Battery diagnosis method and battery diagnosis apparatus

Abstract

A battery diagnosis method according to the present invention includes the steps of: obtaining diagnosis target information including a target full cell profile of a battery cell associated with a first electrical stimulus; modifying, on the basis of a predetermined overpotential profile, the target full cell profile to be associated with a second electrical stimulus that is different from the first electrical stimulus; determining first diagnosis result information, which is a preliminary diagnosis result for the charge and discharge performance of the battery cell, by applying a diagnosis logic to the modified target full cell profile; and determining second diagnosis result information, which is a preliminary diagnosis result for the charge and discharge performance of the battery cell, by applying a diagnosis logic to the modified target full cell profile; and determining second diagnosis result information, which is a precise diagnosis result for the charge and discharge performance of the battery cell, on the basis of at least one preliminary diagnosis factor included in the first diagnosis result information by using a factor correction model.

Description

Battery diagnostic method and battery diagnostic device

[0001] The present invention relates to a battery diagnostic device and method for non-destructively diagnosing the charge / discharge performance of a battery.

[0002] This application claims priority to Korean Patent Application No. 10-2023-0165769, filed on November 24, 2023, and Korean Patent Application No. 10-2024-0065906, filed on May 21, 2024, the entire contents of which are disclosed in the specification and drawings of the aforementioned applications are incorporated herein by reference.

[0003] Recently, as the demand for portable electronic products such as laptops, video cameras, and mobile phones has rapidly increased, and the development of electric vehicles, energy storage batteries, robots, and satellites has been in full swing, research into high-performance batteries capable of repeated charging and discharging is actively being conducted.

[0004] Currently commercialized batteries include nickel-cadmium batteries, nickel-hydrogen batteries, nickel-zinc batteries, and lithium batteries. Among these, lithium batteries are receiving attention for their advantages of being able to charge and discharge freely, having a very low self-discharge rate, and having a high energy density, as they have almost no memory effect compared to nickel-based batteries.

[0005] In general, the actual charge / discharge performance of a battery may fall short of the normal charge / discharge performance due to reasons such as manufacturing defects or deterioration from use. Therefore, it is necessary to accurately diagnose the charge / discharge performance of a battery to improve the lifespan and safety of the battery.

[0006] Conventionally, measurements of a battery's voltage and capacity were recorded while a low-rate electrical stimulus (e.g., low-rate charge or discharge) was applied to the battery. Based on these recorded measurements, a full-cell profile representing the relationship between voltage and capacity was generated, which was used to diagnose the battery's condition. However, because the battery's capacity and voltage change slowly while a low-rate electrical stimulus (e.g., low-rate charge or discharge) is applied, battery diagnosis has the limitation of taking a long time.

[0007] In terms of shortening the diagnosis time, high-level electrical stimulation is naturally advantageous over low-level electrical stimulation. However, when a battery is subjected to high electrical stimulation (e.g., high-rate charging or discharging), the proportion of overpotential in the battery voltage can become excessively large. More specifically, the greater the current flowing through the battery, the more severe the polarization phenomenon, and the overpotential is caused by the polarization phenomenon. Since the battery voltage can be viewed as the sum of the open circuit voltage (OCV) and overpotential, the higher the level of electrical stimulation applied to the battery, the greater the difference between the battery voltage and the actual OCV.

[0008] Since the closer the battery voltage approaches the actual OCV, the more accurately the battery's charge / discharge performance can be diagnosed, overvoltage acts as a type of noise that reduces diagnostic accuracy. Therefore, the results of charge / discharge performance diagnostics based on full-cell profiles acquired using high electrical stimulation can significantly differ from the battery's actual charge / discharge performance.

[0009] The present invention has been devised to solve the above problems, and the purpose of the present invention is to provide a battery diagnosis method and a battery diagnosis device capable of simultaneously shortening the diagnosis time and ensuring diagnosis accuracy by obtaining charge / discharge information ('first target full cell profile' in the claims) by applying a high electric stimulus to a battery, and removing noise caused by overvoltage included in the obtained charge / discharge information using a machine learning-based factor correction model.

[0010] Other objects and advantages of the present invention can be understood through the following description and will be more clearly understood through the embodiments of the present invention. Furthermore, it will be readily apparent that the objects and advantages of the present invention can be realized by the means and combinations thereof set forth in the claims.

[0011] A battery diagnosis method according to one aspect of the present invention comprises the steps of: obtaining diagnosis target information including a target full-cell profile of a battery cell associated with a first electrical stimulus; modifying the target full-cell profile to be associated with a second electrical stimulus different from the first electrical stimulus based on a predetermined overvoltage profile; applying a diagnosis logic to the modified target full-cell profile to determine first diagnosis result information which is a preliminary diagnosis result for charge / discharge performance of the battery cell; and using a factor correction model, determining second diagnosis result information which is a precise diagnosis result for charge / discharge performance of the battery cell based on at least one preliminary diagnosis factor included in the first diagnosis result information.

[0012] The instantaneous voltage change induced in the battery cell when the second electrical stimulus is applied may be smaller than the instantaneous voltage change induced in the battery cell when the first electrical stimulus is applied.

[0013] The first electrical stimulation may be a charging current greater than or equal to a first current rate, and the second electrical stimulation may be a charging current less than or equal to a second current rate that is less than the first current rate.

[0014] The first electrical stimulation may be a discharge current greater than or equal to a first current rate, and the second electrical stimulation may be a discharge current less than or equal to a second current rate that is less than the first current rate.

[0015] The above overvoltage profile may represent a difference between a first reference full-cell profile and a second reference full-cell profile. The first reference full-cell profile may be predetermined as a correspondence between a capacity factor and a voltage of the reference cell while the first electrical stimulus is applied. The second reference full-cell profile may be predetermined as a correspondence between a capacity factor and a voltage of the reference cell while the second electrical stimulus is applied.

[0016] The step of modifying the target full-cell profile may generate the modified target full-cell profile by subtracting the overvoltage profile from the target full-cell profile.

[0017] The above diagnostic target information may further include temperature information of the battery cell measured during the application period of the first electrical stimulation. The temperature information may be input into the factor correction model together with the at least one preliminary diagnostic factor.

[0018] The above diagnostic target information may further include impedance information of the battery cell measured during the application period of the first electrical stimulation. The impedance information may be input into the factor correction model together with the at least one preliminary diagnostic factor.

[0019] The above factor correction model may be a machine learning model that has been trained using a training data set that includes pairs of first diagnostic result information and second diagnostic result information of each of a plurality of test cells having different charge / discharge performances.

[0020] The first diagnostic result information of each of the plurality of test cells may be previously acquired by applying the diagnostic logic to each of the plurality of modified test full-cell profiles. The plurality of modified test full-cell profiles may be individually acquired by modifying the plurality of first test full-cell profiles associated with the first electrical stimulation based on the overvoltage profile. The second diagnostic result information of each of the plurality of test cells may be previously acquired by applying the diagnostic logic to the plurality of second test full-cell profiles associated with the second electrical stimulation.

[0021] The step of determining the second diagnostic result information may be executed under the condition that the performance indicator of the factor correction model is evaluated to be above a threshold value.

[0022] The above battery diagnosis method may further include a step of transmitting a message to a user device notifying that additional learning of the factor correction model is required when the performance indicator of the factor correction model is evaluated to be less than the threshold.

[0023] The second diagnostic result information may include at least one type of diagnostic factor among a positive electrode participation start point, a positive electrode participation end point, and a positive electrode scale factor related to the charge / discharge performance of the positive electrode of the battery cell, as a precision diagnostic factor of the battery cell.

[0024] The second diagnostic result information may include at least one type of diagnostic factor among a negative electrode participation start point, a negative electrode participation end point, and a negative electrode scale factor related to the charge / discharge performance of the negative electrode of the battery cell, as a precision diagnostic factor of the battery cell.

[0025] The second diagnostic result information may include at least one type of diagnostic factor among a positive electrode loading amount related to the charge / discharge performance of the positive electrode of the battery cell, a negative electrode loading amount related to the charge / discharge performance of the negative electrode of the battery cell, and a NP ratio related to the charge / discharge performance of both the positive and negative electrodes of the battery cell, as a precision diagnostic factor of the battery cell.

[0026] According to another aspect of the present invention, a battery diagnosis device includes a data acquisition unit configured to acquire diagnosis target information including a target full-cell profile of a battery cell associated with a first electrical stimulus; and a control circuit configured to modify the target full-cell profile to be associated with a second electrical stimulus different from the first electrical stimulus based on a predetermined overvoltage profile. The control circuit is configured to apply a diagnosis logic to the modified target full-cell profile to determine first diagnosis result information, which is a preliminary diagnosis result for charge / discharge performance of the battery cell. The control circuit is configured to determine second diagnosis result information, which is a precise diagnosis result for charge / discharge performance of the battery cell, based on at least one diagnosis factor included in the first diagnosis result information, using a factor correction model.

[0027] The control circuit may be configured to generate the modified target full-cell profile by subtracting the overvoltage profile from the target full-cell profile.

[0028] The above factor correction model may be a machine learning model that has been trained using a training data set that includes pairs of first diagnostic result information and second diagnostic result information of each of a plurality of test cells having different charge / discharge performances.

[0029] A battery pack according to another aspect of the present invention includes the battery diagnostic device.

[0030] A battery system according to another aspect of the present invention includes the battery diagnostic device.

[0031] According to at least one embodiment of the present invention, even if the charge / discharge information of the battery cell being diagnosed (the "target full cell profile" in the claims) is acquired by applying a high level of electrical stimulation to the battery cell, the charge / discharge performance of the battery can be accurately diagnosed. Accordingly, compared to a diagnostic method using a low level of electrode stimulation (e.g., a low rate of charge or discharge), the time required to diagnose the charge / discharge performance of the battery cell can be shortened.

[0032] In addition, according to at least one of the embodiments of the present invention, by modifying the charge / discharge information so that an overvoltage component caused by high electrical stimulation is at least partially removed, and then analyzing the modified charge / discharge information ('modified target full-cell profile' in the claims), the accuracy of diagnosis of charge / discharge performance can be improved.

[0033] In addition, according to at least one of the embodiments of the present invention, by correcting preliminary diagnosis result information indicating charge / discharge performance determined from the modified charge / discharge information using a machine learning-based factor correction model, it is possible to secure precise diagnosis result information having a high degree of consistency with the actual charge / discharge performance of the battery.

[0034] The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description of the claims.

[0035] The following drawings attached to this specification illustrate preferred embodiments of the present invention, and together with the detailed description of the invention described below, serve to further understand the technical idea of ​​the present invention, and therefore, the present invention should not be interpreted as being limited to matters described in such drawings.

[0036] FIG. 1 is a drawing exemplarily showing the configuration of a battery system according to the present invention.

[0037] Figure 2 is a drawing referenced to explain the relationship between electrical stimulation and full cell profile.

[0038] FIG. 3 is a diagram schematically illustrating an overvoltage profile obtainable from the first reference full-cell profile and the second reference full-cell profile of FIG. 2.

[0039] FIG. 4 is a graph referenced to explain the relationship between the first target full-cell profile, the modified target full-cell profile, and the second target full-cell profile.

[0040] FIG. 5 is a graph for reference in explaining an example of each of a modified target full-cell profile, a second reference full-cell profile, a reference anode profile, and a reference cathode profile.

[0041] FIGS. 6 to 8 are drawings for reference in explaining an example of a procedure for generating a comparison full-cell profile according to diagnostic logic.

[0042] FIGS. 9 to 11 are drawings for reference in explaining another example of a procedure for generating a comparison full-cell profile according to diagnostic logic.

[0043] Figure 12 is a drawing referenced to explain the function of the factor correction model.

[0044] Figures 13 to 21 are diagrams for reference in explaining a training data set provided for learning a factor correction model.

[0045] Figure 22 is a diagram illustrating an example of the neural network structure of the factor correction model of Figure 12.

[0046] Figure 23 is a diagram showing an example of correlation coefficients between diagnostic factors that can be obtained through learning a factor correction model.

[0047] Figure 24 is a flowchart schematically illustrating a battery diagnosis method according to another embodiment of the present invention.

[0048] FIG. 25 is a flowchart schematically illustrating an additional battery diagnosis method related to the battery diagnosis method of FIG. 24.

[0049] Figure 26 is a drawing referenced in explaining the method according to Figure 25.

[0050] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the attached drawings. Prior to this, it should be noted that the terms and words used in this specification and claims should not be construed as limited to their conventional or dictionary meanings. Based on the principle that the inventor can appropriately define the concepts of terms to best explain his or her invention, they should be interpreted in a way that aligns with the technical spirit of the present invention.

[0051] Accordingly, the embodiments described in this specification and the configurations illustrated in the drawings are merely the most preferred embodiments of the present invention and do not represent all of the technical ideas of the present invention. Therefore, it should be understood that there may be various equivalents and modified examples that can replace them at the time of filing this application.

[0052] Terms that include ordinal numbers, such as first, second, etc., are used to distinguish one of the various components from the rest, and are not used to limit the components by such terms.

[0053] Throughout the specification, when a part is said to "include" a component, this does not exclude other components, unless otherwise stated, but rather implies that other components may be included. Furthermore, terms such as "unit" used throughout the specification mean a unit that processes at least one function or operation, and may be implemented using hardware, software, or a combination of hardware and software.

[0054] Additionally, throughout the specification, when we say that a part is "connected" to another part, this includes not only cases where it is "directly connected" but also cases where it is "indirectly connected" with other elements in between.

[0055] FIG. 1 is a drawing exemplarily showing the configuration of a battery system according to the present invention.

[0056] Referring to FIG. 1, the battery system (1) includes a system controller (2), a battery pack (10), a relay (20), an inverter (30), and an electrical load (40).

[0057] The charge / discharge terminals (P+, P-) of the battery pack (10) can be electrically connected to an inverter (30) and / or a charger (3) via a charging cable or the like. The charger (3) may be included in the battery system (1) or provided in a charging station.

[0058] The system controller (2) (e.g., ECU: Electronic Control Unit) is configured to transmit a key-on signal to the battery diagnosis device (100) in response to a start button (not shown) provided in the battery system (1) being turned to the ON position by a user. The system controller (2) is configured to transmit a key-off signal to the battery diagnosis device (100) in response to a start button being turned to the OFF position by a user. The charger (3) can communicate with the system controller (2) to supply charging power in a constant current charging mode, a constant voltage charging mode, and / or a constant power charging mode to the battery (11) through the charge / discharge terminals (P+, P-) of the battery pack (10).

[0059] The battery pack (10) includes a battery (11). The battery pack (10) may further include a battery diagnostic device (100).

[0060] The battery (11) includes at least one battery cell (BC). The battery (11) includes a plurality of battery cells (BC1 to BC N , N is a natural number greater than or equal to 2), these multiple battery cells can be connected in series, parallel, or a mixture of series and parallel.

[0061] The type of the battery cell (BC) is not particularly limited, as long as it is capable of repeated charging and discharging, such as a lithium ion cell. The battery cell (BC) may include at least one unit cell. A unit cell is an electrochemical device that can be independently recharged. When the battery cell (BC) includes multiple unit cells, the multiple unit cells may be connected in series, parallel, or a series-parallel combination. The battery cell (BC) may be a new battery cell that requires verification as to whether it is a good product, or a battery cell that has deteriorated after being verified as a good product and is no longer a new product. Hereinafter, the battery cell (BC) may be referred to as a "target battery cell" or a "target cell."

[0062] The relay (20) is electrically connected in series to the battery (11) via a power path connecting the battery (11) and the inverter (30). In Fig. 1, the relay (20) is illustrated as being connected between the positive terminal of the battery (11) and the charge / discharge terminal (P+). The relay (20) is turned on and off in response to a switching signal from the battery diagnostic device (100). The relay (20) may be a mechanical contactor that is turned on and off by the magnetic force of a coil, or a semiconductor switch such as a MOSFET (Metal Oxide Semiconductor Field Effect transistor).

[0063] The inverter (30) can convert direct current from the battery (11) into alternating current in response to a command from the battery diagnostic device (100) or the system controller (2).

[0064] The electric load (40) is driven using AC power from the inverter (30). For example, a three-phase AC motor (40) can be used as the electric load (40).

[0065] The battery diagnostic device (100) includes a control circuit (130) and a memory (131). The battery diagnostic device (100) may further include at least one of a sensing unit (110) and a communication circuit (150). The 'data acquisition unit' described in the claims of the present invention may refer to at least one of the sensing unit (110) and the communication circuit (150), or may be a configuration that collectively refers to the sensing unit (110) and the communication circuit (150).

[0066] The sensing unit (110) includes a voltage sensor (111) and a current sensor (112).

[0067] The voltage sensor (111) is connected in parallel to the battery (11), measures the battery voltage, which is the voltage across both terminals of the battery (11), and is configured to generate a voltage signal representing the measured battery voltage.

[0068] Of course, the voltage sensor (111) is connected to the positive terminal and negative terminal of each battery cell (BC) included in the battery (11), measures the cell voltage (which may be referred to as 'full cell voltage'), which is the voltage across both ends of each battery cell (BC), and may output an additional voltage signal representing the measured cell voltage (i.e., a measurement of the full cell voltage) to the control circuit (130).

[0069] The current sensor (112) is connected in series to the battery (11) via a current path between the battery (11) and the inverter (30). The current sensor (112) is configured to detect battery current, which is a current flowing through the battery (11), and generate a current signal representing the detected battery current. The current sensor (112) may be implemented using one or a combination of two or more of known current detection elements, such as a shunt resistor, a Hall effect element, etc.

[0070] A temperature sensor (113) is provided to measure the cell temperature, which is the temperature of a battery cell (BC). The temperature sensor (113) can detect the cell temperature of a target cell (BC) periodically or aperiodically while a first electrical stimulus is applied to the battery cell (BC). The temperature sensor (113) can generate a temperature signal representing the detected cell temperature.

[0071] The communication circuit (150) is configured to support wired or wireless communication between the control circuit (130) and the system controller (2). The wired communication may be, for example, CAN (controller area network) communication, and the wireless communication may be, for example, Zigbee or Bluetooth communication. Of course, as long as it supports wired or wireless communication between the control circuit (130) and the system controller (2), the type of communication protocol is not particularly limited. The communication circuit (150) may include an output device (e.g., a display, a speaker) that provides information received from the control circuit (130) and / or the system controller (2) in a form recognizable to the user.

[0072] The control circuit (130) is operably coupled to a relay (20), a voltage sensor (111), a current sensor (112), and a communication circuit (150). The fact that the two components are operably coupled means that the two components are directly or indirectly connected so as to be capable of transmitting and receiving signals in one or both directions.

[0073] The control circuit (130) can collect a voltage signal from a voltage sensor (111) and / or a current signal from a current sensor (112). The control circuit (130) can convert and record each analog signal collected from the sensors (111, 112) into a digital value using an ADC (Analog to Digital Converter) provided therein.

[0074] The control circuit (130) may be referred to as a 'control circuit' or a 'battery controller', and may be implemented in hardware using at least one of application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), microprocessors, and other electrical units for performing functions.

[0075] The memory (131) may include at least one type of storage medium among, for example, a flash memory type, a hard disk type, an SSD (Solid State Disk type), an SDD (Silicon Disk Drive type), a multimedia card micro type, a random access memory (RAM), a static random access memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), and a programmable read-only memory (PROM). The memory (131) may store data and a program required for an operation by the control circuit (130). The memory (131) may store data representing a result of an operation by the control circuit (130). In FIG. 1, the memory (131) is illustrated as being physically independent from the control circuit (130), but may also be built into the control circuit (130).

[0076] The control circuit (130) can turn on the relay (20) in response to a key-on signal. The control circuit (130) can turn off the relay (20) in response to a key-off signal. The key-off signal is a signal that induces a transition from a cycle state to a rest state. Alternatively, the on / off control of the relay (20) may be performed by the system controller (2) instead of the control circuit (130).

[0077] When the inverter (30) or charger (3) operates while the relay (20) is turned on, the battery (11) enters a cycle state. Conversely, when the relay (20) is turned off or the operation of the inverter (30) and charger (3) is stopped, the battery (11) enters a rest state.

[0078] The cycle state refers to a state in which the battery (11) is being charged and discharged, and the rest state refers to a state in which the charging and discharging of the battery (11) has stopped. The fact that the battery (11) is in a cycle state or rest state means that each battery cell (BC) included in the battery (11) is also in a cycle state or rest state.

[0079] The control circuit (130) can determine a voltage detection value and a current detection value based on a voltage signal and a current signal while the battery cell (BC) is in a cycle state and / or an idle state, and then determine (estimate) the state of charge (SOC) of the battery cell (BC) based on the voltage detection value and / or the current detection value.

[0080] If the charger (3) is operating in a constant current charging mode, the current rate (which may be referred to as 'C-rate') of the charging current supplied to the battery cell (BC) is a predetermined constant value, so that in estimating the SOC of the battery cell (BC), the current value of the constant current output from the charger (3) can be used instead of the current detection value obtained using the current sensor (112).

[0081] State of Charge (SOC) is the ratio of the remaining capacity to the full charge capacity (maximum capacity) of a battery cell (BC), and is typically measured in the range of 0 to 1 or 0 to 100%. Known methods such as ampere counting, the Open Circuit Voltage (OCV)-SOC curve, and / or the Kalman filter can be used to determine SOC.

[0082] The communication circuit (150) may obtain a first target full-cell profile (corresponding to the 'target full-cell profile' in the claims) associated with the first electrical stimulus from a separate computing device provided externally (e.g., electric vehicle 1) via wired communication and / or wireless communication. Alternatively, the sensing unit (110) may directly generate the first target full-cell profile of the target cell (BC), which is a battery cell to be diagnosed, based on a measurement signal including a current signal and a voltage signal of the target cell (BC). Alternatively, the control circuit (130) may collect a measurement signal including a current signal and a voltage signal of the target cell (BC) from the sensing unit (110), and then generate the first target full-cell profile of the target cell (BC) based on the collected measurement signal. The measurement signal may further include a temperature signal.

[0083] The control circuit (130) can generate a first target full-cell profile of the target cell (BC) based on the collected measurement signal.

[0084] The first target full-cell profile may represent a correspondence between a capacity factor and a voltage of a target cell (BC) while a first electrical stimulus is applied to the target cell (BC). The capacity factor may be a residual capacity or a state of charge (SOC).

[0085] The first electrical stimulus is an electrical stimulus that causes an instantaneous voltage change in the target cell (BC) when it is applied, which exceeds an allowable level, and is referred to as a “high electrical stimulus.” The second electrical stimulus is an electrical stimulus that causes an instantaneous voltage change in the target cell (BC) when it is applied, which is referred to as a “low electrical stimulus.” For example, the first electrical stimulus may be a charging current having a first current rate (e.g., 1.0 C, 0.33 C, etc.), and the second electrical stimulus may be a charging current having a second current rate (e.g., 0.05 C) that is less than the first current rate. For another example, the first electrical stimulus may be a discharging current having the first current rate, and the second electrical stimulus may be a discharging current having the second current rate.

[0086] The first target full-cell profile may be a profile representing the correspondence between the full-cell voltage and capacity of the target cell (BC) while being charged or discharged at a constant current over a predetermined voltage range (e.g., 3.0 to 4.0 V) or a predetermined SOC range (e.g., 0 to 100% SOC).

[0087] The data acquisition unit can further acquire temperature information of the target cell (BC) during the application period of the first electrical stimulus associated with the first target full-cell profile. The temperature information of the target cell (BC) includes at least one of a start temperature, an end temperature, an average temperature, a maximum temperature, and a minimum temperature. The start temperature represents the temperature at the start of the application period of the first electrical stimulus. The end temperature represents the temperature at the end of the application period of the first electrical stimulus. The average temperature represents the average temperature over the application period of the first electrical stimulus. The maximum temperature represents the maximum temperature during the application period of the first electrical stimulus. The minimum temperature represents the minimum temperature during the application period of the first electrical stimulus.

[0088] Below, before describing the first target full-cell profile obtained using the target cell (BC) of the present invention, the first reference full-cell profile and the second reference full-cell profile will first be described.

[0089] Figure 2 is a drawing referenced to explain the relationship between electrical stimulation and full cell profile.

[0090] The first reference full-cell profile (R1) and the second reference full-cell profile (R2) illustrated in FIG. 2 can be obtained in advance through a pre-experimental procedure of individually applying the first electrical stimulus and the second electrical stimulus to the reference battery cell.

[0091] A reference battery cell is a battery cell that has been previously verified as normal and can have the same positive and negative performance as a new battery cell verified as good. A reference battery cell may also be simply referred to as a "reference cell." A reference cell may be a coin-type cell containing a positive half-cell and a negative half-cell, or a three-electrode cell.

[0092] A "new battery cell" refers to a battery cell in its new condition. "New" condition is synonymous with "Beginning of Life" (BOL). For example, the period from the time of manufacturing completion until the cumulative charge / discharge capacity reaches a predetermined set capacity can be considered BOL, while the period from the time the cumulative charge / discharge capacity reaches the set capacity can be considered MOL (Middle of Life).

[0093] In the graph of Fig. 2, the horizontal axis (X-axis) represents capacity (Ah) and the vertical axis (Y-axis) represents voltage (V).

[0094] The first reference full-cell profile (R1) illustrates the relationship between the capacity and voltage of the reference cell during application of the first electrical stimulus (e.g., charging using the first current rate). The second reference full-cell profile (R2) illustrates the relationship between the capacity and voltage of the reference cell during application of the second electrical stimulus (e.g., charging using the second current rate). The first reference full-cell profile (R1) may be obtained by performing charging using the first current rate while the OCV of the reference cell is set equal to the lower limit of a predetermined voltage range (e.g., 3.0 V). The second reference full-cell profile (R2) may be obtained by performing charging using the second current rate while the OCV of the reference cell is set equal to the lower limit of a predetermined voltage range. Therefore, in FIG. 2, the starting points of the first reference full-cell profile (R1) and the second reference full-cell profile (R2) are approximately identical, whereas the ending points are distinctly different.

[0095] The first reference full-cell profile (R1) and the second reference full-cell profile (R2) can represent the correspondence between the full-cell voltage and capacity of the reference cell over at least a voltage range of interest (e.g., 3.0 to 4.0 V). The lower and upper limits of the voltage range of interest can be represented by the first set voltage (3.0 V in FIG. 2) and the second set voltage (4.0 V in FIG. 2).

[0096] The SOC may be set to 0% when the full cell voltage of any battery cell is equal to the first set voltage, and the SOC may be set to 100% when the full cell voltage is equal to the second set voltage. That is, the first set voltage and the second set voltage may be lower and upper limits of the battery cell voltage corresponding to the SOC (State Of Charge) of 0% to 100% of any battery cell, including the reference cell.

[0097] The starting capacity (Qi) may refer to the remaining capacity when the full-cell voltage of any battery cell is equal to the first set voltage. The ending capacity (Qf) may refer to the remaining capacity when the full-cell voltage of any battery cell is equal to the second set voltage.

[0098] The first reference full-cell profile (R1) may be based on a voltage time series and a current time series (or a capacity time series) obtained by periodically measuring the full-cell voltage and current of the reference cell while the first electrical stimulus is applied.

[0099] The second reference full-cell profile (R2) may be based on a voltage time series and a current time series acquired by periodically measuring the full-cell voltage and current of the reference cell while the second electrical stimulus is applied.

[0100] Here, the first reference full-cell profile (R1) may include an overvoltage in the voltage value corresponding to the same capacity value when compared to the second reference full-cell profile (R2). Therefore, the difference in voltage between the first reference full-cell profile (R1) and the second reference full-cell profile (R2) for the same capacity value may be calculated as an overvoltage.

[0101] Specifically, by removing the second reference full-cell profile (R2) based on the second electrical stimulus from the first reference full-cell profile (R1) based on the first electrical stimulus (calculating the voltage difference by capacity), an overvoltage profile representing the overvoltage by capacity can be generated.

[0102] FIG. 3 is a diagram schematically illustrating an overvoltage profile (OP) obtainable from the first reference full-cell profile (R1) and the second reference full-cell profile (R2) of FIG. 2.

[0103] The overvoltage profile (OP) may be a profile representing a correspondence between capacity and overvoltage. The overvoltage profile (OP) may be a profile representing a voltage difference by capacity between a first reference full-cell profile (R1) and a second reference full-cell profile (R2).

[0104] The capacity range (Qi~Qf) of the overvoltage profile (OP) may be a common capacity range between the first reference full-cell profile (R1) and the second reference full-cell profile (R2). In Fig. 2, the capacity range of the first reference full-cell profile (R1) is 5~47 Ah, and the capacity range of the second reference full-cell profile (R2) is 5~50 Ah, so Qi may be 5 Ah and Qf may be 47 Ah.

[0105] Figure 4 is a graph referenced to explain the relationship between the first target full-cell profile (M), the modified target full-cell profile (E), and the second target full-cell profile (N).

[0106] In FIGS. 2 to 4, Ah is used as the unit of the horizontal axis, but this unit may be expressed in other forms. For example, instead of Ah, a percentage % indicating SOC (State Of Charge) may be used as the unit of the horizontal axis.

[0107] Referring to FIG. 4, the control circuit (130) can generate a first target full-cell profile (M) representing a correspondence between the full-cell voltage and the capacity of the target cell (BC) while a first electrical stimulus is applied to the target cell (BC). The first target full-cell profile (M) can represent a correspondence between the full-cell voltage and the capacity of the target cell (BC) at least over a voltage range of interest.

[0108] Therefore, since the reference cell and the target cell (BC) have different charge / discharge performances, there is bound to be some difference between the first target full cell profile (M) and the first reference full cell profile (R1).

[0109] For example, in the same voltage range of interest (e.g., 3.0 to 4.0 V), the capacity range of the first reference full-cell profile (R1) illustrated in FIG. 2 is 5 to 47 Ah, whereas the capacity range of the first target full-cell profile (M) is 5 to 45 Ah.

[0110] The control circuit (130) can generate a modified target full-cell profile (E) by modifying the first target full-cell profile (M) to be associated with the second electrical stimulation based on the overvoltage profile (OP). Specifically, the control circuit (130) can generate the modified target full-cell profile (E) by subtracting the overvoltage profile (OP) from the first target full-cell profile (M). As a result, at the same capacity value, the voltage value of the modified target full-cell profile (E) can be lower than the voltage value of the first target full-cell profile (M).

[0111] The control circuit (130) can obtain a modified target full-cell profile (E) by subtracting the capacity-specific overvoltage of the overvoltage profile (OP) from the capacity-specific voltage of the first target full-cell profile (M) in the common capacity range of the first target full-cell profile (M) and the overvoltage profile (OP). In this case, a capacity range of 45 to 47 Ah among the entire capacity range of the overvoltage profile (OP) may not be utilized. That is, the modified target full-cell profile (E) can be obtained by removing the capacity-specific overvoltage of the overvoltage profile (OP) corresponding to the capacity-specific voltage of the first target full-cell profile (M).

[0112] Alternatively, the control circuit (130) can generate an adjusted overvoltage profile (not shown in the drawing) by reducing the overvoltage profile (OP) along the horizontal axis so that the capacity range of the overvoltage profile (OP) matches the capacity range of the first target full-cell profile (M). Then, the control circuit (130) can generate a modified target full-cell profile (E) by subtracting the overvoltage value of the adjusted overvoltage profile from the voltage value of the first target full-cell profile (M). That is, the modified target full-cell profile (E) can be obtained by removing the capacity-specific overvoltage of the adjusted overvoltage profile from the capacity-specific voltage of the first target full-cell profile (M).

[0113] The second target full-cell profile (N) is an example of a profile representing the correspondence between the capacity factor and voltage of the target cell (BC), which would be expected to be acquired instead of the first target full-cell profile (M), if the second electrical stimulus had been applied to the target cell (BC) instead of the first electrical stimulus.

[0114] The modified target full-cell profile (E) is an estimation result of the second target full-cell profile (N) based on the first target full-cell profile (M) and the overvoltage profile (OP).

[0115] Referring to FIG. 4, since the modified target full-cell profile (E) is a profile obtained by subtracting the overvoltage profile (OP) from the first target full-cell profile (M), the modified target full-cell profile (E) is more similar to the second target full-cell profile (N) than the first target full-cell profile (M). Therefore, in diagnosing the charge / discharge performance of the target cell (BC), it is advantageous in terms of diagnostic accuracy to utilize the modified target full-cell profile (E) instead of the first target full-cell profile (M).

[0116] Meanwhile, since the modified target full-cell profile (E) does not completely match the second target full-cell profile (N), there may still be a significant difference between the diagnostic results of charge / discharge performance based on the modified target full-cell profile (E) and the actual charge / discharge performance. A method for reducing the error in the diagnostic results of charge / discharge performance will be described later with reference to FIG. 12.

[0117] The control circuit (130) can determine first diagnostic result information on the charge / discharge performance of the target cell (BC) by applying diagnostic logic to the modified target full-cell profile (E). The first diagnostic result information can be said to be a preliminary diagnostic result on the charge / discharge performance of the target cell (BC). The first diagnostic result information includes a first diagnostic factor group, and the first diagnostic factor group can include at least one preliminary diagnostic factor.

[0118] The first diagnostic factor group may include at least one of a positive engagement start point, a positive engagement end point, a positive scale factor, a negative engagement start point, a negative engagement end point, a negative scale factor, a positive loading amount, a negative loading amount, and a NP ratio as a preliminary diagnostic factor.

[0119] In the present specification, the positive electrode participation start point on the positive electrode profile of any battery cell represents the positive electrode voltage and positive electrode capacity (or positive electrode SOC) when the full cell voltage of the battery cell matches a first set voltage. The positive electrode voltage of the positive electrode participation start point may be referred to as a “positive electrode starting potential.” In addition, the negative electrode participation start point on the negative electrode profile of the battery cell represents the negative electrode voltage and negative electrode capacity (or negative electrode SOC) when the full cell voltage of the battery cell matches a first set voltage. The negative electrode voltage of the negative electrode participation start point may be referred to as a “negative electrode starting potential.” Therefore, the voltage difference between the positive electrode participation start point and the negative electrode participation start point may be equal to the first set voltage.

[0120] In addition, the positive participation end point on the positive profile of any battery cell represents the positive voltage and positive capacity when the full-cell voltage of the battery cell matches the second set voltage. The positive voltage of the positive participation end point may be referred to as the 'positive termination potential'. In addition, the negative participation end point on the negative profile of the battery cell represents the negative voltage and negative capacity when the full-cell voltage of the battery cell matches the second set voltage. The negative voltage of the negative participation end point may be referred to as the 'negative termination potential'. Therefore, the voltage difference between the positive participation end point and the negative participation end point may be equal to the second set voltage.

[0121] In the present specification, the anode capacity (capacity value) of a specific point on the anode profile of any battery cell may mean the capacity difference between either one of the two end points of the anode profile and the specific point. The anode SOC of a specific point on the anode profile of any battery cell may mean the ratio of the capacity difference between either one of the two end points of the anode profile (e.g., a low-capacity point) and the specific point to the capacity difference between the two end points of the anode profile.

[0122] Similarly, the negative capacity (capacity value) of a specific point on the negative profile of any battery cell may mean the capacity difference between either one of the two end points of the negative profile (or the positive profile) and the specific point. The negative SOC of a specific point on the negative profile of any battery cell may mean the ratio of the capacity difference between either one of the two end points of the negative profile (e.g., a low-capacity point) and the specific point to the capacity difference between either one of the two end points of the negative profile (or the positive profile).

[0123] The positive scale factor of any battery cell may represent the ratio of the capacity difference between the positive engagement start point and the positive engagement end point of the battery cell to the reference positive capacity of the reference cell. The negative scale factor of any battery cell may represent the ratio of the capacity difference between the negative engagement start point and the negative engagement end point of the battery cell to the reference negative capacity of the reference cell.

[0124] In the memory (131), information indicating the voltage and capacity of each of the reference positive electrode participation start point, reference positive electrode participation end point, reference negative electrode participation start point, and reference negative electrode participation end point, which represent the charge / discharge performance of the reference cell, may be recorded in advance.

[0125] From now on, with reference to FIGS. 5 to 11, the diagnostic processes included in the diagnostic logic will be described.

[0126] Fig. 5 is a graph for reference in explaining an example of each of a modified target full-cell profile (E), a second reference full-cell profile (R2), a reference positive electrode profile (Rp), and a reference negative electrode profile (Rn). In the graph of Fig. 5, the horizontal axis (X-axis) represents capacity, and the vertical axis (Y-axis) represents voltage. The modified target full-cell profile (E) and the second reference full-cell profile (R2) are the same as in Fig. 2.

[0127] Referring to FIG. 5, the reference anode profile (Rp) may be a profile representing the correspondence between the anode voltage and capacity while a second electrical stimulus is being applied to the reference cell. The anode voltage of the reference cell refers to the potential difference between the potential of the reference electrode (not shown) and the potential of the anode of the reference cell.

[0128] The reference cathode profile (Rn) may be a profile representing the correspondence between the cathode voltage and the capacity when a second electrical stimulus is applied to the reference cell. The cathode voltage of the reference cell refers to the potential difference between the potential of the reference electrode and the potential of the cathode of the reference cell.

[0129] The potential of the reference electrode may be, for example, the redox potential of lithium. The anode voltage may be referred to simply as the anode potential, and the cathode voltage may be referred to simply as the cathode potential.

[0130] The reference positive profile (Rp) and the reference negative profile (Rn) may be pre-stored in the memory (131).

[0131] At least one of the reference positive electrode profile (Rp) and the reference negative electrode profile (Rn) can be aligned along the horizontal axis so that the result of synthesis of a portion of the common capacity range (5 to 50 Ah in FIG. 5) of the reference positive electrode profile (Rp) and the reference negative electrode profile (Rn) matches the second reference full cell profile (R2).

[0132] In Fig. 5, a case is illustrated in which the reference cathode profile (Rn) is aligned by shifting to the right, with the starting point (point corresponding to capacity 0) of the reference anode profile (Rp) as the reference point.

[0133] It can be seen from Fig. 5 that the ends of the reference positive electrode profile (Rp) and the reference negative electrode profile (Rn) are misaligned. That is, the capacity range of the reference positive electrode profile (Rp) and the capacity range of the reference negative electrode profile (Rn) do not match and may only partially overlap. Therefore, the second reference full-cell profile (R2) may represent the full-cell voltage of the reference cell in a portion of the capacity range common to the reference positive electrode profile (Rp) and the reference negative electrode profile (Rn).

[0134] The control circuit (130) may be configured to compare the modified target full-cell profile (E) with at least one comparison full-cell profile. The comparison full-cell profile may be a result of generating an adjusted anode profile and an adjusted cathode profile by adjusting each of the reference anode profile (Rp) and the reference cathode profile (Rn) stored in the memory (131), and then synthesizing (combining) the adjusted anode profile and the adjusted cathode profile.

[0135] That is, when the second reference full-cell profile (R2) is a result of subtracting a portion of the reference cathode profile (Rn) from a portion of the reference anode profile (Rp), the comparison full-cell profile can be said to be a result of subtracting a portion of the adjusted cathode profile from a portion of the adjusted anode profile.

[0136] The control circuit (130) can generate at least one comparison full-cell profile by directly adjusting the reference positive electrode profile (Rp) and the reference negative electrode profile (Rn). Alternatively, the at least one comparison full-cell profile can be pre-secured based on the reference positive electrode profile (Rp) and the reference negative electrode profile (Rn) and stored in the memory (131). In this case, the control circuit (130) can obtain the comparison full-cell profile by accessing the memory (131) and reading it.

[0137] The control circuit (130) can generate multiple comparison full-cell profiles from the reference anode profile (Rp) and the reference cathode profile (Rn) by repeating the adjustment procedure of adjusting and then synthesizing each of the reference anode profile (Rp) and the reference cathode profile (Rn) to several levels. The comparison full-cell profile may also be referred to as an 'adjusted reference full-cell profile'.

[0138] The control circuit (130) can specify a comparison full-cell profile among a plurality of comparison full-cell profiles that has a minimum error with respect to the modified target full-cell profile (E). Then, the control circuit (130) can determine that the adjusted positive profile and the adjusted negative profile mapped to the specified comparison full-cell profile are the positive profile and the negative profile of the target cell (BC).

[0139] In this regard, various methods known at the time of filing of the present invention can be employed to determine the error between two profiles as a set of data points each expressible in a two-dimensional coordinate system. For example, the absolute integral of the area between the two profiles or the Root Mean Square Error (RMSE) can be used as the error between the two profiles.

[0140] According to this configuration of the present invention, various state information about the target cell (BC) can be acquired based on the finally determined adjusted anode profile and adjusted cathode profile. The finally determined adjusted anode profile and adjusted cathode profile may be mapped to one of the comparison full-cell profiles that has the minimum error with respect to the modified target full-cell profile (E) among the plurality of comparison full-cell profiles. In particular, the comparison full-cell profile based on the finally determined adjusted anode profile and adjusted cathode profile can be said to be almost identical to the modified target full-cell profile (E) in terms of shape, etc.

[0141] FIGS. 6 to 8 are drawings for reference in explaining an example of a procedure for generating a comparison full-cell profile.

[0142] The procedure for generating a comparative full-cell profile to be described with reference to FIGS. 6 to 8 may be performed in the following order: a first routine (see FIG. 6) for setting four points (positive engagement start point, positive engagement end point, negative engagement start point, negative engagement end point) to correspond to a voltage range of interest, a second routine (see FIG. 7) for performing profile shifting, and a third routine (see FIG. 8) for performing capacity scaling. That is, the procedure for generating a comparative full-cell profile according to one embodiment of the present invention may include the first to third routines.

[0143] The reference anode profile (Rp) and reference cathode profile (Rn) illustrated in Fig. 6 are identical to those illustrated in Fig. 5.

[0144] The control circuit (130) can determine the positive engagement start point (pi), the positive engagement end point (pf), the negative engagement start point (ni), and the negative engagement end point (nf) on the reference positive profile (Rp) and the reference negative profile (Rn).

[0145] Either the positive engagement initiation point (pi) or the negative engagement initiation point (ni) depends on the other.

[0146] For example, the control circuit (130) may divide the positive voltage range from the starting point of the reference positive profile (Rp) to the ending point (or the second set voltage) into a plurality of micro-voltage sections, and then set the boundary points of two adjacent micro-voltage sections among the plurality of micro-voltage sections as positive engagement start points (pi). Each micro-voltage section may have a predetermined size (e.g., 0.01 V). Then, the control circuit (130) may set a point on the reference negative profile (Rn) that is smaller than the positive engagement start point (pi) by the first set voltage (e.g., 3 V) as the negative engagement start point (ni).

[0147] As another example, the control circuit (130) may divide the negative voltage range from the start point to the end point of the reference negative profile (Rn) into a plurality of micro-voltage sections of a predetermined size, and then set the boundary point of two adjacent micro-voltage sections among the plurality of micro-voltage sections as a negative engagement start point (ni). Then, the control circuit (130) may search for a point that is greater than the negative engagement start point (ni) by a first set voltage (e.g., 3 V) from the reference positive profile (Rp), and set the searched point as the positive engagement start point (pi).

[0148] Either the positive engagement end point (pf) or the negative engagement end point (nf) depends on the other.

[0149] For example, the control circuit (130) may divide the voltage range from the second set voltage to the end point of the reference positive electrode profile (Rp) into a plurality of micro-voltage sections of a predetermined size, and then set the boundary point of two adjacent micro-voltage sections among the plurality of micro-voltage sections as the positive electrode participation end point (pf). Then, the control circuit (130) may set the point on the reference negative electrode profile (Rn) that is smaller by the second set voltage (e.g., 4 V) than the positive electrode participation end point (pf) as the negative electrode participation end point (nf).

[0150] As another example, the control circuit (130) may divide the negative voltage range from the start point to the end point of the reference negative profile (Rn) into a plurality of micro-voltage sections of a predetermined size, and then set the boundary point of two adjacent micro-voltage sections among the plurality of micro-voltage sections as a negative participation end point (nf). Then, the control circuit (130) may search for a point from the reference positive profile (Rp) that is greater than the negative participation end point (nf) by a second set voltage (e.g., 4 V), and set the searched point as the positive participation end point (pf).

[0151] Once the determination of the positive engagement start point (pi), the positive engagement end point (pf), the negative engagement start point (ni), and the negative engagement end point (nf) is completed, the control circuit (130) shifts at least one of the reference positive profile (Rp) and the reference negative profile (Rn) to the left or right along the horizontal axis.

[0152] Referring to FIG. 6, the control circuit (130) can shift the reference anode profile (Rp) to the left (low capacity side), shift the reference cathode profile (Rn) to the right (high capacity side), or perform both, so that the capacitance values ​​of the anode participation start point (pi) and the cathode participation start point (ni) match.

[0153] Alternatively, the control circuit (130) may shift the reference anode profile (Rp) to the left, shift the reference cathode profile (Rn) to the right, or both, so that the capacitance values ​​of the anode engagement end point (pf) and the cathode engagement end point (nf) match.

[0154] Fig. 7 illustrates a situation where only the reference anode profile (Rp) is shifted to the left to generate an adjusted reference anode profile (Rp'), and as a result, the capacity value of the anode engagement start point (pi') matches the capacity value of the cathode engagement start point (ni). The adjusted reference anode profile (Rp') may be a result of applying an adjustment procedure to the reference anode profile (Rp) to shift it to the left by the difference in capacity between the anode engagement start point (pi) and the cathode engagement start point (ni). Therefore, the two points (pi, pi') may differ only in capacity value and have the same voltage. In addition, the two points (pf, pf') may differ only in capacity value and have the same voltage.

[0155] When the adjustment result profiles (Rp', Rn) in which at least one of the reference positive profile (Rp) and the reference negative profile (Rn) is shifted are secured, the control circuit (130) can scale the capacity range of at least one of the adjustment result profiles (Rp', Rn).

[0156] According to the example illustrated in FIG. 7, the control circuit (130) can perform an additional adjustment procedure to contract or expand at least one of the adjusted reference anode profile (Rp') and the reference cathode profile (Rn) along the horizontal axis.

[0157] Referring to FIG. 8, the control circuit (130) can generate an adjusted reference anode profile (Rp') by shrinking or expanding the adjusted reference anode profile (Rp') so that the size of the capacity range between two points (pi', pf') of the adjusted reference anode profile (Rp') matches the size of the capacity range of the modified target full-cell profile (E). At this time, one of the two points (pi', pf') can be fixed. Accordingly, the capacity difference between the two points (pi', pf'') of the adjusted reference anode profile (Rp'') can match the capacity range of the modified target full-cell profile (E).

[0158] In addition, the control circuit (130) can generate an adjusted reference cathode profile (Rn') by shrinking or expanding the reference cathode profile (Rn) so that the size of the capacity range between two points (ni, nf) of the reference cathode profile (Rn) matches the size of the capacity range of the modified target full-cell profile (E). At this time, one of the two points (ni, nf) can be fixed. Accordingly, the capacity difference between the two points (ni, nf') of the adjusted reference cathode profile (Rn') can match the capacity range of the modified target full-cell profile (E).

[0159] In Fig. 8, the adjusted reference anode profile (Rp'') is a result of shrinking the adjusted reference anode profile (Rp') shown in Fig. 7, and the adjusted reference cathode profile (Rn') is a result of expanding the reference cathode profile (Rn) shown in Fig. 7.

[0160] The positive participation end point (pf'') on the adjusted reference positive profile (Rp'') corresponds to the positive participation end point (pf) on the adjusted reference positive profile (Rp'). The negative participation end point (nf') on the adjusted reference negative profile (Rn') corresponds to the negative participation end point (nf) on the reference negative profile (Rn).

[0161] The capacity difference between the positive engagement start point (pi') and the positive engagement end point (pf'') of the adjusted reference positive electrode profile (Rp'') matches the size of the capacity range of the modified target full-cell profile (E). Similarly, the capacity difference between the negative engagement start point (ni) and the negative engagement end point (nf') of the adjusted reference negative electrode profile (Rn') matches the size of the capacity range of the modified target full-cell profile (E).

[0162] In addition, the capacity range by the two points (pi', pf'') of the adjusted reference positive electrode profile (Rp'') matches the capacity range by the two points (ni, nf') of the adjusted reference negative electrode profile (Rn'). The control circuit (130) can generate a comparison full-cell profile (S) by subtracting the portion between the two points (pi, pf') of the adjusted reference positive electrode profile (Rp'') from the portion between the two points (ni, nf') of the adjusted reference negative electrode profile (Rn').

[0163] The control circuit (130) can calculate the error (profile error) between the comparison full-cell profile (S) and the modified target full-cell profile (E).

[0164] The control circuit (130) can record in memory (131) at least two of the adjusted reference positive profile (Rp''), the adjusted reference negative profile (Rn'), the positive engagement start point (pi'), the positive engagement end point (pf''), the negative engagement start point (ni), the negative engagement end point (nf'), the positive scale factor, the negative scale factor, the comparative full-cell profile (S), and the profile error by mutually mapping them.

[0165] The anode scale factor of the adjusted reference anode profile (Rp'') can represent the ratio of the capacity difference between two points (pi', pf'') to the capacity difference between two points (pi0, pf0). Alternatively, the anode scale factor of the adjusted reference anode profile (Rp'') can represent the ratio of the anode capacity difference between two points (pi', pf'') to the anode capacity difference between two points (pi0, pf0). Alternatively, the anode scale factor of the adjusted reference anode profile (Rp'') can represent the ratio of the anode SOC difference between two points (pi', pf'') to the anode SOC difference between two points (pi0, pf0).

[0166] The cathode scale factor of the adjusted reference cathode profile (Rn') can represent the ratio of the capacity difference between two points (ni, nf') to the capacity difference between two points (ni0, nf0). Alternatively, the cathode scale factor of the adjusted reference cathode profile (Rn') can represent the ratio of the cathode capacity difference between two points (ni, nf') to the cathode capacity difference between two points (ni0, nf0). Alternatively, the cathode scale factor of the adjusted reference cathode profile (Rn') can represent the ratio of the cathode SOC difference between two points (ni, nf') to the cathode SOC difference between two points (ni0, nf0).

[0167] Hereinafter, ps may be used as a symbol indicating a positive scale factor, and ns may be used as a symbol indicating a negative scale factor.

[0168] Meanwhile, as described above, when the anode voltage range of the reference anode profile (Rp) is divided into a plurality of micro-voltage sections, the boundary points of two adjacent micro-voltage sections among the plurality of micro-voltage sections can be set as the anode participation start point (pi).

[0169] For example, if the anode voltage range of the reference anode profile (Rp) is divided into 100 microvoltage ranges, there may be 100 boundary points that can be set as anode participation start points (pi). Furthermore, if the voltage range that is higher than the second set voltage in the reference anode profile (Rp) is divided into 40 microvoltage ranges, there may be 40 boundary points that can be set as anode participation end points (pf). In this case, up to 4,000 different comparison full-cell profiles can be generated.

[0170] Of course, those skilled in the art will easily understand that as the size of the micro-voltage section decreases, the number of comparative full-cell profiles that can be generated at maximum increases, and conversely, as the size of the micro-voltage section increases, the number of comparative full-cell profiles that can be generated at maximum decreases.

[0171] The control circuit (130) can identify a minimum value among the profile errors of a plurality of comparative full-cell profiles generated as described above, and then obtain a first diagnostic factor group as information mapped to the minimum profile error (e.g., at least one of a positive engagement start point, a positive engagement end point, a negative engagement start point, a negative engagement end point, a positive scale factor, and a negative scale factor) from the memory (131).

[0172] FIGS. 9 to 11 are diagrams that are referenced to explain other examples of a procedure for generating a comparison full-cell profile according to diagnostic logic. Note that the embodiments according to FIGS. 9 to 11 are independent of the embodiments according to FIGS. 6 to 8 . Therefore, terms or symbols commonly used in describing the embodiments according to FIGS. 6 to 8 and the embodiments according to FIGS. 9 to 11 should be understood as being limited to each embodiment.

[0173] The generation procedure of the comparative full-cell profile (U) to be described with reference to FIGS. 9 to 11 may be performed in the following order: a fourth routine (see FIG. 9) for performing capacity scaling, a fifth routine (see FIG. 10) for setting four points (positive participation start point, positive participation end point, negative participation start point, negative participation end point), and a sixth routine (see FIG. 11) for performing profile shifting. That is, the generation procedure of the comparative full-cell profile according to another embodiment of the present invention may include the fourth to sixth routines.

[0174] Referring to FIG. 9, the control circuit (130) can apply a positive scale factor and a negative scale factor selected from a scaling value range to a reference positive profile (Rp) and a reference negative profile (Rn), respectively, to generate an adjusted reference positive profile (Rp') and an adjusted reference negative profile (Rn').

[0175] The scaling value range may be predetermined or may vary depending on the ratio of the size of the capacity range of the modified target full-cell profile (E) to the size of the capacity range of the second reference full-cell profile (R2). For example, when values ​​spaced by 0.1% of the scaling value range (e.g., 90-99%) (i.e., 90%, 90.1%, 90.2%, ... 98.9%, 99%) can be selected as the positive scale factor and the negative scale factor, 91 values ​​can be selected as the positive scale factor and the negative scale factor, respectively. In this case, a maximum of 8,281 adjusted profile pairs (Rp', Rn') can be generated according to 91 × 91 = 8,281 adjustment levels (combinations of positive scale factors and negative scale factors). A tuned profile pair means a combination of a tuned positive profile (Rp') and a tuned negative profile (Rn').

[0176] The adjusted reference anode profile (Rp') and the adjusted reference cathode profile (Rn') illustrated in FIG. 9 illustrate the results of applying a cathode scale factor and anode scale factor of less than 100% to the reference anode profile (Rp) and the reference cathode profile (Rn), respectively.

[0177] Since the positive scale factor and the negative scale factor are less than 100%, the adjusted reference positive profile (Rp') is the reference positive profile (Rp) shrunk along the horizontal axis, and the adjusted reference negative profile (Rn') is also the reference negative profile (Rn) shrunk along the horizontal axis. To facilitate understanding, the starting points of the reference positive profile (Rp) and the reference negative profile (Rn) are fixed, and only the remaining portion is shrunk to the left along the horizontal axis, as an example.

[0178] Referring to FIG. 10, the control circuit (130) can determine the positive engagement start point (pi'), the positive engagement end point (pf'), the negative engagement start point (ni'), and the negative engagement end point (nf') on the adjusted reference positive profile (Rp') and the adjusted reference negative profile (Rn').

[0179] Either the positive engagement start point (pi') or the negative engagement start point (ni') may depend on the other. Furthermore, either the positive engagement end point (pf') or the negative engagement end point (nf') may depend on the other. Furthermore, either the positive engagement start point (pi') or the positive engagement end point (pf') may be set based on the other.

[0180] That is, when any one of the positive engagement start point (pi'), positive engagement end point (pf'), negative engagement start point (ni') and negative engagement end point (nf') is set, the remaining three points can be automatically set by the size of the capacity range of the first set voltage, the second set voltage and / or the modified target full-cell profile (E) (e.g., 45 Ah - 5 Ah = 40 Ah in FIG. 4).

[0181] For example, the control circuit (130) may divide the positive voltage range from the starting point to the ending point (or the second set voltage) of the adjusted reference positive voltage profile (Rp') into a plurality of micro-voltage sections, and then set the boundary points of two adjacent micro-voltage sections among the plurality of micro-voltage sections as positive engagement start points (pi'). Then, the control circuit (130) may set a point on the adjusted reference negative voltage profile (Rn') that is smaller by the first set voltage than the positive engagement start point (pi') as the negative engagement start point (ni').

[0182] As another example, the control circuit (130) may divide the negative voltage range from the start point to the end point of the adjusted reference negative profile (Rn') into a plurality of micro-voltage sections of a predetermined size, and then set the boundary point of two adjacent micro-voltage sections among the plurality of micro-voltage sections as a negative engagement start point (ni'). Then, the control circuit (130) may search for a point that is larger than the negative engagement start point (ni') by a first set voltage from the adjusted reference positive profile (Rp'), and set the searched point as the positive engagement start point (pi').

[0183] As another example, the control circuit (130) may divide the voltage range from the second set voltage to the end point of the adjusted reference positive electrode profile (Rp') into a plurality of micro-voltage sections of a predetermined size, and then set the boundary point of two adjacent micro-voltage sections among the plurality of micro-voltage sections as the positive electrode participation end point (pf'). Then, the control circuit (130) may search for a point in the adjusted reference negative electrode profile (Rn') that is smaller than the positive electrode participation end point (pf') by the second set voltage (e.g., 4 V), and set the searched point as the negative electrode participation end point (nf').

[0184] As another example, the control circuit (130) may divide the negative voltage range from the start point to the end point of the adjusted second reference negative profile (Rn') into a plurality of micro-voltage sections of a predetermined size, and then set the boundary point of two adjacent micro-voltage sections among the plurality of micro-voltage sections as a negative participation end point (nf'). Then, the control circuit (130) may search for a point that is larger than the negative participation end point (nf') by a second set voltage from the adjusted reference positive profile (Rp'), and set the searched point as the positive participation end point (pf').

[0185] The control circuit (130) can additionally determine the remaining three points based on the determined point when one of the positive participation start point (pi'), the positive participation end point (pf'), the negative participation start point (ni'), and the negative participation end point (nf') is determined.

[0186] For example, when the positive participation start point (pi') is first determined, the control circuit (130) can set a point on the adjusted reference positive profile (Rp') that has a capacity value that is greater than the capacity value of the positive participation start point (pi') by the size of the capacity range of the modified target full-cell profile (E) as the positive participation end point (pf'). In addition, the control circuit (130) can search for a point that is lower than the positive participation start point (pi') by a first set voltage from the adjusted reference negative profile (Rn') and set the searched point as the negative participation start point (ni'). In addition, the control circuit (130) can set a point on the adjusted reference negative profile (Rn') that has a capacity value that is greater than the capacity value of the negative participation start point (ni') by the size of the capacity range of the modified target full-cell profile (E) as the negative participation end point (nf').

[0187] As another example, the control circuit (130) may set a point on the adjusted reference positive electrode profile (Rp') that has a capacity value that is smaller by the size of the capacity range of the modified target full-cell profile (E) than the capacity value of the positive electrode participation end point (pf') when the positive electrode participation end point (pf') is first determined, as the positive electrode participation start point (pi'). In addition, the control circuit (130) may search for a point from the adjusted reference negative electrode profile (Rn') that is lower by a second set voltage than the positive electrode participation end point (pf'), and set the searched point as the negative electrode participation end point (nf'). In addition, the control circuit (130) may set a point on the adjusted reference negative electrode profile (Rn') that has a capacity value that is smaller by the size of the capacity range of the modified target full-cell profile (E) than the capacity value of the negative electrode participation end point (nf') as the negative electrode participation start point (ni').

[0188] As another example, when the negative participation start point (ni') is determined, the control circuit (130) can set a point on the adjusted reference negative profile (Rn') that has a capacity value that is greater than the capacity value of the negative participation start point (ni') by the size of the capacity range of the modified target full-cell profile (E) as the negative participation end point (nf'). In addition, the control circuit (130) can search for a point from the adjusted reference positive profile (Rp') that is higher than the negative participation start point (ni') by a first set voltage, and set the searched point as the positive participation start point (pi'). In addition, the control circuit (130) can set a point on the adjusted reference positive profile (Rp') that has a capacity value that is greater than the capacity value of the positive participation start point (pi') by the size of the capacity range of the modified target full-cell profile (E) as the positive participation end point (pf').

[0189] As another example, when the negative participation end point (nf') is determined, the control circuit (130) can set a point on the adjusted reference negative profile (Rn') that has a capacity value that is smaller by the size of the capacity range of the modified target full-cell profile (E) than the capacity value of the negative participation end point (nf') as the negative participation start point (ni'). In addition, the control circuit (130) can search for a point from the adjusted reference positive profile (Rp') that is higher by a second set voltage than the negative participation end point (nf') and set the searched point as the positive participation end point (pf'). In addition, the control circuit (130) can set a point on the adjusted reference positive profile (Rp') that has a capacity value that is smaller by the size of the capacity range of the modified target full-cell profile (E) than the capacity value of the positive participation end point (pf') as the positive participation start point (pi').

[0190] When the determination of the positive engagement start point (pi'), the positive engagement end point (pf'), the negative engagement start point (ni') and the negative engagement end point (nf') is completed based on the pair of positive scale factors and negative scale factors, the control circuit (130) can shift at least one of the adjusted reference positive profile (Rp') and the adjusted reference negative profile (Rn') to the left or right along the horizontal axis so that the capacitance values ​​of the positive engagement start point (pi') and the negative engagement start point (ni') match or so that the capacitance values ​​of the positive engagement end point (pf') and the negative engagement end point (nf') match.

[0191] The adjusted reference cathode profile (Rn'') illustrated in Fig. 11 is only the adjusted reference cathode profile (Rn') illustrated in Fig. 10 shifted to the right. Accordingly, the capacity values ​​of the positive participation start point (pi') and the negative participation start point (ni'') are matched with each other on the horizontal axis. In this regard, since the capacity difference between the positive participation start point (pi') and the positive participation end point (pf') is the same as the capacity difference between the negative participation start point (ni') and the negative participation end point (nf'), when the capacity values ​​of the positive participation start point (pi') and the negative participation start point (ni'') are matched with each other on the horizontal axis, the capacity values ​​of the positive participation end point (pf') and the negative participation end point (nf') are also matched with each other on the horizontal axis.

[0192] Referring to FIG. 11, the control circuit (130) can generate a comparison full-cell profile (U) by subtracting a partial profile between two points (pi', pf') of an adjusted reference positive profile (Rp') from a partial profile between two points (ni'', nf'') of an adjusted reference negative profile (Rn'').

[0193] The control circuit (130) can calculate the error (profile error) between the comparison full-cell profile (U) and the modified target full-cell profile (E).

[0194] The control circuit (130) can record in memory (140) at least two of the adjusted reference positive profile (Rp'), the adjusted reference negative profile (Rn''), the positive engagement start point (pi'), the positive engagement end point (pf'), the negative engagement start point (ni''), the negative engagement end point (nf''), the positive scale factor, the negative scale factor, the comparative full-cell profile (U), and the profile error by mutually mapping them.

[0195] As described above, the control circuit (130) can generate a corresponding comparison full-cell profile (U) for each pair of positive scale factors and negative scale factors selected from the scaling value range. Since there are multiple pairs of positive scale factors and negative scale factors, it is obvious that a plurality of comparison full-cell profiles (U) will also be generated.

[0196] The control circuit (130) can identify the minimum value among the profile errors of a plurality of comparison full cell profiles, and then obtain information mapped to the minimum profile error from the memory (131).

[0197] As described above, the control circuit (130) can execute diagnostic logic to generate a comparison full-cell profile having a minimum error with respect to the modified target full-cell profile (E) based on the reference positive profile (Rp) and the reference negative profile (Rn).

[0198] The control circuit (130) can determine a first diagnostic factor group including at least one of a positive engagement start point, a positive engagement end point, a positive scale factor, a negative engagement start point, a negative engagement end point, and a negative scale factor, each mapped to a minimum profile error, as a preliminary diagnostic factor.

[0199] Meanwhile, the first diagnostic factor group can be said to more accurately represent the actual charge / discharge performance of the target cell (BC) than the result of applying the diagnostic logic to the first target full-cell profile (M), since it is the result of applying the diagnostic logic to the modified target full-cell profile (E).

[0200] However, since the overvoltage profile (OP) is related to the reference cell and not the target cell (BC), there may still be a considerable difference between the charge / discharge performance indicated by the first diagnostic result information and the actual charge / discharge performance of the target cell (BC).

[0201] Therefore, in order to narrow the gap between the charge / discharge performance indicated by the first diagnosis result information and the actual charge / discharge performance, it is desirable to perform a procedure for correcting at least one preliminary diagnosis factor of the first diagnosis result information, which can be achieved by a factor correction model described below. The correction procedure for the first diagnosis result information is performed to determine the second diagnosis result information as a precise diagnosis result for the charge / discharge performance of the target cell (BC).

[0202] FIG. 12 is a drawing for reference in explaining the function of the factor correction model, FIGS. 13 to 21 are drawings for reference in explaining a training data set provided for learning the factor correction model, FIG. 22 is a drawing exemplarily showing a neural network structure of the factor correction model of FIG. 12, and FIG. 23 is a drawing showing an example of a correlation coefficient between diagnostic factors that can be obtained through learning the factor correction model.

[0203] Referring to FIG. 12, the control circuit (130) can determine second diagnostic result information as a precise diagnostic result for the charge / discharge performance of the target cell (BC) by applying a factor correction model (200) to at least one preliminary diagnostic factor of the first diagnostic factor group (1210), which is a preliminary diagnostic result for the charge / discharge performance of the target cell (BC). The second diagnostic result information includes a second diagnostic factor group (1220), and the second diagnostic factor group (1220) can include at least one precise diagnostic factor.

[0204] The first diagnostic factor group (1210) may include at least one type of diagnostic factor among the positive participation start point, negative participation start point, positive participation end point, negative participation end point, positive scale factor, negative scale factor, positive loading amount, negative loading amount, and NP ratio of the target cell (BC), as a preliminary diagnostic factor, determined based on the modified target full-cell profile (E).

[0205] The second diagnostic factor group (1220) may include at least one type of diagnostic factor among the positive participation start point, negative participation start point, positive participation end point, negative participation end point, positive scale factor, negative scale factor, positive loading amount, negative loading amount, and NP ratio of the target cell (BC) as a precision diagnostic factor.

[0206] At least one precision diagnostic factor of the second diagnostic factor group (1220) may be a result of at least one preliminary diagnostic factor of the first diagnostic factor group (1210) being corrected by the factor correction model (200) so that an error between the charge / discharge performance indicated by the first diagnostic factor group (1210) and the actual charge / discharge performance of the target cell (BC) is reduced. The second diagnostic factor group (1220) may represent an estimated result of the charge / discharge performance of the target cell (BC) that would have been determined if the diagnostic logic were applied to the second target full-cell profile (N). For example, a specific diagnostic factor (e.g., a positive electrode engagement start point) of the second diagnostic factor group (1220) may be the specific diagnostic factor of the first diagnostic factor group (1210) that is corrected to be close to the actual charge / discharge performance of the target cell (BC).

[0207] The factor correction model (200) may be a machine learning model trained using a training data set including pairs of first and second diagnostic result information for each of a plurality of test cells. The training data set may further include temperature information and / or impedance information for each of the plurality of test cells.

[0208] A plurality of test cells are prepared in advance for the purpose of training the factor correction model (200). At least one of the plurality of test cells may be a new battery cell verified as good. Each of the remaining test cells may have at least one of the positive and negative electrodes forcibly degraded from a new state through charge / discharge cycling different from that of the other test cells.

[0209] The first diagnostic result information of a specific test cell may include a first diagnostic factor group, which is previously acquired by applying diagnostic logic to a modified test full-cell profile of the test cell. The modified test full-cell profile of the specific test cell may be a first test full-cell profile of the test cell that is modified based on an overvoltage profile (OP). The first test full-cell profile may indicate a correspondence between a capacity factor and a voltage of the test cell while a first electrical stimulus is applied to the test cell. The first diagnostic factor group of the specific test cell may include at least one type of diagnostic factor among a positive engagement start point, a negative engagement start point, a positive engagement end point, a negative engagement end point, a positive scale factor, and a negative scale factor as a preliminary diagnostic factor of the specific test cell.

[0210] The temperature information of a specific test cell includes at least one of a start temperature, an end temperature, an average temperature, a maximum temperature, and a minimum temperature during the application period of the first electrical stimulus to the test cell.

[0211] The second diagnostic result information of a specific test cell may include a second diagnostic factor group, which is acquired in advance by applying diagnostic logic to a second test full-cell profile of the test cell. The second test full-cell profile of the specific test cell may indicate a correspondence between a capacity factor and a voltage of the test cell while a second electrical stimulus is applied to the test cell under a certain external environment. The second diagnostic factor group of the specific test cell may include at least one type of diagnostic factor among a positive participation start point, a negative participation start point, a positive participation end point, a negative participation end point, a positive scale factor, a negative scale factor, a positive loading amount, a negative loading amount, and a NP ratio as a precision diagnostic factor.

[0212] The graphs illustrated in Figures 13 to 21 are individually associated with nine types of diagnostic factors, and represent multiple data points representing pairs of the same type of diagnostic factor included in the training data set, marked on two-dimensional coordinates. The number of data points marked for each graph may be equal to the number of test cells.

[0213] Each data point is defined by two estimates for a specific type of diagnostic factor. That is, the X-axis coordinate of each data point represents the value of a preliminary diagnostic factor included in the first diagnostic result information as an estimate of the specific diagnostic factor, and the Y-axis coordinate represents the value of a precise diagnostic factor included in the second diagnostic result information as another estimate of the specific diagnostic factor.

[0214] The X-axis and Y-axis of Fig. 13 are exemplified as representing the anode SOC at the anode participation start point, respectively. The X-axis and Y-axis of Fig. 14 are exemplified as representing the anode SOC at the anode participation end point, respectively. The X-axis and Y-axis of Fig. 15 are exemplified as representing the anode scale factor, respectively. The X-axis and Y-axis of Fig. 16 are exemplified as representing the anode loading amount, respectively. For reference, the anode loading amount included in the first diagnostic factor group may be equal to the product of the anode scale factor included in the first diagnostic factor group and the reference anode loading amount.

[0215] The X-axis and Y-axis of Fig. 17 are exemplified as representing the negative SOC at the negative participation start point, respectively. The X-axis and Y-axis of Fig. 18 are exemplified as representing the negative SOC at the negative participation end point, respectively. The X-axis and Y-axis of Fig. 19 are exemplified as representing the negative scale factor, respectively. The X-axis and Y-axis of Fig. 20 are exemplified as representing the negative loading amount, respectively. The X-axis and Y-axis of Fig. 21 are exemplified as representing the NP ratio, respectively.

[0216] Referring to Figures 13 to 21, the data points of the training data set are distributed so as to exhibit a learnable tendency. That is, the correlation between the values ​​included in the first diagnostic result information and the values ​​included in the second diagnostic result information as estimates of the type-specific diagnostic factors can be learned by the factor correction model (200).

[0217] The factor correction model (200) can be learned based on the correlation between two estimates for a specific diagnostic factor, and the correlation information between the two estimates obtained through learning can be expressed as the correlation coefficient of Fig. 23. A detailed description will be provided later.

[0218] Among the nine types of diagnostic factors described above, three (i.e., anodic loading, cathodic loading, and NP ratio) can be said to be dependent on the remaining six diagnostic factors. Therefore, the following explanation will focus on the six diagnostic factors.

[0219] Referring to FIG. 22, the neural network of the factor correction model (200) may include an input layer (1000), an intermediate layer (2000), and an output layer (3000).

[0220] In the factor correction model (200), the number of nodes included in each layer, the connections between nodes, the functions of each node included in the intermediate layer (2000), etc. may be predetermined. In addition, the weights for each connection between nodes may be automatically determined through a machine learning process using a training data set.

[0221] The input layer (1000) may include first to sixth input nodes (I1 to I6). The input layer (1000) may include at least one additional input node (I TEMP , I IMP ) may be included.

[0222] When i is a natural number less than or equal to 6, the ith input node (Ii) may be associated with one diagnostic factor among the first diagnostic factor group (1210). In Fig. 22, for convenience of explanation, it is assumed that the first to sixth input nodes (I1 to I6) are associated with six types of diagnostic factors that may be included in the first diagnostic factor group, namely, a positive participation start point, a positive participation end point, a positive scale factor, a negative participation start point, a negative participation end point, and a negative scale factor, respectively.

[0223] An ith input node (Ii) may be provided with an ith input data set (Xi), which is diagnostic factor data associated with it. For example, a first input node (I1) may be provided with a first input data set (X1). The first input data set (X1) may include values ​​relating to the positive engagement start points of a plurality of test cells (e.g., positive starting potential and capacitance values).

[0224] Additional input node (I TEMP ), an additional input data set (X) associated with it TEMP ) may be provided. Additional input data sets (X TEMP ) may include temperature information of multiple test cells (e.g., at least one of a start temperature, an end temperature, an average temperature, a maximum temperature, and a minimum temperature).

[0225] Additional input node (I IMP ), an additional input data set (X) associated with it IMP ) may be provided. Additional input data sets (X IMP ) may include impedance information of a plurality of test cells. The impedance information for each test cell may include at least one of the first to third impedances.

[0226] The first impedance may be the impedance during high-rate charging as the first electrical stimulus. The second impedance may be the impedance during high-rate discharging as the first electrical stimulus. The third impedance may be the result of the first impedance or the second impedance being corrected by cell temperature. Each impedance may be the impedance at the start or end of the application period of the first electrical stimulus. Alternatively, each impedance may be the average impedance during the application period of the first electrical stimulus.

[0227] For reference, it is common knowledge that the impedance of a battery cell has a characteristic that is highly dependent on temperature. Therefore, the first impedance (or second impedance) can be compensated based on the temperature difference between the cell temperature measured at the same timing as the determination time of the first impedance (or second impedance) and the reference temperature. Predetermined relationship data between the correction amount for the impedance and the temperature difference can be recorded in memory. Meanwhile, it is widely known that the impedance of a battery cell can be determined using Ohm's law, electrochemical impedance spectroscopy (EIS), etc., and therefore, a description thereof is omitted herein.

[0228] The output layer (3000) may include at least one of the first to sixth output nodes (O1 to O6). When j is a natural number less than or equal to 6, the j-th output node (Oj) may be associated with any one of a positive participation start point, a positive participation end point, a positive scale factor, a negative participation start point, a negative participation end point, and a negative scale factor. In Fig. 22, for convenience of explanation, it is assumed that the first to sixth output nodes (O1 to O6) are associated with six types of diagnostic factors, namely a positive participation start point, a positive participation end point, a positive scale factor, a negative participation start point, a negative participation end point, and a negative scale factor, respectively. The positive participation start point, the positive participation end point, the positive scale factor, the negative participation start point, the negative participation end point, and the negative scale factor may be referred to as the first to sixth diagnostic factors in that order.

[0229] Input nodes (I1~I6, I TEMP, I IMP ) input data set (X1~X6, X TEMP, X IMP ) is input, the jth output data set (Zj) can be output from the jth output node (Oj). For example, if the first output node (O1) is associated with the positive participation start point, the third output data set (Z3) can include the results of correcting the values ​​of the first input data set (X1).

[0230] In Fig. 22, the input layer (1000) includes the first to sixth input nodes (I1 to I6), and the output input layer (2000) includes the first to sixth output nodes (O1 to O6), but this is only an example. That is, the input layer (1000) may include at least one of the first to sixth input nodes (I1 to I6), and the output layer (3000) may also include at least one of the first to sixth output nodes (O1 to O6). For example, the input data set (X1 to X6, X TEMP, X IMP) If all of them are provided to the input layer (1000), the output layer (3000) may output only one of the first to sixth output data sets (Z1 to Z6).

[0231] Which output data set among the first to sixth output data sets (Z1 to Z6) will be output by the factor correction model (200) can be determined by the connections between nodes, the weights for each connection between nodes, and the functions of each node included in the intermediate layer (2000), and is not particularly limited.

[0232] The intermediate layer (2000) may include the first to mth intermediate nodes (F1 to Fm, where m is a natural number greater than or equal to 2). When k is a natural number less than or equal to m, the kth intermediate node (Fk) may include input nodes (I1 to I6, I TEMP, I IMP ) and at least one of the first to sixth output nodes (O1 to O6). The k-th intermediate node (Fk) may have a form of a function determined through a learning process, and may transmit an estimate calculated based on input values ​​from each input node connected thereto to each output node connected thereto. The j-th output node (Oj) may output an estimate equal to the sum of the estimates received from each intermediate node connected thereto as a correction result of the first diagnostic factor group.

[0233] The function of the kth intermediate node (Fk) can be generated based on the correlation coefficient between the diagnostic factors associated with each node of the input layer (1000) connected to the kth intermediate node (Fk) and the diagnostic factors associated with each node of the output layer (3000) connected to the kth intermediate node (Fk). For reference, the correlation coefficient is a real number between -1 and +1, and a correlation coefficient closer to -1 indicates a negative correlation between two factors, and a correlation coefficient closer to +1 indicates a positive correlation between two factors.

[0234] The function for each intermediate node of the intermediate layer (2000) may be a weighted average function. In this case, a correlation coefficient indicating the degree of correlation between the estimates of the first to sixth diagnostic factors included in the first diagnostic factor group and the estimates of at least one of the first to sixth diagnostic factors included in the second diagnostic factor group may be used as the weight of the function for each intermediate node of the intermediate layer (2000).

[0235] Accordingly, the factor correction model (200) may include at least one of the first to sixth machine learning models. The first to sixth machine learning models may be models that sequentially provide precise diagnosis results for the first to sixth diagnostic factors.

[0236] When the target cell (BC) is in the MOL state, the control circuit (130) can determine at least one degradation parameter based on a second diagnostic factor group. Table 1 below summarizes the degradation parameters and formulas that can be used to determine each degradation parameter. For reference, the second diagnostic factor group when the target cell (BC) was in a new state may already be recorded in the memory (131).

[0237] Table 1

[0238]

[0239] Each of the variables listed in Table 1 is a diagnostic factor that can be included in the second diagnostic factor group described above. The definitions of the deterioration parameters and variables in Table 1 may be as follows.

[0240] <Deterioration parameters>

[0241] P SOH : The anode SOH (State Of Heat) of the target cell (BC)

[0242] N SOH : Cathode SOH of the target cell (BC)

[0243] L SOH : Available lithium SOH of target cell (BC)

[0244] F SOH : Full-cell SOH of the target cell (BC)

[0245] P LOSS : Anode loss rate of target cell (BC)

[0246] N LOSS : Cathode loss rate of target cell (BC)

[0247] L LOSS : Available lithium loss rate of target cell (BC)

[0248] F LOSS : Full cell loss rate of target cell (BC)

[0249] P loading_MOL : Anode loading of the target cell (BC)

[0250] N loading_MOL : Cathode loading of target cell (BC)

[0251] N / P _MOL : NP ratio of target cell (BC)

[0252]

[0253] As a battery cell deteriorates, at least one of the total positive electrode capacity, total negative electrode capacity, available lithium content, and total full cell capacity of the battery cell may gradually decrease from the value at the beginning of life (BOL) state. The total full cell capacity may represent the capacity difference between the two end points of the full cell profile. For example, the total full cell capacity may mean the full charge capacity (FCC). The available lithium content may represent the total amount of lithium that can contribute to charging and discharging the battery cell. P SOH can represent the retention rate of the total anode capacity. N SOH can represent the retention rate of the total cathode capacity. L SOH can represent the retention rate of available lithium. F SOH can represent the retention rate of the total full cell capacity.

[0254] P SOH and P LOSSThe sum of, N SOH and N LOSS The sum of L SOH and L LOSS The sum of, F SOH and F LOSS The sum of each of F can be equal to 1. LOSS is P LOSS and L LOSS may be equal to the sum of

[0255] The positive electrode loading of a battery cell refers to the amount of positive electrode active material (or available capacity) per unit area of ​​the positive electrode of the battery cell. The negative electrode loading of a battery cell refers to the amount of negative electrode active material (or available capacity) per unit area of ​​the negative electrode of the battery cell. The unit of loading is mAh / cm 2 or mg / cm 2 It can be. In Table 1, P loading_ref represents the standard anode loading, and N loading_ref represents the reference cathode loading.

[0256] The reference anode loading is a predetermined value indicating the amount of anode active material (or available capacity) per unit area of ​​the anode of a reference cell. The reference anode loading is the reference anode capacity (Q P_ref ) may be a value obtained by dividing the reference anode area by the reference anode capacity. Here, the reference anode capacity may be a value preset as the total anode capacity of the reference cell. The reference anode area may be a value preset as the area of ​​the anode of the reference cell.

[0257] The reference negative electrode loading is a predetermined value indicating the amount of negative active material (or available capacity) per unit area of ​​the negative electrode of the reference cell. The reference negative electrode loading is the reference negative electrode capacity (Q N_ref ) may be a value obtained by dividing the reference cathode area by the reference cathode capacity. Here, the reference cathode capacity may be a value preset as the total cathode capacity of the reference cell. The reference cathode area may be a value preset as the area of ​​the cathode of the reference cell.

[0258] At least one of the degradation parameters in Table 1 may be included in the second diagnostic factor group as an estimation result for an additional diagnostic factor of the target cell (BC).

[0259]

[0260] <variables>

[0261] pi BOL : Anode capacity at the start of anode participation when the target cell (BC) was in BOL state (anode SOC)

[0262] pi MOL : The anode capacity (anode SOC) of the current anode participation starting point of the target cell (BC) (e.g., pi' shown in Fig. 8).

[0263] pf BOL : Anode capacity (anode SOC) at the anode participation end point when the target cell (BC) was in BOL state

[0264] pf MOL : The anode capacity (anode SOC) of the current anode participation end point (e.g., pf'' shown in Fig. 8) of the target cell (BC).

[0265] ni BOL : Cathode capacity (cathode SOC) at the start of cathode participation when the target cell (BC) was in BOL state

[0266] ni MOL : Cathode capacity (cathode SOC) of the current cathode participation starting point of the target cell (BC) (e.g., ni as shown in Fig. 8)

[0267] nf BOL : Cathode capacity (cathode SOC) at the end point of cathode participation when the target cell (BC) was in BOL state

[0268] nf MOL : The cathode capacity (cathode SOC) of the current cathode participation end point (e.g., nf' shown in Fig. 8) of the target cell (BC).

[0269] ps BOL : Bipolar scale factor when the target cell (BC) was in BOL state

[0270] ps MOL: Current bipolar scale factor of the target cell (BC)

[0271] ns BOL : Cathode scale factor when the target cell (BC) was in BOL state

[0272] ns MOL : Current cathode scale factor of the target cell (BC)

[0273]

[0274] The NP ratio can also be expressed as the N / P ratio, N:P ratio, etc. The NP ratio of the target cell (BC) is (i) the anode loading amount (P) of the target cell (BC) loading_MOL ) for cathode loading (N) loading_MOL ) or (ii) a value representing the ratio of the total cathode capacity to the total anode capacity of the target cell (BC). The control circuit (130) may be ... MOL and Q P_ref The total anode capacity of the target cell (BC) can be determined by multiplying the voltage by the voltage. The control circuit (130) is ns MOL and Q N_ref The total anode capacity of the target cell (BC) can be determined by multiplying the same amount by the same amount.

[0275] The process of determining the second diagnostic factor group of the target cell (BC) can be repeated periodically or aperiodically throughout the life of the target cell (BC).

[0276] Figure 23 shows an example of correlation information between the first diagnostic factor group and the second diagnostic factor group, obtained through learning for the factor correction model (200), in matrix form.

[0277] Referring to Fig. 23, the first to sixth rows represent the first to sixth diagnostic factors of the first diagnostic factor group provided as a training data set in order. The six columns represent the first to sixth diagnostic factors of the second diagnostic factor group provided as a training data set in order. In Fig. 23, the seventh to ninth rows correspond to the start temperature (T_ini), end temperature (T_end), and maximum temperature (T_max) included in the temperature information, and the tenth to twelfth rows correspond to the first impedance (z_c), the second impedance (z_d), and the third impedance (z_temp) included in the impedance information.

[0278] In Fig. 23, pi_A[1], pf_A[2], ps_A[3], ni_A[4], nf_A[5] and ns_A[6] represent, in order, six types of preliminary diagnostic factors included in the first diagnostic factor group of the training data set: positive involvement start point, positive involvement end point, positive scale factor, negative involvement start point, negative involvement end point and negative scale factor. In addition, pi_B[1], pf_B[2], ps_B[3], ni_B[4], nf_B[5] and ns_B[6] represent, in order, six types of precise diagnostic factors included in the second diagnostic factor group of the training data set: positive involvement start point, positive involvement end point, positive scale factor, negative involvement start point, negative involvement end point and negative scale factor.

[0279] When p and q are natural numbers less than or equal to 6, the values ​​of the pth row and the qth column represent the correlation coefficients between the pth diagnostic factor included in the first diagnostic factor group and the qth diagnostic factor included in the second diagnostic factor group.

[0280] For example, the correlation coefficient between the first diagnostic factor (pi_A[1]) in row 1 and the second diagnostic factor (pf_B[2]) in column 2 is -0.52. As another example, the correlation coefficient between the fifth diagnostic factor (nf_A[5]) in row 5 and the fourth diagnostic factor (ni_B[4]) in column 4 is 0.46.

[0281] The values ​​in the 7th row and the qth column represent the correlation coefficient between the starting temperature (T_ini) and the qth diagnostic factor. The values ​​in the 8th row and the qth column represent the correlation coefficient between the ending temperature (T_end) and the qth diagnostic factor. The values ​​in the 9th row and the qth column represent the correlation coefficient between the maximum temperature (T_max) and the qth diagnostic factor.

[0282] The values ​​in the 10th row and the qth column represent the correlation coefficient between the first impedance (z_c) and the qth diagnostic factor. The values ​​in the 11th row and the qth column represent the correlation coefficient between the second impedance (z_d) and the qth diagnostic factor. The values ​​in the 12th row and the qth column represent the correlation coefficient between the third impedance (z_temp) and the qth diagnostic factor.

[0283] Meanwhile, Fig. 23 shows six types of diagnostic factors, but the present invention is not limited thereto. For example, the second diagnostic factor group of the training data set may further include three types of precision diagnostic factors corresponding to positive loading, negative loading, and NP ratio. In this case, the correlation information between the first diagnostic factor group and the second diagnostic factor group may be expressed as a 6 × 9 matrix.

[0284] Fig. 24 is a flowchart schematically illustrating a battery diagnosis method according to another embodiment of the present invention. The method of Fig. 24 can be executed by a battery diagnosis device (100).

[0285] Referring to FIGS. 1 to 24, in step S2410, the control circuit (130) collects, from the sensing unit (110), a measurement signal representing the measured values ​​of the voltage and current of the target cell (BC) during the application period of the first electrical stimulus to the target cell (BC). The measurement signal may further represent the temperature of the target cell (BC) during the application period of the first electrical stimulus.

[0286] In step S2420, the control circuit (130) generates a target full-cell profile (M) associated with the first electrical stimulation based on the measurement signal collected in step S2410. The target full-cell profile (M) may represent a correspondence between a capacity factor and a voltage of the target cell (BC) while the first electrical stimulation is applied. In step S2420, at least one of temperature information and impedance information of the target cell (BC) may be further generated. In the present specification, the diagnostic target information includes the target full-cell profile (M) and may further include at least one of temperature information and impedance information.

[0287] Steps S2410 and S2420 may be replaced with a procedure in which the data acquisition unit acquires the diagnostic target information directly or externally.

[0288] In step S2430, the control circuit (130) modifies the target full-cell profile (M) to be associated with the second electrical stimulus based on the overvoltage profile (OP). As a result, a modified target full-cell profile (E) is generated.

[0289] In step S2440, the control circuit (130) applies the diagnostic logic (see FIGS. 4 to 11) to the modified target full cell profile (E) to determine first diagnostic result information as a preliminary diagnostic result for the charge / discharge performance of the target cell (BC).

[0290] In step S2450, the control circuit (130) determines second diagnosis result information as a precise diagnosis result for the charge / discharge performance of the target cell (BC) based on at least one preliminary diagnosis factor included in the first diagnosis result information using the factor correction model (200). Referring to FIG. 22, the second diagnosis result information is information output from the factor correction model (200) when correction target information is input to the factor correction model (200). The correction target information includes at least one preliminary diagnosis factor and may further include at least one of temperature information and impedance information. The factor correction model (200) may determine at least one precise diagnosis factor by correcting at least one preliminary diagnosis factor using at least one of the temperature information and the impedance information as reference information.

[0291] The second diagnostic factor group (1220) included in the second diagnostic result information represents an estimated result of charge / discharge performance that can be determined if the above diagnostic logic is applied to the target full-cell profile (N) instead of the modified target full-cell profile (E).

[0292] The target full-cell profile (N) represents the correspondence between the capacity factor and the voltage of the target cell (BC) during the application period of the second electrical stimulus, which is different from the first electrical stimulus. The target full-cell profile (N) may not be acquired by actually applying the second electrical stimulus to the target cell (BC). In other words, the target full-cell profile (N) may represent the correspondence between the capacity factor and the voltage of the target cell (BC) that would be expected to be acquired if the second electrical stimulus, instead of the first electrical stimulus, were applied to the target cell (BC).

[0293] Following step S2450, the control circuit (130) may apply a mathematical operation to the second diagnostic result information to determine at least one degradation parameter of the target cell (BC) (see Table 1). The control circuit (130) may limit at least one of an allowable voltage range, an allowable SOC range, and an allowable charge / discharge current for the target cell (BC) based on the at least one degradation parameter. The memory (131) may pre-store relationship data indicating a correspondence between at least one restriction item (i.e., an allowable voltage range, an allowable SOC range, and / or an allowable charge / discharge current) and at least one degradation parameter. As an example, a specific type of degradation parameter (e.g., P LOSS , N LOSS , L LOSS , F LOSS ) from the BOL state and / or other types of degradation parameters (e.g., P SOH , N SOH , L SOH , F SOH ) the greater the decrease from the BOL state, the greater the limits on the allowable voltage range, allowable SOC range and / or allowable charge / discharge current.

[0294]

[0295] Meanwhile, if the amount of the training data set used for learning the factor correction model (200) is insufficient, the correction performance of the factor correction model (200) may deteriorate. If the correction performance of the factor correction model (200) is low, the second diagnostic factor group output from the factor correction model (200) may not properly represent the actual charge / discharge performance of the target cell (BC).

[0296] In addition, if the aforementioned degradation parameters are determined based on output data from a factor compensation model (200) whose compensation performance is lower than a certain level, incorrect control of the allowable voltage range, allowable SOC range, and / or allowable current range of the target cell (BC) may be executed. As a result, the lifespan of the target cell (BC) may be accelerated, or it may become difficult to fully utilize the charge / discharge performance of the target cell (BC).

[0297] Fig. 25 is a flowchart schematically illustrating an additional battery diagnosis method related to the battery diagnosis method of Fig. 24, and Fig. 26 is a drawing referenced in explaining the method according to Fig. 25. The method of Fig. 25 is for evaluating the performance of a factor correction model (200) and can be executed by a battery diagnosis device (100) prior to execution of the method according to Fig. 24.

[0298] Referring to FIG. 25, in step S2510, the control circuit (130) evaluates the performance indicators of the factor correction model (200). The performance indicator evaluation may be based on an evaluation data set. The evaluation data set may be received from an external source by a data acquisition unit or may be pre-stored in a memory (131). The evaluation data set may include a plurality of first diagnostic factor groups and a plurality of comparison factor groups that are paired one-to-one with each other. The evaluation data set may be prepared in advance through a method common to the method of acquiring the training data set described above.

[0299] Performance evaluation information representing the evaluated performance indicators may represent the classification performance when the factor correction model (200) is used as a classification model. The performance evaluation information may include the values ​​of at least one type of performance indicator.

[0300] The control circuit (130) can compare the second diagnostic factor group output from the factor correction model (200) when inputting each first diagnostic factor group of the evaluation data set into the factor correction model (200) with each comparison factor group of the evaluation data set associated with the input first diagnostic factor group, and determine the difference between the two. At this time, two estimates for the same type of diagnostic factor can be compared with each other. For example, the difference between the value of the first diagnostic factor included in the comparison factor group of the evaluation data set and the value of the first diagnostic factor included in the second diagnostic factor group output from the factor correction model (200) can be determined.

[0301] If the difference between two values ​​included in the second diagnostic factor group and the comparison factor group as a specific type of diagnostic factor is within a predetermined tolerance range, the factor correction model (200) may be determined to have output a correct answer. On the other hand, if the difference between the two values ​​is outside the tolerance range, the factor correction model (200) may be determined to have output an incorrect answer. The tolerance range may be individually predetermined for each type of diagnostic factor.

[0302] Figure 26 illustrates performance evaluation information in the form of a data table. Referring to Figure 26, precision, recall, accuracy, and F1-score, which are widely used for performance evaluation of machine learning models, are illustrated as four performance indicators.

[0303] In step S2520, the control circuit (130) determines whether the performance indicator of the factor correction model (200) is greater than or equal to a threshold. The threshold may be individually preset for each performance indicator. For example, if the threshold for precision is 0.70, the precision illustrated in FIG. 26 is 0.76923, and therefore the value of step S2520 may be output as "Yes." The value of step S2520 being "Yes" indicates that each performance indicator of the performance evaluation information is greater than or equal to the threshold. If the value of step S2520 is "Yes," the process may proceed to step S2530. If the value of step S2520 is "Yes," the process may proceed to step S2540.

[0304] A performance index exceeding a threshold value indicates that the factor correction model (200) has been sufficiently trained to estimate the charge / discharge performance of the target cell (BC) with a predetermined level of accuracy or higher. A performance index falling below the threshold value indicates that the factor correction model (200) has not been sufficiently trained.

[0305] In step S2530, the control circuit (130) generates a first evaluation result value. If the first evaluation result value is generated, execution of step S2450 according to FIG. 24 may be permitted.

[0306] In step S2540, the control circuit (130) generates a second evaluation result value. If the second evaluation result value is generated, execution of step S2450 according to FIG. 24 may be prohibited. In addition, the control circuit (130) may transmit a message notifying that additional learning of the factor correction model (200) is required to a user device, such as a battery system (1), using the communication circuit (150).

[0307] The embodiments of the present invention described above are not implemented only through devices and methods, but may also be implemented through a program that realizes a function corresponding to the configuration of the embodiments of the present invention or a recording medium on which the program is recorded, and such implementation can be easily implemented by an expert in the technical field to which the present invention belongs based on the description of the embodiments described above.

[0308] Although the present invention has been described above with reference to limited embodiments and drawings, the present invention is not limited thereto, and it is obvious that various modifications and variations are possible within the scope of the technical idea of ​​the present invention and the equivalent scope of the patent claims to be described below by a person having ordinary skill in the art to which the present invention pertains.

[0309] In addition, the present invention described above is not limited to the above-described embodiments and the attached drawings, and all or part of each embodiment may be selectively combined and configured so that various modifications can be made, as those skilled in the art can make various substitutions, modifications, and changes within the scope of the technical idea of ​​the present invention.

Claims

1. A step of obtaining diagnostic target information including a target full-cell profile of a battery cell associated with a first electrical stimulus; A step of modifying the target full-cell profile to be associated with a second electrical stimulus different from the first electrical stimulus based on a predetermined overvoltage profile; A step of applying diagnostic logic to the above modified target full-cell profile to determine first diagnostic result information, which is a preliminary diagnostic result for the charge / discharge performance of the battery cell; and A step of determining second diagnosis result information, which is a precise diagnosis result for the charge / discharge performance of the battery cell, based on at least one preliminary diagnosis factor included in the first diagnosis result information, using a factor correction model; A battery diagnostic method comprising:

2. In paragraph 1, A battery diagnosis method, wherein an instantaneous voltage change induced in the battery cell when the second electrical stimulus is applied is smaller than an instantaneous voltage change induced in the battery cell when the first electrical stimulus is applied.

3. In paragraph 1, The above first electrical stimulation is a charging current greater than or equal to the first current rate, A battery diagnostic method, wherein the second electrical stimulation is a charging current having a second current rate lower than the first current rate.

4. In paragraph 1, The above first electrical stimulation is a discharge current having a first current rate or higher, A battery diagnostic method, wherein the second electrical stimulation is a discharge current having a second current rate or lower, which is lower than the first current rate.

5. In paragraph 1, The above overvoltage profile represents the difference between the first reference full-cell profile and the second reference full-cell profile, The above first reference full-cell profile is a predetermined relationship between the capacity factor and voltage of the reference cell while the first electrical stimulus is applied. A battery diagnosis method, wherein the second reference full-cell profile is predetermined as a correspondence between the capacity factor and voltage of the reference cell while the second electrical stimulus is applied.

6. In paragraph 1, The steps for modifying the above target full cell profile are: A battery diagnosis method, wherein the modified target full-cell profile is generated by subtracting the overvoltage profile from the target full-cell profile.

7. In paragraph 1, The above diagnostic target information further includes temperature information of the battery cell measured during the application period of the first electrical stimulation, A battery diagnosis method, wherein the temperature information is input into the factor correction model together with the at least one preliminary diagnosis factor.

8. In paragraph 1, The above diagnostic target information further includes impedance information of the battery cell measured during the application period of the first electrical stimulation, A battery diagnosis method, wherein the above impedance information is input into the factor correction model together with the at least one preliminary diagnosis factor.

9. In paragraph 1, The above factor correction model is, A battery diagnosis method, wherein the machine learning model is trained by a training data set including pairs of first diagnosis result information and second diagnosis result information of each of a plurality of test cells having different charge / discharge performances.

10. In paragraph 9, The first diagnostic result information of each of the plurality of test cells is acquired by applying the diagnostic logic to each of the plurality of modified test full cell profiles, The above multiple modified test full-cell profiles are individually modified first test full-cell profiles associated with the first electrical stimulus based on the above overvoltage profile, A battery diagnosis method, wherein the second diagnostic result information of each of the plurality of test cells is acquired by applying the diagnostic logic to a plurality of second test full cell profiles associated with the second electrical stimulation.

11. In paragraph 1, A battery diagnosis method, wherein the step of determining the second diagnosis result information is executed under the condition that the performance indicator of the factor correction model is evaluated to be above a threshold value.

12. In paragraph 11, A step of transmitting a message to a user device notifying that additional learning of the factor compensation model is required when the performance indicator of the factor compensation model is evaluated to be less than the threshold; A battery diagnostic method further comprising:

13. In paragraph 1, The above second diagnostic result information is, A battery diagnosis method, comprising at least one type of diagnostic factor among a cathode participation start point, a cathode participation end point, and a cathode scale factor associated with the charge / discharge performance of the cathode, as a precision diagnostic factor of the battery cell.

14. In paragraph 1, The above second diagnostic result information is, A battery diagnosis method, comprising at least one type of diagnostic factor among a negative electrode participation start point, a negative electrode participation end point, and a negative electrode scale factor related to the charge / discharge performance of the negative electrode of the battery cell, as a precision diagnostic factor of the battery cell.

15. In paragraph 1, The above second diagnostic result information is, A battery diagnosis method, comprising at least one type of diagnostic factor among a positive electrode loading amount associated with the charge / discharge performance of the positive electrode of the battery cell, a negative electrode loading amount associated with the charge / discharge performance of the negative electrode of the battery cell, and a NP ratio associated with the charge / discharge performance of both the positive and negative electrodes of the battery cell, as a precision diagnostic factor of the battery cell.

16. A data acquisition unit configured to acquire diagnostic target information including a target full-cell profile of a battery cell associated with the first electrical stimulation; and A control circuit configured to modify the target full-cell profile to be associated with a second electrical stimulus different from the first electrical stimulus, based on a predetermined overvoltage profile; The above control circuit, By applying diagnostic logic to the above modified target full-cell profile, first diagnostic result information, which is a preliminary diagnostic result for the charge / discharge performance of the battery cell, is determined, A battery diagnosis device configured to determine second diagnosis result information, which is a precise diagnosis result for charge / discharge performance of the battery cell, based on at least one diagnosis factor included in the first diagnosis result information, using a factor correction model.

17. In paragraph 16, The above control circuit, A battery diagnostic device configured to generate the modified target full-cell profile by subtracting the overvoltage profile from the target full-cell profile.

18. In paragraph 16, The above factor correction model is, A battery diagnostic device, wherein the machine learning model is trained by a training data set including pairs of first diagnostic result information and second diagnostic result information of each of a plurality of test cells having different charge / discharge performances.

19. A battery pack comprising a battery diagnostic device according to any one of claims 16 to 18.

20. A battery system comprising a battery diagnostic device according to any one of claims 16 to 18.

Images