A battery low vacuum baking cavity gas leakage detection method and system based on air pressure high frequency oscillation characteristics

Abstract

This invention discloses a method and system for detecting air leakage in a low-vacuum baking chamber of a battery based on high-frequency oscillation feature recognition of gas pressure, relating to the field of lithium-ion battery manufacturing technology. The method includes: acquiring gas pressure waveform data within the baking chamber during the low-vacuum baking stage and preprocessing the gas pressure waveform data; extracting time-series features from the preprocessed gas pressure waveform data, including oscillation frequency per unit time, average rise time, root mean square deviation of rise time, and maximum fluctuation amplitude; determining leakage based on the time-series features, and if leakage occurs, issuing an alarm and recording the corresponding time-series features and their corresponding gas pressure waveform data within the specified unit time. This invention comprehensively depicts the dynamic law of gas pressure during leakage by real-time acquisition and preprocessing of gas pressure waveform data within the baking chamber, followed by time-series feature extraction of the acquisition results. Based on the extracted time-series features, multi-dimensional leakage determination is performed, enabling accurate and real-time leakage detection.

Description

Technical Field

[0001] This invention relates to the field of lithium-ion battery manufacturing technology, and in particular to a method and system for detecting air leakage in a low-vacuum baking cavity of a battery based on high-frequency oscillation characteristics of gas pressure. Background Technology

[0002] In the lithium-ion battery manufacturing process, the baking process is a crucial step in removing residual moisture from the cell electrodes and separators, directly affecting the battery's cycle life, safety performance, and electrochemical performance. Industry standards clearly require that the final moisture content of power battery cells be controlled below 300 ppm. If the moisture content exceeds the standard, it may cause safety accidents such as cell swelling and thermal runaway.

[0003] The low-vacuum baking stage is a core part of the baking process. Its goal is to maintain a low vacuum environment of 60-180 Pa to accelerate the evaporation and removal of moisture inside the battery cell. The stability of the vacuum level in this stage directly determines the moisture removal efficiency. Currently, the industry often adopts a configuration mode of "multi-oven shared air pump system" (e.g., one air pump powers 4-6 ovens) to reduce equipment investment costs. However, this mode has significant risks: if a sealing defect occurs in the cavity of an oven (e.g., aging of the sealing ring, cracking of the cavity weld), it will cause abnormal pressure fluctuations in the low-vacuum stage (e.g., ...). Figure 1 As shown in Figure 2, this not only affects the baking effect of the oven, but also continuously consumes the output flow of the air pump, causing the vacuum degree of other ovens sharing the pump to fail to meet the standard, resulting in a large-area battery cell baking quality problem.

[0004] Existing pressure monitoring technologies for the low-vacuum baking stage mainly fall into two categories: one is "static threshold monitoring," which sets a fixed upper limit for pressure (e.g., 500 Pa). When the pressure continuously exceeds the threshold, an alarm is triggered. This approach cannot identify hidden leaks where the pressure is within the acceptable range of 60-180 Pa but there are high-frequency oscillations. As a result, leaks are not detected in their early stages, and by the time the pressure exceeds the standard, a batch of battery cells has already exceeded the moisture limit (some cells even have a moisture content exceeding 400 ppm). The other type is "low-frequency sampling monitoring," which uses a sampling frequency of less than 0.1 Hz (collecting no more than 6 data points per minute). This approach cannot capture the high-frequency oscillation process where "single climb time is less than 1 minute," resulting in a false negative rate as high as 30%.

[0005] As battery manufacturing evolves towards higher consistency and faster production cycles, traditional monitoring technologies are no longer sufficient to address the risk of "local problems spreading to the whole system" caused by cavity leaks. The rework rate of cells due to undetected cavity leaks accounts for over 40% of the total rework rate in the baking process, with rework costs reaching tens of thousands of yuan per batch. Therefore, there is an urgent need for a technical solution that can identify high-frequency pressure oscillations in the low-vacuum stage in real time and quickly trigger leak alarms, enabling early detection and handling of leaks to prevent further quality losses. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and system for detecting air leakage in a battery low-vacuum baking cavity based on high-frequency oscillation feature recognition of air pressure, thereby solving the technical problem that the accuracy and real-time performance of the existing technical solutions for detecting air leakage in a battery low-vacuum baking cavity cannot meet the needs of battery manufacturing development.

[0007] To achieve the above objectives, the present invention is implemented using the following technical solution:

[0008] In a first aspect, the present invention provides a method for detecting air leakage in a battery low-vacuum baking cavity based on high-frequency oscillation characteristics of gas pressure, comprising:

[0009] Acquire the gas pressure waveform data inside the baking chamber during the low vacuum baking stage, and preprocess the gas pressure waveform data.

[0010] The preprocessed air pressure waveform data is subjected to time series feature extraction, which includes oscillation frequency, average rise time, root mean square error of rise time, and maximum fluctuation amplitude per unit time.

[0011] Based on the aforementioned timing characteristics, a leak is determined. If a leak occurs, an alarm is triggered and the timing characteristics and corresponding air pressure waveform data for the corresponding unit time period are recorded.

[0012] This invention obtains complete gas pressure waveform data by real-time acquisition and preprocessing of gas pressure waveform data in the baking chamber during the low vacuum baking stage. Then, it extracts time-series features, which can comprehensively depict the dynamic law of gas pressure when there is leakage. Based on the extracted time-series features, multi-dimensional leakage judgment is performed, which can achieve accurate and real-time leakage detection.

[0013] Optionally, acquiring the gas pressure waveform data inside the baking chamber during the low-vacuum baking stage includes:

[0014] A pressure sensor is installed in the baking cavity near the vacuum interface.

[0015] During the low-vacuum baking stage, high-frequency sampling is performed by increasing the output rate of the pressure sensor or the sampling frequency of the back-end ADC.

[0016] When exiting the low vacuum baking stage, a low-frequency standby is performed by reducing the output rate of the pressure sensor or the sampling frequency of the back-end ADC.

[0017] This invention, by placing the pressure sensor within the baking chamber near the vacuum interface, avoids the influence of chamber wall temperature changes on pressure measurement, thus improving sampling accuracy. Simultaneously, through high-frequency sampling and low-frequency standby, it ensures the acquisition of accurate and detailed pressure data, providing a higher-quality data foundation for subsequent analysis, while also reducing energy consumption.

[0018] Optionally, the preprocessing of the air pressure waveform data includes:

[0019] The air pressure waveform data was screened using the 3σ principle to remove outliers and replace them with the moving average of the first 3 data points;

[0020] Using the start time of entering the low vacuum baking stage as the time reference, the gas pressure waveform data in each baking chamber is timestamped and aligned.

[0021] This invention employs the 3σ principle for data filtering, objectively and adaptively identifying and handling outliers based on the statistical characteristics of the data itself, thereby improving data quality and analytical reliability. Timestamp alignment ensures that the pressure data from different ovens are comparable over time, facilitating subsequent collaborative monitoring of multiple ovens.

[0022] Optionally, the step of extracting time-series features from the preprocessed air pressure waveform data includes:

[0023] The preprocessed air pressure waveform data is traversed, and all peaks and troughs are identified based on amplitude changes.

[0024] The number of oscillation cycles is identified by the sequence of peaks and troughs within a unit of time, and the oscillation frequency is determined based on the number of oscillation cycles.

[0025] The average climb time is obtained by taking the average of the peak and trough times within a unit of time.

[0026] Calculate the mean squared error of climb time based on the climb time for each climb and the average climb time per unit time.

[0027] The maximum fluctuation range is calculated based on the maximum and minimum air pressure amplitudes per unit time.

[0028] This invention extracts four core features from air pressure data: oscillation frequency, average rise time, mean square error of rise time, and maximum fluctuation amplitude. This allows for a comprehensive depiction of the dynamic pattern of air pressure during leaks, providing data support for subsequent multi-dimensional leak detection.

[0029] Optionally, identifying all peaks and troughs based on amplitude changes includes:

[0030] If data points Simultaneously satisfy:

[0031]

[0032] Then data points At the peak of the wave;

[0033] If data points Simultaneously satisfy:

[0034]

[0035] Then data points Located at a trough;

[0036] In the formula, The first part of the preprocessed air pressure waveform data is shown below. The magnitude of each data point For the set air pressure threshold, The unit is Pa.

[0037] This invention identifies all peaks and troughs based on amplitude changes, and has advantages such as simplicity, speed, and no need for preset parameters. It is suitable for scenarios with high real-time requirements, limited computing resources, or the need for basic rapid detection, and can effectively improve the real-time performance of leak detection.

[0038] Optionally, the step of determining leakage based on the timing characteristics includes:

[0039] If the aforementioned time-series characteristics simultaneously satisfy:

[0040]

[0041] If the amplitude of each data point within a unit of time is within the preset air pressure range, then it is determined that an air leak has occurred.

[0042] In the formula, The first The oscillation frequency, average rise time, mean squared deviation of rise time, and maximum fluctuation amplitude per unit time; For the corresponding judgment threshold, The unit is Hz. The unit is s. The unit is Pa.

[0043] This invention improves the accuracy of leak detection, the rate of identifying hidden leaks, and reduces the false alarm rate by using multi-dimensional leak determination.

[0044] Optionally, the step of determining leakage based on the timing characteristics further includes:

[0045] If a leak is detected in the current unit of time, the leak detection continues for multiple consecutive units of time.

[0046] If multiple consecutive time units are determined to be leaking, the determination result for the current time unit is confirmed; otherwise, the determination result for the current time unit is modified to occasional fluctuation.

[0047] In order to prevent false alarms when a leak is detected, the present invention further incorporates continuous verification, which can effectively reduce the false alarm rate.

[0048] Secondly, the present invention provides a battery low-vacuum baking cavity leakage detection system based on high-frequency oscillation feature recognition of gas pressure, comprising:

[0049] The data processing module is configured to acquire the gas pressure waveform data inside the baking chamber during the low vacuum baking stage, and to preprocess the gas pressure waveform data.

[0050] The feature extraction module is configured to extract time-series features from the preprocessed air pressure waveform data. The time-series features include the oscillation frequency per unit time, the average rise time, the root mean square error of the rise time, and the maximum fluctuation amplitude.

[0051] The air leakage detection module is configured to detect air leakage based on the time sequence characteristics. If an air leakage occurs, an alarm is triggered and the time sequence characteristics and their corresponding air pressure waveform data within the corresponding unit time are recorded.

[0052] Optionally, the leakage detection module is further configured to:

[0053] If a leak is detected in the current unit of time, the leak detection continues for multiple consecutive units of time.

[0054] If multiple consecutive time units are determined to be leaking, the determination result for the current time unit is confirmed; otherwise, the determination result for the current time unit is modified to occasional fluctuation.

[0055] Thirdly, the present invention provides an electronic device, including a processor and a storage medium;

[0056] The storage medium is used to store instructions;

[0057] The processor is configured to operate according to the instructions to perform the steps according to the method described above.

[0058] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the above-described method.

[0059] Fifthly, the present invention provides a computer program product, including a computer program / instructions that, when executed by a processor, implement the steps of the above-described method.

[0060] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0061] The present invention provides a method and system for detecting air leakage in a battery low-vacuum baking chamber based on high-frequency oscillation feature recognition of air pressure. By real-time acquisition and preprocessing of air pressure waveform data in the baking chamber during the low-vacuum baking stage, complete air pressure waveform data is obtained. Then, time-series features are extracted from the data. The time-series features can comprehensively depict the dynamic law of air pressure during leakage. Based on the extracted time-series features, multi-dimensional leakage judgment is performed, which can achieve accurate and real-time leakage detection. Attached Figure Description

[0062] Figure 1 This is a diagram showing the air pressure fluctuation when air leakage occurs in the oven chamber, as provided in the background section of this invention.

[0063] Figure 2 This is the fluctuation diagram of the oven cavity under normal conditions provided in the background art of this invention;

[0064] Figure 3 This is a flowchart of a battery low-vacuum baking cavity leakage detection method provided in an embodiment of the present invention. Detailed Implementation

[0065] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0066] Example 1

[0067] like Figure 1 As shown, this embodiment of the invention provides a method for detecting air leakage in a battery low-vacuum baking cavity based on high-frequency oscillation characteristics of air pressure, including the following steps:

[0068] Step S1: Obtain the gas pressure waveform data inside the baking cavity during the low vacuum baking stage, and preprocess the gas pressure waveform data.

[0069] Specifically, in this embodiment, the hardware selection and deployment are as follows:

[0070] a. Sensor selection: A high-precision capacitive vacuum gauge is used, with a measurement range of 1-100000Pa, an accuracy of ±1Pa, and a sampling frequency that is adjustable from 0.1-10Hz. Here, it is set to 0.2Hz to ensure high-frequency data acquisition.

[0071] b. Deployment location: The vacuum gauge is installed inside the oven cavity near the vacuum interface to avoid the influence of cavity wall temperature changes on the pressure measurement.

[0072] c. Data transmission: The air pressure data (including timestamp, accurate to milliseconds) collected by the vacuum gauge is transmitted in real time with a transmission delay of ≤100ms to avoid feature loss due to data delay.

[0073] d. Sampling trigger logic: Through the process signal linkage of the oven PLC system, when the oven enters the low vacuum baking stage (PLC sends "LowVacuum_Start" signal), 0.2Hz high-frequency sampling is automatically started; when exiting the low vacuum stage (PLC sends "LowVacuum_End" signal), sampling stops and switches to 0.1Hz low-frequency standby mode to reduce energy consumption.

[0074] Data preprocessing:

[0075] a. Outlier Removal: The raw air pressure data is screened using the 3σ principle. The mean μ (approximately 120 Pa) and standard deviation σ (approximately 25 Pa) of historical normal air pressure data are calculated. If a data point exceeds the range of [μ-3σ, μ+3σ] (i.e., 45-195 Pa), it is determined to be an outlier caused by sensor malfunction (e.g., a sudden jump to 1000 Pa). Then, the moving average of the previous three data points is used to replace it, ensuring data continuity. The 3σ principle for data screening, based on the statistical characteristics of the data itself, objectively and adaptively identifies and processes outliers, thereby improving data quality and analytical reliability.

[0076] b. Timestamp alignment: Using the time when the oven PLC sends the "LowVacuum_Start" signal as the time base (t=0), the air pressure data of each oven is timestamped to ensure that the air pressure data of different ovens can be compared in the time dimension, which facilitates subsequent multi-oven collaborative monitoring.

[0077] Step S2: Extract time-series features from the preprocessed air pressure waveform data. The time-series features include the oscillation frequency per unit time, the average rise time, the root mean square error of the rise time, and the maximum fluctuation amplitude.

[0078] Specifically, in this embodiment, the time-series feature extraction of the preprocessed air pressure waveform data includes:

[0079] Step S2.1: Traverse the preprocessed air pressure waveform data and identify all peaks and troughs based on amplitude changes.

[0080] If data points Simultaneously satisfy:

[0081]

[0082] Then data points At the peak of the wave;

[0083] If data points Simultaneously satisfy:

[0084]

[0085] Then data points Located at a trough;

[0086] In the formula, The first part of the preprocessed air pressure waveform data is shown below. The magnitude of each data point For the set air pressure threshold, The unit is Pa.

[0087] This invention identifies all peaks and troughs based on amplitude changes, and has advantages such as simplicity, speed, and no need for preset parameters. It is suitable for scenarios with high real-time requirements, limited computing resources, or the need for basic rapid detection, and can effectively improve the real-time performance of leak detection.

[0088] Step S2.2: Identify the number of oscillation cycles based on the sequence of peaks and troughs per unit time, and determine the oscillation frequency based on the number of oscillation cycles.

[0089] If we consider "trough → peak → trough" as a complete oscillation cycle (i.e., the process of air pressure rising from the lowest value to the highest value and then falling back to the lowest value), then the number of oscillation cycles per minute is the oscillation frequency f. For example, if three complete oscillation cycles are identified within a minute, then f = 3.

[0090] Step S2.3: Identify the time of each climb based on the timestamps of the peaks and troughs within a unit of time, and obtain the average climb time by taking the average of each climb time.

[0091] For example: starting from the identified trough, track to the next peak; this interval is the pressure rise segment (i.e., the process of pressure rising from the trough to the peak); extract the timestamp of the starting trough point of the rise segment. Timestamp of the end peak point Calculate the time difference between the two, which is the time for a single climb. The formula is: If there are 3 oscillation cycles within a 3-minute period, then there are 3 climbing times. Calculate their average value as the average climbing time.

[0092] Step S2.4: Calculate the mean squared error of the climb time based on the climb time for each climb per unit time and the average climb time.

[0093] For example: based on k single climb times within a certain 3-minute period. ;

[0094] First calculate the average climb time. Then, the fluctuation of the climb time is quantified using the mean squared error formula, which is:

[0095]

[0096] When there is an air leak, the cavity sealing defects remain stable (such as aging of the fixed sealing ring), the air pressure oscillation pattern is consistent, and the rise time fluctuation is small. During normal process fluctuations (such as fine-tuning of the air pump pressure), the climb time varies greatly depending on the control settings. This feature can effectively distinguish between two types of fluctuations.

[0097] Step S2.5: Calculate the maximum fluctuation amplitude based on the difference between the maximum and minimum air pressure amplitudes per unit time.

[0098] This invention extracts four core features from air pressure data: oscillation frequency, average rise time, mean square error of rise time, and maximum fluctuation amplitude. This allows for a comprehensive depiction of the dynamic pattern of air pressure during leaks, providing data support for subsequent multi-dimensional leak detection.

[0099] Step S3: Determine if there is a leak based on the timing characteristics. If a leak occurs, issue an alarm and record the timing characteristics and corresponding air pressure waveform data within the corresponding unit time.

[0100] Based on 500 historical cases of cavity leakage and 1000 normal low-vacuum baking process data, this invention determines four-dimensional judgment thresholds through statistical analysis (such as box plots and ROC curves). At the same time, combined with the core gas pressure range of low-vacuum baking, a complete judgment standard is formed, as shown in Table 1.

[0101] Table 1: Judgment Criteria

[0102]

[0103] Based on the above criteria, leakage determination according to time-series characteristics includes:

[0104] If the time series characteristics simultaneously satisfy:

[0105]

[0106] If the amplitude of each data point within a unit of time is within the preset air pressure range, then it is determined that an air leak has occurred.

[0107] In the formula, The first The oscillation frequency, average rise time, mean squared deviation of rise time, and maximum fluctuation amplitude per unit time; For the corresponding judgment threshold, The unit is Hz. The unit is s. The unit is Pa. According to the above standard... The corresponding frequencies are 1Hz, 60s, 5s, and 50Pa.

[0108] This invention improves the accuracy of leak detection, the rate of identifying hidden leaks, and reduces the false alarm rate by using multi-dimensional leak determination.

[0109] This invention further supplements the false positive suppression mechanism, which can effectively reduce the false positive rate:

[0110] (1) Continuous verification:

[0111] If a leak is detected in the current unit of time, the leak detection continues for multiple consecutive units of time.

[0112] If multiple consecutive time units are determined to be leaking, the determination result for the current time unit is confirmed; otherwise, the determination result for the current time unit is modified to occasional fluctuation.

[0113] (2) Process status verification:

[0114] Check the current process status of the oven (such as whether it is in a special stage such as "vacuum fine adjustment" or "cavity preheating"). If it is in a special stage, temporarily disable the air leakage judgment (fluctuations in special stages are normal). Restart the judgment after normal low vacuum baking is restored.

[0115] After confirming that a leak has occurred, this embodiment of the invention also provides some examples of subsequent processing:

[0116] (1) Air pump linkage control:

[0117] Once air leakage in the cavity is confirmed, a control command is sent to the air pump control system in real time to adjust the flow distribution strategy.

[0118] a. Flow limit: Reduce the pump flow allocation ratio of the leaking oven from the original 30% (normal allocation ratio) to 5%, only maintain the basic vacuum level, and avoid excessive occupation of pump resources.

[0119] b. Priority adjustment: Increase the flow rate allocation priority of other normal ovens to the highest level, and stabilize the vacuum level within the target range.

[0120] c. Recovery Mechanism: Once the maintenance personnel confirm the air leak repair (by manually entering the "Repair Complete" command in the system), the flow rate will be automatically restored to the normal ratio.

[0121] (2) Alarm:

[0122] a. Alarm Trigger: A red warning light illuminates on the control panel of the leaking oven, alerting on-site operators to take immediate action.

[0123] b. Remote push: Push alarm information to process engineers and equipment maintenance personnel through the MES system. The content includes: the number of the leaking oven, the start time of the leak, the current air pressure characteristics (such as "oscillation frequency 4 times / minute, climb time 45 seconds"), and the numbers of other affected ovens.

[0124] c. Production line display screen notification: Mark the leaking oven with a red box on the production line monitoring screen and display the words "Cavity leak, pump flow has been limited" to remind management personnel to pay attention.

[0125] (3) Tracing the impact of air leakage:

[0126] a. Data recording: Automatically records leak-related data, adds "mean squared deviation of rise time every 3 minutes", and improves feature vector recording;

[0127] b. Moisture prediction: Based on the "leakage duration - cell moisture" correlation model (using historical data), moisture is predicted by the leakage time.

[0128] In summary, the battery low-vacuum baking chamber leakage detection method of the present invention can achieve the following technical effects: (1) Significantly improved latent leakage identification rate: Through 0.2Hz high-frequency sampling and multi-dimensional feature judgment, the latent leakage of "high-frequency oscillation within the qualified range of 60-180Pa" in the low vacuum stage can be accurately identified, the identification rate is increased from 70% of the traditional technology to 99%, and the false detection rate is reduced to less than 1%, avoiding the excessive moisture content of batch cells caused by the failure to detect leakage in the early stage. After the creepage time variance constraint, "leakage oscillation" can be accurately distinguished. ) and "normal process fluctuations" This avoids misjudging normal fluctuations such as "high frequency but unstable rise time" as air leakage, further improving the identification accuracy by 1%-2%.

[0129] (2) The response speed of gas leak alarm has been greatly accelerated: the traditional "alarm after continuous gas pressure exceeds the standard" (later 20-30 minutes) has been optimized to "confirm gas leak and alarm within 12 minutes", the alarm lag time has been shortened by 50%, and maintenance personnel can intervene before the gas leak impact expands.

[0130] (3) Minimize the impact of the air pump system: By limiting the flow rate and adjusting the priority when there is air leakage, the air leakage oven is avoided from occupying pump resources excessively, and the occurrence rate of large-area battery cell quality problems is reduced.

[0131] (4) Significantly improved alarm accuracy: After the addition of the mean square error constraint to the multi-dimensional judgment, the false alarm rate dropped from 3% to below 2%, especially reducing false alarms caused by "high-frequency fluctuations due to equipment fine-tuning" (60% of the original false alarm cases were due to...). (Exceeding the standard was ruled out), further reducing unnecessary downtime.

[0132] (5) Leakage impact traceability and process optimization: Rise time mean square data can help distinguish the type of leakage (fixed seal defect / dynamic seal wear), such as Most leaks are at fixed sealing points. Most of the wear is dynamic seal wear, providing precise direction for oven maintenance, improving the targeted replacement of seals, and reducing maintenance costs by 5%.

[0133] Example 2

[0134] This invention provides a battery low-vacuum baking cavity leakage detection system based on high-frequency pressure oscillation feature recognition, comprising:

[0135] The data processing module is configured to acquire the gas pressure waveform data inside the baking chamber during the low vacuum baking stage and to preprocess the gas pressure waveform data.

[0136] The feature extraction module is configured to extract time-series features from the preprocessed air pressure waveform data. The time-series features include the oscillation frequency per unit time, the average rise time, the root mean square error of the rise time, and the maximum fluctuation amplitude.

[0137] The air leakage detection module is configured to detect air leakage based on timing characteristics. If an air leakage occurs, an alarm is triggered and the timing characteristics and corresponding air pressure waveform data within the corresponding unit time are recorded.

[0138] Specifically, the leak detection module is also configured as follows:

[0139] If a leak is detected in the current unit of time, the leak detection continues for multiple consecutive units of time.

[0140] If multiple consecutive time units are determined to be leaking, the determination result for the current time unit is confirmed; otherwise, the determination result for the current time unit is modified to occasional fluctuation.

[0141] Example 3

[0142] Based on the battery low-vacuum baking cavity leakage detection method provided in Embodiment 1, this embodiment of the invention provides an electronic device, including a processor and a storage medium;

[0143] Storage media are used to store instructions;

[0144] The processor is used to perform operations according to instructions to execute the steps according to the method described above.

[0145] Example 4

[0146] Based on the battery low-vacuum baking cavity leakage detection method provided in Embodiment 1, this embodiment of the invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above method.

[0147] Example 5

[0148] Based on the battery low-vacuum baking cavity leakage detection method provided in Embodiment 1, this embodiment of the invention provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the above method.

[0149] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0150] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0151] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0152] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0153] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for detecting air leakage in a battery low-vacuum baking cavity based on high-frequency oscillation characteristics of air pressure, characterized in that, include: Acquire the gas pressure waveform data inside the baking chamber during the low vacuum baking stage, and preprocess the gas pressure waveform data. The preprocessed air pressure waveform data is subjected to time series feature extraction, which includes oscillation frequency, average rise time, rise time root mean square error, and maximum fluctuation amplitude per unit time. Based on the aforementioned timing characteristics, a leak is determined. If a leak occurs, an alarm is triggered and the timing characteristics and corresponding air pressure waveform data for the corresponding unit time period are recorded.

2. The method for detecting air leakage in a battery low-vacuum baking cavity based on high-frequency oscillation characteristics of air pressure as described in claim 1, characterized in that, The acquisition of gas pressure waveform data within the baking chamber during the low-vacuum baking stage includes: A pressure sensor is installed in the baking cavity near the vacuum interface. During the low-vacuum baking stage, high-frequency sampling is performed by increasing the output rate of the pressure sensor or the sampling frequency of the back-end ADC. When exiting the low vacuum baking stage, a low-frequency standby is performed by reducing the output rate of the pressure sensor or the sampling frequency of the back-end ADC.

3. The method for detecting leakage in a battery low-vacuum baking cavity based on high-frequency oscillation characteristics of gas pressure as described in claim 1, characterized in that, The preprocessing of the air pressure waveform data includes: The air pressure waveform data was screened using the 3σ principle to remove outliers and replace them with the moving average of the first 3 data points; Using the start time of entering the low vacuum baking stage as the time reference, the gas pressure waveform data in each baking chamber is timestamped and aligned.

4. The method for detecting air leakage in a battery low-vacuum baking cavity based on high-frequency oscillation characteristics of air pressure as described in claim 1, characterized in that, The step of extracting time-series features from the preprocessed air pressure waveform data includes: The preprocessed air pressure waveform data is traversed, and all peaks and troughs are identified based on amplitude changes. The number of oscillation cycles is identified by the sequence of peaks and troughs within a unit of time, and the oscillation frequency is determined based on the number of oscillation cycles. The average climb time is obtained by taking the average of the peak and trough times within a unit of time. Calculate the mean squared error of climb time based on the climb time for each climb and the average climb time per unit time. The maximum fluctuation range is calculated based on the maximum and minimum air pressure amplitudes per unit time.

5. The method for detecting leakage in a battery low-vacuum baking cavity based on high-frequency oscillation characteristics of gas pressure as described in claim 4, characterized in that, The method of identifying all peaks and troughs based on amplitude changes includes: If data points Simultaneously satisfy: Then data points At the peak of the wave; If data points Simultaneously satisfy: Then data points Located at a trough; In the formula, The first part of the preprocessed air pressure waveform data is shown below. The magnitude of each data point For the set air pressure threshold, The unit is Pa.

6. The method for detecting air leakage in a battery low-vacuum baking cavity based on high-frequency oscillation characteristics of air pressure as described in claim 1, characterized in that, The step of determining air leakage based on the time sequence characteristics includes: If the aforementioned time-series characteristics simultaneously satisfy: If the amplitude of each data point within a unit of time is within the preset air pressure range, then it is determined that an air leak has occurred. In the formula, The first The oscillation frequency, average rise time, mean squared deviation of rise time, and maximum fluctuation amplitude per unit time; For the corresponding judgment threshold, The unit is Hz. The unit is s. The unit is Pa.

7. The method for detecting leakage in a battery low-vacuum baking cavity based on high-frequency oscillation characteristics of gas pressure as described in claim 1, characterized in that, The method of determining leakage based on the time sequence characteristics also includes: If a leak is detected in the current unit of time, the leak detection continues for multiple consecutive units of time. If multiple consecutive time units are determined to be leaking, the determination result for the current time unit is confirmed; otherwise, the determination result for the current time unit is modified to occasional fluctuation.

8. A battery low-vacuum baking cavity leakage detection system based on high-frequency oscillation feature recognition of gas pressure, characterized in that, include: The data processing module is configured to acquire the gas pressure waveform data inside the baking chamber during the low vacuum baking stage, and to preprocess the gas pressure waveform data. The feature extraction module is configured to extract time-series features from the preprocessed air pressure waveform data. The time-series features include the oscillation frequency per unit time, the average rise time, the root mean square error of the rise time, and the maximum fluctuation amplitude. The air leakage detection module is configured to detect air leakage based on the time sequence characteristics. If an air leakage occurs, an alarm is triggered and the time sequence characteristics and their corresponding air pressure waveform data within the corresponding unit time are recorded.

9. The battery low-vacuum baking cavity leakage detection system based on high-frequency oscillation feature recognition of gas pressure according to claim 8, characterized in that, The leakage detection module is also configured to: If a leak is detected in the current unit of time, the leak detection continues for multiple consecutive units of time. If multiple consecutive time units are determined to be leaking, the determination result for the current time unit is confirmed; otherwise, the determination result for the current time unit is modified to occasional fluctuation.

10. An electronic device, characterized in that, Including processor and storage media; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1-7.

11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1-7.

12. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instructions are executed by the processor, they implement the steps of the method described in any one of claims 1-7.

Images