Storage battery deterioration diagnosis device, system, method, and program

The diagnostic device and system address the inaccuracies in conventional battery diagnosis by analyzing charge/discharge data and material information to accurately assess storage battery health and lithium precipitation risk, enhancing the reliability and safety of energy storage systems.

WO2024232353A1PCT designated stage expired Publication Date: 2024-11-14ASAHI KASEI KOGYO KABUSHIKI KAISHA

Patent Information

Application Number
PCT/JP2024/016945
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2023-10-30
Filing Date
2024-05-07
Publication Date
2024-11-14

AI Technical Summary

Technical Problem

Conventional storage battery deterioration diagnosis methods based on state of health (SoH) and state of safety (SoS) indicators often lead to erroneous determinations, causing storage batteries to be incorrectly deemed unusable or in need of immediate replacement, resulting in operational disruptions and potential safety hazards.

Method used

A diagnostic device and system that utilizes a trained model to analyze charge/discharge data, material information, and operating conditions to accurately diagnose material deterioration and lithium precipitation risk in storage batteries, providing a more precise assessment of their state and predicting future deterioration trends.

Benefits of technology

Enables more accurate storage battery deterioration diagnosis, reducing the likelihood of premature replacement and operational disruptions, while ensuring safety by providing reliable and timely maintenance insights.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure JP2024016945_14112024_PF_FP_ABST
    Figure JP2024016945_14112024_PF_FP_ABST
Patent Text Reader

Abstract

The purpose of the present disclosure is to provide a storage battery deterioration diagnosis device, a system, a method, and a program that can more accurately diagnose deterioration of the storage battery. A diagnosis device according to the present disclosure is used in an energy storage system. The diagnosis device comprises: a storage unit that stores a trained model trained to output material deterioration information about a storage battery when charge / discharge data is input; an acquisition unit that acquires charge / discharge data about the storage battery; a diagnosis unit that diagnoses the material deterioration information about the storage battery obtained by inputting the charge / discharge data about the storage battery to the trained model; and an output unit that outputs, as a diagnosis result, the material deterioration information obtained from the diagnosis unit.
Need to check novelty before this filing date? Find Prior Art

Description

Battery deterioration diagnosis device, system, method, and program

[0001] The present disclosure relates to a storage battery deterioration diagnosis device, system, method, and program.

[0002] In recent years, from the viewpoint of preserving the global environment or effectively utilizing energy to conserve resources, attention has been focused on power smoothing systems for wind power generation, late-night power storage systems, power storage systems used in conjunction with solar power generation, power storage systems for electric vehicles, etc. Batteries used in these power storage systems are required to have high energy density. As a promising candidate for a high-energy density battery that can meet such requirements, the development of storage batteries, typified by lithium-ion secondary batteries (hereinafter referred to as LiBs), has been actively pursued.

[0003] LiBs generally consist of a positive electrode having a positive electrode active material layer containing a positive electrode active material, a negative electrode having a negative electrode active material layer containing a negative electrode active material, a separator disposed between the positive electrode and the negative electrode, and an electrolyte. LiBs are produced by assembling a cell and then doping lithium ions into the negative electrode active material through a charging process.

[0004] Due to their high energy density, LiBs have recently been used in electric vehicles and energy storage systems (hereinafter referred to as ESS). ESSs are often used to adjust the power output of renewable energy sources, whose power generation is affected by weather. Without power adjustment by ESSs, fluctuations in voltage and frequency can cause problems such as unstable device operation, breakdowns, or power generation equipment shutting down and causing power outages. Therefore, it is important for ESSs to maintain a certain level of capacity and output and maintain a state in which they can always adjust to fluctuations in renewable energy. It is also necessary to periodically diagnose the deterioration state of storage batteries (deterioration diagnosis) and, if necessary, perform maintenance work such as replacing the storage batteries.

[0005] As a technical approach to this problem, a battery management system (hereinafter referred to as BMS) technology has been developed that detects and evaluates the state of charge (SoC), state of health (SoH), and safety indicators (SoS), etc. of the storage battery, controls or stops the charging and discharging of the storage system before an accident occurs, and balances the SoC of multiple cells when there is a deviation between the SoCs of the individual cells. For example, the SoH evaluation is performed by integrating the discharge power (W) over time (hr) from 100% to 0% of the storage system's SoC to obtain the discharged energy E (Wh), and then calculating this as the initial discharged energy E 0 Divide by (SoH = E / E 0 × 100 (%)). By monitoring the SoH of a power storage system, the system operator can optimize operation plans and take appropriate measures such as replacing storage batteries. The SoS can also be measured by, for example, voltage and temperature detection. For example, when LiBs are held at a high voltage of 4.5 V or higher, there is a risk of gas swelling and fire due to the decomposition of the electrolyte solvent and structural destruction of the positive electrode active material. To prevent this, BMSs are equipped with a mechanism that predetermines the upper voltage limit at which LiBs can be used and stops charging when the upper voltage limit is reached.

[0006] However, the above-described method merely quantifies electrochemical data into management indices such as SoH and SoS regarding the safety and reliability of a storage battery, but does not provide information on the deterioration of storage battery materials (hereinafter referred to as "material deterioration information"). As a result, even when a storage battery to be diagnosed is actually in a usable condition, it is often erroneously determined to be unusable based on indices such as SoH and SoS.

[0007] Several studies have been conducted on methods for diagnosing deterioration of storage batteries in these energy storage systems. For example, Patent Document 1 describes a method in which a predetermined voltage is applied to the energy storage system, the current response is detected, and the amount of lithium (Li) deposited inside the battery is estimated based on the detected current. However, this method only estimates the amount of Li deposited from electrochemical data, and does not know the actual state of Li generated in the energy storage system, making it impossible to accurately diagnose deterioration of the storage battery.

[0008] Patent Document 2 describes a method for determining an internal short circuit from the electrochemical behavior of the internal short circuit itself. While this method directly evaluates the phenomenon of an internal short circuit, it is not possible to diagnose the deterioration of the storage battery with high accuracy due to the type of LiB and the influence of noise during measurement.

[0009] Patent Document 3 describes a technique for determining whether lithium deposition has occurred by detecting the presence or absence of a minimum point where the charging current changes from a decrease to an increase during constant voltage charging. This method indirectly evaluates lithium deposition, and therefore cannot accurately diagnose the deterioration of a storage battery.

[0010] Patent Document 4 discloses a technology that creates a database of correlations between overvoltages at the input and output of a storage battery and SoH, and then uses this database as training data to construct a machine learning model to estimate SoH in real time from charge / discharge data during actual ESS operation. While this method is excellent for estimating the degradation state of a storage battery in real time, it is not possible to understand the mode that the degradation state is based on. Therefore, it is not possible to accurately diagnose the degradation of a storage battery.

[0011] JP 2010-86901 A JP 2009-32506 A JP 2012-3863 A International Publication No. 2017 / 022037

[0012] Zachary M. Konz, et al., “Detecting the Onset of Lithium Plating and Monitoring Fast Charging Performance with Voltage Relaxation”, American Chemical Society Energy Letter., (2020), 5(6), pp.1750-1757

[0013] As described above, conventional diagnoses based on SoS and SoH merely estimate the state of a storage battery from electrochemical data, and often have the problem of making erroneous judgments. As a result, a storage battery may be erroneously judged as having reached the end of its life even when it is actually in a usable condition, or may be erroneously judged as still usable even when it actually needs to be replaced immediately. Furthermore, if the diagnosis reveals that a storage battery has deteriorated to the point where it needs to be replaced immediately, continued operation could lead to operational malfunctions or fires due to the battery ignition, so the ESS must be temporarily stopped and maintenance (such as battery replacement) performed. However, from the perspective of the ESS operator, this means stopping the ESS at an unplanned time, resulting in a significant loss of opportunity.

[0014] Therefore, one of the objects of the present disclosure is to provide a storage battery deterioration diagnosis device, system, method, and program that are capable of more accurate storage battery deterioration diagnosis.

[0015] Examples of embodiments of the present disclosure are listed in the following items [1] to

[45] . [1] A diagnostic device for a storage battery used in an energy storage system, comprising: a memory unit that stores a trained model A that has been trained to output material degradation information of the storage battery when charge / discharge data is input; an acquisition unit that acquires charge / discharge data of the storage battery; a diagnostic unit that diagnoses the material degradation information of the storage battery obtained by inputting the charge / discharge data of the storage battery to the trained model A; and an output unit that outputs the material degradation information obtained from the diagnostic unit as a diagnostic result. [2] The diagnostic device according to item 1, wherein the material information of the storage battery is information on at least one material selected from the group consisting of a positive electrode, a negative electrode, a separator, and an electrolyte of the storage battery. [3] The diagnostic device according to item 2, further comprising: a calculation unit that calculates a risk X of lithium precipitation of the storage battery based on information including the material degradation information, and the diagnostic result includes the risk X of lithium precipitation. [4] The diagnostic device according to item 3, wherein the calculation unit calculates a risk X of lithium precipitation of the storage battery based on information including the material deterioration information and at least one selected from the group consisting of the charge / discharge data, material information of the storage battery, and operating conditions of the energy storage system. [5] The calculation unit calculates the risk X of lithium precipitation inferred from the material deterioration information. chem. A weighting coefficient α according to the material information of the negative electrode of the storage battery is added to chem. and the lithium deposition risk X estimated from the charge / discharge data. elec. A weighting coefficient α according to the material information of the negative electrode of the storage battery is added to elec.and optionally multiplying the sum by at least one selected from the group consisting of a weighting factor β corresponding to material information of a positive electrode of the storage battery and a weighting factor γ corresponding to the operating conditions, thereby calculating a risk of lithium precipitation X of the storage battery. [6] The diagnostic device according to any one of items 1 to 5, further including a calculation unit that calculates a state of degradation of the storage battery based on a state of health (SoH) of the storage battery estimated based on charge and discharge data of the storage battery and the material degradation information, and the output unit outputs the state of degradation as a diagnostic result. [7] The diagnostic device according to item 6, wherein the calculation unit is configured to calculate a risk of lithium precipitation X of the storage battery based on information including the material degradation information, and label the state of degradation of the storage battery based on one or more predetermined thresholds for each of the SoH and the risk of lithium precipitation X, and the diagnostic result includes a result of the labeling. [8] The diagnostic device according to item 7, wherein the output unit transmits the diagnostic result to a terminal device of a user. [9] The calculation unit calculates the material deterioration information and the current time (t 1 ) and, optionally, at least one selected from the group consisting of material information of the storage battery and operating conditions of the energy storage system, 1 ) SoH and lithium deposition risk Xt 1

[10] The diagnostic device according to item 9, wherein the material information of the storage battery is information on at least one material selected from the group consisting of a positive electrode, a negative electrode, a separator, and an electrolyte of the storage battery.

[11] The calculation unit calculates a lithium deposition risk X chem. A weighting coefficient α according to the material information of the negative electrode of the storage battery is added to chem. and the lithium deposition risk X estimated from the charge / discharge data. elec. A weighting coefficient α according to the material information of the negative electrode of the storage battery is added to elec.and optionally multiplying the sum by at least one selected from the group consisting of a weighting factor β according to the material information of the positive electrode of the storage battery and a weighting factor γ according to the operating conditions, thereby obtaining the current time (t 1 ) Lithium deposition risk Xt 1

[12] The storage unit calculates the current time (t 1 ) SoH and lithium deposition risk Xt 1 and a trained model B that has been trained to output a transition of the future risk of lithium deposition X when information including past charge / discharge data and a transition of the past SoH is input, and the diagnosis device further includes a prediction unit, and the prediction unit is configured to predict the current time (t 1 ) SoH and lithium deposition risk Xt 1 and information including past charge / discharge data and past changes in SoH are input to the trained model B, and a future change in the lithium deposition risk X is predicted. The calculation unit calculates a time point (t 2 ) Lithium deposition risk Xt 2 and the lithium deposition risk Xt 2

[13] The diagnostic device according to item 7, further comprising a control unit that controls the storage battery based on a result of the labeling.

[14] The diagnostic device according to item 13, wherein the control unit controls the storage battery based on the result of the labeling and a cause of deterioration of the storage battery.

[15] The diagnostic device according to item 13, wherein the control unit controls the storage battery based on the result of the labeling and a target operating period of the energy storage system.

[16] The diagnostic device according to item 13, wherein the storage battery is composed of a plurality of storage battery modules arranged in series and / or parallel, each of which includes a plurality of cells arranged in series and / or parallel, and the diagnosing unit and calculating unit diagnose and label the state of deterioration of each storage battery module, and the control unit controls charge and discharge conditions for each storage battery module.

[17] The material deterioration information of the storage battery includes the area of ​​deposited lithium metal analyzed from an image of the negative electrode surface of the storage battery, the shape of the deposited region, the weight of the deposited lithium metal, the distribution state and color tone, the shape of the lithium metal observed by cross-sectional SEM observation of the negative electrode, the chemical shift value (eV) and / or its area obtained from the Li(s1) spectrum obtained by XPS analysis, 7Item 6. The diagnostic device according to Item 6, wherein the charge / discharge data input to the trained model A is at least one selected from the group consisting of a chemical state expressed by a chemical shift value (ppm) and / or its area obtained from a Li-NMR spectrum, and a g value and / or line width obtained from an ESR spectrum.

[18] The diagnostic device according to Item 6, wherein the charge / discharge data input to the trained model A is at least one selected from the group consisting of a time change in voltage when the storage battery is rested immediately after being charged to an arbitrary SoC (dOCV / dt-t plot), a time change in voltage during pulse charging and / or discharging at a constant current (V-t plot), a time change in current value observed during constant voltage charging (It plot), a charge capacity and a discharge capacity at a constant current or constant power, measurement data obtained by an AC impedance method, and data obtained by mathematically processing these.

[19] The diagnostic device according to item 6, wherein the SoH is estimated based on the capacity when the storage battery is charged and discharged at a constant current and a constant voltage, based on the internal resistance obtained by an AC impedance method, based on the internal resistance obtained from the slope of the current-voltage characteristics (IV characteristics), based on the internal resistance obtained from the voltage transient response when the current is changed in a stepwise manner, based on fitting of charging curve data using a regression equation or a model, based on an equivalent circuit or a physical model that behaves similarly to the charge and discharge data of the storage battery, or a combination of these.

[20] The diagnostic device according to item 19, wherein the SoH is estimated based on the internal resistance obtained from the slope of the current-voltage characteristics (IV characteristics).

[21] A diagnostic device for a storage battery used in an energy storage system, comprising: an acquisition unit that acquires charge / discharge data of the storage battery; a receiving unit that receives a trained model A that outputs material deterioration information from the charge / discharge data of the storage battery; a diagnostic unit that diagnoses the material deterioration information of the storage battery obtained by inputting the charge / discharge data of the storage battery into the received trained model A; and an output unit that outputs the material deterioration information obtained from the diagnostic unit as a diagnostic result.

[22] The diagnostic device according to item 21, further including a calculation unit that estimates a state of health (SoH) of the storage battery based on charge / discharge data of the storage battery and calculates a degradation state of the storage battery based on the SoH and the material degradation information, and the output unit outputs the degradation state as a diagnostic result.

[23] The diagnostic device according to item 22, wherein the calculation unit is configured to calculate a lithium deposition risk X of the storage battery based on information including the material degradation information and label the degradation state of the storage battery based on one or more predetermined thresholds for each of the SoH and the lithium deposition risk X, and the diagnostic result includes a result of the labeling.

[24] The diagnostic device according to item 23, wherein the output unit transmits the diagnostic result to a user's terminal device.

[25] The diagnostic device according to item 23 or 24, further including a control unit that controls the storage battery based on the labeling result.

[26] A diagnostic system for a storage battery used in an energy storage system, comprising: a storage device that stores a learned model A that has been trained to output material degradation information of the storage battery when charge / discharge data is input, an acquisition device that acquires charge / discharge data of the storage battery, a diagnostic device that diagnoses the material degradation information of the storage battery obtained by inputting the charge / discharge data of the storage battery to the learned model A, and an output device that outputs the material degradation information obtained from the diagnostic device as a diagnostic result.

[27] The diagnostic system according to item 26, further comprising a calculation device that estimates a state of health (SoH) of the storage battery based on the charge / discharge data of the storage battery and calculates a degradation state of the storage battery based on the SoH and the material degradation information, and the output device outputs the degradation state as a diagnostic result.

[28] The diagnostic system according to item 27, wherein the calculation device is configured to calculate a lithium precipitation risk X of the storage battery based on information including the material degradation information, and to label a degradation state of the storage battery based on one or more predetermined thresholds for each of the State of Health and the lithium precipitation risk X, and the diagnostic result includes a result of the labeling.

[29] The diagnostic system according to Item 28, wherein the diagnostic system further includes a display device, wherein the output device transmits the diagnostic result to the display device, and the display device displays the diagnostic result.

[30] The diagnostic system according to Item 28 or 29, further including a control device that controls the storage battery based on the labeling result.

[31] A diagnostic system for a storage battery to be used in an energy storage system, including a server and a diagnostic device, wherein the server has: a memory unit that stores a trained model A that has been trained to output material deterioration information of the storage battery when charge / discharge data is input; and a transmission unit that transmits the trained model A stored in the memory unit to a diagnostic device that diagnoses the material deterioration information of the storage battery, and the diagnostic device has: an acquisition unit that acquires charge / discharge data of the storage battery, a receiving unit that receives the trained model A, a diagnostic unit that diagnoses the material deterioration information of the storage battery obtained by inputting the charge / discharge data of the storage battery to the received trained model A, and an output unit that outputs the material deterioration information obtained from the diagnostic unit as a diagnostic result.

[32] The diagnostic system according to Item 31, wherein the server further includes a calculation unit that estimates a state of health (SoH) of the storage battery based on charge / discharge data of the storage battery and calculates a degradation state of the storage battery based on the SoH and the material degradation information, and the output unit outputs the degradation state as a diagnostic result.

[33] The diagnostic system according to Item 32, wherein the calculation unit is configured to calculate a lithium deposition risk X of the storage battery based on information including the material degradation information and label the degradation state of the storage battery based on one or more predetermined thresholds for each of the SoH and the lithium deposition risk X, and the diagnostic result includes a result of the labeling.

[34] The diagnostic system according to Item 33, wherein the diagnostic system further includes a display device, and the output unit transmits the diagnostic result to the display device, and the display device displays the diagnostic result.

[35] The diagnostic system according to Item 33 or 34, further including a control unit that controls the storage battery based on the labeling result.

[36] A diagnostic method for a storage battery used in an energy storage system, the method comprising the steps of: acquiring charge and discharge data of the storage battery, acquiring material degradation information of the storage battery by inputting the charge and discharge data of the storage battery into a trained model A that has been trained to output material degradation information of the storage battery when the charge and discharge data is input, diagnosing the obtained material degradation information of the storage battery, and outputting the material degradation information as a diagnostic result.

[37] The diagnostic method according to item 36, further comprising the steps of: estimating a state of health (SoH) of the storage battery based on the charge and discharge data of the storage battery, calculating a degradation state of the storage battery based on the SoH and the material degradation information, and outputting the degradation state as a diagnostic result.

[38] The diagnostic method according to Item 37, wherein calculating the degradation state includes calculating a lithium deposition risk X of the storage battery based on information including the material degradation information, and labeling the degradation state of the storage battery based on one or more predetermined thresholds for each of the SoH and the lithium deposition risk X, and the diagnostic result includes a result of the labeling.

[39] The diagnostic method according to Item 38, further comprising a step of displaying the output diagnostic result on a display unit.

[40] The diagnostic method according to Item 38 or 39, further comprising a step of controlling the storage battery based on the result of the labeling.

[41] A diagnostic program for a storage battery used in an energy storage system, the diagnostic program causing a computer to execute the following steps: acquire charge and discharge data of the storage battery; diagnose material degradation information of the storage battery obtained by inputting the charge and discharge data of the storage battery into a trained model A that has been trained to output material degradation information of the storage battery when the charge and discharge data is input; and output the material degradation information as a diagnostic result.

[42] The diagnostic program according to Item 41, further causing a computer to: estimate a state of health (SoH) of the storage battery based on charge and discharge data of the storage battery, calculate a degradation state of the storage battery based on the SoH and material degradation information of the storage battery obtained by inputting the charge and discharge data of the storage battery into a trained model A that has been trained to output material degradation information of the storage battery when the charge and discharge data is input, and output the degradation state as a diagnostic result.

[43] The diagnostic program according to Item 42, further causing a computer to: calculate a lithium deposition risk X of the storage battery based on information including the material degradation information, and label the degradation state of the storage battery based on one or more predetermined thresholds for each of the SoH and the lithium deposition risk X, the diagnostic result including a result of the labeling.

[44] The diagnostic program according to Item 43, further causing a computer to display the output diagnostic result on a display unit.

[45] The diagnostic program according to item 43 or 44, further causing a computer to control the storage battery based on a result of the labeling.

[0016] According to the present disclosure, it is possible to provide a storage battery deterioration diagnosis device, system, method, and program that enable more accurate deterioration diagnosis of a storage battery.

[0017] FIG. 1 is a diagram showing a schematic configuration of a diagnostic system 1 according to an embodiment. FIG. 2 is a diagram showing a schematic configuration of an ESS 10. FIG. 3 is a diagram showing a schematic configuration of a storage battery stack 200. FIG. 4 is a diagram showing a schematic configuration of a local BMS 300. FIG. 5 is a diagram showing a schematic configuration of a cloud BMS 400 as a diagnostic device and server, which can be combined with the local BMS 300 of FIG. 4. FIG. 6 is a flowchart showing an example of the operation of a learning process of a trained model A. FIG. 7 is a flowchart showing an example of the operation of a learning process of a trained model B. FIG. 8 is a flowchart showing an example of the operation of a diagnostic process. FIG. 9 is a flowchart showing an example of the operation of a calculation and control process of a lithium deposition risk X. FIG. 10 is a flowchart showing an example of the operation of a calculation and labeling process of a lithium deposition risk X. FIG. 11 is a flowchart showing an example of the operation of predicting a future lithium deposition risk X. FIG. 12 is a flowchart showing an example of a process of labeling the degradation state of a storage battery based on a future lithium deposition risk X. FIG. 13 is a flowchart showing an example of the operation of a control process of a storage battery based on a diagnosis result. Fig. 14 is a flowchart showing an example of the operation of the control processing of the storage battery based on the diagnosis result and the target operating period. Fig. 15 is a diagram showing a schematic configuration of a cloud BMS 400 as a diagnostic device and server, which can be combined with the local BMS 300 of Fig. 4. Fig. 16 is a diagram showing a schematic configuration of the local BMS 300. Fig. 17 is a diagram showing a schematic configuration of a cloud BMS 400 as a diagnostic device and server, which can be combined with the local BMS 300 of Fig. 16. Fig. 18 is a diagram showing a schematic configuration of the local BMS 300 as a diagnostic device. Fig. 19 is a diagram showing a schematic configuration of a cloud BMS 400 as a server, which can be combined with the local BMS 300 of Fig. 18.

[0018] Hereinafter, a deterioration diagnosis device, system, method, and program according to one aspect of the present disclosure will be described with reference to the drawings. However, please note that the technical scope of the present invention is not limited to the embodiments, but extends to the inventions set forth in the claims and their equivalents.

[0019] <Diagnostic System and Diagnostic Device> FIG. 1 is a diagram illustrating a schematic configuration of a diagnostic system 1 according to an embodiment. As illustrated in FIG. 1, the diagnostic system 1 includes one or more ESSs 10, a cloud BMS 400, one or more terminal devices 500, and the like. In FIG. 1, connections indicated by dashed dotted lines indicate electrical connections, and connections indicated by solid lines indicate connections that allow communication. The ESS 10 includes a power conditioner (PCS) 100 including a power converter electrically connected to a power grid, a storage battery stack 200 electrically connected to the PCS, and a local BMS 300 communicatively connected to the PCS and the storage battery stack. The local BMS 300, the cloud BMS 400, and the terminal device 500 are communicatively connected to each other via a network N. The network N is a wired network such as the Internet or an intranet. The network N may also be a wireless network such as a wireless local area network (LAN).

[0020] <Energy Storage System (ESS)> Fig. 2 is a diagram showing a schematic configuration of the ESS 10. As shown in Fig. 2, the ESS 10 includes a PCS 100 and a battery stack 200 electrically connected to a power grid, a local BMS 300 communicatively connected to the PCS 100 and a battery stack 200, and an energy management system (EMS) 240. Typically, the PCS 100, the battery stack 200, the local BMS 300, and the EMS 240 are housed in a single container. In Fig. 2, connections indicated by thick solid lines indicate electrical connections, and connections indicated by thin solid lines indicate connections capable of communication.

[0021] The battery stack 200 includes one or more battery modules 210 arranged in series and / or parallel, each including a plurality of batteries (cells 211) arranged in series and / or parallel, and an end BMS 230 connected to sensors (not shown) of the individual batteries to acquire charge / discharge data, etc. The batteries constituting one battery module are typically connected in series, and the battery modules are connected in parallel with one another.

[0022] The power grid is connected to power sources, loads (power consumers), and facilities (not shown) that perform power transformation, power transmission, power distribution, power smoothing, etc. Power sources include power generation facilities such as wind power generation and solar power generation. Loads include power consumers such as offices, homes, commercial facilities, and factory facilities, as well as power consuming devices such as electric vehicles.

[0023] <Power Conditioner (PCS)> In FIG. 2 , the PCS 100 has the function of adjusting the power charged and discharged from the storage battery. The PCS 100 may be of any form as long as it is functionally commonly used. Generally, the power grid is controlled by alternating current (AC) and the storage battery stack 200 is controlled by direct current (DC). Therefore, the PCS typically includes a power converter that performs AC-DC conversion. If the power on the power grid side is DC, the power converter is a converter that performs DC-DC conversion. The PCS can change the charging and discharging conditions of the storage battery stack 200 in response to external charging and discharging condition commands, including maintaining, changing, or suspending the charging and discharging conditions. The charging and discharging condition commands may be received directly from the local BMS 300 or an energy management system (EMS), or indirectly from the cloud BMS 400 via the network N.

[0024] 2 , the EMS 240 is communicatively connected to the local BMS 300 and the PCS 100. The EMS is communicatively connected to a power supply source and a load (not shown) via a network N, and outputs charge / discharge condition commands to the PCS according to the amount of power generated by the power supply source and the state of the load, thereby controlling the input / output of energy between the power supply source, the load, and the ESS. The EMS is also communicatively connected to the cloud BMS 400 via the network N and may receive charge / discharge condition commands from the cloud BMS 400. The EMS can control the local BMS 300 and / or the PCS 100 based on the received charge / discharge condition commands.

[0025] <Storage Battery Stack> Figure 3 is a diagram showing a schematic configuration of the storage battery stack 200. The storage battery stack 200 has the function of receiving charging and discharging DC power provided by the power converter of the PCS 100 and storing and discharging electrical energy. The storage battery stack 200 includes one or more storage battery modules 210 including one or more storage batteries (cells 211), a sensor unit 220 that measures charging and discharging data from each storage battery that constitutes the storage battery module, and an end BMS 230 connected to each storage battery module.

[0026] The storage battery is, for example, a secondary battery that can be repeatedly charged and discharged, such as a lithium-ion secondary battery (LiB), a lead-acid battery, or a nickel-metal hydride secondary battery, and preferably an LiB. LiBs can exhibit high energy density and high output, thereby reducing the size and cost of the ESS. Any type of LiB may be used as long as it operates according to the principle of lithium-ion secondary battery. Examples of LiBs include ternary lithium-ion secondary batteries (NCM111), high-nickel lithium-ion secondary batteries (NCM811, NCM622, NCM523), iron phosphate lithium-ion secondary batteries (LFP), and manganese lithium-ion secondary batteries. From the perspective of increasing capacity, NCM-based and high-nickel-based batteries with high energy density are preferred, while LFP-based batteries with stable positive electrode materials are preferred from the perspective of safety.

[0027] The end BMS 230 includes an acquisition unit 231 that acquires charge / discharge data from the sensor unit, a monitoring unit 232 (IC) that monitors the charge / discharge states of the storage batteries, and a balance adjustment / stop unit 233 that can balance (equalize) the SoCs of the storage batteries and change charge / discharge conditions, such as stopping charge / discharge. The balance adjustment / stop unit 233 is a mechanism in a current circuit that, upon receiving a command from the local BMS 300, can equalize the SoCs among multiple storage batteries by, for example, dissipating the power of a storage battery with a high SoC as heat from a resistor, and can stop charging / discharging of a storage battery in response to a charge / discharge stop command from the local BMS. Charge / discharge condition commands from the cloud BMS 400 are transmitted to the EMS 240 and then to the local BMS 300 via the network N, or are transmitted to the local BMS 300 via the network N. The local BMS 300 controls each end BMS 230 based on the charge / discharge condition commands, thereby controlling the charge / discharge conditions of the storage batteries. The end BMS may also be commonly referred to as a cell management unit (CMU).

[0028] The charge / discharge data measured by the sensor unit includes, for example, the voltage, current, and temperature of the storage battery, their changes over time, and mathematically processed data. More specifically, the charge / discharge data includes, for example, the voltage change over time when the storage battery is rested immediately after charging to an arbitrary SoC (dOCV / dt-t plot), the voltage change over time during pulse charging and / or discharging at a constant current (V-t plot), the current change over time observed during constant voltage charging (I-t plot), the charge capacity and discharge capacity at a constant current or constant power, measurement data obtained by AC impedance analysis, and mathematically processed data. By using charge / discharge data acquired during actual operation of the ESS, material degradation information can be inferred with even greater accuracy. For example, the charge / discharge data may be data related to the voltage, current, and temperature when the storage battery is charged or discharged due to operations such as inertial force or frequency adjustment, response to instantaneous voltage drops, peak shifting, or arbitrage in a stationary energy storage system. Alternatively, the charge / discharge data may be data obtained from charging / discharging according to a charge / discharge program separately designated for inspection, maintenance, etc. of the ESS, separate from the normal operating conditions of the ESS.

[0029] The sensor unit may measure charge / discharge data using any method. For example, voltage measurements can be performed by attaching voltage sensor terminals to the positive and negative terminals of the target storage battery. Current measurements can be performed by attaching an ammeter to the target wiring. Temperature measurements can be performed by attaching a temperature sensor to the target storage battery. The measurement resolution of voltage, current, and temperature may be the same as that of a typical sensor, such as 8-bit, 12-bit, 16-bit, or 24-bit. The measurement accuracy and sampling interval of charge / discharge data can be set arbitrarily depending on the feature related to the material degradation information (hereinafter referred to as electrochemical feature) to be acquired. For example, if the direct current resistance (DCIR) of a storage battery is to be acquired, a short sampling interval is preferable, such as 1 msec or more and 100 msec or less. Furthermore, when measuring at a low C rate, such as the SoC-OCV curve of a storage battery, the sampling interval may be relatively long, such as 100 msec or more and 10 sec or less.

[0030] The end BMS 230 acquires charge / discharge data of the storage battery measured by the sensor unit and transmits the data directly to the local BMS 300 via a wired or wireless connection. Alternatively, the charge / discharge data may be transmitted to the local BMS 300 via the network N.

[0031] <Local BMS> Fig. 4 is a diagram showing a schematic configuration of the local BMS 300. As shown in Fig. 4, the local BMS 300 includes an acquisition unit 310, a transmission unit (transmission / reception unit 320), a reception unit (transmission / reception unit 320), a storage unit 330, a calculation unit 342, a control unit 344, and an output unit 350. The acquisition unit, transmission / reception unit, storage unit, calculation unit, control unit, output unit, etc. are connected to each other via a bus, an interface unit, etc.

[0032] Acquisition unit: The acquisition unit 310 acquires charge / discharge data of the storage battery from the end BMS 230. The acquisition unit 310 may acquire the charge / discharge data directly from the end BMS 230 via a wired or wireless connection, or may acquire (receive) the charge / discharge data from the end BMS 230 via the network N. The acquired charge / discharge data is stored in the storage unit 330.

[0033] Transmitter / receiver: The transmitter / receiver 320 has a wired communication interface circuit that complies with a communication protocol such as TCP / IP (Transmission Control Protocol / Internet Protocol), and is communicatively connected to the network N in accordance with a communication standard such as Ethernet (registered trademark). The transmitter / receiver 320 may also have an antenna that transmits and receives wireless signals, and a wireless communication interface circuit that complies with a communication protocol such as wireless LAN, and be communicatively connected to the network N in accordance with a communication standard such as wireless LAN.

[0034] The transmitter (transmitter / receiver 320) transmits the charge / discharge data of the storage battery to the cloud BMS 400, which serves as a diagnostic device, via the EMS 240. The transmitter (transmitter / receiver 320) may transmit the charge / discharge data of the storage battery to the cloud BMS 400 without going through the EMS. When the calculator 342 of the local BMS 300 calculates the SoH, SoC, SoP, SoS, etc., the transmitter may transmit the calculated SoH, etc. to the cloud BMS 400, either via or without going through the EMS.

[0035] The receiving unit (transmitting / receiving unit 320) can receive charge / discharge condition commands and the like transmitted from the cloud BMS 400 via or without the EMS. The received charge / discharge condition commands and the like are stored in the storage unit 330 and are used to control the charge / discharge conditions of the storage battery. Transmission and reception include transmission and reception via a wired connection or a wireless connection, and transmission and reception via the network N (the same applies below).

[0036] Storage Unit: The storage unit 330 includes a memory device such as a RAM (Random Access Memory) or a ROM (Read Only Memory), a fixed disk device such as a hard disk, or a portable storage device such as a flexible disk or optical disk. The storage unit 330 can store computer programs used for various processes of the local BMS 300, acquired charge / discharge data, calculated SoH, SoC, SoP, and SoS, etc. The computer programs may be installed into the storage unit 330 from a computer-readable portable recording medium using a known setup program or the like. The portable recording medium may be, for example, a CD-ROM (Compact Disc Read Only Memory) or a DVD-ROM (Digital Versatile Disc Read Only Memory). The computer programs may be stored on a recording medium owned by a specific server and installed via the network N.

[0037] Processing Unit: The processing unit 340 operates based on a program stored in advance in the storage unit 330. The processing unit 340 is, for example, a CPU. A DSP (digital signal processor), an LSI (large scale integration), an ASIC (application specific integrated circuit), an FPGA (field-programmable gate array), or the like may be used as the processing unit 340. The processing unit 340 may include a calculation unit 342 and a control unit 344, and is connected to an acquisition unit, a transmission / reception unit, a storage unit, an output unit, and the like, and controls each functional unit.

[0038] The calculation unit 342 may calculate the SoH, SoC, SoP, SoS, etc. from the charge / discharge data of the storage battery. In this aspect, the calculated SoH, etc. are transmitted to the cloud BMS 400, which serves as a diagnostic device, and are used to calculate diagnostic results in the cloud BMS 400. The method of calculating the SoH will be described in detail later.

[0039] The control unit 344 controls the storage battery based on the charge / discharge condition command received from the cloud BMS 400. More specifically, the control unit 344 controls the ESS by transmitting the received charge / discharge condition command to the PCS 100 or the end BMS 230 via the output unit 350. The control method will be described in detail later.

[0040] Output Unit: The output unit 350 outputs (transmits) charge / discharge condition commands from the control unit 344 to the PCS 100 and / or the end BMS 230. When the local BMS 300 receives the State of Health (SoH), material degradation information, the degradation state of the storage battery, labeling results, and the current and future risk of lithium deposition X (hereinafter, "diagnosis results, etc.") from the cloud BMS 400, the output unit 350 may output (transmit) the received diagnosis results, etc. to a user's terminal device 500 having a display unit. The diagnosis results, etc. can be transmitted to the terminal device 500 from the transmission unit via a network. The local BMS 300 may further have a display unit (not shown), in which case the output unit 350 may output (display) the diagnosis results, etc. on the display unit.

[0041] The display unit has a display such as a liquid crystal display or an organic electroluminescence (EL) display, and an interface circuit that outputs image data to the display, and can visually display diagnostic results, etc. This allows the deterioration state of the storage battery and the risk of lithium deposition to be immediately grasped at the ESS operation site.

[0042] <Cloud BMS> Figure 5 is a diagram showing the schematic configuration of a cloud BMS 400, which functions as a diagnostic device and a server and is combined with the local BMS 300 of Figure 4. The cloud BMS 400 functions as a diagnostic device and a server, generating and updating a trained model, diagnosing degradation using the trained model, and outputting the degradation diagnosis. The cloud BMS 400 includes an acquisition unit (a receiver in the transceiver unit 420), a memory unit 430, a diagnosis unit 442, a calculation unit 443, a prediction unit 444, a control unit 445, and an output unit (a transmitter in the transceiver unit 420). The transceiver unit, memory unit, diagnosis unit, calculation unit, prediction unit, and control unit are interconnected via a bus, an interface unit, and the like. The server may be a single server or multiple servers interconnected via a network N.

[0043] Transmitter / receiver: The transmitter / receiver 420 has a wired communication interface circuit that complies with a communication protocol such as TCP / IP (Transmission Control Protocol / Internet Protocol), and is communicatively connected to the network N in accordance with a communication standard such as Ethernet (registered trademark). The transmitter / receiver 420 may also have an antenna that transmits and receives wireless signals, and a wireless communication interface circuit that complies with a communication protocol such as wireless LAN, and be communicatively connected to the network N in accordance with a communication standard such as wireless LAN.

[0044] Acquisition unit: The acquisition unit (reception unit in the transmission / reception unit 420) acquires (receives) charge / discharge data of the storage battery from the local BMS 300. The acquisition unit may also receive the SoH, SoC, SoP, SoS, etc. calculated (estimated) by the local BMS 300 based on the charge / discharge data. The acquired charge / discharge data and SoH, etc. are stored in the storage unit 430.

[0045] Storage unit: The storage unit 430 includes a memory device such as a RAM (Random Access Memory) or a ROM (Read Only Memory), a fixed disk device such as a hard disk, or a portable storage device such as a flexible disk or an optical disk. The storage unit 430 stores the trained model 431, computer programs used for various processes of the cloud BMS 400, acquired charge / discharge data, calculated SoH, SoC, SoP, and SoS, an integrated database 432, tables, and the like. The computer program may be installed into the storage unit 430 from a computer-readable portable recording medium using a known setup program or the like. The portable recording medium is, for example, a CD-ROM (compact disc read only memory), a DVD-ROM (digital versatile disc read only memory), or the like. The computer program may be stored in a recording medium owned by a predetermined server and installed via the network N.

[0046] The trained model 431 includes a trained model A that can identify one or more pieces of material deterioration information when charge / discharge data is input. The trained model 431 also includes a current SoH and a current lithium precipitation risk Xt 1 The system further includes a trained model B that outputs the future transition of the lithium deposition risk X when information including past charge / discharge data and past transitions of SoH is input. Trained models A and B may each be composed of a single trained model, or preferably, are composed of a combination of multiple trained models that differ for each type of input charge / discharge data.

[0047] The integrated database 432 includes a dataset A containing one or more electrochemical features obtained through experiments as training data and one or more pieces of material degradation information associated therewith. The dataset A is used to generate and update the trained model A. The integrated database 432 also includes a dataset B containing charge / discharge data, a transition in SoH, and a transition in the risk of lithium precipitation, all of which are linked together as training data. The dataset B is used to generate and update the trained model B. The integrated database 432 is updated as needed by accumulating charge / discharge data of sampled storage batteries in various states of degradation, material degradation information, SoH, and risk of lithium precipitation X obtained by analyzing the sampled storage batteries. The accuracy of the trained model 431 improves as the amount of database data collected in the integrated database 432 increases. Details of the integrated database 432 will be described later.

[0048] Processing Unit: The processing unit 440 operates based on a program stored in advance in the storage unit 430. The processing unit 440 is, for example, a CPU. A DSP (digital signal processor), an LSI (large scale integration), an ASIC (application specific integrated circuit), an FPGA (field-programmable gate array), or the like may be used as the processing unit 440. The processing unit 440 includes a generation unit 441, a diagnosis unit 442, a calculation unit 443, a prediction unit 444, and a control unit 445, and is connected to the transmission / reception unit, the storage unit, etc. to control each functional unit.

[0049] The generation unit 441 generates and re-trains the trained model 431 including trained models A and B. Details of the training process will be described later.

[0050] The diagnosis unit 442 diagnoses (identifies) material deterioration information of the storage battery by inputting charge / discharge data of the storage battery to the trained model A. When the trained model A receives the charge / discharge data, it extracts one or more electrochemical feature quantities from the charge / discharge data and can identify one or more pieces of material deterioration information that are highly relevant to the extracted electrochemical feature quantities. Details of the diagnosis process will be described later.

[0051] The calculation unit 443 calculates the degradation state of the storage battery based on information including the estimated SoH and material degradation information obtained from the trained model A. While conventionally, storage battery degradation diagnosis has been performed using parameters based solely on charge / discharge data such as the SoH, the diagnostic device disclosed herein combines the SoH with material degradation information within the storage battery, thereby enabling more accurate diagnosis of the degradation state of the storage battery. The SoH may be calculated by the calculation unit 443 of the cloud BMS 400 based on charge / discharge data of the storage battery acquired from the local BMS 300, or may be calculated by the calculation unit 342 of the local BMS 300. When the calculation unit 342 of the local BMS 300 calculates the SoH, the SoH is transmitted to the cloud BMS 400 together with the charge / discharge data. Details of the SoH calculation will be described later.

[0052] The calculation unit 443 can calculate the lithium deposition risk X of the storage battery based on information including material degradation information inside the storage battery. The material degradation information of the storage battery includes material degradation information of at least one material selected from the group consisting of the positive electrode, negative electrode, separator, and electrolyte of the storage battery. The calculation unit 443 receives information including the material degradation information output from the diagnosis unit and calculates the lithium deposition risk of the corresponding storage battery as a numerical value (lithium deposition risk X). Information used to calculate the lithium deposition risk X can include, in addition to material degradation information, charge / discharge data, material information of the storage battery, and operating conditions of the ESS. The material information of the storage battery includes material information of at least one material selected from the group consisting of the positive electrode, negative electrode, separator, and electrolyte of the storage battery. Details of the method for calculating the lithium deposition risk X will be described later.

[0053] The calculation unit 443 can label the degradation state of the storage battery based on one or more predetermined thresholds for the estimated SoH and the lithium deposition risk X. While conventionally, battery degradation diagnosis was performed using parameters based solely on charge / discharge data, such as the SoH, the diagnostic device of the present disclosure can divide the degradation state of the storage battery into multiple quadrants based on the SoH and information on material degradation within the storage battery, and label the state according to the degree of degradation. This aspect therefore has the advantage of making it possible to more accurately and easily grasp the degradation state of the storage battery. Details of the labeling process will be described later.

[0054] The prediction unit 444 calculates the current SoH and the current lithium deposition risk Xt 1 , as well as information including past charge / discharge data and past changes in SoH, are input to the trained model B to predict a future change in the risk of lithium deposition X. Then, the calculation unit 443 can label the deterioration state of the storage battery based on the information including the future change in the risk of lithium deposition X. Details of the prediction process and labeling process will be described later.

[0055] The control unit 445 controls the charge / discharge conditions of the storage battery based on the diagnosis results. The diagnosis results include the deterioration state of the storage battery and may also include the result of labeling the storage battery. The control unit 445 can generate a charge / discharge condition command, including maintaining, changing, or suspending the charge / discharge conditions, based on the diagnosis results. Details of the control process will be described later.

[0056] Output unit: The output unit (transmitter in the transceiver unit 420) outputs (transmits) the calculated degradation state as a diagnostic result to the user's terminal device 500. The diagnostic result includes the degradation state of the storage battery and may further include the State of Health (SoH), material degradation information, labeling results, and current and future lithium deposition risk X ("diagnosis results, etc."). The output unit may transmit the diagnostic results, etc. to the EMS 240 and / or the local BMS 300.

[0057] The output unit (transmitter / receiver 420) transmits the charge / discharge condition command generated by the control unit 445 to the local BMS 300, either via or without the EMS 240. The local BMS 300 controls the end BMS 230 based on the charge / discharge condition command, thereby controlling the charge / discharge conditions of the storage battery.

[0058] <Terminal Device> The terminal device 500 shown in FIG. 1 is a user terminal and may be, for example, an information processing device such as a personal computer, notebook PC, tablet PC, or multi-function mobile phone (so-called smartphone). The terminal device 500 includes a receiving unit and a display unit (not shown), which are interconnected via a bus and an interface unit. The receiving unit can receive diagnostic results, including the deterioration state of the storage battery, as well as State of Health (SoH), material deterioration information, labeling results, and current and future lithium deposition risk X (hereinafter referred to as "diagnosis results"), transmitted from the local BMS 300 or the cloud BMS 400. The display unit has a display, such as a liquid crystal display (LCD) or organic electroluminescence (EL) display, and an interface circuit that outputs image data to the display, allowing the diagnostic results to be visually displayed on the display. This allows the terminal device 500 to instantly notify the user of the diagnostic results of the storage battery installed in the currently operating ESS.

[0059] Users include ESS operators, ESS suppliers, battery or battery material manufacturers, and battery recyclers. By instantly notifying users of diagnostic results, users can immediately identify issues related to the ESS operating conditions and battery specifications without dismantling and inspecting the batteries. As a result, each user can enjoy the following benefits, for example: ESS operators can develop highly predictable maintenance plans based on the battery degradation state, enabling safe and efficient ESS operation. Furthermore, even if the diagnostic results indicate a battery that requires immediate replacement and would otherwise require immediate ESS operation shutdown, ESS operation can be continued and planned maintenance can be performed by changing the charging and discharging conditions to safe conditions that take the battery degradation state into account. For example, if an ESS is used to operate equipment in a factory, etc., the operational efficiency of the ESS can be maximized by aligning the ESS battery replacement period with the equipment's scheduled repair period. Furthermore, for example, the operational efficiency of an ESS can be maximized by controlling the optimal charge / discharge conditions of an operating storage battery in accordance with the market conditions of new or reused batteries to be replaced (the timing of procurement of batteries with the desired performance and quantity). Furthermore, for example, the profits from the difference in electricity rates can be maximized by controlling the optimal charge / discharge conditions of an operating storage battery in accordance with fluctuations in the electricity rate market. ESS suppliers can devise ESS designs and operating conditions that extend the life of storage batteries based on diagnostic results, feed back diagnostic results into the development of new ESSs, and identify ESSs or storage batteries that may be subject to recall more quickly. Furthermore, reducing the risk of lithium deposition in ESSs and / or extending their lifespans can reduce the frequency of ESS maintenance and prevent battery overloading, which is expected to reduce the cost of storage batteries used in ESSs. Manufacturers of storage batteries or storage battery materials can predict the delivery dates of storage batteries or storage battery materials in advance based on maintenance plans, allowing them to supply storage batteries or storage battery materials in a planned manner. Furthermore, the diagnostic results can be fed back into the development of new storage batteries or storage battery materials, enabling each customer to be provided with products of higher value.Storage battery recyclers can determine in advance the delivery dates of batteries to be recycled based on a maintenance plan, as well as the delivery dates of replacement reuse batteries, enabling them to create efficient recycling plans. Furthermore, they can check the safety and material degradation information of the batteries to be recycled in advance, enabling safer and more efficient recycling work.

[0060] 6 is a flowchart showing an example of the operation of the learning process of the trained model A. Note that the flow of the operation described below is based on a program stored in advance in the storage unit 430 of the cloud BMS 400 and is executed mainly by the generation unit 441 included in the processing unit 440 of the cloud BMS 400 in cooperation with each element.

[0061] First, the cloud BMS 400 acquires charge / discharge data and material degradation information as training data for generating the trained model A (step S101). The training data preferably includes charge / discharge data and material degradation information for the same type of storage battery as that used in the ESS to be diagnosed and controlled. The charge / discharge data and material degradation information as training data can be obtained by charging and discharging storage batteries in various degradation states in advance, for example, through accelerated experiments, dismantling the storage batteries, and visually, chemically, and physically analyzing each component, such as the positive electrode, negative electrode, separator, and electrolyte. Alternatively, the training data can be obtained by receiving, from the local BMS, charge / discharge data for storage batteries in various degradation states sampled from the ESS to be diagnosed and controlled, and then experimentally analyzing the sampled storage batteries to obtain material degradation information. Details of the experiment will be described later.

[0062] Next, the electrochemical feature quantities of the charge / discharge data and the material degradation information are compiled into a database (step S102). More specifically, the electrochemical feature quantities are extracted from the charge / discharge data of the storage battery, and dataset A is created by associating the extracted electrochemical feature quantities with material degradation information obtained by dismantling and analyzing the storage battery. In the integrated database 432, dataset A is recorded as one or more tables in which the electrochemical feature quantities and material degradation information are integrated.

[0063] Finally, a trained model A is generated based on the created table and stored in the storage unit 430 (step S103), thereby completing the series of steps. The trained model A can be generated by pre-training a machine learning model using supervised learning, such as a neural network or a support vector machine. More specifically, when charge / discharge data is input, one or more electrochemical feature quantities are extracted from the charge / discharge data, and the model is pre-trained to identify one or more pieces of material degradation information that are highly correlated with the extracted electrochemical feature quantities. When electrochemical feature quantities previously extracted from charge / discharge data are input, the model may be pre-trained to identify one or more pieces of material degradation information that are highly correlated with the input electrochemical feature quantities. Details of the method for extracting electrochemical feature quantities will be described later. It is preferable to generate a different trained model for each type of input charge / discharge data and combine multiple trained models to construct a trained model A that can comprehensively identify one or more pieces of material degradation information from a combination of one or more pieces of charge / discharge data.

[0064] 7 is a flowchart showing an example of the operation of the learning process of the trained model B. The operation flow described below is based on a program stored in advance in the storage unit 430 of the cloud BMS 400 and is executed mainly by the generation unit 441 included in the processing unit 440 of the cloud BMS 400 in cooperation with each element.

[0065] First, the cloud BMS 400 acquires charge / discharge data, an SoH estimated from the charge / discharge data, and a lithium deposition risk X calculated from the charge / discharge data through a diagnostic process using the trained model A, which serve as training data for generating the trained model B (step S201). Details of the SoH estimation and the calculation of the lithium deposition risk X will be described later. The training data preferably includes charge / discharge data, etc., of the same type of storage battery used in the ESS to be diagnosed and controlled. The charge / discharge data, etc., as training data can also be acquired by previously charging and discharging storage batteries that have been placed in various states of degradation through an accelerated experiment, etc. Alternatively, the charge / discharge data, etc., as training data can also be obtained by receiving, from the local BMS, charge / discharge data of storage batteries in various states of degradation sampled from the ESS to be diagnosed and controlled.

[0066] Next, the acquired charge / discharge data, SoH, and lithium deposition risk X are compiled into a database (step S202). More specifically, charge / discharge data, changes in SoH, and changes in lithium deposition risk X at various measurement points (degraded states) from the initial stage of the storage battery until the end of its life (replacement) are acquired, and these are correlated to create dataset B. In the integrated database 432, dataset B is recorded as one or more integrated tables that integrate the charge / discharge data, changes in SoH, and changes in lithium deposition risk X.

[0067] Finally, a trained model B is generated based on the created table and stored in the storage unit 430 (step S203), thereby completing the series of steps. The trained model B can be generated by pre-training a machine learning model using supervised learning, such as a neural network or a support vector machine. More specifically, when information including the SoH and lithium deposition risk X at a certain point in time, as well as charge / discharge data from that point in time and the transition of the SoH, is input, the trained model is pre-trained to output the lithium deposition risk X from that point in time. It is preferable to generate a different trained model for each type of input charge / discharge data and combine multiple trained models to construct a trained model B that can comprehensively output the future lithium deposition risk X from one or a combination of multiple charge / discharge data.

[0068] <<Integrated Database>> The integrated database is a database including a data set A that integrates electrochemical feature quantities of charge / discharge data of a storage battery and corresponding material degradation information, and a data set B that integrates information including charge / discharge data, a transition in SoH, and a transition in lithium deposition risk X. An example of a method for creating the integrated database will be described below for the case where an LiB is used as the storage battery.

[0069] <Dataset A> Extraction of electrochemical features: First, LiBs in various states of degradation, whose usage history, such as charging and discharging, is known in advance, are prepared. To acquire the training data, preferably, the same type of LiBs as those used in the ESS to be diagnosed and controlled, more preferably, LiBs actually used in the ESS to be diagnosed and controlled, are used. The degradation state may be adjusted in various ways, such as through accelerated experiments. These LiBs are subjected to charging and discharging operations, and the resulting charging and discharging data (e.g., changes in current, voltage, and temperature over time) are acquired. The charging and discharging data of the storage battery can be obtained by electrically charging and discharging the storage battery using a commonly available charging and discharging device and recording the changes in voltage, current, and temperature over time obtained in response. The charging and discharging device can have voltage and current ranges that match the voltage (V) and current capacity (Ah) of the storage battery to be diagnosed and controlled, and can arbitrarily set the charging and discharging conditions. Examples include the ACD-01 charge / discharge test equipment 5V / 10A (1A)-12ch system (charge / discharge unit model: ACDS-10APS (01)-05N) manufactured by Asuka Electronics Co., Ltd., and the regenerative charge / discharge test equipment [5V 360A 1.8kW 6ch] manufactured by Fujitsu Telecom Network Limited.

[0070] The measured charge / discharge data includes, for example, the voltage change over time when the storage battery is charged to an arbitrary SoC and then immediately paused (dOCV / dt-t plot), the voltage change over time during pulse charging and / or discharging at a constant current (V-t plot), the current change over time observed during constant voltage charging (I-t plot), the charge capacity and discharge capacity at a constant current or constant power, measurement data obtained by the AC impedance method, and mathematically processed data thereof. Next, the obtained charge / discharge data is mathematically converted to extract features (electrochemical features) related to material degradation information of the LiB. Examples of electrochemical feature extraction are given below.

[0071] (1) Change in voltage over time when the battery is rested immediately after charging As a method for electrochemically detecting Li metal deposition in LiB, the method described in Zachary M. Konz, et al., "Detecting the Onset of Lithium Plating and Monitoring Fast Charging Performance with Voltage Relaxation", American Chemical Society Energy Letters, (2020), 5(6), pp.1750-1757 can be used. In this method, LiB discharged to SoC = 0% is charged at a constant current (for example, C rate = 3C), charging is stopped at an arbitrary SoC, and the time course of voltage relaxation is measured when the LiB is rested in a no-load state. If Li deposition has occurred, the negative electrode contains Li metal and Li(C) intercalated in graphite. 6 Li) coexist, and at that time, a plateau occurs as a mixed potential of the negative electrode. As a result, a plateau occurs in the cell voltage (hereinafter also referred to as open circuit voltage, or OCV (Open Circuit Voltage)) in a no-load state immediately after charging. Using this OCV time transition (hereinafter also referred to as OCV-t plot), an electrochemical feature quantity related to Li metal deposition can be obtained. For example, a dOCV / dt-t plot is created in which dOCV / dt, which is the time derivative of the OCV-t plot, is plotted on the vertical axis and time t (sec) is plotted on the horizontal axis. As a result, when a plateau occurs in the OCV-t plot, the dOCV / dt-t plot will have a profile with a maximum value. This maximum value (d 2 OCV / dt 2 The presence or absence of the maximum value (d = 0 point) and the time t (sec) at the maximum value can be used as the electrochemical feature. 2 OCV / dt 2 The presence of a peak (point where t = 0) indicates that a plateau has occurred, i.e., that Li metal has precipitated. In addition, the longer the time t (sec) at the maximum value, the longer the plateau, i.e., the more Li metal has precipitated.

[0072] (2) Time-Varying Voltage During Pulse Charge and / or Discharge at Constant Current By analyzing plots of voltage drop (e.g., V-t plots) observed when applying pulse charge and / or discharge at constant current to LiB, information suggesting the state of Li metal deposition can be obtained. When LiB is subjected to constant-current pulse charge or pulse discharge at a relatively high C rate, a sudden fluctuation in voltage (overvoltage) occurs according to Ohm's law (V = IR) due to the internal resistance of LiB. LiB overvoltages can be broadly divided into resistance overvoltages derived from electronic resistance, activation overvoltages due to chemical reactions occurring at the interface between the electrode active material and the electrolyte, and concentration overvoltages generated by the concentration gradient of ions moving through the electrolyte or electrode active material. Because these three types of overvoltages have different reaction rates, when plotted as voltage versus time, it is observed that the overvoltage changes depending on the time period. Roughly speaking, the time periods are divided into short periods of 0.1 seconds or less for resistance overpotential, short periods of 0.1 to several seconds for activation overpotential, and long periods of several to several thousand seconds for concentration overpotential. The reaction in which Li metal deposits mainly falls under the activation overpotential, so by analyzing the transition of overpotential in the time range of 0.1 to several seconds, it is possible to consider the deposition of Li metal.

[0073] (3) Time change of current value observed during constant voltage charging The state of Li metal deposition can be analyzed by analyzing the time dependence of the current observed when LiB is charged at a constant voltage at a high C rate (hereinafter referred to as the It plot). When Li deposition occurs during constant voltage charging at a high C rate, the It plot shows a behavior in which the current value reaches a minimum value once, and then gradually increases with a delay. This is because the charge applied to LiB is consumed for Li metal deposition without being used to charge the positive and negative electrodes, temporarily reducing the convergence current value, and then the charge is charged to the positive and negative electrodes with a delay, gradually increasing the convergence current value. The depth of the minimum value in the It plot (minimum current value: A) and its position (sec) can be used as electrochemical feature quantities. Note that the voltage during constant voltage charging when determining Li metal deposition is preferably performed in a high SoC region where the negative electrode potential is low.

[0074] (4) Charge / Discharge Efficiency (Coulomb Efficiency) The charge / discharge efficiency during low-rate charge / discharge can be used as an electrochemical feature. In this method, the charge / discharge efficiency of a target LiB can be obtained by, for example, dividing the discharge current capacity (Ah) obtained by discharging a LiB adjusted to SoC = 100% at a constant current of 0.25 C by the charge current capacity (Ah) obtained by charging a LiB adjusted to SoC = 0% at a constant current of 0.25 C. If a large amount of the charged current capacity is consumed by side reactions that do not actually contribute to charging, the discharge capacity will be small, resulting in a decrease in charge / discharge efficiency. An example of a side reaction is the formation reaction of Li dendrites. Li dendrites may penetrate the separator, which serves to separate and insulate the positive and negative electrodes, and cause an internal short circuit. Therefore, the charge / discharge efficiency reflects information about internal short circuits in the LiB and is therefore an important electrochemical feature for determining the safety of the LiB. The lower the charge / discharge efficiency, the more Li metal is deposited and the more Li dendrites are formed.

[0075] (5) Nyquist Plot Electrochemical feature values ​​can be obtained from a Nyquist plot obtained by electrochemical impedance spectroscopy (EIS). For example, a FRA module for cell test systems (module code: 1455) manufactured by SOLARTRON ANALITICAL can be used to measure EIS. In this method, a Nyquist plot is created by sweeping a LiB adjusted to an arbitrary SoC, for example, in a range from 20 kHz to 0.1 Hz, and acquiring the response data. The Nyquist plot can obtain inductance (H), capacitance (F), and DC resistance (Ω) corresponding to an equivalent circuit model of the LiB, which can be attributed to the material components inside the LiB. For example, if Li metal is precipitated, the component attributed to Li metal can be extracted as its parasitic impedance, and this can be used as an electrochemical feature value.

[0076] As described above, there are several methods for acquiring electrochemical characteristic quantities related to the safety of LiB, particularly Li metal deposition, but the electrochemical characteristic quantities are not limited to these.

[0077] Obtaining Material Degradation Information: Material degradation information as training data can be obtained by dismantling LiBs in various states of degradation after completing the charge / discharge operations required to obtain the electrochemical features described above, and extracting and analyzing the internal materials (positive electrodes, negative electrodes, electrolytes, separators, etc.). It is preferable to collect information on various degradation states for each type of storage battery prepared for training data acquisition. Therefore, it is preferable to prepare a large number of storage batteries in various states of degradation by conducting accelerated tests such as cycle tests, high-temperature storage tests, and float tests. Furthermore, material degradation information can be inferred with even greater accuracy by using the results of analyzing the constituent materials of LiBs obtained by acquiring and dismantling LiBs actually used in ESS operation. For example, when LiBs are charged and discharged due to operations such as inertial force or frequency adjustment, response to momentary voltage drops, peak shifting, or arbitrage in a stationary energy storage system, defective LiBs can be extracted, dismantled, and analyzed to obtain material degradation information.

[0078] The method for dismantling an LiB may be any method depending on the shape of the LiB (cylindrical, laminated, square, etc.) Generally, the LiB can be dismantled by peeling off the exterior material with general-purpose tools (nippers, scissors, etc.), removing the electrode assembly (a laminated or wound structure of a positive electrode, a negative electrode, and a separator) inside, and further separating the electrode assembly into its individual components.

[0079] Prior to analysis, the disassembled positive electrode, negative electrode, and separator are preferably subjected to a pretreatment process in which they are washed with a solvent to remove any deposits not immobilized on the surface. The solvent may be any solvent capable of washing away the electrolyte adhering to the surface, and carbonate solvents such as dimethyl carbonate, ethyl methyl carbonate, and diethyl carbonate are preferred. For example, the washing method involves immersing the target sample in diethyl carbonate solvent in an amount 50 to 100 times the weight of the target sample for 10 minutes or more, then replacing the solvent and immersing the target sample again. The target sample is then removed from the diethyl carbonate and vacuum dried (temperature: 0 to 200°C, pressure: 0 to 20 kPa, time: 1 to 40 hours, under conditions such that the remaining diethyl carbonate in the target sample is 1% by mass or less).

[0080] Next, the pretreated positive electrode, negative electrode, and separator are subjected to visual, physical, and chemical analysis, and the obtained information can be used as material degradation information for the LiB. For example, in visual analysis, surface images of the positive electrode, negative electrode, and separator are obtained using a digital camera, a scanning electron microscope (SEM), a transmission electron microscope (TEM), or the like, and the images are analyzed using binarization, color analysis, or the like to obtain information on the area, distribution, shape, and the like of by-products formed due to the deterioration of the storage battery. Examples of physical analysis include, for example, observing the weight of by-products formed on the positive electrode, negative electrode, and separator as a difference from the weight of the material before deterioration. Examples of chemical analysis include structural analysis of the active materials, etc., using X-ray diffraction (XRD), and analysis of the chemical state of the materials forming the positive electrode, negative electrode, and separator using X-ray photoelectron spectroscopy (XPS), nuclear magnetic resonance (NMR), electron spin resonance (ESR), or the like.

[0081] When LiB is used as the storage battery, the material degradation information preferably includes information on the deposition of Li metal on the negative electrode, which is directly related to the safety of LiB. More specifically, the material degradation information includes the area of ​​deposited lithium metal analyzed from an image of the negative electrode surface of the storage battery, the shape of the deposition region, the deposition weight, distribution state and color tone, the shape of lithium metal observed by cross-sectional SEM observation of the negative electrode, the chemical shift value (eV) and / or its area obtained from the Li(s1) spectrum obtained by XPS analysis, 7 It includes chemical states expressed by chemical shift values ​​(ppm) and / or their areas obtained from Li-NMR spectra, and g values ​​and / or line widths obtained from ESR spectra.

[0082] (1) Image analysis of Li metal formed on a negative electrode As an example of material deterioration information, an image analysis of Li metal formed on a negative electrode is given. However, the image analysis is not limited to the state of Li metal on a negative electrode, and can be performed on any image, such as the collapse of a positive electrode active material, SEI formed on the surface of a negative electrode, or black spots on a separator.

[0083] First, the negative electrode disassembled and pretreated by the above method is photographed with a camera that detects visible light, such as a digital camera, a single-lens reflex camera, or a smartphone camera. Graphite is generally used as the negative electrode active material for LiB, and graphite doped with Li ions (C 6Li) is generally yellow, but the areas where Li metal is deposited generally exhibit a white, gray, or silver color with a metallic luster. Next, by analyzing the photographed image of the negative electrode surface, information such as the presence or absence of Li metal deposition, the area of ​​the deposited area, and the quality of the deposited Li metal is obtained. For example, as the form in which Li metal is deposited on the negative electrode surface, island-like Li metal deposition areas may be formed on the surface of the negative electrode. In this case, an image is obtained in which island-like white areas exist on a yellow background. The image is processed by binarization analysis or the like, and information such as the presence or absence of Li metal deposition areas, the deposition area, and the deposition shape can be analyzed from the color difference and area. Furthermore, information regarding the quality of the deposited Li metal can be obtained from the color tone of the Li metal deposition area. For example, plate-shaped Li metal has a bright color tone with the inherent metallic luster of Li metal when viewed with the naked eye, but Li dendrites, which can cause internal short circuits, have a fine and complex dendritic structure and therefore appear dark gray when viewed with the naked eye. Information on the precipitated Li metal obtained from these areas and colors can be obtained by computer analysis at a level equivalent to that of visual judgment using a trained model that has undergone machine learning or deep learning on the acquired images and images processed by binary analysis, color analysis, etc.

[0084] Since image analysis is affected by the conditions under which the image is captured, it is desirable to keep the conditions constant during capture. The capture conditions include, for example, the charge state of the LiB when the negative electrode is obtained, the conditions for preparing the negative electrode, such as the cleaning and drying state of the negative electrode, and the conditions during capture, such as the color and luminous intensity of the light source in the capture environment, magnification and contrast, color correction, and the distance and angle between the camera and the negative electrode.

[0085] (2) Scanning Electron Microscope (SEM) SEM observation allows for the microscopic morphology of by-products formed on the surface of a material. Furthermore, by combining it with energy dispersive X-ray analysis (EDX), qualitative elemental analysis of the structures formed on the surface can be performed. Furthermore, by mapping the elemental analysis results, the spatial distribution of elements on the surface can be observed. An example of material degradation is the morphology of Li metal formed on the surface of the negative electrode. When LiB is charged, Li ions cannot be absorbed into the negative electrode active material, and Li metal is precipitated on the negative electrode surface. The morphology of the precipitated Li metal affects the risk of internal short-circuiting of LiB. Generally, if the Li metal has a dendritic morphology, there is a higher risk of short-circuiting the positive and negative electrodes through the microscopic holes in the separator. Therefore, by understanding the morphology of Li metal through SEM observation, the risk of internal short-circuiting can be estimated.

[0086] A specific example of the SEM observation method is shown below. First, gold is sputtered onto the surface of the negative electrode in a vacuum of 10 Pa to coat the surface with a gold film several nanometers thick. Next, an SEM image of the surface of the negative electrode active material layer can be taken under atmospheric exposure under the conditions shown below. (SEM Measurement Conditions) Measuring device: Hitachi High-Technologies Corporation, Field Emission Scanning Electron Microscope S-4700, Acceleration voltage: 1 kV, Emission current: 10 μA, Measurement magnification: 20,000x, Detector: Secondary electron detector, Electron beam incident angle: 90°. When taking the image, it is preferable to adjust the brightness and contrast so that no pixels in the SEM image reach the maximum brightness value and the average brightness value is within the range of 40% to 60% of the maximum brightness value.

[0087] (3) X-ray Photoelectron Spectroscopy (XPS) The bonding state of lithium can be determined by analyzing the electronic state of lithium using XPS. Measurements can be performed under the following conditions: X-ray source: monochromated AlKα, X-ray beam diameter: 100 μmφ (25 W, 15 kV), narrow scan pass energy: 58.70 eV, with charge neutralization, narrow scan sweep count: 10 times (carbon, oxygen), 20 times (fluorine), 30 times (phosphorus), 40 times (lithium), 50 times (silicon), and narrow scan energy step: 0.25 eV. It is preferable to clean the surface of the positive electrode by sputtering before XPS measurement. Sputtering conditions include, for example, an acceleration voltage of 1.0 kV, a 2 mm x 2 mm area, and 1 minute of sputtering (SiO 2 The surface of the positive electrode can be cleaned under the condition of a rate of 1.25 nm / min (equivalent to 1.25 nm / min). 2 or Li-C bond, and the peak at 55 to 60 eV is LiF, Li 2 CO 3 , Li x P.O. y F z (x, y, z are integers from 1 to 6), the C1s binding energy peak of 285 eV is assigned to a C-C bond, the 286 eV peak is assigned to a C-O bond, the 288 eV peak is assigned to COO, and the 290 to 292 eV peak is assigned to CO 3 2- , C—F bond, and the peak at O1s binding energy 527-530 eV is O 2- (Li 2 O), and the peak at 531-532 eV is CO, CO 3 , OH, PO x (x is an integer of 1 to 4), SiO x (x is an integer from 1 to 4), and the peak at 533 eV is attributed to C—O and SiO x (x is an integer of 1 to 4), the peak at 685 eV of F1s binding energy is attributed to LiF, the peak at 687 eV is attributed to C—F bond, Li x P.O. y F z(x, y, and z are integers from 1 to 6), PF 6 - The P2p binding energy is assigned to PO. x (x is an integer from 1 to 4), and the peak at 134 to 136 eV is PF x (x is an integer from 1 to 6), and for the Si2p binding energy, the peak at 99 eV is Si silicide, and the peak at 101 to 107 eV is Si x O y (x and y are any integers). When peaks overlap in the obtained spectrum, it is preferable to separate the peaks by assuming a Gaussian function or a Lorentzian function and assign the spectrum. The chemical state of the lithium present can be determined from the results of the measurement of the electronic state and the results of the ratio of the elements present obtained above.

[0088] (4) Nuclear Magnetic Resonance (NMR) NMR can provide chemical information about the components that make up the positive electrode, negative electrode, separator, and electrolyte. 7 A measurement method for Li will be described. First, the composite layer is collected from the negative electrode after the above-described pretreatment and weighed. Next, the obtained negative electrode composite layer is used as a sample and subjected to solid 7 Li-NMR measurement is performed. An example of the measurement device is ECA700 ( 7 The resonance frequency of Li-NMR is 272.1 MHz. Measurement is performed by the single pulse method under a room temperature environment, with the magic angle spinning rotation speed set to 14.5 kHz and the irradiation pulse width set to a 45° pulse. A 1 mol / L lithium chloride aqueous solution is used as the shift reference, and the shift position measured separately as an external standard is set to 0 ppm. When measuring the lithium chloride aqueous solution, the sample is not rotated, and the irradiation pulse width is set to a 45° pulse, and measurement is performed by the single pulse method. When measuring, a sufficient waiting time is taken between measurements, and the waiting time is set to 3000 seconds. The solid of the negative electrode composite layer obtained by this method 7The Li-NMR spectrum allows us to understand the chemical state of Li. For example, in the case of Li metal, which poses a risk of ignition, the chemical shift occurs around 270 ppm. On the other hand, in the case of a solid electrolyte interface (SEI), such as an SEI, which does not lead to the ignition of LiB, the shift occurs around 0 ppm.

[0089] Although some examples of specific methods for acquiring material deterioration information have been described above, material deterioration information is not limited to these items and methods.

[0090] Other information: Dataset A preferably includes information on each component used in the LiB, linked to the electrochemical feature amount and material degradation information described above. This allows the trained model A to identify material degradation information with even higher accuracy. Examples of component information include positive electrode manufacturing conditions and negative electrode manufacturing conditions. Positive electrode manufacturing conditions include the basis weight (g / m 2 ), the thickness (μm) of the positive electrode active material layer, the bulk density (g / cc) of the positive electrode active material layer, the type and amount (mass %) of the positive electrode active material in the positive electrode active material layer, the amount (mass %) of lithium carbonate in the positive electrode active material layer in the positive electrode precursor, the type and amount (mass %) of the binder in the positive electrode active material layer, the type and amount (mass %) of the conductive material in the positive electrode active material layer, and the type and amount (mass %) of the dispersant in the positive electrode active material layer. 2 ), the thickness (μm) of the negative electrode active material layer, the bulk density (g / cc) of the negative electrode active material layer, the type and amount (mass %) of the negative electrode active material in the negative electrode active material layer, the type and amount (mass %) of the binder in the negative electrode active material layer, the type and amount (mass %) of the conductive material in the negative electrode active material layer, and the type and amount (mass %) of the dispersant in the negative electrode active material layer.

[0091] It is preferable that the data set A includes design information of the LiB linked to the electrochemical feature quantity and material deterioration information described above. This allows the trained model A to identify material deterioration information with even higher accuracy. The design information includes, for example, the stacking form (flat type, wound type, etc.), reaction area (m 2 ), cell thickness (mm), separator thickness (μm), separator resistance (mΩ·m 2), the type and concentration of the salt in the electrolyte (M), the volume of the electrolyte (g), and the volume of excess electrolyte (g). 2 "Excess amount of electrolyte (g)" is the amount of electrolyte not retained in the electrode body.

[0092] Updating Dataset A: Dataset A can be updated at any time by receiving charge / discharge data from the local BMS 300 of storage batteries in various states of degradation that are actually in use and have been used in the ESS that is the target of diagnosis and control, and experimentally analyzing the sampled storage batteries to obtain material degradation information.

[0093] <Dataset B> First, LiBs in various states of degradation are prepared, whose usage history, such as charging and discharging, is known in advance. To acquire the training data, preferably, the same type of LiB as that used in the ESS to be diagnosed and controlled, more preferably, LiBs actually used in the ESS to be diagnosed and controlled, are used. The degradation state may be adjusted in various ways, such as through accelerated experiments. Charging and discharging operations are performed on these LiBs, and the resulting charging and discharging data (such as changes in current, voltage, and temperature over time) are acquired. The charging and discharging data for the storage battery can be acquired in the same way as Dataset A described above.

[0094] Next, the SoH is calculated based on the acquired charge / discharge data. The SoH can be calculated by a conventional method. For example, the storage battery is charged / discharged at a constant current and a constant voltage, the current capacity is measured, and the SoH is calculated as the retention rate [%] of the full charge capacity [Ah] at the time of degradation (= full charge capacity [Ah] at the time of degradation / initial full charge capacity [Ah] × 100) using the initial full charge capacity [Ah] as a reference. Details are as described below in <Calculation of SoH>.

[0095] Then, from the acquired charge / discharge data, material degradation information of the storage battery is output through diagnostic processing by the trained model A, and the lithium deposition risk X is calculated based on the output material degradation information. Details of the calculation of the lithium deposition risk X are as described below in "Calculation and Labeling Process of Lithium Deposition Risk X." The charge / discharge data, SoH, and lithium deposition risk X at various measurement points (degraded states) are measured and calculated to construct a dataset B.

[0096] Dataset B preferably includes information on each component used in the LiB, linked to the above-described charge / discharge data, SoH, and lithium precipitation risk X. This enables trained model B to predict the future lithium precipitation risk X with even higher accuracy. The information on the components is the same as that used in dataset A described above.

[0097] Dataset B preferably includes LiB design information linked to the above-described charge / discharge data, SoH, and lithium precipitation risk X. This enables trained model B to predict the future lithium precipitation risk X with even higher accuracy. The design information is the same as that used in dataset A described above.

[0098] Data set B can be updated at any time by receiving charge / discharge data from local BMS 300 of storage batteries in various states of deterioration that are actually in use and have been used in the ESS that is the target of diagnosis and control, and obtaining information on the progress of SoH and the progress of lithium deposition risk X.

[0099] <<Diagnostic Processing>> Fig. 8 is a flowchart showing an example of the operation of the diagnostic processing. An example of the operation of the diagnostic processing will be described below with reference to the flowchart shown in Fig. 8. Note that the flow of the operation described below is executed based on a program pre-stored in the storage unit 430 of the cloud BMS 400, mainly by the diagnostic unit 442 included in the processing unit 440 of the cloud BMS 400 in cooperation with each element.

[0100] First, the acquisition unit (transmitter / receiver 420) of the cloud BMS 400 acquires charge / discharge data of the storage battery from the local BMS 300 (step S301). The acquisition unit 420 may store the charge / discharge data in the storage unit 430. The diagnosis unit 442 extracts features (electrochemical features) related to material deterioration information from the charge / discharge data (step S302). The extraction of the electrochemical features may be performed by calculation using the trained model A. The diagnosis unit 442 inputs the charge / discharge data or the electrochemical features extracted from the charge / discharge data into the trained model A, and diagnoses (identifies) material deterioration information within the storage battery that is highly correlated with the electrochemical features of the charge / discharge data (steps S303 to S304). Then, the output unit (transmitter / receiver 420) outputs the diagnosed (identified) material deterioration information (step S305), thereby completing the series of steps.

[0101] The electrochemical feature is a feature that has a technical relationship with the material degradation information of the storage battery, and is obtained by mathematically processing the charge / discharge data (electrochemical data). More specifically, for example, when the time change of the voltage when the storage battery is rested immediately after being charged to an arbitrary SoC (dOCV / dt-t plot) is used as the charge / discharge data, the maximum value of dOCV / dt in the dOCV / dt-t plot, the local maximum value (d 2 OCV / dt 2 The presence or absence of a point where dI / dt = 0, and the time t (sec) at the maximum value are extracted. When the time change of voltage (V-t plot) during pulse charging and / or discharging at a constant current is used as charge / discharge data, the transition of overvoltage in the time range of 0.1 seconds to several seconds is extracted. When the time change of current value observed during constant voltage charging (It plot) is used as charge / discharge data, the minimum value (d 2 I / dt 2The presence or absence of a minimum current (A) and the time t (sec) at the minimum value are extracted. When the charge capacity and discharge capacity at a constant current or constant power are used as the charge / discharge data, the charge / discharge efficiency (Coulomb efficiency) and the like are extracted as the electrochemical feature. When measurement data obtained by the AC impedance method are used as the charge / discharge data, a group of parameters such as the resistance component (Ω) and / or capacitance component (F) derived from Li metal in the Nyquist plot are extracted as the electrochemical feature. The method for extracting the electrochemical feature will be described in detail later.

[0102] The trained model A comprehensively considers a combination of one or more electrochemical feature quantities as listed above and identifies one or more pieces of highly relevant material degradation information. The material degradation information is information relating to the degradation of components such as the positive electrode, negative electrode, separator, and electrolyte used in the storage battery. The material information includes, for example, visual, physical, and chemical information of the components. The material degradation information preferably includes information relating to the risk of lithium deposition. The information relating to the risk of lithium deposition includes the area of ​​deposited lithium metal (Li deposition area ratio (%)), the shape of the Li deposition region, the weight of Li deposition (mg / m), and the like, analyzed from an image of the surface of the negative electrode of the storage battery. 2 ), the distribution state and color tone of the Li deposition region, the shape of the Li metal observed by cross-sectional SEM observation of the negative electrode (Li morphology), the chemical shift value (eV) and / or its area obtained from the Li(s1) spectrum obtained by XPS analysis, 7 The information includes information such as chemical shift values ​​(ppm) and / or their areas obtained from the Li-NMR spectrum, and chemical states expressed by g-values ​​and / or line widths obtained from the ESR spectrum. Details of the material degradation information will be described later. More preferably, the material degradation information includes information such as the Li deposition area ratio (%), Li deposition weight (mg / m 2 ), and Li forms.

[0103] The material deterioration information is output (transmitted) from the transmitter (transmitter / receiver 420) to the user's terminal device 500 and / or the local BMS 300 via the network N. In addition, the diagnosed (specified) material deterioration information is used in the calculation process of the lithium deposition risk X, which will be described below.

[0104] <Calculation of the Degradation State of the Storage Battery> The calculation unit 443 can acquire (identify) one or more pieces of material degradation information from the diagnosis unit through the diagnostic processing operations described above, and calculate the degradation state of the storage battery based on the separately calculated SoH and material degradation information. The material degradation information is as described above, and preferably includes information on the risk of lithium deposition and information obtained by mathematically processing this information. Examples of calculation methods include weighting each of the SoH and material degradation information based on material information of the storage battery used in the ESS and future operating conditions of the ESS to index the degradation state of the storage battery, or labeling the degradation state of the storage battery based on one or more predetermined thresholds for each of the SoH and material degradation information, and combinations thereof.

[0105] For example, as described below, the calculation unit 443 labels the degradation state of the storage battery based on one or more predetermined thresholds for the lithium deposition risk X calculated based on the SoH and material degradation information. The lithium deposition risk X is weighted according to material information of the storage battery used in the ESS, future operating conditions of the ESS, etc.

[0106] 10 is a flowchart showing an example of the operation of the calculation and labeling process of the lithium deposition risk X. The operation flow described below is executed based on a program stored in advance in the storage unit 430 of the cloud BMS 400, mainly by the calculation unit 443 included in the processing unit 440 of the cloud BMS 400 in cooperation with each element.

[0107] First, the calculation unit 443 acquires (specifies) material deterioration information from the diagnosis unit through the operation of the diagnosis process described above (step S401). Next, the calculation unit 443 calculates the lithium precipitation risk X based on information including the specified material deterioration information (step S402). The lithium precipitation risk X is calculated based on the lithium precipitation risk X estimated from the material deterioration information. chem. , and the risk of lithium deposition X estimated from charge / discharge data elec.In addition, the lithium deposition risk X may be calculated based on the material information of the storage battery used in the ESS and the future operating conditions of the ESS. chem. and lithium deposition risk X elec. By weighting these factors, it is possible to reflect in the lithium deposition risk X the difference in safety depending on the type of storage battery and the likelihood of deterioration depending on the future charge / discharge pattern of the storage battery.

[0108] In this embodiment, the storage unit 430 stores the material deterioration information of the storage battery and the lithium deposition risk X estimated from the material deterioration information. chem. , charge / discharge data, and the risk of lithium precipitation X estimated from the charge / discharge data elec. The diagnostic unit 443 stores in advance one or more tables linking material information of the storage battery with a weighting factor corresponding to the material information, and information relating to operating conditions of the ESS with a weighting factor corresponding to the operating conditions. The calculation unit 443 can refer to the table and calculate the lithium deposition risk X of the corresponding storage battery from the material deterioration information of the storage battery input from the diagnostic unit. An example of a specific method for calculating the lithium deposition risk X is shown below.

[0109] Lithium Deposition Risk X due to Material Degradation Information chem. Calculation of lithium deposition risk X estimated from material degradation information chem. is the lithium deposition risk X corresponding to the material deterioration information of item n chem. (n) and make a comprehensive judgment. chem. The average value of (n) can be used. That is, it can be calculated by the following formula (1). n means the nth information of the material deterioration information output from the diagnosis unit, and the number of output items is 1 to n. max There are pieces.

[0110] The storage unit 430 stores the lithium deposition risk X corresponding to each item of the material deterioration information. chem. (n) is stored in a table in which the X chem.The average value of (n) can be calculated. The setting method can be based on empirical rules, but it can also be quantified by building a trained model based on a database that links the safety tests and ignition records of the target storage battery with material deterioration information. Table 1 below shows the lithium deposition risk X corresponding to each item of material deterioration information. chem. The following table 1 shows an example of a table in which the lithium deposition risk X (n) is set. chem. The types and number of items to be included in the average value can be set arbitrarily.

[0111]

[0112] Lithium deposition risk based on charge / discharge data X elec. Calculation of Lithium Deposition Risk X Inferred from Charge / Discharge Data elec. is the lithium deposition risk X corresponding to the electrochemical feature of item m extracted from the charge / discharge data. elec. (m) and make a comprehensive judgment. elec. The average value of (m) can be used. That is, it can be calculated by the following formula (2). m means the m-th information of the electrochemical feature amount output from the acquisition unit, and the number of output items is 1 to m max It is said that there are pieces.

[0113] The storage unit 430 stores the lithium deposition risk X for each item of the electrochemical characteristic amount. elec. (m) A preset table is stored, and the calculation unit refers to the table to calculate X elec. The average value of (m) can be calculated. The setting method can be based on empirical rules, but it can also be quantified by building a trained model based on a database that links safety tests and ignition records of the target storage battery with material deterioration information. Table 2 below shows the lithium deposition risk X corresponding to each item of the electrochemical feature. elec. The following table 2 shows an example of a table in which the lithium deposition risk X (m) is set. elec. The types and number of items to be included in the average value can be set arbitrarily.

[0114]

[0115] <Weighting factor setting> Lithium deposition risk X chem. and X elec. For example, when considering the risk of Li metal deposition on the negative electrode surface, the lithium deposition risk X may be calculated by weighting the above factors according to the type of storage battery and the future charge / discharge conditions of the ESS. chem. and X elec. , and a weighting coefficient α according to the information of the negative electrode of the storage battery. chem. and α elec. For example, the weighting coefficient α chem. is set as α, and the weighting coefficient α of the electrochemical feature elec. Assuming that is 1-α, as shown in the following formula (3), X based on material deterioration information chem. and X based on electrochemical features elec. can be weighted. α is a weighting coefficient depending on the type of negative electrode active material, and is 0<α<1. α can be set according to the type of negative electrode active material used in the target storage battery. As an electrochemical feature of the negative electrode, for example, a peak in the dQ / dV value occurs due to a discontinuous change in the negative electrode potential accompanying a phase transition of the negative electrode active material. When the negative electrode active material is graphite-based, this peak is easy to determine and the electrochemical feature is easy to capture, so the importance of the electrochemical feature is high. On the other hand, when the negative electrode active material is hard carbon-based and does not involve a phase transition, it is difficult to capture the electrochemical feature, and the importance of material degradation information is high. Therefore, when the negative electrode active material is hard carbon-based, the lithium deposition risk X based on material degradation information is lower than when it is graphite-based. chem. The lithium deposition risk X based on the electrochemical feature elec. Therefore, when focusing on the negative electrode active material as material information of the negative electrode, if the material is a hard carbon type, the weighting coefficient α of the material deterioration information can be set relatively large.

[0116] The weighting coefficient α can be set based on empirical rules, or it can be quantified by constructing a trained model based on a database that links the safety tests and fire records of the target storage battery with material degradation information. Table 3 below shows an example of a table in which weighting coefficients are set according to the type of negative electrode active material.

[0117]

[0118] Furthermore, a weighting factor can be set based on information about the positive electrode of the storage battery, for example, the type of positive electrode active material. When the weighting factor based on the type of positive electrode active material is β, the lithium deposition risk X can be calculated, for example, by the following formula (4). α is the same as in equation (3). β is a weighting coefficient depending on the type of positive electrode active material, and is 0<β<1. For example, in an LiB, if a high nickel-based material (such as NCM622 or NCM811) is used as the positive electrode active material, the weighting coefficient β can be set relatively large because the material is prone to ignition in the event of an internal short circuit. On the other hand, if the positive electrode active material is a lithium iron phosphate-based material (LFP), the weighting coefficient β can be set relatively small.

[0119] The weighting factor β can be set based on empirical rules, or it can be quantified by constructing a trained model based on a database that links the safety tests and fire records of the target storage battery with material degradation information. Table 4 below shows an example of a table in which weighting factors are set according to the type of positive electrode active material.

[0120]

[0121] The risk of lithium deposition may be weighted by a weighting factor according to the operating conditions of the ESS. Examples of the operating conditions of the ESS include the C rate of charge and discharge, the number of charge and discharge cycles performed in a day, and the battery temperature. The weighting factor according to the operating conditions of the ESS is γ, and the weighting factor corresponding to the C rate of charge and discharge is γ. time , and the weighting coefficient corresponding to the number of charge / discharge cycles performed in a day is γ cycle , and the weighting coefficient corresponding to the battery temperature is γ temp. In this case, the risk of lithium precipitation X can be calculated, for example, by the following formula (5). α and β are the same as in the above formulas (3) and (4). time is the weighting coefficient corresponding to the C rate of charging and discharging, γ cycle is the weighting coefficient corresponding to the number of charge / discharge cycles performed in a day, γ temp. are weighting coefficients corresponding to the battery temperature, and 0<γ time <1, 0<γ cycle <1, 0<γ temp. <1. If the C rate of charge / discharge is high, there is a risk of causing further deterioration of the LiB. For example, at a high C rate during charging, the insertion reaction of Li ions into the negative electrode active material cannot keep up, and there is a risk of Li metal being deposited on the negative electrode surface. Therefore, the higher the C rate of charge / discharge, the greater the γ time can be set relatively large. In addition, when the number of charge / discharge cycles per day is large, the LiB is given a large number of charge / discharge histories, which makes it more susceptible to deterioration. Therefore, the more the number of charge / discharge cycles per day, the more γ cycle can be set relatively large. If the battery temperature is too low, the insertion reaction of Li ions into the negative electrode during charging of the LiB is difficult, and Li metal is likely to precipitate. On the other hand, if the battery temperature is too high, high resistance occurs due to the acceleration of SEI formation in the negative electrode active material, and the insertion reaction of Li ions is inhibited, and Li metal is likely to precipitate. Therefore, the higher or lower the battery temperature is relative to room temperature, the higher the risk of Li metal precipitation, and γ temp. can be set relatively large.

[0122] Weighting coefficient γ time , γ cycle , and γ temp. The weighting coefficients can be set based on empirical rules, but they can also be quantified by building a trained model based on a database that links safety tests and fire records of the target storage battery with material degradation information. Tables 5 to 7 below show examples of tables in which weighting coefficients are set according to the type of positive electrode active material.

[0123]

[0124]

[0125]

[0126] As mentioned above, the weighting coefficient α estimated from the material degradation information and electrochemical feature quantity chem. and α elec. The lithium deposition risk X can be calculated comprehensively by weighting the coefficients in accordance with the material information of the storage battery and the operating conditions of the ESS, etc. However, the method for calculating the lithium deposition risk X is not limited to the above specific example, as long as appropriate coefficient items can be set in accordance with the type of storage battery, the operating conditions of the ESS, etc.

[0127] <Estimation of Lithium Deposition Risk X Using the Results of Safety Tests of Storage Batteries> As an example of setting the lithium deposition risk X, an estimation using a trained model C that has learned material degradation information and the results of safety tests of the target storage batteries will be described below.

[0128] A safety test is a test that checks whether a dangerous event such as heat generation, opening, gas emission, or fire occurs when a storage battery is subjected to heat or external stress. Examples of safety tests for storage batteries include various methods such as an overcharge test, an overdischarge test, a heating test, a nail penetration test, a crush test, a differential pressure test, a drop test, an external short circuit test, and an internal short circuit test. Standards for safety test methods are established, such as JIS C8715-2 "Industrial lithium secondary battery cells and battery systems - Part 2: Safety requirements -", ​​IEC 62660-3 "Lithium-ion secondary batteries for electric vehicle propulsion - Safety requirements -", ​​and ISO 12405-1 "Electric vehicles - Test specifications for lithium-ion battery pack systems - Part 3 Safety requirements". When the materials that make up the battery are in a certain state of degradation, these events are likely to be dangerous. Therefore, linking the state of material degradation with the results of safety tests provides important insight into understanding the dangers posed by ESS batteries when they are in a certain state of degradation.

[0129] As an example of material degradation information, the deposition of Li metal will be described. It is known that deposited Li metal is generally active and reacts with the electrolyte when exposed to high temperatures, causing localized heat generation near the deposited Li metal. When localized heat generation occurs inside a storage battery, the positive electrode active material present in the vicinity undergoes thermal decomposition, leading to thermal runaway and the risk of fire. Thus, by knowing in advance the amount, form, and chemical state of the deposited Li metal, as well as events such as heat generation and fire that occur during a heating test, it is possible to predict the risk of fire in a storage battery operated with an ESS.

[0130] The material degradation information learned by the trained model C for estimating the lithium deposition risk X may be the same as the material degradation information learned by the trained model A for diagnosing the material degradation information described above. That is, information on the deposition of Li metal on the negative electrode, which is directly related to the safety of the LiB, is preferred, and includes the area of ​​the deposited lithium metal analyzed from an image of the negative electrode surface of the storage battery, the shape of the deposition region, the deposition weight, distribution state and color tone, the shape of the lithium metal observed by cross-sectional SEM observation of the negative electrode, the chemical shift value (eV) and / or its area obtained from the Li(s1) spectrum obtained by XPS analysis, 7 Examples include chemical states expressed by chemical shift values ​​(ppm) and / or their areas obtained from Li-NMR spectra, and g values ​​and / or line widths obtained from ESR spectra.

[0131] As the results of safety tests learned by the trained model C for estimating the lithium deposition risk X, if the purpose is to focus on learning about Li metal deposition, the results of tests such as overcharge tests, heating tests, nail penetration tests, crush tests, drop tests, external short circuit tests, and internal short circuit tests, which are indicated by hazard levels (for example, hazard levels defined by EUCAR Hazard Level), and battery data such as temperature, current, and voltage obtained at the time of an event can be used.

[0132] A method for creating a trained model C for estimating the lithium deposition risk X includes creating a database of material degradation information obtained by disassembling and analyzing a storage battery in a given state of degradation, prepared for acquiring training data, and information obtained from the results of safety tests on storage batteries in a similar state of degradation; training a machine learning model using these as training data and learning the correlation between at least one piece of material degradation information and the information on the safety test results. It is preferable to collect information on various degradation states for a single storage battery type when preparing the training data. Therefore, it is preferable to prepare a large number of storage batteries in various degradation states by conducting accelerated tests such as cycle tests, high-temperature storage tests, and float tests. Furthermore, by using the results of analyzing the constituent materials of LiBs obtained by acquiring and disassembling LiBs actually used in ESS operation, the lithium deposition risk X can be estimated with even greater accuracy.

[0133] <SoH Calculation> The calculation unit 443 of the cloud BMS 400 calculates (estimates) the state of health (SoH) of the storage battery based on the acquired charge / discharge data (step S403). The SoH is generally expressed as the maintenance rate [%] of the full charge capacity [Ah] at the time of degradation, based on the initial full charge capacity [Ah] (= full charge capacity [Ah] at the time of degradation [Ah] / initial full charge capacity [Ah] × 100), or as the increase rate [%] of the internal resistance [Ω] at the time of degradation, based on the initial internal resistance [Ω] (= internal resistance [Ω] at the time of degradation [Ω] / initial internal resistance [Ω] × 100). Either calculation standard may be used. The SoH calculation (estimation) may be performed by the calculation unit 342 of the local BMS 300. The SoH can be estimated by conventional methods, such as (1) estimation based on the capacity when the storage battery is charged and discharged at a constant current and a constant voltage, (2) estimation based on the internal resistance obtained by the AC impedance method, (3) estimation based on the internal resistance obtained from the slope of the current-voltage characteristics (IV characteristics), (4) estimation based on the internal resistance obtained from the voltage transient response when the current is changed in a stepwise manner, (5) estimation by fitting the charging curve data using a regression equation or a model, or (6) estimation based on an equivalent circuit or physical model that behaves in the same way as the charging and discharging data of the storage battery, or a combination of these.

[0134] (1) More specifically, in the estimation based on the capacity when a storage battery is charged and discharged at a constant current and a constant voltage, the storage battery is first charged and discharged at a constant current and a constant voltage, and the current capacity is measured. Then, the initial full charge capacity [Ah] is used as a reference, and the SoH is calculated as the maintenance rate [%] of the full charge capacity [Ah] at the time of degradation (= full charge capacity [Ah] at the time of degradation [Ah] / initial full charge capacity [Ah] × 100). Depending on the specifications of the storage battery, it may be possible to evaluate the SoH by accumulating a charge and discharge history.

[0135] (2) Estimation based on internal resistance obtained by the AC impedance method involves applying AC current of a single or multiple frequencies to a storage battery using a dedicated measuring device to measure the impedance at each frequency. The frequency characteristics of the storage battery's internal resistance are obtained (electrochemical impedance spectroscopy), and the internal resistance can be separated by fitting it to an equivalent circuit (determining the circuit constants). Then, using the initial internal resistance [Ω] as the reference, the increase rate [%] of internal resistance [Ω] during degradation (= internal resistance during degradation [Ω] / initial internal resistance [Ω] × 100) is calculated as the SoH.

[0136] (3) Estimation based on the internal resistance obtained from the slope of the current-voltage characteristic (IV characteristic) is performed by creating a current-voltage characteristic (IV characteristic) for each SoH from the charge / discharge curve (rate characteristic) of the storage battery. The IV characteristic is linearly approximated, and the internal resistance is calculated from the slope. Then, using the initial internal resistance [Ω] as a reference, the rate of increase [%] of the internal resistance [Ω] during degradation (= internal resistance during degradation [Ω] / initial internal resistance [Ω] × 100) is calculated as the SoH. Because the IV characteristic cannot always be linearly approximated, it is preferable to make corrections based on the temperature and current value of the storage battery to ensure estimation accuracy.

[0137] (4) In estimation based on internal resistance obtained from the voltage transient response when the current is changed stepwise, a specific current pattern that changes stepwise is applied to the storage battery, and the voltage response is measured. The internal resistance can be calculated by dividing the voltage drop [V] from the open circuit voltage by the current value [A]. Then, using the initial internal resistance [Ω] as the reference, the SoH is calculated as the rate of increase [%] of internal resistance [Ω] during degradation (= internal resistance during degradation [Ω] / initial internal resistance [Ω] × 100).

[0138] (5) Estimation by fitting charging curve data using a regression equation or model is a method of estimating the degree of deterioration and actual capacity from the profile of the charging and discharging curve. The charging and discharging curve data can be fitted (circuit constants are determined) using a regression equation or an equivalent circuit to estimate the internal resistance. Then, the SoH is calculated as the percentage increase in internal resistance (Ω) during deterioration (= internal resistance (Ω) during deterioration (Ω) / initial internal resistance (Ω) × 100) based on the initial internal resistance (Ω).

[0139] (6) Estimation based on an equivalent circuit or physical model that behaves similarly to the charge / discharge data of a storage battery is a method of estimating SOH by expressing a virtual storage battery (model) using an equivalent circuit or equations based on physical phenomena. Model construction and parameter determination are required for each storage battery specification. Model construction involves modeling internal battery phenomena, such as electrochemical reactions, Li transport (migration, diffusion) in the electrolyte, Li transport (diffusion) in the active material, and degradation reactions (internal resistance increase, electrolyte decomposition). The internal resistance of the constructed virtual storage battery is defined as the degraded internal resistance [Ω], and the increase rate [%] from the initial internal resistance [Ω] (= degraded internal resistance [Ω] / initial internal resistance [Ω] × 100) is calculated as the SOH.

[0140] Among these, because of their excellent accuracy in estimating the SoH, it is preferable to calculate (estimate) the SoH by (1) estimating based on the capacity when the storage battery is charged and discharged at a constant current and a constant voltage, or (3) estimating based on the internal resistance obtained from the slope of the current-voltage characteristics (IV characteristics).

[0141] <Labeling Process> The calculation unit 443 of the cloud BMS 400 labels the degradation state of the storage battery based on one or more predetermined thresholds for the SoH calculated as described above and the lithium deposition risk X (step S404). The output unit (transmitter / receiver 420) then outputs the labeling result (step S405), thereby completing the series of steps. The labeling can be performed based on a labeling table pre-stored in the storage unit 430. As illustrated in the labeling table of FIG. 10 , the SoH, expressed as the capacity maintenance ratio, may be divided into three levels: 70% or less, greater than 70% to less than 90%, and greater than 90% to 100%. The lithium deposition risk X may also be divided into four levels: "none" (less than 0.1), "low (0.1 or greater but less than 5.0)," "medium" (5.0 or greater but less than 10.0), and "high" (10.0 or greater). In the example of the labeling table in FIG. 10, the deterioration state of the storage battery is classified into 12 different quadrants, A1 to A4, B1 to B4, and C1 to C4, and the deterioration state of the storage battery is labeled as follows. The labeling results correspond to the predicted replacement time of the storage battery if operation continues under the current operating conditions of the ESS. Deterioration state 1 (A1): No replacement required for more than one year Deterioration state 2 (A2, B1, B2): Replacement required for more than three months but within one year Deterioration state 3 (A3, B3, C1, C2): Replacement required within three months Deterioration state 4 (A4, B4, C3, C4): Immediate replacement

[0142] 10 is merely an example, and the number and values ​​of thresholds, the number of labels and the method of classifying quadrants, and the replacement timings corresponding to the labeling can be set arbitrarily depending on the operating conditions of the ESS, the material information and design information of the storage battery, the calculation results of the future risk X of lithium deposition described below, etc. The labeling table can be set based on empirical rules, but the labeling table may also be constructed based on a database linking the material information, design information, degradation state, charge / discharge conditions, and period until replacement of the target storage battery, etc.

[0143] The labeling result is output (transmitted) from the transmitter (transmitter / receiver 420) to the user's terminal device 500 via the network N. The labeling result may be output (transmitted) to the EMS 240 and / or the local BMS 300. The labeling result is also used in the control process of the storage battery described below.

[0144] 11 is a flowchart showing an example of an operation for predicting a future lithium deposition risk X. The operation flow described below is based on a program stored in advance in the storage unit 430 of the cloud BMS 400 and is executed mainly by a calculation unit 443 and a prediction unit 444 included in the processing unit 440 of the cloud BMS 400 in cooperation with each element.

[0145] First, the calculation unit 443 calculates the current time, that is, the time when the diagnosis is performed (t 1 ) estimated and calculated from the charge and discharge data at the current time (t 1 ) (step S501). Next, the calculation unit 443 acquires (specifies) material deterioration information based on the charge / discharge data at the specified current time (t 1 ) based on information including material deterioration information at the current time (t 1 ) Lithium deposition risk Xt 1 (Step S502). Then, the calculation unit 443 calculates the current time (t 1 Based on the charge and discharge data at the current time (t 1 These steps S501 to S503 are carried out based on the fact that the charge / discharge data used is the current value (t 1 ), which are essentially the same as steps S401 to S403 in FIG. 10 described above.

[0146] Next, the prediction unit 444 acquires information including past charge / discharge data and past changes in SoH (step S504). The information including past charge / discharge data and past changes in SoH can be acquired from an integrated database stored in the storage unit 430 of the cloud BMS 400, or may be acquired from a database stored in the storage unit 330 of the local BMS 300. In addition to the past charge / discharge data and past SoH, the information may also include past operating conditions of the ESS. Then, the prediction unit 444 inputs the current SoH and the current lithium deposition risk Xt 1 , as well as information including past charge / discharge data and past changes in SoH, and predicts the future change in the risk of lithium precipitation X, which has a high correlation with the information (steps S505 to S506). Then, the output unit (transmitter / receiver 420) outputs the prediction result (step S507), and the series of steps ends. As illustrated in the graph of FIG. 10, the future change in the risk of lithium precipitation X is 1 ) is represented as a curve to the right of

[0147] The prediction result is output (transmitted) from the transmitter (transmitter / receiver 420) to the user's terminal device 500 via the network N. The prediction result may also be output (transmitted) to the EMS 240 and / or the local BMS 300. The prediction result is also used in the control process of the storage battery described below.

[0148] 12 is a flowchart showing an example of a process for labeling the degradation state of a storage battery based on the future risk of lithium deposition X. The flow of operations described below is based on a program stored in advance in the storage unit 430 of the cloud BMS 400 and is executed mainly by the calculation unit 443 and the prediction unit 444 included in the processing unit 440 of the cloud BMS 400 in cooperation with each element.

[0149] First, the prediction unit 444 predicts the future transition of the lithium deposition risk X by the operation of the prediction process described above (step S601). Next, the calculation unit 443 calculates the future transition of the lithium deposition risk X at a certain time point (t2 ) Lithium deposition risk Xt 2 Calculate the lithium deposition risk Xt 2 The state of health of the battery is labeled based on one or more predetermined thresholds for

[0150] More specifically, first, the calculation unit 443 calculates the current lithium deposition risk Xt 1 Obtain Xt 1 is a preset threshold X 1 It is determined whether or not it is equal to or greater than Xt (steps S602 to S603). 1 is X 1 If the result is Yes, the calculation unit 443 labels the storage battery as being in a degraded state 4 (replace immediately) (step S604). 1 is X 1 If the result is less than three months (t 3M ) Lithium deposition risk Xt 3M Obtain Xt 3M is X 1 It is determined whether or not it is equal to or greater than Xt (steps S605 to S606). 3M is X 1 If the result is YES, the calculation unit 443 labels the storage battery as being in a deteriorated state 3 (replace within three months) (step S607). 3M is X 1 If the result is less than t (No), the calculation unit 443 next calculates the risk of lithium deposition X one year from the transition of the risk of lithium deposition X in the future (t 1Y ) Lithium deposition risk Xt 1Y Obtain Xt 1Y is X 1 It is determined whether or not it is equal to or greater than Xt (steps S608 to S609). 1Y is X 1 If the result is equal to or greater than Xt (Yes), the calculation unit 443 labels the storage battery as being in a deteriorated state 2 (replace within more than three months and one year) (step S610). 1Y is X 1If the result is No, the calculation unit 443 labels the storage battery as being in a degradation state 1 (no replacement required for over one year) (step S611). Then, the output unit (transmitter / receiver 420) outputs the labeling result (step S612), and the series of steps ends.

[0151] The graph in FIG. 12 shows the transition of the future lithium deposition risk X used in this labeling process and Xt 1 , Xt 3M , Xt 1Y , and X 1 In this example, Xt 3M is X 1 Less than (No), but Xt 1Y is X 1 Since the result is equal to or greater than this (Yes), the calculation unit 443 labels the storage battery as being in a degradation state 2 (replace within more than three months and one year).

[0152] The labeling process in FIG. 12 is merely an example, and the number and values ​​of thresholds can be set arbitrarily depending on the operating conditions of the ESS, material information of the storage battery, design information, etc. For example, the threshold X 1 As shown in the labeling table of Figure 10, "Battery 1" indicates a "medium" to "high" risk of lithium deposition, indicating that battery replacement is necessary. The number of labels, the classification method, and the replacement timing corresponding to the labeling can be set arbitrarily depending on the type of storage battery, etc. The labeling table can be set based on empirical rules, but it may also be constructed based on a database linking information such as material information, design information, deterioration state, charge / discharge conditions, and the period until replacement of the target storage battery.

[0153] The labeling result is output (transmitted) from the transmitter (transmitter / receiver 420) to the user's terminal device 500 via the network N. The labeling result may be output (transmitted) to the EMS 240 and / or the local BMS 300. The labeling result is also used in the control process of the storage battery described below.

[0154] Control Processing

[0155] <Control Processing Form 1: Control Based on Lithium Deposition Risk X> An example of the operation of the control processing based on the lithium deposition risk X will be described with reference to Fig. 9. The flow of the operation described below is based on a program stored in advance in the storage unit 430 of the cloud BMS 400 and is executed mainly by the control unit 445 included in the processing unit 440 of the cloud BMS 400 in cooperation with each element.

[0156] The processing unit 440 further includes a control unit 445 that controls the charge / discharge conditions of the storage battery based on the lithium precipitation risk X described above. The control unit 445 controls the ESS by transmitting a charge / discharge condition change command or a charge / discharge stop command to the power converter 100 or the end BMS 230 based on the lithium precipitation risk X. Transmission includes transmission from the cloud BMS 400 to the power converter 100 or the end BMS 230 via a wired or wireless connection, and transmission via the network N. The output lithium precipitation risk X, like the material deterioration information, is transmitted to the user's terminal device 500 via the network N. Furthermore, an alert can be issued to the user's terminal device 500 via the network N in response to the lithium precipitation risk X. The alert may also be issued to the local BMS 300.

[0157] First, by the operation of the diagnostic process described above, material deterioration information is acquired (identified) from the diagnostic unit (step S901). Next, the calculation unit calculates the lithium deposition risk X based on information including the identified material deterioration information (step S902).

[0158] Based on the lithium deposition risk X calculated in the calculation of the lithium deposition risk X (step S902), the control unit compares the lithium deposition risk X with a preset threshold value of the lithium deposition risk X. In the example of FIG. 9, X is used as the threshold value of the ignition risk. 1 and X 2 (However, X 1 <X 2 ) Two thresholds are set.

[0159] The control unit first calculates the lithium deposition risk X and the threshold value X 1 Compared with, the risk of lithium deposition is X1 It is determined whether the risk of lithium deposition X is equal to or less than X (step S903). 1 If the calculated value is equal to or less than the planned value, the control unit maintains the operating conditions as planned (step S904) and ends the series of steps.

[0160] Lithium deposition risk is X 1 If it exceeds the threshold value X, the control unit then calculates the lithium deposition risk X and the threshold value X. 2 Compared with, the risk of lithium deposition is X 2 It is determined whether the risk of lithium deposition X is equal to or less than X (step S905). 1 Exceeding and X 2 If it is below this level, it is assumed that the deterioration has progressed to a certain extent, and a command to change (relax) the operating conditions of the storage battery to reduce the risk of fire is output to the power converter 100 (step S906), and the series of steps ends. Any method for relaxing the operating conditions may be used as long as it can suppress the deterioration of the storage battery, and examples of such methods include narrowing the operating voltage range, reducing the maximum power, reducing the C rate during charging at a high SoC%, reducing the charging SoC%, and optimizing the operating temperature.

[0161] Lithium deposition risk is X 2 If the deterioration level of the storage battery exceeds this limit, the degree of deterioration is deemed to be unacceptable from a safety standpoint, and the control unit outputs a command to the end BMS 230 to stop operation of the storage battery in question, and also issues an alert from the transmission unit (transmission / reception unit 420) to the user's terminal device 500 via the network N (step S907), thereby completing the series of steps.

[0162] As described above, the ESS can be operated more safely by controlling the storage battery in accordance with the risk of lithium deposition X. In the example of Fig. 9, the case where the threshold value is set to two levels has been described, but the threshold value can be set to any number of levels so that the ESS can be controlled or stopped more safely.

[0163] 13 is a flowchart showing an example of the operation of the control process of the storage battery based on the labeling result. Note that the flow of the operation described below is based on a program stored in advance in the storage unit 430 of the cloud BMS 400 and is executed mainly by the control unit 445 included in the processing unit 440 of the cloud BMS 400 in cooperation with each element.

[0164] First, the control unit 445 acquires the labeling results described above and determines whether the storage battery is in degradation state 4 (replace immediately) (steps S701-S702). If the storage battery is in degradation state 4 (Yes), the control unit 445 changes the charge / discharge conditions of the storage battery to control mode α (step S703) and ends the series of steps. If the storage battery is not in degradation state 4 (No), the control unit 445 determines whether the storage battery is in degradation state 3 (replace within three months) (step S704). If the storage battery is in degradation state 3 (Yes), the control unit 445 changes the charge / discharge conditions of the storage battery to control mode β (step S705) and ends the series of steps. If the storage battery is not in degradation state 3 (No), the control unit 445 determines whether the storage battery is in degradation state 2 (replace within more than three months but one year) (step S706). If the storage battery is in degradation state 2 (Yes), the control unit 445 maintains the charge / discharge conditions of the storage battery (step S707), and ends the series of steps. If the storage battery is not in degradation state 2 (No), that is, if the storage battery is in degradation state 1 (no replacement required for more than one year), the control unit 445 changes the charge / discharge conditions of the storage battery to control mode γ (step S708), and ends the series of steps. Here, the control modes α to γ ​​respectively represent the following conditions. Control mode α: conditions under which lithium deposition hardly progresses Control mode β: conditions under which lithium deposition does not progress easily Control mode γ: conditions under which output is improved within a safe range

[0165] Control modes α and β may be any method capable of suppressing lithium deposition. Examples include narrowing the voltage range during use, reducing the maximum power during use, reducing the C-rate during charge and discharge, reducing the charge State of Charge (SoC) (%), and optimizing the operating temperature. These methods may also be combined. Among these, reducing the C-rate during charge and discharge and lowering the voltage during use (reducing the upper limit voltage) are effective for suppressing lithium deposition. For example, in control mode α, the C-rate and upper limit voltage before control are set to the reference (100%), and the C-rate is reduced to 25% to 50% and the upper limit voltage is reduced to 60% to 80%. In control mode β, the C-rate and upper limit voltage before control are set to the reference (100%), and the C-rate is reduced to 50% to 75% and the upper limit voltage is reduced to 80% to 90%. Control mode γ may be any method capable of improving the output of the storage battery within a safe range. For example, methods include widening the voltage range during use, increasing the maximum power during use, increasing the C rate during charge and discharge, increasing the charge SoC (%), and optimizing the operating temperature, and these may be combined. For example, in control mode γ, the C rate and upper limit voltage before control are set to the reference (100%), and the C rate is increased to 101% to 120% and the upper limit voltage is increased to 101% to 120%. The above control modes are merely examples, and specific conditions can be set arbitrarily depending on the material information and design information of the storage battery. The control mode can be set based on empirical rules, or it can be set based on a database linking the deterioration state of the target storage battery, charge and discharge conditions, and the period until replacement.

[0166] The control mode can be further modified depending on the cause of battery degradation. As described above, the battery material degradation information obtained by the trained model A includes information on the degradation of materials (such as the positive electrode, negative electrode, electrolyte, and separator) inside the battery. Therefore, by identifying the material (component) that is causing battery degradation, the control mode can be further optimized to suppress the degradation of that material (component). If the cause of negative electrode degradation, for example, a decrease in the amount of lithium in the negative electrode, is the cause of degradation, it is effective not to use a range near the lower limit of the SoC (%) in order to prevent negative electrode degradation due to non-uniformity in the intercalation reaction caused by charging and discharging in a low SoC (%) range. Furthermore, if the cause of positive electrode degradation, for example, destruction of the positive electrode active material, is the cause of degradation, it is effective not to use a range near the upper limit of the SoC (%) in order to suppress positive electrode interface degradation due to oxidative decomposition of the electrolyte that occurs at the interface between the positive electrode and the electrolyte when a high voltage is applied.

[0167] In a preferred embodiment, the storage battery is composed of a plurality of storage battery modules 210 arranged in series and / or parallel, each including a plurality of cells arranged in series and / or parallel. The diagnosing unit 442 and the calculating unit 443 diagnose and label the deterioration state of each storage battery module, and the control unit 445 controls the charge / discharge conditions for each storage battery module. As described above, the plurality of storage batteries constituting one storage battery module are typically connected in series, and the plurality of storage battery modules are connected in parallel with each other. Therefore, from a safety standpoint, when replacing batteries, each storage battery module, each consisting of a plurality of storage batteries connected in series, is shut down and replaced. Therefore, diagnosing, labeling, and controlling each storage battery module is rational and efficient in terms of actual ESS operation.

[0168] 14 is a flowchart showing an example of the operation of the control process of the storage battery based on the diagnosis result and the target operation period. Note that the operation flow described below is based on a program stored in advance in the storage unit 430 of the cloud BMS 400 and is executed mainly by the control unit 445 included in the processing unit 440 of the cloud BMS 400 in cooperation with each element.

[0169] First, the control unit 445 acquires the labeling results described above (step S801). Next, the control unit 445 acquires information regarding the target operating period of the ESS (step S802). The target operating period of the ESS can be arbitrarily determined based on various factors. For example, if the ESS is used to operate equipment in a factory, the target operating period may be set to the time of scheduled maintenance of the equipment (e.g., two months from now). Alternatively, the target operating period may be set to the time when batteries with the desired performance and quantity can be procured (e.g., two months from now) based on the market conditions of new storage batteries or reused batteries to be replaced. Furthermore, based on forecast information on electricity market prices, the target operating period can be determined to minimize opportunity losses, for example, by aligning the ESS maintenance period with a period when the electricity market price is expected to rise by a predetermined amount or more.

[0170] Next, the control unit 445 determines whether the storage battery is in degradation state 4 (replace immediately) (step S803). If the storage battery is in degradation state 4 (Yes), the control unit 445 changes the charge / discharge conditions of the storage battery to control mode α (step S804) and ends the series of steps. If the storage battery is not in degradation state 4 (No), the control unit 445 determines whether the storage battery is in degradation state 3 (replace within three months) (step S805). If the storage battery is in degradation state 3 (Yes), the control unit 445 changes the charge / discharge conditions of the storage battery to control mode β (step S806) and ends the series of steps. If the storage battery is not in degradation state 3 (No), the control unit 445 determines whether the storage battery is in degradation state 2 (replace within more than three months but one year) (step S807). If the storage battery is in degradation state 2 (Yes), the control unit 445 maintains the charge / discharge conditions of the storage battery (step S808) and ends the series of steps. If the storage battery is not in degradation state 2 (No), that is, if it is in degradation state 1 (no replacement required for over one year), the control unit 445 changes the charge / discharge conditions of the storage battery to control mode γ (step S809), and ends the series of steps. Here, control modes α to γ ​​respectively represent the following conditions. Control mode α: conditions under which lithium deposition hardly progresses and operation for two months is possible. Control mode β: conditions under which lithium deposition does not progress easily and operation for two months is possible. Control mode γ: conditions under which output is improved within a safe range and operation for two months is possible.

[0171] As explained above in "Control Based on Diagnostic Results," the control modes α to γ ​​can be set in the same way, except that the target operating period is taken into consideration. The control modes can also be modified in the same way depending on the cause of battery deterioration. It is also preferable to diagnose, label, and control each battery module.

[0172] <<Modifications>> The above description mainly focuses on an example of an embodiment in which the cloud BMS 400 performs all of the diagnostic processing, the calculation and prediction processing of the current and future lithium deposition risk X, the labeling processing, and the control processing of the first and second embodiments. In this embodiment, reducing the functions of the local BMS 300 is expected to reduce the equipment costs of the ESS. However, in the diagnostic system 1, any of the devices constituting the system may function as a diagnostic device as long as the diagnostics and other processes of the present disclosure can be performed as an entire system interconnected via a network. Furthermore, the diagnostic device (diagnostic unit), calculation device (calculation unit), and control device (control unit) may be configured as a single device or as multiple devices interconnected via a network N.

[0173] <Variation 1> Fig. 15 is a diagram showing a schematic configuration of a cloud BMS 400 as a diagnostic device and server, which can be combined with the local BMS 300 of Fig. 4. The cloud BMS 400 includes an acquisition unit (a receiver in the transceiver unit 420), an output unit (a transmitter in the transceiver unit 420), a storage unit 430, and a processing unit 440. The storage unit 430 stores a trained model 431 including trained models A and / or B, and an integrated database 432. The processing unit 440 includes a generation unit 441, a diagnosis unit 442, a calculation unit 443, and a prediction unit 444. The transmission / reception unit, storage unit, processing unit, etc. are connected to each other via a bus, an interface unit, etc.

[0174] In this aspect, the cloud BMS 400 in FIG. 15 performs only the diagnostic process and the calculation and prediction process of the current and future lithium deposition risk X using a diagnostic unit, a calculation unit, and a prediction unit, and transmits the results of the diagnostic process and the calculation and prediction process to the local BMS 300. The local BMS 300 in FIG. 4 can perform the labeling process and the control process using a calculation unit and a control unit. Alternatively, the cloud BMS 400 in FIG. 15 performs only the diagnostic process, the calculation and prediction process of the current and future lithium deposition risk X, and the labeling process using a diagnostic unit, a calculation unit, and a prediction unit, and transmits the results of the diagnostic process and the calculation and prediction process to the local BMS 300. The local BMS 300 in FIG. 4 can perform the control process using a calculation unit and a control unit.

[0175] <Variation 2> Fig. 16 is a diagram illustrating a schematic configuration of the local BMS 300. The local BMS 300 includes an acquisition unit 310, a transmission unit (transmission / reception unit 320), a reception unit (transmission / reception unit 320), a storage unit 330, a processing unit 340, and an output unit 350. The reception unit receives a trained model 331 including a trained model B transmitted from the cloud BMS 400 and stores it in the storage unit 330. If an existing trained model is stored in the storage unit 330, the received trained model is reflected in the existing trained model. The storage unit 330 stores the trained model 331 including the received (reflected) trained model B, etc. The processing unit 340 includes a calculation unit 342, a prediction unit 343, and a control unit 344. The acquisition unit, transmission / reception unit, storage unit, processing unit, output unit, etc. are connected to each other via a bus, an interface unit, etc.

[0176] Fig. 17 is a diagram showing a schematic configuration of a cloud BMS 400 as a diagnostic device and server, which can be combined with the local BMS 300 in Fig. 16. The cloud BMS 400 includes an acquisition unit (a receiver in the transceiver unit 420), an output unit (a transmitter in the transceiver unit 420), a storage unit 430, and a processing unit 440. The storage unit 430 stores a trained model 431 including trained models A and B, and an integrated database 432. The processing unit 440 includes a generation unit 441, a diagnosis unit 442, and a calculation unit 443. The transmission / reception unit, storage unit, processing unit, etc. are connected to each other via a bus, an interface unit, etc.

[0177] In this aspect, the cloud BMS 400 in FIG. 17 generates a trained model 431 including trained models A and B, and transmits and reflects trained model B to the local BMS 300 in FIG. 16 . The cloud BMS 400 in FIG. 17 performs only the diagnosis process using the diagnosis unit and calculation unit, and transmits the results of the diagnosis process to the local BMS 300. The local BMS 300 in FIG. 16 can then perform calculation and prediction processes of the current and future lithium deposition risk X, labeling processes, and control processes using the calculation unit, prediction unit, and control unit. Alternatively, the cloud BMS 400 in FIG. 17 performs only the diagnosis process and calculation processes of the current lithium deposition risk X using the diagnosis unit and calculation unit, and transmits the results of the diagnosis process and calculation processes to the local BMS 300. The local BMS 300 in FIG. 16 can then perform prediction processes of the future lithium deposition risk X, labeling processes, and control processes using the calculation unit, prediction unit, and control unit.

[0178] <Variation 3> Fig. 18 is a diagram illustrating a schematic configuration of a local BMS 300 as a diagnostic device. The local BMS 300 includes an acquisition unit 310, a transmission unit (transmission / reception unit 320), a reception unit (transmission / reception unit 320), a storage unit 330, a processing unit 340, and an output unit 350. The reception unit receives a trained model 331 including trained models A and / or B transmitted from the cloud BMS 400 and stores it in the storage unit 330. If an existing trained model is stored in the storage unit 330, the received trained model is reflected in the existing trained model. The storage unit 330 stores the received (reflected) trained model 331 including trained models A and / or B. The processing unit 340 includes a diagnosis unit 341, a calculation unit 342, a prediction unit 343, and a control unit 344. The acquisition unit, transmission / reception unit, storage unit, processing unit, output unit, etc. are connected to one another via a bus, an interface unit, etc.

[0179] Fig. 19 is a diagram showing a schematic configuration of a cloud BMS 400 as a server that can be combined with the local BMS 300 of Fig. 18. The cloud BMS 400 includes an acquisition unit (a receiving unit in the transceiver unit 420), an output unit (a transmitting unit in the transceiver unit 420), a storage unit 430, and a processing unit 440. The storage unit 430 stores a trained model 431 including trained models A and / or B, and an integrated database 432. The processing unit 440 includes a generation unit 441. The transceiver unit, storage unit, processing unit, etc. are connected to each other via a bus, an interface unit, etc.

[0180] In this aspect, the cloud BMS 400 in Figure 19 generates a trained model 431 including trained models A and / or B, and transmits and reflects this to the local BMS 300 in Figure 18. The local BMS 300 in Figure 18 can then perform all of the diagnostic processing, calculation and prediction processing of the current and future lithium deposition risk X, labeling processing, and control processing using the transmitted and reflected trained models A and / or B. In this aspect, the local BMS 300 as a diagnostic device can diagnose and control the storage battery independently of the network N, which has the advantage of enabling stable ESS operation even in a poor network environment.

[0181] For details of the functional units and processes in Modifications 1 to 3, the descriptions of the functional units and processes described with reference to FIGS. 4 and 5 are cited.

[0182] 1 Diagnostic system, 10 Energy storage system, 100 PCS, 200 Battery stack, 210 Battery module, 220 Sensor unit, 230 End BMS, 231 Acquisition unit, 232 Monitoring unit, 233 Balance adjustment / stop unit, 240 EMS, 300 Local BMS, 310 Acquisition unit, 320 Transmitting / receiving unit, 330 Memory unit, 331 Received (reflected) learned model, 340 Processing unit, 341 Diagnosis unit, 342 Calculation unit, 343 Prediction unit, 344 Control unit, 350 Output unit, 400 Cloud BMS, 420 Transmitting / receiving unit, 430 Memory unit, 431 Learned model, 432 Integrated database, 440 Processing unit, 441 Generation unit, 442 Diagnosis unit, 443 Calculation unit, 444 Prediction unit, 445 Control unit, 450 Output unit, 500 terminal device

Claims

1. A diagnostic device for a storage battery used in an energy storage system, comprising: a memory unit that stores a trained model A that has been trained to output material deterioration information of the storage battery when charge / discharge data is input; an acquisition unit that acquires charge / discharge data of the storage battery; a diagnosis unit that diagnoses material deterioration information of the storage battery obtained by inputting the charge / discharge data of the storage battery to the trained model A; and an output unit that outputs the material deterioration information obtained from the diagnosis unit as a diagnosis result.

2. The diagnostic device according to claim 1, wherein the material information of the storage battery is information on at least one material selected from the group consisting of a positive electrode, a negative electrode, a separator, and an electrolyte of the storage battery.

3. The diagnostic device according to claim 2, further comprising a calculation unit that calculates a risk X of lithium precipitation in the storage battery based on information including the material deterioration information, and the diagnostic result includes the risk X of lithium precipitation.

4. The diagnostic device according to claim 3, wherein the calculation unit calculates a risk X of lithium precipitation of the storage battery based on information including the material deterioration information and at least one selected from the group consisting of the charge / discharge data, material information of the storage battery, and operating conditions of the energy storage system.

5. The calculation unit calculates the lithium precipitation risk X estimated from the material deterioration information. chem. A weighting coefficient α according to the material information of the negative electrode of the storage battery chem. and the lithium precipitation risk X estimated from the charge / discharge data. elec. A weighting coefficient α according to the material information of the negative electrode of the storage battery elec. and optionally multiplying the sum by at least one selected from the group consisting of a weighting coefficient β corresponding to material information of a positive electrode of the storage battery and a weighting coefficient γ corresponding to the operating conditions, thereby calculating a risk X of lithium precipitation of the storage battery.

6. The diagnostic device according to any one of claims 1 to 5, further comprising a calculation unit that calculates a degradation state of the storage battery based on a state of health (SoH) of the storage battery estimated based on charge / discharge data of the storage battery and the material degradation information, and the output unit outputs the degradation state as a diagnostic result.

7. The diagnostic device according to claim 6, wherein the calculation unit is configured to calculate a lithium precipitation risk X of the storage battery based on information including the material deterioration information, and to label a deterioration state of the storage battery based on one or more predetermined threshold values ​​for each of the SoH and the lithium precipitation risk X, and the diagnostic result includes a result of the labeling.

8. The diagnostic device according to claim 7, wherein the output unit transmits the diagnostic result to a user's terminal device.

9. The calculation unit calculates the material deterioration information and the current time (t 1 ) and, optionally, at least one selected from the group consisting of material information of the storage battery and operating conditions of the energy storage system, 1 ) SoH and lithium precipitation risk Xt 1 The diagnostic device according to claim 7, further comprising:

10. The diagnostic device according to claim 9, wherein the material information of the storage battery is information on at least one material selected from the group consisting of a positive electrode, a negative electrode, a separator, and an electrolyte of the storage battery.

11. The calculation unit calculates the lithium precipitation risk X estimated from the material deterioration information. chem. A weighting coefficient α according to the material information of the negative electrode of the storage battery chem. and the lithium precipitation risk X estimated from the charge / discharge data. elec. A weighting coefficient α according to the material information of the negative electrode of the storage battery elec. and optionally multiplying the sum by at least one selected from the group consisting of a weighting factor β according to the material information of the positive electrode of the storage battery and a weighting factor γ according to the operating conditions, thereby obtaining the current time (t 1 ) Lithium precipitation risk Xt 1 The diagnostic device according to claim 10, further comprising:

12. The storage unit stores the current time (t 1 ) SoH and lithium precipitation risk Xt 1 and a trained model B that has been trained to output a future transition of the risk of lithium precipitation X when information including past charge / discharge data and a past transition of the SoH is input, and the diagnosis device further includes a prediction unit, and the prediction unit is configured to predict the current time (t 1 ) SoH and lithium precipitation risk Xt 1 and information including past charge / discharge data and past changes in SoH are input to the trained model B, and a future change in the lithium precipitation risk X is predicted. The calculation unit calculates a time point (t 2 ) Lithium precipitation risk Xt 2 Calculate the lithium precipitation risk Xt 2 The diagnostic device according to claim 11 , further comprising: a step of labeling the state of deterioration of the storage battery based on one or more predetermined thresholds for 13. The diagnostic device according to claim 7, further comprising a control unit that controls the storage battery based on a result of the labeling.

14. The diagnostic device according to claim 13, wherein the control unit controls the storage battery based on the result of the labeling and a cause of deterioration of the storage battery.

15. The diagnostic device according to claim 13, wherein the control unit controls the storage battery based on the result of the labeling and a target operating period of the energy storage system.

16. The diagnostic device according to claim 13, wherein the storage battery is composed of a plurality of storage battery modules arranged in series and / or parallel, each of which includes a plurality of cells arranged in series and / or parallel, the diagnosing unit and calculating unit diagnose and label the deterioration state of each of the storage battery modules, and the control unit controls charging and discharging conditions for each of the storage battery modules.

17. The material deterioration information of the storage battery includes the area of ​​deposited lithium metal analyzed from an image of the negative electrode surface of the storage battery, the shape of the deposited region, the weight of the deposit, the distribution state and color tone, the shape of the lithium metal observed by cross-sectional SEM observation of the negative electrode, the chemical shift value (eV) and / or its area obtained from the Li(s1) spectrum obtained by XPS analysis, 7 The diagnostic device according to claim 6, wherein the chemical state is at least one selected from the group consisting of a chemical shift value (ppm) and / or its area obtained from a Li-NMR spectrum, and a g value and / or line width obtained from an ESR spectrum.

18. The diagnostic device of claim 6, wherein the charge / discharge data input to the trained model A is at least one selected from the group consisting of voltage change over time when the storage battery is rested immediately after being charged to an arbitrary SoC (dOCV / dt-t plot), voltage change over time during pulse charging and / or discharging at a constant current (V-t plot), current change over time observed during constant voltage charging (I-t plot), charge capacity and discharge capacity at constant current or constant power, measurement data obtained by an AC impedance method, and data obtained by mathematically processing these.

19. The diagnostic device of claim 6, wherein the SoH is estimated based on the capacity when the storage battery is charged and discharged at a constant current and constant voltage, based on the internal resistance obtained by an AC impedance method, based on the internal resistance obtained from the slope of the current-voltage characteristics (IV characteristics), based on the internal resistance obtained from the voltage transient response when the current is changed in a step manner, by fitting charging curve data using a regression equation or a model, based on an equivalent circuit or physical model that behaves in a similar manner to the charging and discharging data of the storage battery, or a combination of these.

20. The diagnostic device according to claim 19, wherein the SoH is estimated based on an internal resistance obtained from the slope of a current-voltage characteristic (IV characteristic).

21. A diagnostic device for a storage battery used in an energy storage system, comprising: an acquisition unit that acquires charge / discharge data of the storage battery; a receiving unit that receives a trained model A that outputs material deterioration information from the charge / discharge data of the storage battery; a diagnosis unit that diagnoses material deterioration information of the storage battery obtained by inputting the charge / discharge data of the storage battery to the received trained model A; and an output unit that outputs the material deterioration information obtained from the diagnosis unit as a diagnosis result.

22. The diagnostic device according to claim 21, further comprising a calculation unit that estimates a state of health (SoH) of the storage battery based on charge / discharge data of the storage battery, and calculates a degradation state of the storage battery based on the SoH and the material degradation information, and the output unit outputs the degradation state as a diagnostic result.

23. The diagnostic device according to claim 22, wherein the calculation unit is configured to calculate a lithium precipitation risk X of the storage battery based on information including the material deterioration information, and to label a deterioration state of the storage battery based on one or more predetermined threshold values ​​for each of the SoH and the lithium precipitation risk X, and the diagnostic result includes a result of the labeling.

24. The diagnostic device according to claim 23, wherein the output unit transmits the diagnostic result to a terminal device of a user.

25. The diagnostic device according to claim 23 or 24, further comprising a control unit that controls the storage battery based on the result of the labeling.

26. A diagnostic system for a storage battery used in an energy storage system, comprising: a memory device that stores a trained model A that has been trained to output material deterioration information of the storage battery when charge / discharge data is input; an acquisition device that acquires charge / discharge data of the storage battery; a diagnostic device that diagnoses the material deterioration information of the storage battery obtained by inputting the charge / discharge data of the storage battery to the trained model A; and an output device that outputs the material deterioration information obtained from the diagnostic device as a diagnostic result.

27. The diagnostic system of claim 26, further comprising a calculation device that estimates a state of health (SoH) of the storage battery based on charge / discharge data of the storage battery, and calculates a degradation state of the storage battery based on the SoH and the material degradation information, and the output device outputs the degradation state as a diagnostic result.

28. The diagnostic system of claim 27, wherein the calculation device is configured to calculate a lithium precipitation risk X of the storage battery based on information including the material degradation information, and to label a degradation state of the storage battery based on one or more predetermined thresholds for each of the SoH and the lithium precipitation risk X, and the diagnostic result includes a result of the labeling.

29. The diagnostic system of claim 28, wherein the diagnostic system further includes a display device, the output device transmits the diagnostic result to the display device, and the display device displays the diagnostic result.

30. The diagnostic system according to claim 28 or 29, further comprising a control device for controlling the storage battery based on the result of the labeling.

31. A diagnostic system for a storage battery used in an energy storage system, comprising a server and a diagnostic device, wherein the server has: a memory unit that stores a trained model A that has been trained to output material deterioration information of the storage battery when charge / discharge data is input; and a transmission unit that transmits the trained model A stored in the memory unit to a diagnostic device that diagnoses the material deterioration information of the storage battery; and the diagnostic device has: an acquisition unit that acquires charge / discharge data of the storage battery; a receiving unit that receives the trained model A; a diagnostic unit that diagnoses material deterioration information of the storage battery obtained by inputting the charge / discharge data of the storage battery to the received trained model A; and an output unit that outputs the material deterioration information obtained from the diagnostic unit as a diagnostic result.

32. The diagnostic system of claim 31, wherein the server further includes a calculation unit that estimates a state of health (SoH) of the storage battery based on charging and discharging data of the storage battery, and calculates a degradation state of the storage battery based on the SoH and the material degradation information, and the output unit outputs the degradation state as a diagnostic result.

33. The diagnostic system of claim 32, wherein the calculation unit is configured to calculate a lithium precipitation risk X of the storage battery based on information including the material degradation information, and to label a degradation state of the storage battery based on one or more predetermined threshold values ​​for each of the SoH and the lithium precipitation risk X, and the diagnostic result includes a result of the labeling.

34. The diagnostic system according to claim 33, further comprising a display device, the output section transmitting the diagnostic result to the display device, and the display device displaying the diagnostic result.

35. The diagnostic system according to claim 33 or 34, further comprising a control unit that controls the storage battery based on the result of the labeling.

36. A diagnostic method for a storage battery used in an energy storage system, the method comprising the steps of: acquiring charge / discharge data of the storage battery; acquiring material deterioration information of the storage battery by inputting the charge / discharge data of the storage battery into a trained model A trained to output material deterioration information of the storage battery when charge / discharge data is input; diagnosing the obtained material deterioration information of the storage battery; and outputting the material deterioration information as a diagnostic result.

37. The diagnostic method according to claim 36, further comprising the steps of: estimating a state of health (SoH) of the storage battery based on charge / discharge data of the storage battery; calculating a degradation state of the storage battery based on the SoH and the material degradation information; and outputting the degradation state as a diagnostic result.

38. The diagnostic method according to claim 37, wherein the calculation of the degradation state includes calculating a lithium precipitation risk X of the storage battery based on information including the material degradation information, and labeling the degradation state of the storage battery based on one or more predetermined threshold values ​​for each of the SoH and the lithium precipitation risk X, and the diagnostic result includes a result of the labeling.

39. The diagnostic method according to claim 38, further comprising the step of displaying the outputted diagnostic result on a display unit.

40. The diagnostic method according to claim 38 or 39, further comprising a step of controlling the storage battery based on the result of the labeling.

41. A diagnostic program for a storage battery used in an energy storage system, the diagnostic program causing a computer to execute the following steps: acquire charge / discharge data of the storage battery; diagnose material deterioration information of the storage battery obtained by inputting the charge / discharge data of the storage battery into a trained model A trained to output material deterioration information of the storage battery when charge / discharge data is input; and output the material deterioration information as a diagnostic result.

42. The diagnostic program of claim 41, further causing the computer to execute the following: estimating a state of health (SoH) of the storage battery based on charge and discharge data of the storage battery; calculating a deterioration state of the storage battery based on the SoH and material deterioration information of the storage battery obtained by inputting the charge and discharge data of the storage battery into a trained model A trained to output material deterioration information of the storage battery when charge and discharge data is input; and outputting the deterioration state as a diagnostic result.

43. The diagnostic program of claim 42, further causing the computer to calculate a lithium precipitation risk X of the storage battery based on information including the material degradation information, and label a degradation state of the storage battery based on one or more predetermined thresholds for each of the SoH and the lithium precipitation risk X, and the diagnostic result includes a result of the labeling.

44. The diagnostic program according to claim 43, further causing the computer to display the outputted diagnostic results on a display unit.

45. The diagnostic program according to claim 43 or 44, further causing a computer to control the storage battery based on a result of the labeling.

Citation Information

Patent Citations

  • Battery safety control method and device

    CN116298910A

  • Manufacturing method of secondary battery, protection circuit operation check system, and secondary battery that intentionally generates micro short circuit

    JP2020149937A

  • Lithium ion secondary battery state determination device

    JP2024075195A

  • Storage battery evaluating device, power storage system and storage battery evaluating method

    WO2017022037A1

Cited By

  • Battery lithium precipitation quality determination method and module, and model training method and module

    CN119782745A

  • Power grid regulation system based on big data

    CN119921319A

  • Battery degradation analysis support device and battery degradation analysis support method

    JP7795676B1