Lifespan prediction device and lifespan prediction method

Abstract

A controller according to the present invention acquires initial data of an initial state of a secondary battery, acquires deterioration data of a deterioration state of the secondary battery used for a certain period of time, generates a state image in which at least the initial data and the deterioration data are expressed, inputs the state image into a trained neural network model trained so as to output a lifespan using the state image as input data, and predicts the lifespan by outputting the lifespan from the trained neural network model.

Description

Life prediction device and life prediction method

[0001] The present invention relates to a lifespan prediction device and a lifespan prediction method for predicting the lifespan of a secondary battery.

[0002] A technology is known in which machine learning of a neural network model is performed using a training image generated based on the AC impedance measurement results of a secondary battery, and the trained neural network model is used to estimate the full charge capacity of the secondary battery from the AC impedance measurement results (Patent Document 1).

[0003] Japanese Patent Application Laid-Open No. 2020-20604

[0004] However, when predicting the lifespan of a secondary battery using the technology described in Patent Document 1, a large number of measurements and learning images are required for each degree of deterioration of the secondary battery in order to improve prediction accuracy, which results in the problem that prediction accuracy cannot be improved efficiently.

[0005] The problem to be solved by the present invention is to provide a lifespan prediction method and a lifespan prediction device that can more efficiently improve the prediction accuracy of the lifespan of a secondary battery.

[0006] The present invention solves the above problem by obtaining initial data on the initial state of a secondary battery, obtaining degradation data on the degradation state of a secondary battery that has been used for a certain period of time, generating a state image representing the initial data and degradation data, inputting the state image into a trained neural network model that has been trained to output a lifespan using the state image as input data, and outputting the lifespan from the trained neural network model to predict the lifespan.

[0007] According to the present invention, the accuracy of predicting the life of a secondary battery can be improved more efficiently.

[0008] Fig. 1 is a block diagram of a lifespan prediction device according to an embodiment of the present invention. Fig. 2 is a diagram showing an example of a state image of a secondary battery in this embodiment. Fig. 3 is a diagram showing a modified example of the state image of a secondary battery in this embodiment. Fig. 4 is an example of training data used for training a neural network model in this embodiment. Fig. 5 is a flowchart showing the control procedure of a lifespan prediction method executed by a controller according to this embodiment.

[0009] Embodiments of a lifespan prediction device and a lifespan prediction method according to the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of a lifespan prediction device 1 according to one embodiment of the present invention. In this embodiment, the lifespan prediction device and the lifespan prediction method will be described based on an example in which the lifespan prediction device 1 is provided in a lifespan prediction system 100. However, the present invention is not limited to this example, and the lifespan prediction device 1 may be any device that predicts the lifespan of a secondary battery based on the deterioration state (degree of deterioration) of the secondary battery. The lifespan prediction system 100 is a system mounted on a vehicle such as an electric vehicle or a hybrid vehicle, and includes a lifespan prediction device 1 and a sensor 2. The lifespan prediction device 1 and the sensor 2 are connected via wire or wirelessly so as to be able to exchange information. The sensor 2 is attached to a secondary battery (not shown) whose lifespan is to be predicted, and detects the state of the secondary battery. The secondary battery is an automotive battery mounted on the vehicle.

[0010] The secondary battery is, for example, a lithium-ion secondary battery. The secondary battery includes a power generating element formed by stacking electrode layers (positive and negative electrode layers) and separators and filling them with an electrolyte, a positive electrode tab connected to the positive electrode layer, a negative electrode tab connected to the negative electrode layer, and an exterior member that houses and seals these. The life prediction device 1 acquires data related to the physical properties of the secondary battery from a sensor 2 and predicts the life of the secondary battery based on the acquired data on the secondary battery. The secondary battery is not limited to a vehicle, and may be installed in any device that operates using the secondary battery as a power source. The life prediction device 1 is not limited to being installed in a device equipped with a secondary battery, such as a vehicle, and may be a cloud server located outside the device equipped with the secondary battery. In this embodiment, the secondary battery whose life is to be predicted is also referred to as a target secondary battery.

[0011] The lifespan prediction device 1 includes a controller 10, a memory 20, and a display device 30. The controller 10 is a processor and includes a ROM (Read Only Memory) storing a program, a CPU (Central Processing Unit) that executes the program stored in the ROM, and a RAM (Random Access Memory) that functions as an accessible storage device. The controller 10 includes functional blocks: a data acquisition unit 11, an image generation unit 12, and a prediction unit 13. The controller 10 executes the functions of the data acquisition unit 11, the image generation unit 12, and the prediction unit 13 by using the CPU to execute the program stored in the ROM. Details of each functional block, such as the data acquisition unit 11, will be described later. The controller 10 is connected to the memory 20 and the display device 30 via a wired or wireless connection so that information can be exchanged between them. The memory 20 and the display device 30 may not necessarily be provided in the lifespan prediction device 1 but may be located outside the lifespan prediction device 1.

[0012] The memory 20 is a storage medium for storing various data. The memory 20 stores initial data of the initial state of the secondary battery. The initial state is the state before deterioration. The initial data is data of physical quantities indicating the physical properties of the secondary battery, which is obtained when the secondary battery in the initial state is charged and discharged. The data of physical quantities indicating the physical properties of the secondary battery in the initial state includes, for example, data related to the temperature, current, voltage, internal resistance, OCV, capacity, and charge / discharge curve when the secondary battery is charged and discharged. In other words, the initial data includes data obtained by measuring changes in these physical quantities when the secondary battery in the initial state is charged and discharged. When the secondary battery is in the initial state, the initial data is obtained and stored in the memory 20.

[0013] The memory 20 may store degradation data on the degradation state of a secondary battery used for a certain period of time. The degradation state refers to the state after degradation. The degradation data is data on physical quantities indicating the physical state of a secondary battery acquired when a secondary battery used for a certain period of time is charged and discharged. The data on physical quantities indicating the physical state of a secondary battery used for a certain period of time includes, for example, data on the temperature, current, voltage, internal resistance, OCV, capacity, and charge / discharge curve of the secondary battery during charging and discharging. In other words, the degradation data includes data measuring changes in these physical quantities when a secondary battery used for a certain period of time is charged and discharged. When degradation data is acquired by the data acquisition unit 11 after the secondary battery is first used, the degradation data is stored in the memory 20. For example, degradation data acquired for each cycle of the secondary battery during use of the secondary battery may be stored in chronological order. A cycle of a secondary battery refers to the period from charging once to discharging once. The number of cycles refers to the number of cycles, where one cycle is defined as charging once to discharging once. Note that storing degradation data is not a required configuration and may be applied as needed.

[0014] The display device 30 is a device such as a liquid crystal display or a head-up display (HUD) that provides information to a user such as a vehicle occupant. The display device 30 displays an image when it receives a control instruction to display an image from the controller 10. In this embodiment, the display device 30 displays an image including the life of a secondary battery. The display device 30 may also display the vehicle's cruising range. The display device 30 may also include an input device that allows the vehicle occupant to input instructions. Examples of the input device include a touch panel and a switch. The input device accepts instructions input by the user. The display device 30 may also include a speaker as an output device.

[0015] Sensor 2 is attached to the target secondary battery and detects the physical properties of the target secondary battery. Sensor 2 includes, for example, a voltage sensor, a current sensor, and a temperature sensor. Sensor 2 detects the physical properties of the target secondary battery at regular intervals. In this embodiment, sensor 2 detects the physical properties of the target secondary battery at regular intervals while the target secondary battery is in an initial state and while the target secondary battery is being charged or discharged after being used for a certain period of time. The detection value of sensor 2 is output to controller 10.

[0016] Next, the functions of the functional blocks included in the controller 10 will be described. The data acquisition unit 11 acquires initial data for the initial state of the target secondary battery. Specifically, the data acquisition unit 11 acquires, from the detection values ​​of the sensor 2, data indicating the physical properties of the target secondary battery in the initial state when the target secondary battery is being charged or discharged, as initial data. The initial data acquired by the data acquisition unit 11 is stored in the memory 20.

[0017] The data acquisition unit 11 also acquires degradation data on the degradation state of a secondary battery that has been used for a certain period of time. For example, after the secondary battery begins to be used, the data acquisition unit 11 acquires, as degradation data, data indicating the physical properties of the secondary battery when the secondary battery in a degraded state is being charged or discharged from the detection value of the sensor 2 at a predetermined timing. The predetermined timing is, for example, when a user inputs an instruction to display the life of the secondary battery. For example, in this embodiment, the latest degradation data is acquired when the user inputs an instruction to display the life of the secondary battery. The predetermined timing may also be a regular cycle. For example, the regular cycle may be every cycle of the secondary battery. The predetermined timing may also be a timing arbitrarily set by the user, for example, when the number of cycles of the secondary battery reaches a preset number.

[0018] Furthermore, in this embodiment, the data acquiring unit 11 is not limited to acquiring the latest degradation data at a predetermined timing, but may also acquire past degradation data stored in the memory 20. That is, the acquired degradation data may include not only degradation data at the current time point but also degradation data at a certain time point in the past. Furthermore, the acquired degradation data may include not only degradation data at a single time point but also degradation data at a plurality of different time points.

[0019] Hereinafter, this embodiment will be described using data on the charge / discharge curve of a secondary battery as an example of the acquired initial data and deterioration data, which is data on physical quantities indicating the physical state of a secondary battery. A charge / discharge curve of a secondary battery is a curve with the horizontal axis representing capacity and the vertical axis representing voltage. The acquired data on the charge / discharge curve of a secondary battery includes at least one of the discharge curve and the charge curve of the secondary battery. The acquired data on the charge / discharge curve of a secondary battery may be data on the charge / discharge curve at a plurality of different points in time.

[0020] In this embodiment, the acquired charge / discharge curve data may be limited to charge / discharge curve data within a predetermined range of SOC. The predetermined range is a range within which secondary battery data can be easily obtained during charging / discharging and is set arbitrarily by the user, but is preferably a range of SOC from 30% to 80%. For example, the data acquisition unit 11 acquires charge data within a predetermined range of SOC for a secondary battery in an initial state as initial data. Furthermore, the data acquisition unit 11 acquires charge data within a predetermined range of SOC for a secondary battery that has been used for a certain period as degradation data. In this way, the data acquisition unit 11 acquires degradation data on the degradation state of the battery after being used for a certain period of time with a single charge.

[0021] The image generation unit 12 generates a status image that represents at least the initial data and the degradation data. That is, the status image is an image in which the initial data and the degradation data are visually depicted on a single image, and includes graphics, such as graphs, that represent each piece of data. For example, if the initial data and the degradation data are charge / discharge curve data, the status image is an image in which the charge / discharge curve of the target secondary battery in its initial state and the charge / discharge curve of the target secondary battery in its degraded state are depicted on the same graph. In this embodiment, the image generation unit 12 generates the status image of the target secondary battery based on the initial data of the target secondary battery stored in the memory 20 and the degradation data of the target secondary battery acquired by the data acquisition unit 11.

[0022] An example of a status image in this embodiment will be described below with reference to FIG. 2 . FIG. 2 is a diagram illustrating an example of a status image of a secondary battery in this embodiment. (A) of FIG. 2 shows charge curves (L1, L2) of a secondary battery in an initial state (BoL) and a secondary battery in a degraded state (MoL). The initial state (BoL) secondary battery is the secondary battery at an initial time (t0). The degraded state (MoL) secondary battery is the secondary battery at a predetermined timing (tn) after the secondary battery has started to be used. (B) of FIG. 2 shows discharge curves (L3, L4) of a secondary battery in an initial state (BoL) and a secondary battery in a degraded state (MoL). (C) of FIG. 2 shows the charge curve (L1) of the secondary battery in the initial state and the charge curve (L2) of the secondary battery in a degraded state in (A) of FIG. 2 , and the discharge curve (L3) of the secondary battery in the initial state and the discharge curve (L4) of the secondary battery in a degraded state in (B) of FIG. 2 , in a single graph.

[0023] In the present embodiment, the image generation unit 12 may also perform processing of the initial data and the degradation data to generate a status image showing the initial data and the degradation data after the processing. The processing is processing of data characteristics through mathematical operations. For example, the image generation unit 12 performs a differential operation on the initial data and the degradation data. In the differential operation, a dQ / dV curve is created from the charge / discharge curve of the secondary battery. The image generation unit 12 generates a status image showing the dQ / dV curve of the secondary battery in the initial state and the dQ / dV curve of the secondary battery in the degradation state created by the differential operation. The processing is not limited to a differential operation and may include, for example, addition, subtraction, multiplication, division, integration, and exponentiation.

[0024] Here, a modified example of the status image according to this embodiment will be described with reference to FIG. 3 . FIG. 3 is a diagram illustrating a modified example of the status image according to this embodiment. In FIG. 3A , a charge curve (L1) of a secondary battery in an initial state and charge curves (L21, L22) of secondary batteries in multiple states of degradation, as well as a discharge curve (L3) of a secondary battery in an initial state and discharge curves (L41, L42) of secondary batteries in multiple states of degradation are plotted in a single graph. The charge curves (L21, L22) of secondary batteries in multiple states of degradation are charge curves obtained when the secondary batteries are charged at different points in time. The discharge curves (L41, L42) of secondary batteries in multiple states of degradation are charge curves obtained when the secondary batteries are discharged at different points in time. Charging and discharging a secondary battery at different points in time refers to, for example, charging and discharging a secondary battery in different cycles. In this embodiment, by displaying degradation data at multiple different points in time on a single image, differences in degradation over time can be reflected as input data. The number of pieces of degraded data displayed in one image is not limited to two pieces of degraded data, but may be three or more pieces of degraded data.

[0025] In FIG. 3B, charge curves for each state within a predetermined range of SOC are plotted based on FIG. 2A. As indicated by the lengths of the dashed arrows, predetermined ranges (Br, Mr) are set for the charge curves (L1, L2) of the secondary battery in the initial state (BoL) and the secondary battery in the deteriorated state (MoL). The predetermined range is, for example, 30% to 80% SOC. In FIG. 3C, a dQ / dV curve (L5) created from initial data for the initial state (BoL) and a dQ / dV curve (L6) created from deterioration data for the deteriorated state (MoL) are plotted in a single graph.

[0026] The prediction unit 13 predicts the life of the target secondary battery based on the acquired data regarding the physical properties of the target secondary battery. In this embodiment, the life of the secondary battery is indicated by the number of cycles of the secondary battery. Specifically, the prediction unit 13 predicts the life of the target secondary battery by inputting a state image of the target secondary battery into a trained neural network model and causing the trained neural network model to output the life of the target secondary battery. The trained neural network model is a model that has been trained to output the life of the secondary battery using the state image of the secondary battery as input data. The prediction unit 13 displays the predicted life of the target secondary battery to the user via the display device 30. Specifically, the prediction unit 13 transmits a control instruction to the display device 30 to display an image including the life of the target secondary battery.

[0027] The trained neural network model is configured by a neural network including an input layer to which input data including a state image of the secondary battery is input, at least one intermediate layer, and an output layer that outputs output data including the life of the secondary battery. The trained neural network model is subjected to machine learning to predict the life of the secondary battery using training data that associates input data with output data. When input data is input to the input layer, the trained neural network model causes a computer to function so that output data corresponding to the input data is output from the output layer. Images such as the specific examples described in FIGS. 2 and 3 are also used as the state images for learning, as appropriate. Note that in this embodiment, the trained neural network model may be generated in the life prediction device 1, or an externally generated trained neural network model may be stored in the life prediction device 1. Furthermore, the trained neural network model is not limited to being stored in the life prediction device 1, but may also be stored on an external cloud server or the like, and the life prediction device 1 may access the trained neural network model via communication as appropriate.

[0028] Here, learning of the neural network model in this embodiment will be described with reference to FIG. 4 . FIG. 4 is an example of training data used in learning of the neural network model in this embodiment. FIG. 4 shows a specific example of training data in which data on state images of secondary batteries are associated with the lifespans of the secondary batteries, and a specific graph for each state image. In the example of FIG. 4 , a lifespan of 1000 cycles for secondary battery A, 700 cycles for secondary battery B, and 400 cycles for secondary battery C are associated with state images of secondary battery A, secondary battery B, and secondary battery C, respectively. Secondary battery A deteriorates more slowly than secondary battery B. Secondary battery C deteriorates more quickly than secondary battery B. The degree of deterioration of secondary batteries varies depending on how they are used. These state images and lifespan data of secondary batteries are data for training secondary batteries, and may include data obtained from secondary batteries that have actually been used and deteriorated, data obtained from secondary batteries used in experiments, and data obtained by simulation. The state images of the secondary batteries used for learning are generated from degradation data of secondary batteries that have reached a predetermined number of cycles. Images A, B, and C depict the discharge curves (L3) of the secondary batteries A, B, and C in their initial states and their discharge curves (L4) in their deteriorated states, respectively, on the same image.

[0029] The trained neural network model according to this embodiment is trained using training data in which such state images and life spans are associated. This allows the difference in changes in the physical properties of a secondary battery over time to be linked to the life span of the secondary battery, thereby enabling the construction of a model with improved life span prediction accuracy. As shown in the example of FIG. 4 , the state image of the target secondary battery is assumed to be state image X. In this case, the prediction unit 13 inputs state image X of the target secondary battery into the trained neural network model and causes the trained neural network model to output the life span (e.g., 500 cycles) of the target secondary battery, thereby predicting the life span of the target secondary battery. In this embodiment, since both data sets, before and after degradation, are included in a single image, the amount of information to be input can be increased without increasing the number of images for each data set. Furthermore, since information on the difference in changes over time from the initial state until a certain period of use is represented in a single image, the amount of information input in a single image can be increased compared to representing the initial data and degradation data in separate images. This allows the construction of a model with improved prediction accuracy more efficiently. In this embodiment, the characteristics of the initial data and the degraded data are expressed independently on a single image, which allows for a wider variety of combinations and less loss of information than when differential data is created from the initial data and the degraded data, thereby increasing the amount of information to be input.

[0030] The prediction unit 13 may also calculate a cruising range of the vehicle based on the predicted lifespan. The cruising range is the distance the vehicle can travel from the current time until the number of cycles of the secondary battery reaches the predicted lifespan (e.g., 550 cycles). For example, the prediction unit 13 calculates the cruising range based on the lifespan, average power consumption, and battery capacity of the secondary battery. If the lifespan of the secondary battery is 550 cycles, the average power consumption is 6 km / kWh, and the battery capacity is 60 kWh, the cruising range is 198,000 km. The prediction unit 13 may also calculate the cruising range based on the lifespan of the secondary battery and the theoretical cruising range when fully charged. If the lifespan of the secondary battery is 550 cycles and the theoretical cruising range when fully charged is 400 km, the cruising range is 220,000 km. The prediction unit 13 displays the cruising range to the user via the display device 30. For example, the prediction unit 13 transmits a control instruction to the display device 30 to display an image including the cruising range. When the life of the secondary battery is predicted periodically, the cruising range is also calculated and displayed each time. This allows the user to check the progress of the cruising range. The current mileage of the vehicle may also be displayed together with the cruising range.

[0031] Next, a processing flow of a life prediction method for predicting the life of a secondary battery by the controller 10 will be described with reference to Fig. 5. Fig. 5 is a flowchart showing the control procedure of the life prediction method executed by the controller according to this embodiment.

[0032] In step S1, the controller 10 acquires initial data for a target secondary battery that is the target of life prediction. The initial data is, for example, data on the charge and discharge curve of the target secondary battery in an initial state. The controller 10 acquires the initial data before the target secondary battery deteriorates. The acquired initial data is stored in memory 20. In step S2, the controller 10 acquires deterioration data for a target secondary battery that has been used for a certain period of time. The deterioration data is, for example, data on the charge and discharge curve of the target secondary battery that has been used for a certain period of time. In step S3, the controller 10 generates a status image that shows the initial data and the deterioration data. For example, the status image is an image in which the charge and discharge curve of the target secondary battery in an initial state and the charge and discharge curve of the target secondary battery in a deteriorated state are plotted on the same graph.

[0033] In step S4, controller 10 inputs the state image generated in step S3 into the trained neural network model. The trained neural network model is a model that has been trained to output the life of a secondary battery using the state image of the secondary battery as input data. In step S5, controller 10 predicts the life of the target secondary battery by outputting the life of the target secondary battery from the trained neural network model. The predicted life of the target secondary battery is displayed to the user via display device 30.

[0034] As described above, in the lifespan prediction method and lifespan prediction device according to this embodiment, the controller acquires initial data on the initial state of the secondary battery, acquires degradation data on the degradation state of the secondary battery after it has been used for a certain period of time, generates a state image representing at least the initial data and the degradation data, inputs the state image into a trained neural network model, and predicts the lifespan by having the trained neural network model output the lifespan, the trained neural network model being a model trained to use the state image as input data and output the lifespan. This makes it possible to more efficiently improve the prediction accuracy of the lifespan of the secondary battery.

[0035] In the lifespan prediction method and lifespan prediction device according to this embodiment, the initial data and degradation data are data of physical quantities that indicate the physical properties of the secondary battery. This allows the physical properties of the secondary battery over time to be linked to the lifespan of the secondary battery, thereby enabling the construction of a neural network model with improved prediction accuracy.

[0036] In the lifespan prediction method and lifespan prediction device according to the present embodiment, the data on physical quantities indicating the physical properties of the secondary battery includes data on the charge and discharge curve of the secondary battery. This allows the charge and discharge curve of the secondary battery over time to be linked to the lifespan of the secondary battery, thereby enabling the construction of a neural network model with improved prediction accuracy.

[0037] In the lifespan prediction method and lifespan prediction device according to this embodiment, the controller acquires charge data of a secondary battery in an initial state within a predetermined range of SOC as initial data, and acquires charge data of a secondary battery that has been used for a certain period of time within a predetermined range of SOC as degradation data. This makes it possible to acquire the secondary battery data required for prediction without any special processing.

[0038] In the lifespan prediction method and device according to this embodiment, the controller processes the initial data and degradation data and generates a status image showing the processed initial data and degradation data. This allows a neural network model to be built with improved prediction accuracy by learning using information that emphasizes subtle differences between the initial data and degradation data.

[0039] In the lifespan prediction method and device according to this embodiment, the controller performs a differential operation on the initial data and the degradation data to generate a status image representing the initial data and the degradation data after the differential operation. This allows a neural network model to be constructed with improved prediction accuracy by learning using information that emphasizes subtle differences between the initial data and the degradation data.

[0040] In the lifespan prediction method and lifespan prediction device according to this embodiment, the secondary battery is an on-board battery mounted in a vehicle, and the controller calculates the cruising range that the vehicle can travel based on the predicted lifespan and outputs the calculated cruising range to the user of the vehicle, thereby enabling the user to grasp the cruising range.

[0041] Although the embodiments of the present invention have been described above, these embodiments are described to facilitate understanding of the present invention and are not described to limit the present invention. Therefore, each element disclosed in the above embodiments is intended to include all design modifications and equivalents that fall within the technical scope of the present invention.

[0042] DESCRIPTION OF SYMBOLS 1... Life prediction device 10... Controller 11... Data acquisition unit 12... Image generation unit 13... Prediction unit 20... Memory 30... Display device

Claims

1. A lifespan prediction method executed by a controller for predicting the lifespan of a secondary battery, wherein the controller: acquires initial data on the initial state of the secondary battery; acquires degradation data on the degradation state of the secondary battery after it has been used for a certain period of time; generates a state image representing at least the initial data and the degradation data; inputs the state image into a trained neural network model; and predicts the lifespan by having the trained neural network model output the lifespan, wherein the trained neural network model is a model that has been trained to use the state image as input data and output the lifespan.

2. A life prediction method according to claim 1, wherein the initial data and the deterioration data are data of physical quantities that indicate the physical properties of the secondary battery.

3. A life prediction method according to claim 2, wherein the data on physical quantities indicating the physical properties of the secondary battery includes data on the charge / discharge curve of the secondary battery.

4. A lifespan prediction method as claimed in claim 3, wherein the controller acquires charging data of the secondary battery in the initial state within a predetermined range of SOC as the initial data, and acquires charging data of the secondary battery that has been used for a certain period of time within the predetermined range of SOC as the degradation data.

5. A lifespan prediction method according to any one of claims 1 to 4, wherein the controller processes the initial data and the deteriorated data, and generates the status image showing the initial data and the deteriorated data after the processing.

6. A lifespan prediction method according to any one of claims 1 to 5, wherein the controller performs a differential operation on the initial data and the deterioration data, and generates the status image representing the initial data and the deterioration data after the differential operation.

7. A lifespan prediction method according to any one of claims 1 to 6, wherein the secondary battery is an on-board battery mounted in a vehicle, and the controller calculates a cruising range that the vehicle can travel based on the predicted lifespan, and outputs the calculated cruising range to a user of the vehicle.

8. A lifespan prediction device comprising a controller for predicting the lifespan of a secondary battery, wherein the controller: acquires initial data on the initial state of the secondary battery; acquires degradation data on the degradation state of the secondary battery after use for a certain period of time; generates a state image representing at least the initial data and the degradation data; inputs the state image into a trained neural network model; and predicts the lifespan by outputting the lifespan from the trained neural network model; the trained neural network model is a model that has been trained to use the state image as input data and output the lifespan.

Images