdes during VSIP charge;

Figure 9 shows typical voltage profde during VSIP with a plurality of voltage stages Vj (here total charge time is about 35 min);

Figure 10 shows detailed voltage and current profdes during VSIP charge showing voltage and current intermittency.

Figure 11 shows detailed voltage and current profdes during VSIP charge showing rest time;

Figure 12 shows Voltage and current profdes during rest time showing a voltage drop;

Figure 13 shows current profde at stage j;

Figure 14 shows current profde at sub-step j,p;

Figure 15 shows Typical AV(j)= Vj+i - Vj vs. Time profde during VSIP charging in ~17 min over many cycles;

Figure 16 shows voltage and gained capacity during VSIP charge in 26 mn;

Figure 17 shows discharge profde of 12 Ah cell after VSIP charge in 26 mn;

Figure 18 shows linear voltammetry vs VSIP;

Figure 19 shows two successive VSIP charge profiles can be different from each other;

Figure 20 shows VSIP charge voltage and current profdes (60 min);

Figure 21 shows VSIP charge voltage and current profdes (45 min); Figure 22 shows VSIP charge voltage and current profiles (30 min);

Figure 23 shows VSIP charge voltage and current profiles (20 min);

Figure 24 shows 80% partial charge with VSIP in ~ 16 min;

Figure 25 shows Temperature profile during VSIP charge in 30 min: Stress test for LIB’ quality control (QC);

Figure 26 shows Temperature profile during VPC in 20 min of a good quality cell;

Figure 27 shows VSIP enhances cell’s capacity;

Figure 28 and 29 show VSIP applies to multi -cell systems in parallel;

Figure 30 and 31 show VSIP applies to multi-cell systems in series;

Figure 32 shows a Cycle performance index;

Figure 33 is a flow diagram of an embodiment of the extended-life fast-charge method, including a Bayesian optimization;

Figure 34 is a schematic view of an extended-life fast-charge system implementing the fast-charge method of Figure 33;

Figure 35 shows 4 cells-in-series voltage profiles measured during a NLV charge in about 30 min.

DETAILED DESCRIPTION OF AN EMBODIMENT

For programming a controller implementing the fast-charging method according to the invention, with an artificial intelligence (Al)-based approach, a list of duty criteria is proposed: fixing the charging time t _C h reaching the target capacity in t _C h keeping temperature under control (<60 °C) achieving the target cycle number insuring battery safety enhancing capacity

The variables in the fast-charging method according to the invention are: the VSIP governing equation A= AV/At =f(i, V, Ai/At, T, SOC, SOH) the charge current limits the current trigger for next voltage step the rest time the temperature limit the voltage limit the target capacity limit

A Bayesian optimization is used to adjust the Non Linear Voltammetry (NLV) variables. The NLV variables are adjusted at each cycle to meet the criteria:

A= — at =f( 'i, V, - at, T, SOC, SOH) '

With reference to Figures 6 and 7, in a fist embodiment, the fast charging (VSIP) method according to the invention is implemented during charge sequences within VSIP charge, CC discharge cycles. In these profiles, the C-rate is representative of the current in the battery cell.

As shown in Figures 8 and 9, a VSIP charge sequence, which has a duration of about 26 min, includes a number of increasing voltage stages, each voltage stage Vi,...,Vj,Vj+i,..Vk including constant voltage plateau.

A shown in Figures 10 and 11, during each voltage plateau in a VSIP charging sequence, the voltage profile is constant and decreases to a low constant voltage between two successive plateaus, while the C-rate profile includes a decrease during each plateau and decreases to zero during the rest period between two plateaus.

During a rest time, as illustrated by Figure 12 showing detailed current and voltage profile, the voltage can be controlled so that has a constant negative value calculated as above described.

As shown in Figure 13, a voltage stage j includes current impulsions 1,2,3, ...nj in response to voltage plateaus applied to the terminal of a battery cell.

During a voltage plateau Vj, the current at sub-step j,p decreases from I ™ to i , as shown in Figure 15.

For a large number of charging cycles operated with the fast-charging method according to the invention, the voltage variations AV experienced between the successive voltage plateau within successive voltage stages Vj, Vj+i. globally decrease with time, as shown in Figure 15.

During a voltage charge VSIP sequence lasting 26 min full charge time as shown in Figure 16, the charge capacity Q _C h continuously increases while the corresponding voltage profile includes successive voltage stages each comprising voltage plateau with rest times. As shown in Figure 17, during a following discharge sequence, the discharge capacity Qdis decreases with the voltage applied to the terminals of the battery cell.

The VSIP fast charging method according to the invention clearly differs from a conventional Linear Voltammetry (LV) method, with respective distinct voltage and current profiles shown in Figure 18. The respective current and voltage profiles can differ from a charge/discharge VSIP cycle to another, as shown in Figure 19.

The variability of voltage and current profiles is also observed when the charge time is modified, for example from 60 min, 45 min, 30 min to 20 min, with reference to respective Figures 20,21,22 and 23. For a 60 min charge time, the charge sequence includes 4 voltage stages (Figure 20), and for a 45 min charge time the charge sequence includes 8 voltage stages (Figure 21). For a 30 min charge time, the charge sequence includes 10 voltage stages (Figure 22) and for a 20 min charge time, the charge sequence includes 4 voltage stages (Figure 23).

As shown in Figure 24, the VSIP charging method according to the invention allows a 80% partial charge of a Lithium-Ion battery cell in about 16 min.

With reference to Figure 25, during a VSIP charge in 30 min, cells A, B and D had temperature raising above the safety limit of 50 °C. These battery cells didn’t pass the VSIP stress test. Only cell C passed the stress test. It means that all LIB cells can’t be fast charged.

Thus, the VSIP charging method according to the invention can also be used as stress quality control (QC) test before using a cell in a system for fast charging

With reference to Figure 26, during a charge sequence of an excellent quality LIB cell, the full charge is reached in about 20 min and the temperature of the cell does not exceed 32 °C.

With reference to Figure 27, by adjusting the VSIP parameters such as the upper voltage limit, the step time, AV and A// At for the voltage step transition, the discharge capacity can be improved without compromising safety and life span.

The VSIP charging method according to the invention can be implemented for charging 4 LIB cells assembled in parallel in about 35 min, as shown in Figure 28 with a CC discharge and in Figure 29 which is a detailed view of the voltage and current profdes during the VSIP charge sequence of Figure 28,

With reference to Figures 30 and 31, the VSIP charging method according to the invention can also be applied for charging 4 e-cig cells in series, in about 35 min.

As shown in Figure 35, the profdes of the voltages VI, V2, V3 and V4, corresponding to 4 cells connected in series and measured during a NLV charge, are very close to each other, which avoids cell balancing.

Note that in this configuration, the VSIP charging method is particularly advantageous, compared to CCCV, as it no longer requires a time-consuming and energy-using active cell balancing.

As shown in Figure 32, the charge and discharge capacity varies as a function of the number of cycles, A fast charge cycle performance index can be calculated as: n 1 Qdisch/Q nom t,- i=l with

<t> = normalized cycle performance index i= cycle number ti=charge time @ 1 ^th cycle (hr)

Qdisch ⁼ discharge capacity @ i ^th cycle (Ah)

Qnom = ⁿ ominal capacity (Ah)

With reference to Figures 33 and 34, an example of an extended-life fast-charge system 100, along with the implemented charging method, is now described.

This extended-life fast-charge system 100 comprises a VSIP charge system 10 including a power electronics converter 11 designed for processing electric energy provided by an external energy source E and supplying a variable voltage V(t) to a battery cell B to be charged. Note that this battery cell B can be replaced by a system of battery cells connected in series and/or in parallel.

The VSIP system 10 further includes a VSIP controller 1 designed for receiving and processing: measurement data provided by a current sensor 13 placed in the current circuit between the power electronics converter 11 and the battery cell B, and by a temperature sensor 12 placed on or in the battery cell B, instruction data collected from a user interface 6, including inputs such as an expected C- Rate, a charge voltage instruction and a charge time instruction.

The extended-life fast-charge system 100 is further adapted to receive the parameter s as an input 14 to the user interface 6, ass

Typically, the parameter s can be equal to 0.002% (average slope of capacity loss per cycle), corresponding to a capacity loss of 20% in 1000 cycles.

An output 15 can be the number of cycles experienced by the charged battery cell with a relative variation in discharge capacity per cycle AQ/Q less than 8.

The VSIP controller 1 is further designed to control power electronics components within the converter 10 so as to generate a charge voltage profde according to the VSIP method until at least of one the termination criteria for ending 9 the charging process are met.

These VSIP termination criteria 5 include:

- minimum C-Rate cut-off,

- safety voltage exceeded,

- charge capacity reached

- over temperature. From inputs “C-Rate”, “Voltage” and “elapsed charge Time” which can be entered as instructions 6 by an user, the VSIP controller 1 first determines an initial K value and a charge step.

Provided that no charge termination criterion is met and a predetermined threshold for C-Rate is not reached, the VSIP controller 1 launches a charge sequence 2 by applying voltage for a charge step duration and C-Rate - which is an image of the current flowing into the battery cell - is measured.

When current reaches a pre-set C-rate value, the VSIP controller 1 commutes to a rest period 3 during which no voltage is applied to the battery cell. The duration of this rest period depends on the measured C-Rate before current decreasing.

If the C shift reaches the determined threshold 8, the VSIP controller 1 calculates a shift voltage 4 required to maintain a sufficient charge of the battery cell. This calculation is based on the NLV equation using K-value and AC -rate. The calculated shift voltage is then applied for applying a new voltage stage to the battery cell.

In the particular embodiment of the fast-charge method shown in Figure 33, this fast-charge method comprises, at the output of the above-described VSIP fast-charge process 30, a step 24 for calculating the relative capacity loss AQ/Q based on previously collected capacity data 23. These capacity data, that include capacity data collected during two successive charge cycles, may have been collected in different ways : from local storages within the VSIP controller or within the battery cell. The AQ/Q value is then compared (step 25) to the threshold s. As long as AQ/Q is less than s, the present VSIP charge parameters are maintained.

Temperature T of the battery cell is monitored (step 21) all along the charge process and compared (step 22) to the predetermined limit of temperature Tum. If measured temperature T exceeds Tii _m , the VSIP charge process is ended.

If AQ/Q exceeds s, the VSIP charge parameters are then modified and applied to the VSIP charge process 30. Adjustment rules can be easily derived from the equations governing the VSIP process as above described. Artificial Intelligence techniques can also be implemented to process previous capacity loss measures in function of a plurality of VSIP parameters.

Of course, the present invention is not limited to the above-described examples and other embodiments can be considered without departing from the scope of the invention.