Method and apparatus for analyzing battery

Abstract

A method and an apparatus for analyzing a battery are disclosed. The battery analysis apparatus applies multiple AC signals having different frequencies to a battery so as to identify impedance for each frequency, identifies a temperature of a place in which a battery impedance is measured, and inputs a tensor including the impedance for each frequency and the temperature to an artificial intelligence model so as to identify whether the battery is defective.

Description

Battery analysis method and apparatus therefor

[0001] An embodiment of the present invention relates to a battery analysis method and a device thereof, and more particularly, to a battery analysis method and a device thereof for determining whether a battery is defective by taking temperature into consideration.

[0002] Electrochemical Impedance Spectroscopy (EIS) is a method for analyzing resistance characteristics by applying AC electrical signals of various frequencies to a battery and measuring its response. The time required to analyze resistance characteristics is inversely proportional to the frequency, and particularly at low frequencies (below a few Hz), analysis of resistance characteristics can take considerable time.

[0003] Another battery testing method is alternating current internal resistance (AC-IR). AC-IR typically measures the battery's response to a 1 kHz AC signal, without changing the frequency.

[0004] The technical problem to be achieved by an embodiment of the present invention is to provide a battery analysis method and device capable of accurately determining whether a battery is defective by considering the characteristic that the pattern of battery impedance changes depending on the temperature of the battery impedance measurement location.

[0005] In order to achieve the above technical task, an example of a battery analysis method according to an embodiment of the present invention includes the steps of: applying a plurality of AC signals having different frequencies to a battery to determine impedance for each frequency; determining the temperature of a location where the battery impedance is measured; and inputting a tensor including the impedance for each frequency and the temperature into an artificial intelligence model to determine whether the battery is defective.

[0006] In order to achieve the above technical task, an example of a battery analysis device according to an embodiment of the present invention includes an impedance identification unit that applies a plurality of AC signals having different frequencies to a battery to identify impedance by frequency; a temperature identification unit that identifies the temperature of a location where battery impedance is measured; and an analysis unit that inputs a tensor including the impedance by frequency and the temperature into an artificial intelligence model to identify whether the battery is defective.

[0007] According to an embodiment of the present invention, it is possible to accurately determine whether a battery is defective by considering changes in the battery impedance pattern according to the temperature of the measurement location.

[0008] FIG. 1 is a drawing showing an example of a battery analysis device according to an embodiment of the present invention;

[0009] FIG. 2 is a diagram illustrating an example of a method for measuring frequency-dependent impedance information of a battery according to an embodiment of the present invention.

[0010] FIG. 3 is a diagram illustrating an example of an artificial intelligence model for determining whether a battery is defective according to an embodiment of the present invention.

[0011] Figures 4 and 5 are diagrams showing experimental examples of frequency-dependent impedance changes according to temperature.

[0012] FIG. 6 is a flowchart illustrating an example of a battery analysis method according to an embodiment of the present invention;

[0013] FIG. 7 is a diagram illustrating an example of a learning method of an artificial intelligence model according to an embodiment of the present invention, and

[0014] FIG. 8 is a diagram illustrating an example configuration of a battery analysis device according to an embodiment of the present invention.

[0015] Hereinafter, a battery analysis method and device according to an embodiment of the present invention will be described in detail with reference to the attached drawings.

[0016] FIG. 1 is a drawing illustrating an example of a battery analysis device according to an embodiment of the present invention.

[0017] Referring to FIG. 1, the battery analysis device (100) receives frequency-dependent impedance information (e.g., EIS information (110)) and temperature information (112) of the battery, and determines whether the battery is defective (120) and outputs the information. The battery analysis device determines whether the battery is defective using an artificial intelligence model, and this will be discussed in detail in FIG. 3 and below.

[0018] FIG. 2 is a diagram illustrating an example of a method for measuring frequency-dependent impedance information of a battery according to an embodiment of the present invention.

[0019] Referring to FIG. 2, the battery analysis device (100) sequentially applies AC signals having multiple different frequencies (210) to the battery (200) to measure the frequency-dependent impedance (220) of the battery. The battery is a rechargeable secondary battery. The battery analysis device (100) can measure the frequency-dependent impedance using the EIS method.

[0020] In another embodiment, the battery analysis device (100) may not directly measure the frequency-dependent impedance of the battery (200), but may receive frequency-dependent impedance information of the battery from a separate EIS measurement device (not shown). However, for the convenience of explanation, the following embodiment assumes that the battery analysis device (100) directly measures the frequency-dependent impedance of the battery (200).

[0021] FIG. 3 is a diagram illustrating an example of an artificial intelligence model for determining whether a battery is defective according to an embodiment of the present invention.

[0022] Referring to Fig. 3, the artificial intelligence model (300) determines whether the battery is defective (320) based on the distribution pattern (310) of impedance by frequency. The artificial intelligence model (300) can be implemented using various types of conventional artificial neural networks such as a CNN (Convolutional Neural Network). In one embodiment, the impedance by frequency can be displayed as a Nyquist plot on a graph including a real axis and an imaginary axis. The battery impedance distribution (310) may differ depending on the temperature of the location where the impedance by frequency of the battery is measured.

[0023] Figures 4 and 5 are diagrams showing experimental examples of frequency-dependent impedance changes according to temperature.

[0024] Referring to Figure 4, a histogram depicts the number of batteries used in the experiment at different temperatures. Temperatures may vary depending on where the battery impedance is measured. For example, a laboratory environment may have a temperature of 22 to 25°C, while a manufacturing line environment may have a temperature of 27 to 29°C.

[0025] Referring to Fig. 5, the results of measuring frequency-dependent impedance for multiple batteries at different temperatures of Fig. 4 are illustrated. It can be seen that the distribution (510) of impedance for multiple batteries (400) measured in a laboratory environment of Fig. 4 is different from the distribution (520) of impedance for multiple batteries (410) measured in a manufacturing line environment. Therefore, when determining whether a battery is defective simply by using frequency-dependent impedance as in Fig. 3, the results may not be accurate.

[0026] FIG. 6 is a flowchart illustrating an example of a battery analysis method according to an embodiment of the present invention.

[0027] Referring to FIG. 6, the battery analysis device (100) measures the impedance of the battery by frequency (S600). The battery analysis device (100) determines the temperature of the measurement location when measuring the impedance by frequency (S610). In one embodiment, the battery analysis device (100) may measure the temperature of the measurement location using a temperature sensor or the like at the time of measuring the impedance of the battery, or may directly receive temperature information from a user or the like. In another embodiment, the battery analysis device (100) may determine the temperature by frequency. That is, the battery analysis device (100) may measure the temperature when measuring the impedance by applying a first frequency and the temperature when measuring the impedance by applying a second frequency, respectively.

[0028] The battery analysis device (100) inputs a tensor containing frequency-specific impedance information and temperature information of the battery into an artificial intelligence model to determine whether the battery is defective (S620). For example, the battery analysis device (100) can determine whether the battery is defective by inputting impedance information and temperature information (either single temperature information regardless of frequency or frequency-specific temperature information) for the first to Nth frequencies into the artificial intelligence model. In order to determine whether the battery is defective using the artificial intelligence model, the artificial intelligence model must first undergo a training process.

[0029] FIG. 7 is a diagram illustrating an example of a learning method of an artificial intelligence model according to an embodiment of the present invention.

[0030] Referring to Fig. 7, at least one defective battery (700) and at least one normal battery (710) are prepared. The battery analysis device determines frequency-specific impedance information for the defective battery at different temperatures. For example, the battery analysis device (100) determines frequency-specific impedance information (A1) for the defective battery (700) at a first temperature, and determines frequency-specific impedance information (A2) for the defective battery (700) at a second temperature. In addition, the battery analysis device (100) determines frequency-specific impedance information for the normal battery (710) at different temperatures. For example, the battery analysis device (100) determines frequency-specific impedance information (B1) for the normal battery (710) at a first temperature, and determines frequency-specific impedance information (B2) for the normal battery (710) at a second temperature.

[0031] The battery analysis device (100) generates learning data that labels frequency-specific impedance information (A1, A2, B1, B2) measured at multiple different temperatures as whether the battery is defective. For example, for a defective battery (700), frequency-specific impedance information (A1, A2) measured at multiple different temperatures is labeled as 'battery defective', and for a normal battery (710), frequency-specific impedance information (B1, B2) measured at multiple different temperatures is labeled as 'battery normal'.

[0032] The battery analysis device (100) trains an artificial intelligence model using learning data. For example, the battery analysis device (100) inputs frequency-specific impedance information (A1, A2, B1, B2) and temperature information (temperature 1, temperature 2) of the learning data into an artificial intelligence model to produce a predicted value (defective or normal) of whether the battery is defective, and trains the artificial intelligence model to reduce the error by comparing the predicted value with the ground truth of the learning data, i.e., the labeled value (defective or normal). Since the method of training an artificial intelligence model using a supervised learning method using labeled learning data is already a widely known configuration, further description thereof will be omitted.

[0033] FIG. 8 is a diagram illustrating an example configuration of a battery analysis device according to an embodiment of the present invention.

[0034] Referring to FIG. 8, the battery analysis device (800) includes an impedance detection unit (810), a temperature detection unit (820), an analysis unit (830), an artificial intelligence model (840), and a training unit (850). In one embodiment, if the artificial intelligence model (840) has been trained in advance, the training unit (850) may be omitted. In another embodiment, the battery analysis device (800) may be implemented as a computing device including a memory, a processor, and an input / output device. In this case, each configuration may be implemented as software (program), loaded into the memory, and then executed by the processor.

[0035] The impedance detection unit (810) applies multiple AC signals of different frequencies to the battery to detect the impedance at each frequency. An example of a method for detecting the impedance of a battery at each frequency is illustrated in FIG. 2.

[0036] The temperature detection unit (820) detects the temperature of the location where the battery impedance is measured.

[0037] The analysis unit (830) inputs a tensor containing impedance and temperature by frequency into an artificial intelligence model to determine whether the battery is defective. At this time, the artificial intelligence model (840) is a model that has been trained in advance.

[0038] When training of the artificial intelligence model (840) is required, the training unit (850) trains the artificial intelligence model (840). Specifically, the training unit (850) generates training data in which impedances for each frequency and the temperature of the measurement location, measured by applying signals of multiple different frequencies at different temperatures to at least one defective battery and at least one normal battery, are labeled as indicating whether the battery is defective, and trains the artificial intelligence model (840) using the training data. An example of a training method for the artificial intelligence model (840) is illustrated in FIG. 7.

[0039] The present invention can also be implemented as computer-readable program code on a computer-readable recording medium. Computer-readable recording media include all types of recording devices that store data that can be read by a computer system. Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disks, and optical data storage devices. Furthermore, computer-readable recording media can be distributed across network-connected computer systems, allowing computer-readable code to be stored and executed in a distributed manner.

[0040] The present invention has been described above, focusing on preferred embodiments thereof. Those skilled in the art will appreciate that the present invention can be implemented in modified forms without departing from its essential characteristics. Therefore, the disclosed embodiments should be considered illustrative rather than limiting. The scope of the present invention is set forth in the claims, not the foregoing description, and all differences within the scope equivalent thereto should be construed as being encompassed by the present invention.

Claims

1. A step of applying multiple AC signals with different frequencies to a battery to determine impedance by frequency; Step for determining the temperature of the location where the battery impedance is measured; and A battery analysis method, characterized by including a step of inputting a tensor including the impedance and temperature for each frequency into an artificial intelligence model to determine whether the battery is defective.

2. In paragraph 1, Further comprising a step of training the above artificial intelligence model; The above training steps are: A step of generating learning data labeled with the impedance by frequency and the temperature of the measurement location by applying the plurality of AC signals at different temperatures to at least one defective battery and at least one normal battery, as a sign of whether the battery is defective; and A battery analysis method, characterized by including a step of training the artificial intelligence model using the above learning data.

3. An impedance detection unit that applies multiple AC signals with different frequencies to the battery and detects impedance by frequency; A temperature detection unit that detects the temperature of the location where the battery impedance is measured; and A battery analysis device characterized by including an analysis unit that inputs a tensor including the impedance and temperature for each frequency into an artificial intelligence model to determine whether the battery is defective.

4. In paragraph 3, Further comprising a training unit for training the above artificial intelligence model; The training unit generates learning data labeling the impedance by frequency and the temperature of the measurement location measured by applying the plurality of AC signals at different temperatures to at least one defective battery and at least one normal battery, and trains the artificial intelligence model using the learning data. A battery analysis device.

5. A computer-readable recording medium having recorded thereon a computer program for performing the method described in Article 1.

Images