Battery leakage identification method and device, computer equipment and storage medium

Abstract

The invention relates to a battery leakage identification method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring response value data sets output by various gas sensors to different gases to be detected in an effective response period, the effective response period being a preset time period from a target sampling moment, the target sampling moment is a moment when the change amplitude of the response value of at least one gas sensor is detected to exceed a preset amplitude threshold value for the first time, extracting change characteristic data of the response value of each gas sensor based on the response value data set, and performing gas classification identification on the change characteristic data of the response value of each gas sensor, and obtaining a gas identification result for detecting whether the battery leaks or not. By adopting the method, the gas identification time can be remarkably shortened, and the influence of the baseline drift phenomenon of the gas sensor on the identification result can be reduced, so that the gas identification result is more accurate.

Description

Technical Field

[0001] The present application relates to the field of battery monitoring technology, and in particular to a battery leakage identification method, device, computer equipment, storage medium and computer program product. Background Art

[0002] Battery leakage refers to the phenomenon that liquid, gas or other substances overflow from the inside of the battery. Factors such as overcharging and discharging, high temperature, mechanical damage or improper use may increase the possibility of battery leakage. In order to detect battery leakage in time and take measures to reduce the occurrence of safety accidents within an effective time, it is very necessary to monitor the battery leakage during battery use.

[0003] At present, commonly used battery leakage identification and detection solutions mainly include spectral or mass spectrometry analysis, the use of sensors or corresponding detection devices, while the detection and classification methods often rely on large-scale detection equipment, which are complex and time-consuming to operate, have slow classification responses, and are still difficult to effectively detect whether the battery has leakage, etc., and repeated detection is required. It can be seen that the current battery leakage identification and detection solutions have the problem of low recognition efficiency. Summary of the invention

[0004] Based on this, it is necessary to provide a battery leakage identification method, apparatus, computer equipment, computer-readable storage medium and computer program product that can improve the efficiency of battery leakage identification in order to solve the above technical problems.

[0005] In a first aspect, the present application provides a battery leakage identification method. The method comprises:

[0006] Acquire a data set of response values ​​output by multiple gas sensors to different gases to be tested during an effective response period, where the effective response period is a preset time period after a target sampling time, and the target sampling time is the time when the change amplitude of the response value of at least one gas sensor is first detected to exceed a preset amplitude threshold;

[0007] Extracting the change characteristic data of the response value of each gas sensor based on the response value data set;

[0008] The change characteristic data of the response value of each gas sensor is used to classify and identify the gas to obtain a gas identification result, which is used to determine whether the battery is leaking.

[0009] In the technical solution of the embodiment of the present application, the moment when the change amplitude of the response value of at least one gas sensor is detected for the first time to exceed the preset amplitude threshold is used as the data sampling start time, and the response value of the gas sensor within the preset time period is collected. Then, the change characteristic data of the collected multiple response values ​​are extracted, and the change characteristic data of the response value of each gas sensor is subjected to gas classification and identification to obtain a gas identification result that can be used to detect whether the battery has leaked. The whole process is different from the traditional practice of obtaining the response value of the complete gas-sensitive response process of the gas sensor in real time, which consumes a lot of testing time and energy. Only the response value of the effective response period is extracted, which can significantly shorten the identification time. Moreover, compared with the method of directly identifying the response value, by identifying the change characteristic data of the response value, the influence of the baseline drift of the gas sensor on the identification result can be reduced, making the identification result more accurate. In summary, the adoption of the above scheme can greatly improve the efficiency of battery leakage identification.

[0010] In some embodiments, based on the response value data set, extracting the change characteristic data of the response value of each gas sensor includes:

[0011] Generate a response value variation curve of each gas sensor based on the response value corresponding to each time point in the response value data set;

[0012] According to the change rate of the response value at different time points, the slope characteristics of each response value change curve are determined, and the change characteristic data includes the slope characteristics.

[0013] In the technical solution of the embodiment of the present application, the slope value of the response value change curve represents the intensity of the reaction between the gas sensor and the gas, and characterizes more features than the response value. By extracting the slope characteristics of the response value change curve of each gas sensor during the effective response period, the change amplitude of the response value can be accurately judged. Compared with directly classifying and identifying the response value, the recognition time can be significantly shortened, the data processing process is simplified, and the processing speed and efficiency are improved. In addition, the slope can be used to directly and effectively divide the classification boundaries of gas categories, reduce the influence of the baseline offset of the gas sensor on the accuracy of the gas identification results, and improve the gas identification results.

[0014] In some embodiments, generating a response value variation curve of each gas sensor based on the response value corresponding to each time point in the response value data set includes:

[0015] Extract the resistance value corresponding to each time point in the response value data set;

[0016] According to the resistance value corresponding to each time point, a resistance value change curve of each gas sensor is generated, and the response value change curve includes the resistance value change curve.

[0017] In the technical solution of the embodiment of the present application, by extracting the slope characteristics of the resistance change curve and classifying and identifying the resistance slope characteristics, the gas category can be identified quickly and accurately, and the battery leakage can be monitored in real time.

[0018] In some embodiments, the change characteristic data of the response value of each gas sensor is subjected to gas classification and identification, and the gas identification result obtained includes:

[0019] Taking the change characteristic data of the response value of each gas sensor as input, calling the trained gas classification model, performing gas classification and identification on the change characteristic data of the response value of each gas sensor, and obtaining a gas identification result;

[0020] Among them, the gas classification model is trained based on the historical response value change characteristic data of multiple gas sensors carrying gas category labels.

[0021] In the technical solution of the embodiment of the present application, by calling the trained gas classification model to classify and identify the change characteristic data of the response value, the gas category can be identified efficiently and accurately, and accurate gas identification results and battery leakage monitoring results can be obtained.

[0022] In some embodiments, the method further comprises:

[0023] Fault identification is performed on the change characteristic data of the response value of each gas sensor to obtain a fault identification result.

[0024] In the technical solution of the embodiment of the present application, by performing fault identification on the change characteristic data of the response value of each gas sensor, a fault identification result is obtained, which can effectively predict whether a battery leakage is found or the change in the gas environment is caused by other uncontrollable factors, so that testers can find the problem in time and take corresponding measures in a targeted manner.

[0025] In some embodiments, fault identification is performed on the change characteristic data of the response value of each gas sensor, and the fault identification result obtained includes:

[0026] Taking the change characteristic data of the response value of each gas sensor as input, the trained fault identification model is called to perform fault identification on the change characteristic data of the response value of each gas sensor to obtain a fault identification result. The fault identification model is trained based on the historical response value change characteristic data of multiple gas sensors carrying fault labels.

[0027] In the technical solution of the embodiment of the present application, by calling the fault identification model, the change characteristic data of the response value of each gas sensor is used for fault identification, so that the fault identification result can be obtained quickly and accurately.

[0028] In some embodiments, the gas classification model is obtained by training a first random forest model using historical response value change characteristic data of multiple gas sensors carrying gas category labels, and the fault identification model is obtained by training a second random forest model using historical response value change characteristic data of multiple gas sensors carrying fault labels.

[0029] In the technical solution of the embodiment of the present application, the random forest model is selected as the classification model for gas identification and fault identification. Compared with commonly used classification models such as neural networks and support vector machines, the operation speed is faster, it is not easy to overfit, and it has good anti-noise ability, so that the efficiency and accuracy of gas identification and fault identification are higher.

[0030] In some embodiments, before performing gas classification and identification on the change characteristic data of the response value of each gas sensor, the method further includes:

[0031] Performing data preprocessing on the change characteristic data of the response value of each gas sensor, wherein the data preprocessing includes at least one of data cleaning, outlier removal and normalization processing;

[0032] Gas classification and identification based on the change characteristic data of the response value of each gas sensor includes:

[0033] The change characteristic data of the response values ​​of each gas sensor after preprocessing are classified and identified to obtain a gas identification result.

[0034] In the technical solution of the embodiment of the present application, by performing data preprocessing on the response value, the standardization and quality of the data can be improved, the adverse effects of noise data and worthless data on the gas identification results can be reduced, and the accuracy of the identification results can be improved.

[0035] In a second aspect, the present application also provides a battery leakage identification device. The device comprises:

[0036] A data acquisition module, used to acquire a data set of response values ​​output by multiple gas sensors to different gases to be tested during an effective response period, wherein the effective response period is a preset time period after a target sampling time, and the target sampling time is a time when the change amplitude of the response value of at least one gas sensor is first detected to exceed a preset amplitude threshold;

[0037] A feature extraction module, used to extract the change feature data of the response value of each gas sensor based on the response value data set;

[0038] The classification and identification module is used to classify and identify gases based on the change characteristic data of the response values ​​of each gas sensor to obtain a gas identification result, which is used to determine whether the battery is leaking.

[0039] In a third aspect, the present application further provides a computer device, wherein the computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps in the above-mentioned battery leakage identification method embodiments are implemented.

[0040] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps in the above-mentioned battery leakage identification method embodiments are implemented.

[0041] In a fifth aspect, the present application further provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned battery leakage identification method embodiments are implemented.

[0042] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 This is a diagram of the application environment of the battery leakage identification method in some embodiments of the present application;

[0044] Figure 2 A schematic diagram of a flow chart of a battery leakage identification method in some embodiments of the present application;

[0045] Figure 3 A schematic diagram of a change process of the resistance output by the gas sensor when a battery leaks in some embodiments of the present application;

[0046] Figure 4 A flowchart of a battery leakage identification method in some other embodiments of the present application;

[0047] Figure 5 A flowchart of a battery leakage identification method in some other embodiments of the present application;

[0048] Figure 6 A schematic diagram of a change process of the resistance output by a gas sensor in an effective response phase in some embodiments of the present application;

[0049] Figure 7 A detailed flowchart of a battery leakage identification method in some embodiments of the present application;

[0050] Figure 8 A detailed flowchart of a battery leakage identification method in some other embodiments of the present application;

[0051] Fig. 9 A detailed flowchart of a battery leakage identification method in some embodiments of the present application;

[0052] Fig.10 A schematic diagram of a change process of a resistance response value of a gas sensor during monitoring in some embodiments of the present application;

[0053] Fig.11 A schematic diagram of a change process of the resistance output by a gas sensor array in an effective response phase in some embodiments of the present application;

[0054] Fig.12 A schematic diagram of the slope characteristic change of the resistance change curve corresponding to the gas sensor in some embodiments of the present application;

[0055] Fig.13 This is a structural block diagram of a battery leakage identification device in some embodiments of the present application;

[0056] Fig.14 A structural block diagram of a battery leakage identification device in some other embodiments of the present application;

[0057] Fig.15 This is a diagram of the internal structure of a computer device in some embodiments of the present application. DETAILED DESCRIPTION

[0058] The following embodiments of the technical solution of the present application are described in detail in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present application, and are therefore only used as examples, and cannot be used to limit the scope of protection of the present application.

[0059] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned figure descriptions and any variations thereof are intended to cover non-exclusive inclusions.

[0060] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0061] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0062] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0063] In the description of the embodiments of the present application, the term "multiple" refers to more than two (including two). Similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).

[0064] Battery leakage refers to the phenomenon that electrolyte, gas or other substances overflow from the inside of the battery. Storing batteries in a high temperature, humid or damp environment will accelerate the corrosion and damage of the battery. Overcharging and discharging, high temperature, mechanical damage to the battery shell, or improper use such as reversed positive and negative poles may increase the possibility of battery leakage. In order to detect battery leakage in time and take measures to reduce the occurrence of safety accidents within an effective time, it is very necessary to monitor the battery for leakage during battery use.

[0065] At present, the commonly used battery leakage identification and detection schemes mainly include the use of sensors or corresponding detection devices. For example, different gas sensors react differently to different gases released by battery leakage, and the gas type is identified by comparing and analyzing the characteristic data in these different reaction processes, and the battery is further judged whether it has leaked by the identified gas type. Alternatively, a bionic human olfactory system is used to quickly provide the overall information of the sample under test with a specific sensor and pattern recognition system, and then the reaction is classified by advanced signal processing technology. However, the first method mentioned above usually requires extracting the response value of the sensor output during the complete gas-sensitive response process. This process consumes a lot of testing time, and in actual applications, as the sampling time increases, the derivative characteristics of the sensor response curves to different gases are basically the same. These redundant information will affect the accuracy of gas category identification. The second method, although it can achieve higher accuracy, these methods often rely on large-scale detection equipment, which is not only costly, time-consuming, and complex to operate, but also makes the identification efficiency low, and it is difficult to achieve real-time application on site. It can be seen that the current battery leakage identification scheme has the problem of low identification efficiency.

[0066] In order to alleviate the problem of low efficiency in battery leakage identification, research has found that the sensor's gas-sensitive response to gas mainly occurs when it begins to contact the test gas. As the sampling time continues to increase, the oxygen ions adsorbed on the material surface are gradually exhausted, and the gas-sensitive response process tends to be flat. The derivative characteristics of the sensor's response curves to different gases are also basically the same, which means that the response values ​​collected during the flat stage of the gas-sensitive response process will cause redundancy and affect the accuracy of identification.

[0067] Based on the above considerations, in order to solve the problem of low efficiency in battery leakage identification, a battery leakage identification scheme is proposed, which can be to fix multiple gas sensors on the battery pack, and the sensors monitor whether the battery is leaking in real time. When the response value of at least one gas sensor is detected to change dramatically for the first time, that is, the change amplitude exceeds the preset amplitude threshold, sampling begins, and response value data sets of multiple gas sensors to different test gases during the effective response period (a preset time period after the target sampling time) are collected. Then, based on the response value data set, the change characteristic data of the response value of each gas sensor is extracted. Finally, the change characteristic data of the response value of each gas sensor is subjected to gas classification and identification to obtain a gas identification result, and the gas identification result can be used to directly determine whether the battery is leaking.

[0068] The above process is different from the traditional method of obtaining the response value of the complete gas-sensitive response process of the gas sensor in real time, which consumes a lot of testing time and energy. Only the response value of the effective response period is extracted, which can significantly shorten the recognition time. In addition, compared with the method of directly identifying the response value, by identifying the changing characteristic data of the response value, the influence of the baseline drift of the gas sensor on the recognition result can be reduced, making the recognition result more accurate. In summary, the above scheme can greatly improve the efficiency of battery leakage recognition.

[0069] The battery leakage identification method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the battery 100 to be tested and multiple gas sensors 102 (such as a gas sensor array 102) are placed in the same space, and the gas sensor 102 communicates with the control end 104 through a network. Specifically, the gas sensor 102 can monitor whether the battery is leaking in real time, and send the measured response value of the gas-sensitive reaction to the control end 104 in real time. When the control end 104 detects for the first time that the change amplitude of the response value of at least one gas sensor 102 exceeds the preset amplitude threshold, sampling begins, and the response value data sets of multiple gas sensors 102 output to different gases to be tested during the effective response period (the preset time period after the target sampling time) are collected. Then, based on the response value data set, the change characteristic data of the response value of each gas sensor 102 is extracted. Finally, the change characteristic data of the response value of each gas sensor 102 is subjected to gas classification and identification processing to obtain a gas identification result, and the gas identification result can be used to directly determine whether the battery is leaking.

[0070] The control terminal 104 may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, IoT devices and portable wearable devices. IoT devices may be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. Portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The control terminal 104 may also be a server or a cloud server. The server or cloud server may be implemented as an independent server or a server cluster consisting of multiple servers.

[0071] In some embodiments, Figure 2 As shown, a battery leakage identification method is provided, and the method is applied to Figure 1 The control terminal 104 in the example is used for explanation. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0072] S200, obtaining a data set of response values ​​output by multiple gas sensors to different gases to be tested during an effective response period, where the effective response period is a preset time period after a target sampling time, and the target sampling time is the time when the change amplitude of the response value of at least one gas sensor is first detected to exceed a preset amplitude threshold.

[0073] The response values ​​output by the gas sensor include data such as current, resistance, load, and impedance. In this embodiment, in order to enhance the accuracy and reliability of gas detection, the multiple gas sensors may be presented in the form of a gas sensor array, and the gas sensor array includes gas sensors with different selective sensitivities to different gases. Taking the battery to be tested as a lithium battery pack as an example, the multiple sensors in the gas sensor array may include multiple gas sensors with different selective sensitivities to specific gases represented by DMC (DimethylCarbonate) and DEC (Diethyl Carbonate), CO, and CO2 in the electrolyte of the lithium battery pack.

[0074] In actual applications, the gas sensor and the battery to be tested are placed in the same space, and the gas sensor array monitors in real time whether the lithium battery to be tested is leaking. When the battery to be tested leaks, it will release specific gases, mainly carbonate gases. The released gas interacts with the gas-sensitive material, causing the impedance of the gas-sensitive material layer to change dramatically in a short period of time. The impedance and load state of the gas sensor will also change accordingly. At this time, the gas sensor will output the corresponding response value such as resistance value in real time, and send the response value to the control end in real time. When the control end detects for the first time that the response value of at least one gas sensor has changed dramatically, that is, the amplitude of the change exceeds the preset amplitude of the change, it is determined that the resistance or impedance of the gas-sensitive material layer of the gas sensor has changed dramatically, and a battery leakage may have occurred. At this time, sampling begins, which can be to collect the response values ​​output by each gas sensor within a period of time after the first detection of the dramatic change in the response value. Specifically, the change in resistance output by the gas sensor when the battery leaks can be seen in Figure 3 In this embodiment, the change amplitude of the response value may be the change rate of the response value, and when the change rate of the response value is greater than a preset change rate threshold value, such as 80%, it is determined that the resistance or impedance of the gas-sensitive material layer of the gas sensor has changed dramatically. It is understandable that the change rate threshold value is determined according to the actual situation.

[0075] After a large number of experiments, it was found that if the battery leaks, the effective stage of the gas sensor's gas-sensitive response to the gas usually occurs within 100 seconds after the gas begins to contact. Therefore, in this embodiment, the response value output by each gas sensor within 100 seconds after the target sampling time can be collected. In other embodiments, the selection of the optimal sampling time can also take the response value of the first 100 seconds in the experimental data, and take the response value of every 10 seconds as a group of experimental data. The response value data of different sampling times are respectively classified into gas categories to obtain the corresponding recognition accuracy, and finally the sampling data with the highest recognition accuracy is taken as the optimal sampling time. It can be understood that in other embodiments, in order to reduce the impact of other uncontrollable factors on the gas environment, it can also include but not be limited to starting sampling and collecting the response values ​​output by each gas sensor when the response values ​​of at least two gas sensors change drastically, or when the response values ​​output by all gas sensors change drastically.

[0076] S400, extracting change characteristic data of the response value of each gas sensor based on the response value data set.

[0077] The response value change characteristic data refers to the data that characterizes the change of the response value, including but not limited to the change amplitude data or the change rate. After the response value data set of each gas sensor is obtained, the response values ​​at different time points in the response value data set can be analyzed to extract the change characteristic data of the response value of each gas sensor. Specifically, the response value data set can be preprocessed, including denoising and deleting outliers, and then the change characteristics of the preprocessed response value data set are extracted by a feature extraction algorithm. For example, a time series method is used to regard the response values ​​at different time points as a time series, and the time series is analyzed by a statistical method to extract the change characteristics of the response value.

[0078] S600, performing gas classification and identification on the change characteristic data of the response value of each gas sensor to obtain a gas identification result, and the gas identification result is used to determine whether the battery is leaking.

[0079] The gas identification result includes data such as the identified gas category, gas concentration, and gas quantity. After obtaining the change characteristic data of the response value of each gas sensor, the change characteristic data can be classified and identified to identify the gas existing in the environment of the current battery to be tested, so as to judge whether the battery is really leaking according to the category and / or quantity of the identified specific gas, or whether the gas environment has changed due to uncontrollable factors such as gas sensor aging, severe vibration, or battery thermal runaway, causing a drastic change in the response value output by the gas sensor. In specific implementation, before classifying and identifying the change characteristic data of the response value, the change characteristic data of the response value of each gas sensor can be preprocessed. The data preprocessing includes at least one of data cleaning, outlier removal, and normalization. Then, the change characteristic data after data preprocessing is classified and identified to obtain the gas identification result.

[0080] In the technical solution of the embodiment of the present application, the moment when the change amplitude of the response value of at least one gas sensor is detected for the first time to exceed the preset amplitude threshold is used as the data sampling start time, and the response value of the gas sensor within the preset time period is collected. Then, the change characteristic data of the collected multiple response values ​​are extracted, and the change characteristic data of the response value of each gas sensor is subjected to gas classification and identification to obtain a gas identification result that can be used to detect whether the battery has leaked. The whole process is different from the traditional practice of obtaining the response value of the complete gas-sensitive response process of the gas sensor in real time, which consumes a lot of testing time and energy. Only the response value of the effective response period is extracted, which can significantly shorten the identification time. Moreover, compared with the method of directly identifying the response value, by identifying the change characteristic data of the response value, the influence of the baseline drift of the gas sensor on the identification result can be reduced, making the identification result more accurate. In summary, the adoption of the above scheme can greatly improve the efficiency of battery leakage identification.

[0081] like Figure 4 As shown, in some embodiments, S400 includes: S420, generating a response value change curve of each gas sensor based on the response value corresponding to each time point in the response value data set, and determining the slope characteristics of each response value change curve according to the change rate of the response value at different time points, and the change characteristic data includes the slope characteristics.

[0082] S600 includes: S602, performing gas classification and identification on the slope characteristics of the response value change curve of each gas sensor to obtain a gas identification result, and the gas identification result is used to detect whether the battery is leaking.

[0083] The rate of change of the response value refers to the amount of change of the response value per unit time. In this embodiment, the slope feature is taken as an example of the change characteristic data. Specifically, the corresponding curve drawing tool can be called according to the response value corresponding to each time point in the acquired response value data set to generate the response value change curve of each gas sensor, and then the rate of change of the response value corresponding to each time point is calculated, and then the slope characteristics of each response value change curve are obtained according to the rate of change of the response value at different time points. Specifically, the slope characteristics of the response value change curve corresponding to each gas sensor can be determined based on the following method:

[0084] K T (t) = (X T+t -X T ) / t

[0085] Where t represents the time interval for calculating the derivative, X T+t and X T Represent the response values ​​at time T+t and T respectively.

[0086] Then, the slope characteristics of the response value change curves of each gas sensor are used as the object of classification and identification processing to obtain the gas category result. It can be understood that the response value change curve includes but is not limited to the resistance change curve, the current change curve and the impedance change curve.

[0087] In the technical solution of the embodiment of the present application, the slope value of the response value change curve represents the intensity of the reaction between the gas sensor and the gas, and characterizes more features than the response value. By extracting the slope characteristics of the response value change curve of each gas sensor during the effective response period, the change amplitude of the response value can be accurately judged. Compared with directly classifying and identifying the response value, the recognition time can be significantly shortened, the data processing process is simplified, and the processing speed and efficiency are improved. In addition, the slope can be used to directly and effectively divide the classification boundaries of gas categories, reduce the influence of the baseline offset of the gas sensor on the accuracy of the gas identification results, and improve the gas identification results.

[0088] like Figure 5 As shown, in some embodiments, S420 includes: S422, extracting the resistance value corresponding to each time point in the response value data set, generating a resistance value change curve of each gas sensor according to the resistance value corresponding to each time point, and determining the slope characteristics of each resistance value change curve according to the rate of change of the resistance value at different time points.

[0089] Since the response value output by the gas sensor includes multi-dimensional monitoring parameters, in actual applications, when the gas-sensitive material of the gas sensor reacts chemically with the gas, the resistance of the gas-sensitive material layer changes more obviously in a short period of time. Therefore, in this embodiment, the resistance in the response value can be extracted to identify the gas category. Specifically, the multi-dimensional response value output by each gas sensor during the effective response period can be used to extract the resistance value corresponding to each time point, and a resistance change curve can be generated based on the resistance value corresponding to each time point. Then, the resistance slope feature of the resistance change curve is extracted, and the resistance slope feature is classified and identified to obtain the gas identification result. Specifically, the slope feature of the response value change curve corresponding to each gas sensor can be determined based on the following method:

[0090] K T (t)=(R T+t -R T ) / t

[0091] Where t represents the time interval for calculating the derivative, R T+t and R T They represent the resistance values ​​at time T+t and time T respectively.

[0092] In other embodiments, the collection of the resistance value during the effective response period may also be based on a preset resistance value range, and the resistance value within the preset resistance value range is collected as the effective resistance value. Figure 6 , set the maximum resistance value of the gas sensor output in dry air to R max The minimum resistance value output by the gas sensor in the gas to be measured is R min After many experiments, it was found that the resistance value range of the gas sensor output in the effective response stage is 0.95*R max ~1.05*R min Therefore, the sample can be collected at 0.95*R max ~1.05*R min The resistance value between 0.95 and 0.95 is taken as the effective resistance value for gas classification and identification. It is understood that the resistance value range includes but is not limited to the above 0.95*R max ~1.05*R min , it can also be other ranges.

[0093] In the technical solution of the embodiment of the present application, by extracting the slope characteristics of the resistance change curve and classifying and identifying the resistance slope characteristics, the gas category can be identified quickly and accurately, and the battery leakage can be monitored in real time.

[0094] like Figure 7As shown, in some embodiments, S600 includes: S620, taking the change characteristic data of the response value of each gas sensor as input, calling the trained gas classification model, performing gas classification and identification on the change characteristic data of the response value of each gas sensor, and obtaining a gas identification result, and the gas identification result is used to determine whether the battery is leaking.

[0095] Among them, the gas classification model is trained based on the historical response value change characteristic data of multiple gas sensors carrying gas category labels.

[0096] The gas category label is the label data for marking the gas category. In the actual training process of the gas classification model, different gases of known gas categories can be input into the gas sensor array. When the response value of the gas sensor, such as the resistance value, changes dramatically for the first time, the resistance value of the subsequent 100s (seconds) is collected to obtain the resistance value change of the gas sensor array to different gases. For each gas sensor, the slope of the time-resistance curve is extracted as a feature based on the resistance value, and then the slope features of the resistance change curves corresponding to different gas sensors are grouped together and labeled with the corresponding gas categories to obtain a certain number of sample sets. For example, at time T, the slope feature of the resistance change curve corresponding to gas sensor A is a, the slope feature of the resistance change curve corresponding to gas sensor B is b, the slope feature of the resistance change curve corresponding to gas sensor C is c, and the slope feature of the resistance change curve corresponding to gas sensor D is d, then a, b, c, d are a group of data, and the corresponding gas categories are labeled. The sample set is then divided into a training set, a test set, and a validation set, and the training set is then input into the pre-built classification model. After self-iterative learning through the machine learning algorithm, when the gas classification model reaches the preset accuracy or converges, the training of the gas classification model is completed, and the trained gas classification model is tested and verified through the test set and validation set to obtain the final gas classification model that can achieve accurate gas category recognition. Classification models may include but are not limited to random classification models, support vector machines, and decision tree models.

[0097] In a specific implementation, after extracting the change characteristics of the response value data set, the change characteristic data of the response value of each gas sensor is used as input, and the trained gas classification model is called to perform gas classification and identification on the change characteristic data of the response value of each gas sensor to obtain a gas identification result, which includes the gas type, gas concentration and / or the number of gas types, etc.

[0098] In the technical solution of the embodiment of the present application, by calling the trained gas classification model to classify and identify the change characteristic data of the response value, the gas category can be identified efficiently and accurately, and accurate gas identification results and battery leakage monitoring results can be obtained.

[0099] like Figure 7 As shown, in some embodiments, after S600, the process further includes: S800, performing fault identification on the change characteristic data of the response value of each gas sensor to obtain a fault identification result.

[0100] The fault identification result refers to the fault cause and fault condition data that cause the response value of the gas sensor to change dramatically. Specifically, the fault identification result includes the fault identification result type and the degree of fault. Common fault types include but are not limited to battery leakage, battery thermal runaway, battery mechanical damage, and severe sensor shaking. After obtaining the change characteristic data of the response value of each gas sensor, the change characteristic data of the response value of each gas sensor can be combined with the preset correspondence between the response value change characteristic data and the fault type to perform fault identification on the fault identification result. Furthermore, the change characteristic data of the response value of each gas sensor can also be used to predict the working condition to obtain the battery working condition prediction result.

[0101] In the technical solution of the embodiment of the present application, by performing fault identification on the change characteristic data of the response value of each gas sensor, a fault identification result is obtained, which can effectively predict whether a battery leakage is found or the change in the gas environment is caused by other uncontrollable factors, so that testers can find the problem in time and take corresponding measures in a targeted manner.

[0102] In other embodiments, the gas identification results may be analyzed and processed to obtain fault identification results. Specifically, after obtaining identification results such as gas types and / or the number of gas types, the gas identification results may be analyzed and processed to further obtain fault identification results. Specifically, common fault conditions may be divided into different fault types in advance, and for each fault type, the gas category that exists when the type of fault occurs is set. For each gas category, the possible fault type is set, and gas leakage fault diagnosis data including the association between the fault type and the gas category is established. In specific implementation, after obtaining the gas identification results, data matching may be performed on the gas leakage fault diagnosis data according to the type of gas, and the corresponding fault type is matched, and then the corresponding fault degree is determined according to the concentration of the gas. For example, if a high concentration of carbonate gas is detected for a lithium battery pack, it is determined that a battery leakage fault has occurred and the degree of the fault is serious.

[0103] In other embodiments, the fault degree corresponding to each fault type may be determined in advance according to the fault type, and a corresponding relationship between the fault degree and the fault type may be established. For example, if the fault type is battery leakage, the fault degree is severe, if the fault type is severe sensor shaking, the fault degree is slight, and if the fault type is mechanical damage to the battery, the fault degree is general. Then, after determining the fault type, the fault degree is directly determined according to the corresponding relationship between the fault degree and the fault type.

[0104] It is understandable that, in other embodiments, the fault identification result may also be obtained by processing the gas identification result using a classification algorithm or a machine learning algorithm.

[0105] In the technical solution of the embodiment of the present application, fault identification is performed through gas identification results to obtain fault identification results. The fault identification results can be obtained simply and quickly, and it can be effectively predicted whether battery leakage occurs or the gas environment changes due to other uncontrollable factors, so that testers can find problems in time and take corresponding measures in a targeted manner.

[0106] like Figure 8 As shown, in some embodiments, S800 includes: S820, taking the change characteristic data of the response value of each gas sensor as input, calling the trained fault identification model to perform fault identification on the change characteristic data of the response value of each gas sensor, and obtaining a fault identification result, the fault identification model is trained based on the historical response value change characteristic data of multiple gas sensors carrying fault labels.

[0107] In this embodiment, the change characteristic data of the response value of each gas sensor is used for fault identification, and the fault identification result can be: the test data of the gas sensor in the gas-sensitive response process for different fault types and fault degrees in the historical experiment is collected, including gas category, fault type and corresponding gas-sensitive performance data, wherein the gas-sensitive performance data includes a certain performance of the gas-sensitive material, such as the degree of change of the resistance when reacting with the gas (i.e., the resistance change data over time), the response and recovery time, the sensitivity and the stability time length, etc. Then, the gas-sensitive performance under different working conditions is integrated to obtain the gas-sensitive database of the fault-marking gas. Then, for the gas-sensitive database of the fault-marking gas, the gas-sensitive performance data carrying the fault type and fault degree labels are constructed, and then the gas-sensitive performance data carrying the fault category and fault degree labels are input as a training set into the pre-constructed classification model for training. The classification model distinguishes the gas-sensitive performance data marked with different fault types and fault degrees by self-iteration. When the classification model reaches the preset accuracy or convergence, the model training is completed, and the fault identification model is obtained.

[0108] In a specific implementation, the change characteristic data of the response value of each gas sensor may be converted into data that can be identified by the fault identification model, and the converted change characteristic data of the response value of each gas sensor is used as input to call the trained fault identification model to perform fault identification on the change characteristic data of the response value of each gas sensor, and obtain the corresponding fault identification result, including data such as the fault type and fault degree. Among them, the fault identification model may include but is not limited to a random forest model, a support vector machine, and a decision tree model.

[0109] In the technical solution of the embodiment of the present application, by calling the fault identification model, the change characteristic data of the response value of each gas sensor is used for fault identification, so that the fault identification result can be obtained quickly and accurately.

[0110] In some embodiments, the gas classification model is obtained by training a first random forest model using historical response value change characteristic data of multiple gas sensors carrying gas category labels, and the fault identification model is obtained by training a second random forest model using historical response value change characteristic data of multiple gas sensors carrying fault labels.

[0111] In practical applications, through a large number of model experiments, it is found that the random forest model performs better than other classification models. Therefore, in this embodiment, the random forest model can be selected as the basic classification model for gas identification and fault identification. The specific process of building a gas classification model can be: integrating the historical response value change feature data carrying the gas category label to determine the feature set for training. Select samples from the feature set in a random manner to create multiple derivative sample sets. The derivative sample set is obtained by random sampling based on the original data set and has the same size as the original data set. For each derivative sample set, a decision tree is constructed, and the data in the derivative sample set is used as the training sample. The selected slope feature is used for training. According to the importance of each feature to the target variable, the best feature is selected for node splitting, and then the decision trees in the sample set are combined to form a gas classification model.

[0112] In other embodiments, the process of building a random forest can also be: based on a data set of historical response value change characteristic data carrying gas category labels, calling the random forest model (Random ForestClassifer library class) in Scikit-learn, setting the initial n_estimators, max_depth, and other parameters, and dividing the corresponding gas classification boundaries, importing the training set data into the random forest model for model training to distinguish the response value change characteristic data carrying different gas category labels, and using the cross-validation method to adjust the parameters, including n_estimators, max_depth, min_samples_leaf, min_samples_leaf and other parameters, so as to achieve an increase in accuracy. The model training can be completed when the random forest model reaches a preset accuracy or convergence. Further, the model can be trained multiple times to observe the model stability.

[0113] Specifically, the model adjustment can be done by using Bayesian optimization parameter adjustment, estimating the posterior distribution of the objective function based on the data set using Bayesian theorem, and then selecting the next sampled hyperparameter combination based on the distribution. Bayesian optimization parameter adjustment uses a Gaussian process, taking into account previous parameter information, and continuously updating prior information. Compared with the grid search method, it has fewer iterations and is faster. After the random forest training is completed, the test set data can be preprocessed by data cleaning, normalization, etc., and then the test set data can be imported for model performance evaluation, that is, the test set data is predicted by the trained random forest model, and the prediction results are compared with the actual labels. The prediction accuracy is calculated until the prediction accuracy of the model meets the preset requirements to obtain the final gas classification model.

[0114] In a specific implementation, the change characteristic data of the response value of each gas sensor may be used as input, and a trained random forest model may be called to perform gas classification and identification on the change characteristic data of the response value of each gas sensor to obtain a gas identification result.

[0115] The construction process of the fault identification model can be based on the historical response value change characteristic data carrying the fault label (including the fault type and fault degree), and adopts the construction method of the gas classification model mentioned above to construct the fault identification model. The specific process is not repeated here. In specific implementation, the change characteristic data of the response value of each gas sensor can be used as input to call the trained fault identification model for fault identification, including identification of the fault type and fault degree, to obtain the corresponding fault identification result.

[0116] In the technical solution of the embodiment of the present application, the random forest model is selected as the classification model for gas identification and fault identification. Compared with commonly used classification models such as neural networks and support vector machines, the operation speed is faster, it is not easy to overfit, and it has good anti-noise ability. In this way, the efficiency and accuracy of gas identification and fault identification are higher.

[0117] like Figure 8 As shown, in some embodiments, before S600, it also includes: S500, performing data preprocessing on the change characteristic data of the response value of each gas sensor, and the data preprocessing includes at least one of data cleaning, outlier elimination and normalization processing.

[0118] S600 includes: S640, classifying and identifying the change characteristic data of the response values ​​of each gas sensor after preprocessing to obtain a gas identification result, and the gas identification result is used to determine whether the battery is leaking.

[0119] In practical applications, since the multiple response values ​​obtained may contain outliers and some useless information, the response values ​​may be preprocessed before extracting the change characteristic data of the response values. Specifically, data preprocessing may include but is not limited to data cleaning, data normalization and feature selection. Then, the change characteristic data of the battery monitoring parameters after preprocessing are extracted. Among them, data cleaning includes checking whether there are null values ​​and outliers in the data. If there are null values ​​or outliers, the null values ​​and outliers are deleted, or the outliers are replaced with the average value. Data normalization includes determining the actual value range of the battery monitoring data, and then scaling the data to normalized data between 0-1.

[0120] In the technical solution of the embodiment of the present application, by performing data preprocessing on the response value, the standardization and quality of the data can be improved, the adverse effects of noise data and worthless data on the gas identification results can be reduced, and the accuracy of the identification results can be improved.

[0121] In order to make a clearer description of the battery leakage monitoring method provided by the present application, the following is combined with the attached Fig. 9 and one A specific embodiment is described, and the specific embodiment includes the following contents:

[0122] S200, obtaining a data set of response values ​​output by multiple gas sensors to different gases to be tested during an effective response period, where the effective response period is a preset time period after a target sampling time, and the target sampling time is the time when the change amplitude of the response value of at least one gas sensor is first detected to exceed a preset amplitude threshold.

[0123] In practical applications, a gas sensor array consisting of four gas sensors can be placed in the same space as the battery to be tested. The gas sensor array monitors in real time whether the lithium battery to be tested is leaking. When the battery to be tested leaks, it will release specific gases, mainly carbonate gases. The released gas interacts with the gas-sensitive material, causing the impedance of the gas-sensitive material layer to change dramatically in a short period of time. The impedance and load state of the gas sensor will also change accordingly. At this time, the gas sensor will output a response value in real time and send the response value to the control end in real time. Specifically, as an example, the change in the response value of the gas sensor can be seen in Fig.10 . When the control end detects for the first time that the response value of at least one gas sensor has changed dramatically, that is, when the change amplitude exceeds the preset change amplitude, it is determined that the resistance or impedance of the gas-sensitive material layer of the gas sensor has changed dramatically, and battery leakage may have occurred. At this time, sampling begins, which can be to collect the response values ​​output by each gas sensor within 100 seconds after the first detection of the dramatic change in the response value. As an example, the change in the resistance value output by each sensor during the effective response stage can be seen in Fig.11 .

[0124] S422, extracting the resistance value corresponding to each time point in the response value data set, generating a resistance value change curve of each gas sensor according to the resistance value corresponding to each time point, and determining the slope characteristics of each resistance value change curve according to the change rate of the resistance value at different time points.

[0125] Specifically, for each gas sensor, multiple response values ​​output during the effective response period are extracted, and the resistance value corresponding to each time point is extracted. According to the resistance value corresponding to each time point, a resistance change curve is generated, and then the resistance slope feature of the resistance change curve is extracted, and the resistance slope feature is classified and identified to obtain the gas identification result. Specifically, the slope feature of the response value change curve corresponding to each gas sensor can be determined based on the following method:

[0126] K T (t)=(R T+t -R T ) / t

[0127] Where t represents the time interval for calculating the derivative, R T+t and R T Represent the resistance values ​​at time T+t and time T respectively. As an example, the slope characteristic changes of the response value change curves of each gas sensor during the effective response stage can be seen in Fig.12 .

[0128] S520, performing data preprocessing on the slope characteristics of the resistance value change curve of each gas sensor, where the data preprocessing includes at least one of data cleaning, outlier elimination, and normalization processing.

[0129] S622, taking the slope characteristics of the resistance value change curve of each gas sensor after preprocessing as input, calling the trained gas classification model, performing gas classification and identification on the slope characteristics of the resistance value change curve of each gas sensor, and obtaining a gas identification result.

[0130] The gas identification results include the identified gas category, gas concentration, gas quantity and other data. Among them, the gas classification model is obtained by training the random forest model using the change feature data of the historical response value carrying the gas category label.

[0131] S822, taking the slope characteristics of the resistance value change curve of each gas sensor as input, calling the trained fault identification model to perform fault identification on the slope characteristics of the resistance value change curve of each gas sensor, and obtaining a fault identification result.

[0132] The fault identification results include the fault identification result type and fault degree, etc. The fault identification model is obtained by training the random forest model using the historical response value change feature data of multiple gas sensors carrying fault labels.

[0133] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0134] Based on the same inventive concept, the embodiment of the present application also provides a battery leakage identification device for implementing the battery leakage identification method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more battery leakage identification device embodiments provided below can refer to the limitations of the battery leakage identification method above, and will not be repeated here.

[0135] In some embodiments, Fig.13As shown, a battery leakage identification device 800 is provided, including: a data acquisition module 810, a feature extraction module 820 and a classification identification module 830, wherein:

[0136] The data acquisition module 810 is used to obtain a data set of response values ​​output by multiple gas sensors to different gases to be tested during an effective response period. The effective response period is a preset time period after a target sampling time. The target sampling time is the time when the change amplitude of the response value of at least one gas sensor is first detected to exceed a preset amplitude threshold.

[0137] The feature extraction module 820 is used to extract the change feature data of the response value of each gas sensor based on the response value data set.

[0138] The classification and identification module 830 is used to classify and identify gases based on the change characteristic data of the response values ​​of each gas sensor to obtain a gas identification result. The gas identification result is used to determine whether the battery is leaking.

[0139] In the technical solution of the embodiment of the present application, the moment when the change amplitude of the response value of at least one gas sensor is detected for the first time to exceed the preset amplitude threshold is used as the data sampling start time, and the response value of the gas sensor within the preset time period is collected. Then, the change characteristic data of the collected multiple response values ​​are extracted, and the change characteristic data of the response value of each gas sensor is subjected to gas classification and identification to obtain a gas identification result that can be used to detect whether the battery has leaked. The whole process is different from the traditional practice of obtaining the response value of the complete gas-sensitive response process of the gas sensor in real time, which consumes a lot of testing time and energy. Only the response value of the effective response period is extracted, which can significantly shorten the identification time. Moreover, compared with the method of directly identifying the response value, by identifying the change characteristic data of the response value, the influence of the baseline drift of the gas sensor on the identification result can be reduced, making the identification result more accurate. In summary, the adoption of the above scheme can greatly improve the efficiency of battery leakage identification.

[0140] In some embodiments, the feature extraction module 820 is also used to generate a response value change curve for each gas sensor based on the response value corresponding to each time point in the response value data set, and determine the slope characteristics of each response value change curve according to the change rate of the response value at different time points. The change characteristic data includes the slope characteristics.

[0141] In some embodiments, the feature extraction module 820 is also used to extract the resistance value corresponding to each time point in the response value data set, and generate a resistance value change curve of each gas sensor according to the resistance value corresponding to each time point, and the response value change curve includes a resistance value change curve.

[0142] In some embodiments, the classification and identification module 830 is further used to use the change characteristic data of the response value of each gas sensor as input, call the trained gas classification model, perform gas classification and identification on the change characteristic data of the response value of each gas sensor, and obtain a gas identification result;

[0143] Among them, the gas classification model is trained based on the historical response value change characteristic data of multiple gas sensors carrying gas category labels.

[0144] like Fig.14 As shown, in some embodiments, the device further includes a fault identification module 840 for performing fault identification on the change characteristic data of the response value of each gas sensor to obtain a fault identification result.

[0145] In some embodiments, the fault identification module 840 is also used to perform statistical analysis on the gas identification results to obtain fault identification results; or, taking the change characteristic data of the response value of each gas sensor as input, calling a trained fault identification model to perform fault identification on the change characteristic data of the response value of each gas sensor to obtain a fault identification result, and the fault identification model is trained based on the historical response value change characteristic data of multiple gas sensors carrying fault labels.

[0146] In some embodiments, the gas classification model is obtained by training a first random forest model using historical response value change characteristic data of multiple gas sensors carrying gas category labels, and the fault identification model is obtained by training a second random forest model using historical response value change characteristic data of multiple gas sensors carrying fault labels.

[0147] like Fig.14 As shown, in some embodiments, the device further includes a data preprocessing module 825 for performing data preprocessing on the change characteristic data of the response value of each gas sensor, and the data preprocessing includes at least one of data cleaning, outlier removal and normalization processing;

[0148] The classification and identification module 830 is also used to perform classification and identification processing on the change characteristic data of the response values ​​of each gas sensor after preprocessing to obtain a gas identification result.

[0149] Each module in the battery leakage identification device can be implemented in whole or in part by software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module.

[0150] In some embodiments, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Fig.15As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected via a system bus, and the communication interface, the display unit and the input device are connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a battery leakage identification method is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0151] Those skilled in the art will understand that Fig.15 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0152] In some embodiments, a computer device is provided, including a memory and a processor. The memory stores a computer program. When the processor executes the computer program, the steps in the above-mentioned battery leakage identification method embodiments are implemented.

[0153] In some embodiments, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned battery leakage identification method embodiments are implemented.

[0154] In some embodiments, a computer program product is provided, including a computer program, which implements the steps in the above-mentioned battery leakage identification method embodiments when executed by a processor.

[0155] It should be noted that the data involved in this application (including but not limited to data used for analysis, stored data, displayed data, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0156] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0157] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application, and they should all be included in the scope of the claims and specification of the present application. In particular, as long as there is no structural conflict, the various technical features mentioned in the various embodiments can be combined in any way. The present application is not limited to the specific embodiments disclosed herein, but includes all technical solutions that fall within the scope of the claims.

Claims

1. A battery leakage identification method, characterized in that: The method comprises: Acquire a data set of response values ​​output by multiple gas sensors to different gases to be tested during an effective response period, wherein the effective response period is a preset time period after a target sampling time, and the target sampling time is a time when the change amplitude of the response value of at least one gas sensor is first detected to exceed a preset amplitude threshold; Extracting the change characteristic data of the response value of each gas sensor based on the response value data set; Gas classification and identification are performed on the change characteristic data of the response value of each gas sensor to obtain a gas identification result, and the gas identification result is used to determine whether the battery is leaking.

2. The method according to claim 1, characterized in that Based on the response value data set, extracting the change characteristic data of the response value of each gas sensor includes: Generate a response value variation curve of each gas sensor based on the response value corresponding to each time point in the response value data set; According to the change rates of the response values ​​at different time points, the slope characteristics of each response value change curve are determined, and the change characteristic data includes the slope characteristics.

3. The method according to claim 2, characterized in that The step of generating a response value variation curve of each gas sensor based on the response value corresponding to each time point in the response value data set includes: Extracting the resistance value corresponding to each time point in the response value data set; According to the resistance value corresponding to each time point, a resistance value change curve of each gas sensor is generated, and the response value change curve includes a resistance value change curve.

4. The method according to any one of claims 1 to 3, characterized in that: The gas classification and identification is performed on the change characteristic data of the response value of each gas sensor to obtain the gas identification result, which includes: Taking the change characteristic data of the response value of each gas sensor as input, calling the trained gas classification model, performing gas classification and identification on the change characteristic data of the response value of each gas sensor, and obtaining a gas identification result; The gas classification model is trained based on historical response value change characteristic data of multiple gas sensors carrying gas category labels.

5. The method according to claim 4, characterized in that The method further comprises: Fault identification is performed on the change characteristic data of the response value of each gas sensor to obtain a fault identification result.

6. The method according to claim 5, characterized in that The fault identification of the change characteristic data of the response value of each gas sensor includes: Taking the change characteristic data of the response value of each gas sensor as input, a trained fault identification model is called to perform fault identification on the change characteristic data of the response value of each gas sensor to obtain a fault identification result. The fault identification model is trained based on the historical response value change characteristic data of multiple gas sensors carrying fault labels.

7. The method according to claim 6, characterized in that The gas classification model is obtained by training a first random forest model using historical response value change characteristic data of multiple gas sensors carrying gas category labels, and the fault identification model is obtained by training a second random forest model using historical response value change characteristic data of multiple gas sensors carrying fault labels.

8. The method according to any one of claims 1 to 3, characterized in that: Before the gas classification and identification is performed on the change characteristic data of the response values ​​of each gas sensor to obtain the gas identification result, the method further includes: Performing data preprocessing on the change characteristic data of the response value of each gas sensor, wherein the data preprocessing includes at least one of data cleaning, outlier removal and normalization processing; The gas classification and identification is performed on the change characteristic data of the response value of each gas sensor to obtain the gas identification result, which includes: Gas classification and identification are performed on the change characteristic data of the response values ​​of each gas sensor after preprocessing to obtain a gas identification result.

9. A battery leakage identification device, characterized in that: The device comprises: A data acquisition module, used to acquire a data set of response values ​​output by multiple gas sensors to different gases to be tested during an effective response period, wherein the effective response period is a preset time period after a target sampling time, and the target sampling time is a time when the change amplitude of the response value of at least one gas sensor is first detected to exceed a preset amplitude threshold; A feature extraction module, used to extract the change feature data of the response value of each gas sensor based on the response value data set; The classification and identification module is used to classify and identify gases based on the change characteristic data of the response values ​​of each gas sensor to obtain a gas identification result, which is used to determine whether the battery is leaking.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

12. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images