Secondary battery state estimation system, secondary battery state estimation method and storage medium

Abstract

A state estimation system, a state estimation method, and a storage medium for a secondary battery are provided, capable of obtaining SOC-OCV characteristics corresponding to a wide variety of secondary batteries. The state estimation system for the secondary battery is configured as follows: a state variable measurement unit that measures state variables, including output current and output voltage, of the operating secondary battery at predetermined intervals; a state variable processing unit that outputs state variable processing data, including charge / discharge quantity, voltage change, voltage, and electrical power, calculated based on the state variables measured by the state variable measurement unit; and a characteristic identification unit that uses the state variable processing data to identify the SOC-OCV characteristics of the secondary battery using a learned characteristic identification model.

Description

Technical Field

[0001] This invention relates to a state estimation system, a state estimation method, and a storage medium. Background Technology

[0002] It is known that there is a technique for learning a model based on time series data involving the State of Charge (SOC) of a battery from a first time point to a second time point later than the first time point and the SOH of the battery at the first time point as input data and the SOH at the second time point as output data (for example, see Japanese Patent Application Laid-Open No. 2019-168453 (hereinafter Patent Document 1)). Summary of the Invention

[0003] In the technology described in Patent Document 1, the SOC-OCV characteristic corresponding to the secondary battery for which SOH is estimated is set. However, the set SOC-OCV characteristic becomes an inherent characteristic in a specific secondary battery. Therefore, in the technology described in Patent Document 1, it is difficult to estimate SOH corresponding to a wide variety of secondary batteries. Thus, there are situations where it is required to obtain SOC-OCV characteristics corresponding to a wide variety of secondary batteries.

[0004] The present invention was made in consideration of such circumstances, and one of its objectives is to provide a state estimation system, state estimation method and storage medium capable of obtaining SOC-OCV characteristics corresponding to a wide variety of secondary batteries.

[0005] To address the aforementioned issues and achieve this objective, the present invention employs the following solution.

[0006] (1): A state estimation system for a secondary battery according to one aspect of the present invention comprises: a state variable measuring unit that measures state variables including output current and output voltage of a secondary battery in operation at each predetermined time; a state variable processing unit that outputs state variable processing data including charge / discharge amount and voltage change amount calculated based on the state variables measured by the state variable measuring unit; and a characteristic identification unit that uses the state variable processing data to identify the SOC-OCV characteristics of the secondary battery through a learned characteristic identification model.

[0007] (2): Based on the state estimation system of the above (1) scheme, the state variable processing unit may set a unit charge / discharge quantity or unit time, and set the nearest interval closest to the current time and one or more past intervals earlier than the nearest interval as the expected interval obtained based on the unit charge / discharge quantity or unit time used in the calculation of the state variable processing data. The state variable processing unit calculates the voltage change obtained based on the state variable as interval data for each of the set intervals, and outputs the calculated interval data in the state variable processing data.

[0008] (3): Based on the state estimation system of the above (2) scheme, the state variable processing unit may set a large charge / discharge quantity as a unit charge / discharge quantity and a small charge / discharge quantity as a unit charge / discharge quantity that is smaller than the large charge / discharge quantity. It may also set a large interval based on the large charge / discharge quantity and a small interval based on the small charge / discharge quantity as the expected interval derived from the large charge / discharge quantity and the small charge / discharge quantity used in the calculation of the state variable processing data.

[0009] (4): Based on the state estimation system of the above (2) scheme, the state variable processing unit may set a large interval based on the unit time and a small interval based on a time shorter than the large interval as the expected interval derived from the unit time used in the calculation of the state variable processing data.

[0010] (5): Based on the state estimation system of the above (2) scheme, the state variable processing unit may set the period during which the specified charge / discharge quantity is calculated by the cumulative calculation of the output current measured by the state variable measuring unit or the period during which the specified value is obtained as the desired interval.

[0011] (6): Based on the state estimation system of the above scheme (2) or (3), it is also possible that at least one of the calculated charge / discharge amount of each desired interval and the calculated voltage change amount of each desired interval is a slope change rate calculated using the least squares method.

[0012] (7): Based on the state estimation system of any of the above schemes (1) to (4), the feature recognition model may also be configured as an RNN (recurrent neural network).

[0013] (8): Based on the state estimation system of the above (5) scheme, the intermediate layer of the RNN can also be configured as LSTM (Long Short-Term Memory) or GRU (Gated Recurrent Unit).

[0014] (9): Based on the state estimation system of any of the above schemes (1) to (4), the feature recognition model may also be configured as a CNN (convolutional neural network).

[0015] (10): A state estimation system for a secondary battery according to one aspect of the present invention comprises: a state variable measuring unit that measures state variables including output current and output voltage of a secondary battery in operation at predetermined times; a state variable processing unit that outputs state variable processing data including charge / discharge quantity and voltage change quantity calculated based on the state variables measured by the state variable measuring unit; a characteristic identification unit that uses the state variable processing data to identify the SOC-OCV characteristics of the secondary battery through a learned characteristic identification model and outputs characteristic identification information indicating the identification result; and a degradation state estimation unit that uses input data including the charge / discharge quantity, the voltage change quantity and the characteristic identification information to estimate the degradation state of the secondary battery in operation through a learned degradation state model.

[0016] (11): A state estimation method of the present invention, wherein a computer in the state estimation system performs the following processing: measuring the state variables of the operating secondary battery, including the output current and the output voltage, at each specified time; outputting state variable processing data, including the charge / discharge amount and voltage change, calculated based on the measured state variables; and using the state variable processing data to identify the SOC-OCV characteristics of the secondary battery through a learned characteristic recognition model.

[0017] (12): A state estimation method according to an embodiment of the present invention, wherein a computer in the state estimation system performs the following processing: measuring state variables including output current and output voltage of a secondary battery in operation at each predetermined time; outputting state variable processing data including charge / discharge quantity and voltage change quantity calculated based on the measured state variables; using the state variable processing data to identify the SOC-OCV characteristics of the secondary battery through a learned characteristic identification model and outputting characteristic identification information representing the identification result; using input data including the charge / discharge quantity, the voltage change quantity and the characteristic identification information to estimate the degradation state of the secondary battery in operation through a learned degradation state model.

[0018] (13): A storage medium of one aspect of the present invention stores a program, wherein the program is used to cause a computer in a state estimation system to perform the following processing: measuring state variables including output current and output voltage of a secondary battery in operation at each predetermined time; outputting state variable processing data including charge / discharge amount and voltage change amount calculated based on the measured state variables; and using the state variable processing data to identify the SOC-OCV characteristics of the secondary battery through a learned characteristic recognition model.

[0019] (14): A storage medium of one aspect of the present invention stores a program, wherein the program is configured to cause a computer in a state estimation system to perform the following processing: measuring state variables including output current and output voltage of a secondary battery in operation at each predetermined time; outputting state variable processing data including charge / discharge quantity and voltage change quantity calculated based on the measured state variables; using the state variable processing data to identify the SOC-OCV characteristics of the secondary battery through a learned characteristic identification model and outputting characteristic identification information representing the identification result; and using input data including the charge / discharge quantity, the voltage change quantity and the characteristic identification information to estimate the degradation state of the secondary battery in operation through a learned degradation state model.

[0020] Based on (1), (11), and (13), the SOC-OCV characteristics of a secondary battery are identified by using a learned characteristic identification model that processes state variable data, including charge / discharge amounts and voltage changes, calculated for the operating secondary battery at each specified time. Thus, SOC-OCV characteristics can be obtained corresponding to a wide variety of secondary batteries.

[0021] According to (2), when identifying the SOC-OCV characteristics of a secondary battery using a learned characteristic identification model, interval data based on the charge / discharge amount and voltage change calculated corresponding to multiple desired intervals are used. Therefore, it is possible to improve the accuracy of the identification of SOC-OCV characteristics based on the learned characteristic identification model.

[0022] According to (3), it is possible to: set a large interval based on a large amount of charge and discharge and a small interval based on a small amount of charge and discharge, and use interval data corresponding to the large interval and the small interval respectively. Thus, it is possible to further improve the accuracy of the recognition of SOC-OCV characteristics based on the learned feature recognition model.

[0023] According to (4), it is possible to: set a large number of intervals based on a unit time and a small number of intervals based on a unit time shorter than the above-mentioned unit time as the desired intervals, and use interval data corresponding to the large number of intervals and the small number of intervals respectively. Thus, it is possible to further improve the accuracy of the recognition of SOC-OCV features based on the learned feature recognition model.

[0024] According to (5), the desired range can be set based on the cumulative current value.

[0025] According to (6), the charge / discharge quantity and voltage change quantity as interval data can be made into charge / discharge quantity and voltage change quantity with high accuracy and reduced noise.

[0026] According to (7), by using RNNs with respect to feature recognition models, we can expect high-precision inference results.

[0027] According to (8), by setting the intermediate layer in the RNN of the feature recognition model to LSTM, we can expect high-precision inference results.

[0028] According to (9), by using CNN with respect to the feature recognition model, we can expect inference results with high accuracy.

[0029] According to (10), (12), and (14), the learned degradation state model uses input data, including characteristic identification information representing the SOC-OCV characteristics identified by the learned characteristic identification model, in addition to the charge / discharge amount and voltage variation, to estimate the degradation state of the secondary battery in operation. This improves the accuracy of the degradation state estimation for the secondary battery. Attached Figure Description

[0030] Figure 1 This is a diagram showing an example of the configuration of the degradation state estimation device according to this embodiment.

[0031] Figure 2 This is a diagram illustrating an example of a scheme for obtaining a large amount of interval data in this embodiment.

[0032] Figure 3 This is a diagram illustrating the SOC-OCV characteristics of this embodiment.

[0033] Figure 4 This diagram illustrates the identification of SOH corresponding to the relationship between the charge / discharge amount and the voltage change amount in this embodiment.

[0034] Figure 5 This diagram illustrates the identification of SOH corresponding to the relationship between the charge / discharge amount and the voltage change amount in this embodiment.

[0035] Figure 6 This diagram illustrates the identification of SOH corresponding to the relationship between the charge / discharge amount and the voltage change amount in this embodiment.

[0036] Figure 7 This is a flowchart illustrating an example of the processing steps performed by the degradation state estimation device in this embodiment in association with the estimation of SOH. Detailed Implementation

[0037] Hereinafter, with reference to the accompanying drawings, embodiments of the state estimation system, state estimation method, and state estimation storage medium of the present invention will be described.

[0038] Figure 1This diagram illustrates an overall configuration example of the degradation state estimation device 100 according to this embodiment. The degradation state estimation device 100 estimates the state of harmonics (SOH) of the secondary battery 200 as the degradation state of the secondary battery 200. The degradation state estimation device 100 shown in this diagram includes a state variable measurement unit 101, a preprocessing unit 102, a second learned model 103 (an example of a degradation state model), and a degradation state estimation unit 104.

[0039] The state variable measurement unit 101 measures the output current and output voltage as state variables of the operating secondary battery 200, and outputs the measured output current Iout and output voltage Vout. The output voltage Vout can also be calculated based on the CCV (Closed Circuit Voltage) detected by the sensor in the secondary battery 200. The output voltage Vout can also be calculated based on OCV (Open Circuit Voltage).

[0040] The preprocessing unit 102 calculates state variable processing data corresponding to the current time t using the output current Iout and output voltage Vout input from the state variable measurement unit 101, and uses this data as preprocessing. The state variable data is used by the characteristic recognition unit 123 in estimating the SOC-OCV characteristics. All data in the state variable processing data except for the electrical data P(t) is output as input data Din(t) corresponding to the current time t, and the degradation state estimation unit 104 uses the input data Din(t) in estimating the SOH.

[0041] The preprocessing unit 102 includes a state variable processing unit 121, a first learned model 122 (an example of a feature recognition model) and a feature recognition unit 123.

[0042] The state variable processing unit 121 receives the output current Iout and output voltage Vout from the state variable measurement unit 101. Based on the input output current Iout and output voltage Vout, the state variable processing unit 121 calculates a large amount of interval data (charge / discharge quantity LCA(t)~LCA(t-2), voltage change LEV(t~LEV(t-2)), a small amount of interval data (charge / discharge quantity SCA(t), voltage change LEV(t)), voltage data V(t), and power data P(t)) corresponding to the current time t, according to each predetermined estimation time.

[0043] The voltage data V(t) can be CCV. Alternatively, the voltage data V(t) can also be OCV.

[0044] The power data P(t) can be either discharging power (power consumption) or charging power. Alternatively, it can replace power data P(t) with information such as current quantity, auxiliary equipment usage, or vehicle usage patterns—information that indicates the extent of power consumption and charging. Such power data P(t) can either be a substitute for power data P(t), for example, it can be data estimated from the cloud, or it can be unique information where the characteristics remain unchanged.

[0045] The interval data includes large interval data and small interval data. Large interval data refers to the voltage change calculated by the state variable processing unit 121 for each corresponding large interval of charge / discharge, such as a unit large charge / discharge. Small interval data refers to the voltage change calculated by the state variable processing unit 121 for each corresponding small interval of charge / discharge, such as a unit small charge / discharge.

[0046] From now on, the "large charge / discharge range" will be abbreviated as "large range", and the "small charge / discharge range" will be abbreviated as "small range".

[0047] Reference Figure 2 The method for calculating a large amount of interval data is explained in the following example. The figure shows the unit charge / discharge quantity Aprd of the secondary battery 200 obtained corresponding to the period from the current time t to a time t-3 earlier than the current time t. The unit charge / discharge quantity Aprd shown in the figure is obtained by accumulating the output current Iout.

[0048] The state variable processing unit 121 sets the period from the current time t back to the time t-1 when the specified cumulative unit current value (the cumulative value of the output current Iout) is obtained as the nearest large interval (the nearest large interval T1). The state variable processing unit 121 sets the period from time t1 back to the time t-1 when the specified cumulative unit current value is obtained as the first large interval past the nearest large interval T1 (the past large interval T2-1). The state variable processing unit 121 sets the period from time t-1 back to the time t-2 when the specified cumulative unit current value is obtained as the past large interval T2-2, which is even further past than the past large interval T2-1.

[0049] In the following explanations, unless otherwise distinguished between past large intervals T2-1 and T2-2, they will be referred to as past large interval T2. Unless otherwise distinguished between recent large interval T1 and past large interval T2, they will be referred to as large interval T.

[0050] As described above, the large interval T is set as the interval for obtaining a unit large charge / discharge quantity based on the cumulative value of each unit current. Therefore, the length of each large interval T can also be different. The large interval T can also be set separately according to a specified unit time.

[0051] A large number of intervals T can be, for example, tens to hundreds of seconds in length.

[0052] The number of past large intervals T2 set by the state variable processing unit 121 is not limited to two; one or more are acceptable.

[0053] In the example of this diagram, three large intervals T are set to be continuous over time. This can be made discontinuous by setting intervals between two consecutive large intervals T. When setting more than three large intervals T, it can be set so that two consecutive large intervals T at certain times are continuous, while two consecutive large intervals T at other times are discontinuous.

[0054] The state variable processing unit 121 calculates the large interval T corresponding to obtaining a certain unit large charge / discharge capacity LCA set as described above. That is, as Figure 2 As shown, the state variable processing unit 121 calculates the actual time of the most recent large-scale interval T1 in a manner corresponding to the unit large-scale charge / discharge quantity LCA = charge / discharge quantity LCA(t), calculates the actual time of the past large-scale interval T2-1 in a manner corresponding to the charge / discharge quantity LCA(t-1) = charge / discharge quantity LCA(t), and calculates the actual time of the past large-scale interval T2-2 in a manner corresponding to the charge / discharge quantity LCA(t-2) = charge / discharge quantity LCA(t-1). In other words, the state variable processing unit 121 calculates the past times (t-1), (t-2), and (t-3) in a manner corresponding to the unit large-scale charge / discharge quantity LCA = LCA(t) = LCA(t-1) = LCA(t-2), and calculates the actual times T1, T2-1, and T2-2. Alternatively, the most recent large-volume interval can be fixed as T1, i.e., T1 = T2-1 = T2-2. The charge / discharge quantity LCA(t) is calculated corresponding to the most recent large-volume interval T1, the charge / discharge quantity LCA(t-1) is calculated corresponding to the past large-volume interval T2-1, and the charge / discharge quantity LCA(t-2) is calculated corresponding to the past large-volume interval T2-2. However, by setting the unit large-volume charge / discharge quantity LCA, the estimation accuracy of SOH (described later) is successfully improved.

[0055] The state variable processing unit 121 can calculate at least one of the charge / discharge quantities LCA(t), LCA(t-1), and LCA(t-2) as a slope change rate using the least squares method. The charge / discharge quantity LCA calculated as a slope change rate in this way improves accuracy by reducing noise. As a result, the accuracy of the SOC-OCV characteristics estimated using the first learned model 122 (described later) and the SOH estimated using the second learned model 103 (described later) using the charge / discharge quantity LCA can be improved.

[0056] State variable processing unit 121 as follows Figure 2 The large number of intervals T calculated as shown are as follows: Figure 3 As shown, the change in the corresponding voltage (output voltage Vout) is calculated. That is, the state variable processing unit 121 calculates the voltage change LEV(t) corresponding to the most recent large interval T1, calculates the voltage change LEV(t-1) corresponding to the past large interval T2-1, and calculates the voltage change LEV(t-2) corresponding to the past large interval T2-2.

[0057] The state variable processing unit 121 can also calculate the slope change rate using the least squares method for at least one of the voltage changes LEV(t), LEV(t-1), and LEV(t-2). In this case, the accuracy of the SOC-OCV characteristic estimated by the first learned model 122 (described later) and the SOH estimated by the second learned model 103 (described later) using the voltage change LEV can also be improved.

[0058] The state variable processing unit 121 can calculate the slope change rate using the least squares method for at least one of the voltage changes LEV(t), LEV(t-1), and LEV(t-2), either for a specified unit large charge / discharge quantity LCA or for a large charge / discharge quantity LCA(t) corresponding to a specified time interval. The voltage change LEV(t) calculated as a slope change rate in this way has improved accuracy by reducing noise. As a result, the accuracy of the SOC-OCV characteristics estimated by the first learned model 122 (described later) and the SOH estimated by the second learned model 103 using the voltage change LEV(t) calculated as a slope change rate can be improved.

[0059] In the following description, unless otherwise specified, the charge / discharge quantities LCA(t), LCA(t-1), and LCA(t-2) will be referred to as charge / discharge quantity LCA.

[0060] In the following description, unless otherwise specified, the voltage changes LEV(t), LEV(t-1), and LEV(t-2) will be referred to as voltage changes LEV.

[0061] In cases where there is no specific distinction between charge / discharge quantity (LCA) and voltage change (LEV), it is also recorded as "a large range of data".

[0062] The state variable processing unit 121 can also calculate, for example, the voltage change per unit of charge / discharge corresponding to the most recent large interval, as large interval data.

[0063] Return to description Figure 1 In addition to calculating the voltage change LEV corresponding to each large interval T, such as the charge / discharge quantity LCA, the state variable processing unit 121 also calculates the actual time and voltage change SEV(t) corresponding to a small interval, such as the charge / discharge quantity SCA(t). The state variable processing unit 121 can also use past values, such as those corresponding to times t-1 and t-2, for the charge / discharge quantity SCA and the voltage change SEV.

[0064] The state variable processing unit 121 can, for example, set a period from the current time t back to the time when a predetermined current accumulation smaller than that corresponding to a large interval is obtained as a small interval. Alternatively, the state variable processing unit 121 can set a period of a predetermined unit current accumulation (charge / discharge amount) smaller than that of a large interval T as a small interval. A small interval can be, for example, a few seconds to tens of seconds.

[0065] In cases where there is no particular distinction between the charge / discharge quantity SCA(t) and the voltage change quantity SEV(t), it is recorded as "a small amount of data in a short period".

[0066] As described above, the small-scale interval is defined as the period during which the accumulated current values ​​are obtained to become the specified unit small-scale charge / discharge quantity (SCA). Therefore, the length of each small-scale interval can also be different.

[0067] The number of past small intervals set by the state variable processing unit 121 is not limited to one; one or more are acceptable.

[0068] In the example shown in the figure, the power change SEV(t) for a single small interval is input. Similar to the large interval T mentioned above, two small intervals can be set, or two small intervals that are consecutive in time can be set, and the power change SEV(t) for each of the set small intervals can be input. The small intervals can be made discontinuous by setting intervals between them. When setting three or more small intervals, it can be configured such that two small intervals consecutive in time are continuous, while two small intervals consecutive in other times are discontinuous. Small intervals can also be set individually according to a specified unit time.

[0069] The state variable processing unit 121 calculates the voltage data V(t) and power data P(t) corresponding to the current time t. The state variable processing unit 121 can set the output voltage Vout at the current time t as the voltage data V(t). The state variable processing unit 121 can calculate the power data P(t) based on the output current Iout and output voltage Vout at the current time t.

[0070] In the following description, without making a special distinction between the large amount of interval data, the small amount of interval data, and the voltage data V(t) and power data P(t) calculated by the state variable processing unit 121, they will be recorded as state variable processing data.

[0071] The state variable processing unit 121 outputs state variable processing data at the current time t, which is updated every predetermined time interval. Therefore, the state variable processing data becomes time series data obtained at each predetermined time interval.

[0072] The first learned model 122 is a learned model that generates an inferred result of the SOC-OCV characteristic by using sample data corresponding to the state variable processing data and SOC-OCV characteristics as teaching data through machine learning. The first learned model 122 can output characteristic recognition information that identifies the inferred SOC-OCV characteristic as the inferred result of the SOC-OCV characteristic.

[0073] The first learned model 122 is configured as an RNN (Recurrent Neural Network). Based on this, the intermediate layers of the first learned model 122, which is an RNN, can be configured as LSTM (Long Short-Term Memory) or GRU (Gated Recurrent Unit). Alternatively, the first learned model 122 can be configured as a CNN (Convolutional Neural Network). The first learned model 122 can treat the estimation of SOC-OCV characteristics as either a regression problem or a classification problem.

[0074] In the following explanation, we will take the case where the first learned model 122 is configured as an RNN with an LSTM in the intermediate layer and the estimation of the SOC-OCV characteristics is treated as a regression problem as an example. In this case, the state variable processing data output by the state variable processing unit 121 is input to the first learned model 122 as LSTM blocks.

[0075] The SOC-OCV characteristic, as a state of the secondary battery, shows the correlation between the SOC and OCV of the secondary battery.

[0076] Figure 4A specific example of the SOC-OCV characteristic is shown. In this figure, the horizontal axis is SOC and the vertical axis is OCV. The figure shows curves C1 to C5 corresponding to five different SOC-OCV characteristics. For example, multiple SOC-OCV characteristics can be set as inferred candidates corresponding to the first learned model 122, and characteristic identification information can be assigned to the SOC-OCV characteristics set as inferred candidates.

[0077] In the example shown in the figure, for the same SOC, curve C1 has the highest OCV value, followed by curves C2, C3, C4, and C5, which decrease in that order. As a relative relationship, for example, when comparing two SOC-OCV characteristics corresponding to curves C1 and C2, the SOC-OCV characteristic corresponding to curve C1 is considered the high characteristic, and the SOC-OCV characteristic corresponding to curve C2 is considered the low characteristic.

[0078] The feature recognition unit 123 inputs the state variable processing data corresponding to the current time t into the first learned model 122. The feature recognition unit 123 obtains the feature recognition information output by the first learned model 122 corresponding to the input of the state variable processing data. The feature recognition unit 123 outputs the obtained feature recognition information as feature recognition information CID(t) corresponding to the current time t.

[0079] The second learned model 103 is a machine learning learned model that uses sample data corresponding to the input data Din(t) and the feature identification information CID(t) and SOH as teaching data. The second learned model 103 outputs SOH corresponding to the input data Din(t) input by the deterioration state estimation unit 104.

[0080] The input data Din(t) includes a large amount of interval data (charge and discharge amount LCA(t), LCA(t-1), LCA(t-2), voltage change LEV(t), LEV(t-1), LEV(t-2)), a small amount of interval data (charge and discharge amount SAC(t), voltage change SEV(t)), voltage data V(t), and characteristic identification information CID(t).

[0081] The input data Din(t) may also include the power data P(t).

[0082] The second learned model 103 is configured as an RNN (Recurrent Neural Network). Based on this, the intermediate layers of the second learned model 103, which is an RNN, can be configured as LSTM (Long Short-Term Memory) or GRU (Gated Recurrent Unit). Alternatively, the second learned model 103 can be configured as a CNN (Convolutional Neural Network). The second learned model 1103 can treat the estimation of SOH as either a regression problem or a classification problem.

[0083] In the following explanation, we will take the case where the second learned model 103 is configured as an RNN with an LSTM in the intermediate layer and the estimation of SOH is treated as a classification problem as an example. In this case, the state variable processing data output by the state variable processing unit 121 is input to the second learned model 103 as LSTM blocks.

[0084] The degradation state estimation unit 104 inputs the input data Din(t) corresponding to the current time t to the second learned model 103. The degradation state estimation unit 104 obtains the SOH value output by the second learned model 103 corresponding to the input of the input data Din(t). The degradation state estimation unit 104 outputs the obtained SOH value as the SOH estimation value Dout.

[0085] Figure 5 An example of the relationship between charge / discharge quantity and voltage change is shown. The figure shows four curves, C11, C12, C21, and C22, corresponding to different estimated SOH values. Curves C11 and C21 show examples where the secondary battery capacity is the same (α(Ah), and curves C12 and C22 show examples where the secondary battery capacity is the same (β(Ah)) but less than α(Ah).

[0086] The estimated SOH value also corresponds to a specific SOC-OCV characteristic. Therefore, these four curves C11, C12, C21, and C22 also correspond to specific SOC-OCV characteristics. As a relative relationship, the SOC-OCV characteristics corresponding to curves C11 and C12 can be set as high characteristics, and the SOC-OCV characteristics corresponding to curves C21 and C22 can be set as low characteristics to distinguish them.

[0087] In the example in this figure, as shown at the intersection point IS, there is a point where the curve corresponding to the SOC-OCV characteristic with the same high characteristic intersects with the curve corresponding to the SOC-OCV characteristic with the same low characteristic.

[0088] exist Figure 6The figure shows three curves, C31, C32, and C31-1, illustrating the relationship between charge / discharge amount and voltage change. In this figure, curves C31 and C32 correspond to different State of Charge (SOH), representing curves obtained when the corresponding secondary battery is started operating from a 100% SOC state. On the other hand, curve C31-1 corresponds to the same SOH as curve C31, but represents curves obtained when the secondary battery is started operating from a 50% SOC state. In the example shown in this figure, curves C31 and C32 do not overlap, but curve C31-1, which corresponds to the same SOC-OCV characteristics as curve C31 but has a different SOC at the start of operation, will overlap with curve C32. That is, there are overlapping intervals between multiple curves due to the SOC at the start of operation.

[0089] Based on the fluctuations of state variables measured by the state variable measurement unit 101, there are also cases where curves corresponding to different SOH values ​​overlap or intersect.

[0090] In the input data Din(t) input to the second learned model 103 by the degradation state estimation unit 104 in this embodiment, multiple charge / discharge quantities LCA(t), LCA(t-1), LCA(t-2) and three voltage changes LEV(t), LEV(t-1), LEV(t-2) corresponding to each of the multiple (3) large intervals (T1, T2-1, T2-2) are input. That is, for the second learned model 103, the charge / discharge quantities and voltage changes are input as long-term history of the large interval data, in addition to short-term history of the small interval data.

[0091] Therefore, even in cases where the estimated SOH values ​​overlap or intersect among the estimated candidates, as described above, the second learned model 103 can distinguish between overlapping and intersecting curves to identify SOH because it uses the amount of charge / discharge and voltage changes corresponding to multiple large intervals that differ in time for estimation. In this case, the second learned model 103 increases the number of input parameters within a single LSTM block, and by increasing the number of input parameters for the history, it is expected to expand the breadth of learning based on past information. As a result, it is possible to improve the accuracy of the estimated SOH value Dout(t) output by the second learned model 103.

[0092] The state variable processing data input by the feature recognition unit 123 to the first learned model 122 also contains the same data as the input data Din(t), including multiple large interval data corresponding to the charging and discharging amount and multiple large interval data corresponding to the voltage change amount.

[0093] By inputting such a large amount of interval data, the first learned model 122 can also make estimations using the charge / discharge amounts and voltage changes corresponding to multiple large intervals that differ in time. Thus, even for example... Figure 5 In cases where the curves corresponding to high and low SOC-OCV characteristics intersect, because the range of the curves can be referenced over a long period rather than instantaneously, the SOC-OCV characteristics can be appropriately identified under both high and low characteristics, avoiding misjudgments. This improves the accuracy of the SOC-OCV characteristics estimated by the first learned model 122.

[0094] The input data Din(t) input by the degradation state estimation unit 104 to the second learned model 103 includes characteristic identification information CID of the SOC-OCV characteristic estimated by the first learned model 122 as described above. That is, the input data Din(t) includes information about the SOC-OCV characteristic.

[0095] By using input data Din(t) that includes information about such SOC-OCV characteristics, when estimating the SOH using the second learned model 103, it is no longer necessary to use the SOC estimation value and the internal resistance estimation value. By not using the SOC estimation value and the internal resistance estimation value, in this embodiment, the second learned model 103 can estimate the SOH without being affected by the error of the SOC estimation value, thus improving accuracy.

[0096] The characteristic identification information CID included in the input data Din(t) used for estimating SOH is not fixedly set to correspond to the target secondary battery 200. That is, the characteristic identification information CID corresponds to the result estimated by the first learned model 122 using state variable processing data based on state variables measured about the target secondary battery 200. Therefore, in this embodiment, the characteristic identification information CID corresponding to the corresponding secondary battery 200 can be estimated with a certain degree of accuracy regardless of changes in the specifications, etc. of the secondary battery 200. Therefore, as the degradation state estimation device 100 of this embodiment, it is possible to appropriately estimate SOH corresponding to a variety of secondary batteries 200.

[0097] In this embodiment, the second learned model 103 can estimate SOH using input data Din(t) that does not include feature identification information CID, provided that certain accuracy conditions are met.

[0098] Reference Figure 7 The flowchart illustrates an example of the processing steps performed by the degradation state estimation device 100 in conjunction with the estimation of SOH. The processing in this diagram begins each time a predetermined time has elapsed.

[0099] The state variable measurement unit 101 measures the state variables of the secondary battery 200 (step S100). The state variables of the secondary battery 200 are the output current Iout and the output voltage Vout corresponding to the current time t.

[0100] In the preprocessing unit 102, the state variable processing unit 121 uses the output current Iout and output voltage Vout measured in step S100 to calculate the state variable processing data corresponding to the current time t (step S102).

[0101] The state variable processing data includes a large amount of interval data, a small amount of interval data, voltage data V(t), and power data P(t). The large amount of interval data consists of charge / discharge quantities LCA(t), LCA(t-1), LCA(t-2), and voltage changes LEV(t), LEV(t-1), LEV(t-2). The small amount of interval data consists of charge / discharge quantities SCA(t) and voltage changes SEV(t).

[0102] The feature recognition unit 123 inputs the state variable data calculated in step S102 into the first learned model 122 (step S104).

[0103] The first learned model 122 outputs characteristic identification information as the estimation result of the SOC-OCV characteristic in accordance with the state variable data input in step S104. The characteristic identification unit 123 obtains the characteristic identification information output by the first learned model 122 and outputs the obtained characteristic identification information as the characteristic identification information CID(t) corresponding to the current time t (step S106).

[0104] The degradation state estimation unit 104 inputs the input data Din(t) into the second learned model 103 (step S108).

[0105] The second learned model 103 estimates the SOH corresponding to the input data Din(t) input in step S108. The degradation state estimation unit 104 outputs the SOH estimated by the second learned model 103 as the SOH estimation value Dout(t) corresponding to the current time t (step S110).

[0106] The application of the secondary battery 200 in this embodiment is not particularly limited. The secondary battery 200 can be installed in a vehicle, for example, to power it. The secondary battery 200 can also be installed in residences, office buildings, etc. The secondary battery 200 can be included in power grids, such as smart grids.

[0107] The degradation state estimation system, which is the function of the degradation state estimation device 100 in this embodiment, can be configured as a degradation state estimation system with functions distributed across multiple devices. As an example, such a degradation state estimation system may include: a terminal device having a secondary battery to which degradation is estimated; and a cloud server communicatively connected to the terminal device. It can be configured such that the terminal device sends the state variables of the secondary battery, measured by the state variable measurement unit 101, to the cloud server, and the cloud server uses the received state variables to output a SOH estimation value through its functions as a preprocessing unit 102, a second learned model 103, and a degradation state estimation unit 104.

[0108] Alternatively, a program for implementing the functions of the aforementioned degradation state estimation device 100 can be recorded on a computer-readable recording medium, and the computer system can read and execute the program recorded on the recording medium to perform the processing described above as the degradation state estimation device 100. Here, "reading and executing the program recorded on the recording medium by the computer system" includes installing the program on the computer system. The term "computer system" here refers to hardware including an operating system and peripheral devices. The "computer system" may also include multiple computer devices connected via a network including communication lines such as the Internet, WAN, LAN, and dedicated lines. "Computer-readable recording medium" refers to removable media such as floppy disks, optical disks, ROMs, and CD-ROMs, and storage devices such as hard disks built into the computer system. In this way, the recording medium storing the program may also be a non-transitory recording medium such as a CD-ROM. The recording medium also includes recording media located internally or externally that can be accessed from a transmission server for transmitting the program. The code of the program stored on the recording medium of the transmission server may also be different from the code of a program that can be executed on a terminal device. That is, as long as it can be downloaded from the delivery server and installed in a form that can be executed on the terminal device, the form in which it is stored on the delivery server is irrelevant. It is also possible to divide the program into multiple parts, download them at different times, and then combine them on the terminal device, with the delivery server for each part of the program being different. Furthermore, the "computer-readable recording medium" is defined as a recording medium that holds the program for a certain period of time, such as a server when the program is sent via a network, or volatile memory (RAM) inside the computer system that acts as a client. The aforementioned program can also be used to implement a portion of the above-described functions. Moreover, it can also be a so-called differential file (differential program) that can achieve the above-described functions by combining it with a program already recorded on the computer system.

[0109] The above description illustrates specific embodiments of the present invention, but the present invention is not limited to such embodiments in any way, and various modifications and substitutions can be made without departing from the spirit of the present invention.

Claims

1. A state estimation system for a secondary battery, wherein, The state estimation system for the secondary battery includes: The state variable measurement unit measures the state variables of the operating secondary battery, including the output current and output voltage, at each specified time. The state variable processing unit outputs state variable processing data, including charge / discharge quantity and voltage change quantity, calculated based on the state variables measured by the state variable measurement unit; and The feature recognition unit uses the state variable processing data to identify the SOC-OCV characteristics of the secondary battery through a learned feature recognition model. The state variable processing unit sets a unit charge / discharge quantity or unit time, and sets the nearest interval to the current time and at least one past interval preceding the nearest interval as a desired interval derived from the unit charge / discharge quantity or unit time used in the calculation of the state variable processing data. For each of the set intervals, the state variable processing unit calculates the voltage change derived from the state variable as interval data, and outputs the calculated interval data, which is then included in the state variable processing data. The state variable processing unit sets a large charge / discharge quantity as a unit charge / discharge quantity and a small charge / discharge quantity as a unit charge / discharge quantity that is smaller than the large charge / discharge quantity. It also sets a large interval based on the large charge / discharge quantity and a small interval based on the small charge / discharge quantity as a desired interval derived from the large charge / discharge quantity and the small charge / discharge quantity used in the calculation of the state variable processing data.

2. The state estimation system according to claim 1, The state variable processing unit sets a large number of intervals based on the unit time used in the calculation of the state variable processing data and a small number of intervals based on a shorter time than the large number of intervals, as a desired interval.

3. The state estimation system according to claim 1, The state variable processing unit sets the period during which the specified charge / discharge quantity is calculated from the output current measured by the state variable measurement unit, or the period during which the specified value is reached, as the desired interval.

4. The state estimation system according to claim 1, At least one of the calculated charge / discharge amount for each desired interval and the calculated voltage change for each desired interval is a slope change rate calculated using the least squares method.

5. The state estimation system according to any one of claims 1 to 4, The feature recognition model is constructed as an RNN, or recurrent neural network.

6. The state estimation system according to claim 5, The intermediate layers of the RNN are configured as LSTM (Long Short-Term Memory) or GRU (Gated Recurrent Unit).

7. The state estimation system according to any one of claims 1 to 4, The feature recognition model is constructed as a CNN, or Convolutional Neural Network.

8. A state estimation system for a secondary battery, wherein, The secondary battery state estimation system includes: The state variable measurement unit measures the state variables of the operating secondary battery, including the output current and output voltage, at each specified time. The state variable processing unit outputs state variable processing data, including charge / discharge quantity and voltage change quantity, calculated based on the state variables measured by the state variable measurement unit. The feature recognition unit uses the state variable to process the data, identifies the SOC-OCV characteristics of the secondary battery through the learned feature recognition model, and outputs feature recognition information representing the recognition result. as well as The degradation state estimation unit uses input data including the charge / discharge amount, the voltage change amount, and the characteristic identification information to estimate the degradation state of the operating secondary battery through a learned degradation state model. The state variable processing unit sets a unit charge / discharge quantity or unit time, and sets the nearest interval to the current time and at least one past interval preceding the nearest interval as a desired interval derived from the unit charge / discharge quantity or unit time used in the calculation of the state variable processing data. For each of the set intervals, the state variable processing unit calculates the voltage change derived from the state variable as interval data, and outputs the calculated interval data, which is then included in the state variable processing data. The state variable processing unit sets a large charge / discharge quantity as a unit charge / discharge quantity and a small charge / discharge quantity as a unit charge / discharge quantity that is smaller than the large charge / discharge quantity. It also sets a large interval based on the large charge / discharge quantity and a small interval based on the small charge / discharge quantity as a desired interval derived from the large charge / discharge quantity and the small charge / discharge quantity used in the calculation of the state variable processing data.

9. A method for estimating the state of a secondary battery, wherein, The computer in the state-predicted system performs the following processing: Measure the state variables of the operating secondary battery, including output current and output voltage, at each specified time. Output state variable processing data, including charge / discharge quantity and voltage change quantity, calculated based on the measured state variables; The state variables are used to process the data to identify the SOC-OCV characteristics of the secondary battery through the learned characteristic recognition model; Set a unit charge / discharge amount or unit time, and set the nearest interval to the current time and one or more past intervals before the nearest interval as the expected interval derived from the unit charge / discharge amount or unit time used in the calculation of the state variable processing data. For each set interval, calculate the voltage change derived from the state variable as interval data, and include the calculated interval data in the state variable processing data and output it. A large charge / discharge quantity is defined as a unit charge / discharge quantity, and a small charge / discharge quantity is defined as a unit charge / discharge quantity that is smaller than the large charge / discharge quantity. A large range based on the large charge / discharge quantity and a small range based on the small charge / discharge quantity are defined as expected ranges derived from the large charge / discharge quantity and the small charge / discharge quantity used in the calculation of the state variable processing data.

10. A method for estimating the state of a secondary battery, wherein, The computer in the state-predicted system performs the following processing: Measure the state variables of the operating secondary battery, including output current and output voltage, at each specified time. Output state variable processing data, including charge / discharge quantity and voltage change quantity, calculated based on the measured state variables; The state variable is used to process the data and identify the SOC-OCV characteristics of the secondary battery through the learned characteristic identification model, and the characteristic identification information representing the identification result is output. The degradation state of the operating secondary battery is estimated using input data including the charge / discharge amount, the voltage change amount, and the characteristic identification information through a learned degradation state model. Set a unit charge / discharge amount or unit time, and set the nearest interval to the current time and one or more past intervals before the nearest interval as the expected interval derived from the unit charge / discharge amount or unit time used in the calculation of the state variable processing data. For each set interval, calculate the voltage change derived from the state variable as interval data, and include the calculated interval data in the state variable processing data and output it. A large charge / discharge quantity is defined as a unit charge / discharge quantity, and a small charge / discharge quantity is defined as a unit charge / discharge quantity that is smaller than the large charge / discharge quantity. A large range based on the large charge / discharge quantity and a small range based on the small charge / discharge quantity are defined as expected ranges derived from the large charge / discharge quantity and the small charge / discharge quantity used in the calculation of the state variable processing data.

11. A storage medium storing a program, wherein, This program causes the computer in the state-predicting system to perform the following processing: Measure the state variables of the operating secondary battery, including output current and output voltage, at each specified time. Output state variable processing data, including charge / discharge quantity and voltage change quantity, calculated based on the measured state variables; The state variables are used to process the data to identify the SOC-OCV characteristics of the secondary battery through the learned characteristic recognition model; Set a unit charge / discharge amount or unit time, and set the nearest interval to the current time and one or more past intervals before the nearest interval as the expected interval derived from the unit charge / discharge amount or unit time used in the calculation of the state variable processing data. For each set interval, calculate the voltage change derived from the state variable as interval data, and include the calculated interval data in the state variable processing data and output it. A large charge / discharge quantity is defined as a unit charge / discharge quantity, and a small charge / discharge quantity is defined as a unit charge / discharge quantity that is smaller than the large charge / discharge quantity. A large range based on the large charge / discharge quantity and a small range based on the small charge / discharge quantity are defined as expected ranges derived from the large charge / discharge quantity and the small charge / discharge quantity used in the calculation of the state variable processing data.

12. A storage medium storing a program, wherein, This program causes the computer in the state-predicting system to perform the following processing: Measure the state variables of the operating secondary battery, including output current and output voltage, at each specified time. Output state variable processing data, including charge / discharge quantity and voltage change quantity, calculated based on the measured state variables; The state variable is used to process the data and identify the SOC-OCV characteristics of the secondary battery through the learned characteristic identification model, and the characteristic identification information representing the identification result is output. The degradation state of the operating secondary battery is estimated using input data including the charge / discharge amount, the voltage change amount, and the characteristic identification information through a learned degradation state model. Set a unit charge / discharge amount or unit time, and set the nearest interval to the current time and one or more past intervals before the nearest interval as the expected interval derived from the unit charge / discharge amount or unit time used in the calculation of the state variable processing data. For each set interval, calculate the voltage change derived from the state variable as interval data, and include the calculated interval data in the state variable processing data and output it. A large charge / discharge quantity is defined as a unit charge / discharge quantity, and a small charge / discharge quantity is defined as a unit charge / discharge quantity that is smaller than the large charge / discharge quantity. A large range based on the large charge / discharge quantity and a small range based on the small charge / discharge quantity are defined as expected ranges derived from the large charge / discharge quantity and the small charge / discharge quantity used in the calculation of the state variable processing data.

Images