Ultrasonic method for detecting electrolyte distribution in a battery

Abstract

The application relates to an ultrasonic detection method for electrolyte distribution in a battery. The method comprises the following steps: providing a battery to be detected, performing ultrasonic wave scanning on the battery to be detected to obtain a group of to-be-detected ultrasonic transmission signals corresponding to an ROI region; extracting a transmission signal characteristic value corresponding to each to-be-detected ultrasonic transmission signal, and generating an electrolyte content calculation value corresponding to the transmission signal characteristic value by using a pre-constructed characteristic-electrolyte content model; determining an electrolyte distribution state of the ROI region in the battery to be detected based on electrolyte calculation contents of all scanning detection points, so as to represent an electrolyte infiltration state of the ROI region in the battery to be detected by using the electrolyte distribution state. The application realizes real-time detection of the electrolyte distribution infiltrated in the battery by using ultrasonic waves, and improves the detection efficiency and accuracy.

Description

Technical Field

[0001] This invention relates to an ultrasonic testing method, and more particularly to an ultrasonic testing method for electrolyte distribution within a battery. Background Technology

[0002] In the manufacturing process of electrochemical energy storage devices such as lithium-ion batteries and emerging solid-state batteries, the injection and full wetting of electrolyte are core process steps that determine battery performance, safety, and batch consistency. If the electrolyte fails to achieve rapid and uniform penetration into the porous structure composed of electrodes and separators, it can easily lead to local ion transport obstruction, manifested as increased internal resistance, decreased effective capacity, and shortened cycle life. In severe cases, it can even cause micro-short circuits or thermal runaway risks due to local dry areas.

[0003] However, effective monitoring of the electrolyte wetting state inside batteries has long been a technological bottleneck in battery manufacturing. Currently, the industry primarily relies on offline, destructive methods for detection. For example, while weighing is simple to operate, it only reflects the total electrolyte volume and cannot reveal the spatial distribution of the electrolyte. Disassembly and observation methods require damaging the battery structure, making them unsuitable for online monitoring and unable to quantify the wetting degree at different locations, limiting their application to limited sampling inspection scenarios. In recent years, some research has attempted to introduce non-destructive testing technologies. While X-ray computed tomography (X-CT) can provide information on the internal three-dimensional structure, the electrolyte is often a low-atomic-number organic solvent (such as carbonates), whose X-ray absorption coefficient is very similar to that of porous carbon / metal oxide electrode materials, resulting in severely insufficient imaging contrast and difficulty in clearly distinguishing the liquid phase distribution. Furthermore, the high cost and slow scanning speed of the equipment make it difficult to integrate into high-speed production lines. Infrared thermal imaging relies on slight changes in thermal conductivity or specific heat capacity before and after immersion. It is easily affected by factors such as ambient temperature fluctuations, battery self-heating, and uneven surface emissivity. It has poor quantitative accuracy and lacks spatial resolution, making it difficult to support closed-loop process control.

[0004] Ultrasonic testing, due to its advantages of being non-destructive, rapid, and sensitive to the acoustic properties of the medium, theoretically has the potential to become an online monitoring tool for electrolyte wetting status. In fact, some studies have already used it to identify macroscopic defects such as bubbles and electrode delamination inside batteries. However, in the specific application scenario of electrolyte wetting, existing ultrasonic methods still have significant limitations: most schemes only make qualitative judgments on "whether wetting has occurred" based on rough changes in sound velocity or attenuation, lacking a physical model that establishes a quantitative mapping relationship between the ultrasonic response signal and the local electrolyte content; more importantly, current ultrasonic testing, based on two-dimensional images of ultrasonic transmission signal intensity, sound velocity, or attenuation coefficient, cannot directly distinguish whether a local area is a bubble or poorly wetted, and cannot quickly and accurately detect the distribution of electrolyte wetting inside the battery.

[0005] As can be seen from the above description, how to use ultrasound to quickly and accurately detect the distribution of electrolyte inside a battery is a technical problem that urgently needs to be solved. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide an ultrasonic detection method for electrolyte distribution in a battery. This method uses ultrasound to detect the distribution of electrolyte in the battery in real time, thereby improving the efficiency and accuracy of the detection.

[0007] According to the technical solution provided by the present invention, an ultrasonic detection method for electrolyte distribution in a battery is provided, the ultrasonic detection method comprising:

[0008] Provide the battery to be tested. The battery under test is subjected to ultrasonic scanning to obtain a set of ultrasonic transmission signals corresponding to the ROI region. The ROI region includes at least the electrolyte wetting area in the battery under test. The set of ultrasonic transmission signals includes several ultrasonic transmission signals, and each ultrasonic transmission signal corresponds to a scanning detection point in the ROI region. Extract the corresponding transmission signal feature value for each ultrasound transmission signal to be tested, and use the pre-constructed feature-electrolyte content model to generate the electrolyte content solution value corresponding to the transmission signal feature value; Based on the electrolyte content calculated at all scanning detection points, the electrolyte distribution state in the ROI region of the battery under test is determined, so as to characterize the electrolyte wetting state in the ROI region of the battery under test using the electrolyte distribution state.

[0009] The characteristic-electrolyte content model, when constructed through data fitting, includes: The mathematical model used in the feature-electrolyte content model is determined, and a fitting dataset is created based on the mathematical model used in the feature-electrolyte content model to fit the data. When creating a fitted dataset, the following steps are included: Provide several reference cells with known electrolyte injection amounts but not entirely identical electrolyte wetting distribution states; Each reference cell was subjected to ultrasonic scanning, and the reference ultrasonic transmission signal group corresponding to the ROI region within the reference cell was obtained. Based on the reference ultrasonic transmission signal sets of all reference cells and the electrolyte injection amount of each reference cell, the battery dataset for each reference cell is calculated, and a fitting dataset is formed based on the battery datasets of all reference cells. Each reference battery's battery dataset includes several fitted data samples. Each fitted data sample includes a reference electrolyte content per unit area and a transmission fitted feature value corresponding to the reference electrolyte content per unit area. The type of the transmission fitted feature value is consistent with the type of the transmission signal feature value. For each reference battery's battery dataset, the number of fitted data samples within the battery dataset is consistent, and the fitted data samples within all reference datasets have a consistent distribution.

[0010] Solving the battery dataset for each reference battery includes: Based on the corresponding transmission reference feature value of each reference ultrasound transmission signal in the reference ultrasound transmission signal group, a reference feature distribution histogram of each reference cell is constructed to characterize the distribution state of the transmission reference feature value of the corresponding reference cell. The type of transmission reference feature value is consistent with the type of transmission signal feature value. For each reference cell, the histogram of reference feature distribution is divided into reference feature intervals, and the area of ​​each reference feature sub-interval is calculated. When dividing the reference feature intervals, the number of reference feature sub-intervals obtained by dividing each reference feature distribution histogram is the same, the reference feature interval ranges of the corresponding reference feature sub-intervals are consistent, and the number of reference feature sub-intervals obtained is no more than the number of reference batteries. For each reference cell, based on the area of ​​each reference feature sub-interval and the electrolyte injection amount of the reference cell, an electrolyte content distribution state equation is constructed to characterize the equivalent relationship between sub-interval area - electrolyte content per unit area - electrolyte injection amount. Based on the state equation of electrolyte content distribution of all reference cells, the electrolyte content per unit area of ​​each reference characteristic sub-interval is calculated. For each reference cell, the transmission fitting feature value of each reference feature sub-interval is determined, and a fitting data sample of the reference cell is formed based on the determined transmission fitting feature value and the corresponding electrolyte content per unit area.

[0011] Within the reference feature distribution histogram, we have: The horizontal axis of the reference feature distribution histogram corresponds to the transmission reference feature value, and the vertical axis of the reference feature distribution histogram represents the number of each transmission reference feature value. When dividing the reference feature interval, the following is included: Based on the transmission reference characteristic values ​​of all reference cells, the distribution range of the transmission reference characteristic values ​​is determined, and reference characteristic interval division parameters are configured based on the determined distribution range. The reference characteristic interval division parameters include at least the number of reference characteristic sub-intervals and the corresponding interval range of each reference characteristic sub-interval.

[0012] When determining the transmission fitting eigenvalues ​​for each reference feature sub-interval, the following steps are included: Based on the distribution of all transmission reference feature values ​​in the aforementioned reference feature sub-interval, transmission fitting feature values ​​are statistically generated, wherein... When generating transmission fitting feature values, the following are included: The average value of all transmission reference feature values ​​in the reference feature sub-interval is calculated, and the result of the average calculation is configured as the transmission fitting feature value. or, The median of all transmission reference feature values ​​within the reference feature sub-interval is statistically determined, and the statistically determined median is configured as the transmission fitting feature value.

[0013] For the feature-electrolyte content model constructed through data fitting, then:

[0014] in, These are characteristic values ​​of the transmitted signal. Electrolyte content per unit area , These are the weighting coefficients. , , , , , These are the parameters determined through data fitting.

[0015] When constructing the feature-electrolyte content model based on machine learning, it also includes: A machine learning-based electrolyte content prediction model is provided, and a training dataset is created for training the model. The training dataset includes several training samples, each training sample including a transmission training feature value and a label of electrolyte content per unit area corresponding to the transmission training feature value. Configure the model training conditions, and train the electrolyte content prediction model using the training dataset under the configured model training conditions until the target training state is reached. After that, configure the electrolyte content prediction model that has reached the target training state as the feature-electrolyte content model.

[0016] After obtaining the calculated electrolyte content value corresponding to each scanning detection point, the wetting state of the battery under test is analyzed. When performing infiltration state analysis, the following are included: Based on the calculated electrolyte content value corresponding to each scanning detection point, the electrolyte wetting amount in the strong region corresponding to the strong ultrasonic signal region is statistically analyzed. Based on the actual amount of electrolyte injected into the battery under test and the amount of electrolyte wetting in the area with strong ultrasonic signal, the amount of electrolyte wetting in the weak area corresponding to the area with weak ultrasonic signal is calculated. The area of ​​the corresponding weak region in the weak ultrasonic signal region is statistically analyzed. Then, based on the electrolyte wetting amount in the weak region and the corresponding weak region area, the electrolyte content per unit area of ​​the weak region is calculated. Based on the electrolyte content per unit area in the weak region calculated above, the wetting state of the battery under test is judged. Specifically, when the electrolyte content per unit area in the weak region is greater than the electrolyte wetting threshold in the weak region, it is determined that air bubbles exist in the ultrasonic weak signal area. When the electrolyte content per unit area in the weak region is not greater than the electrolyte wetting threshold in the weak region, it is determined that there is insufficient wetting in the weak ultrasonic signal region.

[0017] After obtaining the set of ultrasonic transmission signals to be examined, the process also includes signal interpolation, in which... When performing signal interpolation, interpolation is performed based on the ultrasonic transmission signals to be tested at least two adjacent scanning detection points to generate the interpolated transmission signal to be tested. The electrolyte content corresponding to each interpolated transmission signal to be detected is generated using a feature-electrolyte content model. Based on the calculated electrolyte content at all scanning detection points and the calculated electrolyte content corresponding to all interpolated transmission signals to be tested, the electrolyte distribution status in the ROI region of the battery under test is determined.

[0018] It also includes a distribution map of the infiltration degree of the ROI region, in which, When generating the infiltration degree distribution map, the following steps are included: Determine the maximum value of all electrolyte solutions and normalize all electrolyte solutions using the maximum value to generate the corresponding wetting normalization value for each electrolyte solution. Based on all the normalized infiltration values, a state map of the infiltration degree distribution is generated.

[0019] The advantages of this invention are as follows: Ultrasonic scanning is performed on the battery under test to obtain a set of ultrasonic transmission signals corresponding to the Region of Interest (ROI). The characteristic values ​​of each ultrasonic transmission signal are extracted, and a pre-constructed feature-electrolyte content model is used to generate electrolyte content calculation values ​​corresponding to the transmission signal characteristic values. Based on the calculated electrolyte content at all scanning detection points, the electrolyte distribution state within the ROI region of the battery under test is determined. This electrolyte distribution state is used to characterize the electrolyte wetting state within the ROI region of the battery under test, thereby enabling real-time detection of the electrolyte distribution within the battery under test using ultrasound, improving detection efficiency and accuracy. Attached Figure Description

[0020] Figure 1 This is a flowchart of an embodiment of the present invention for ultrasonic detection of electrolyte distribution in a battery.

[0021] Figure 2 This is a schematic diagram of the first embodiment of the present invention, which generates a transmission image of the battery under test by ultrasonic scanning.

[0022] Figure 3 Based on Figure 2 A schematic diagram of an embodiment for generating the corresponding wetting degree distribution state from the transmission image of the battery under test.

[0023] Figure 4 This is a schematic diagram of a second embodiment of the present invention, in which a transmission image of the battery under test is generated by ultrasonic scanning.

[0024] Figure 5 Based on Figure 4 A schematic diagram of an embodiment for generating the corresponding wetting degree distribution state from the transmission image of the battery under test. Detailed Implementation

[0025] The present invention will be further described below with reference to specific accompanying drawings and embodiments.

[0026] To enable real-time detection of electrolyte distribution within a battery using ultrasound, thereby improving detection efficiency and accuracy, this invention provides an ultrasonic detection method for electrolyte distribution within a battery. The ultrasonic detection method includes: Provide the battery to be tested. The battery under test is subjected to ultrasonic scanning to obtain a set of ultrasonic transmission signals corresponding to the ROI region. The ROI region includes at least the electrolyte wetting area in the battery under test. The set of ultrasonic transmission signals includes several ultrasonic transmission signals, and each ultrasonic transmission signal corresponds to a scanning detection point in the ROI region. Extract the corresponding transmission signal feature value for each ultrasound transmission signal to be tested, and use the pre-constructed feature-electrolyte content model to generate the electrolyte content solution value corresponding to the transmission signal feature value; Based on the electrolyte content calculated at all scanning detection points, the electrolyte distribution state in the ROI region of the battery under test is determined, so as to characterize the electrolyte wetting state in the ROI region of the battery under test using the electrolyte distribution state.

[0027] Figure 1 The illustration shows an embodiment of the present invention for ultrasonic testing of electrolyte distribution within a battery. As shown in the figure, when performing ultrasonic testing, a battery to be tested should be provided. The battery to be tested is the target object for electrolyte distribution testing according to the present invention. The battery to be tested should be a type of battery that requires electrolyte injection, that is, the type of battery to be tested should be a type that meets the requirements for electrolyte distribution testing. For example, the battery to be tested can be the aforementioned power battery. The types of batteries to be tested will not be listed here.

[0028] It should be understood that when performing ultrasonic testing on the electrolyte distribution within the battery under test, the battery should first be ultrasonically scanned. The method and process of ultrasonic scanning can be consistent with existing technologies. After ultrasonic scanning, ultrasonic scanning information of the battery under test can be obtained. This ultrasonic scanning information reflects the state of the battery after the ultrasonic signal passes through it. In other words, during ultrasonic scanning, a transmission scan should be performed on the battery under test. Therefore, the method of ultrasonic scanning should be designed to meet the requirements of transmission scan.

[0029] As explained above, after the electrolyte is injected into the battery under test, it will be distributed within a predetermined area, which is the desired electrolyte wetting area of ​​the battery under test. To improve the accuracy of electrolyte distribution detection, after obtaining the ultrasonic scanning information of the battery under test, a Region of Interest (ROI) should be selected based on the predetermined area within the battery. Subsequently, the ultrasonic transmission signal set corresponding to the ROI area is obtained. In specific implementation, commonly used techniques can be used to obtain the ultrasonic transmission signal set corresponding to the ROI area. For example, ultrasonic imaging can be performed based on the ultrasonic scanning information of the battery under test. Then, based on the desired electrolyte wetting area of ​​the battery under test, the required ROI area can be selected and determined, and the ultrasonic transmission signal set corresponding to the selected ROI area can be obtained. Generally, the ROI area can be consistent with or slightly larger than the desired electrolyte wetting area.

[0030] Similar to existing technologies, the ultrasonic transmission signal group to be tested includes several ultrasonic transmission signals to be tested. The number of ultrasonic transmission signals to be tested should be related to the size of the selected ROI region and the parameters selected for the ultrasonic scanning. Each ultrasonic transmission signal to be tested should correspond to a scanning detection point within the ROI region, which is a location point within the ROI region. Since the ROI region corresponds to the desired electrolyte region, a single ultrasonic transmission signal to be tested here can reflect the transmission state of the ultrasonic signal to the desired electrolyte region.

[0031] To determine the electrolyte distribution within the battery under test, the transmission signal characteristics of each ultrasonic transmission signal should be extracted. For each scanning detection point, the intensity of the ultrasonic signal passing through that point will vary depending on the wetting state. Therefore, the transmission signal characteristics should at least reflect the intensity characteristics of the ultrasonic signal passing through the corresponding scanning detection point, thus reflecting the electrolyte content at that point. Alternatively, the transmission signal characteristics can reflect the time dimension state characteristics of the ultrasonic signal passing through the corresponding scanning detection point. Specifically, when the transmission signal characteristics reflect the intensity characteristics of the ultrasonic signal passing through the corresponding scanning detection point, the transmission signal characteristics can be one of the peak-to-peak value (PPV), amplitude, attenuation coefficient, or envelope area of ​​the ultrasonic transmission signal under test. When the transmission signal characteristics reflect the time dimension state characteristics of the ultrasonic signal at the corresponding scanning detection point, the transmission signal characteristics can be the time of transit (TOF). Of course, other forms reflecting the frequency variation characteristics of the ultrasonic transmission signal can also be selected. The type of transmission signal characteristics can be chosen as needed, and will not be listed here.

[0032] It should be noted that when the transmitted signal characteristic is the PPV value, the PPV value is the amplitude of the maximum point of the transmitted signal waveform minus the amplitude of the corresponding minimum point. When the transmitted signal characteristic is the amplitude, the corresponding amplitude can be extracted by obtaining the maximum value of the ultrasonic transmitted signal under test. When the transmitted signal characteristic is the attenuation coefficient, it can be obtained using existing commonly used calculation methods. The attenuation coefficient reflects the attenuation characteristics of the battery under test for ultrasonic signals, and the unit of measurement is dB / cm / MHz. When the transmitted signal characteristic is the envelope area, the envelope area refers to the integral of the envelope curve over time. The envelope processing is consistent with existing technology and will not be elaborated here. The transit time refers to the total time it takes for the ultrasonic wave to travel from the transmitting transducer, through the battery under test, to being detected by the receiving transducer. The transit time can also be called the time of flight.

[0033] Depend on Figure 1As can be seen, after extracting the transmission signal features of each ultrasonic transmission signal to be inspected, the feature-electrolyte content model should be used to calculate the electrolyte content value corresponding to the transmission signal feature value. Generally, when inspecting the electrolyte distribution of the battery under inspection, the feature-electrolyte content model should be pre-constructed and generated. For each ultrasonic transmission signal to be inspected, the corresponding transmission signal feature corresponds to the electrolyte content of the corresponding scanning detection point. Therefore, the electrolyte content value corresponding to each transmission signal feature can be calculated using the feature-electrolyte content model. The method of constructing the feature-electrolyte content model and the method of calculating the electrolyte content value will be explained in detail below.

[0034] It should be understood that the above method can determine the electrolyte content at each scanning detection point. Subsequently, based on the electrolyte content at all scanning detection points, the electrolyte distribution state within the ROI region of the battery under test is determined, thus characterizing the electrolyte wetting state of the ROI region within the battery under test. Specifically, the electrolyte content can characterize the electrolyte content at the corresponding scanning detection point. Therefore, based on the electrolyte content at all scanning detection points, the electrolyte wetting state of the ROI region can be determined. Thus, the electrolyte wetting state primarily determines the electrolyte distribution in different areas within the ROI region, thereby achieving the detection of electrolyte distribution within the battery under test.

[0035] As explained above, for the battery under test, after performing ultrasonic scanning and acquiring the ultrasonic transmission signal group to be tested, the electrolyte content corresponding to each ultrasonic transmission signal to be tested can be calculated by constructing a feature-electrolyte content model. Based on all the electrolyte content calculations, the electrolyte distribution detection in the battery under test can be realized. The detection has good real-time performance and can improve the efficiency and accuracy of electrolyte distribution detection.

[0036] In one embodiment of the present invention, when the feature-electrolyte content model is constructed through data fitting, it includes: The mathematical model used in the feature-electrolyte content model is determined, and a fitting dataset is created based on the mathematical model used in the feature-electrolyte content model to fit the data. When creating a fitted dataset, the following steps are included: Provide several reference cells with known electrolyte injection amounts but not entirely identical electrolyte wetting distribution states; Each reference cell was subjected to ultrasonic scanning, and the reference ultrasonic transmission signal group corresponding to the ROI region within the reference cell was obtained. Based on the reference ultrasonic transmission signal sets of all reference cells and the electrolyte injection amount of each reference cell, the battery dataset for each reference cell is calculated, and a fitting dataset is formed based on the battery datasets of all reference cells. Each reference battery's battery dataset includes several fitted data samples. Each fitted data sample includes a reference electrolyte content per unit area and a transmission fitted feature value corresponding to the reference electrolyte content per unit area. The type of the transmission fitted feature value is consistent with the type of the transmission signal feature value. For each reference battery's battery dataset, the number of fitted data samples within the battery dataset is consistent, and the fitted data samples within all reference datasets have a consistent distribution.

[0037] In practical implementation, the characteristic-electrolyte content model can take the form of a mathematical model. In this case, the characteristic-electrolyte content model should be constructed using data fitting methods. Specifically, when constructing the characteristic-electrolyte content model using data fitting methods, the mathematical model used for the characteristic-electrolyte content model should be determined. After determining the mathematical model used for the characteristic-electrolyte content model, a fitting dataset adapted to the mathematical model should be created. Adaptation to the mathematical model specifically means that it can satisfy the requirement of data fitting to the mathematical model and achieve the goal of constructing the characteristic-electrolyte content model.

[0038] In one embodiment of the present invention, when creating the fitting dataset, multiple reference cells should be provided. For each reference cell, the amount of electrolyte injected is known, but the electrolyte wetting distribution should not be completely identical. The electrolyte wetting distribution within each reference cell can be referred to the above description. The amount of electrolyte injected specifically refers to the mass of the corresponding electrolyte injected into the reference cell. In practice, all reference cells should have slightly different electrolyte wetting distributions. For example, commonly used techniques can be used to determine the corresponding electrolyte wetting distribution for each reference cell, and reference cells with slightly different electrolyte wetting distributions can be selected.

[0039] After determining the reference battery, each reference battery should be ultrasonically scanned. The method of ultrasonic scanning the reference battery should be consistent with the method of ultrasonic scanning the battery under test, as detailed in the ultrasonic scanning instructions above. It is understood that after performing a transmission scan on the reference battery, the reference ultrasonic transmission signal set corresponding to the ROI region within the reference battery can be obtained using the above method. Here, the ROI region within the reference battery can be consistent with the ROI region of the battery under test; that is, the ROI region of the reference battery should also include the desired electrolyte wetting area within the reference battery. For each reference battery's reference ultrasonic transmission signal set, the reference ultrasonic transmission signal set should include several reference ultrasonic transmission signals. The details of the reference ultrasonic transmission signals can be found in the corresponding description of the ultrasonic transmission signals under test above, and will not be repeated here.

[0040] In practice, for each reference cell, a corresponding cell dataset can be generated based on the reference ultrasonic transmission signal set and the corresponding electrolyte injection volume. The required fitting dataset can then be created based on the cell datasets of all reference cells. Specifically, the method for generating the cell dataset is the same for each reference cell. Each cell dataset should include fitting data samples, and the number of fitting data samples should be the same across different reference cells.

[0041] In one embodiment of the present invention, each fitted data sample includes a reference electrolyte content per unit area and a transmission fitting feature value corresponding to the reference electrolyte content per unit area. Furthermore, the type of the transmission fitting feature value is consistent with the type of the transmission signal feature value. For example, if the transmission signal feature value uses a PPV value, then the transmission fitting feature value should also be the PPV value of the reference ultrasonic transmission signal. When the transmission signal feature value is of other types, the transmission fitting feature value should use the corresponding type. Examples will not be provided here. It should be understood that when generating the fitted data samples, since a reference electrolyte content per unit area and the corresponding transmission fitting feature value are used, the type of the reference battery may be different from the type of the battery under test. In specific implementation, the reference battery is preferably the same type as the battery under test. In addition, all fitted data samples within the reference dataset have a consistent distribution. The consistent distribution of the fitted data samples will be specifically explained below in conjunction with the method of generating the battery dataset.

[0042] In one embodiment of the present invention, solving the battery dataset for each reference battery includes: Based on the corresponding transmission reference feature value of each reference ultrasound transmission signal in the reference ultrasound transmission signal group, a reference feature distribution histogram of each reference cell is constructed to characterize the distribution state of the transmission reference feature value of the corresponding reference cell. The type of transmission reference feature value is consistent with the type of transmission signal feature value. For each reference cell, the histogram of reference feature distribution is divided into reference feature intervals, and the area of ​​each reference feature sub-interval is calculated. When dividing the reference feature intervals, the number of reference feature sub-intervals obtained by dividing each reference feature distribution histogram is the same, the reference feature interval ranges of the corresponding reference feature sub-intervals are consistent, and the number of reference feature sub-intervals obtained is no more than the number of reference batteries. For each reference cell, based on the area of ​​each reference feature sub-interval and the electrolyte injection amount of the reference cell, an electrolyte content distribution state equation is constructed to characterize the equivalent relationship between sub-interval area - electrolyte content per unit area - electrolyte injection amount. Based on the state equation of electrolyte content distribution of all reference cells, the electrolyte content per unit area of ​​each reference characteristic sub-interval is calculated. For each reference cell, the transmission fitting feature value of each reference feature sub-interval is determined, and a fitting data sample of the reference cell is formed based on the determined transmission fitting feature value and the corresponding electrolyte content per unit area.

[0043] It should be understood that when multiple reference batteries exist, a battery dataset for each reference battery should be calculated and generated separately. When generating the battery dataset for each reference battery, a reference feature distribution histogram for that reference battery should be constructed. In one embodiment of the present invention, the construction of the reference feature distribution histogram includes: The horizontal axis of the reference feature distribution histogram corresponds to the transmission reference feature value, and the vertical axis of the reference feature distribution histogram represents the number of each transmission reference feature value.

[0044] In practice, when constructing a histogram of reference feature distribution for each reference cell, the transmission reference feature value of each reference ultrasonic transmission signal belonging to the reference cell should be used. The meaning of the type of transmission reference feature value being consistent with the type of transmission signal feature value can be found in the above explanation of the consistency between the type of transmission fitting feature value and the type of transmission signal feature value, which will not be repeated here.

[0045] When constructing a reference feature distribution histogram, the horizontal axis of the histogram can be configured to correspond to the transmission reference feature value, while the vertical axis represents the number of each transmission reference feature value. For each reference cell, after determining the transmission reference feature value of each reference ultrasonic transmission signal, the number of such reference transmission feature values ​​should be counted, and the counted number should be used as the vertical axis of each reference transmission feature value. In specific implementation, the corresponding reference feature distribution histogram for each reference cell can be constructed using the same method.

[0046] After constructing the reference feature distribution histogram for each reference battery, reference feature intervals should be divided, that is, the range of values ​​for the abscissa of each reference feature distribution histogram should be divided into intervals. In one embodiment of the present invention, the reference feature interval division includes: Based on the transmission reference characteristic values ​​of all reference cells, the distribution range of the transmission reference characteristic values ​​is determined, and reference characteristic interval division parameters are configured based on the determined distribution range. The reference characteristic interval division parameters include at least the number of reference characteristic sub-intervals and the corresponding interval range of each reference characteristic sub-interval.

[0047] It should be noted that after determining the corresponding transmission reference feature value for each reference cell, the distribution range of the transmission reference feature value can be obtained. This distribution range is then used as the range of values ​​for the corresponding horizontal coordinates of all reference cells. In other words, for all reference cells, the range of coordinates corresponding to the horizontal coordinates is the same when constructing the corresponding reference feature distribution histogram. After determining the distribution range of the transmission reference feature value, specific reference feature intervals can be divided, such as determining the number of reference feature sub-intervals and the corresponding interval range for each reference feature sub-interval. Subsequently, each reference feature distribution histogram can be divided into reference feature intervals. For example, if there are 8 reference feature sub-intervals, the interval range of each reference feature sub-interval can be determined based on the distribution range of the transmission reference feature value. This allows for the division of each reference feature histogram. At this point, for the battery dataset of each reference cell, it can be guaranteed that the corresponding fitted data samples have a consistent distribution.

[0048] In practical implementation, when dividing the reference feature intervals, the number of reference feature sub-intervals remains consistent across all reference feature distribution histograms, and the corresponding interval ranges of the reference feature sub-intervals are also identical. That is, all reference feature distribution histograms are divided into reference feature intervals under the same reference feature interval division parameters. Furthermore, the number of reference feature sub-intervals should not exceed the number of reference cells. For example, if eight reference cells are provided, the number of reference feature intervals can be eight or less. The interval range of each reference feature sub-interval can generally be selected as needed. For instance, it can be evenly divided according to the number of reference feature intervals based on the distribution range of all transmission reference feature values. Of course, other methods can also be used to determine the interval range of each reference feature sub-interval; these will not be elaborated upon here.

[0049] For each reference cell, the area of ​​the corresponding reference feature sub-interval can be determined based on the number of transmitted reference feature values ​​within each reference feature sub-interval. Since the electrolyte injection amount of the reference cell is known, an electrolyte content distribution state equation characterizing the equivalent relationship between sub-interval area - electrolyte content per unit area - electrolyte injection amount is constructed, and its expression is: Where n is the number of reference feature intervals, For the first The area of ​​each reference feature subinterval, measured in mm. 2 (square millimeters) For the first The electrolyte content per unit area corresponding to each reference characteristic sub-interval, measured in g / mm². 2 (grams per square millimeter) The electrolyte injection amount for the current reference battery is measured in grams (g).

[0050] Since the corresponding range of each reference characteristic sub-interval is consistent, and the number of reference characteristic sub-intervals is no more than the number of reference cells, the electrolyte content per unit area corresponding to each reference characteristic sub-interval can be calculated using the commonly used calculation method based on the state equation of electrolyte content distribution of all reference cells. The specific calculation method and process can be consistent with the existing technology, and will not be elaborated here.

[0051] After determining the electrolyte content per unit area for each reference feature sub-interval, for each reference cell, a transmission fitting feature value for each reference feature sub-interval is determined. Based on the determined transmission fitting feature value and the corresponding electrolyte content per unit area, a fitting data sample for the reference cell is formed. For example, multiplying the determined transmission fitting feature value by the corresponding electrolyte content per unit area yields a fitting data sample. As explained above, the number of fitting data samples within the cell dataset should be consistent with the number of reference feature sub-intervals obtained above.

[0052] In one embodiment of the present invention, determining the transmission fitting feature value for each reference feature sub-interval includes: Based on the distribution of all transmission reference feature values ​​in the aforementioned reference feature sub-interval, transmission fitting feature values ​​are statistically generated, wherein... When generating transmission fitting feature values, the following are included: The average value of all transmission reference feature values ​​in the reference feature sub-interval is calculated, and the result of the average calculation is configured as the transmission fitting feature value. or, The median of all transmission reference feature values ​​within the reference feature sub-interval is statistically determined, and the statistically determined median is configured as the transmission fitting feature value.

[0053] Since the electrolyte wetting distribution of the reference cells is not completely identical, the corresponding transmission fitting characteristic values ​​of different reference cells in their respective reference characteristic sub-intervals may differ. Therefore, the transmission fitting characteristic value of each reference cell in all reference characteristic sub-intervals should be calculated. Thus, the transmission fitting characteristic value should correspond to the corresponding reference characteristic sub-interval for each reference cell. In specific implementation, when determining the transmission fitting characteristic value corresponding to a reference characteristic sub-interval, the average or median of all transmission reference characteristic values ​​within that sub-interval can be used as the corresponding transmission fitting characteristic value. Of course, other methods can also be used to determine the transmission fitting characteristic value of each reference characteristic sub-interval; these will not be illustrated here.

[0054] It should be noted that after creating / building the fitting dataset, existing commonly used techniques can be used to fit the mathematical model used in the feature-electrolyte content model. Methods such as least squares fitting can be used for data fitting, and the feature-electrolyte content model can be completed after fitting. The specific data fitting method can be selected as needed, and will not be listed here.

[0055] To improve the detection accuracy of electrolyte distribution, a double-weighted logistic composite model can be used for the feature-electrolyte content model. In one embodiment of this invention, when using the double-weighted logistic composite model, for the feature-electrolyte content model constructed through data fitting, the following applies:

[0056] in, These are characteristic values ​​of the transmitted signal. Electrolyte content per unit area , These are the weighting coefficients. , , , , , These are the parameters determined through data fitting.

[0057] Specifically, This is the first logistic term in the double-weighted logistic composite model. Parameter A is the magnitude of the first logistic term, determining the "plateau height" of its curve. A larger parameter A results in a higher value for the first logistic term. Parameter B is the rate of change of the first logistic term, controlling its steepness: B > 0 indicates an increasing logistic term, while B < 0 indicates a decreasing one. Parameter C is the inflection point of the first logistic term, determining its "center position." A larger parameter C indicates a more rightward (delayed) inflection point.

[0058] This is the second Logistic term in the double-weighted Logistic composite model. The parameters D, E, and F can be found in the descriptions of parameters A, B, and C above, and will not be repeated here.

[0059] It should be understood that when fitting the above-mentioned double-weighted Logistic composite model to the fitted dataset, the main focus is on determining the values ​​of parameters A to F. That is, the corresponding values ​​of parameters A to F are determined through data fitting. Therefore, once the values ​​of parameters A to F are determined, the feature-electrolyte content model is constructed. Regarding the weighting coefficients... Weighting coefficients The settings can be determined using methods such as experience.

[0060] During operation, for each ultrasonic transmission signal to be inspected, the corresponding transmission signal feature values ​​are first extracted. Then, the electrolyte content per unit area is calculated using the method described above. Subsequently, the calculated electrolyte content per unit area is multiplied by the pixel area corresponding to each scanning detection point to obtain the electrolyte content value corresponding to the current scanning detection point. It is understood that the pixel area corresponding to each scanning detection point is related to the ultrasonic scanning equipment, and the size of the pixel area corresponding to each scanning detection point can be determined using existing techniques, thereby allowing the calculation of the aforementioned electrolyte content value.

[0061] In one embodiment of the present invention, when constructing the feature-electrolyte content model based on machine learning, the method further includes: A machine learning-based electrolyte content prediction model is provided, and a training dataset is created for training the model. The training dataset includes several training samples, each training sample including a transmission training feature value and a label of electrolyte content per unit area corresponding to the transmission training feature value. Configure the model training conditions, and train the electrolyte content prediction model using the training dataset under the configured model training conditions until the target training state is reached. After that, configure the electrolyte content prediction model that has reached the target training state as the feature-electrolyte content model.

[0062] In practice, the feature-electrolyte content model can also be constructed using machine learning. When constructing the feature-electrolyte content model using machine learning, a machine learning-based electrolyte content prediction model should be provided. The electrolyte content prediction model can use existing commonly used neural network models. The neural network model used can be selected as needed, and will not be elaborated here.

[0063] Similar to existing technologies, after determining the electrolyte content prediction model, a training dataset should be provided for training the model. This training dataset includes several training samples, each in the same format, such as transmission training feature values ​​and corresponding unit area electrolyte content labels. The transmission training feature values ​​should correspond to the aforementioned transmission signal feature values, and are used only for model training. The details of the transmission training feature values ​​and their generation methods are explained above and will not be repeated here. For the unit area electrolyte content label, commonly used techniques can be employed to measure the corresponding label. The specific method for measuring the corresponding electrolyte content label will not be detailed here. Furthermore, it is understood that a feature-electrolyte content model obtained through mathematical fitting can also be used to generate the corresponding unit area electrolyte content label.

[0064] Similar to existing technologies, model training conditions should also be configured during model training. These conditions typically include necessary parameters such as the loss function, optimizer, number of training iterations, and training termination conditions. The specific details of the training conditions can be selected as needed to meet the required model training needs. The loss function can be a commonly used type, such as mean squared error. Once the target training state is reached, the electrolyte content prediction model that has reached the target training state is configured as a feature-electrolyte content model. Specifically, reaching the target training state may include reaching a set number of training iterations or a corresponding training termination condition. The specific conditions for reaching the target training state can be selected as needed and will not be elaborated further here.

[0065] During inference, for each ultrasonic transmission signal to be tested, the transmission signal feature value is extracted, and the extracted transmission signal feature value is loaded into the trained electrolyte content prediction model. The electrolyte content prediction model outputs the electrolyte content per unit area corresponding to the transmission signal feature value. Then, the electrolyte content calculation value corresponding to the current ultrasonic signal to be tested can be determined by the above method. The specific method and process for determining the corresponding electrolyte content calculation value can be referred to the above description, which will not be repeated here.

[0066] In one embodiment of the present invention, after obtaining the ultrasonic transmission signal group to be tested, a signal interpolation operation is further included, wherein... When performing signal interpolation, interpolation is performed based on the ultrasonic transmission signals to be tested at least two adjacent scanning detection points to generate the interpolated transmission signal to be tested. The electrolyte content corresponding to each interpolated transmission signal to be detected is generated using a feature-electrolyte content model. Based on the calculated electrolyte content at all scanning detection points and the calculated electrolyte content corresponding to all interpolated transmission signals to be tested, the electrolyte distribution status in the ROI region of the battery under test is determined.

[0067] It should be understood that, due to the limitations of ultrasonic scanning, the ultrasonic transmission signals within the test signal set cannot fully cover all points within the ROI region; that is, some ROI regions lack scanned detection points. To further improve the accuracy of ultrasonic detection of electrolyte distribution, after obtaining the test signal set for the battery under test, signal interpolation can be performed. During signal interpolation, interpolation is performed based on the test signal corresponding to any two adjacent scanned detection points. After interpolation, a test interpolated transmission signal can be generated. In other words, interpolating two adjacent test ultrasonic transmission signals generates a test interpolated transmission signal. The interpolation method can be consistent with existing technologies, such as linear interpolation. The interpolation method can be selected as needed and will not be elaborated here.

[0068] As explained above, a scanning detection point specifically refers to a location point in an ultrasonic scan where ultrasonic signals can pass through. Adjacent scanning detection points specifically refer to points that are adjacent within the ROI region, such as being adjacent in the same row or column. The adjacent status can be determined by the ultrasonic scanning process.

[0069] After generating the interpolated transmission signal to be tested, the corresponding transmission signal feature values ​​should be extracted. The method for extracting these feature values ​​can be found in the above description. After obtaining the corresponding transmission signal feature values, the corresponding electrolyte content per unit area should be generated using the feature-electrolyte content model. Subsequently, the corresponding electrolyte content is calculated. The specific method for obtaining the electrolyte content can be one of the methods mentioned above, as detailed in the above description.

[0070] After signal interpolation, the electrolyte distribution in the ROI region of the battery under test should be determined based on the calculated electrolyte content at all scanning detection points and the calculated electrolyte content corresponding to all interpolated transmission signals under test.

[0071] To output the electrolyte distribution in a quantitative or graphical manner, one embodiment of the present invention further includes generating a wetting degree distribution map of the ROI region, wherein... When generating the infiltration degree distribution map, the following steps are included: Determine the maximum value of all electrolyte solutions and normalize all electrolyte solutions using the maximum value to generate the corresponding wetting normalization value for each electrolyte solution. Based on all the normalized infiltration values, a state map of the infiltration degree distribution is generated.

[0072] For the reference battery, after obtaining multiple calculated electrolyte contents using the above method, the maximum value of the calculated electrolyte contents can be determined. Then, the maximum value of the calculated electrolyte contents can be used to normalize all calculated electrolyte contents to generate a corresponding normalized wetting value for each calculated electrolyte content. During normalization, each calculated electrolyte content can be divided by a normalized reference value, and the quotient is used as the corresponding normalized wetting value. The normalized reference value can be set as follows: specifically, the porosity of the corresponding material in the wetting region and the electrolyte capacity per unit volume of voids in the wetting region are obtained. Then, the obtained porosity and the electrolyte capacity per unit volume of voids in the wetting region are multiplied, and the result of the product is used as the normalized reference value.

[0073] It should be noted that porosity is an inherent characteristic of materials, generally calculated by dividing the volume of pores within the material by the total volume. For the electrolyte capacity per unit volume of voids, the volume of electrolyte that can be held per unit volume of the wetting region (i.e., the effective void volume fraction) should generally be calculated first, and then multiplied by the electrolyte density to obtain the capacity of the wetting region to hold electrolyte per unit volume of voids. The method for calculating the electrolyte capacity per unit volume of the wetting region can be consistent with existing technologies, such as by summing the volume of interparticle voids and the volume of voids within the particle structure per unit volume of the wetting region.

[0074] After obtaining the normalized values ​​of the calculated electrolyte content, wetting degree distribution map can be generated using existing technologies. The wetting degree distribution map can be used to determine the wetting state of different locations within the ROI region.

[0075] Figure 2 The diagram illustrates a first embodiment of generating a transmission map of a battery under test based on a set of ultrasonic transmission signals under test. Specifically, when generating the transmission map of the battery under test, the corresponding PPV value of the ultrasonic transmission signal under test is mainly used. The value on the right is the quantization range of the PPV value. The larger the value, the stronger the received ultrasonic signal, and the smaller the value, the weaker the received ultrasonic signal. The colors in the diagram correspond to the corresponding PPV values, which is a visualization mapping based on the PPV values. Figure 3 Based on Figure 2The transmission image of the battery under test is detected using the above method to generate a schematic diagram of the wetting degree distribution. Specifically, when generating the wetting degree distribution diagram, the corresponding wetting degree of the battery under test is used. The value on the right represents the quantitative range of the wetting degree. Low values ​​represent low electrolyte wetting degree and high values ​​represent high electrolyte wetting degree. The color in the diagram corresponds to the corresponding wetting degree. Specifically, it is based on the visualization mapping of the electrolyte wetting degree. Figure 2 and Figure 3 The column and row coordinates represent the column and row indices of the pixels within the battery ROI region, respectively. As explained above, this means... Figure 2 and Figure 3 The images show the distribution of PPV and immersion level in the pixel coordinate system within the battery ROI area.

[0076] Figures 4-5 Each of these forms a set of transmission maps of the battery under test and a corresponding wetting degree distribution map. The difference is... Figures 2-3 The corresponding degree of infiltration is relatively low. Figures 4-5 The diagram shows the situation when the degree of wetting is high. For other situations, please refer to the corresponding descriptions above, which will not be elaborated here.

[0077] In practice, after obtaining the calculated electrolyte content value corresponding to each scanning detection point, the wetting state analysis of the battery under test can also be performed. This wetting state analysis includes: Based on the calculated electrolyte content value corresponding to each scanning detection point, the electrolyte wetting amount in the strong region corresponding to the strong ultrasonic signal region is statistically analyzed. Based on the actual amount of electrolyte injected into the battery under test and the amount of electrolyte wetting in the area with strong ultrasonic signal, the amount of electrolyte wetting in the weak area corresponding to the area with weak ultrasonic signal is calculated. The area of ​​the corresponding weak region in the weak ultrasonic signal region is statistically analyzed. Then, based on the electrolyte wetting amount in the weak region and the corresponding weak region area, the electrolyte content per unit area of ​​the weak region is calculated. Based on the electrolyte content per unit area in the weak region calculated above, the wetting state of the battery under test is judged. Specifically, when the electrolyte content per unit area in the weak region is greater than the electrolyte wetting threshold in the weak region, it is determined that air bubbles exist in the ultrasonic weak signal area. When the electrolyte content per unit area in the weak region is not greater than the electrolyte wetting threshold in the weak region, it is determined that there is insufficient wetting in the weak ultrasonic signal region.

[0078] In practical implementation, the strong ultrasonic signal region generally refers to the area where the calculated electrolyte content is greater than the strong electrolyte wetting threshold, corresponding to the scanning detection point. The strong electrolyte wetting threshold can generally be set based on the electrolyte content corresponding to the minimum ultrasonic response of the wetting region. For example, based on the statistical value of the calculated electrolyte content, the calculated electrolyte content corresponding to a normalized wetting value greater than 40% can be used as the strong electrolyte wetting threshold. Of course, other methods can also be used to set the strong electrolyte wetting threshold based on statistical results; specific settings will not be illustrated here. After determining the strong electrolyte wetting threshold, all calculated electrolyte content values ​​greater than the strong electrolyte wetting threshold are summed to obtain the strong region electrolyte wetting amount. It can be understood that the strong region electrolyte wetting amount here represents the total electrolyte wetting amount within the strong region, i.e., the total mass or volume of electrolyte wetting within the strong region.

[0079] Since the electrolyte injection amount for each battery under test is known, the electrolyte wetting amount in the weak region can be calculated after statistically determining the electrolyte wetting amount in the strong region. This can be achieved by subtracting the actual electrolyte injection amount from the statistically determined electrolyte wetting amount in the strong region; the difference is the electrolyte wetting amount in the weak region. Specifically, the electrolyte wetting amount in the weak region here refers to the electrolyte wetting amount in the ultrasonic weak signal region. The ultrasonic weak signal region is one of two regions opposite to the ultrasonic strong signal region. Therefore, based on the above description of the ultrasonic strong signal region, the corresponding meaning of the ultrasonic weak signal region can be determined. It can be understood that after determining the meaning of the ultrasonic weak signal region, the area of ​​the weak region can be statistically determined using the above method. The area of ​​the weak region is also the sum of the pixel areas corresponding to the scanning detection points within the ultrasonic weak signal region.

[0080] In practice, the electrolyte wetting amount in the weak region is divided by the area of ​​the weak region, and the corresponding quotient is the electrolyte content per unit area of ​​the weak region, i.e., W2. After obtaining the electrolyte content per unit area of ​​the weak region, it can be compared with the electrolyte wetting threshold of the weak region.

[0081] In practical implementation, when setting the electrolyte weak area wetting threshold, one feasible approach is as follows: After determining the aforementioned weak ultrasonic signal region, for each scanning detection point within the weak ultrasonic signal region, the corresponding electrolyte content calculation value is extracted. Then, based on all the electrolyte content calculation values ​​within the weak ultrasonic signal region and the corresponding pixel areas of all scanning detection points, the corresponding arithmetic mean is calculated. This calculated arithmetic mean is used as the electrolyte weak region wetting threshold, i.e., W1. Of course, other methods can also be used to set the electrolyte weak region wetting threshold; specific setting methods will not be illustrated here.

[0082] In practice, W2 represents the electrolyte content per unit area in the weak region, and W1 represents the electrolyte wetting threshold in the weak region. When W2 > W1, it is determined that the weak ultrasonic signal region contains gas; when W2 ≤ W1, the weak ultrasonic signal region is insufficiently wetting but does not contain gas.

[0083] It should be understood that when an ultrasonic signal propagates inside a battery, it will be strongly scattered and attenuated if it encounters air bubbles, resulting in a weaker received signal. Consequently, the calculated electrolyte content value for a given scanning point will be lower than expected. Therefore, weak ultrasonic signal areas may be caused by two reasons: first, insufficient actual electrolyte, i.e., inadequate wetting; second, the presence of air bubbles, i.e., sufficient actual electrolyte wetting but the signal is blocked by air bubbles.

[0084] It should be noted that the electrolyte wetting amount in the weak region is the actual electrolyte amount separated from the total injected amount. It is not affected by bubble attenuation. Therefore, the electrolyte content per unit area W2 in the weak region, obtained by dividing this actual total amount by the area of ​​the weak region, is the true average content characterizing the electrolyte content per unit area in the weak region. The electrolyte wetting threshold W1 in the weak region is the arithmetic mean of the electrolyte content per unit area calculated from the detection points within the weak region; it is the measured average content characterizing the electrolyte content per unit area in the weak region. Specifically, in comparison: if W2 > W1, it indicates that the true average content is higher than the measured average, meaning the measured value is underestimated. This is likely due to signal attenuation caused by bubbles; therefore, it is determined that bubbles exist in the weak ultrasonic signal region. If W2 ≤ W1, it indicates that the true average content is not higher than the measured average, and the measured value is not significantly underestimated; therefore, it is determined to be insufficient wetting (no bubbles).

[0085] When determining that a weak ultrasonic signal region contains gas, further judgments can be made based on other conditions: if W2 / W1 ≥ 2, the weak ultrasonic signal region contains a large amount of gas; if 2 > W2 / W1 ≥ 1.2, the weak ultrasonic signal region contains a considerable amount of gas; if 1.2 > W2 / W1 ≥ 1, the weak ultrasonic signal region contains a small amount of gas. It should be noted that the values ​​of 2 and 1.2 mentioned above are merely exemplary reference thresholds. In practical engineering applications, these thresholds need to be calibrated and adjusted according to the specific battery design, electrolyte characteristics, and detection accuracy. Determining a reasonable threshold for a specific battery system usually requires first conducting a set of calibration experiments (such as electrolyte injection gradient experiments or gas content gradient experiments) to fit an accurate judgment boundary. This part belongs to existing technology and will not be elaborated here.

[0086] When it is determined that gas is present in the area of ​​weak ultrasonic signal, a more detailed classification can be performed based on the ratio to further analyze the gas content. The larger the W2 / W1 ratio, the greater the degree to which the measured value is underestimated, that is, the more severe the attenuation of the ultrasonic signal by bubbles, thus indirectly reflecting the number or size of bubbles. This classification helps to quantitatively evaluate the quality of the impregnation process.

Claims

1. An ultrasonic detection method for electrolyte distribution within a battery, characterized in that, The ultrasonic testing method includes: Provide the battery to be tested. The battery under test is subjected to ultrasonic scanning to obtain a set of ultrasonic transmission signals corresponding to the ROI region. The ROI region includes at least the electrolyte wetting area in the battery under test. The set of ultrasonic transmission signals includes several ultrasonic transmission signals, and each ultrasonic transmission signal corresponds to a scanning detection point in the ROI region. Extract the corresponding transmission signal feature value for each ultrasound transmission signal to be tested, and use the pre-constructed feature-electrolyte content model to generate the electrolyte content solution value corresponding to the transmission signal feature value; Based on the electrolyte content calculated at all scanning detection points, the electrolyte distribution state in the ROI region of the battery under test is determined, so as to characterize the electrolyte wetting state in the ROI region of the battery under test using the electrolyte distribution state.

2. The ultrasonic detection method for electrolyte distribution within a battery according to claim 1, characterized in that, The characteristic-electrolyte content model, when constructed through data fitting, includes: The mathematical model used in the feature-electrolyte content model is determined, and a fitting dataset is created based on the mathematical model used in the feature-electrolyte content model to fit the data. When creating a fitted dataset, the following steps are included: Provide several reference cells with known electrolyte injection amounts but not entirely identical electrolyte wetting distribution states; Each reference cell was subjected to ultrasonic scanning, and the reference ultrasonic transmission signal group corresponding to the ROI region within the reference cell was obtained. Based on the reference ultrasonic transmission signal sets of all reference cells and the electrolyte injection amount of each reference cell, the battery dataset for each reference cell is calculated, and a fitting dataset is formed based on the battery datasets of all reference cells. Each reference battery's battery dataset includes several fitted data samples. Each fitted data sample includes a reference electrolyte content per unit area and a transmission fitted feature value corresponding to the reference electrolyte content per unit area. The type of the transmission fitted feature value is consistent with the type of the transmission signal feature value. For each reference battery's battery dataset, the number of fitted data samples within the battery dataset is consistent, and the fitted data samples within all reference datasets have a consistent distribution.

3. The ultrasonic detection method for electrolyte distribution within a battery according to claim 2, characterized in that, Solving the battery dataset for each reference battery includes: Based on the corresponding transmission reference feature value of each reference ultrasound transmission signal in the reference ultrasound transmission signal group, a reference feature distribution histogram of each reference cell is constructed to characterize the distribution state of the transmission reference feature value of the corresponding reference cell. The type of transmission reference feature value is consistent with the type of transmission signal feature value. For each reference cell, the histogram of reference feature distribution is divided into reference feature intervals, and the area of ​​each reference feature sub-interval is calculated. When dividing the reference feature intervals, the number of reference feature sub-intervals obtained by dividing each reference feature distribution histogram is the same, the reference feature interval ranges of the corresponding reference feature sub-intervals are consistent, and the number of reference feature sub-intervals obtained is no more than the number of reference batteries. For each reference cell, based on the area of ​​each reference feature sub-interval and the electrolyte injection amount of the reference cell, an electrolyte content distribution state equation is constructed to characterize the equivalent relationship between sub-interval area - electrolyte content per unit area - electrolyte injection amount. Based on the state equation of electrolyte content distribution of all reference cells, the electrolyte content per unit area of ​​each reference characteristic sub-interval is calculated. For each reference cell, the transmission fitting feature value of each reference feature sub-interval is determined, and a fitting data sample of the reference cell is formed based on the determined transmission fitting feature value and the corresponding electrolyte content per unit area.

4. The ultrasonic detection method for electrolyte distribution within a battery according to claim 3, characterized in that, Within the reference feature distribution histogram, we have: The horizontal axis of the reference feature distribution histogram corresponds to the transmission reference feature value, and the vertical axis of the reference feature distribution histogram represents the number of each transmission reference feature value. When dividing the reference feature interval, the following is included: Based on the transmission reference characteristic values ​​of all reference cells, the distribution range of the transmission reference characteristic values ​​is determined, and reference characteristic interval division parameters are configured based on the determined distribution range. The reference characteristic interval division parameters include at least the number of reference characteristic sub-intervals and the corresponding interval range of each reference characteristic sub-interval.

5. The ultrasonic detection method for electrolyte distribution within a battery according to claim 3, characterized in that, When determining the transmission fitting eigenvalues ​​for each reference feature sub-interval, the following steps are included: Based on the distribution of all transmission reference feature values ​​in the aforementioned reference feature sub-interval, transmission fitting feature values ​​are statistically generated, wherein... When generating transmission fitting feature values, the following are included: The average value of all transmission reference feature values ​​in the reference feature sub-interval is calculated, and the result of the average calculation is configured as the transmission fitting feature value. or, The median of all transmission reference feature values ​​within the reference feature sub-interval is statistically determined, and the statistically determined median is configured as the transmission fitting feature value.

6. The ultrasonic detection method for electrolyte distribution within a battery according to claim 2, characterized in that, For the feature-electrolyte content model constructed through data fitting, then: in, These are characteristic values ​​of the transmitted signal. Electrolyte content per unit area , These are the weighting coefficients. , , , , , These are the parameters determined through data fitting.

7. The ultrasonic detection method for electrolyte distribution within a battery according to claim 1, characterized in that, When constructing the feature-electrolyte content model based on machine learning, it also includes: A machine learning-based electrolyte content prediction model is provided, and a training dataset is created for training the model. The training dataset includes several training samples, each training sample including a transmission training feature value and a label of electrolyte content per unit area corresponding to the transmission training feature value. Configure the model training conditions, and train the electrolyte content prediction model using the training dataset under the configured model training conditions until the target training state is reached. After that, configure the electrolyte content prediction model that has reached the target training state as the feature-electrolyte content model.

8. The ultrasonic detection method for electrolyte distribution in a battery according to any one of claims 1 to 7, characterized in that, in After obtaining the calculated electrolyte content value corresponding to each scanning detection point, the wetting state of the battery under test is analyzed. When performing infiltration state analysis, the following are included: Based on the calculated electrolyte content value corresponding to each scanning detection point, the electrolyte wetting amount in the strong region corresponding to the strong ultrasonic signal region is statistically analyzed. Based on the actual amount of electrolyte injected into the battery under test and the amount of electrolyte wetting in the area with strong ultrasonic signal, the amount of electrolyte wetting in the weak area corresponding to the area with weak ultrasonic signal is calculated. The area of ​​the corresponding weak region in the weak ultrasonic signal region is statistically analyzed. Then, based on the electrolyte wetting amount in the weak region and the corresponding weak region area, the electrolyte content per unit area of ​​the weak region is calculated. Based on the electrolyte content per unit area in the weak region calculated above, the wetting state of the battery under test is judged. Specifically, when the electrolyte content per unit area in the weak region is greater than the electrolyte wetting threshold in the weak region, it is determined that air bubbles exist in the ultrasonic weak signal area. When the electrolyte content per unit area in the weak region is not greater than the electrolyte wetting threshold in the weak region, it is determined that there is insufficient wetting in the weak ultrasonic signal region.

9. The ultrasonic detection method for electrolyte distribution within a battery according to any one of claims 1 to 7, characterized in that, After obtaining the set of ultrasonic transmission signals to be examined, the process also includes signal interpolation, in which... When performing signal interpolation, interpolation is performed based on the ultrasonic transmission signals to be tested at least two adjacent scanning detection points to generate the interpolated transmission signal to be tested. The electrolyte content corresponding to each interpolated transmission signal to be detected is generated using a feature-electrolyte content model. Based on the calculated electrolyte content at all scanning detection points and the calculated electrolyte content corresponding to all interpolated transmission signals to be tested, the electrolyte distribution status in the ROI region of the battery under test is determined.

10. The ultrasonic detection method for electrolyte distribution within a battery according to any one of claims 1 to 7, characterized in that, It also includes a distribution map of the infiltration degree of the ROI region, in which, When generating the infiltration degree distribution map, the following steps are included: Determine the maximum value of all electrolyte solutions and normalize all electrolyte solutions using the maximum value to generate the corresponding wetting normalization value for each electrolyte solution. Based on all the normalized infiltration values, a state map of the infiltration degree distribution is generated.

Images