Solid-state battery thermal fault diagnosis and online early warning method

Abstract

The invention discloses a solid-state battery thermal fault diagnosis and online early warning method, and relates to the technical field of battery fault detection, and the method comprises the steps: collecting a solid-state battery temperature distribution image, and generating an image feature parameter and a temperature gradient map; according to the quantized image characteristic parameters and the environment parameters, obtaining a solid-state battery use state evaluation value; and setting a screening temperature, and executing a corresponding strategy according to the distribution area of the high-temperature image. According to the solid-state battery thermal fault diagnosis and on-line early warning method, the image characteristic parameters of the battery in the use state are detected and analyzed in a thermal imaging mode to obtain the state evaluation value Bauyt of the battery, then the state evaluation value Bauyt is combined with the detection result obtained by subsequent execution of the detection strategy, and the state evaluation value Bauyt and the detection result obtained by subsequent execution of the detection strategy are combined to obtain the thermal fault diagnosis and on-line early warning method of the solid-state battery. Therefore, the adaptive capacity of the battery is comprehensively predicted, a corresponding battery use mode is conveniently formulated according to the bearing capacity of the battery, and the remaining service life of the battery can be fully utilized.

Description

Technical Field

[0001] The present invention relates to the technical field of battery fault detection, and in particular to a solid-state battery thermal fault diagnosis and online early warning method. Background Art

[0002] Lithium-ion batteries are one of the most widely used rechargeable batteries. Their high energy density, long life and lightweight advantages make them widely used in electric vehicles, portable electronic devices and energy storage systems. However, the normal operation of lithium-ion batteries will be affected by temperature. Excessive charging and discharging, internal and external short circuits, mechanical damage and long-term operation in high temperature environments may cause thermal runaway of the battery and even cause fire or explosion. Therefore, equipping lithium-ion batteries with appropriate thermal management systems and performing real-time online temperature diagnosis can avoid the occurrence of irreversible thermal runaway and ensure the safe and reliable operation of lithium-ion batteries.

[0003] At present, the existing Chinese patent application number is: "CN202310041018.5" discloses a battery thermal runaway fault detection method and system. At high temperatures, the decomposition of battery materials will lead to gas generation and battery expansion. The battery structure is designed to rupture and discharge the accumulated electrolytic gas at a critical gas pressure to prevent explosive force. Although this patent can detect the battery expansion force and gas to obtain the type of short-circuit fault in the battery, it is inaccurate to judge only by "expansion". Taking mobile phones as an example, during the use of existing mobile phone batteries, although the internal ions are constantly transformed, the gas and expansion that can be generated when the battery is working are negligible. When we find that the mobile phone battery has faults such as "bulging", it means that the battery damage degree has reached more than 60% and needs to be stopped. For electric vehicles, they cannot be stopped immediately when a fault occurs, which causes users to drive vehicles with safety hazards for a period of time. In addition, the existing Chinese patent application number is: "CN202211467164.6" discloses a power lithium battery thermal runaway fault classification and risk prediction method and system, including: a module-level power battery model fault injection method, a random fault generation and labeling method, a power lithium-ion battery fault multi-classification model based on a deep learning method, and a transfer learning method for applying the model to a real vehicle. Although this application can diagnose the battery by simulating automobile usage, in actual use, it is often necessary to make a prediction in advance. Moreover, due to different battery heating conditions, deviations are likely to occur if the same detection method is used for testing.

[0004] However, during the implementation of the above technical solution, it was found that there were at least the following technical problems:

[0005] Most existing battery thermal fault detection systems use overall calculations, that is, they need to be executed from beginning to end according to a predetermined program, and during the calculation process, normal batteries or circuits will also be tested, resulting in each test taking a lot of time to calculate. Secondly, since it is impossible to ensure that all influencing factors in the battery production process remain consistent, and the actual use environment is also different, the tolerance of each battery cell is also dynamically changing. Therefore, only the data obtained through the actual use of the battery can best reflect the battery situation;

[0006] There are two common types of battery thermal failures. One is battery core failure. During the use of the battery, chemical reactions are constantly taking place inside the battery core, causing electrical energy to be lost in the form of heat. The other is line failure. When a short circuit occurs in the line, it will not only damage the circuit but also generate a large amount of heat energy. Regardless of the type of thermal failure, it will eventually manifest itself in the form of heat, that is, a large amount of heat energy will be generated at the location where the failure occurs. Therefore, the location of the thermal failure can be preliminarily determined by detecting the distribution of heat, which is convenient for subsequent inspections. To this end, we propose a solid-state battery thermal failure diagnosis and online early warning method. Summary of the invention

[0007] 1. Technical issues to be solved

[0008] In view of the shortcomings of the prior art, the present invention provides a solid-state battery thermal fault diagnosis and online early warning method, which uses thermal imaging to detect the indicated temperature, temperature change rate and temperature diffusion rate of the battery under use, and analyzes to obtain a state evaluation value Bauy that can intuitively judge the actual tolerance of the battery. t , and then according to the state evaluation value Bauy t The position relationship with the state threshold interval can be used to call the corresponding solution, thereby providing a basis for battery replacement and adjustment. It can also extend the battery life when the replacement cycle is missed, providing sufficient time for battery replacement. Secondly, the state evaluation value Bauy t Combined with the detection results obtained by the subsequent execution of the detection strategy, the adaptive estimate Bauy is obtained x , thereby comprehensively predicting the adaptability of the battery and facilitating the formulation of a corresponding battery usage mode based on the battery's tolerance, thereby making full use of the remaining service life of the battery and solving the technical problems of low computational efficiency and inaccurate prediction results in the existing solid-state battery thermal fault diagnosis method.

[0009] (II) Technical solution

[0010] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0011] A solid-state battery thermal fault diagnosis and online early warning method, the method comprising:

[0012] After normalizing the acquired solid-state battery temperature distribution image, it is input into the feature extraction network to generate image feature parameters and temperature gradient map, where the image feature parameters include the indicated temperature, temperature change rate and temperature diffusion speed of each area;

[0013] According to the quantified image feature parameters and the environmental parameters of the location of the solid-state battery, the solid-state battery usage status evaluation value is obtained, and then the current usage of the solid-state battery is evaluated according to the position of the solid-state battery usage status evaluation value in the preset state threshold interval;

[0014] Set a screening temperature, delete the images in the temperature gradient map that are lower than the screening temperature, record the remaining images as high-temperature images, mark the locations of the high-temperature images in the pre-recorded solid-state battery internal structure distribution map, and execute the corresponding strategy according to the distribution area of ​​the high-temperature images;

[0015] When the high-temperature image is distributed at the location of the battery cell, a maintenance warning is issued and the battery detection strategy is executed;

[0016] When high temperature images are distributed at the location of the line, a maintenance warning is issued and the line detection strategy is executed;

[0017] When the high-temperature image is distributed in a location other than the circuit or the battery core in the solid-state battery, the temperature distribution image of the solid-state battery during use is collected again, and the solid-state battery use status assessment values ​​obtained from the two collections are compared. If the areas where the results of the two collections are located are consistent, no response is given, otherwise a verification warning is issued;

[0018] Combined with the health prediction coefficient and the solid-state battery usage status assessment value obtained in the detection strategy process, the adaptability estimation value of the solid-state battery under the current environmental parameters and the solid-state battery thermal failure prediction value corresponding to the battery adaptability estimation value are calculated.

[0019] Furthermore, the process of obtaining image feature parameters is as follows:

[0020] According to the preset image segmentation size, the temperature distribution image is divided into n image blocks of the same size, and m image points are randomly selected in the same image block, the RGB values ​​corresponding to the image points are retrieved, and the temperature ratio matching the RGB values ​​of the image points is obtained from the database. The obtained value is recorded as the indicated temperature Qc of the corresponding image block;

[0021] The formula for generating the indicated temperature Qc of the corresponding image block is as follows:

[0022]

[0023] In the formula, Del represents the average color coefficient of all image points in the same image block, Qed x Orn represents the preset temperature ratio, Orn represents the exponent of the temperature ratio, δ represents the correction factor of the average color perception coefficient, ψ Del is the battery temperature influence coefficient during detection, λ Del is the environmental temperature influence coefficient during detection, ε1 and ε2 represent the preset battery temperature influence coefficient and the weight of the battery temperature influence coefficient respectively, ε1+ε2=1, R i , G i , B i Represents the RGB value of the i-th image point, Rcl j , Glid j 、Bibo j Represents the weight corresponding to the RGB value of the i-th image point, 0<i≤m, e is a natural constant;

[0024]

[0025] Where W q Indicates the temperature change rate of the same image block, T Q Indicates the preset detection time, Qc t represents the indicated temperature at the end of the test time, and Qc0 represents the indicated temperature at the beginning of the test time;

[0026] After the indicated temperature calculation is completed, multiple image blocks with the same indicated temperature are randomly selected and marked as the starting image block. The temperature at this time is recorded as the approved temperature. Then the indicated temperature of the image blocks around the starting image block is obtained, and the image blocks with a temperature lower than the indicated temperature are recorded as the image blocks to be tested. The time T when the indicated temperature of the image blocks to be tested reaches the approved temperature is recorded. L ;

[0027]

[0028] In the formula, B t represents the temperature change rate, and L represents the distance between two adjacent image blocks.

[0029] Furthermore, the environmental parameters of the location of the solid-state battery include air temperature, air humidity, and air thermal conductivity, and the impact of the external environment of the solid-state battery is analyzed. The specific analysis process is as follows:

[0030] Extract the environmental adaptability information of the solid-state battery stored in the database, obtain the environmental parameter range that the solid-state battery can tolerate, and further obtain the temperature range, humidity range and air thermal conductivity range that the solid-state battery can tolerate;

[0031] The temperature range that the solid-state battery can withstand is compared with the maximum indicated temperature obtained during the actual test process, and the adaptation coefficient between the solid-state battery's temperature tolerance and the maximum indicated temperature is analyzed and recorded as

[0032] Similarly, according to the analysis method of the matching coefficient between the solid-state battery tolerance temperature and the maximum indicated temperature, the adaptation coefficients of the solid-state battery tolerance humidity and tolerance air thermal conductivity to the humidity and air thermal conductivity in the actual detection process are obtained, and they are recorded as

[0033] By analyzing the formula Get the environmental parameter impact value Eltq of solid-state batteries t , where φ represents the correction factor of the preset environmental parameter impact value, W t , S t 、Dr t represent the preset weight factors of temperature, humidity and air thermal conductivity, respectively, and W t >S t >Dr t >0.

[0034] Further, the solid-state battery usage status evaluation value Bauy is generated t The formula is as follows:

[0035]

[0036] In the formula, Qc max Indicates the maximum temperature of all indicated temperatures in the temperature distribution image, Qc x represents the indicated temperature of the x-th image block, x is a positive integer, a1, a2, a3 represent the weights of the indicated temperature, temperature diffusion coefficient and environmental parameter influence coefficient respectively, a1+a2+a3=1, G1 represents the correction coefficient.

[0037] Furthermore, the state threshold interval includes the state threshold upper limit Bauy ↑ and the lower threshold of the state Bauy ↓ ;

[0038] When the state evaluation value Bauy t ≥ Upper threshold of status Bauy ↑ If , no response is given;

[0039] When the state threshold is upper Bauy ↑ >Status evaluation value Bauy t ≥Bauy lower threshold ↓ When , a replacement instruction is issued;

[0040] When the state threshold is lowered ↓>Status evaluation value Bauy t When the status evaluation value Bauy is reached, a replacement command is issued and the t The corresponding low-power output mode limits the maximum output voltage and maximum output current of the solid-state battery.

[0041] Furthermore, the specific process of executing the battery detection strategy is as follows:

[0042] Set the duration of the monitoring cycle, obtain the maximum and minimum values ​​of the current and voltage during the solid-state battery monitoring cycle, and obtain the rate of change of the current and voltage during the solid-state battery monitoring cycle;

[0043] Connect a preset detection load to the circuit, re-monitor according to the preset monitoring period, and record the rate of change of current and voltage during the monitoring period;

[0044] According to the data obtained in the two monitoring cycles, the formula Get the estimated health coefficient K of the solid battery cell 电池 , where maxI0, minI0, maxI1, minI1 represent the maximum and minimum current values ​​in the monitoring period before and after the load is connected, maxU0, minU0, maxU1, minU1 represent the maximum and minimum voltage values ​​in the monitoring period before and after the load is connected, T d It represents the duration of the preset monitoring cycle, Ies and Upg represent the weight factors of the current change rate and the voltage change rate respectively, Ies>0, Upg>0, and G2 represents the correction coefficient.

[0045] Furthermore, the specific process of executing the line detection strategy is as follows:

[0046] Input the preset current and voltage to the line and keep T s After a certain time, record the average temperature of the area where the circuit is located in the solid battery;

[0047] By applying the analytical formula Get the line health estimation coefficient C of the solid-state battery 线路 , where θ represents the correction factor of the preset line health estimate, I q ,I b Respectively represent the line input current before and after adjustment, U q , U b They represent the line input voltages before and after adjustment respectively, Ice and Uce represent the weight factors of current and voltage respectively, and Ice+Uce=1.

[0048] Furthermore, when a verification warning is received, the shooting frequency of the solid-state battery temperature distribution image is adjusted, and the adjusted shooting frequency must be 5 times or more than the original shooting frequency; the coordinates of the area block corresponding to the maximum temperature indicated in the solid-state battery temperature distribution image are obtained, and all the coordinates are substituted into the two-dimensional coordinate axis to generate a coordinate change curve; the obtained coordinate change curve is compared with the historical coordinate change curve in the database; wherein, when the degree of overlap between the two curves reaches more than 80%, no response is given; otherwise, a maintenance instruction is issued.

[0049] Furthermore, the formula for generating the adaptive estimate is as follows:

[0050]

[0051] In the formula, Bauy x represents the adaptive estimate, Bauy x ≥0, when Bauy x When <0, the temperature distribution image of the solid-state battery is collected;

[0052] The formula used to generate the solid-state battery thermal failure prediction value is as follows:

[0053] Waty t =Bauy x ×γ-Eltq t ×μ

[0054] In the formula, Waty t Represents the solid-state battery thermal failure prediction value, γ represents the proportional coefficient of the adaptability estimation value, μ represents the proportional coefficient of the environmental parameter influence value, and is compared with the solid-state battery thermal failure prediction value obtained last time. When the difference is less than the prediction difference interval, the result is output; otherwise, the solid-state battery temperature distribution image is re-collected.

[0055] Further, an evaluation threshold interval of the thermal fault prediction value is set, and the evaluation threshold interval includes a qualified threshold interval, a risk threshold interval, and a temperature drop threshold interval;

[0056] Wherein, when the thermal fault prediction value is within the qualified threshold range, no response is made and the result is output;

[0057] When the fault prediction value is within the risk threshold range, a maintenance instruction is issued and the auxiliary cooling system is turned on to cool the solid-state battery;

[0058] When the thermal fault prediction value is within the temperature reduction threshold range, a danger warning is issued and the auxiliary cooling system is turned on to cool the solid-state battery. At the same time, the output power of the solid-state battery is reduced until the indicated temperature on the surface of the solid-state battery no longer increases.

[0059] (III) Beneficial effects

[0060] 1. The present invention uses thermal imaging to detect the indicated temperature, temperature change rate and temperature diffusion rate of the battery under use, and analyzes to obtain a state evaluation value Bauy that can intuitively judge the actual tolerance of the battery. t , and then according to the state evaluation value Bauy t The position relationship with the state threshold interval can be used to call the corresponding solution, thereby providing a basis for battery replacement and adjustment. It can also extend the battery life when the replacement cycle is missed, providing sufficient time for battery replacement. Secondly, the state evaluation value Bauy t Combined with the detection results obtained by the subsequent execution of the detection strategy, the adaptive estimate Bauy is obtained x , thereby comprehensively predicting the adaptability of the battery and facilitating the formulation of a corresponding battery usage mode based on the battery's tolerance, thereby making full use of the remaining battery life.

[0061] 2. The present invention detects the distribution of the indicated temperature, finds the heat source of the battery, and then calls and executes the corresponding detection strategy according to the location of the heat source, which can improve the detection efficiency and the accuracy of the calculation results. In addition, the results and state evaluation values ​​obtained by executing the detection strategy can be calculated. t Comprehensively predict the battery life.

[0062] 3. The present invention obtains the battery's tolerance range to temperature, humidity and air thermal conductivity, and calculates the corresponding adaptation coefficient and environmental parameter impact value Eltq t , thereby judging the degree of impact of the environment on the battery and the adaptability of the battery in the current environment, which is not only convenient for subsequent status evaluation t The calculation excludes environmental interference and combines the environmental parameter influence value Eltq recorded in historical records t , reminding users of the impact of the current environment on the battery, making it easier for users to understand intuitively, and can be coordinated with the vehicle system for regulation, thereby increasing the battery life.

[0063] In summary, the system designed in the present invention can realize a comprehensive evaluation and analysis of the battery pack, improve the energy utilization efficiency, battery health management and maintenance capabilities, and the ability to evaluate the adaptability of the battery pack. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings.

[0065] Figure 1It is a flowchart of the overall process of an embodiment of the present invention;

[0066] Figure 2 It is a schematic diagram of a high-temperature image of a solid-state battery in an embodiment of the present invention. DETAILED DESCRIPTION

[0067] The embodiment of the present application provides a solid-state battery thermal fault diagnosis and online early warning method to solve the technical problems of low computing efficiency and inaccurate prediction results of existing solid-state battery thermal fault diagnosis methods when used. When the solid-state battery thermal fault diagnosis method is used, the present invention uses thermal imaging to detect the indicated temperature, temperature change rate and temperature diffusion rate of the battery under use, and analyzes to obtain a state evaluation value Bauy that can intuitively judge the actual tolerance of the battery. t , and then according to the state evaluation value Bauy t The position relationship with the state threshold interval can be used to call the corresponding solution, thereby providing a basis for battery replacement and adjustment. It can also extend the battery life when the replacement cycle is missed, providing sufficient time for battery replacement. Secondly, the state evaluation value Bauy t Combined with the detection results obtained by the subsequent execution of the detection strategy, the adaptive estimate Bauy is obtained x , so as to comprehensively predict the adaptability of the battery and make it convenient to formulate a corresponding battery usage mode according to the battery's tolerance, so as to make full use of the remaining service life of the battery; the present invention detects the distribution of the indicated temperature, finds the heat source of the battery, and then calls and executes the corresponding detection strategy according to the location of the heat source, which can improve the detection efficiency and the accuracy of the calculation result. In addition, the results and state evaluation value Bauy obtained by executing the detection strategy can be t Comprehensively predict the battery life.

[0068] Example 1: Please refer to Figure 1 to Figure 2 The technical solution in the embodiment of the present application is to solve the technical problems that the existing solid-state battery thermal fault diagnosis method has low computational efficiency and inaccurate prediction results when used. The overall idea is as follows:

[0069] In view of the problems existing in the prior art, the present invention provides a solid-state battery thermal fault diagnosis and online early warning method, which is divided into two parts: hardware (mainly thermal imager, other equipment can use vehicle-mounted equipment) and software. The thermal fault detection of the battery is completed through the cooperation of hardware and software. At the same time, the corresponding detection strategy can be called according to the detection results, which improves the detection accuracy and facilitates subsequent maintenance. The specific structure is as follows:

[0070] hardware:

[0071] The NV384 thermal imager is used as the acquisition device for the temperature distribution image of the solid-state battery. The NV384 thermal imager is a high-sensitivity uncooled detector with a stable imaging circuit, low image noise, frame rate, resolution, temperature measurement accuracy, low power consumption, and stable and reliable operation. It has a compact structure and is easy to install. It has an external power supply and a reserved RS232 communication interface for easy system integration. It is suitable for night assisted driving of various vehicles. It is connected to the vehicle-mounted system through the RS232 communication interface to transmit the acquired solid-state battery temperature distribution image to the vehicle-mounted system for calculation.

[0072] Installation: Set up at least four groups of NV384 thermal imagers, and install them 20-30cm away from the solid-state battery. The lens of the thermal imager is tilted toward the direction of the solid-state battery. It is best if all solid-state batteries appear within the shooting range of the thermal imager. When it is impossible to cover all of them, a combination method is adopted, that is, by setting the number and position of NV384 thermal imagers, the solid-state battery is divided into a corresponding number of areas, and the NV384 thermal imager is matched with the area to be photographed (the corresponding shooting area needs to be fully covered).

[0073] software:

[0074] Analyze the images collected by the hardware (NV384 thermal imager) to obtain the data brought by the images, and then deduce the battery conditions through the image data.

[0075] Image data acquisition: The temperature changes on the battery surface are monitored in real time by a thermal imager, and the thermal images are saved at the same time interval. The saved surface thermal images are batch processed and cropped into images of the same size as the input of the feature extraction network. The image needs to be normalized before input. The feature extraction network first extracts key point information from the image, converts it into a digital signal and inputs it into the network. After that, the extracted information is forward propagated to complete classification, recognition and positioning. The key data in the temperature distribution image of the solid-state battery is extracted by the feature extraction network to obtain image feature parameters that can reflect the temperature characteristics of the battery and a temperature gradient map that can reflect the temperature distribution. The image feature parameters include the indicated temperature, temperature change rate and temperature diffusion rate of each area. The above-mentioned feature extraction network architecture can adopt the common image network architecture on the market. For example, the journal "Intelligent Fault Location Technology of Lithium-ion Batteries Based on Thermal Imaging" published by the School of Electrical Automation and Information Engineering of Tianjin University in February 2024 pointed out the feature extraction network architecture.

[0076] Indicative temperature: According to the pre-set image segmentation size, the solid-state battery temperature distribution image collected by the thermal imager is divided into several small images (image blocks), and then several points (image points) are randomly selected in an image block, and the RGB values ​​of the corresponding points are extracted according to the image feature extraction network, and the temperature proportional coefficient matching the RGB value of the image point is retrieved from the database, where the number of temperature distribution image segmentations is n, and the number of image points randomly selected in the same image block is m. The value obtained by importing is recorded as the indicated temperature Qc of the corresponding image block.

[0077] Get the average color perception coefficient Del of all image points in the same image block, and use the average color perception coefficient Del as the color of the image block to participate in subsequent calculations. The formula for the average color perception coefficient Del is as follows:

[0078]

[0079] In the formula, R i , G i , B i Represents the RGB value of the i-th image point, Rcl j , Glid j 、Bibo j Represents the weight corresponding to the RGB value of the i-th image point, 0<i≤m.

[0080] Using the average color perception coefficient Del, combined with the preset temperature ratio Qed x Orn Then the indicated temperature Qc of the corresponding image block can be obtained, where the average color perception coefficient Del will increase as the color deepens, and the temperature setting ratio Qed is calculated by the temperature represented by different average color perception coefficients Del. x Orn , since the image is cyan at the beginning of the temperature, it gradually changes to yellow and then to red as the temperature continues to rise. However, when the temperature continues to rise, the color change is not obvious, and it may change from light red to dark red. For this reason, we set the temperature ratio Qed x Orn It changes in exponential form to display the temperature in different colors. The specific formula is as follows:

[0081]

[0082] In the formula, Del represents the average color coefficient of all image points in the same image block, Qed x Orn represents the preset temperature ratio, Orn represents the exponent of the temperature ratio, δ represents the correction factor of the average color perception coefficient, ψ Del is the battery temperature influence coefficient during detection, λDel is the ambient temperature influence coefficient during detection, ε1 and ε2 represent the preset battery temperature influence coefficient and the weight of the battery temperature influence coefficient respectively, ε1+ε2=1, and e is a natural constant.

[0083] Temperature change rate: By monitoring the change of the indicated temperature of the same image block during the detection time, the temperature change of a certain image block during the detection process can be obtained. The specific formula is as follows:

[0084]

[0085] Where W q represents the temperature change rate of the same image block, T represents the preset detection time, Qc t Qc0 indicates the indicated temperature at the end of the detection time, and Qc1 indicates the indicated temperature at the start of the detection time.

[0086] For example, setting the detection time T Q = 1s, the indicated temperature at the start is Qc0 = 23°C, the indicated temperature at the end of the detection time is Qc t =25℃, then By observing W q The positive and negative sides of the same image block determine whether it is rising or falling, and then use W q The size of reflects the change of battery surface temperature.

[0087] When W q When the temperature is ≥8℃ / s, the on-board battery cooling system needs to be turned on to cool the solid-state battery;

[0088] When W q When the temperature is ≥14℃ / s, it is necessary to turn on the on-board battery cooling system and reduce the output power of the solid-state battery. The reduction range can be set according to needs, and the minimum reduction is to 8.5% of the power before adjustment, and an alarm command is sent to the driver.

[0089] Temperature change rate: After the indicated temperature is calculated, randomly select multiple image blocks with the same indicated temperature and mark them as the starting image blocks. The temperature at this time is recorded as the approved temperature. Then obtain the indicated temperature of the image blocks around the starting image block, and record the image blocks with a temperature lower than the indicated temperature as the image blocks to be tested. Record the time T when the indicated temperature of the image blocks to be tested reaches the approved temperature. L ;

[0090]

[0091] In the formula, B trepresents the temperature change rate, L represents the distance between two adjacent image blocks, the distance between two adjacent image blocks is equal to the distance between the center points of the two image blocks, that is, the coordinates of the center points of the two image blocks are (X0, Y0) and (X1, Y1), respectively. Since the size of the image segmentation is fixed, the distance between the center points of two adjacent image blocks is also fixed, and is only related to the preset segmentation size, and can be calculated by the above method.

[0092] When selecting the starting image block, multiple image blocks are randomly selected first, and the indicated temperatures of the image blocks around the randomly selected image block are obtained. The surrounding image blocks with indicated temperatures greater than the selected image block are discarded; the rest are counted as image blocks to be tested. When a certain image block to be tested reaches the indicated temperature of the selected image block in the starting state, the counting is stopped, and this time period is counted as T L .

[0093] During the solid-state battery testing process, the environment around the solid-state battery will also affect it, so it is also necessary to analyze the environment where the solid-state battery is located, and to obtain the tolerance range of the solid-state battery to temperature, humidity and air thermal conductivity, and calculate the corresponding adaptation coefficient and environmental parameter impact value Eltq t , thereby judging the degree of impact of the environment on the battery and the adaptability of the battery in the current environment, which is not only convenient for subsequent status evaluation t The calculation excludes environmental interference and combines the environmental parameter influence value Eltq recorded in historical records t , reminding users of the impact of the current environment on the battery, facilitating users' intuitive understanding, and can be coordinated with the vehicle system to increase the battery life. The impact of the external environment on solid-state batteries is analyzed, and the specific analysis process is as follows:

[0094] Extract the environmental adaptability information of the solid-state battery stored in the database, obtain the environmental parameter range that the solid-state battery can tolerate, and further obtain the temperature range, humidity range and air thermal conductivity range that the solid-state battery can tolerate;

[0095] The temperature range that the solid-state battery can withstand is compared with the maximum indicated temperature obtained during the actual test process, and the adaptation coefficient between the solid-state battery's temperature tolerance and the maximum indicated temperature is analyzed and recorded as

[0096] Similarly, according to the analysis method of the matching coefficient between the solid-state battery tolerance temperature and the maximum indicated temperature, the adaptation coefficients of the solid-state battery tolerance humidity and tolerance air thermal conductivity to the humidity and air thermal conductivity in the actual detection process are obtained, and they are recorded as

[0097] By analyzing the formula Get the environmental parameter impact value Eltq of solid-state batteries t , where φ represents the correction factor of the preset environmental parameter impact value, W t , S t 、Dr t represent the preset weight factors of temperature, humidity and air thermal conductivity, respectively, and W t >S t >Dr t >0.

[0098] Evaluate the use status of solid-state batteries: Based on the quantified image feature parameters and the environmental parameters of the location of the solid-state battery, obtain the use status evaluation value of the solid-state battery and generate the solid-state battery use status evaluation value Bauy t The formula is as follows:

[0099]

[0100] In the formula, Qc max Indicates the maximum temperature of all indicated temperatures in the temperature distribution image, Qc x represents the indicated temperature of the x-th image block, x is a positive integer, a1, a2, a3 represent the weights of the indicated temperature, temperature diffusion coefficient and environmental parameter influence coefficient respectively, a1+a2+a3=1, G1 represents the correction coefficient, represents the temperature diffusion coefficient.

[0101] Among them, since the use of the battery will be affected by the environment, it is necessary to t Remove the influence of environmental parameters Eltq t , so the state evaluation value Bauy t It can best reflect the current usage status of solid batteries.

[0102] After obtaining the status evaluation value Bauy reflecting the battery usage t After that, it is necessary to compare it with the preset state threshold interval, and evaluate the current use of the solid-state battery according to the position of the solid-state battery usage state evaluation value in the preset state threshold interval, where the state threshold interval includes the state threshold upper limit Bauy ↑ (Qualifying line) and lower limit of status threshold Bauy ↓ (Scrap Line);

[0103] When the state evaluation value Bauy t ≥ Upper threshold of status Bauy ↑ If the battery is in a state of inactivity, it means that the battery life and usage status meet the usage requirements and can continue to be used normally. Otherwise, no response is given.

[0104] When the state threshold is upper Bauy↑ >Status evaluation value Bauy t ≥Bauy lower threshold ↓ When the battery is about to reach the scrap standard, it needs to be replaced in time. If the battery can still be used at this stage, a replacement instruction will be issued;

[0105] When the state threshold is lowered ↓ >Status evaluation value Bauy t When the battery life reaches a critical value, it may be dangerous if it is not replaced in time. At this time, it is necessary to replace it in time and retrieve the status evaluation value Bauy. t The corresponding low-power output mode (the output upper limit of the solid-state battery is set in a preset manner, so that when it is used in a low-power state, the minimum power must be reduced to 8.5% of the power before adjustment) limits the maximum output voltage and maximum output current of the solid-state battery, thereby ensuring that the solid-state battery will not operate at high load.

[0106] Thermal fault location screening: Set the screening temperature, delete all images below the screening temperature in the temperature gradient map, and record the remaining images as high-temperature images, such as Figure 2 As shown, the location of the high-temperature image is marked in the pre-recorded solid-state battery internal structure distribution map, and the corresponding strategy is executed according to the distribution area of ​​the high-temperature image. The solid-state battery internal structure distribution map adopts a partitioning method, and the location of the circuit and the location of the battery core are marked with lines of different colors on the surface of the solid-state battery to facilitate subsequent distinction. The execution of the corresponding strategy according to the distribution area of ​​the high-temperature image is divided into the following four cases:

[0107] Case 1: When the high temperature image is distributed at the location of the battery core, it means that the temperature rise point is generated by the battery core. And from the difference in color in the image, it can be judged that the temperature of the battery core at different locations is different. For the battery core with higher temperature, a battery maintenance warning is issued and the battery detection strategy is executed. The specific process of the battery detection strategy is as follows:

[0108] Set the duration of the monitoring cycle, obtain the maximum and minimum values ​​of the current and voltage during the solid-state battery monitoring cycle, and obtain the rate of change of the current and voltage during the solid-state battery monitoring cycle;

[0109] Connect a preset detection load to the circuit, re-monitor according to the preset monitoring period, and record the rate of change of current and voltage during the monitoring period;

[0110] According to the data obtained in the two monitoring cycles, the formula Get the estimated health coefficient K of the solid battery cell 电池, where maxI0, minI0, maxI1, minI1 represent the maximum and minimum current values ​​in the monitoring period before and after the load is connected, maxU0, minU0, maxU1, minU1 represent the maximum and minimum voltage values ​​in the monitoring period before and after the load is connected, T d It represents the duration of the preset monitoring cycle, Ies and Upg represent the weight factors of the current change rate and the voltage change rate respectively, Ies>0, Upg>0, and G2 represents the correction coefficient.

[0111] Since the health estimation coefficient K 电池 and state assessment value Bauy t Comparatively, the service life of battery cells and solid-state batteries are evaluated separately.

[0112] Case 2: When the high temperature image is distributed at the location of the line, it means that the temperature rise point is caused by the line, such as short circuit, line aging, etc., and the line needs to be repaired in time, so a maintenance warning is issued and the line detection strategy is executed. The specific process of executing the line detection strategy is as follows:

[0113] Input the preset current and voltage to the line and keep T s After a certain time, record the average temperature of the area where the circuit is located in the solid battery;

[0114] By applying the analytical formula Get the line health estimation coefficient C of the solid-state battery 线路 , where θ represents the correction factor of the preset line health estimate, I q ,I b Respectively represent the line input current before and after adjustment, U q , U b They represent the line input voltages before and after adjustment respectively, Ice and Uce represent the weight factors of current and voltage respectively, and Ice+Uce=1.

[0115] The obtained results are then compared with the standard values, including the upper and lower loss thresholds, to determine whether the line is damaged, as follows:

[0116] When the line is normal, the preset current and voltage are input to the line for a period of T s After time, the line health prediction coefficient C′ is obtained 线路 , and when C′ 线路 -C 线路 When the value is ≥ the upper loss threshold, it means that the influence of temperature on the line is negligible, and no response is made;

[0117] When the loss lower limit threshold>C′ 线路 -C 线路When the value is greater than or equal to the lower loss threshold, it indicates that the current temperature has interfered with the line and the line load and the amount of data calculation on the circuit board need to be reduced.

[0118] When C′ 线路 -C 线路 When the loss is less than the lower limit threshold, it means that the line has suffered irreversible damage and the use of the solid-state battery needs to be stopped and repaired in time.

[0119] Case 3: When the high-temperature images are distributed in the circuits and battery cells inside the solid-state battery, the battery detection strategy and the circuit detection strategy are executed simultaneously.

[0120] Case 4: When the high-temperature image is distributed in a location other than the circuit or the battery cell inside the solid-state battery, the temperature distribution image of the solid-state battery during use is collected again, and the solid-state battery usage status assessment values ​​obtained from the two collections are compared. When the areas where the results of the two collections are located are consistent, no response is given, otherwise a verification warning is issued.

[0121] Among them, when the areas where the results of the two acquisitions are located are consistent, the image needs to be sent to the background management system for manual review by the staff.

[0122] When receiving a verification warning, adjust the shooting frequency of the solid-state battery temperature distribution image. The adjusted shooting frequency must be 5 times or more greater than the original shooting frequency.

[0123] Obtain the coordinates of the area block corresponding to the maximum temperature in the solid-state battery temperature distribution image, and substitute all the coordinates into the two-dimensional coordinate axis to generate a coordinate change curve;

[0124] The acquired coordinate change curve is compared with the historical coordinate change curve in the database; when the overlap degree of the two curves reaches more than 80%, no response is made, and the coordinate change curve acquired this time is stored in the database as the historical coordinate change curve; otherwise, a maintenance instruction is issued.

[0125] Battery adaptability prediction: Combine the health estimation coefficient obtained in the detection strategy process and the solid-state battery usage status assessment value to calculate the adaptability estimation value of the solid-state battery under the current environmental parameters. Due to different detection strategies, the formula for calculating the adaptability estimation value will also be different. The formula for generating the adaptability estimation value is as follows:

[0126]

[0127] In the formula, Bauy x represents the adaptive estimate, Bauy x ≥0, when Bauy x When <0, the solid-state battery temperature distribution image is collected.

[0128] Among them, when the battery detection strategy and the line detection strategy are executed, the average of the adaptability estimates obtained by the two strategies is, for example: But when the battery detection strategy gets Bauy x Bauy obtained with line detection strategy x When the difference between them is greater than the error threshold, a maintenance instruction is issued.

[0129] Then, the corresponding solid-state battery thermal failure prediction value is calculated based on the battery adaptability estimate, and the solid-state battery thermal failure prediction value Waty is obtained. t Compare with the evaluation threshold range. The formula for generating the solid-state battery thermal failure prediction value is as follows:

[0130] Waty t =Bauy x ×γ-Eltq t ×μ

[0131] In the formula, Waty t represents the solid-state battery thermal failure prediction value, γ represents the proportional coefficient of the adaptability estimate value, μ represents the proportional coefficient of the environmental parameter influence value, and is compared with the solid-state battery thermal failure prediction value obtained last time, that is, the solid-state battery thermal failure prediction value Waty t -The last solid-state battery thermal failure prediction value Waty t When the predicted difference interval is less than , the result is output; otherwise, the solid-state battery temperature distribution image is re-collected.

[0132] Setting an evaluation threshold interval of a thermal fault prediction value, the evaluation threshold interval includes a qualified threshold interval, a risk threshold interval, and a temperature drop threshold interval;

[0133] Wherein, when the thermal fault prediction value is within the qualified threshold range, no response is made and the result is output;

[0134] When the fault prediction value is within the risk threshold range, a maintenance instruction is issued and the auxiliary cooling system is turned on to cool the solid-state battery;

[0135] When the thermal fault prediction value is within the temperature reduction threshold range, a danger warning is issued and the auxiliary cooling system is turned on to cool the solid-state battery. At the same time, the output power of the solid-state battery is reduced until the indicated temperature on the surface of the solid-state battery no longer increases.

[0136] Finally, it should be noted that: Obviously, the above embodiments are only examples for clearly explaining the present invention, and are not intended to limit the implementation methods. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation methods here. The obvious changes or modifications derived from this are still within the scope of protection of the present invention.

Claims

1. A solid-state battery thermal fault diagnosis and online early warning method, characterized in that: The method includes: After normalizing the acquired solid-state battery temperature distribution image, it is input into the feature extraction network to generate image feature parameters and temperature gradient map, where the image feature parameters include the indicated temperature, temperature change rate and temperature diffusion speed of each area; According to the quantified image feature parameters and the environmental parameters of the location of the solid-state battery, the solid-state battery usage status evaluation value is obtained, and then the current usage of the solid-state battery is evaluated according to the position of the solid-state battery usage status evaluation value in the preset state threshold interval; Set a screening temperature, delete the images in the temperature gradient map that are lower than the screening temperature, record the remaining images as high-temperature images, mark the locations of the high-temperature images in the pre-recorded solid-state battery internal structure distribution map, and execute the corresponding strategy according to the distribution area of ​​the high-temperature images; When the high-temperature image is distributed at the location of the battery cell, a maintenance warning is issued and the battery detection strategy is executed; When high temperature images are distributed at the location of the line, a maintenance warning is issued and the line detection strategy is executed; When the high-temperature image is distributed in a location other than the circuit or the battery core in the solid-state battery, the temperature distribution image of the solid-state battery during use is collected again, and the solid-state battery use status assessment values ​​obtained from the two collections are compared. If the areas where the results of the two collections are located are consistent, no response is given, otherwise a verification warning is issued; Combined with the health prediction coefficient and the solid-state battery usage status assessment value obtained in the detection strategy process, the adaptability estimation value of the solid-state battery under the current environmental parameters and the solid-state battery thermal failure prediction value corresponding to the battery adaptability estimation value are calculated.

2. The solid-state battery thermal fault diagnosis and online early warning method according to claim 1, characterized in that: The process of obtaining image feature parameters is as follows: According to the preset image segmentation size, the temperature distribution image is divided into n image blocks of the same size, and m image points are randomly selected in the same image block, the RGB values ​​corresponding to the image points are retrieved, and the temperature ratio matching the RGB values ​​of the image points is obtained from the database. The obtained value is recorded as the indicated temperature Qc of the corresponding image block; The formula for generating the indicated temperature Qc of the corresponding image block is as follows: In the formula, Del represents the average color coefficient of all image points in the same image block, Qed x Orn represents the preset temperature ratio, Orn represents the exponent of the temperature ratio, δ represents the correction factor of the average color perception coefficient, ψ Del is the battery temperature influence coefficient during detection, λ Del is the environmental temperature influence coefficient during detection, ε1 and ε2 represent the preset battery temperature influence coefficient and the weight of the battery temperature influence coefficient respectively, ε1+ε2=1, R i , G i , B i Represents the RGB value of the i-th image point, Rcl j , Glid j 、Bibo j Represents the weight corresponding to the RGB value of the i-th image point, 0<i≤m, e is a natural constant; Where W q Indicates the temperature change rate of the same image block, T Q Indicates the preset detection time, Qc t represents the indicated temperature at the end of the test time, and Qc0 represents the indicated temperature at the beginning of the test time; After the indicated temperature calculation is completed, multiple image blocks with the same indicated temperature are randomly selected and marked as the starting image block. The temperature at this time is recorded as the approved temperature. Then the indicated temperature of the image blocks around the starting image block is obtained, and the image blocks with a temperature lower than the indicated temperature are recorded as the image blocks to be tested. The time T when the indicated temperature of the image blocks to be tested reaches the approved temperature is recorded. L ; In the formula, B t represents the temperature change rate, and L represents the distance between two adjacent image blocks.

3. The solid-state battery thermal fault diagnosis and online early warning method according to claim 1, characterized in that: The environmental parameters of the location of the solid-state battery include air temperature, air humidity, and air thermal conductivity. The impact of the external environment of the solid-state battery is analyzed. The specific analysis process is as follows: Extract the environmental adaptability information of the solid-state battery stored in the database, obtain the environmental parameter range that the solid-state battery can tolerate, and further obtain the temperature range, humidity range and air thermal conductivity range that the solid-state battery can tolerate; The temperature range that the solid-state battery can withstand is compared with the maximum indicated temperature obtained during the actual test process, and the adaptation coefficient between the solid-state battery's temperature tolerance and the maximum indicated temperature is analyzed and recorded as Similarly, according to the analysis method of the matching coefficient between the solid-state battery tolerance temperature and the maximum indicated temperature, the adaptation coefficients of the solid-state battery tolerance humidity and tolerance air thermal conductivity to the humidity and air thermal conductivity in the actual detection process are obtained, and they are recorded as By analyzing the formula Get the environmental parameter impact value Eltq of solid-state batteries t , where φ represents the correction factor of the preset environmental parameter impact value, W t , S t 、Dr t represent the preset weight factors of temperature, humidity and air thermal conductivity, respectively, and W t >S t >Dr t >0.

4. The solid-state battery thermal fault diagnosis and online early warning method according to claim 3, characterized in that: Generate solid-state battery usage status assessment value Bauy t The formula used is as follows: In the formula, Qc max Indicates the maximum temperature of all indicated temperatures in the temperature distribution image, Qc x represents the indicated temperature of the x-th image block, x is a positive integer, a1, a2, a3 represent the weights of the indicated temperature, temperature diffusion coefficient and environmental parameter influence coefficient respectively, a1+a2+a3=1, G1 represents the correction coefficient.

5. The solid-state battery thermal fault diagnosis and online early warning method according to claim 4, characterized in that: The state threshold interval includes the state threshold upper limit Bauy ↑ and the lower threshold of the state Bauy ↓ ; When the state evaluation value Bauy t ≥ Upper threshold of status Bauy ↑ If , no response is given; When the state threshold is upper Bauy ↑ >Status evaluation value Bauy t ≥Bauy lower threshold ↓ When , a replacement instruction is issued; When the state threshold is lowered ↓ >Status evaluation value Bauy t When the status evaluation value Bauy is reached, a replacement command is issued and the t The corresponding low-power output mode limits the maximum output voltage and maximum output current of the solid-state battery.

6. The solid-state battery thermal fault diagnosis and online early warning method according to claim 1, characterized in that: The specific process of executing the battery detection strategy is: Set the duration of the monitoring cycle, obtain the maximum and minimum values ​​of the current and voltage during the solid-state battery monitoring cycle, and obtain the rate of change of the current and voltage during the solid-state battery monitoring cycle; Connect a preset detection load to the circuit, re-monitor according to the preset monitoring period, and record the rate of change of current and voltage during the monitoring period; According to the data obtained in the two monitoring cycles, the formula Get the estimated health coefficient K of the solid battery cell 电池 , where maxI0, minI0, maxI1, minI1 represent the maximum and minimum current values ​​in the monitoring period before and after the load is connected, maxU0, minU0, maxU1, minU1 represent the maximum and minimum voltage values ​​in the monitoring period before and after the load is connected, T d It represents the duration of the preset monitoring cycle, Ies and Upg represent the weight factors of the current change rate and the voltage change rate respectively, Ies>0, Upg>0, and G2 represents the correction coefficient.

7. The solid-state battery thermal fault diagnosis and online early warning method according to claim 6, characterized in that: The specific process of executing the line detection strategy is as follows: Input the preset current and voltage to the line and keep T s After a certain time, record the average temperature of the area where the circuit is located in the solid battery; By applying the analytical formula Get the line health estimation coefficient C of the solid-state battery 线路 , where θ represents the correction factor of the preset line health estimate, I q ,I b Respectively represent the line input current before and after adjustment, U q , U b They represent the line input voltages before and after adjustment respectively, Ice and Uce represent the weight factors of current and voltage respectively, and Ice+Uce=1.

8. The solid-state battery thermal fault diagnosis and online early warning method according to claim 1, characterized in that: When a verification warning is received, the shooting frequency of the solid-state battery temperature distribution image is adjusted, and the adjusted shooting frequency must be 5 times or more than the original shooting frequency; the coordinates of the area block corresponding to the maximum temperature indicated in the solid-state battery temperature distribution image are obtained, and all the coordinates are substituted into the two-dimensional coordinate axis to generate a coordinate change curve; the obtained coordinate change curve is compared with the historical coordinate change curve in the database; when the overlap degree of the two curves reaches more than 80%, no response is given; otherwise, a maintenance instruction is issued.

9. The solid-state battery thermal fault diagnosis and online early warning method according to claim 7, characterized in that: The formula used to generate the adaptability estimate is as follows: In the formula, Bauy x represents the adaptive estimate, Bauy x ≥0, when Bauy x When <0, the temperature distribution image of the solid-state battery is collected; The formula used to generate the solid-state battery thermal failure prediction value is as follows: Watts t =Bauy x ×γ-Eltq t ×μ In the formula, Waty t Represents the solid-state battery thermal failure prediction value, γ represents the proportional coefficient of the adaptability estimation value, μ represents the proportional coefficient of the environmental parameter influence value, and is compared with the solid-state battery thermal failure prediction value obtained last time. When the difference is less than the prediction difference interval, the result is output; otherwise, the solid-state battery temperature distribution image is re-collected.

10. The solid-state battery thermal fault diagnosis and online early warning method according to claim 9, characterized in that: Setting an evaluation threshold interval of a thermal fault prediction value, the evaluation threshold interval includes a qualified threshold interval, a risk threshold interval, and a temperature drop threshold interval; Wherein, when the thermal fault prediction value is within the qualified threshold range, no response is made and the result is output; When the fault prediction value is within the risk threshold range, a maintenance instruction is issued and the auxiliary cooling system is turned on to cool the solid-state battery; When the thermal fault prediction value is within the temperature reduction threshold range, a danger warning is issued and the auxiliary cooling system is turned on to cool the solid-state battery. At the same time, the output power of the solid-state battery is reduced until the indicated temperature on the surface of the solid-state battery no longer increases.

Images