Method for detecting internal short-circuited battery cells

Abstract

The method for detecting an internal short-circuited battery cell according to various embodiments includes the following steps: periodically generating voltage values ​​and current values ​​corresponding to a plurality of battery cells that are electrically connected to each other between a first terminal and a second terminal and are in use; updating a G parameter value and an H parameter value in real time according to the voltage value and current value of the corresponding battery cell by using an adaptive filter, the G parameter value and the H parameter value being obtained by digitizing a G parameter and an H parameter indicating a current state of the corresponding battery cell, respectively; calculating a G parameter representative value and an H parameter representative value representing the G parameter value and the H parameter value of the battery cell, respectively; and determining whether an internal short circuit occurs in the corresponding battery cell based on the G parameter value and the H parameter value of the corresponding battery cell, the G parameter representative value, and the H parameter representative value. The G parameter is a parameter indicating the sensitivity of the voltage of the battery cell to the current change, and the H parameter is a parameter indicating the effective potential determined by the distribution of the resistance distribution and the distribution of the local equilibrium potential within the battery cell.

Description

Technical Field

[0001] The present disclosure relates to a method of accurately and quickly detecting a battery cell in which an internal short circuit occurs among battery cells in use. Background Art

[0002] Batteries have high applicability compared to other energy storage devices, and due to characteristics such as relatively high energy and power density, have been widely used in electric vehicles (EV) or hybrid electric vehicles (HEV) driven by electric drive sources and portable devices. In particular, when strong output is required, a battery pack in which a plurality of batteries are connected in series and in parallel can also be used.

[0003] Battery management is important for energy saving and safe use of batteries or battery-powered electric devices, and for this purpose, accurate estimation and diagnosis of the state of the battery is essential. In particular, when the battery is used, an internal short circuit may occur between the positive and negative electrodes of the battery cell. For stable use of the battery, it is very important to quickly and accurately detect the internal short circuit.

[0004] In the prior art, the correlation between the measurable variables (e.g., current (I), voltage (V), and temperature (T)) of the battery to be estimated and the internal state is experimentally found in advance, and then the correlation is created in a tabular form, and the internal state of the battery (such as resistance, capacitance, or open circuit voltage (OCV)) is estimated. Because the internal state is affected by various variables, in order to accurately estimate the internal state, various variables such as the state of charge (SOC), current (I), voltage (V), and temperature (T) of the battery need to be combined, and data needs to be collected for each of them. Therefore, the higher the required accuracy, the more data is required, resulting in increased time and money costs. In addition, the problem with such a conventional method is that the accuracy is unreliable because the battery should be stopped, specific conditions or environments should be met, or sensor errors or estimation errors are accumulated. Summary of the invention

[0005] Technical issues

[0006] The present disclosure provides a method for accurately estimating whether an internal short circuit occurs in each battery cell in real time by using a voltage value and a current value obtained by measuring the voltage and current of each battery cell in use, so as to overcome the shortcomings of the conventional method. The present disclosure provides a method for accurately and quickly detecting a battery cell in which an internal short circuit occurs among battery cells by using a G parameter and an H parameter indicating the internal state of the battery cell. The present disclosure provides an internal short circuit battery cell detection algorithm that can be loaded into a battery management system (BMS).

[0007] Technical Solution

[0008] According to one aspect of the present disclosure, an internal short-circuited battery cell detection method includes the following steps: detecting the voltage and current of each of a plurality of battery cells that are electrically connected to each other between a first terminal and a second terminal and are in use, so as to periodically generate a voltage value and a current value of each of the plurality of battery cells; updating a G parameter value and an H parameter value in real time according to the voltage value and the current value of each of the plurality of battery cells by using an adaptive filter, the G parameter value and the H parameter value being obtained by digitizing a G parameter and an H parameter indicating a current state of each of the plurality of battery cells; calculating a G parameter representative value representing the G parameter value of the plurality of battery cells and calculating an H parameter representative value representing the H parameter value of the plurality of battery cells; and determining whether a short circuit occurs in each of the plurality of battery cells based on the G parameter value and the H parameter value of each of the plurality of battery cells, the G parameter representative value, and the H parameter representative value. The G parameter is a parameter indicating the sensitivity of the voltage of each of the plurality of battery cells to a current change, and the H parameter is a parameter indicating an effective potential determined by a local equilibrium potential distribution and a resistance distribution in each of the plurality of battery cells.

[0009] Beneficial Effects

[0010] The internal short-circuit battery cell detection method according to various embodiments of the present disclosure is greatly improved in cost, scalability, and adaptability compared to conventional methods. In particular, because the internal short-circuit battery cell detection method is not an estimation method based on estimation such as an existing experimental model, but estimates the internal state of the battery by using a value that can be directly measured, the internal short-circuit battery cell detection method has higher accuracy than conventional methods. In addition, the internal short-circuit battery cell detection method of the present disclosure can be loaded into a data battery management system (BMS) as an algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 A configuration of a battery system for performing an internal short-circuited battery cell detection method according to an embodiment is shown.

[0012] Figure 2 is a flow chart illustrating an internal short-circuited battery cell detection method according to an embodiment.

[0013] Figure 3 It is shown Figure 2 A detailed flow chart of some operations of the internal short circuit battery cell detection method.

[0014] Figure 4 Parameter values ​​measured and estimated in a battery pack including normal battery cells are shown.

[0015] Figure 5 Shown in Figure 4The measurement or estimate obtained after replacing a normal battery cell with an internally shorted battery cell in a battery pack. DETAILED DESCRIPTION

[0016] With reference to the disclosed embodiments described in detail below in conjunction with the accompanying drawings, the advantages and features of the present disclosure and the methods for achieving the same will become apparent. However, the present disclosure is not limited to the embodiments presented below, but may be embodied in various forms, and it will be understood that all changes, equivalents and substitutes that do not depart from the spirit and technical scope of the present disclosure are included in the present disclosure. On the contrary, these embodiments are provided so that the present disclosure will be thorough and complete, and the scope of the present disclosure will be fully conveyed to those of ordinary skill in the art. In the specification of the present disclosure, when it is believed that the omission of certain detailed explanations of the related art will not necessarily obscure the essence of the present disclosure, certain detailed explanations of the related art are omitted.

[0017] The terms used in this application are only used to describe specific embodiments, and are not intended to limit the present disclosure. Unless the context clearly indicates otherwise, the singular forms "one", "one (kind / person)" and "the (described)" are also intended to include plural forms. In addition, as used in this application, the terms "including", "having" and their cognates can be understood to represent a certain feature, quantity, step, operation, constituent element, assembly or a combination thereof, but may not be interpreted as excluding the presence or addition of one or more other features, quantity, step, operation, constituent element, assembly or a combination thereof. It will be understood that, although the terms "first", "second" etc. can be used here to describe various components, these components should not be limited by these terms. These terms are only used to distinguish one component from another component.

[0018] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings, wherein the same or corresponding elements are denoted by the same reference numerals throughout and repeated descriptions thereof are omitted.

[0019] Figure 1 A configuration of a battery system for performing an internal short-circuit battery detection method according to an embodiment is shown.

[0020] Reference Figure 1 , the battery system 100 may include a battery 110 , a voltage measuring unit 120 , a current measuring unit 130 , a microprocessor 140 , and a memory 150 .

[0021] The battery 110 for storing electricity includes a plurality of battery cells 111 electrically connected between a first terminal 101 and a second terminal 102. The battery cells 111 may be connected in series, in parallel, or in a combination of the two. The battery cells 111 may have the same capacity and may release and charge the same amount of current. However, in reality, the internal states of the battery cells 111 may be different from each other. For example, the battery cells 111 may have different internal resistances and electromotive forces. The battery cells 111 may have different G parameter values ​​and H parameter values.

[0022] Each battery cell 111 may include a rechargeable secondary battery. For example, the battery cell 111 may include a nickel-cadmium battery, a lead storage battery, a nickel metal hydride (NiMH) battery, a lithium ion battery, or a lithium polymer battery. The number of battery cells 111 constituting the battery 110 may be determined according to the capacity, output voltage, and output current required by the battery 110.

[0023] Despite Figure 1 One battery 110 is shown in the figure, but a plurality of batteries 110 may be connected in parallel and / or in series, and may be connected to a load and / or a charging device through a first terminal 101 and a second terminal 102. Figure 1 1 is not shown, but the battery 110 connected to the load and / or the charging device is being used. That is, the battery 110 can be discharged by supplying current to the load, or can be charged by receiving power from the charging device.

[0024] The voltage measuring unit 120 may be connected to two terminals of each battery cell 111 being used, and may measure the voltage of each battery cell 111 to periodically generate a voltage value of each battery cell 111. When the battery cells 111 are connected in parallel, the voltage measuring unit 120 may measure the voltage of only one battery cell 111 because the battery cells 111 connected in parallel have the same voltage.

[0025] The voltage measuring unit 120 can measure the voltage of each battery cell 111 within a preset time period Δt. The time period Δt can be, for example, 1 second. The time period Δt can be set to another time length, for example, 0.1 seconds, 0.5 seconds, 2 seconds, 5 seconds, or 10 seconds. The time period Δt can be appropriately set according to the electric system connected to the battery system 100.

[0026] The voltage value currently or most recently measured is referred to as a current voltage value and is represented by V(t). The voltage value measured before a time period Δt is referred to as an immediately preceding voltage value and is represented by V(t-1).

[0027] The voltage values ​​of the battery cells 111 may be different from each other according to the internal state of the battery cells 111. The current voltage value and the immediately previous voltage value of the j-th battery cell 111 are represented by V_j(t) and V_j(t-1), respectively. Here, j may indicate the number of battery cells 111, and may be equal to or less than the total number of battery cells 111 included in the battery 110.

[0028] The current measuring unit 130 may measure the current of each battery cell 111 being used to periodically generate a current value of each battery cell 111. When the battery cells 111 are connected in series, the current measuring unit 130 may measure the current of only one battery cell 111 because the battery cells 111 connected in series have the same current.

[0029] The voltage measuring unit 120 and the current measuring unit 130 may be synchronized with each other and may simultaneously measure the voltage and current of each battery cell 111. The current measuring unit 130 may measure the current of each battery cell 111 within a time period Δt. The current value measured by the current measuring unit 130 may be displayed as positive (+) when it is a charging current, and may be displayed as negative (-) when it is a discharging current. The current value currently or most recently measured is referred to as the current current value and is represented by I(t), and the current value measured before the time period Δt is referred to as the immediately preceding current value and is represented by I(t-1).

[0030] The current value of each battery cell 111 may be different from each other according to the internal state of the battery cell 111. The present current value and the immediately preceding current value of the j-th battery cell 111 are represented by I_j(t) and I_j(t-1), respectively.

[0031] The microprocessor 140 may update the G parameter value and the H parameter value in real time according to the voltage value of each battery cell 111 provided by the voltage measuring unit 120 and the current value of each battery cell 111 provided by the current measuring unit 130, and the G parameter value and the H parameter value are obtained by digitizing the G parameter and the H parameter respectively indicating the current state of each battery cell 111. The G parameter is a parameter indicating the sensitivity of the voltage of the battery cell 111 being used to the current change, and the H parameter is a parameter indicating the effective potential determined by the local equilibrium potential distribution and the resistance distribution in the battery cell 111 being used. The G parameter and the H parameter of the j-th battery cell 111 are represented by G_j and H_j, respectively.

[0032] The G parameter value and the H parameter value of each battery cell 111 may be generated in real time by using an adaptive filter according to the voltage value and the current value of each battery cell 111. The G parameter value and the H parameter value of the jth battery cell 111 are represented by G_j(t) and H_j(t), respectively.

[0033] The microprocessor 140 may use an adaptive filter to generate the G parameter value and the H parameter value of each battery cell 111. The adaptive filter may be a filter using a recursive least squares (RLS) method or a filter using a weighted least squares (WLS) method. An embodiment in which the microprocessor 140 uses an RLS filter will now be described in detail.

[0034] The microprocessor 140 can identify an internal short-circuited battery cell in which an internal short circuit occurs among the battery cells 111 by using the G parameter value and the H parameter value of each battery cell 111. The term "internal short-circuited battery cell" refers to a battery cell in which physical direct contact occurs between the positive electrode and the negative electrode. When the battery cell is used for a long time in a high temperature environment or an external shock occurs, an internal short circuit may occur in the battery cell. The internal short circuit may occur in a gradual manner, or the internal short circuit may occur and then disappear in the battery cell depending on the external environment. In this specification, a battery cell in which an internal short circuit occurs is referred to as an internal short-circuited battery cell or an internally short-circuited battery cell.

[0035] According to an embodiment, the microprocessor 140 may calculate a G parameter representative value according to the G parameter value of each battery cell 111. The G parameter representative value may be determined as a value that may represent the G parameter. For example, the G parameter representative value may be an average value or a median value of the G parameter values.

[0036] The microprocessor 140 may calculate an H parameter representative value according to the H parameter value of each battery cell 111. The H parameter representative value may be determined as a value that may represent the H parameter. For example, the H parameter representative value may be an average value or a median value of the H parameter values.

[0037] The microprocessor 140 may determine whether an internal short circuit occurs in each battery cell 111 by using the G parameter value of each battery cell 111 , the H parameter value of each battery cell 111 , the G parameter representative value, and the H parameter representative value.

[0038] Because the microprocessor 140 uses only four simple arithmetic operations to detect internal short-circuited battery cells, the microprocessor 140 may be included in a battery management system (BMS) of the battery pack. According to another example, the internal short-circuited battery cell detection method according to the present embodiment may be executed by a microcontroller or an electronic control unit (ECU) in the BMS of an electric vehicle. According to another example, the internal short-circuited battery cell detection method according to the present embodiment may be executed by an integrated controller of an energy storage system. According to another example, the internal short-circuited battery cell detection method according to the present embodiment may be executed by a processor of a server that communicates with a battery system or an energy storage system.

[0039] The memory 150 may store commands and data required to execute the internal short-circuited battery cell detection method according to the present embodiment. According to the present embodiment, since the G parameter value and the H parameter value of each battery cell 111 may be generated based on the voltage value and the current value of each battery cell 111 generated in each time period Δt, and the internal short-circuited battery cell may be detected based on the G parameter value and the H parameter value of each battery cell 111, the current voltage value, the current current value, and the immediately preceding current value of each battery cell 111 may be stored in the memory 150, and other voltage data and current data may not need to be stored in the memory 150. That is, the memory 150 does not need to store a large amount of voltage data and current data.

[0040] According to the internal short-circuited battery cell detection method of the present embodiment, the immediately previous value and current value of the state vector including the G parameter and the H parameter of each battery cell 111 and the immediately previous value and current value of the covariance matrix required by the RLS method can be stored in the memory 150. The G parameter representative value and the H parameter representative value can be stored in the memory 150. Since the memory 150 does not need to store a large amount of commands and data, the memory 150 can be implemented as a memory with a small size. For example, the memory 150 can be implemented as a memory within the microprocessor 140.

[0041] Examples of important estimates that indicate the current state of a battery include the battery state of charge (SOC), the battery state of health (SOH), and the battery state of power capability (SOP). Conventional battery internal state estimation methods involve experimentally discovering in advance the correlation between measurable variables (e.g., current (I), voltage (V), and temperature (T)) of a battery cell to be estimated and the internal state, tabulating the correlation, and estimating the internal state of the battery cell (such as internal resistance, internal capacitance, or open circuit voltage (OCV)). Such battery state estimation methods are deficient in terms of cost, scalability, and adaptability.

[0042] According to the traditional battery internal state estimation method, data should be collected in advance through experiments, and the internal state of the battery should be estimated based on the collected data. Because the internal state is affected by various variables, in order to accurately estimate the internal state, various variables such as the state of charge (SOC), current (I), voltage (V) and temperature (T) of the battery need to be combined, and data needs to be collected for each of them. Therefore, the higher the required accuracy, the more data is required, resulting in increased time and money costs. In addition, when the internal state is estimated by performing extrapolation on values ​​that exceed the range of the variable measured experimentally, it may be difficult to rely on the estimate. In addition, when the internal state estimated by experiment is inaccurate, the utility of the estimated internal state is reduced. The method of estimating state variables by experimentally discovering correlations in advance and using correlations also has similar problems.

[0043] In order to extract the state information of the battery pack, experimental data about the battery pack is required. However, due to the high cost required to collect experimental data of the battery pack, the state information of the battery pack is usually estimated by using the data measured in the battery cell. However, this method is based on the assumption that the relationship between the variables measured in the battery cell is a scalable mechanism for a larger system such as a battery pack. Therefore, unless additional verification is performed in advance to prove the scalability of the mechanism between systems, the reliability and accuracy of the state information of the battery pack estimated in the above manner may be questioned.

[0044] Traditional internal state estimation methods have poor adaptability. Batteries have different characteristics over time. This is called the aging effect. The correlation between measurable variables such as the voltage (V), current (I) and temperature (T) of the battery and the internal state will change over time due to the aging effect. However, when a table summarizing the correlation between the internal state and the measurable variables is completed in advance by experiment, it is difficult to reflect the aging effect. In order to reflect the aging effect, many experiments should be performed on the aging effect in advance, or the aging effect should be predicted in advance and reflected in the experimental model. However, the problem with the former is that the cost increases, and the problem with the latter is that the reliability of the state estimation is poor when the aging effect is not accurately predicted in advance. Not only time but also design changes of battery cells will lead to changes in battery cell characteristics. However, when these changes are not predicted in advance and reflected in the experimental model, if design changes of battery cells such as the active materials used in the battery cells or the shape of the battery cells occur in the future, the existing experimental model cannot be used.

[0045] The internal state of the battery can also be estimated by using an electrochemical model (e.g., the Newman model) that takes into account the electrochemical phenomena and thermal behavior in the battery. The use of an electrochemical model has the following advantages: physical states that are difficult to measure directly can be used as control parameters of the BMS. However, because the disadvantage of the electrochemical model is that it requires excessive development time and cost and the resources required for calculation are too many, it is difficult to use the electrochemical model in the BMS.

[0046] In order to solve such problems of the conventional method, the present disclosure provides a method for distinguishing a normal battery cell from an internal short-circuited battery cell by using a G parameter and an H parameter as parameters indicating the current state of the battery cell. The internal short-circuited battery cell detection method according to the present disclosure can be relatively simple to perform, sufficient for use in a BMS, and can have high accuracy even without additional operating conditions.

[0047] The G parameter is a state quantity indicating the sensitivity of the terminal voltage of a battery cell to changes in the current applied to the battery cell in use, and has the unit of resistance. The H parameter is an effective potential determined by the local equilibrium potential distribution and the resistance distribution in the battery cell in use. The G parameter and H parameter of a battery cell can be quantified by using a theoretical model through an explicit correlation between the battery material properties and the design variables. The G parameter and H parameter of a battery cell will now be described.

[0048] In a battery cell, it may be assumed that a voltage V and a current i have a relationship such as V=f(i;x,p). Here, x is a physical quantity indicating an internal state of the battery cell, and p is a parameter.

[0049] When the function f as a nonlinear implicit function can be divided into a fast-changing amount g and a slow-changing amount h, the above relationship equation can be expressed as V=g(i,x,p)+h(i,x,p).

[0050] When it is assumed that there is a function G(i; x, p)=dg / di that changes slowly relative to the current i, the above relationship equation can be expressed as V=G(i; x, p)i+H(i; x, p).

[0051] In the above relationship equation, dG / di and dH / di have very small values. In other words, when the above assumptions are met, G and H are functions that change slowly relative to the current i, so the function f indicating the nonlinear relationship between the voltage V and the current i can be expressed as a quasi-linear relationship.

[0052] Here, G is referred to as a G parameter and H is referred to as an H parameter. When current i is a charge / discharge current and Ueq is an equilibrium potential of a battery cell, by using the G parameter G and the H parameter H, the discharge overvoltage can be expressed as Ueq-V=-G·i+(Ueq-H).

[0053] Here, -G·i is the overvoltage generated by the battery cell to make the current flow through the terminal, and includes the reaction dynamic polarization amount and the electronic and ionic resistance polarization amount. (Ueq-H) is the overvoltage generated when the local thermodynamic equilibrium state of the battery cell deviates from the overall system equilibrium state. That is, (Ueq-H) represents the inefficiency caused by the thermodynamic inhomogeneity inside the battery cell, and when the internal system of the battery cell reaches the thermodynamic equilibrium state, the H parameter H is equal to the equilibrium potential Ueq.

[0054] An internal short-circuited battery cell detection method according to an embodiment of the present disclosure involves: extracting G parameters G and H parameters H of each battery cell directly from voltage values ​​and current values ​​measured in each battery cell by using, for example, an RLS method, and detecting an internal short-circuited battery cell by using the G parameters G and H parameters H of each battery cell.

[0055] Figure 2 is a flow chart illustrating an internal short-circuited battery cell detection method according to an embodiment.

[0056] Reference Figure 1 and Figure 2 , the microprocessor 140 can perform the Figure 2 Internal short circuit battery cell detection method.

[0057] The microprocessor 140 can detect the voltage and current of each of the plurality of battery cells 111 that are electrically connected to each other between the first terminal 101 and the second terminal 102 and are in use by using the voltage measuring unit 120 and the current measuring unit 130. The microprocessor 140 can periodically generate a voltage value V and a current value I of each battery cell 111 (S110). A voltage value V_j and a current value I_j of the j-th battery cell (e.g., Cell_j) can be generated. The j-th battery cell (e.g., Cell_j) refers to any one of the battery cells 111 included in the battery 110, and j is equal to or greater than 1, and can be equal to or less than the number of battery cells 111 included in the battery 110.

[0058] The microprocessor 140 can update the G parameter value G_j(t) and the H parameter value H_j(t) in real time according to the voltage value V_j and the current value I_j of the corresponding battery cell (e.g., Cell_j) by using the adaptive filter stored in the memory 150. The G parameter value G_j(t) and the H parameter value H_j(t) are obtained by digitizing the G parameter and the H parameter indicating the current state of the corresponding battery cell (e.g., Cell_j) (S120). The operation S120 in which the G parameter value G_j(t) and the H parameter value H_j(t) are updated in real time will now be described in more detail.

[0059] The microprocessor 140 may calculate a G parameter representative value G_ref representing the G parameter value G(t) of the battery cell 111, and may calculate an H parameter representative value H_ref representing the H parameter value H(t) of the battery cell 111 (S130). According to an embodiment, the G parameter representative value G_ref may be an average value of the G parameter values ​​G(t), and the H parameter representative value H_ref may be an average value of the H parameter values ​​H(t). According to another embodiment, the G parameter representative value G_ref may be a median value of the G parameter values ​​G(t), and the H parameter representative value H_ref may be a median value of the H parameter values ​​H(t).

[0060] The microprocessor 140 may determine whether the corresponding battery cell (e.g., Cell_j) is an internal short-circuited battery cell based on the G parameter value G_j(t) and the H parameter value H_j(t) of the corresponding battery cell (e.g., Cell_j), the G parameter representative value G_ref, and the H parameter representative value H_ref. For example, for the corresponding battery cell (e.g., Cell_j), the microprocessor 140 may calculate the G parameter deviation value G_ref-G_j(t) of the corresponding battery cell (e.g., Cell_j) by subtracting the G parameter value G_j(t) of the corresponding battery cell (e.g., Cell_j) from the G parameter representative value G_ref, and may calculate the H parameter deviation value H_ref-H_j(t) of the corresponding battery cell (e.g., Cell_j) by subtracting the H parameter value H_j(t) of the corresponding battery cell (e.g., Cell_j) from the H parameter representative value H_ref. The microprocessor 140 can determine whether the G parameter deviation value G_ref-G_j(t) of the corresponding battery cell (e.g., Cell_j) is greater than the preset G parameter setting value G_th, and can determine whether the H parameter deviation value H_ref-H_j(t) of the corresponding battery cell (e.g., Cell_j) is greater than the preset H parameter setting value H_th (S140).

[0061] When the G parameter deviation value G_ref-G_j(t) of the corresponding battery cell (e.g., Cell_j) is greater than the preset G parameter setting value G_th and the H parameter deviation value H_ref-H_j(t) of the corresponding battery cell (e.g., Cell_j) is greater than the preset H parameter setting value H_th, the microprocessor 140 may determine that the battery cell (e.g., Cell_j) is an internal short-circuited battery cell (S150). For example, the G parameter setting value G_th may be preset to about 0.1, and the H parameter setting value H_th may be preset to about 0.01. These values ​​are examples, and the G parameter setting value G_th and the H parameter setting value H_th may be set to other values ​​according to the characteristics of the battery cell 111.

[0062] When the G parameter value G_j(t) and the H parameter value H_j(t) of the battery cell (e.g., Cell_j) do not satisfy the condition of operation S140, the microprocessor 140 may determine whether another battery cell (e.g., Cell_k) satisfies the condition of operation S140 to determine whether the battery cell Cell_k is an internal short-circuited battery cell. In this way, the microprocessor 140 may determine whether an internal short circuit occurs in each of all the battery cells 111 in the battery 110.

[0063] When all battery cells 111 do not satisfy the conditions of operation S140, the microprocessor 140 may not proceed to operation S110. The microprocessor 140 may quickly and accurately determine whether an internal short circuit occurs in each battery cell 111 by repeatedly performing operations S110 to S140 after a preset period of time Δt.

[0064] According to an embodiment, the microprocessor 140 may perform the RLS filter stored in the memory 150 by using Figure 2 operation.

[0065] Figure 3 It is shown Figure 2 Flowchart of operations S110 to S120 of the internal short-circuited battery cell detection method.

[0066] Reference Figures 1 to 3 , when the microprocessor 140 uses the RLS filter, the microprocessor 140 may initialize the state vector Θ=[G, H] and the covariance matrix P of each battery cell 111 (S101). According to the present embodiment, there may be a state vector Θ_j=[G_j, H_j] and a covariance matrix P_j corresponding to the j-th battery cell Cell_j. In operation S101, the microprocessor 140 may initialize the state vector Θ_j=[G_j, H_j] and the covariance matrix P_j of the j-th battery cell (e.g., Cell_j).

[0067] Figure 2 Operation S110 may include Figure 3Operations S111, S112, and S113. In operation S111, the microprocessor 140 may generate a current voltage value V(t) and a current current value I(t) of each battery cell 111. The current voltage value V(t) and the current current value I(t) of each battery cell 111 generated before the time period Δt are the immediately preceding voltage value V(t-1) and the immediately preceding current value I(t-1) of each battery cell 111. For example, the microprocessor 140 generates the immediately preceding voltage value V_j(t-1) and the immediately preceding current value I_j(t-1) of the j-th battery cell (e.g., Cell_j), and generates the current voltage value V_j(t) and the current current value I_j(t) of the j-th battery cell (e.g., Cell_j) after the time period Δt.

[0068] In operation S112, the microprocessor 140 may calculate a current difference ΔI between a current current value I(t) and an immediately preceding current value I(t-1) of each battery cell 111. For example, the microprocessor 140 may calculate a current difference ΔI_j between a current current value I_j(t) and an immediately preceding current value I_j(t-1) of a j-th battery cell (e.g., Cell_j).

[0069] In operation S113, the microprocessor 140 may compare the current difference ΔI of each battery cell 111 with a preset range. The preset range may be defined by using a lower limit value ΔImin and an upper limit value ΔImax. For example, the microprocessor 140 may determine whether the current difference ΔI_j of the jth battery cell (e.g., Cell_j) is between the lower limit value ΔImin and the upper limit value ΔImax.

[0070] When the current difference ΔI_j of the j-th battery cell (e.g., Cell_j) is between the lower limit value ΔImin and the upper limit value ΔImax, in operation S121, the state vector θ_j and the covariance matrix P_j of the j-th battery cell (e.g., Cell_j) may be updated. The state vector θ_j of the j-th battery cell (e.g., Cell_j) includes the G parameter value G_j(t) and the H parameter value H_j(t) of the j-th battery cell (e.g., Cell_j), and when the state vector θ_j is updated in operation S121, the G parameter value G_j(t) and the H parameter value H_j(t) are also updated (S120).

[0071] Next, the microprocessor 140 may execute Figure 2 Operation S130.

[0072] According to this embodiment, the state vector and covariance matrix P(t). A method of updating the G parameter value G_j(t) and the H parameter value H_j(t) of any one of the battery cells 111 (e.g., Cell_j) in real time will now be described. Since the G parameter value G_k(t) and the H parameter value H_k(t) are calculated in the same manner for another battery cell (e.g., Cell_k), the number (e.g., j or k) used to identify each battery cell 111 is omitted.

[0073] State Vector Including G parameters and H parameter and can be defined as

[0074] The covariance matrix P(t) includes the first value P1(t) and the second value P2(t) and can be defined as

[0075] The voltage value V(t) and the current value I(t) of the battery cell 111 are generated in each time period Δt, and the state vector is also updated in each time period Δt by using a recursive method. And the covariance matrix P(t). The G parameters are also updated in each time period Δt and H parameter

[0076] In operation S101, the microprocessor 140 may make the state vector and the covariance matrix P(t). For example, the state vector Can be initialized as And the covariance matrix P(t) can be initialized as In this example, although the state vector The elements of and the elements of the covariance matrix P(t) are all initialized to 1, but this is an example and the elements may be initialized to other values.

[0077] The microprocessor 140 can measure the voltage and current of each battery cell 111 by using the voltage measuring unit 120 and the current measuring unit 130, and can periodically generate the voltage value and current value of each battery cell 111, for example, in each time period Δt. The microprocessor 140 can generate the immediately preceding voltage value V(t-1) and the immediately preceding current value I(t-1) of each battery cell 111, and can generate the current voltage value V(t) and the current current value I(t) of each battery cell 111 after the time period Δt.

[0078] The microprocessor 140 can update the G parameter in each time period Δt by using the current voltage value V(t) and the current current value I(t) generated in each time period Δt using the RLS method. and H parameter

[0079] The microprocessor 140 can calculate the current value I(t) based on the current current value I(t) and the previous value of the state vector To calculate the current voltage estimate of the battery cell 111 The immediate previous value of the state vector Calculated based on the immediately preceding voltage value V(t-1) and the immediately preceding current value I(t-1) before the time period Δt, and including the immediately preceding value of the G parameter and the immediately preceding value of the H parameter Current voltage estimation of battery cell 111 This can be done by changing the previous value of the H parameter The current value I(t) and the previous value of the G parameter The product of is added and calculated as

[0080] The microprocessor 140 can update the gain matrix L(t) based on the current value I(t) and the immediately preceding value P(t-1) of the covariance matrix. The immediately preceding value P(t-1) of the covariance matrix is ​​calculated based on the immediately preceding voltage value V(t-1) and the immediately preceding current value I(t-1) before the time period Δt, and according to the definition of the covariance matrix P(t), includes the first immediately preceding value P1(t-1) of the covariance matrix and the second immediately preceding value P2(t-1) of the covariance matrix. Corresponding to each battery cell 111, there is a gain matrix L(t).

[0081] When updating the state vector When the covariance matrix P(t) is calculated, the gain matrix L(t) is used. The gain matrix L(t) includes a first value L1(t) of the gain matrix and a second value L2(t) of the gain matrix, and can be calculated as follows.

[0082]

[0083] Here, λ1 is the first forgetting factor and is related to the G parameter. λ2 is the second forgetting factor and is related to the H parameter. The first forgetting factor λ1 and the second forgetting factor λ2 are the factors indicating the calculation of the G parameter. and H parameter The past voltage and current values ​​of the G parameter are and the current value of the H parameter The closer the first forgetting factor λ1 and the second forgetting factor λ2 are to 1, the more they will affect the current value of the G parameter in the long term. and the current value of the H parameter And the closer the first forgetting factor λ1 and the second forgetting factor λ2 are to 0, the more they will affect the current value of the G parameter in a short period of time. and the current value of the H parameter

[0084] According to an embodiment, the first forgetting factor λ1 and the second forgetting factor λ2 may be equal to or greater than 0.9 and equal to or less than 1. According to another embodiment, the first forgetting factor λ1 may be set to be equal to or greater than the second forgetting factor λ2. For example, the first forgetting factor λ1 may be set to 0.99999, and the second forgetting factor λ2 may be set to 0.95. These setting values ​​may vary according to the characteristics of the battery cell 111.

[0085] The inventors of the present disclosure have found in experiments performed on specific battery cells that high reliability results were obtained when the first forgetting factor λ1 and the second forgetting factor λ2 were 0.99999 and 0.95, respectively. However, the above values ​​are examples, and other values ​​may be set according to the characteristics of the battery cell 111. For example, the first forgetting factor λ1 may be set to 0.9999, and the second forgetting factor λ2 may be set to 0.98.

[0086] In another example, both the first forgetting factor λ1 and the second forgetting factor λ2 may be set to 1. In this case, it may be considered that the first forgetting factor λ1 and the second forgetting factor λ2 are not applied.

[0087] The microprocessor 140 may update the covariance matrix P(t) based on the current current value I(t), the gain matrix L(t), and the immediately previous value P(t-1) of the covariance matrix. The covariance matrix P(t) may be calculated as follows.

[0088]

[0089] The microprocessor 140 may be configured to To calculate the current voltage value V(t) and the current voltage estimate calculated earlier The voltage error e(t) between them.

[0090] The microprocessor 140 may be based on the immediately previous value of the state vector The current value of the gain matrix L(t) and the voltage error e(t) are used to update the state vector Since the state vector is updated So the current value of the G parameter can be generated and the current value of the H parameter

[0091] The current value of the state vector This can be done by multiplying the current value of the gain matrix L(t) and the voltage error e(t) by the immediately previous value of the state vector Add to calculate.

[0092]

[0093] The state vector is recursively expressed as The above equation can be derived as follows.

[0094] First, the loss function ε to which the first forgetting factor λ1 and the second forgetting factor λ2 are applied is defined as follows.

[0095]

[0096] Here, V(i) is the i-th voltage value, I(i) is the i-th current value. V(t) and I(t) are the current voltage value and the current current value, respectively, and V(t-1) and I(t-1) are the immediately preceding voltage value and the immediately preceding current value, respectively.

[0097] G(i) and H(i) are the real values ​​of the i-th G parameter and the i-th H parameter, respectively, and and They are the current value estimates of the G parameter and the current value estimates of the H parameter respectively.

[0098] When the loss function ε is used and When the result of differentiation is 0, the loss function ε is and is minimized.

[0099] The loss function ε is used to The result of taking the derivative is 0 It can be obtained as follows.

[0100]

[0101]

[0102] When summarizing the above equation, as follows.

[0103]

[0104] The loss function ε is used to The result of taking the derivative is 0 It can be obtained as follows.

[0105]

[0106]

[0107] When summarizing the above equation, as follows.

[0108]

[0109] For real-time estimation, the above obtained and By using the state vector This is summarized in recursive form as follows.

[0110]

[0111] As mentioned above, because the current voltage is estimated is defined as And the voltage error e(t) is defined as So the state vector can be expressed as follows.

[0112]

[0113] Here, as described above, the gain matrix L(t) and the covariance matrix P(t) can be calculated as follows.

[0114]

[0115]

[0116] Because the internal short-circuit battery cell detection method of this embodiment uses a recursive method, the memory 150 may only store the current voltage value V(t), the current current value I(t), the state vector and the covariance matrix P(t). According to another example, the first forgetting factor λ1 and the second forgetting factor λ2 may also be stored in the memory 150. In addition, the calculation is very simple and is possible even with the memory 150 having a small size of several kB. Because the state vector The covariance matrix P(t) is updated periodically, so the voltage and current fluctuations of each battery cell 111 can be reflected in the G parameter in real time. and H parameter middle.

[0117] The microprocessor 140 may repeatedly perform operations S120 to S150 in each time period Δt. When an internal short-circuited battery cell occurs among the battery cells 111, the microprocessor 140 may detect the internal short-circuited battery cell in real time.

[0118] according to Figure 3 Example, G parameter and H parameter It may not be updated in every time period Δt.

[0119] As described above, the microprocessor 140 may compare the current difference ΔI with a specific range defined by a lower limit value ΔImin and an upper limit value ΔImax to determine whether the current difference ΔI is within the specific range (S113). The lower limit value ΔImin and the upper limit value ΔImax may be set according to the charge / discharge current mode. The lower limit value ΔImin and the upper limit value ΔImax may be set based on the charge / discharge rate of the battery system 100. For example, the lower limit value ΔImin may be set to 0.2C, and the upper limit value ΔImax may be set to 1C.

[0120] When the current difference ΔI is too small or too large, the sensing error between the voltage value and the current value causes the G parameter to be and H parameter When the current difference ΔI is within a specific range, the method may proceed to operation S121, and the state vector may be updated based on the current voltage value V(t) and the current current value I(t). and covariance matrix P(t). However, when the current difference ΔI exceeds a certain range, the method may return to operation S111 and the state vector may not be updated. That is, when the current difference ΔI exceeds a certain range, the current voltage value V(t) and the current current value I(t) will not affect the state vector and the covariance matrix P(t).

[0121] By updating the G parameter only when the current difference ΔI is within a certain range and H parameter G parameter and H parameter The reliability can be kept high.

[0122] Now refer to Figure 4 and Figure 5 Describe the results obtained by measuring or estimating parameters of normal cells and internally shorted cells.

[0123] Figure 4 Graphs a to d are shown, which respectively show a battery pack current and a battery pack voltage measured in a battery pack including normal battery cells, and a G parameter value and an H parameter value estimated for the battery pack.

[0124] Figure 4 Graphs a to d of FIG. 1 use current / voltage actual operation mode data of a 20Ah energy storage system (ESS). Four battery cells are connected in series, and the time period Δt is 1 second. The current flowing through a normal battery cell is the same as the battery pack current. Figure 4 Graphs a to d show the variation of a specific parameter value as a function of time (seconds).

[0125] Figure 5 Shown in Figure 4 The measurement or estimate obtained after replacing a normal battery cell with an internally shorted battery cell in a battery pack.

[0126] Graph e shows the cell voltage of each of the normal cell and the internal short-circuited cell, and graph f shows the state of charge (SOC) of the cell. Graph g shows the G parameter value of each of the normal cell and the internal short-circuited cell, and graph h shows the H parameter value of each of the normal cell and the internal short-circuited cell.

[0127] As shown in graph g, the G parameter value of the internal short-circuited battery cell is less than the G parameter value of the normal battery cell. In addition, as shown in graph h, the H parameter value of the internal short-circuited battery cell is less than the H parameter value of the normal battery cell. That is, both the G parameter value and the H parameter value of the internal short-circuited battery cell are estimated to be less than the G parameter value and the H parameter value of the normal battery cell. Therefore, the method of the present embodiment for determining whether a battery cell is an internal short-circuited battery cell based on the G parameter value and the H parameter value can be reliable.

[0128] The spirit of the present disclosure is not limited to the above-described embodiments, and all scopes equivalent to the claims or changed equivalently thereto, except for the described claims, belong to the scope of the spirit of the present disclosure.

Claims

1. A method for detecting an internal short-circuited battery cell, the method comprising: detecting a voltage and a current of each of a plurality of battery cells that are electrically connected to each other between a first terminal and a second terminal and are in use, to periodically generate a voltage value and a current value of each of the plurality of battery cells; updating a G parameter value and an H parameter value in real time according to the voltage value and the current value of each of the plurality of battery cells by using an adaptive filter, the G parameter value and the H parameter value being obtained by digitizing a G parameter and an H parameter indicating a current state of each of the plurality of battery cells; calculating a G parameter representative value representing the G parameter values ​​of the plurality of battery cells, and calculating an H parameter representative value representing the H parameter values ​​of the plurality of battery cells; as well as determining whether an internal short circuit occurs in each of the plurality of battery cells based on the G parameter value and the H parameter value of each of the plurality of battery cells, the G parameter representative value, and the H parameter representative value, wherein the G parameter is a parameter indicating the sensitivity of the voltage of each of the plurality of battery cells to the current change, and the H parameter is a parameter indicating the effective potential determined by the local equilibrium potential distribution and the resistance distribution in each of the plurality of battery cells, The step of periodically generating the voltage value and the current value of each of the plurality of battery cells comprises: selecting a first battery cell as one of the plurality of battery cells; generating an immediately preceding voltage value and an immediately preceding current value of the first battery cell; and generating a current voltage value and a current current value of the first battery cell after a preset period of time, and The method further comprises the following steps: Initializing a plurality of state vectors and a plurality of covariance matrices corresponding to the plurality of battery cells, respectively, each of the plurality of state vectors including the G parameter value and the H parameter value of a corresponding battery cell among the plurality of battery cells; calculating a current difference of the first battery cell between the present current value of the first battery cell and the immediately previous current value of the first battery cell; and comparing the current difference of the first battery cell with a preset range, The G parameter value and the H parameter value of the first battery cell are updated by using the current voltage value and the current current value of the first battery cell only when the current difference of the first battery cell is within the preset range.

2. The internal short-circuit battery cell detection method according to claim 1, wherein: The step of determining whether an internal short circuit occurs in each of the plurality of battery cells includes: selecting a first battery cell as one of the plurality of battery cells; calculating a G parameter deviation value of the first battery cell by subtracting the G parameter value of the first battery cell from the G parameter representative value; calculating an H parameter deviation value of the first battery cell by subtracting the H parameter value of the first battery cell from the H parameter representative value; and When the G parameter deviation value of the first battery cell is greater than a preset G parameter setting value and the H parameter deviation value of the first battery cell is greater than a preset H parameter setting value, it is determined that the first battery cell is an internal short-circuited battery cell.

3. The internal short-circuit battery cell detection method according to claim 1, wherein: The G parameter representative value is an average value of the G parameter values ​​of the plurality of battery cells, and the H parameter representative value is an average value of the H parameter values ​​of the plurality of battery cells.

4. The internal short-circuit battery cell detection method according to claim 1, wherein: The G parameter representative value is a median value of the G parameter values ​​of the plurality of battery cells, and the H parameter representative value is a median value of the H parameter values ​​of the plurality of battery cells.

5. The internal short-circuit battery cell detection method according to claim 1, wherein: The adaptive filter is a filter using a recursive least squares (RLS) method.

6. The internal short-circuit battery cell detection method according to claim 1, wherein: The step of updating the G parameter value and the H parameter value of each of the plurality of battery cells comprises: calculating a current voltage estimate for the first cell based on the current value of the first cell and an immediately previous value of a state vector corresponding to the first cell; Based on the current current value of the first battery cell and the immediately previous value of the covariance matrix corresponding to the first battery cell, updating the gain matrix corresponding to the first battery cell and the covariance matrix corresponding to the first battery cell; calculating a voltage error of the first battery cell between the present voltage value of the first battery cell and the present voltage estimate of the first battery cell; and The state vector corresponding to the first battery cell is updated based on the immediately previous value of the state vector corresponding to the first battery cell, the current value of the gain matrix corresponding to the first battery cell, and the voltage error of the first battery cell to generate the current value of the G parameter of the first battery cell and the current value of the H parameter of the first battery cell.

7. The internal short-circuit battery cell detection method according to claim 6, wherein: The present voltage estimate of the first cell is calculated by adding an immediately previous value of the H parameter of the first cell to a product of the present current value of the first cell and an immediately previous value of the G parameter value of the first cell.

8. The internal short-circuit battery cell detection method according to claim 6, wherein: A current value of the state vector corresponding to the first battery cell is calculated by adding a product of the current value of the gain matrix corresponding to the first battery cell and the voltage error of the first battery cell to an immediately previous value of the state vector corresponding to the first battery cell.

9. The internal short-circuit battery cell detection method according to claim 6, wherein: The step of updating the gain matrix corresponding to the first battery cell and the covariance matrix corresponding to the first battery cell includes applying a first forgetting factor associated with the G parameter of the first battery cell and a second forgetting factor associated with the H parameter of the first battery cell.

10. The internal short-circuit battery cell detection method according to claim 9, wherein: The gain matrix corresponding to the first battery cell is calculated according to the following formula: and The covariance matrix corresponding to the first battery cell is calculated according to the following formula: Wherein, L(t) is the current value of the gain matrix corresponding to the first battery cell, L(t-1) is the immediately previous value of the gain matrix corresponding to the first battery cell, P(t) is the current value of the covariance matrix corresponding to the first battery cell, P(t-1) is the immediately previous value of the covariance matrix corresponding to the first battery cell, I(t) is the current current value of the first battery cell, λ1 is the first forgetting factor, and λ2 is the second forgetting factor.

Images