FR3021169A1 - SYSTEM AND METHOD FOR CHARGING A TRACTION BATTERY REDUCING LEAKAGE CURRENTS - Google Patents

Abstract

The invention relates to a method for charging a traction battery of a vehicle comprising a non-isolated charger able to recharge the battery from an external electrical network, comprising: a step (E1) for starting a battery; charge at nominal load power (Pinit); - step (E2) for comparing at least one estimate of a component (Ifuite) of a leakage current at a cutoff threshold (Scoupure), said method being characterized in that it further comprises a step (E3) of decreasing the load power (Pcharge) at a load power lower than said rated load power (Pinit), when said at least one estimate is greater than or equal to cut-off threshold (Scoupure).

Description

BACKGROUND OF THE INVENTION The present invention relates generally to the fields of electricity and the automobile, and more specifically relates to a method of charging a traction battery. for an electric or hybrid vehicle.

The on-board chargers for electric or hybrid vehicles make it possible to recharge the traction battery of such a vehicle from an external power supply network. Most of these chargers are isolated from the power grid by a galvanic isolation transformer. However such a transformer is expensive, heavy and bulky. This is why some chargers are uninsulated from the external power supply network, and require specific protection against leakage currents that may appear on the vehicle chassis and turn into touch current, which are limited to 3.5mA (milliAmpères) by the international standard IEC (International Electrotechnical Commission) 61851-2. Thus the French patent FR2966652 describes a protection device against leakage currents and touch currents, arranged at the input of a non-insulated charger of electric vehicle, and to prohibit the charging of the vehicle battery as soon as an estimate of the leakage current or touch current in the vehicle reaches certain thresholds. Another protection device for a non-isolated loader is described in French patent application number FR1262469, in which a current load of the traction battery of a vehicle is cut as soon as certain frequency components of the leakage current exceed thresholds associated with these components. These thresholds are set so that a circuit breaker present upstream of the charger is not blinded by these frequency components of the leakage current. The large leakage currents thus reduce the availability of a non-isolated charger for electric or hybrid vehicles, and therefore reduce the possibilities for a user of such a vehicle to recharge his traction battery, depending on the quality of the network. food that he has at his disposal. The patent application US20130176650 proposes to reduce the leakage currents on a non-isolated charger of electric or hybrid vehicle, which increases the availability thereof. For this, a capacity is added at the input of the charger between the neutral of the supply network, and the chassis of the vehicle. However, this solution requires the addition of components to the vehicle charger, and is not applicable to all power networks, the neutral of a power network is not always available. It is an object of the invention to overcome at least some of the disadvantages of the prior art by providing a method and system for charging a traction battery for an uninsulated charging vehicle to increase availability. charger especially in the case where the supply network available to a user of the vehicle is very disturbed. To this end, the invention proposes a method of charging a traction battery of an at least partially electric traction vehicle, said vehicle comprising a non-insulated charger able to recharge said traction battery from an electrical network. external, said method comprising: a step of starting a charge of said traction battery at a nominal charging power, a step of comparing at least one estimate of a component of a leakage current at a threshold of switching, said method being characterized in that it further comprises a step of decreasing a charging power in progress at a charging power lower than said current charging power, when said at least one estimate is greater than or equal to threshold cutoff.

Thanks to the invention, it avoids the user a first cut or a new cut of the current load because of too large leakage currents, the charging system according to the invention cutting the load immediately or after one to two minutes depending on the nature of these leakage currents too high. The charging system according to the invention prohibits the resumption of the charge in progress after a predetermined number of charge interruptions. According to an advantageous characteristic of the charging method according to the invention, in said comparison step several component estimates of said leakage current chosen from a list comprising a DC component, a low-frequency component lower than 400 Hz (Hertz), a component are compared. a high frequency component corresponding to the frequency of hashing of the charger, at corresponding cut-off thresholds, said reduction step being carried out as soon as one of said estimates is greater than or equal to its cutoff threshold. By distinguishing between the different types of cutoff possible, an immediate cutoff due to a too high low frequency component of the leakage current or a non-immediate cutoff due to the other components of the leakage current, increases the availability of the load for the user. adapting the step of decreasing the load power to these different types of cutoff. The decrease is indeed greater in the case where the too high component of the leakage current causes an immediate interruption of the load. According to another advantageous characteristic, said step of decreasing the current charging power is carried out only when said at least one estimate remains above said cut-off threshold during a first predetermined duration and when said at least one estimate corresponds to a component said leakage current which is a DC or high frequency component of said leakage current. This high frequency component corresponds to a frequency greater than 400 Hz.

Thus, when the excessively high component of the leakage current does not lead to an immediate interruption of the current load, it is ensured that this fault is not due to a temporary disruption of the network, such as for example the start-up of a consumer load using the network, before decreasing the load power, in order to optimize it.

According to another advantageous characteristic, said step of decreasing the current charging power is iterated when said at least one estimate remains above said cut-off threshold for a second predetermined duration and when said at least one estimate corresponds to a component of said leakage current which is a continuous or high frequency component of said leakage current. This prevents the load from being cut off in the event that the excessively high component of the leakage current does not lead to an immediate interruption of the current load, while maintaining optimum charging time and power.

According to another advantageous characteristic, said step of decreasing the current charging power is preceded by a step of calculating said lower charging power, in which said current charging power is multiplied by a gain less than a capacity to converging said current load power to a load power inducing leakage current component values lower than their respective cutoff thresholds in less than ten iterations. The duration of the load is thus optimized while reducing the risk of a definitive interruption of the current load by the protections against the touch currents implemented in the system according to the invention.

According to another advantageous characteristic, when said at least one estimate corresponds to an estimate of a low frequency component of the leakage current, said comparison step is followed by a step of breaking the load of said traction battery, and said reduction step is performed during a recovery step of said load at said lower load power.

This implementation of the method ensures the safety of the user in the case of dangerous currents for the user. According to another advantageous characteristic, said step of decreasing said current charging power is preceded by a step of analyzing frequency deviations of a signal for measuring the supply frequency of the external electrical network, and is conditioned by said analyzing step, said step of taking up said load being performed without said step of decreasing when said frequency differences are below a disturbance threshold. Indeed, when the network is not disturbed the reduction step is unnecessary, and the recovery of the load at its nominal power is then preferable for the user. According to another advantageous characteristic, said step of taking up said load is followed by a step of analyzing the difference between said estimate of a low frequency current component and said cutoff threshold. Monitoring the level of the mains disturbances determines to what extent the load power can be increased to its rated power. According to another advantageous characteristic, when said difference between said estimate and said cutoff threshold is greater than a safety margin for a third predetermined duration, said analysis step is followed by a step of increasing the charging power by Classes. This optimizes the duration of the current load. According to another advantageous characteristic, said step of increasing the current charging power is iterated when said difference remains above said safety margin for a fourth predetermined duration. This optimization avoids a new power failure by slowly raising the load power in the case of transient disturbances of the network. According to another advantageous characteristic, when said step of increasing is followed by a step of breaking said load, said step of breaking is followed by a step of taking back said load at the load power preceding the last step of increase. This increases the availability of the load for the user, optimally. According to another advantageous characteristic, when said at least one estimate remains below said cut-off threshold after said step of decreasing for a fifth predetermined duration, said step of decreasing is followed by a step of increasing the charging power in progress at said rated load power. This increase step optimizes the duration of the load in the case of a fault resulting in an immediate or non-immediate cut, and in the case of transient disturbances of the network. The invention also relates to a charging system of a traction battery of an at least partially electric traction vehicle, said vehicle comprising a non-insulated charger able to recharge said traction battery from an external electrical network, said characterized in that it comprises means for implementing the charging method according to the invention. The invention also relates to a computer program comprising instructions for implementing the charging method according to the invention, when it is executed on one or more processors. The charging system according to the invention and the computer program according to the invention have advantages similar to those of the method according to the invention. Other features and advantages will become apparent on reading a preferred embodiment described. with reference to the figures in which: - Figure 1 shows a charging system according to the invention, in this embodiment of the invention - Figure 2 shows steps of the charging process according to the invention, in the case a fault in the leakage current does not cause an immediate interruption of the current load, in this embodiment of the invention; FIG. 3 represents possible first steps of the charging method according to the invention, in the case of a leakage current fault resulting in an immediate cut of the current load, in this embodiment of the invention; and FIG. 4 represents possible second stages of the charging process. according to the invention, in the case of a leakage current fault resulting in an immediate interruption of the current load, in this embodiment of the invention. According to a preferred embodiment of the invention shown in Figure 1, a charging system S according to the invention is integrated in an electric vehicle VE or hybrid. The charging system S implements the charging method according to the invention in a software and hardware way. It is able to be connected to a network RES of external power supply, via an electric charging cable leading the supply phases (p1, (p2, (p3 of the network RES, and a socket of the network RES, at the input of the charging system S which comprises a measurement and filtering device DMF This RES network is here a three-phase network but alternatively the charging system uses a single-phase supply network or The DMF measuring and filtering device at the input of the charging system S comprises: means for estimating the leakage current and the touching current of the charging system S. These means include, for example, a measurement torus comprising the phases and the earth and for measuring the magnetic flux of the leakage current of the three phases, and a mass current detection circuit, as described in the application FR2966652. flight provides an estimate of the touch current; alternatively the measurement of the leakage current is carried out using a measurement resistor also called shunt resistor at the ground connection of the vehicle VE. means for filtering the leakage current measured over time, supplied by the estimation means in the form of a measurement signal; these means are detailed in the patent application filing number FR1262469, and allow to obtain estimates of the following components of the leakage current: o a DC component, o a low frequency component below 400Hz, o a high frequency component included between 1 kHz and 150 kHz, and a high frequency component corresponding to the hash frequency of the charging system which is, in this embodiment of the invention, 10 kHz. These values are obtained by appropriate low-pass and bandpass filtering, after amplification, rectification or clipping of the measurement signal if necessary.

The charging system S also comprises a non-isolated charger CNI, for example of the type described in the patent application FR2943188, which transforms the alternating current delivered by the supply phases (p1, (p2, (p3) into a direct current which it delivers to the traction battery BAT of the vehicle after smoothing by a smoothing ability.

The charge of the traction battery BAT is supervised at the level of the battery BAT by a BMS supervision module communicating with the main computer ECU of the vehicle VE. The VE VEU's main ECU computer also communicates with the DMF measuring and filtering device and the non-isolated CNI charger. The continuous estimation of the leakage current and possibly of the touch current supplied by the measuring and filtering device DMF to the ECU computer enables it to cut off a current load when a value of the leakage current reaches a threshold This cut-off threshold corresponds, for example, to an acceptable maximum touch current. Indeed the measurement of the touch current is not always available, it establishes a cutoff threshold applicable to the leakage current estimated or measured, and responding to various electrical safety constraints. This interruption of the charge is effected by the opening of contactors of the battery BAT, and / or by communication with the charging terminal, and / or by opening contactors upstream of the charging system S in the vehicle VE. More precisely, the ECU calculator compares the values of the different components of the leakage current estimated by the measuring and filtering device DMF with a corresponding threshold of breaking of the load: the value of the cutoff threshold for the DC component of the leakage current is for example 5.5 mA. The value of the cutoff threshold for the low frequency component of the leakage current is, for example, 17 mA. The value of the cut-off threshold for the high-frequency component between 1 kHz and 150 kHz of the leakage current is, for example, 220 mA. The value of the cut-off threshold for the high frequency component at 10 kHz of the leakage current is, for example, 200 mA.

When the estimate of the low frequency component of the leakage current exceeds its cut-off threshold of 17mA, the ECU calculator immediately cuts the current load, that is to say within 60ms (milliseconds) after this comparison, because this low frequency component is considered dangerous for a user. In the case of such an immediate interruption of the load, the ECU computer controls resumption of the interrupted load in the optics of not stopping the load only for a transient disturbance. However, after three power outages, the ECU computer no longer restarts the load that has been interrupted. When at least one of the other components of the leakage current exceeds its cut-off threshold without the low-frequency component of the leakage current exceeding its cutoff threshold, then the ECU calculator cuts off the current load after only two minutes if this situation continues, because these leakage current components are not immediately dangerous for the user, but blind the circuit breakers upstream of the VE vehicle power supply. In the case of such a non-immediate interruption of the load, the ECU computer controls resumption of the interrupted load, within the limit of four load interruptions, that is to say a fourth recovery of the load. is not allowed. The main computer ECU vehicle VE also implements the charging method according to the invention in a software manner.

With reference to FIG. 2, steps of the charging method according to the invention, in the case of a fault in the leakage current that does not lead to an immediate interruption of the current charge, are represented in the form of a algorithm comprising steps E1 to E6. Stage E1 is the starting of a load of the traction battery at a rated load power Ft, possibly negotiated with a charging terminal on which the vehicle VE is connected. This nominal charging power Pt corresponds to the nominal power of charge, that is to say to the load power expected to supply the battery when no mains disturbance or fault of the charger is detected. This nominal power Pt possibly varies as the battery is charged. Step E2 is a comparison of at least one estimate of a component Ifude of a leakage current at a cut-off threshold S'upure- This cutoff threshold depends on the nature of this component Ifude the leakage current as described upper. This step is performed continuously by the charging system S as soon as a BAT battery charge is in progress. More precisely in this step E2, the ECU calculator compares each estimate of leakage current components measured by the measuring and filtering device DMF, to their respective cut-off thresholds. It should be noted that, alternatively, it is possible to use other estimates of components of the leakage current, for example an estimate of the absolute value of the leakage current, or integration of the leakage current over a range. frequency, is also to be considered as an estimate of a component of the leakage current. In this case of use of the charging method according to the invention, it is assumed that at least one estimate of a component of the leakage current measured by the device DMF, other than the estimation of the low frequency component, is greater than or equal to its cutoff threshold for a first predetermined duration of 10 seconds. Step E3 is then the reduction of the charging power in progress charging (i.e. the charging power in use, or current charging power) to a lower charging power, after which this first predetermined duration. This step E3 does not take place if, before the end of the first predetermined duration, all the estimates of the components of the leakage current stabilize below their respective breaking thresholds. This step E3 is preceded by a step of calculating the lower load power to be applied, not shown, in which the current load power is multiplied P - therefore charges the rated load power Pt if it is a first step of decreasing, by a gain of less than one, equal to 0.9 in this embodiment of the invention.

Following this step E3: - If the ECU computer determines during the step E2 executed continuously, that an estimate of a component of the leakage current other than its low frequency component continues to be greater than or equal to its cut-off threshold during a second predetermined duration, then iterates the reduction step E3. During this iteration, a new lower load power equal to, for example, 0.9 times the charging power in charge is calculated. As a variant, the new lower load power is equal to 0.8 times the current charging power P - load, the gain less than one used in this step E3 being configured to converge the charging power in progress Charging to a power charge generating an acceptable leakage current in less than ten iterations. The second predetermined duration is set at 10 seconds in this embodiment of the invention, but alternatively it is less than or greater than the first predetermined duration. If the ECU calculator determines during the step E2 executed continuously, that before the end of the second predetermined duration all the estimates of the components of the leakage current stabilize below their respective cut-off thresholds, then a counter is initialized. To zero in a step E4. If step E4 has just been carried out, the ECU calculator rechecks during the step E2 executed continuously, that for a new duration of 10 seconds all the estimates of the components of the leakage current remain below their respective cut-off thresholds. . If this condition is verified then the counter "Counter" is incremented in a step E5, otherwise the reduction step E3 is repeated. - When the counter is incremented in a step E5, it is then verified in a step E6 that it is less than a predetermined value, for example 60, corresponding to a fifth predetermined time without fault leakage current of about 10 minutes. If this predetermined value is reached, then the current charging power Pcharge is increased to its nominal power Ft because the mains disturbances may have disappeared, and the charging process is resumed just after the starting step E1. If this predetermined value is not reached then it loops back to the verification step E2 immediately following step E4.

With reference to FIGS. 3 and 4, steps of the charging method according to the invention, in the case of a leakage current fault resulting in an immediate interruption of the current charge, are represented in the form of an algorithm comprising steps E1 to E3 and E7 to E1.

Step E1 is the start of a load of the traction battery at its nominal load power Finit, possibly negotiated as explained above with reference to FIG. 2. Step E2 is the comparison of each of the component component estimates. leakage current measured by the DM F measuring and filtering device, at their respective cut-off thresholds, as described above with reference to FIG. 2. However, it is assumed in this new use case of the charging method according to the invention , that the estimate of the low frequency component of the leakage current is greater than or equal to its cutoff threshold.

The next step E7 is then a first cut of the current load controlled by the ECU computer. The next step E8 is a first load recovery, also controlled by the ECU computer, at the rated load power Finit. This step E8 is followed by a comparison step E2 of the estimation of the low frequency component IBF of the leakage current, delivered continuously by the measuring and filtering device DMF, at its cutoff threshold. If the estimate of the low-frequency component is greater than its cut-off threshold, then the ECU calculates a second cut of the load in a step E9, otherwise the ECU calculator continues the step E2 of comparison of the estimates of the components of leakage current and the current load is not interrupted. When step E9 of second load break is performed, the ECU calculator in a step El 0 analyzes the deviations of a signal for measuring the frequency of the supply of the external electrical network. For this, the ECU calculator uses a phase-locked loop implemented, for example, in the DMF measurement and filtering device, and compares a standard deviation 4 f of variation of the measurement signal of the supply frequency of the supply network. compared to its theoretical value of 50Hz, provided by the phase locked loop. Indeed, the inventors have noticed that when the power supply network is disturbed, even if its average frequency is still 50 Hz, the phase-locked loop measuring its frequency has a signal at its output that has much greater oscillations with respect to this frequency. average value only when the network is not disturbed. If this standard deviation 4f is less than a disturbance threshold Sf, then the charging procedure is continued according to the steps following the reference A. If this standard deviation 4f is greater than the disturbance threshold Sf, then the power of the load in progress P - charge at 0.2 times the nominal load power Pt according to step E3, which is performed during a step El 1 of second recovery of the load that has just been interrupted. The threshold Sf disturbance is set for example at about 5Hz.

It should be noted that in this step El 0 the standard deviation is used, but alternatively it is possible to use the average of the frequency deviations or their maximum amplitude. This reduction step E3 therefore does not take place if the network has not been identified as very disturbed in step E10. Indeed, in the case of an undisturbed network, the steps referenced A shown in FIG. 4 consist of: a step E12 of second recovery of the load at its nominal power Pt, followed by a comparison step E2, in particular estimation of the low-frequency component IBF of the leakage current, delivered continuously by the measuring and filtering device DMF, at its cut-off threshold Scoupure. If the estimate of the low-frequency component is greater than its cut-off threshold, then the ECU calculator permanently interrupts the load in a step E13, otherwise the ECU calculator continues the comparison step E2 of the leakage current component estimates. and the current load is not interrupted. When the supply network has been identified as very disturbed in step El 0 and the load recovery step El 1 has been accompanied by a step E3 of reducing the load power in progress P load, then we go to the steps referenced B shown in FIG. 4. The step El 1 is followed by a comparison step E2, in particular the estimation of the low frequency component IBF of the leakage current, delivered continuously by the measuring device DMF. and filtering, at its cutoff threshold Scoupure. If the estimate of the low frequency component is greater than its cutoff threshold Scoupure, then the load is permanently interrupted according to the step E13, otherwise we go to the step E14 of analyzing the difference Al between the estimate of the IBF component continuously delivered by the measuring and filtering device DMF and the cutoff threshold Scoupure. If in the analysis step E14 the ECU calculator determines that the difference A1 is greater than a margin M of security for a third predetermined duration, set at 10 seconds in this embodiment of the invention, then we go to a step E15 of moderate increase in charging power in progress Charge, that is to say, it multiplies for example by 1.1. Alternatively, the third predetermined duration is less than or greater than 10 seconds.

If, on the contrary, at the analysis step E14, the ECU calculator determines that the difference A1 does not remain above the safety margin M over the third predetermined duration, then the ECU calculator continues to compare the estimates of the components. leakage current according to step E2 and the analysis step E14 if no immediate cut is necessary. When the step E15 of increasing the charging power in progress Pcharge is carried out, the comparison step E2 is continued, in particular of the IBF component with its cut-off threshold. If the estimate of the IBF component is less than its cutoff threshold S'upure, then we move to a new step E16 analysis of the gap Al between the estimate of the IBF component continuously delivered by the measuring device and filtering DMF and the cutoff threshold Scoupure If, on the contrary, the estimate of the IBF component is greater than its cut-off threshold S'upure, then the ECU computer controls a third cut of the load in a step E17. The step E17 of breaking is followed by a step of recovery E18 of the load at the load power preceding the last step of increase E15, that is to say that the load is taken again at the load power in progress Charge divided by 1.1. Step E18 then continues according to the comparison step E2 which precedes the end-of-charge step E13, performed when the squelch cut-off threshold is crossed again. In the new analysis step E16, if the ECU calculator determines that the difference A1 is greater than a margin M of security for a fourth predetermined duration, set at 10 seconds in this embodiment of the invention, then we return in step E15 of moderate increase of the load power in progress P - charge- This increase is limited to the attainment of the rated load power Pt which can not be exceeded. As a variant, in this step E16 the fourth predetermined duration is fixed at a duration well in excess of 10 seconds, for example 10 minutes, and if at the end of this fourth predetermined duration the difference A1 remains greater than a safety margin M, then it returns to the rated load power Pt during a last new step E15. The margin M of security is fixed for example to 4mA approximately. If, on the other hand, at the new analysis step E16, the ECU calculator determines that the difference A1 does not remain above the safety margin M over the fourth predetermined duration, then the ECU calculator returns to the previous step E2. in particular comparing the IBF component with its cutoff threshold Scoupure- Other embodiments of this embodiment of the invention are of course conceivable. Thus, for example, in the case of a fault in the leakage current resulting in an immediate interruption of the load, it is possible to carry out the reduction step E3 as soon as the load is taken up again, and / or to consider a single step E15 increase the load, so as to return directly to the nominal power Pt load after a determined third period of several minutes. It should be noted moreover that each step E3 for decreasing the current charging power, and / or each step E15 for increasing the current charge is accompanied by an update of a calculation of the time of charge of the battery, which is communicated to the user via a man-machine interface.

Claims (14)

REVENDICATIONS1. A method for charging a traction battery (BAT) of a vehicle (VE) with at least partially electrical traction, said vehicle (VE) comprising a non-isolated charger (CNI) able to recharge said traction battery (BAT) at from an external electrical network (RES), said method comprising: - a step (E1) of starting a charge of said traction battery (BAT) at a nominal charging power (Finit), - a step (E2 ) of comparing at least one estimate of a component (Ifuite) of a leakage current to a cutoff threshold (pure Scou), said method being characterized in that it further comprises a step (E3) of decreasing a charging power in progress (Pcharge) to a charging power lower than said current charging power, when said at least one estimate is greater than or equal to said cut-off threshold (S'u pure) - 2. The charging method as claimed in claim 1, characterized in that in said comparison step (E2) several component estimates of said leakage current chosen from a list comprising a DC component, a low frequency component lower than 400 Hz, a high frequency component between 1 kHz and 150 kHz, and a high frequency component corresponding to the charger hash frequency, at corresponding cut-off thresholds, said step (E3) of reduction being carried out as soon as one of said estimates is greater or equal to its cutoff threshold (pure Scou) - 3. Charging method according to claim 1 or 2, characterized in that said step (E3) for reducing the current charging power (Pcharge) is performed only when said at least one estimate remains above said cutoff threshold (S'upure,) for a first predetermined duration and when said at least one estimate corresponds to a component of said leakage current which is a DC or high frequency component of said leakage current. 4. Charging method according to any one of claims 1 to 3, characterized in that said step (E3) for decreasing the current charging power is iterated when said at least one estimate remains above said cutoff threshold (S'upure,) for a second predetermined duration and when said at least one estimate corresponds to a component of said leakage current which is a DC or high frequency component of said leakage current. 5. A charging method according to claim 4, characterized in that said step (E3) for reducing the current charging power is preceded by a step of calculating said lower charging power, wherein said power is multiplied by charge in progress (Charge) charge) by a gain less than one capable of converging said current charging power (Pcharge) to a charging power inducing leakage current component values lower than their respective cutoff thresholds less ten iterations. 6. Charging method according to claim 1 or 2, characterized in that when said at least one estimate corresponds to an estimate of a low frequency component (IBF) of the leakage current, said comparison step (E2) is followed by a step (E7) of cutting the load of said traction battery (BAT), and said step (E3) of reduction is carried out during a step of recovery (E11) of said load to said lower load power . 7. The charging method as claimed in claim 6, characterized in that said step (E3) of decreasing said current charging power (Pcharge) is preceded by a step (E10) of frequency difference analysis (4f ) a signal for measuring the supply frequency of the external electrical network (RES), and is conditioned by said analysis step (E10), said step (E12) of taking up said load being carried out without said step of decreasing when said frequency deviations are below a disturbance threshold. The charging method according to claim 7, wherein said step (El 1) of recovery of said load is followed by a step (E14) of analyzing the difference (AI) between said estimate of a component of low frequency current (IBF) and said cutoff threshold (pure Scou) - The charging method according to claim 8, wherein when said deviation (AI) between said estimate and said cutoff threshold (S'upure) 1 is greater than a safety margin (M) for a third predetermined duration, said step (E14) is followed by a step of increasing (E15) the current charging power (P, - charge). 10. The charging method as claimed in claim 9, characterized in that said step of increasing (E15) the current charging power is iterated when said difference (AI) remains above said safety margin (M) during a fourth predetermined duration. 11. A charging method according to claim 9 or 10, characterized in that when said step (E15) of increase is followed by a step (E17) of cutting said load, said step (E17) of cut is followed by a step (E18) of taking back said load at the charging power preceding the last step of increasing (E15). 12. A charging method according to any one of claims 1 to 11, characterized in that when said at least one estimate remains below said cut-off threshold (S'upure,) after said step (E3) decrease for a fifth time predetermined, said decreasing step (E3) is followed by a step of increasing (E15) the current charging power to said rated load power (Finit). 13. System for charging (S) a traction battery (BAT) of a vehicle (VE) with at least partially electrical traction, said vehicle (VE) comprising a non-isolated charger (CNI) able to recharge said battery of traction (BAT) from an external electrical network (RES), said system (S) being characterized in that it comprises means for implementing the charging method according to any one of claims 1 to 12. 14. Computer program comprising instructions for implementing the charging method according to any one of claims 1 to 12, when executed on one or more processors. 20