Battery liquid leakage detection method, electronic device, and storage medium
By standardizing battery data and training a long short-term memory model, a leakage detection model was established, which solved the problem of untimely battery leakage detection in existing technologies and achieved the effect of early leakage detection.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- EVE POWER CO LTD
- Filing Date
- 2025-03-18
- Publication Date
- 2026-06-18
AI Technical Summary
Existing battery leakage detection solutions cannot detect cell leakage in a timely manner during battery pack charging and discharging, resulting in a delayed response from the fire protection system and posing a safety hazard.
A long short-term memory model was used to train the sample battery data. After standardization and division into training and test sets, a leakage detection model was established. The trained model was then used to perform real-time detection on the batteries to be tested.
It improves the timeliness of battery leakage detection, enabling early detection of leaks and reducing safety risks.
Smart Images
Figure CN2025083186_18062026_PF_FP_ABST
Abstract
Description
Battery leakage detection methods, electronic devices, and storage media
[0001] This application claims priority to Chinese Patent Application No. 202411804485X, filed on December 9, 2024, the entire contents of which are incorporated herein by reference. Technical Field
[0002] This application relates to the field of battery testing technology, specifically to a battery leakage detection method, electronic device, and storage medium. Background Technology
[0003] Lithium iron phosphate (LFP) batteries have become the mainstream choice for energy storage systems due to their advantages such as high energy density, mature mass production technology, long cycle life, and deep discharge. Battery energy storage systems are typically housed in a standard 20-foot shipping container. Within this limited space, a large number of batteries are connected in series and parallel to form battery packs. The energy storage capacity of these packs is often in the MWh range, with a conversion power ranging from hundreds of kW to MW. DC-side battery energy storage systems are characterized by high voltage, high operating power, and deep charge / discharge capabilities, all of which place higher demands on the safe operation of the batteries.
[0004] The energy storage container is equipped with a fire suppression system that can monitor and respond quickly to fires inside the container in real time. The system uses multiple sensors for smoke, temperature, and gas detection. When a battery leaks, the reaction of the positive and negative electrolytes generates VOC gases, which can be detected by these sensors. However, the detection of VOC gases indicates that the gas has already diffused into the container, suggesting a severe battery leak. While the fire suppression system primarily ensures fire safety at the energy storage system level, battery leakage is one aspect that needs to be considered. In reality, during the charging and discharging process of the battery pack, leakage from individual cells may not be detected in time by the fire suppression system, demonstrating a problem with the current battery leakage detection methods in terms of timeliness. Technical issues
[0005] The energy storage container is equipped with a fire suppression system that can monitor and respond quickly to fires inside the container in real time. The system uses multiple sensors for smoke, temperature, and gas detection. When a battery leaks, the reaction of the positive and negative electrolytes generates VOC gases, which can be detected by these sensors. However, the detection of VOC gases indicates that the gas has already diffused into the container, suggesting a severe battery leak. While the fire suppression system primarily ensures fire safety at the energy storage system level, battery leakage is one aspect that needs to be considered. In reality, during the charging and discharging process of the battery pack, leakage from individual cells may not be detected in time by the fire suppression system, demonstrating a problem with the current battery leakage detection methods in terms of timeliness. Technical solutions
[0006] This application provides a battery leakage detection method, including:
[0007] Acquire sample battery data when battery leakage occurs;
[0008] The sample battery data is standardized, and the processed sample battery data is divided into a training set and a test set.
[0009] The preset long short-term memory model is trained based on the training set, and the leakage detection model is determined from the trained long short-term memory model based on the test set.
[0010] The battery leakage detection model is used to detect battery leakage in the battery under test.
[0011] Secondly, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program as described in any of the methods above.
[0012] Thirdly, this application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the methods described above. Beneficial effects
[0013] The beneficial effects provided by this application are: by using sample battery data when battery leakage occurs in advance, a pre-set long short-term memory model is trained, and a leakage detection model is determined in the trained long short-term memory model based on the test set. When battery leakage detection is actually performed, the trained leakage detection model can be directly used to detect battery leakage in the battery under test, thereby improving the timeliness of battery leakage detection. Attached Figure Description
[0014] Figure 1 is a schematic flowchart of the battery leakage detection method provided in the embodiment of this application;
[0015] Figure 2 is a schematic diagram of the battery leakage detection system provided in an embodiment of this application;
[0016] Figure 3 is a schematic diagram of the structure of the electronic device provided in an embodiment of this application.
[0017] Implementation methods of this application
[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0019] This application provides a battery leakage detection method, an electronic device, and a storage medium.
[0020] Specifically, the battery leakage detection method can be applied to a terminal, which may include a tablet computer or a personal computer (PC). The terminal can establish a wired or wireless connection with a server, which may include a standalone server, a distributed server, or a server cluster consisting of multiple servers.
[0021] The following sections provide detailed descriptions of each example. It should be noted that the order in which the embodiments are described is not intended to limit the priority of the embodiments.
[0022] A battery leakage detection method includes: acquiring sample battery data when battery leakage occurs; standardizing the sample battery data and dividing the processed sample battery data into a training set and a test set; training a preset long short-term memory model based on the training set, and determining a leakage detection model in the trained long short-term memory model based on the test set; and performing battery leakage detection on the battery to be tested according to the leakage detection model.
[0023] Please refer to Figure 1, which is a schematic flowchart of the battery leakage detection method provided in this embodiment. The specific flow of the battery leakage detection method is as follows:
[0024] 101. Obtain sample battery data when battery leakage occurs.
[0025] For example, data such as battery voltage, temperature, and pressure are collected from the battery energy storage system during normal operation and when leakage occurs. To ensure the collected data accurately reflects whether leakage has occurred, sample battery testing or actual leakage events are needed, or leakage conditions can be simulated under laboratory conditions. Specifically, data logging devices or systems can be used to capture battery data, including voltage, current, temperature, and pressure. The collected data is then formatted for analysis, such as time series data, where each time point has corresponding parameter values. Optionally, in some embodiments of this application, the data can be preprocessed, such as noise removal, missing value imputation, and smoothing, to ensure data quality. Furthermore, the data can be labeled to indicate which data represents normal battery operation and which represents leakage.
[0026] 102. Standardize the sample battery data and divide the processed sample battery data into training and testing sets.
[0027] Data standardization involves scaling the data proportionally to fit it into a small, specific range, most commonly [0, 1]. For battery data, the specifics are as follows:
[0028] Data standardization involves scaling the data proportionally to fit it into a small, specific range, most commonly [0, 1]. For battery data, the specifics are as follows:
[0029] a. Calculate the mean and standard deviation
[0030] Mean: Calculates the average of all sample data.
[0031] Standard deviation: Calculates the degree of dispersion of data, i.e. the width of the data distribution.
[0032] b. Apply standardized formulas
[0033] Standardize each data point x using the following formula:
[0034] ,in, It is the standardized value. It is the mean. It is the standard deviation.
[0035] In addition, any missing data points need to be processed or imputed before standardization to avoid problems when calculating the mean and standard deviation.
[0036] In addition, any missing data points need to be processed or filled before standardization to avoid problems when calculating the mean and standard deviation.
[0037] Optionally, in some embodiments of this application, the step "standardizing the sample battery data" may specifically include:
[0038] The average value of the sample battery data during leakage is determined as a first preset value, and;
[0039] The variance of the sample battery data during leakage is determined as a second preset value.
[0040] The sample battery data can specifically include voltage, temperature, and pressure. When the battery leaks, the mean values of these data can be calculated separately, and the mean values of voltage, temperature, and pressure can be set as first preset values (e.g., 0), representing typical values of these parameters under leakage conditions. Similarly, when the battery leaks, the variances of the voltage, temperature, and pressure data can be calculated separately, and the variances of voltage, temperature, and pressure can be set as second preset values (e.g., 1).
[0041] After standardization, the processed data is divided into training and testing sets, with a ratio of 7:3 (70% for training and 30% for testing). However, this ratio can be adjusted based on specific circumstances. Random sampling can be used to divide the dataset into training and testing sets, ensuring that samples in each set are randomly selected to avoid any potential bias. Optionally, in some embodiments of this application, additional processing can be performed after dividing the dataset, such as:
[0042] Data augmentation: If the amount of data is small, data augmentation techniques can be used to increase the sample size.
[0043] Feature selection: Determine which features are most important for training the model, which may require removing some irrelevant or redundant features.
[0044] 103. Train the preset long short-term memory model based on the training set, and determine the leakage detection model from the trained long short-term memory model based on the test set.
[0045] Long Short-Term Memory (LSTM) is a special type of Recurrent Neural Network (RNN) architecture that excels at processing sequential data, especially time series analysis. LSTM was originally designed to address the vanishing or exploding gradient problems that traditional RNNs encounter when processing long sequences of data, enabling LSTM to capture long-term dependencies.
[0046] First, a pre-defined LSTM can be obtained. The architecture of this LSTM may include one or more LSTM layers and a final output layer. The activation function of the output layer depends on the nature of the problem (e.g., for binary classification problems, the sigmoid activation function is typically used). Then, the training set data is used as input features, and the corresponding leakage labels are used as target outputs. During training, a callback function (such as ModelCheckpoint) can be used to save the model weights. This callback function can be set to save the current best model weights after each epoch if the performance on the validation set has improved. Optionally, in some embodiments of this application, the step "training the pre-defined long short-term memory model based on the training set and determining the leakage detection model in the trained long short-term memory model based on the test set" may specifically include:
[0047] Obtain the preset long short-term memory model;
[0048] The long short-term memory model is trained using the training set;
[0049] The model weights are stored in the short-term memory model during the training period according to the preset callback function;
[0050] The leakage detection model is determined based on the model weights and the test set in the trained long short-term memory model.
[0051] For example, specifically, when training a Long Short-Term Memory (LSTM) model using data from the training set, the LTM model learns the relationship between the input data and the leak labels. During training, a callback function (such as ModelCheckpoint in TensorFlow or Keras) is used to save the model's weights. This callback function can be set to save the current best model weights after each epoch if performance on the validation set has improved. This way, even if training is interrupted, training can continue from the saved best weights without starting from scratch. Finally, the saved best model weights are used to initialize a new or identical LSTM model. This model with the best weights is evaluated on the test set to determine its performance. Since the test set is data not used during training, it can provide important information about the model's generalization ability. Specifically, the final leak detection model can be determined based on performance metrics on the test set (such as accuracy, recall, F1 score, etc.).
[0052] Optionally, in some embodiments of this application, the step "training the long short-term memory model using a training set" may specifically include:
[0053] Predicted leakage data based on the training set using a long short-term memory model;
[0054] The long short-term memory model is trained based on the actual leakage data and the estimated leakage data corresponding to the training set.
[0055] For example, preliminary leakage prediction data is generated based on training set features. Then, based on the actual leakage data and predicted leakage data corresponding to the training set, the loss of the Long Short-Term Memory (LSTM) model is generated, and the model weights are adjusted based on this loss. Simultaneously, callback functions, such as Early Stopping or Model Checkpointing, are used to save the best-performing model during training. The model's performance is evaluated on the validation set to check for overfitting or underfitting. Based on the performance evaluation results, the model's hyperparameters, such as learning rate, batch size, number of LSTM layers, or number of units, are adjusted.
[0056] Optionally, in some embodiments of this application, the step "determining the leakage detection model in the trained long short-term memory model based on model weights and the test set" may specifically include:
[0057] Determine the model weights and model loss of the Long Short-Term Memory model at time n, where n is an integer greater than 0;
[0058] Update the model weights of the Long Short-Term Memory (LSTM) model at time n+1, and train the LSM model with updated weights based on the training set.
[0059] The model loss corresponding to the long short-term memory model at time n+1 is determined based on the test set.
[0060] Based on the model loss of the long short-term memory model at time n and the model loss at time n+1, a leakage detection model is determined in the trained long short-term memory model.
[0061] Model weights are a set of internal parameters of the model that determine how the model maps input data to output. Model loss is a metric that measures the difference between the model's predictions and the true labels, such as mean squared error or cross-entropy loss. For example, at the end of the first training epoch, the model weights and the loss on the validation set are recorded. Then, the model weights at the second time step are updated based on the model loss, and the long short-term memory model with updated weights is trained based on the training set. The model loss corresponding to the long short-term memory model at the second time step is determined based on the test set. If the model loss at the second time step is less than that at the first time step, it indicates that the model is learning and improving its performance. If the model loss decreases, the training process continues; if the model loss increases or no longer decreases significantly, the model can be adjusted or training can be stopped to avoid overfitting. Optionally, in some embodiments of this application, if the model loss remains unchanged when the number of iterations meets a preset condition, the long short-term memory model trained in the last iteration is determined as the leakage detection model.
[0062] Optionally, in some embodiments of this application, the step "determining the leakage detection model in the trained long short-term memory model based on the model loss of the long short-term memory model at time n and the model loss at time n+1" may specifically include:
[0063] Compare the magnitude of the model loss of the Long Short-Term Memory model at time n and time n+1;
[0064] The long short-term memory model with the least model loss was selected as the leakage detection model.
[0065] For example, at the end of the nth training cycle, record the model's loss on the validation set, denoted as L. n At the end of the (n+1)th training cycle, the model's loss on the validation set is recorded again, denoted as L. n+1 If L n >L n+1 This indicates that the model's performance has improved in the new training cycle. If L n+1 >L n If the result is negative, it indicates that the model performance has not improved, or there is a risk of overfitting.
[0066] By comparing the model losses across different training epochs, the model with the lowest loss is selected as the optimal model, and the model weights with the lowest loss are saved at the end of each training epoch. This way, even if the training process is interrupted, training can continue from the optimal model weights, or the optimal model can be used directly for prediction.
[0067] Optionally, in some embodiments of this application, the battery leakage detection method of this application may further include:
[0068] The effectiveness of model training is tested based on the model loss of the Long Short-Term Memory model at time n and the model loss at time n+1.
[0069] If the test result is valid, the error between the estimated leakage data of the sample battery predicted by the latest long short-term memory model and the actual leakage data of the sample battery is calculated.
[0070] When the error meets the preset conditions, the latest long short-term memory model is determined as the leakage detection model.
[0071] For example, after training, plot the training loss and validation loss curves as a function of the number of training epochs. If the loss curves show a decreasing trend, the training is effective; save and output the model. If the loss curves show a non-decreasing trend, the training is ineffective, and the model needs to be retrained. Calculate the root mean square error (RMSE) and mean absolute percentage error (MAPE) between the model's predicted leakage data and the actual leakage data. The smaller the error, the higher the model's accuracy.
[0072] 104. Perform battery leakage detection on the battery under test according to the leakage detection model.
[0073] For example, specifically, real-time data of the battery under test is collected, including parameters such as voltage, temperature, and pressure. This real-time data is input into a leakage detection model, which predicts whether the battery under test is in a leakage state. If leakage is detected, an alarm can be triggered or maintenance personnel can be notified to take appropriate measures. That is, optionally, in some embodiments of this application, the step "performing battery leakage detection on the battery under test according to the leakage detection model" may specifically include:
[0074] Battery data of the battery under test is acquired by a sensor installed inside the battery under test;
[0075] The battery data is processed based on the leakage detection model to predict whether the battery under test is in a leakage state.
[0076] If the battery under test is in a state of leakage, the system will control the battery to stop charging and discharging, and output the corresponding fault information for battery leakage.
[0077] For example, sensors are deployed inside the battery under test to monitor key parameters such as voltage, temperature, and pressure in real time. Then, a pre-trained leak detection model is loaded, and the data is input into it. The leak detection model will estimate whether the battery is leaking based on the input data. If a leak is detected, a control command is sent to the battery management system (BMS) to instruct it to stop the battery's charging and discharging operations. Simultaneously, corresponding fault information is output, including a leak alarm, battery identification, timestamp, and sensor readings.
[0078] The battery energy storage system (BMS) has a three-level architecture, consisting of a master control unit (MBMU), a main control unit (SBMU), and slave control units (VCMU). The slave control units (VCMU) collect the temperature and voltage of the battery cells and upload the data to the master and main control units. The master and main control units then determine the status and faults of the battery cells and issue control commands.
[0079] When a battery cell leaks, the negative electrode aluminum foil corrodes, causing a drop in the voltage of the individual cell. During leakage, a chemical reaction occurs inside the cell, releasing a large amount of heat and raising the cell temperature. The normal internal pressure of the cell is atmospheric pressure (101.325 kPa), and the chemical reaction also generates VOC gases, leading to an increase in internal pressure. Based on the phenomenon of battery cell leakage, a sensor is added at the location of the cell's explosion-proof valve to collect the cell's voltage, temperature, and internal pressure. Based on these three data points, an intelligent algorithm is developed to centrally determine whether a cell is leaking. Therefore, this application provides a battery leakage detection system, as shown in Figure 2. This system includes a main control Bluetooth module, a battery pack, and the battery cell. The BMS can continuously detect cell leakage. Through Bluetooth technology, the sensor transmits the collected data to a paired Bluetooth receiver board inside the battery pack. Since the reliability and stability of Bluetooth communication are related to the communication distance, both the Bluetooth receiver board and the VCMU are located inside the battery pack. Short-range Bluetooth communication is more reliable, ensuring that cell data is not lost. It's important to note that the VCMU's function is to collect the temperature and voltage of the battery cells. The data collected by the sensors is not used as the actual temperature and voltage output of the cells, but only to determine whether the cells are leaking. Therefore, the Bluetooth receiver board transmits the data in real time via the CAN communication line to the Bluetooth module designed in the VCMU software. No filtering is required. The VCMU then sends this data to the Bluetooth module in the main control SBMU software, where the SBMU determines the leakage status of the battery cells.
[0080] The sensors are located inside the battery cell, collecting data on cell voltage, internal temperature, and internal pressure every 1000ms. Each sensor is equipped with a Bluetooth transmitter, paired with a corresponding Bluetooth receiver board. The collected data is transmitted via the Bluetooth transmitter to the Bluetooth receiver board inside the battery pack. The Bluetooth receiver board then transmits the data to the VCMU via CAN communication. The Bluetooth module in the VCMU software sends the data to the main controller. To reflect the true state of each cell, the data is not processed and is directly sent to the Bluetooth module of the main controller SBMU. The SBMU determines whether the cell is leaking. If there is no leakage, the system charges and discharges normally; if there is leakage, the main controller MBMU issues a command to stop charging and discharging the battery pack containing the leaking cell and issues a cell leakage fault message.
[0081] A Bluetooth module has been added to the VCMU and SBMU software. The VCMU's Bluetooth module receives cell leakage data, checks the voltage, temperature, and pressure data of each cell, and matches them with the cell number. If data is missing, that set of data is not transmitted; the next set of collected data overwrites this data and is sent to the SBMU. The SBMU's newly added Bluetooth module receives this data. This Bluetooth module contains intelligent algorithm code that determines whether there is leakage based on the cell data. If leakage is detected, a cell leakage fault is reported; otherwise, the system operates normally.
[0082] This application's embodiment, after acquiring sample battery data when leakage occurs, standardizes the sample battery data and divides it into a training set and a test set. Then, it trains a pre-defined Long Short-Term Memory (LSTM) model based on the training set, and determines a leakage detection model from the trained LSM model using the test set. Finally, it performs battery leakage detection on the battery under test based on the leakage detection model. The battery leakage detection scheme provided in this application pre-trains a pre-defined LSM model using sample battery data when leakage occurs, and determines a leakage detection model from the trained LSM model using the test set. During actual battery leakage detection, the trained leakage detection model can be directly used to detect leakage in the battery under test, thereby improving the timeliness of battery leakage detection.
[0083] Furthermore, this application also provides an electronic device, as shown in FIG3, which illustrates a structural schematic diagram of the electronic device involved in this application embodiment. Specifically:
[0084] The electronic device may include components such as a processor 301 with one or more processing cores, a memory 302 with one or more computer-readable storage media, a power supply 303, and an input unit 304. Those skilled in the art will understand that the electronic device structure shown in FIG3 does not constitute a limitation on the electronic device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein:
[0085] The processor 301 is the control center of the electronic device. It connects various parts of the electronic device via various interfaces and lines. By running or executing software programs and / or modules stored in the memory 302, and by calling data stored in the memory 302, it performs various functions and processes data, thereby providing overall monitoring of the electronic device. Optionally, the processor 301 may include one or more processing cores; preferably, the processor 301 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, and the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 301.
[0086] The memory 302 can be used to store software programs and modules. The processor 301 executes various functional applications and battery leakage detection by running the software programs and modules stored in the memory 302. The memory 302 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 302 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 302 may also include a memory controller to provide the processor 301 with access to the memory 302.
[0087] The electronic device also includes a power supply 303 that supplies power to various components. Preferably, the power supply 303 can be logically connected to the processor 301 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The power supply 303 may also include one or more DC or AC power supplies, recharging systems, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components.
[0088] The electronic device may also include an input unit 304, which can be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal inputs related to user settings and function control.
[0089] Although not shown, the electronic device may also include a display unit, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 301 in the electronic device loads the executable files corresponding to the processes of one or more applications into the memory 302 according to the following instructions, and the processor 301 runs the applications stored in the memory 302 to realize various functions, as follows:
[0090] Acquire sample battery data when battery leakage occurs; standardize the sample battery data and divide the processed sample battery data into training set and test set; train a preset long short-term memory model based on the training set, and determine the leakage detection model in the trained long short-term memory model based on the test set; perform battery leakage detection on the battery to be tested according to the leakage detection model.
[0091] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.
[0092] This application's embodiment, after acquiring sample battery data when leakage occurs, standardizes the sample battery data and divides it into a training set and a test set. Then, it trains a pre-defined Long Short-Term Memory (LSTM) model based on the training set, and determines a leakage detection model from the trained LSM model using the test set. Finally, it performs battery leakage detection on the battery under test based on the leakage detection model. The battery leakage detection scheme provided in this application pre-trains a pre-defined LSM model using sample battery data when leakage occurs, and determines a leakage detection model from the trained LSM model using the test set. During actual battery leakage detection, the trained leakage detection model can be directly used to detect leakage in the battery under test, thereby improving the timeliness of battery leakage detection.
[0093] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.
[0094] Therefore, embodiments of this application provide a storage medium storing a plurality of instructions that can be loaded by a processor to execute steps in any of the battery leakage detection methods provided in embodiments of this application. For example, the instructions can execute the following steps:
[0095] Acquire sample battery data when battery leakage occurs; standardize the sample battery data and divide the processed sample battery data into training set and test set; train a preset long short-term memory model based on the training set, and determine the leakage detection model in the trained long short-term memory model based on the test set; perform battery leakage detection on the battery to be tested according to the leakage detection model.
[0096] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.
[0097] The storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.
[0098] Since the instructions stored in the storage medium can execute the steps of any of the battery leakage detection methods provided in the embodiments of this application, the beneficial effects that any of the battery leakage detection methods provided in the embodiments of this application can achieve can be realized. For details, please refer to the previous embodiments, which will not be repeated here.
[0099] The battery leakage detection method, electronic device, and storage medium provided in the embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. A method for detecting battery leakage, comprising: Acquire sample battery data when battery leakage occurs; The sample battery data is standardized, and the processed sample battery data is divided into a training set and a test set. The preset long short-term memory model is trained based on the training set, and the leakage detection model is determined from the trained long short-term memory model based on the test set. The battery leakage detection model is used to detect battery leakage in the battery under test.
2. The battery leakage detection method according to claim 1, wherein, The step of training a preset long short-term memory model based on the training set, and determining a leakage detection model from the trained long short-term memory model based on the test set, includes: Obtain the preset long short-term memory model; The long short-term memory model is trained using the training set. The model weights of the long short-term memory model are stored during training according to a preset callback function; Based on the model weights and the test set, a leakage detection model is determined in the trained long short-term memory model.
3. The battery leakage detection method according to claim 2, wherein, The step of determining the leakage detection model in the trained long short-term memory model based on the model weights and the test set includes: Determine the model weights and model loss of the long short-term memory model at time n, where n is an integer greater than 0; Update the model weights of the long short-term memory model at time n+1, and train the long short-term memory model with updated weights based on the training set; Based on the test set, determine the model loss corresponding to the long short-term memory model at time n+1; Based on the model loss of the long short-term memory model at time n and the model loss at time n+1, a leakage detection model is determined in the trained long short-term memory model.
4. The battery leakage detection method according to claim 3, wherein, The step of determining the leakage detection model in the trained long short-term memory model based on the model loss at time n and the model loss at time n+1 includes: Compare the magnitude of the model loss of the Long Short-Term Memory model at time n and time n+1; The long short-term memory model with the least model loss was selected as the leakage detection model.
5. The battery leakage detection method according to claim 4 further includes: The effectiveness of the model training is tested based on the model loss of the Long Short-Term Memory model at time n and the model loss at time n+1. If the detection result is valid, then calculate the error between the estimated leakage data corresponding to the sample battery predicted by the latest long short-term memory model and the actual leakage data corresponding to the sample battery. When the error meets the preset conditions, the latest long short-term memory model is determined as the leakage detection model.
6. The battery leakage detection method according to claim 2, wherein, The step of training the long short-term memory model using the training set includes: Based on the long short-term memory model, the predicted leakage data corresponding to the training set is estimated. The long short-term memory model is trained based on the actual leakage data and the estimated leakage data corresponding to the training set.
7. The battery leakage detection method according to any one of claims 1 to 6, wherein, The standardization process for the sample battery data includes: The average value of the sample battery data during leakage is determined as a first preset value, and; The variance of the sample battery data during leakage is determined as a second preset value.
8. The battery leakage detection method according to any one of claims 1 to 6, wherein, The step of performing battery leakage detection on the battery under test according to the leakage detection model includes: Battery data of the battery under test is acquired by a sensor installed inside the battery under test; The battery data is processed according to the leakage detection model to estimate whether the battery under test is in a leakage state. If the battery under test is in a leaking state, control the battery under test to stop charging and discharging, and output the fault information corresponding to the battery leakage; The detection module is used to perform battery leakage detection on the battery under test according to the leakage detection model.
9. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein, When the processor executes the program, it implements the steps of the battery leakage detection method as described in any one of claims 1-8.
10. A computer-readable storage medium, characterized in that, It stores a computer program, wherein when the computer program is executed by a processor, it implements the steps of the battery leakage detection method as described in any one of claims 1-8.