A method and system for detecting water content in lithium batteries

Abstract

The present invention relates to the technical field of lithium battery water content detection, and in particular to a lithium battery water content detection method and system. The method comprises the following steps: obtaining a thermal imaging image of the lithium battery surface; identifying the temperature gradient distribution of the thermal imaging image of the lithium battery surface, and detecting the sudden change of local distribution changes to extract the temperature imbalance area and the temperature normal area; calculating the best penetrating electromagnetic wave frequency according to the temperature imbalance area and the temperature normal area, and performing a multi-band electromagnetic wave penetration test to collect electromagnetic reflection signals from the temperature imbalance area and the temperature normal area; calculating the reflection signal strength of multiple frequency bands of the electromagnetic reflection signal and mining the comprehensive reflection characteristics of each area to obtain the signal reflection characteristics of the temperature imbalance area and the signal reflection characteristics of the temperature normal area. The present invention realizes efficient and accurate detection of the water content of lithium batteries.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium battery water content detection, and in particular to a lithium battery water content detection method and system. Background Art

[0002] As the core energy device of modern portable electronic devices, electric vehicles and energy storage systems, lithium batteries are widely used in consumer electronics, transportation and new energy fields. Lithium batteries have high energy density, long service life and high charge and discharge efficiency, so they have been widely used in various industries. However, with the extension of lithium battery usage time and changes in the external environment, the moisture content inside the battery has gradually become an important factor affecting battery performance.

[0003] The moisture in lithium batteries mainly comes from the introduction of moisture during the battery manufacturing process or changes in humidity in the external environment. The presence of moisture will significantly affect the electrochemical reaction process of the battery, and thus affect the battery's capacity, charge and discharge efficiency, cycle life and safety. Excessive moisture content will not only cause chemical changes in the electrolyte inside the battery, but may also cause battery expansion, leakage, short circuit, and even thermal runaway and other safety issues. Therefore, accurate detection of the moisture content in lithium batteries is of great significance for battery performance evaluation, quality control and safety monitoring.

[0004] Traditional methods for detecting water content in lithium batteries mostly rely on direct sampling and chemical reagent analysis. Common techniques include gravimetric method, chemical titration method and ion chromatography. These methods usually require disassembly or special treatment of the battery, and have problems such as high operational complexity, long time consumption and limited detection accuracy. With the increasing requirements for lithium battery health status management, traditional water content detection methods can no longer meet the needs of real-time, non-destructive and accurate detection. Therefore, in order to effectively solve the limitations in lithium battery water content detection, a new intelligent detection method is urgently needed. Summary of the invention

[0005] In order to solve the above technical problems, the present invention proposes a method and system for detecting water content in a lithium battery to solve at least one of the above technical problems.

[0006] To achieve the above object, the present invention provides a method for detecting water content of a lithium battery, comprising the following steps:

[0007] Step S1: Acquire a thermal imaging image of the surface of a lithium battery; identify the temperature gradient distribution of the thermal imaging image of the surface of the lithium battery, and perform a mutation detection of local distribution changes to extract a temperature imbalance area and a temperature normal area;

[0008] Step S2: Calculate the optimal penetrating electromagnetic wave frequency according to the temperature imbalance area and the temperature normal area, and perform a multi-band electromagnetic wave penetration test to collect electromagnetic reflection signals from the temperature imbalance area and the temperature normal area;

[0009] Step S3: calculating the reflection signal strength of multiple frequency bands of the electromagnetic reflection signal and mining the comprehensive reflection characteristics of each region to obtain the signal reflection characteristics of the temperature imbalance region and the signal reflection characteristics of the temperature normal region;

[0010] Step S4: identifying the difference of the signal reflection characteristics of the temperature imbalance area and the signal reflection characteristics of the temperature normal area in each frequency band, and mining the regional feature deviation to obtain the reflection deviation characteristics of the temperature imbalance area;

[0011] Step S5: performing nonlinear coupling analysis of electromagnetic parameters according to the reflection deviation characteristics of the temperature imbalance area, and performing correlation analysis of the dynamic change of water content, so as to construct a dynamic electromagnetic water content interference model;

[0012] Step S6: reverse the water content in the temperature imbalance area using the dynamic electromagnetic water content interference model, and conduct a comprehensive evaluation to generate a lithium battery water content detection report.

[0013] The present invention can quickly identify abnormal temperature areas on the battery surface through thermal imaging images. These areas are often related to problems such as abnormal moisture content and heat accumulation. When there is moisture in a certain area inside the battery, the heat generated in this area during operation will be inconsistent with the heat in other parts, resulting in uneven temperature gradients. Therefore, detecting the temperature distribution can quickly locate potential water-containing parts, and there is no need to contact the battery or disassemble the battery structure to avoid physical damage to the battery, while providing real-time monitoring capabilities. The optimal frequency range is calculated through temperature information to improve the penetration ability of electromagnetic waves and ensure that the impact of internal moisture content on electromagnetic signals can be effectively detected. Multi-band electromagnetic wave testing can capture electromagnetic responses at different levels of the battery, thereby improving the accuracy of moisture detection. By finely calculating the signal strength of multiple frequency bands, the detailed characteristics of electromagnetic reflection in each area can be revealed, providing an accurate signal basis for subsequent analysis. The reflection characteristics of temperature imbalance areas and conventional areas can be effectively distinguished, thereby providing clearer clues for moisture analysis. Through difference analysis, the changes in reflection characteristics of temperature imbalance areas and conventional areas can be accurately distinguished. These difference characteristics are usually closely related to moisture content, temperature fluctuations and changes in the internal structure of the battery. The deviation characteristics of the temperature imbalance area provide an important basis for further analysis of the moisture content and can identify possible moisture anomalies in advance. The electromagnetic penetration strength of the area with moisture is different from that of the area without moisture, which leads to inconsistent characteristics of the reflected electromagnetic signal. By performing electromagnetic parameter nonlinear coupling analysis on the reflection deviation characteristics, a relationship model between the temperature imbalance area and the moisture content can be established. The dynamic change of moisture has a strong nonlinear effect on the reflection characteristics of electromagnetic waves. Therefore, by establishing a dynamic electromagnetic moisture content interference model, the impact of moisture changes on electromagnetic wave signals can be more accurately quantified. It can handle the complex nonlinear relationship between electromagnetic wave propagation and moisture changes, thereby improving the accuracy of moisture content detection. Based on the established electromagnetic interference model and in-depth analysis, accurate inference of the moisture content in the temperature imbalance area can be achieved. By combining the intensity, phase change and frequency band difference of the electromagnetic reflection signal, the moisture content data of each area is generated. Finally, a comprehensive lithium battery moisture content detection report is generated through a comprehensive evaluation of the moisture content of all areas. The report can not only accurately reflect the moisture status of each area of ​​the battery, but also provide a basis for the health status and performance optimization of the battery.

[0014] In this specification, a lithium battery water content detection system is provided, which is used to execute the lithium battery water content detection method as described above, including:

[0015] The temperature mutation detection module is used to obtain the thermal imaging image of the lithium battery surface; identify the temperature gradient distribution of the thermal imaging image of the lithium battery surface, and perform mutation detection of local distribution changes to extract the temperature imbalance area and the temperature normal area;

[0016] The optimal frequency module is used to calculate the optimal penetrating electromagnetic wave frequency according to the temperature imbalance area and the temperature normal area, and to perform multi-band electromagnetic wave penetration test to collect electromagnetic reflection signals from the temperature imbalance area and the temperature normal area;

[0017] A reflection feature mining module is used to calculate the reflection signal strength of multiple frequency bands of the electromagnetic reflection signal and mine the comprehensive reflection features of each region to obtain the signal reflection features of the temperature imbalance region and the signal reflection features of the temperature normal region;

[0018] The reflection deviation module is used to identify the difference of the signal reflection characteristics of the temperature imbalance area and the signal reflection characteristics of the temperature normal area in each frequency band, and to mine the regional feature deviation to obtain the reflection deviation characteristics of the temperature imbalance area;

[0019] The water content interference module is used to perform nonlinear coupling analysis of electromagnetic parameters based on the reflection deviation characteristics of the temperature imbalance area, and to perform correlation analysis of the dynamic change of water content, so as to construct a dynamic electromagnetic water content interference model;

[0020] The moisture content inversion module is used to infer the moisture content in the temperature imbalance area of ​​the dynamic electromagnetic moisture content interference model and conduct a comprehensive evaluation to generate a lithium battery moisture content detection report.

[0021] The present invention can efficiently identify the abnormal change area of ​​battery surface temperature through thermal imaging images. The temperature imbalance area is often closely related to the uneven distribution of moisture inside the battery or the heat accumulation inside the battery. Thermal imaging is a non-contact detection method, which avoids the damage that may be caused by direct contact with the battery surface, and is helpful for long-term safety monitoring of the battery. The temperature mutation detection can identify the potential problems of the battery in real time, take protective measures in time, and improve the battery safety. By analyzing the characteristics of the temperature imbalance area and the conventional area, the optimal penetration frequency range is calculated to maximize the penetration ability of electromagnetic waves. The application of multi-band electromagnetic waves can comprehensively and detailedly capture the reflection signals of different areas on the battery surface and inside, ensuring the accuracy of the moisture detection results. By analyzing the intensity of multiple frequency band signals, more abundant electromagnetic reflection features can be extracted to provide strong support for subsequent signal analysis, and the characteristic differences of different areas (temperature imbalance area and conventional area) can be distinguished, making the subsequent moisture content analysis more accurate. By comprehensively mining the signal strength and characteristics, the quality of data acquisition can be effectively improved, thereby improving the reliability of the entire detection system, and effectively identifying the temperature imbalance area and the conventional area. The signal deviation in the conventional area can accurately distinguish the reflection feature changes caused by moisture or other factors. By mining the reflection deviation features, the abnormal area inside the battery can be quickly identified to help locate the possible moisture accumulation area. Through the detailed analysis of the reflection deviation, the tiny electromagnetic changes can be captured, thereby improving the sensitivity and accuracy of the detection system. Through nonlinear coupling analysis, the complex nonlinear relationship between the moisture content change and the electromagnetic wave reflection signal can be revealed. This modeling method can more accurately capture the subtle effects of moisture changes. The model can not only describe the effect of moisture content on electromagnetic wave propagation, but also reflect the dynamic process of moisture changes in real time, providing an accurate basis for subsequent analysis. Through the dynamic coupling of electromagnetic parameters and moisture content, accurate moisture content detection data can be provided for each area to reduce errors. Through the reverse algorithm, the moisture content of each area inside the battery can be accurately obtained, avoiding the sampling error in the traditional method. It comprehensively considers a variety of electromagnetic parameters (such as reflection intensity, phase change, etc.), making the moisture content detection results more comprehensive and accurate, and can automatically generate a detailed moisture content detection report, providing direct support for battery health management and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A schematic diagram of the steps of a method for detecting water content in a lithium battery according to the present invention;

[0023] Figure 2 Detailed implementation flow chart of step S1;

[0024] Figure 3 Detailed implementation flow chart of step S2;

[0025] Figure 4 Detailed implementation flow chart of step S3. DETAILED DESCRIPTION

[0026] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0027] This application example provides a lithium battery water content detection method and system. The execution subject of the lithium battery water content detection method and system includes but is not limited to: mechanical equipment, data processing platform, cloud server node, network upload device, etc. equipped with the system can be regarded as the general computing node of this application, and the data processing platform includes but is not limited to: at least one of an audio image management system, an information management system, and a cloud data management system.

[0028] See also Figures 1 to 4 The present invention provides a method for detecting the water content of a lithium battery, the method comprising the following steps:

[0029] Step S1: Acquire a thermal imaging image of the surface of a lithium battery; identify the temperature gradient distribution of the thermal imaging image of the surface of the lithium battery, and perform a mutation detection of local distribution changes to extract a temperature imbalance area and a temperature normal area;

[0030] Step S2: Calculate the optimal penetrating electromagnetic wave frequency according to the temperature imbalance area and the temperature normal area, and perform a multi-band electromagnetic wave penetration test to collect electromagnetic reflection signals from the temperature imbalance area and the temperature normal area;

[0031] Step S3: calculating the reflection signal strength of multiple frequency bands of the electromagnetic reflection signal and mining the comprehensive reflection characteristics of each region to obtain the signal reflection characteristics of the temperature imbalance region and the signal reflection characteristics of the temperature normal region;

[0032] Step S4: identifying the difference of the signal reflection characteristics of the temperature imbalance area and the signal reflection characteristics of the temperature normal area in each frequency band, and mining the regional feature deviation to obtain the reflection deviation characteristics of the temperature imbalance area;

[0033] Step S5: performing nonlinear coupling analysis of electromagnetic parameters according to the reflection deviation characteristics of the temperature imbalance area, and performing correlation analysis of the dynamic change of water content, so as to construct a dynamic electromagnetic water content interference model;

[0034] Step S6: reverse the water content in the temperature imbalance area using the dynamic electromagnetic water content interference model, and conduct a comprehensive evaluation to generate a lithium battery water content detection report.

[0035] The present invention can quickly identify abnormal temperature areas on the battery surface through thermal imaging images. These areas are often related to problems such as abnormal moisture content and heat accumulation. When there is moisture in a certain area inside the battery, the heat generated in this area during operation will be inconsistent with the heat in other parts, resulting in uneven temperature gradients. Therefore, detecting the temperature distribution can quickly locate potential water-containing parts, and there is no need to contact the battery or disassemble the battery structure to avoid physical damage to the battery, while providing real-time monitoring capabilities. The optimal frequency range is calculated through temperature information to improve the penetration ability of electromagnetic waves and ensure that the impact of internal moisture content on electromagnetic signals can be effectively detected. Multi-band electromagnetic wave testing can capture electromagnetic responses at different levels of the battery, thereby improving the accuracy of moisture detection. By finely calculating the signal strength of multiple frequency bands, the detailed characteristics of electromagnetic reflection in each area can be revealed, providing an accurate signal basis for subsequent analysis. The reflection characteristics of temperature imbalance areas and conventional areas can be effectively distinguished, thereby providing clearer clues for moisture analysis. Through difference analysis, the changes in reflection characteristics of temperature imbalance areas and conventional areas can be accurately distinguished. These difference characteristics are usually closely related to moisture content, temperature fluctuations and changes in the internal structure of the battery. The deviation characteristics of the temperature imbalance area provide an important basis for further analysis of the moisture content and can identify possible moisture anomalies in advance. The electromagnetic penetration strength of the area with moisture is different from that of the area without moisture, which leads to inconsistent characteristics of the reflected electromagnetic signal. By performing electromagnetic parameter nonlinear coupling analysis on the reflection deviation characteristics, a relationship model between the temperature imbalance area and the moisture content can be established. The dynamic change of moisture has a strong nonlinear effect on the reflection characteristics of electromagnetic waves. Therefore, by establishing a dynamic electromagnetic moisture content interference model, the impact of moisture changes on electromagnetic wave signals can be more accurately quantified. It can handle the complex nonlinear relationship between electromagnetic wave propagation and moisture changes, thereby improving the accuracy of moisture content detection. Based on the established electromagnetic interference model and in-depth analysis, accurate inference of the moisture content in the temperature imbalance area can be achieved. By combining the intensity, phase change and frequency band difference of the electromagnetic reflection signal, the moisture content data of each area is generated. Finally, a comprehensive lithium battery moisture content detection report is generated through a comprehensive evaluation of the moisture content of all areas. The report can not only accurately reflect the moisture status of each area of ​​the battery, but also provide a basis for the health status and performance optimization of the battery.

[0036] In the embodiment of the present invention, refer to Figure 1 , is a schematic flow chart of the steps of a method for detecting water content in a lithium battery of the present invention. In this example, the steps of the method for detecting water content in a lithium battery include:

[0037] Step S1: Acquire a thermal imaging image of the surface of a lithium battery; identify the temperature gradient distribution of the thermal imaging image of the surface of the lithium battery, and perform a mutation detection of local distribution changes to extract a temperature imbalance area and a temperature normal area;

[0038] In this embodiment, a suitable thermal imaging camera is selected, such as FLIR, Seek Thermal, etc. These devices should have high resolution (such as 640x480 pixels) and a good temperature measurement range (-20°C to +120°C) to ensure that the thermal distribution on the surface of the lithium battery can be accurately captured. Before conducting the experiment, ensure that the thermal imaging camera is configured correctly, including focal length, exposure time, and gain settings. In order to obtain the best image quality, the equipment can be calibrated in a laboratory environment to ensure that the temperature accuracy and sensitivity of the camera meet the standards. Before acquiring thermal imaging images, ensure that the lithium battery is in proper working condition. The temperature changes can be monitored during the charging and discharging process to increase the changes in the battery surface temperature, thereby improving the image contrast. Use a thermal imaging camera to shoot the lithium battery, ensuring that the camera is kept at a certain distance from the battery surface to avoid heat interference. Different shooting angles can be selected to obtain more comprehensive thermal distribution information. The ambient temperature, humidity, and working status of the battery should be recorded for each shot for subsequent analysis. Use image processing software (such as MATLAB, OpenCV) to preprocess the acquired thermal imaging images, including denoising, contrast enhancement, and color mapping. Denoising can use techniques such as median filtering or Gaussian filtering to enhance the temperature contrast in the image so that the temperature gradient can be more clearly identified. Based on the camera's calibration data, the temperature value of each pixel in the image is corrected to ensure accurate temperature distribution information. A known standard temperature source can be used for comparison to ensure that the temperature value of each pixel is consistent with the actual situation. By calculating the temperature gradient of each pixel in the thermal imaging image, areas with large temperature changes can be identified. A threshold is set to determine the significance of the temperature gradient. By binarizing the gradient amplitude map, areas with significant temperature changes are extracted. These areas may indicate thermal imbalance on the surface of the lithium battery. Using change detection algorithms (such as Canny edge detection or clustering-based methods), analyze the temperature gradient image and find mutation areas. The Canny edge detection algorithm can effectively identify mutation edges in the image and help determine the temperature imbalance area. Use the Canny algorithm for edge detection, set appropriate high and low thresholds to extract the mutation edges of temperature changes, and divide the areas in the thermal imaging image into temperature imbalance areas and regular areas based on the results of mutation detection. Region growing algorithms or K-means clustering algorithms can be used to cluster areas with similar temperatures.

[0039] Step S2: Calculate the optimal penetrating electromagnetic wave frequency according to the temperature imbalance area and the temperature normal area, and perform a multi-band electromagnetic wave penetration test to collect electromagnetic reflection signals from the temperature imbalance area and the temperature normal area;

[0040] In this embodiment, when selecting the optimal penetration frequency, the interaction characteristics between electromagnetic waves and lithium battery materials, including the influence of dielectric constant, conductivity and temperature on these parameters, need to be considered. The choice of frequency should be based on the electromagnetic properties of lithium battery materials, especially the differences between the temperature imbalance area and the conventional area. Select a frequency range (e.g., 300 MHz to 30 GHz) and analyze the penetration depth of electromagnetic waves at different frequencies through theoretical models or numerical simulations (e.g., finite element method). According to the calculation results, select a frequency that can obtain the best signal reflection in the temperature imbalance area and the conventional area, such as selecting a specific frequency (e.g., 2.4 GHz or 5 GHz) for subsequent experimental testing. Equipped with a suitable signal generator and receiver to generate and receive electromagnetic wave signals of different frequency bands. Ensure that the equipment can cover the selected frequency range and has sufficient sensitivity and resolution. Test in a controlled environment to reduce external interference. You can choose to perform it in a shielded room or laboratory, place the lithium battery in a fixed position, and ensure that the incident angle of the electromagnetic wave is consistent with the environmental conditions. According to the preset frequency range, electromagnetic waves of different frequency bands are gradually emitted. When testing each frequency band, record the signal transmission power and the response signal of the receiver. The transmission time and scanning time of each frequency band can be set to ensure the continuity of data acquisition. Use the receiver to collect the reflected signal in real time and record the electromagnetic reflection intensity of each frequency band. Ensure that the sampling frequency and time interval of the data acquisition system are accurate enough to capture small changes in the reflected signal. Post-process the collected reflected signal data, including denoising, smoothing and normalization. Use data analysis software (such as MATLAB or Python) to perform spectral analysis on the signal to extract the amplitude and phase information of the reflected signal. Analyze the difference in reflected signals between the temperature imbalance area and the conventional area to determine the change in reflection intensity under different frequency bands. Compare the reflected signals of different frequency bands to find the frequency band that can best distinguish the temperature imbalance area. Organize the experimental results into a report, including comparative analysis charts of reflected signal intensity, temperature imbalance area and conventional area under different frequency bands. Record the best penetration frequency and its corresponding reflected signal characteristics to provide a basis for subsequent analysis.

[0041] Step S3: calculating the reflection signal strength of multiple frequency bands of the electromagnetic reflection signal and mining the comprehensive reflection characteristics of each region to obtain the signal reflection characteristics of the temperature imbalance region and the signal reflection characteristics of the temperature normal region;

[0042] In this embodiment, the electromagnetic reflection signal data collected from the multi-band test is organized into a structured format. The data of each frequency band should include the reflection signal strength, phase information, temperature information, and the corresponding area identification (such as the temperature imbalance area and the temperature normal area). Clean the data, remove outliers and noise, and ensure the accuracy of the data. For example, use statistical methods to detect and remove reflection signal values ​​that exceed the normal range to improve the reliability of subsequent analysis. Calculate the reflection signal strength of different frequency bands, select multiple frequency bands (such as 2.4 GHz, 5 GHz, 10 GHz, etc.), and calculate the reflection signal strength of each frequency band one by one. Ensure that the signal strength data of each frequency band is processed in the same way for subsequent comparison. For each area (temperature imbalance area and temperature normal area), extract key signal features, including average reflection intensity, peak reflection intensity, reflection intensity standard deviation, and phase change. These features help to identify and distinguish the electromagnetic characteristics of different areas. Group the signal data of each area and calculate the relevant features. Calculate the phase change of the signal and analyze the phase characteristics of different areas. The differences in phase changes may be related to temperature changes and changes in material properties, which can be evaluated by calculating the average phase value and phase standard deviation for each region.

[0043] Step S4: identifying the difference of the signal reflection characteristics of the temperature imbalance area and the signal reflection characteristics of the temperature normal area in each frequency band, and mining the regional feature deviation to obtain the reflection deviation characteristics of the temperature imbalance area;

[0044] In this embodiment, the signal reflection feature data of the temperature imbalance area and the temperature normal area are sorted. The reflection signal characteristics of each frequency band should include key information such as reflection intensity, phase information, frequency, etc. to ensure the integrity and consistency of the data. The collected reflection features are standardized to eliminate the dimensional differences between different frequency bands. This can be calculated by subtracting the mean value of each feature value from the feature and dividing it by the standard deviation, and using statistical analysis methods (such as t-test or variance analysis) to identify the reflection feature differences between different regions. By comparing the reflection intensity and phase of each frequency band, it is determined whether there is a significant difference. For each frequency band, the average reflection intensity and phase values ​​of the temperature imbalance area and the temperature normal area are calculated. The p value of each frequency band is recorded to determine whether there is a significant difference. Generally, a p value less than 0.05 is considered to be statistically significant. The results are organized into a table to clearly show the comparison results of different frequency bands. The reflection deviation feature is mined by calculating the reflection feature deviation between the temperature imbalance area and the temperature normal area. For each frequency band, the deviation of the reflection intensity and phase between the temperature imbalance area and the normal area is calculated. Record the deviation value for each frequency band and analyze its changing trend. For example, if the reflection intensity is significantly reduced in a certain frequency band, it may indicate that this frequency band is particularly sensitive to the temperature imbalance area.

[0045] Step S5: performing nonlinear coupling analysis of electromagnetic parameters according to the reflection deviation characteristics of the temperature imbalance area, and performing correlation analysis of the dynamic change of water content, so as to construct a dynamic electromagnetic water content interference model;

[0046] In this embodiment, in the interaction between electromagnetic waves and materials, especially in the temperature imbalance area of ​​lithium batteries, electromagnetic parameters (such as dielectric constant, conductivity, etc.) may show a nonlinear relationship. This nonlinear characteristic may be related to the physical state of the material (such as temperature, humidity and water content). Therefore, establishing a nonlinear coupling model of electromagnetic parameters is the key to understanding its dynamic changes. Select relevant electromagnetic parameters for analysis, including reflection coefficient, dielectric constant and conductivity. The variation data of these parameters under different temperature and water content conditions can be obtained through experimental measurement or literature review. The experimental parameters may include reflection intensity and temperature variation range (e.g., 20°C to 60°C) at different frequencies. Using nonlinear regression analysis methods, a relationship model between electromagnetic parameters and temperature and water content is constructed. Machine learning methods such as polynomial regression or support vector machine (SVM) can be used for model training and parameter optimization. By minimizing the error, the best fitting model is found, and the water content data of lithium batteries under different working conditions are collected, which can be obtained by chemical analysis or electromagnetic wave measurement technology. The water content data is matched with the reflection deviation characteristics to ensure the timeliness and consistency of the data. Use time series analysis or correlation analysis to explore the dynamic relationship between reflection deviation characteristics and water content. The Pearson correlation coefficient or Spearman rank correlation coefficient can be used to quantify the degree of association between them. Based on the analysis results, a dynamic electromagnetic water content interference model is constructed. The model should be able to predict the water content changes of lithium batteries in real time, taking into account the influence of temperature and reflection signals. This can be achieved using a linear regression model or a neural network model, and the model is trained to make accurate water content predictions under different conditions. Use an independent experimental data set to verify the validity and accuracy of the constructed model. By comparing the difference between the water content predicted by the model and the actual measured value, the accuracy and error indicators of the model (such as root mean square error RMSE) are calculated. According to the verification results, the model is adjusted and optimized. The prediction ability of the model can be improved by adding more input features, choosing a different model architecture, or adjusting the model parameters.

[0047] Step S6: reverse the water content in the temperature imbalance area using the dynamic electromagnetic water content interference model, and conduct a comprehensive evaluation to generate a lithium battery water content detection report.

[0048] In this embodiment, first, the dynamic electromagnetic moisture content interference model constructed in step S5 is selected. Ensure that the model is fully verified and has good predictive ability. Prepare relevant input data, including temperature, reflection signal strength and other electromagnetic parameters, which should be concentrated in the temperature imbalance area. Collect the latest lithium battery test data, with special attention to the electromagnetic reflection signal in the temperature imbalance area. These data should include reflection intensity and phase information under different temperature and humidity conditions to ensure their timeliness and accuracy. For example, the reflection signals at different temperatures such as 25°C, 35°C, and 45°C are recorded to ensure that the typical operating range of lithium batteries is covered. Use the dynamic electromagnetic moisture content interference model to reverse the moisture content. Input the collected reflection signal and temperature data, and calculate the corresponding moisture content value through the model. Perform statistical analysis on the moisture content obtained by reverse calculation, and calculate its key indicators such as mean and standard deviation. Evaluate the change trend of the current moisture content by comparing with historical data. For example, the change of moisture content reversed under different temperature conditions can be calculated to ensure the statistical significance of the data. Set appropriate evaluation criteria to determine whether the water content of lithium batteries is within a safe range. You can refer to industry standards or previous research results, such as setting water content thresholds. There are usually clear guidance in the literature on lithium battery design and use. Analyze and reverse the relationship between water content and battery performance (such as charge and discharge efficiency, cycle life, etc.). The degree of correlation between the two can be quantified by calculating the correlation coefficient. This analysis can reveal the impact of changes in water content on lithium battery performance and provide a basis for subsequent optimization.

[0049] In this embodiment, refer to Figure 2 , is a flowchart of detailed implementation steps of step S1. In this embodiment, the detailed implementation steps of step S1 include:

[0050] Step S11: using a thermal imager to perform a comprehensive scan and temperature measurement on the lithium battery to be inspected, so as to generate a thermal imaging image of the lithium battery surface;

[0051] Step S12: performing adaptive noise filtering on the thermal imaging image of the lithium battery surface to construct an adaptive filtered thermal imaging image;

[0052] Step S13: performing temperature gradient distribution recognition on the adaptively filtered thermal imaging image to generate image temperature gradient distribution data;

[0053] Step S14: performing gradient distribution change analysis on the image temperature gradient distribution data to obtain temperature gradient distribution change characteristics;

[0054] Step S15: Based on the temperature gradient distribution change characteristics, the adaptive filtered thermal imaging image is subjected to local distribution change mutation detection to extract the temperature imbalance area and the temperature normal area.

[0055] In this embodiment, a high-performance thermal imager is selected, such as FLIR T640 or Seek ThermalCompactPRO. These devices should have a high resolution of 640x480 pixels, can provide clear thermal imaging images, and have a wide temperature measurement range of -20°C to +400°C to ensure that the temperature changes of the lithium battery can be captured. Before measuring, ensure that the thermal imager is calibrated to ensure the measurement accuracy of the equipment. In terms of environmental conditions, choose an environment with stable room temperature to avoid the influence of strong light, strong wind, humidity changes, etc. on the measurement results. Place the lithium battery to be tested on a sturdy and shockproof workbench to ensure that the battery does not move during the scanning process. Ensure that the lithium battery is in normal working condition (for example, charging or discharging) to obtain the true temperature distribution. Set the measurement parameters of the thermal imager, including measurement mode (such as focus mode, automatic adjustment mode), temperature range (according to the operating temperature of the lithium battery), and image acquisition frequency (usually 30 frames per second are selected to capture rapidly changing temperatures). Use a thermal imager to fully scan the surface of the lithium battery to ensure that every part of the battery is covered. Manual scanning or automated scanning with a fixed thermal imager can be used to ensure the integrity and continuity of the image and avoid missing any hot spots. Store the acquired thermal image data in a computer in an appropriate file format (such as TIFF or JPEG) for subsequent processing and analysis. Each image should contain relevant temperature data and shooting parameters, such as time, ambient temperature, measurement conditions, etc. Select a suitable adaptive filtering algorithm, such as adaptive median filtering or bilateral filtering, which can effectively remove random noise from thermal images while retaining important edge information. Adaptive median filtering is particularly suitable for processing thermal images because it can dynamically adjust the filtering parameters based on local statistical characteristics. Implement the adaptive filtering algorithm in image processing software (such as MATLAB, OpenCV or Python's PIL library).

[0056] Load the thermal image into the program and process each pixel. For each pixel, calculate the median of its neighboring pixels and adaptively adjust the size of the filter window according to the noise level. In specific implementation, the window size (for example, 3x3 or 5x5) can be set and dynamically adjusted according to the local pixel value of the image to ensure that a larger window is used in areas with more noise and a smaller window is used in areas with less noise. By comparing the images before and after filtering, observe the reduction of noise. Indicators such as signal-to-noise ratio (SNR) or root mean square error (RMSE) can be calculated to verify the degree of improvement in the filtering effect. Ideally, the filtered image should show clearer temperature features and reduce interfering noise. Use image processing software to calculate the temperature gradient of the adaptively filtered thermal image. The Sobel operator or Prewitt operator can be used to calculate the gradient of the image and identify the area of ​​temperature change. In specific implementation, the filtered image is converted to a grayscale image and a gradient operator is applied to obtain the gradient value of each pixel. These values ​​can reflect the intensity and direction of temperature change. Calculate the gradient in the horizontal and vertical directions. Visualize the calculated gradient data to generate a gradient distribution map. Different gradient values ​​can be represented by different colors through pseudo-color mapping, clearly showing the temperature distribution characteristics of the lithium battery surface. Record the temperature gradient value of each area, analyze its change trend, and compare it with the temperature under normal working conditions to identify possible abnormal areas.

[0057] Select appropriate statistical analysis methods (such as mean, variance analysis) and signal processing methods (such as wavelet transform) to analyze the changes in temperature gradient distribution data. These methods can help identify the changing characteristics and trends of temperature gradients. Calculate the mean and standard deviation of the temperature gradient to analyze the fluctuations of temperature gradients in different regions. You can set a threshold to identify areas where the temperature gradient changes significantly. In specific analysis, you can use statistical software (such as R or Python's SciPy library) for data processing and calculate the p value to determine the significant change in the temperature gradient. Extract the time and frequency domain characteristics of the temperature gradient through methods such as wavelet transform to help identify potential temperature anomalies. Using wavelet transform, you can obtain the performance of the temperature gradient at different frequencies, thereby identifying periodic changes and mutation points. Select a suitable mutation detection algorithm, such as the Cumulative Sum (CUSUM) method or the Change Point Detection algorithm, which can effectively identify temperature imbalance areas and regular areas. For the CUSUM method, you can quickly locate the area of ​​temperature anomaly by monitoring the accumulation and change of the temperature gradient. Input the temperature gradient change data into the mutation detection algorithm to identify the mutation point of the temperature distribution. According to the set threshold, mark the areas within the normal threshold range and those outside the range. The specific implementation of the algorithm can use Python's statsmodels library or custom functions to quickly process data. According to the test results, mark the temperature imbalance area (such as where the temperature rises abnormally) and the temperature regular area (such as where the temperature is stable). Record these results in the database for subsequent analysis and processing. Visualize the mutation detection results and mark the temperature imbalance area and the regular area on the thermal imaging image so that they can be clearly displayed to relevant personnel. Different colors can be used to identify abnormal areas to ensure that the information is intuitive and easy to understand. Generate a detailed analysis report, including the characteristics of the temperature imbalance area, possible causes and recommended measures, to provide a reference for subsequent lithium battery safety testing and management.

[0058] In this embodiment, the specific steps of step S12 are:

[0059] Obtain the humidity parameters of the environment to be detected;

[0060] Performing a dark channel priori calculation on the thermal imaging image of the lithium battery surface according to the humidity parameter of the environment to be detected to obtain an atmospheric shielding factor;

[0061] The transmission map value of each pixel of the thermal imaging image of the lithium battery surface is estimated, thereby generating the transmission map value of each pixel;

[0062] Perform thermal imaging defogging correction according to the atmospheric shielding factor and the transmission map value of each pixel to construct a defogging corrected thermal imaging image;

[0063] Perform image noise detection on the dehazed and corrected thermal imaging image and mark all image noise points;

[0064] Calculating the grayscale distribution of all the image noise points to generate a grayscale distribution value for each noise point;

[0065] Performing geometric feature analysis on all the image noise points one by one to extract the geometric features of each noise point;

[0066] The noise point depth semantic classification is performed based on the grayscale distribution value of each noise point and the geometric features of each noise point, so as to obtain the type characteristics of each noise point;

[0067] Based on the type characteristics of each noise point, the dehazed and corrected thermal imaging image is adaptively filtered to construct an adaptive filtered thermal imaging image.

[0068] In this embodiment, a suitable humidity sensor is selected, such as a digital thermometer and hygrometer, which can provide accurate relative humidity values. Make sure that the sensor is in good working condition and has been calibrated. Before performing thermal imaging detection of lithium batteries, record the relative humidity parameters of the environment to be detected, usually expressed as a percentage (for example, 40% relative humidity). Make sure that the sensor is away from heat sources and fans during measurement to avoid affecting the readings. The recorded humidity parameters are input into the data processing system as reference data for subsequent image processing. This step is the key to ensuring the accuracy of subsequent defogging processing. Dark channel prior is a commonly used defogging technique based on the statistical characteristics of natural images. It is assumed that in the absence of fog, the color channel values ​​of certain areas will be very low. Calculating the dark channel can help estimate the atmospheric illumination and transmittance in the image. Select the thermal imaging image after adaptive filtering and convert it to RGB format for subsequent processing. For each pixel, calculate its dark channel value. The dark channel value is usually calculated by taking the minimum value of the neighboring pixels. ,in, is the neighborhood of pixel X, and estimates the atmospheric shielding factor based on the ambient humidity parameter. The transmission map value reflects the degree of light propagation in the atmosphere and can be calculated by the relationship between the dark channel prior and the atmospheric illumination. t(x)=1−ω⋅ (x) / A, where ω is a constant, usually set to 0.95, indicating the amount of transmission map retention. For each pixel of the filtered thermal imaging image, the transmission map value is calculated using the above formula. The transmission map value is recorded in a new image matrix to generate the transmission map value of each pixel. The thermal imaging image is dehazed based on the atmospheric shielding factor and the transmission map value of each pixel. The correction formula is as follows: , where J(x) is the corrected image, I(x) is the original image, and A is the atmospheric illumination. Implement the defogging correction algorithm in the image processing software, apply the above formula pixel by pixel, and obtain the defogging corrected thermal imaging image. Record the corrected image data and compare it with the original image to observe the defogging effect. Use histogram analysis to confirm whether the brightness and contrast of the corrected image have improved. Use image processing algorithms (such as Otsu algorithm or adaptive threshold algorithm) to detect noise in the defogging corrected image. Select an appropriate threshold to distinguish between noise and normal signals. Convert the defogging corrected image to a grayscale image, apply the noise detection algorithm, and generate a binary image, where the noise area is represented by white and other areas are represented by black. Mark all image noise locations and record the coordinates and size of the noise for subsequent analysis. Extract the grayscale value of each noise area one by one and calculate its grayscale distribution. The grayscale distribution histogram can be generated by counting the grayscale values ​​of pixels in the noise area. For each noise area, calculate its statistical characteristics such as average grayscale value and standard deviation, and record these values. Python's NumPy library can be used for fast calculations. Collect the grayscale distribution values ​​of all noise points and form a data set for subsequent analysis and comparison. Perform geometric feature analysis on each noise point, including the area, perimeter, shape factor, etc. of the noise point. Use image processing libraries (such as OpenCV) to extract the boundaries of each noise point and calculate its geometric features. Record the geometric features of each noise point, including position, size, and shape. Store these feature data in a database for subsequent analysis and classification. Select a suitable deep learning model (such as a convolutional neural network CNN) for semantic classification of noise points. First, prepare a training data set, including the grayscale distribution values ​​and geometric features of the noise points as input, and the noise type as output. Train the deep learning model, adjust hyperparameters (such as learning rate, batch size, etc.), and evaluate the model performance through cross-validation. After training, use the test set to evaluate the accuracy of the model. Use the trained model to classify each noise point and identify the type characteristics of the noise point, such as "hot noise", "cold noise", or "other interference". According to the type characteristics of each noise point, select the appropriate adaptive filtering algorithm (such as adaptive mean filtering or adaptive Gaussian filtering) to effectively remove different types of noise points. Input the dehazed and corrected thermal imaging image into the adaptive filtering algorithm to process different types of noise points respectively. For hot noise points, a stronger filtering factor can be used, while for cold noise points, a weaker filtering factor is used. Apply the filtering algorithm pixel by pixel to generate an adaptive filtered thermal imaging image. Ensure that while removing noise points, the details and temperature characteristics of the image are preserved as much as possible.

[0069] In this embodiment, refer to Figure 3 , is a flowchart of detailed implementation steps of step S2. In this embodiment, the detailed implementation steps of step S2 include:

[0070] Step S21: extracting lithium battery shell material parameters based on lithium battery production log;

[0071] Step S22: Calculate the relative dielectric constant and conductivity according to the material parameters of the lithium battery shell to generate the relative dielectric constant and conductivity parameters of the lithium battery;

[0072] Step S23: mining the electromagnetic characteristics of the shell based on the relative dielectric constant and conductivity parameters to obtain electromagnetic characteristic information of the lithium battery shell;

[0073] Step S24: Calculate the best penetrating electromagnetic wave frequency according to the electromagnetic characteristic information of the lithium battery shell, so as to obtain the best penetrating electromagnetic detection frequency range;

[0074] Step S25: performing a multi-band electromagnetic wave penetration test on the temperature imbalance area and the temperature normal area based on the optimal penetration electromagnetic detection frequency range, and collecting electromagnetic reflection signals from the temperature imbalance area and the temperature normal area.

[0075] In this embodiment, the production logs of lithium batteries are accessed. These logs usually include production date, batch number, shell material type, thickness, density, component ratio, production process (such as injection molding, casting, etc.) and related control parameters (such as temperature, pressure, etc.). Use data processing software (such as Excel or Python's data analysis library Pandas) to extract relevant fields from the logs. Make sure to classify according to production date and batch number for subsequent analysis. During the data sorting process, check and clean the data to remove duplicate, erroneous records and missing values ​​to ensure the integrity and accuracy of the data. Analyze the extracted material data to determine the basic physical and chemical properties of different lithium battery shell materials. For common polymer materials, such as polypropylene (PP) and polyamide (PA), you can consult the material manual to obtain data such as their relative dielectric constant and conductivity. Record the characteristic data of each material, including its relative dielectric constant, conductivity, thermal conductivity, tensile strength and heat resistance. These parameters are crucial for evaluating the electromagnetic properties of lithium batteries. Relative dielectric constant It is an important parameter to measure the dielectric properties of materials. According to the molecular structure of the material and experimental data, it can be calculated using the following formula: ,in, is the dielectric constant of the material, is the dielectric constant of vacuum (about 8.854 × 10^-12F / m). Electrical conductivity is usually expressed as conductivity (σ), which can be calculated by resistivity (ρ): σ = 1 / ρ, where ρ is the resistivity of the material. For lithium battery shell materials, these parameters can be obtained through resistance measurement experiments. The calculated relative dielectric constant and conductivity data are recorded in the database to ensure that the data format is consistent for subsequent analysis. The calculation results can be verified using statistical software (such as MATLAB or Python's SciPy library) to ensure their accuracy. The electromagnetic properties of the lithium battery shell are analyzed based on the relative dielectric constant and conductivity. Finite element analysis (FEA) or electromagnetic field simulation software (such as COMSOL Multiphysics, ANSYS HFSS) can be used for simulation. Establish a geometric model of the lithium battery shell in the simulation software and input the relative dielectric constant and conductivity data. Set appropriate boundary conditions, such as the incident angle of the electromagnetic wave, and select a suitable excitation source type (such as a plane wave or a pulse signal). Run simulations to analyze the response of electromagnetic waves of different frequencies to the shell material, and extract key electromagnetic characteristic data, such as reflection coefficient (S11), transmission coefficient (S21), and absorption coefficient. Generally, electromagnetic performance within a specific frequency range may show different characteristics. According to the electromagnetic characteristics of the lithium battery shell, calculate the optimal penetrating electromagnetic wave frequency. This usually involves a comprehensive analysis of the reflection coefficient and the transmission coefficient to find the frequency range with the minimum reflection and the maximum transmission. By analyzing the simulation results, plotting the reflection and transmission characteristic curves, observing the responses at different frequencies, and determining the optimal frequency range for penetrating electromagnetic waves. Linear regression or optimization algorithms can be used to find the optimal frequency. Record the optimal penetration frequency range, such as 5 GHz to 10 GHz, to ensure that electromagnetic waves within this range can effectively penetrate the battery shell. Perform statistical analysis on the calculation results to ensure the accuracy and reliability of the frequency selection. In order to obtain a more comprehensive understanding, the frequency responses under different material characteristics can be compared and further feature extraction can be performed. Design an electromagnetic wave penetration test plan, select a suitable frequency source (such as a signal generator) and a receiving device (such as a spectrum analyzer) to ensure that the test can be performed within the optimal frequency range. Place the lithium battery on a dedicated test platform to ensure that it is stable and motionless. Configure the electromagnetic wave transmitting and receiving device, and set the frequency source to perform scanning tests within the optimal frequency range. Transmit electromagnetic waves at each frequency and record the reflected signals, paying special attention to the reflected signals in the temperature imbalance area and the temperature normal area. You can set multiple frequency bands (such as 5 GHz, 7 GHz, 9 GHz, etc.) for testing to ensure that the optimal penetration frequency range is covered. Use a spectrum analyzer to collect reflected signal data and record signal strength and changes. Analyze the signal reflection characteristics at different frequencies, compare them with temperature data, and identify the electromagnetic characteristics of the temperature imbalance area. Record the results in a database and generate a test report for subsequent analysis and decision-making.The report should include the amplitude of the reflected signal, the frequency response characteristics, and the correspondence between the temperature zones.

[0076] In this embodiment, refer to Figure 4 , is a flowchart of detailed implementation steps of step S3. In this embodiment, the detailed implementation steps of step S3 include:

[0077] Step S31: calculating the reflected signal strength of multiple frequency bands of the electromagnetic reflected signal to generate the reflected signal strength of each frequency band;

[0078] Step S32: Calculate the mean of the reflected signal amplitude based on the reflected signal strength of each frequency band, so as to obtain the mean of the multi-band reflected amplitude;

[0079] Step S33: performing amplitude variation trend analysis of the reflected signal strength in each frequency band to generate signal amplitude variation trends in different frequency bands;

[0080] Step S34: calculating the phase of the electromagnetic reflection signal in each frequency band, and extracting the phase characteristics of each frequency band;

[0081] Step S35: performing phase drift amplitude analysis based on the phase characteristics of each frequency band, thereby generating a phase drift range for each frequency band;

[0082] Step S36: mining the comprehensive reflection characteristics of each region for the mean value of the multi-band reflection amplitude, the signal amplitude variation trend of different frequency bands, and the phase drift range of each frequency band to obtain the signal reflection characteristics of the temperature imbalance region and the signal reflection characteristics of the temperature normal region.

[0083] In this embodiment, a spectrum analyzer is used to record electromagnetic reflection signals in different frequency bands (such as 5 GHz, 7 GHz, 9 GHz, etc.). Ensure that the environmental conditions remain consistent during the signal acquisition process in each frequency band to reduce noise interference. During the data acquisition process, set a suitable sampling rate and bandwidth to ensure the integrity of the signal. For example, the sampling rate can be set to twice the frequency to satisfy the Nyquist sampling theorem. The reflection signal data of each frequency band is processed and its signal strength is calculated. The calculated reflection signal strength of each frequency band is recorded in the data table for subsequent analysis. Ensure that the signal strength data of each frequency band has a consistent format and unit for easy comparison. After collecting the signal strength of each frequency band, calculate the mean amplitude of the reflection signal of each frequency band. Use programming tools such as Python or MATLAB to write a script to automatically calculate the mean of each frequency band. Enter all reflection signal strength data into the program to obtain the mean result. Record the amplitude mean of each frequency band and generate relevant charts (such as bar charts) to intuitively display the mean of the reflection signal of different frequency bands. For the reflection signal strength of each frequency band, use time series analysis or statistical methods to analyze its amplitude change trend. The rate of change of the signal strength of each frequency band can be calculated, R = S(t)-S(t-1) / S(t-1)×100%, where S(t) and S(t-1) are the signal strengths at the current moment and the previous moment, respectively. Use data analysis software (such as R or Python's Pandas library) to process the signal strength data, calculate the rate of change of the signal strength of each frequency band, and generate a change trend graph. Use a chart to show the change trend of the reflected signal strength of each frequency band, and observe whether there is an obvious upward or downward trend. Record the trend analysis results for subsequent comprehensive reflection feature mining. The phase characteristics of the reflected signal can be obtained by phase difference measurement. ϕ=arctan(I / Q), where I is the imaginary part of the signal and Q is the real part of the signal. The phase information of the reflected signal of each frequency band is extracted in the spectrum analyzer, and the phase characteristics of each frequency band are recorded. Tools such as MATLAB can be used for phase calculation. The obtained phase characteristic data is sorted to ensure that the phase data of each frequency band has a consistent format for subsequent analysis and comparison. The phase drift amplitude can be obtained by calculating the phase change. Phase drift usually refers to the change in the phase of a signal during transmission, which can be calculated using the following formula: Δϕ=ϕ(final)−ϕ(initial), where ϕ(final) is the final phase of the signal and ϕ(initial) is the initial phase of the signal. The phase data of each frequency band is analyzed to calculate the phase drift amplitude and record the phase change range. This calculation can be performed using programming tools such as Python or MATLAB. Generate a statistical report of the phase drift amplitude and record the phase drift range of each frequency band for subsequent comprehensive analysis.Based on the obtained multi-band reflection amplitude mean, amplitude change trend, and phase drift range, the signal reflection characteristics of the temperature imbalance area and the temperature normal area are comprehensively analyzed. Machine learning or pattern recognition algorithms (such as K-means clustering or support vector machine) can be used to distinguish the reflection characteristics of different areas. All calculation results are summarized to construct a feature matrix containing the amplitude mean, change trend, and phase drift data of each frequency band. A comprehensive reflection feature map is generated through visualization tools (such as Python's Matplotlib or Seaborn library) to show the differences in reflection characteristics of different areas. For the temperature imbalance area, focus on analyzing the differences between its reflection characteristics and normal areas to identify potential faults or abnormalities.

[0084] In this embodiment, step S4 includes the following steps:

[0085] Step S41: identifying the difference of the signal reflection characteristics of the temperature imbalance area and the signal reflection characteristics of the temperature normal area in each frequency band, and extracting the electromagnetic wave frequency band of each characteristic difference;

[0086] Step S42: performing multi-time-frequency decomposition according to each electromagnetic wave frequency band with characteristic differences, thereby obtaining frequency domain data of reflection signals of the difference frequency bands;

[0087] Step S43: performing regional feature deviation mining on the frequency domain data of the reflected signal in the difference frequency band to obtain the reflection deviation characteristics of the temperature imbalance area.

[0088] In this embodiment, the signal reflection characteristic data of the temperature imbalance area and the temperature normal area are extracted from the previous analysis. These data should include the reflection amplitude mean, phase characteristics, and amplitude change trend of multiple frequency bands. These data are organized into a table to ensure that the reflection characteristics of each frequency band can be easily compared. In order to identify the characteristic differences between different regions, statistical test methods (such as t-tests or analysis of variance ANOVA) can be used to compare the signal reflection characteristics of the two regions in each frequency band. This can help determine which frequency bands have significant differences. For example, use the SciPy library in Python to perform a t-test:

[0089] from scipy import stats

[0090] t_statistic, p_value = stats.ttest_ind(signal_region1, signal_region2), where signal_region1 and signal_region2 are the reflected signal data of the temperature imbalance region and the temperature normal region, respectively. After identifying the frequency bands with significant differences, record the specific frequency values ​​of these frequency bands. For example, assuming that the p-values ​​in the 7 GHz and 9 GHz frequency bands are less than 0.05, it can be considered that there are significant differences in the reflection characteristics of these two frequency bands. These difference frequency bands are organized into a table to facilitate subsequent analysis and discussion.

[0091] For the extracted characteristic difference frequency band, multi-time-frequency decomposition is performed to obtain the frequency domain data of its reflection signal. Methods such as wavelet transform (Wavelet Transform) or short-time Fourier transform (STFT) can be used. Wavelet transform performs better when processing non-stationary signals.

[0092] import pywt

[0093] coeffs = pywt.wavedec(signal_data, 'db4', level=5)

[0094] Among them, signal_data is the reflection signal data of a specific frequency band, 'db4' is the selected wavelet basis, and level is the number of decomposition layers. Through the wavelet coefficients, the frequency components on different time scales can be obtained, so as to analyze the frequency domain characteristics of the signal in a specific frequency band. The calculated frequency domain data is sorted to extract the main frequency components and amplitude information of the difference frequency band. A spectrum diagram can be generated to compare the reflection characteristics of different frequency bands in the temperature imbalance and normal areas. Ensure that the frequency domain data has clear identification for subsequent analysis and visualization. For the frequency domain data of each difference frequency band, calculate the deviation characteristics of the reflection signal between the temperature imbalance area and the temperature normal area. This can be achieved by comparing the mean and standard deviation of the two sets of frequency domain data. Calculate the mean and standard deviation of the frequency domain data of the temperature imbalance area and the temperature normal area in each difference frequency band:

[0095] mean_region1 = np.mean(freq_domain_data_region1)

[0096] std_region1 = np.std(freq_domain_data_region1)

[0097] By comparing the mean and standard deviation, the deviation characteristics of the temperature imbalance area can be identified. For example, if the mean value of the imbalance area in a certain frequency band is significantly higher than the mean value of the normal area, it can be considered that there is a reflection characteristic deviation in this frequency band.

[0098] In this embodiment, step S5 includes the following steps:

[0099] Step S51: performing electromagnetic distribution analysis on the electromagnetic reflection signals in the temperature imbalance area and the temperature normal area to generate an electromagnetic distribution field on the surface of the lithium battery;

[0100] Step S52: performing nonlinear coupling analysis of electromagnetic parameters according to the electromagnetic distribution field on the surface of the lithium battery to obtain nonlinear coupling characteristics of the electromagnetic field;

[0101] Step S53: performing battery water content interference analysis on the electromagnetic field nonlinear coupling characteristics according to the reflection deviation characteristics of the temperature imbalance area, so as to obtain the water content interference law of the temperature imbalance area;

[0102] Step S54: Perform correlation analysis on dynamic changes of moisture content based on the moisture content interference law in the temperature imbalance area, so as to construct a dynamic electromagnetic moisture content interference model.

[0103] In this embodiment, the electromagnetic reflection signal data of the temperature imbalance area and the temperature normal area are extracted from the previous analysis. These data should include information such as the reflection signal intensity, phase, and spectrum of each frequency band. Ensure the accuracy of the data and avoid data deviations caused by environmental noise or equipment errors. Use numerical simulation software (such as ANSYS HFSS or COMSOL Multiphysics) to create an electromagnetic distribution field model of a lithium battery. Enter the geometric parameters of the model, including the size of the battery, material properties (such as dielectric constant, conductivity, etc.), and set appropriate boundary conditions. For the incident conditions of the electromagnetic wave, select a suitable incident angle and polarization mode. For example, you can choose plane wave incidence and set the incident angle to 0 degrees to simulate the case of vertical incidence. After completing the model setting, run the electromagnetic field simulation to obtain the electromagnetic field distribution map on the surface of the lithium battery. During the simulation process, ensure that the calculation time and accuracy are set reasonably to obtain high-quality results. Generate distribution maps of electric field intensity, magnetic field intensity, and electromagnetic wave reflectivity through post-processing tools. Use visualization tools such as Matplotlib to convert electromagnetic distribution field data into heat maps or three-dimensional maps to show the changes in electromagnetic field intensity in different regions. Pay special attention to the difference between the electromagnetic field in the temperature imbalance area and the conventional area, so as to provide basic data for the subsequent nonlinear coupling analysis. The nonlinear coupling analysis of electromagnetic parameters involves the mutual influence between the electric field, temperature and humidity. According to the Maxwell equations and the heat conduction equation, the establishment of a coupling model can better describe the changes of the electromagnetic field under different environmental conditions. When constructing the electromagnetic field model, temperature and humidity are introduced into the model as variables. You can use multi-physics simulation software such as COMSOL Multiphysics to create a coupling model that includes electromagnetic fields, thermal fields and fluid fields. Set different environmental parameters, such as temperature range (such as 25°C to 85°C) and humidity range (such as 20% to 100%), and perform multiple groups of simulations to obtain the response of the electromagnetic field to different environmental conditions. Run the coupling analysis and record data such as electromagnetic field intensity, reflection characteristics and phase change. Make sure to perform multiple simulations under different environmental conditions to obtain reliable statistical data. Generate a nonlinear coupling characteristic diagram of the electromagnetic field related to the environmental conditions, and record the relationship between the electric field intensity and temperature and humidity. These data can be processed by data analysis software (such as MATLAB or Python) to generate a relationship diagram between the reflection coefficient and temperature and humidity. Use visualization tools to display nonlinear coupling characteristics, paying special attention to the nonlinear response in the temperature imbalance area to provide a basis for subsequent interference analysis. Water content usually affects the dielectric properties and conductivity of the battery, which in turn affects the electromagnetic reflection characteristics. It is key to establish a relationship model between water content and electromagnetic reflection. Collect reflection deviation feature data in the temperature imbalance area and match it with the moisture content data measured in the laboratory. Ensure that the time and environmental conditions of the data are consistent to reduce interference factors.

[0104] Use statistical analysis methods (such as regression analysis) to establish the relationship between reflection deviation and water content. For example, use a linear regression model to calculate the correlation coefficient between reflection signal deviation and water content:

[0105] from sklearn.linear_model import LinearRegression

[0106] model = LinearRegression()

[0107] model.fit(X, y) # X is the water content data, y is the reflection deviation feature, analyze the goodness of fit of the model to ensure that the model can effectively describe the relationship between reflection deviation and water content. According to the water content interference law in the temperature imbalance area, use the time series analysis method to observe the influence of water content changing over time on electromagnetic reflection characteristics.

[0108] Collect real-time monitoring data of temperature imbalance areas, including electromagnetic reflection characteristics and moisture content change data.

[0109] import pandas as pd

[0110] data = pd.read_csv('dataset.csv') # Import data

[0111] The ARIMA (Autoregressive Integrated Moving Average) model is used for time series analysis to capture the dynamic characteristics of water content: from statsmodels.tsa.arima.model import ARIMA

[0112] model = ARIMA(data['water content'], order=(p, d, q)) # p, d, q are model parameters

[0113] model_fit = model.fit()

[0114] Through model fitting, the dynamic electromagnetic moisture content interference model is obtained, the model parameters and prediction results are recorded, and the model is verified to ensure that the model can accurately predict the changes in moisture content.

[0115] In this embodiment, step S6 includes the following steps:

[0116] Step S61: estimating the water interference intensity of the temperature imbalance area using the dynamic electromagnetic water content interference model to obtain a water interference intensity value;

[0117] Step S62: quantifying the interference phase deviation range of the dynamic electromagnetic water content interference model, thereby generating a water content interference phase deviation range;

[0118] Step S63: performing deep iterative learning according to the water interference intensity value and the water interference phase deviation range, thereby generating a dynamic water interference learning sample;

[0119] Step S64: performing reverse calculation of the moisture content in the temperature imbalance region based on the dynamic moisture interference learning sample to obtain the moisture content in each imbalance region;

[0120] Step S65: Comprehensively evaluate the water content of each unbalanced area to generate a lithium battery water content detection report.

[0121] In this embodiment, ensure that the dynamic electromagnetic moisture interference model has been established and verified. The model should include electromagnetic reflection characteristics, moisture content data and environmental condition parameters (such as temperature, humidity, etc.) in the temperature imbalance area. Collect and organize relevant data to ensure its consistency and accuracy. Interference intensity estimation method: I (interference) = α⋅S (reflection) + β, where I (interference) is the interference intensity, α and β are regression coefficients, and S (reflection) is the reflection signal strength, which is obtained by linear regression or polynomial regression. Use the scikit-learn library in Python for linear regression analysis: from sklearn.linear_modelimport LinearRegression

[0122] model = LinearRegression()

[0123] model.fit(X, y) # X is the reflected signal intensity, y is the interference intensity, and the water interference intensity can be estimated by calculating the relationship between the electromagnetic wave reflected signal intensity and the water content. By fitting the model, the water interference intensity value of the temperature imbalance area is calculated. Make sure to perform multiple estimates at different temperatures and humidities to obtain reliable intensity values. In the dynamic electromagnetic moisture interference model, the phase of the reflected signal with different moisture contents is measured. Use a spectrum analyzer or signal processing software (such as MATLAB or Python) to extract the phase information. Calculate the phase deviation range and record the phase deviation data for each imbalance area. The following code can be used for phase calculation:

[0124] import numpy as np

[0125] phase_actual = np.angle(signal_actual)

[0126] phase_reference = np.angle(signal_reference)

[0127] phase_deviation = np.abs(phase_actual - phase_reference)

[0128] Organize the phase deviation range into a table, indicating the phase deviation value for each imbalance region. Generate a visual chart to show the distribution of the water interference phase deviation range for intuitive analysis and comparison. Deep iterative learning can be achieved by building a neural network model. Choose a suitable deep learning framework (such as TensorFlow or PyTorch) to build the neural network. The model input is the water interference intensity value and the phase deviation, and the output is the predicted water content value. Prepare a training data set, including the estimated water interference intensity value and the corresponding phase deviation range. Divide the data into a training set and a test set for model training and verification.

[0129] Build a neural network model and set the appropriate number of layers and nodes, such as a simple fully connected network:

[0130] import tensorflow as tf

[0131] model = tf.keras.Sequential([

[0132] tf.keras.layers.Dense(64, activation='relu', input_shape=(2,)),#Input 2 features

[0133] tf.keras.layers.Dense(64, activation='relu'),

[0134] tf.keras.layers.Dense(1) # Output is water content ])

[0136] model.compile(optimizer='adam', loss='mean_squared_error') Train the model using the training set to minimize the loss function (e.g., mean squared error). During the training process, the test set is used regularly to verify the performance of the model. After training, the accuracy and loss value of the model are recorded to ensure that the model can effectively predict the moisture content. Use the trained deep learning model to predict the new samples to infer the moisture content of each temperature imbalance area. This process is performed by inputting the known moisture interference intensity value and phase deviation range. Input the new sample data (including the moisture interference intensity value and phase deviation range) into the trained model to predict the moisture content: predictions = model.predict(new_samples) # new_samples is the new sample data, record the predicted moisture content value of each imbalance area, and ensure the accuracy and reliability of the data. The inverse moisture content data is organized into a table, listing the predicted moisture content value of each imbalance area. Generate relevant charts to show the moisture content distribution in different areas for subsequent comprehensive evaluation. A comprehensive assessment is conducted on the moisture content data of each imbalanced area, including statistical indicators such as the mean, standard deviation, maximum and minimum values ​​of the moisture content, to ensure the comprehensiveness of the assessment results.

[0137] Use statistical analysis tools (such as Python's Pandas library) to process the moisture content data and calculate various statistical indicators:

[0138] import pandas as pd

[0139] water_content_stats = pd.DataFrame({

[0140] 'Mean': predictions.mean(),

[0141] 'Standard deviation': predictions.std(),

[0142] 'Maximum value': predictions.max(),

[0143] 'Minimum': predictions.min()

[0144] }) Based on the statistical results, write a lithium battery water content test report, describing in detail the water content test results of each imbalance area. The report should include statistical analysis of water content, possible influencing factors and recommended measures. Generate a visual chart to show the distribution of water content in each imbalance area, so that relevant personnel can quickly understand the test results.

[0145] In this embodiment, a lithium battery water content detection system is provided, which is used to execute the lithium battery water content detection method as described above, including:

[0146] The temperature mutation detection module is used to obtain the thermal imaging image of the lithium battery surface; identify the temperature gradient distribution of the thermal imaging image of the lithium battery surface, and perform mutation detection of local distribution changes to extract the temperature imbalance area and the temperature normal area;

[0147] The optimal frequency module is used to calculate the optimal penetrating electromagnetic wave frequency according to the temperature imbalance area and the temperature normal area, and to perform multi-band electromagnetic wave penetration test to collect electromagnetic reflection signals from the temperature imbalance area and the temperature normal area;

[0148] A reflection feature mining module is used to calculate the reflection signal strength of multiple frequency bands of the electromagnetic reflection signal and mine the comprehensive reflection features of each region to obtain the signal reflection features of the temperature imbalance region and the signal reflection features of the temperature normal region;

[0149] The reflection deviation module is used to identify the difference of the signal reflection characteristics of the temperature imbalance area and the signal reflection characteristics of the temperature normal area in each frequency band, and to mine the regional feature deviation to obtain the reflection deviation characteristics of the temperature imbalance area;

[0150] The water content interference module is used to perform nonlinear coupling analysis of electromagnetic parameters based on the reflection deviation characteristics of the temperature imbalance area, and to perform correlation analysis of the dynamic change of water content, so as to construct a dynamic electromagnetic water content interference model;

[0151] The moisture content inversion module is used to infer the moisture content in the temperature imbalance area of ​​the dynamic electromagnetic moisture content interference model and conduct a comprehensive evaluation to generate a lithium battery moisture content detection report.

[0152] The present invention can efficiently identify the abnormal change area of ​​battery surface temperature through thermal imaging images. The temperature imbalance area is often closely related to the uneven distribution of moisture inside the battery or the heat accumulation inside the battery. Thermal imaging is a non-contact detection method, which avoids the damage that may be caused by direct contact with the battery surface, and is helpful for long-term safety monitoring of the battery. The temperature mutation detection can identify the potential problems of the battery in real time, take protective measures in time, and improve the battery safety. By analyzing the characteristics of the temperature imbalance area and the conventional area, the optimal penetration frequency range is calculated to maximize the penetration ability of electromagnetic waves. The application of multi-band electromagnetic waves can comprehensively and detailedly capture the reflection signals of different areas on the battery surface and inside, ensuring the accuracy of the moisture detection results. By analyzing the intensity of multiple frequency band signals, more abundant electromagnetic reflection features can be extracted to provide strong support for subsequent signal analysis, and the characteristic differences of different areas (temperature imbalance area and conventional area) can be distinguished, making the subsequent moisture content analysis more accurate. By comprehensively mining the signal strength and characteristics, the quality of data acquisition can be effectively improved, thereby improving the reliability of the entire detection system, and effectively identifying the temperature imbalance area and the conventional area. The signal deviation in the conventional area can accurately distinguish the reflection feature changes caused by moisture or other factors. By mining the reflection deviation features, the abnormal area inside the battery can be quickly identified to help locate the possible moisture accumulation area. Through the detailed analysis of the reflection deviation, the tiny electromagnetic changes can be captured, thereby improving the sensitivity and accuracy of the detection system. Through nonlinear coupling analysis, the complex nonlinear relationship between the moisture content change and the electromagnetic wave reflection signal can be revealed. This modeling method can more accurately capture the subtle effects of moisture changes. The model can not only describe the effect of moisture content on electromagnetic wave propagation, but also reflect the dynamic process of moisture changes in real time, providing an accurate basis for subsequent analysis. Through the dynamic coupling of electromagnetic parameters and moisture content, accurate moisture content detection data can be provided for each area to reduce errors. Through the reverse algorithm, the moisture content of each area inside the battery can be accurately obtained, avoiding the sampling error in the traditional method. It comprehensively considers a variety of electromagnetic parameters (such as reflection intensity, phase change, etc.), making the moisture content detection results more comprehensive and accurate, and can automatically generate a detailed moisture content detection report, providing direct support for battery health management and maintenance.

[0153] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0154] The above is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for detecting water content in a lithium battery, characterized in that: The following steps are involved: Step S1: Acquire a thermal imaging image of the lithium battery surface; Identify the temperature gradient distribution of the thermal imaging image of the lithium battery surface and detect the sudden change of local distribution changes to extract the temperature imbalance area and the normal temperature area; Step S2: Calculate the optimal penetrating electromagnetic wave frequency according to the temperature imbalance area and the temperature normal area, and perform a multi-band electromagnetic wave penetration test to collect electromagnetic reflection signals from the temperature imbalance area and the temperature normal area; Step S3: calculating the reflection signal strength of multiple frequency bands of the electromagnetic reflection signal and mining the comprehensive reflection characteristics of each region to obtain the signal reflection characteristics of the temperature imbalance region and the signal reflection characteristics of the temperature normal region; Step S4: identifying the difference of the signal reflection characteristics of the temperature imbalance area and the signal reflection characteristics of the temperature normal area in each frequency band, and mining the regional feature deviation to obtain the reflection deviation characteristics of the temperature imbalance area; Step S5: performing nonlinear coupling analysis of electromagnetic parameters according to the reflection deviation characteristics of the temperature imbalance area, and performing correlation analysis of the dynamic change of water content, so as to construct a dynamic electromagnetic water content interference model; Step S6: reverse the water content in the temperature imbalance area of ​​the dynamic electromagnetic water content interference model, and conduct a comprehensive evaluation to generate a lithium battery water content detection report; Among them, the specific steps of step S3 are: Step S31: calculating the reflected signal strength of multiple frequency bands of the electromagnetic reflected signal to generate the reflected signal strength of each frequency band; Step S32: Calculate the mean of the reflected signal amplitude based on the reflected signal strength of each frequency band, so as to obtain the mean of the multi-band reflected amplitude; Step S33: performing amplitude variation trend analysis of the reflected signal strength in each frequency band to generate signal amplitude variation trends in different frequency bands; Step S34: calculating the phase of the electromagnetic reflection signal in each frequency band, and extracting the phase characteristics of each frequency band; Step S35: performing phase drift amplitude analysis based on the phase characteristics of each frequency band, thereby generating a phase drift range for each frequency band; Step S36: mining the comprehensive reflection characteristics of each region for the mean value of the multi-band reflection amplitude, the signal amplitude variation trend of different frequency bands, and the phase drift range of each frequency band, so as to obtain the signal reflection characteristics of the temperature imbalance region and the signal reflection characteristics of the temperature normal region; The specific steps of step S4 are: Step S41: identifying the difference of the signal reflection characteristics of the temperature imbalance area and the signal reflection characteristics of the temperature normal area in each frequency band, and extracting the electromagnetic wave frequency band of each characteristic difference; Step S42: performing multi-time-frequency decomposition according to each electromagnetic wave frequency band with characteristic differences, thereby obtaining frequency domain data of reflection signals of the difference frequency bands; Step S43: performing regional feature deviation mining on the frequency domain data of the reflected signal in the difference frequency band to obtain the reflection deviation characteristics of the temperature imbalance area.

2. The method for detecting water content of a lithium battery according to claim 1, characterized in that: The specific steps of step S1 are: Step S11: using a thermal imager to perform a comprehensive scan and temperature measurement on the lithium battery to be inspected, so as to generate a thermal imaging image of the lithium battery surface; Step S12: performing adaptive noise filtering on the thermal imaging image of the lithium battery surface to construct an adaptive filtered thermal imaging image; Step S13: performing temperature gradient distribution recognition on the adaptively filtered thermal imaging image to generate image temperature gradient distribution data; Step S14: performing gradient distribution change analysis on the image temperature gradient distribution data to obtain temperature gradient distribution change characteristics; Step S15: Based on the temperature gradient distribution change characteristics, the adaptive filtered thermal imaging image is subjected to local distribution change mutation detection to extract the temperature imbalance area and the temperature normal area.

3. The method for detecting water content of a lithium battery according to claim 2, characterized in that: The specific steps of step S12 are: Obtain the humidity parameters of the environment to be detected; Performing a dark channel priori calculation on the thermal imaging image of the lithium battery surface according to the humidity parameter of the environment to be detected to obtain an atmospheric shielding factor; The transmission map value of each pixel of the thermal imaging image of the lithium battery surface is estimated, thereby generating the transmission map value of each pixel; Perform thermal imaging defogging correction according to the atmospheric shielding factor and the transmission map value of each pixel to construct a defogging corrected thermal imaging image; Perform image noise detection on the dehazed and corrected thermal imaging image and mark all image noise points; Calculating the grayscale distribution of all the image noise points to generate a grayscale distribution value for each noise point; Performing geometric feature analysis on all the image noise points one by one to extract the geometric features of each noise point; The noise point depth semantic classification is performed based on the grayscale distribution value of each noise point and the geometric features of each noise point, so as to obtain the type characteristics of each noise point; Based on the type characteristics of each noise point, the dehazed and corrected thermal imaging image is adaptively filtered to construct an adaptive filtered thermal imaging image.

4. The method for detecting water content of a lithium battery according to claim 1, characterized in that: The specific steps of step S2 are: Step S21: extracting lithium battery shell material parameters based on lithium battery production log; Step S22: Calculate the relative dielectric constant and conductivity according to the material parameters of the lithium battery shell to generate the relative dielectric constant and conductivity parameters of the lithium battery; Step S23: mining the electromagnetic characteristics of the shell based on the relative dielectric constant and conductivity parameters to obtain electromagnetic characteristic information of the lithium battery shell; Step S24: Calculate the best penetrating electromagnetic wave frequency according to the electromagnetic characteristic information of the lithium battery shell, so as to obtain the best penetrating electromagnetic detection frequency range; Step S25: performing a multi-band electromagnetic wave penetration test on the temperature imbalance area and the temperature normal area based on the optimal penetration electromagnetic detection frequency range, and collecting electromagnetic reflection signals from the temperature imbalance area and the temperature normal area.

5. The method for detecting water content of a lithium battery according to claim 1, characterized in that: The specific steps of step S5 are: Step S51: performing electromagnetic distribution analysis on the electromagnetic reflection signals in the temperature imbalance area and the temperature normal area to generate an electromagnetic distribution field on the surface of the lithium battery; Step S52: performing nonlinear coupling analysis of electromagnetic parameters according to the electromagnetic distribution field on the surface of the lithium battery to obtain nonlinear coupling characteristics of the electromagnetic field; Step S53: performing battery water content interference analysis on the electromagnetic field nonlinear coupling characteristics according to the reflection deviation characteristics of the temperature imbalance area, so as to obtain the water content interference law of the temperature imbalance area; Step S54: Perform correlation analysis on dynamic changes of moisture content based on the moisture content interference law in the temperature imbalance area, so as to construct a dynamic electromagnetic moisture content interference model.

6. The method for detecting water content of a lithium battery according to claim 1, characterized in that: The specific steps of step S6 are: Step S61: estimating the water interference intensity of the temperature imbalance area using the dynamic electromagnetic water content interference model to obtain a water interference intensity value; Step S62: quantifying the interference phase deviation range of the dynamic electromagnetic water content interference model, thereby generating a water content interference phase deviation range; Step S63: performing deep iterative learning according to the water interference intensity value and the water interference phase deviation range, thereby generating a dynamic water interference learning sample; Step S64: performing reverse calculation of the moisture content in the temperature imbalance region based on the dynamic moisture interference learning sample to obtain the moisture content in each imbalance region; Step S65: Comprehensively evaluate the water content of each unbalanced area to generate a lithium battery water content detection report.

7. A lithium battery water content detection system, characterized in that: Used to perform the lithium battery water content detection method as claimed in claim 1, comprising: The temperature mutation detection module is used to obtain the thermal imaging image of the lithium battery surface; identify the temperature gradient distribution of the thermal imaging image of the lithium battery surface, and perform mutation detection of local distribution changes to extract the temperature imbalance area and the temperature normal area; The optimal frequency module is used to calculate the optimal penetrating electromagnetic wave frequency according to the temperature imbalance area and the temperature normal area, and to perform multi-band electromagnetic wave penetration test to collect electromagnetic reflection signals from the temperature imbalance area and the temperature normal area; A reflection feature mining module is used to calculate the reflection signal strength of multiple frequency bands of the electromagnetic reflection signal and mine the comprehensive reflection features of each region to obtain the signal reflection features of the temperature imbalance region and the signal reflection features of the temperature normal region; The reflection deviation module is used to identify the difference of the signal reflection characteristics of the temperature imbalance area and the signal reflection characteristics of the temperature normal area in each frequency band, and to mine the regional feature deviation to obtain the reflection deviation characteristics of the temperature imbalance area; The water content interference module is used to perform nonlinear coupling analysis of electromagnetic parameters based on the reflection deviation characteristics of the temperature imbalance area, and to perform correlation analysis of the dynamic change of water content, so as to construct a dynamic electromagnetic water content interference model; The moisture content inversion module is used to infer the moisture content in the temperature imbalance area of ​​the dynamic electromagnetic moisture content interference model and conduct a comprehensive evaluation to generate a lithium battery moisture content detection report.

Images