TECHNICAL FIELD

The present disclosure generally pertains to testing of rechargeable battery cells, commonly called secondary cells. More specifically, the disclosure relates to diagnosing secondary cells arranged in batteries of electrical vehicles (EV) using EV chargers.

BACKGROUND

Electric vehicles, EV, with electric drive are becoming increasingly important technically and economically. In operation, EVs are supplied with energy from batteries comprising one or more battery packs. One battery pack typically comprises a plurality of interconnected cells. The cells are secondary cells, i.e., rechargeable electric cells that can convert chemical energy into electrical energy by a reversible chemical reaction. For simplicity, the secondary cells are herein referred to as simply cells.

The usability and safety of electric vehicles depend to a considerable extent by the State of Health, SOH, of their batteries. Battery diagnosing of an EV is typically performed on the entire battery by connecting the EV to test equipment. Hence, individual cells of battery packs can typically not be diagnosed after they are put in use, as diagnosing is typically performed on battery packs or on the entire battery.

As of today, individual cells are typically tested and traced before they are put into use, as illustrated in Fig. 1 . For example, manufacturing data d _man is typically collected in the cell factory during cell manufacturing 101. The manufacturing data may be provided to the Performance and Lifecycle, P&L, testing lab 102 to trace possible manufacturing problems. Data d _P&L , d _PA is also collected during P&L testing 102 and during battery/battery pack assembly 103. All this data, herein referred to as production data d _procL is available to the cell manufacturer. However, the currently used process, there is no traceability of individual cells once the cells are installed in EVs 1 . This can be devastating as health issues of individual cells typically affect SOH of the entire battery. Hence, there is a need for improved methods for tracing and diagnosing individual cells after they are put in use.

SUMMARY

The present disclosure aims at facilitating tracing and diagnosing of individual cells after they are put in use. This is achieved by the proposed method and control arrangement.

According to a first aspect the disclosure relates to a charging infrastructure for EVs, comprising a plurality of EV chargers and a control arrangement. The plurality of EV chargers are configured to record energy signatures representing energy consumed and/or supplied by the EV chargers while charging and/or discharging battery packs arranged in the EVs. The control arrangement is configured to obtain, production data associated with cells of the battery packs, to receive the energy signatures from the EV chargers, and to diagnose the cells by evaluating the obtained energy signatures and obtained production data according to predetermined cell health criteria. By analysing data recorded by EV chargers together with production data (such as with measurements performed during slurry production), it is possible to identify cells, or batches of cells, which may have poor performance or reduced expected lifetime. Hence, diagnosis may be performed without any additional hardware e.g. being embedded in the battery packs by a manufacturer. Instead, the charging infrastructure is utilised for real-life testing.

In some embodiments, the diagnosing comprises analysing energy signatures to determine whether charging and/or discharging meets charging criteria, such as energy consumption and/or charging time. By analysing energy signatures, an in particular changes in energy signatures, cells with low performance can be detected.

In some embodiments, the energy signatures comprise a two-dimensional representation of energy consumed during a time period when the battery packs are charged and/or discharged. Hence, unexpected changes in the two-dimensional representation may be an indication that a battery pack comprises one or more unhealthy cells. In some embodiments, the production data is used to identify health issues of individual cells of battery packs whose energy signatures indicate a health issue. Hence, the production data can be used to track down a health issue on battery pack level to one or more individual cells.

In some embodiments, the EV chargers comprise temperature sensors and wherein the control arrangement is configured to diagnose the cells based on temperature data representing temperature while discharging and charging the battery packs. Thereby, energy signatures may be compensated for variations that depend on temperature.

In some embodiments, the control arrangement is configured to jointly analyse signatures of a plurality of battery packs in order to identify batches of cells, distributed among the battery packs, having poor health. By analysing many cells comprising cells grouped into batched, it is possible to get insights on quality on batch level and to track health issues of a pack down to the individual cell.

In some embodiments, the EV chargers are configured to collect the energy signatures during of charging and/or discharging cycles at different points in time during life cycles of the cells. Thereby, changes over time may be analysed and sudden or unexpected changes may be identified.

In some embodiments, the EV chargers are configured to collect the energy signatures while running specific tests including predefined cycles of charging and discharging. By running specific test programs more accurate diagnosis is possible, as the energy signature of a predefined test is more foreseeable than an energy signature of a normal charging.

In some embodiments, the EV chargers are configured to perform the specific tests based on detecting an unexpected change in an energy signature recorded during normal charging of a battery packs. Hence, specific tests may be performed on demand, when an energy signature recorded normal charging indicate that something may be wrong.

In some embodiments, the control arrangement is configured to diagnose the cells by comparing several energy signatures collected at different times, whereby abnormal changes can be detected. Hence, previously recorded energy signatures may be used as reference data when evaluating an energy signature.

In some embodiments, the control arrangement is configured to diagnose the cells by evaluating properties of production data and wherein the evaluating comprises evaluating whether parameters of the production data are within a certain distance from acceptance levels, wherein data within the certain distance indicate increased risk of a health issues. If parameters of production data are close to acceptance levels, this may be an indication that a cell the root cause of a suspicious energy signature.

In some embodiments, the control arrangement is configured to, upon the diagnosing indicating a health issue of a cell, identifying other cells associated with the same production batch that may be affected by same health issue. If one bad performing cell is identified and there are more cells of the same batch that have also been flagged as poorly performing, then one might suspect that something has gone wrong in that particular batch.

According to a second aspect the disclosure relates to a corresponding method for diagnosing individual cells of battery packs arranged in electrical vehicles, EVs. The method comprises recording, by the EV chargers, energy signatures of the respective battery packs, wherein the energy signatures represent energy consumed and/or supplied by the EV chargers while discharging and charging the battery packs using the EV chargers and diagnosing the cells by evaluating the obtained energy signatures and obtained production data according to predetermined cell health criteria. The method is associated with the same effects as the charging infrastructure according to the first aspect.

In an alternative, the second aspect relates to a method for diagnosing individual cells of battery packs arranged in electrical vehicles, EVs. The method comprises obtaining, production data associated with cells of the battery packs. The method further comprises, upon the EVs being connected to EV chargers for charging the cells, obtaining, from the EV chargers, energy signatures of the respective battery packs. The energy signatures are recorded while discharging and charging the battery packs using the EV chargers. The method further comprises diagnosing the cells by evaluating the obtained energy signatures and obtained production data according to predetermined cell health criteria.

In some embodiments, the diagnosing comprises identifying, based on the production data, health issues of an individual cell of a battery pack whose energy signatures indicate a health issue.

In some embodiments, the method comprises measuring temperature while discharging and charging the battery packs and diagnosing the cells based on the temperature.

In some embodiments, the diagnosing comprises jointly analysing signatures of a plurality of battery packs in order to identify batches of cells, distributed among the battery packs, having poor health.

In some embodiments, the recording comprises collecting energy signatures during of charging and/or discharging cycles at different points in time during life cycles of the cells.

In some embodiments, the recording is performed while the EV chargers are running specific tests including predefined cycles of charging and discharging.

In some embodiments, the recording comprises performing the specific tests based on detecting an unexpected change in an energy signature recorded during normal charging of a battery packs.

In some embodiments, the diagnosing comprises comparing several energy signatures collected at different times, whereby abnormal changes can be detected.

In some embodiments, the diagnosing comprises evaluating properties of production data and wherein the evaluating comprises evaluating whether parameters of the production data are within a certain distance from acceptance levels, wherein data within the certain distance indicate increased risk of a health issues.

In some embodiments, the method comprises, upon the diagnosing indicating a health issue of a cell, identifying other cells associated with the same production batch that may be affected by same health issue. According to a third aspect, the disclosure relates to a control arrangement configured to control the methods according to the first aspect.

According to a fourth aspect, the disclosure relates to a computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method according to the first aspect.

According to a fifth aspect, the disclosure relates to a computer-readable medium comprising instructions which, when executed by a computer, cause the computer to carry out the method according to the first aspect.

BRIEF DESCRIPTION OF THE DRAWINGS

The embodiments disclosed herein are illustrated by way of example, and by not by way of limitation, in the figures of the accompanying drawings. Like reference numerals refer to corresponding parts throughout the drawings, in which

Fig. 1 illustrates diagnosing of cells during cell manufacturing and assembly according to prior art.

Fig. 2 illustrates a battery pack comprising cells.

Fig. 3 illustrates the cell manufacturing process performed in the cell factory of Fig. 1 in further detail.

Fig. 4 illustrates a charging infrastructure for electrical vehicles.

Fig. 5 conceptually illustrates the proposed technique of diagnosing cells through EV charging.

Fig. 6A and 6B illustrates energy signatures of healthy and un-healthy cells.

Fig. 7 is a flow chart of the proposed method for diagnosing cells.

Fig. 8 illustrates an example implementation of the proposed method

Fig. 9 illustrates an EV charger for use by the proposed method.

Fig. 10 illustrates a control arrangement configure to perform the proposed method.

Fig. 11 illustrates the slurry production process of step a) in Fig. 3 in further detail.

DETAILED DESCRIPTION

In conventional Electric Vehicles, EV, the battery can be charged via a charging connector. EV charging may also be utilized as a mechanism to monitor the electrical performance of the battery once the vehicle is in the ownership of the end consumer. For example, patent publication CN202770974 II shows an electric vehicle power battery automatic test and diagnosis system. The system comprises a charging and discharging device. In the maintenance process of the battery pack, the system can determine a life state of the battery pack, diagnose problems in the battery pack and provide a comprehensive and effective guidance for a maintenance operation. Also acquiring actual available capacity can be performed. However, this kind of testing can only be used to individual test battery packs, and not individual cells.

Due to foreseen energy demand caused by extensive use of electric vehicles, improved solutions for smart EV charging are currently being developed. Smart EV charging is the intelligent charging of EVs, where charging can be shifted based on grid loads and in accordance with the vehicle owner’s needs. For example, a utility can offer EV owners monetary and/or non-monetary benefits in exchange for enrolment in a program that permits controlled charging at the times when curtailment capacity is needed for the grid. In the future, residential chargers may also be smart chargers, whereby the EVs are charged in the most beneficial manner from an economical and environmental perspective.

The introduction of smart charging systems implies that the charging can be controlled by a third party, such as by the vehicle and/or battery manufacturer. In view of this, the inventor has realized that smart charging may also solve post-production traceability problems mentioned above. More specifically, problems detected during smart charging can be correlated with production information to achieve more reliable information about what has gone wrong. In other words, smart charging makes it possible, not only to perform diagnosing of batteries at suitable times, but also to collect data from such diagnosing to perform more advanced diagnosing on cell level.

This disclosure proposes a technique that makes it possible to trace individual cells not only through P&L testing and battery pack assembly, but also after the cells have been put in use. The technique is based on the insight that the production, and in particular the slurry production process, is exposed to many variables that may affect slurry quality and that there is a strong relationship between poor cell performance and production, in particular poor slurry quality. More specifically, it is proposed to collect data from electrical vehicles in the field using EV chargers and to analyse the collected data together with production data (such as with data from measurements performed during slurry production). In this way, it will be possible to identify batches of cells that have inferior performance or reduced expected lifetime. Thereby, the root cause of an identified health issue may be identified. In the long run, this will make it possible to identify poor batches of cells (e.g., poor slurry batches) and to recall batteries that are expected to fail. The data can also be used to improve production, such as for tuning acceptance thresholds in P&L testing.

Embodiments of the present disclosure will now be described more fully hereinafter, with reference to Figs 2 to 10. The same reference numbers are used throughout the figures. The invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided by way of example so that this disclosure will be thorough and complete, and will fully convey the scope of the invention to those persons skilled in the art.

Fig. 2 illustrates a battery pack 20 comprising a plurality of secondary cells 21 , hereinafter simply called cells 20. An EV may comprise one or more battery packs 20. The EV can be connected to an EV charger 104 (fig. 9) in order to charge (and discharge) the battery packs 20.

Cell manufacturing is a complicated procedure involving many steps, where something may go wrong. Fig. 3 illustrates some main steps of cell manufacturing process in further detail. In step a), electrode slurry is produced. The slurry is a dispersion composed of conductive additives, polymer binders, and electrochemically active material particles that serve as reservoirs for lithium. The slurry is then coated onto conductive substrates and dried to form porous electrodes, step b). In the calendering, step c) the electrodes are compressed by passing through pairs of heated rolls, called calenders. Thereafter, notches or electrode tabs are formed along long sides of the electrode sheets, step d). The electrode tabs are used for forming and/or attaching the terminals of the cell. The cells are then assembled, step e). More specifically, the electrode sheets are rolled with an isolating layer in-between and inserted into enclosures that are filled with electrolyte before they are sealed. The Formation and Aging, F&A, step f) process is the last step in the cell manufacturing process. In F&A the cells 20 go through a series of charging, discharging and aging cycles to activate the working materials and evaluate the cells performance.

In all the steps, data may be collected and saved. In manufacturing, data may not be assigned to individual cells, but rather to batches of cells 21 or battery packs 20. For example, various data is collected in the slurry production process (step a). The production data is typically stored by the manufacturer and may involve various data such as sensor measurements (temperature, humidity, conductivity etc.), staff data, hardware data (which machine), time data, etc. When cells 21 are produced, the slurry batch used is registered together with cell identifiers for traceability. Some production data, such as F&A data may be registered for individual cells. The data from the various steps comprises information that is valuable when evaluating why an individual cell is not performing as expected, as it may indicate deviations in the production process.

Fig. 4 illustrates a charging infrastructure 100 for electrical vehicles, EVs, 1 configured to diagnose individual cells 21 of battery packs 20 arranged in EVs 1. The charging infrastructure 100 comprises a plurality of EV chargers 104 and a control arrangement 10. When the EVs 1 need to charge their battery packs they drive to one of the EV chargers 104 for charging. The EV chargers 104 may be private chargers, typically only used by a limited number of EVs or public chargers. The EV chargers 104 may be controlled by one charging supplier, or by several suppliers in cooperation.

The EV chargers 104 are configured to charge EVs. However, the EV chargers may in addition be configured to discharge the EVs (e.g. during night) for test purposes, or simply to utilise the energy. For example, the EV chargers 104 are configured to perform specific tests involving predetermined cycles of charging and discharging. The EV chargers 104 may also be configured to record various data. In this example, the EV chargers 104 are configured to record energy signatures representing energy consumed and/or supplied by the EV chargers 104 while charging and/or discharging the battery packs 20, see Fig. 6A. The control arrangement 10 is configured to control operation of a plurality of EV chargers. The control arrangement 10 may handle power supply and billing for the charging. It is herein proposed that the control arrangement 10 is also configured to receive production data d _prod from a manufacturer of the cells 21. It is herein also proposed that the control arrangement 10 is configured to receive the energy signatures 61 , 62 recorded by the EV chargers, and to diagnose the cells by evaluating the obtained energy signatures and obtained production data according to predetermined cell health criteria, as will be further explained in Fig. 7 and 8.

Fig. 5 conceptually illustrates the proposed technique of diagnosing cells 21 through EV charging. The proposed technique involves analysing production data, d _procL , collected during production of battery packs 20 (such as during manufacturing 101 , P&L testing 102 and battery pack assembly 103 (see Fig. 1 )) together with data d _EV captured by EV chargers 104. Typically, data is collected from a plurality of EV chargers 104 and from a plurality of EVs 1 . A huge amount of data may be collected representing charging and discharging of a plurality of vehicles over time. By collecting a lot of data from many vehicles over time, patterns representing normal behaviour of healthy cells may be identified, as well as patterns of unhealthy cells. The analysis is performed by a control arrangement 10, which may be a part of a virtual power plant 105. More specifically, the technique comprises monitoring the battery packs 20 through EV chargers 104 to spot anomalies through the energy signature and trace/diagnose issues associated with abnormal energy signatures within the manufacturing process. The technique can additionally or alternatively be used to initiate trace for product recalls/recycling.

In the proposed method energy signatures are analysed. An energy signatures are basically a curve that illustrate energy consumed or supplied, for example in a two- dimensional space. In other words, the energy signatures 61 , 62 comprises a two- dimensional representation of energy consumed during a time period when the battery packs 20 are charged discharged and/or charging. For example, an energy signature may represent energy consumed (or supplied) by the battery pack during one (or a few) charging and/or discharging cycle(s). Various unexpected changes in energy signatures of cells 21 may indicate that something is wrong. It is expected that energy signatures that indicate that capacity of a battery pack 20 has suddenly changed will be the strongest indicator of a battery pack 20 having poor health (i.e. a health issue). For example, if a battery pack 20 is fully charged too fast, it indicates that capacity has gone down. This is illustrated by the diagram in Fig. 6A which illustrates two cycles of discharging a batter pack to 0% and thereafter charging the pack to 100%, while measuring time/energy consumption. In the diagram the y-axis is consumed power in kWh and the x-axis represent time. In other words, when the EV is charging it consumes 7 KW, provided it is a 7 KW charger. For a 11 KW charger it would be 11 KW etc.

In this example, the dashed curve 62 represents an energy signature of a healthy battery pack 20, while the solid curve 61 represents an unhealthy battery pack 20. The first declining line is attempting to show that the remaining charge being drawn out of the battery, in order to start the charging from 0% state of charge. The battery should then rest after the discharge process, which is illustrated by a zero consumed power. Thereafter, energy consumption raises as the starting begins and stays at about 7 kW during the charging, which takes several hours. The unhealthy cell 61 stays at 7 kW for a shorter period of time, than the healthy cell. In other words, in this example, the solid curve (unhealthy 61 ), consumes much less energy than expected (healthy 62) for a healthy battery pack 20, which indicate that the available capacity of the battery pack 20 is reduced and therefore the time to charge up the battery pack 20 is reduced. Hence, the solid curve indicates a failure or issue with performance and life of the battery pack 20. The drop in energy consumption could alternatively or in addition manifest itself as a lower power consumption. In this example, a power consumption of 6,5 kW during the charging could be an indicator of an error. Hence, drop in energy consumption is one key indication - but it is expected that there will be various different signatures that may indicate problems.

It should be noted that every decrease in power consumption does not necessarily mean that there is an error, as a slow decrease in energy consumption KWH corresponds to natural ware. However, an energy drop above a certain level could typically be used as an indicator of premature failure of a cell. Data for determining what is normal wear and what is an error can be collected by studying behaviour of a plurality of cells 21 and/or battery packs 20 over time. In this context, battery packs 20 with deviating behaviour would be considered suspicious Fig. 6B illustrates an example where energy signatures recorded every month are compared. In this example there is a significant change in energy consumed (dotted line) the fifth month. Such a sudden change could be an indicator of an error, which differs from normal wear which would typically be visible as a slow progressive change.

Fig. 7 is a flow chart of the proposed method for diagnosing individual cells 21 of battery packs 20 arranged in electrical vehicles, EVs, 1 . The method is performed by a control arrangement 10 of the charging infrastructure. Parts of the method may be controlled by a control arrangement 40 of the EV chargers 104. The method may be implemented as a computer program comprising instructions which, when the program is executed by a computer (e.g., a processor in the control arrangement 10, 40 (Fig. 10)), cause the computer to carry out the method. According to some embodiments the computer program is stored in a computer-readable medium (e.g., a memory or a compact disc) that comprises instructions which, when executed by a computer, cause the computer to carry out the method. The method is performed as EVs comprising the cells 21 are put in use such that they are charges by EV chargers 104. One or more of the steps of the method may be repeated as more data from EV chargers 104 is available.

The method comprises obtaining S1 , production data d _prod associated with cells 21 of the battery packs 20. In other words, data recorded during production of the cells 21 or battery packs 20, are retrieved or collected for example from the cell manufacturer. Production data d _prod herein refers to any data collected or recorded in connection with production of battery packs before they are installed in the EVs. If the method is performed by a virtual power plant, VPP, 105 operated by the cell manufacturer, the data will be readily available. However, the data may also be stored in a standardised or agreed format and transferred between parties involved in manufacturing, operation, diagnosing and charging of cells 21 .

The production data may comprise various sorts of data, from different parts of the production process. In some embodiments, the production data d _prod comprises one or more of cell manufacturing data d _man , P&L test data d _P&L , packet assembly data d _ass and production staff data. Different data may be obtained from different sources. The more data that is available, the more information about what went wrong can be revealed. For example, production data d _procL indicative of sensor measurements captured during production and/or data indicative of particular hardware (such as information identifying a certain machine or tool) used in production is obtained.

The production data d _prod may also include sensor data from the slurry production process (Fig. 10). In other words, in some embodiments, the cell manufacturing data d _man comprises at-line and/or inline measurements of the slurry production process. In other words, data may be recorded either directly in the process (inline) or in the lab next door (at-line).

Some production data may be associated with individual cells 21. Other production data may be associated with a batch, or set, of cells, such as with all cells 21 comprising slurry from a certain slurry mixing batch, see Fig. 10. In some embodiments, each cell 21 belongs to one or more production batches, wherein the production data comprises batch data representing individual production batches of the cells 21. That a cell 21 belong to a batch means that it has some similar properties as other cells of the batch. The batch may be defined by use of the same hardware, the same staff or that it comprises a compound of the same batch (such as the same slurry batch). In some embodiments, a combination of individual cell data and batch data is used.

When the EVs 1 are operated their batteries (or battery packs 20) will regularly be charged. Some of the charging is typically performed by EV chargers 104, that may be connected to (controlled by or accessible by) the control arrangement 10. Such a connection may be indirect, i.e., via one or more intermediate nodes. Connected to does not imply that there has to be a continual connection, but it may be enough that the control arrangement 10 can communicate with the EV charger 104 (at least at some points in time) in order to exchange information. The EV chargers 104 may in addition to charging the cells 21 , perform specific tests including predefined cycles of charging and discharging, See Fig. 6A. This is typically performed during night, so that the operator is not negatively affected. The method further comprises, upon the EVs 1 being connected to EV chargers 104 for charging the cells 21 (or battery packs 20), recording S2 by the EV chargers 104, energy signatures of the respective battery packs 20. In other words, the method comprises, upon the EVs 1 being connected to EV chargers 104 for charging the cells 21 (or battery packs 20), obtaining, from the EV chargers 104, energy signatures of the respective battery packs 20. The energy signatures are recorded while discharging and/or charging the battery packs 20 using the EV chargers 104. The recording is performed using energy meters 41 arranged in the EV chargers 104. For example, energy signature data is communicated over a communication interface between the EV chargers 104 and the control arrangement 10. Data is typically collected from a plurality of EV chargers 104. Typically, energy signatures are collected from a plurality of charging and/or discharging cycles at different points in time during the life cycles of the cells 21 . For example, the energy signatures are collected every month. Hence, the energy signatures are collected in on ongoing manner during charging and discharging and communicated to the control arrangement 10. The energy signatures may also be stored in the EV chargers 104 (e.g. in control arrangement 40) before it is communicated to the control arrangement 10. Hence, for each cell 21 several energy signatures may be collected and compared (for example every month or week), whereby abnormal changes can be detected. This would typically take place when running specific test program. Alternatively, the energy signatures may be recorded during normal charging activity. A significant change in energy signature may possibly be detected during normal charging of the battery packs 20, which may trigger the EV charger 104 to perform more advances testing.

As explained before specific tests may be performed to enable more accurate comparison over time. In this way, the recorded energy signatures may provide more information than if energy signatures were only collected during normal operation as the predefined tests will normally result in predictable energy signatures, where variations are typically a result of a health issue. Also, a specific charging range, e.g. zero to a certain level (e.g. full), may reveal more variations caused by poor health, than just a random charging initiated by a user. In other words, in some embodiments, the recording S2 is performed while the EV chargers 104 are running specific tests including predefined cycles of charging and discharging.

The EV chargers 104 may also be configured to perform such a specific test program based on detecting a significant change in an energy signature recorded during normal charging of a battery packs 20. Hence, more detailed data may be obtained for specific cells based on demand, such as when a deviation is detected or when it is suspected that there is an error in a specific batch of cells 20. In other words, in some embodiments, the recording S2 comprises performing the specific tests based on detecting an unexpected change in an energy signature recorded during normal charging of a battery packs 20.

The data collected in the first two steps S1 , S2 is then used for diagnosing the cells 21 . In other words, the method comprises diagnosing S3 the cells 21 by evaluating the obtained energy signatures and obtained production data d _prod according to (or based on) predetermined cell health criteria. In other words, the cell health criteria correspond to a principle or standard by which health of the cells may be judged or decided. The diagnosing typically involves identifying correlations between poorly performing cells 21 , battery packs 20 or batteries and corresponding production data d _prod . The predetermined cell health criteria define how the obtained data shall be analysed. The predetermined cell health criteria may define rules or an evaluation method, as described in Fig. 8. The health criteria may consider properties of the battery pack, such as energy signature and cell age. In addition, the health criteria may consider properties of production data d _prod of the pack. In some embodiments, the health criteria comprise rules that can be used or applied to determine a state of health based on the obtained energy signatures and obtained production data d _prod , and possibly also other parameters. In other words, the health criteria can be used to trace the root cause of failing cells. For example, the diagnosing S3 may result in that a certain problem is detected. This may be crucial to determine the seventy of the problem and to take appropriate measures. The health criteria may also be referred to as diagnosing rules or health rules.

The diagnosing may be performed in various different ways. In some embodiments, the diagnosing S3 comprises analysing energy signatures to determine whether charging and/or discharging meets charging criteria, such as energy consumption and/or charging time. For example, a suspicious charging speed or capacity may be an indicator that something is wrong, and that further analysis of production data dprod is required. Hence, in some embodiments, the diagnosing S3 comprises evaluating properties of production data d _prod . Properties that may be relevant may be temperature, times, staff, and various sensor data associated with production. A factor that may influence the energy signatures is temperature, basically outside temperature during the charging. Hence, when diagnosing cells 21 the energy signatures captured at different points in time the curves may be compensated based on temperature. In other words, in some embodiments, the method comprises measuring temperature while discharging and charging the battery packs 20 and diagnosing S3 the cells 21 based on the temperature. Temperature is typically measured by a sensor 46 arranged in the EV charger 104.

Another way to identify the cell 21 causing problem of a battery packet is to analyse a large number of battery packs 20. By identifying many unhealthy battery packs 20, common features of the unhealthy packs 20 may be revealed. For example, it may turn out that many battery packs 20 comprises one or more cells 21 belonging to a certain batch. In such a scenario one may deduce the problem to the batch. In other words, in some embodiments, the diagnosing S3 comprises jointly analysing signatures of a plurality of battery packs 20 in order to identify batches of cells 21 , distributed among the battery packs 20, having poor health.

Some production data, such as test data, may be reviewed to see if it deviates from an expected or normal level, even if it is still above acceptance levels. In other words, in some embodiments, the evaluating comprises evaluating whether parameters of the production data d _prod are within a certain distance from acceptance levels, wherein data within the certain distance indicate increased risk of a health issues. This type of analysis may in the long run be used to change acceptance levels to avoid similar errors in the future.

The diagnosing may reveal S4 health issues of batteries or battery packs 20. However, to identify which cell(s) 21 are causing the issue the production data may be used. For example, production data of all cells 21 of an unhealthy pack are investigated to check which one may be the root cause of the identified problem. It is expected that unhealthy cells may have some parameters within the certain distance (i.e. close to failing) from acceptance levels. In other words, in some embodiments, the production data is used to identify health issues of individual cells 21 of battery packs 20 whose energy signatures indicate a health issue. If a health issue is detected in one cell 21 belonging to a particular batch, then it may be relevant to check the health of all affected cells, i.e., all cells of the batch. This may be the case, when the expected cause of a health issue is in the slurry production process. If there are already one or more other cells 21 of the same batch that have similar health issue, then one may draw the conclusion that reason was that the slurry batch was poor and that all cells should be called back and replaced. In other words, in some embodiments, the diagnosing S3 comprises, investigating if there are identified health issues of other cells 21 pertaining to the same production batches as the cells.

In the same way, if one cell 21 is bad and slurry production data looks suspicious, then it may make sense to flag the batch to track the health of other cells of the batch. In other words, in some embodiments, the method comprises, upon the diagnosing S3 indicating a health issue of a cell, identifying S5 other cells 21 associated with the same production batch that may be affected by same health issue.

Hence, health issues that may affect many cells 21 could be detected S4 by tracking all the cells 21 of the batch. For example, one may put one warning flag on the batch for each failure. When there is a certain number of warning flags on a certain batch, then one may call in all affected EVs to the workshop to exchange the cells 21 before they start performing badly. In the same way well performing cells may be flagged. If a health issue is detected and there are several other cells 21 in the same batch that are flagged as well performing, then one may conclude that the problem is not a batch issue but an individual problem of the battery/battery pack.

By analysing the production data d _prod , the cause of a health issue may be determined, and various measures may be taken depending on what is the cause. Sometimes it may be possible to identify a remedy, such as reconfiguration of battery or charging parameters. In other cases, it may be required to instruct the driver to take the EV to a service station. In other words, in some embodiments, the method comprises, upon the diagnosing indicating S3 a health issue of the cell 21 , identifying S6 remedy, based on one or more of; the obtained production data ( d _prod .), the measured energy signature and diagnosis of other cells of the same batch. In some embodiments, the proposed method is repeated every X month. For each repetition one may compare energy consumption and charging time with previous tests to determine health of pack. For example, if takes Y times less to charge then health of pack is suspect.

Fig. 8 illustrates an example implementation of the proposed method and in particular of the step of diagnosing S3 health of the secondary cells based on the obtained production data d _prod and energy signatures using predetermined health criteria. In other words, the evaluation of the predetermined health criteria may involve the following.

When one battery pack 20 has an energy signature that deviates more than a certain amount from an expected (normal) energy signature, the battery pack 20 is considered failing 81 . An 8D root cause investigation of available data is then initiated, based on available data. The first step 82 is to check Formation and Ageing, F&A, data and Performance and Lifecycle testing, P&L, data. One may for example start by analyzing F&A data of cells in the failing pack. Formation data is recorded after slurry production but before the cells are sent to the P&L lab. Data from formation gives good electrical insights about batches of cells, such as cell capacity, Direct current internal resistance, DCIR, cell voltage, Open-circuit voltage etc. Next, reference P&L data for representative slurry batches of cells 21 of the pack 20 are checked to see if there were any special circumstances in P&L testing. If the reference P&L data looks OK 83, the conclusion is that the problem is an isolated problem associated with the pack assembly or Battery Management System, BMS. The next step is then to diagnose 84 the battery package assembly data d _PA and BMS of this particular battery pack 20. The failure may also be considered a natural depreciation.

However, if the reference P&L data looks suspect 85, the conclusion is that there might be something suspicious with the production, such as with the slurry batch. The next step is then to check production data 86. In practice, all relevant production data that is available may be checked. The most important may be slurry production data. Other relevant data may be data recorded by inline quality control systems. If the available production data looks OK 87 the conclusion is that the problem is an isolated problem associated with the pack assembly or Battery Management System, BMS. The next step is then to diagnose 88 the battery package assembly data d _PA and BMS of this particular battery pack 20. The failure may also be considered a natural depreciation.

However, if the production data also looks suspect 89, then the production is the suspected cause of the failure 90. The remedy would in this case be to look up 91 all suspected EVs on the EV networks to investigate if they all have the same problem. If several cells of the same batch have the same problem, then all batteries may be replaced even if the degradation is not visible in all of them. The findings of the error are sent back to production to see if there is anything in the process that can be changed to avoid similar problems in the future.

Fig. 9 illustrates an EV charger 104 in further detail. The EV charger 104 comprises an energy meter 41 , DC protection circuitry 42, a communication interface 43, a communication interface 44, an EV connector 45 and one or more sensors 46 and control circuitry 40.

The energy meter 41 is typically a “MID meter”. A “MID meter” is a power meter device that enables recording energy consumption of a charging installation or a specific charger, in conformity with the Measuring Instruments Directive (MID). The energy meter 41 is configured to record energy signatures of batteries, or of individual battery packs 20, during charging (and discharging) of EVs.

The DC protection circuitry 42 comprises circuitry that provides protection against DC residual fault currents, and against AC residual fault currents. In some embodiments, the DC protection circuitry 42 comprises a Type B RCD.

The communication interface 44, enables the EV charger 104 to communicate with other units, such as for providing energy signatures recorded while discharging and charging battery packs 20 to the control arrangement 10 of the virtual power plant 105 (Fig. 5). The communication interface 44, may utilise any standardised or proprietary communication protocol, such as 4G, 5G or ethernet.

The EV connector 45 is basically a connector that connects the battery packs 20 in the EV to a power source/storage of the EV charger 104. The EV connector 45 may involve communication means whereby certain data may be communicated between the EV 1 and the EV connector 45.

The one or more sensors 46 are sensors configured to measure environmental parameters during charging (and discharging). In particular the one or more sensors comprises a temperature sensor, a humidity sensor, a motion sensor etc.

The control circuitry 40 is basically a computer that controls operation of the EV charger 104, such as charging and discharging of EVs 1. The control circuitry 40 comprises at least one processor 11 and memory 12. The control circuitry 40 is configured to identify battery packs being charged using electronic identifiers, scanned QR codes or similar for identification. The control circuitry 40 is configured to capturing and store (i.e. , record) accurate energy signatures and profiles of battery packs 20 of connected vehicles 1 using the energy meter 41 . The control circuitry 40 is configured to provide the captured signatures to the control arrangement 10, e.g., using the communication interface 44.

Fig. 10 illustrates a control arrangement 10 configure to perform or control the proposed method. The control arrangement 10 is for example located at premises provided or controlled by the cell manufacturer, such as at a virtual power plant, VPP, 105 used to charge electrical vehicles, EV, 1. A virtual power plant is a system that integrates several types of power sources to give a reliable overall power supply.

The control arrangement 10 comprises at least one processor 11 and memory 12. In general, the electronic user device 2, is configured to perform all embodiments of the method described herein. This might e.g., be achieved by the processor 11 executing software stored in the memory 202. More specifically, the control arrangement 10 is configured to obtain, production data associated with cells 21 of the battery packs 20. The control arrangement 10 is further configured to, upon the EVs 1 being connected to EV chargers 104 for charging the cells; obtain from the EV chargers 104, energy signatures of the respective battery packs 20, and diagnose the cells by evaluating the obtained energy signatures and obtained production data d _prod according to predetermined cell health criteria. Fig. 11 illustrates the slurry production process (step 101 a in Fig. 3) in further detail. Slurry production is a part of the cell production process that involves several steps and has been identified as a source of risk for SOH. Slurry production is therefore conceptually illustrated for better understanding of sources of potential cell failure and to illustrate how production data can be collected to detect such failures. The main raw material of the electrode slurry is the active component (or powder) containing lithium ions. In the slurry production process, the active component is mixed with other raw material such as binder and conductive material (e.g., carbon). For each raw material there may be a loading step, a mixing step and possibly also a buffer step.

Fig. 11 illustrate how the raw materials are first loaded in individual hoppers. In this simplified figure there is only one hopper per raw material, i.e., one binder powder hopper 31 , one active powder hopper 32 and one conductive powder hopper 33, but in a real implementation several hoppers may be used for each raw material. There may also be different types of the raw materials, such as different types of binder powder. The binder powder and conductive powder are mixed (with a solvent) in respective mixers (binder mixer 38 and conductive mixer 39) to produce binder and conductive paste. After the mixers there is typically a respective buffer step 34, 36, where binder and conductive paste is stored. In these intermediate buffers 34, 36 various production data may be collected.

The slurry is then produced by mixing agent powder with binder and conductive paste (and possibly solvent 30) in a slurry mixer 35. Various production data, may be redorded at the slurry mixer 35. If there are several slurry mixers 35, an identity of the mixer may also be recorded and stored. In addition, data associated with staff and hardware involved in the production may be registered.

The produced slurry is then stored in one or more intermediate slurry storages 37 before it is applied to electrode sheets. In the storage production data may also be recorded.

Throughout the slurry production process data is recorded and mapped to one or more batches. A batch may be defined in different ways. For example, a batch of slurry is the accumulative sum of slurry found in a storgare, after a refilling from the minimum to the maximum level before any consumption from this tank. However, a batch may also refer to the slurry being mixed as a batch in the PD mixer. Same applies to a batch of binder or conductive paste. With that said, each cell may be associated with one or more batches, depending on what slurry was used to produce the electrodes. During production each cell is uniquely identified and relevant batched are registered, such that corresponding production data can be collected at a later point in time.

Production sata that may be recorded include, but is not limited to, composition, chemical properties, conductivity, weight, level, storing time, temperature, humidity, etc. In addition, data associated with staff and hardware involved in the process may be registered.

The proposed technique has been described with reference to lithium-ion cells, but it should be appreciated that method for other types of cells including cells made from solid state materials, such as graphene. Such cells are expected to be more commonly used in the future.

The terminology used in the description of the embodiments as illustrated in the accompanying drawings is not intended to be limiting of the described method, control arrangement or computer program. Various changes, substitutions and/or alterations may be made, without departing from disclosure embodiments as defined by the appended claims.

The term “or” as used herein, is to be interpreted as a mathematical OR, i.e. , as an inclusive disjunction; not as a mathematical exclusive OR (XOR), unless expressly stated otherwise. In addition, the singular forms "a", "an" and "the" are to be interpreted as “at least one”, thus also possibly comprising a plurality of entities of the same kind, unless expressly stated otherwise. It will be further understood that the terms "includes", "comprises", "including" and/ or "comprising", specifies the presence of stated features, actions, integers, steps, operations, elements, and/ or components, but do not preclude the presence or addition of one or more other features, actions, integers, steps, operations, elements, components, and/ or groups thereof. A single unit such as e.g. a processor may fulfil the functions of several items recited in the claims.