A mine explosion-proof lithium ion battery life prediction method and system

Abstract

The present application relates to the technical field of image data processing, and more particularly to a mine-used explosion-proof lithium ion battery life prediction method and system, which comprises the following steps: collecting a thermal image of the surface of the battery at the constant voltage stage of the end of charging; calculating geometric thermal complexity based on the gray extremum difference in the neighborhood of the pixel point; identifying high gradient pixel points, combining their distance distribution characteristics, geometric thermal complexity and global gradient characteristics to construct a thermal intensity factor; calculating the initial fractal dimension of the thermal image, and correcting it to obtain an adaptive fractal dimension according to the thermal intensity factor; and predicting the life based on the mapping relationship between the adaptive fractal dimension and the battery health state. The present application uses the thermal intensity factor to correct the fractal dimension, effectively distinguishes the thermal accumulation interference caused by the heat dissipation lag of the explosion-proof shell from the thermal abnormality of the battery cell aging, and solves the prediction misjudgment and omission of the battery life under complex thermal environment.

Description

Technical Field

[0001] This invention relates to the field of image data processing technology, and in particular to a method and system for predicting the lifespan of explosion-proof lithium-ion batteries used in mining. Background Technology

[0002] Explosion-proof lithium-ion batteries are the core power source for mining electrical equipment. Due to the special working conditions underground, these batteries must be encapsulated in a heavy explosion-proof casing, resulting in extremely poor heat dissipation. This leads to continuous heat accumulation and uneven temperature distribution inside the module, accelerating cell aging and even causing thermal runaway. Therefore, accurately predicting the lifespan of explosion-proof lithium-ion batteries for mining, without compromising the explosion-proof structure, is crucial for ensuring safe production underground.

[0003] Currently, non-contact monitoring based on infrared thermal imaging is the mainstream method for assessing battery status. Existing technologies typically employ the box-dimensional method from fractal theory, treating the thermal distribution on the battery surface as a geometric surface and evaluating the geometric complexity of the thermal texture by calculating its fractal dimension. It is generally believed that the higher the degree of battery aging, the more uneven its surface temperature distribution, the higher the geometric complexity of the texture, and the larger the calculated fractal dimension.

[0004] However, in the actual application scenarios of explosion-proof batteries in mining, the traditional box-counting method mentioned above struggles to obtain accurate prediction results. This is because the unique heat dissipation-constrained environment inside the explosion-proof casing often results in large-area, diffuse thermal accumulation on the battery surface. This thermal accumulation, caused by the lag in environmental heat dissipation, manifests as extremely rich geometric textures in thermal images, leading to a significantly higher geometric fractal dimension calculated by traditional algorithms. This can easily misjudge a healthy battery as being in a severely aged state. Simultaneously, under the interference of complex background thermal textures, early point-like thermal anomalies truly caused by micro-short circuits or aging within the cell are often submerged due to their small proportion in the geometric statistics, leading to missed detections. Summary of the Invention

[0005] To address the technical problem of complex thermal accumulation interference caused by the delayed heat dissipation of the explosion-proof casing, which cannot be effectively separated from the actual aging characteristics by the geometric morphology statistical method, resulting in misjudgment or omission in the prediction of the life of mining batteries, the present invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides a method for predicting the lifespan of explosion-proof lithium-ion batteries used in mining, the method comprising the steps of:

[0007] A thermal image of the battery surface is acquired during the constant voltage stage at the end of charging. Based on the grayscale extreme value difference within the neighborhood of each pixel in the thermal image, the geometric thermal complexity of the thermal image is calculated. A predetermined number of pixels with the largest gradient amplitude in the thermal image are designated as high-gradient pixels. The sum of the gradient amplitudes of all pixels in the thermal image is multiplied by the geometric thermal complexity of the thermal image to obtain an initial intensity value. The negative exponential function value of the distance distribution characteristics of the high-gradient pixels is multiplied by the initial intensity value to obtain the thermal intensity factor of the thermal image. The initial fractal dimension of the thermal image is calculated, and the initial fractal dimension is corrected based on the thermal intensity factor of the thermal image to obtain an adaptive fractal dimension. Based on the mapping relationship between the adaptive fractal dimension and the battery health status, the battery life is predicted.

[0008] This invention acquires thermal images during the constant-voltage phase at the end of charging and uses these images as a basis to calculate geometric thermal complexity, thermal intensity factor, and adaptive fractal dimension to predict battery life. The invention introduces the thermal intensity factor as a key indicator, which integrates the amplitude and intensity of pixel gradients with the spatial distance distribution characteristics of high-gradient pixels to achieve quantitative characterization and qualitative identification of the texture features of thermal images. Based on this, the initial fractal dimension is adaptively corrected using the thermal intensity factor: when there are clustered high-energy hot spots, the dimension weight is increased to highlight aging characteristics; when there is only diffuse heat accumulation, the dimension weight is suppressed to eliminate interference. Finally, a lifespan mapping is established based on the corrected adaptive fractal dimension, thereby achieving high-precision prediction of the internal aging state of the battery without damaging the explosion-proof structure.

[0009] Preferably, the step of calculating the geometric thermal complexity of the thermal image based on the gray-level extreme value difference in the neighborhood of each pixel in the thermal image includes: accumulating and averaging the gray-level extreme value differences in the neighborhood of all pixels in the thermal image to obtain the geometric thermal complexity of the thermal image.

[0010] This invention assesses the geometric thermal complexity of a thermal image by averaging the gray-level extreme differences in the neighborhood of all pixels. This calculation method captures the degree of gray-level fluctuations in every tiny region of the thermal image surface from a microscopic perspective. The gray-level extreme differences reflect the severity of local temperature changes. By averaging the values ​​across the entire image, the texture details of a two-dimensional image can be transformed into macroscopic geometric indicators that reflect surface roughness, thereby providing accurate geometric benchmark data for distinguishing complex background thermal textures from real aging features.

[0011] Preferably, the thermal intensity factor of the thermal image Satisfying the relation:

[0012] ;

[0013] in, It is the geometric thermal complexity of the thermal image; , These are the width and height of the thermal image, respectively. , They are pixels The horizontal gradient magnitude and the vertical gradient magnitude; It is the one with the largest gradient amplitude in the thermal image. The average Euclidean distance between pixels It is the preset quantity; This is an adjustment factor for the value. It is a standard normalized function; It is a natural exponential function.

[0014] This invention clarifies the specific calculation formula for the thermal intensity factor, introduces a standard normalization function and a value adjustment coefficient, and utilizes the natural exponential function to process the average Euclidean distance of high-gradient pixels to construct a nonlinear spatial filter. By adjusting the coefficient, the factor value is mapped to a suitable range, ensuring its mathematical stability as a correction coefficient.

[0015] Preferably, the distance distribution characteristic of the high-gradient pixels is the average Euclidean distance between the high-gradient pixels.

[0016] Preferably, calculating the initial fractal dimension of the thermal image includes: using the difference box dimension method to calculate the fractal dimension of the thermal image at multiple preset grid scales.

[0017] Preferably, the step of correcting the initial fractal dimension based on the thermal intensity factor of the thermal image to obtain an adaptive fractal dimension includes: multiplying the fractal dimension of the thermal image by the thermal intensity factor of the thermal image to obtain the adaptive fractal dimension.

[0018] This invention uses a multiplication operation to correct the initial fractal dimension using a thermal intensity factor. This mechanism utilizes the principle of signal amplification: if the thermal intensity factor indicates a high-risk hotspot (i.e., a large factor value), the multiplication operation significantly amplifies the fractal dimension, causing its value to exceed the normal range, thus issuing an alarm signal; if the thermal intensity factor indicates background noise (i.e., a small factor value), the multiplication operation compresses the fractal dimension, bringing it back to a safe range. This non-linear weighting method improves the signal-to-noise ratio of the adaptive fractal dimension to aging characteristics, reducing the false alarm rate.

[0019] Preferably, predicting battery life based on the mapping relationship between the adaptive fractal dimension and battery health status includes: normalizing the adaptive fractal dimension between a preset baseline dimension and a threshold dimension to obtain the predicted battery life value.

[0020] Preferably, acquiring the thermal image of the battery surface during the constant voltage stage at the end of charging includes: monitoring the charging state of the battery and acquiring the thermal image when the battery is in the constant voltage charging stage at the end of charging.

[0021] In a second aspect, the present invention provides a life prediction system for explosion-proof lithium-ion batteries used in mining. The system includes a memory and a processor. The memory stores computer program instructions, which, when executed by the processor, implement the life prediction method for explosion-proof lithium-ion batteries used in mining according to the first aspect of the present invention.

[0022] By adopting the above technical solution, a computer program for predicting the lifespan of a mine-use explosion-proof lithium-ion battery according to the first aspect of the present invention is generated and stored in a memory so that it can be loaded and executed by a processor. A terminal device can then be made based on the memory and the processor for convenient use.

[0023] The beneficial effects of this invention are as follows: First, this invention acquires thermal images at the end of battery charging. Addressing the problem of severe heat accumulation in mining explosion-proof batteries due to the sluggish heat dissipation caused by the heavy casing, it introduces a thermal intensity factor based on the spatial distance distribution of high-gradient pixels, building upon the computational geometric thermal complexity. A negative exponential function is used to construct spatial constraints, effectively distinguishing between clustered aging hot spots and discrete environmental noise. This thermal intensity factor is then used to dynamically weight and correct the initial fractal dimension, achieving targeted enhancement of aging features and suppression of interfering textures. Finally, the lifetime is predicted based on the corrected adaptive fractal dimension, improving the accuracy and anti-interference capability of lifetime prediction for mining explosion-proof lithium-ion batteries under complex operating conditions, effectively preventing misjudgments and missed judgments. Attached Figure Description

[0024] Figure 1 A flowchart of a method for predicting the lifespan of an explosion-proof lithium-ion battery for mining, provided in an embodiment of the present invention;

[0025] Figure 2 The original infrared thermal image of a mining explosion-proof lithium-ion battery during the constant voltage stage at the end of charging, provided for an embodiment of the present invention;

[0026] Figure 3 This is a schematic diagram of the result after processing by a method for predicting the lifespan of explosion-proof lithium-ion batteries used in mining, provided in an embodiment of the present invention.

[0027] Figure 4 A comparison chart of lifetime prediction differences before and after fractal dimension correction provided in an embodiment of the present invention;

[0028] Figure 5 This is a structural block diagram of a mining explosion-proof lithium-ion battery life prediction system provided in an embodiment of the present invention. Detailed Implementation

[0029] The first aspect of this invention provides a method for predicting the lifespan of explosion-proof lithium-ion batteries used in mining, such as... Figure 1 As shown, the method includes steps S100-S400:

[0030] Step S100: Collect a thermal image of the battery surface during the constant voltage stage at the end of charging.

[0031] It should be noted that the thermal characteristics of lithium-ion batteries are highly dependent on operating conditions. Under stable conditions such as quiescent or low-rate discharge, the difference in thermal behavior between aged and healthy cells is minimal and easily masked by environmental noise, making it difficult to use as an effective basis for lifespan prediction. Therefore, this step aims to capture the optimal observation time when the internal electrochemical reactions of the battery are most active and the differences in internal resistance heat generation are most significant, thereby obtaining thermal imaging data with high signal-to-noise ratio and clear aging characteristics, laying a data foundation for subsequent feature extraction.

[0032] Specifically, firstly, a micro infrared thermal imaging module with explosion-proof certification is pre-integrated inside the explosion-proof battery box for mining to monitor the charging and discharging status of the battery in real time. When the battery is detected to be in the constant voltage stage at the end of charging, an acquisition command is triggered to obtain an infrared thermal image of the surface of the battery module.

[0033] like Figure 2 The image shown is an infrared thermogram of a mining explosion-proof lithium-ion battery during the constant voltage stage at the end of charging. In the image, due to the lag in heat dissipation caused by the explosion-proof casing, there is a large area of ​​diffuse gray in the background; there are tiny bright spots in the center of the image.

[0034] The acquired infrared thermal images are then preprocessed, such as denoising, before being used as input data for subsequent calculations. Furthermore, considering that subsequent fractal dimension and gradient calculations are based on single-channel grayscale data, while the original thermal image data is typically a temperature matrix or a pseudo-color image, grayscale processing of the acquired data is also necessary.

[0035] Specifically, the effective temperature range of the battery surface is linearly mapped to... The grayscale space is configured as follows: pixels below the lower temperature limit are assigned a value of 0, pixels above the upper temperature limit are assigned a value of 255, and the grayscale value in the middle area is interpolated proportionally. The upper temperature limit can be set to 80. The lower temperature limit is set to 20. The implementers can also configure it according to their needs. Through this grayscale mapping, the physical temperature field is transformed into a standard 8-bit grayscale image matrix, thereby eliminating the influence of differences in temperature measurement ranges of different devices on texture feature extraction and ensuring the standardization of subsequent algorithm inputs.

[0036] At this point, a thermal image of the battery module surface has been obtained.

[0037] Step S200: Calculate the geometric thermal complexity of the thermal image based on the gray-level extreme value difference in the neighborhood of each pixel in the thermal image.

[0038] It should be noted that in fractal theory, the roughness or complexity of a thermal image surface is usually positively correlated with its fractal dimension, reflecting the entropy state. However, directly calculating the fractal dimension of the entire image is computationally intensive and lacks local detail. To intuitively assess the disorder of thermal distribution on the battery surface, this step introduces a geometric thermal complexity index. This index, from a microscopic perspective, indirectly characterizes the complexity of the global thermal texture by accumulating the grayscale fluctuations in tiny regions of the thermal image surface. Under this logic, the more drastic the grayscale changes in the neighborhood of each pixel, the more texture fluctuations and the more complex the shape of the thermal image surface, resulting in a greater geometric thermal complexity. This provides a geometric baseline value for subsequent dimension correction.

[0039] First, construct the neighborhood of each pixel in the thermal image. It should be noted that the size of the neighborhood determines the scale sensitivity of texture feature extraction. If it is too small, it will be greatly affected by noise, while if it is too large, it may smooth out key small thermal features.

[0040] As a preferred implementation, the neighborhood size is set to The pixel designation is based on the structural characteristics of mining explosion-proof batteries, which have thick casings and poor heat dissipation. Under normal operating conditions, heat accumulation tends to be distributed over a large, flat area, while abnormal heat generation caused by cell aging manifests as tiny, dot-like protrusions. The neighborhood size can maximize the highlighting of the dramatic grayscale changes of a single pixel relative to its surroundings, and keenly capture early aging features; compared with a larger neighborhood size, this setting can effectively avoid the smoothing effect and prevent dot-like aging signals from being submerged by background heat accumulation.

[0041] Then, based on the gray-level extreme value difference in the neighborhood of each pixel in the thermal image, the geometric thermal complexity of the thermal image is calculated. It should be noted that, to evaluate the gray-level changes in the neighborhood of each pixel, the thermal image is first considered as a three-dimensional gray-level surface, where... , The axis represents pixel coordinates. The axes represent grayscale values. On this three-dimensional surface, the difference in grayscale extreme values ​​within the neighborhood is calculated to assess the steepness of the local micro-surface.

[0042] Based on the above logic, the geometric thermal complexity of thermal images Satisfying the relation:

[0043] ;

[0044] in, , These are the width and height of the thermal image, respectively. , They are in pixels The maximum and minimum gray values ​​within the center's neighborhood; It is the gray level of the thermal image. Due to the influence of the accuracy of the infrared detector, the gray level of the infrared thermal imager commonly used in industry is 255. Therefore, the present invention preferably sets it to 255. It is a standard normalization function used to quantize calculation results to... For the interval, specific methods such as minimum-maximum normalization and Z-score standardization can be used, which are all existing technologies and will not be elaborated on here.

[0045] In this relation, This is used to calculate the grayscale difference in a local area, reflecting the magnitude of geometric undulations in that area. By summing the results over the entire image using a double summation method, the average value is essentially used to represent the relative surface area of ​​the grayscale surface in the thermal image through the cumulative average of local grayscale fluctuations. The greater the geometric thermal complexity of the thermal image, the more undulations and richer the texture within it.

[0046] Thus, the geometric thermal complexity has been obtained.

[0047] Step S300: The pixels with the largest gradient magnitude in the thermal image are recorded as high gradient pixels; the thermal intensity factor of the thermal image is constructed by combining the distance distribution characteristics of the high gradient pixels, the geometric thermal complexity of the thermal image, and the global gradient features of the thermal image.

[0048] It should be noted that geometric complexity alone is insufficient to distinguish between harmless complex textures caused by heat dissipation lag and dangerous high-energy hot spots caused by cell aging. To address the issue of high-dimensional false alarms or missed small hot spots due to the special operating conditions of explosion-proof enclosures, this step introduces thermal gradient intensity analysis. By constructing a thermal intensity factor to correct the geometric indices, accurate identification of aging characteristics can be achieved.

[0049] First, the pixels with the largest gradient amplitude in the thermal image, representing a predetermined number, are designated as high-gradient pixels. It should be noted that in explosion-proof battery monitoring, aging hot spots typically exhibit point-like high-energy characteristics, meaning they have extremely high center temperatures and large edge gradients, and are spatially clustered. In contrast, environmental noise, such as isolated noise points generated by electromagnetic interference or ordinary poor heat dissipation, usually shows a discrete spatial distribution or a lower gradient amplitude. Therefore, by analyzing the spatial distance between several pixels with the largest gradient amplitudes, it is possible to effectively distinguish between genuine aging heat sources and random noise.

[0050] Specifically, the Sobel operator is used to calculate the horizontal and vertical gradient magnitudes of each pixel in the thermal image, and the pixel with the largest gradient magnitude is identified. For each pixel, calculate the average Euclidean distance of these pixels on the thermal image plane. As a preferred implementation, The preferred value is set to 10. Value too small, such as Distance calculations are easily affected by the randomness of single-point noise; if Value is too large, such as If background pixels from non-hotspot regions are included, the clustering characteristics of the hotspots may be diluted. Therefore, selecting the top 10 points with the largest gradients can maintain high computational robustness while ensuring the capture of hotspot features.

[0051] Then, combining the distance distribution characteristics of high-gradient pixels, the geometric thermal complexity of the thermal image, and the global gradient features of the thermal image, a thermal intensity factor is constructed. It should be noted that the purpose of this factor is to strengthen the weight of true aging characteristics and suppress the influence of background interference. If the overall gradient amplitude of the thermal image is large and high-gradient points are clustered, it indicates the existence of real and strong local hot spots, and their weight in lifetime prediction should be increased. Conversely, if the gradient amplitude is small or high-gradient points are scattered, it may be noise, and its influence should be reduced.

[0052] Based on the above logic, the thermal intensity factor of a thermal image satisfies the following relationship:

[0053] ;

[0054] in, It is the thermal intensity factor of the thermal image; It is the geometric thermal complexity of the thermal image; , These are the width and height of the thermal image, respectively. , They are pixels The horizontal gradient magnitude and the vertical gradient magnitude; It is the one with the largest gradient amplitude in the thermal image. The average Euclidean distance between pixels; The adjustment factor for the value can be set to... This is used to adjust the range of values ​​for the thermal intensity factor in a thermal image to a certain interval. This is to achieve bidirectional correction of the fractal dimension; It is a standard normalized function; It is a natural exponential function.

[0055] In this relation, Assess the thermal intensity of the entire graph; a higher value indicates more drastic temperature changes and is more consistent with aging characteristics. The term utilizes the properties of negative exponential functions to introduce space constraints: when The smaller the value, the more clustered the high gradient points are. Approaching 1, the influence of the strength term is retained; when The larger the value, the more discrete the high gradient points are, suggesting potential noise. The value approaches 0, thus suppressing the graph's contribution to the results.

[0056] Thus, the thermal intensity factor of the thermal image was obtained.

[0057] Step S400: Calculate the initial fractal dimension of the thermal image, correct the initial fractal dimension according to the thermal intensity factor of the thermal image to obtain the adaptive fractal dimension, and predict the battery life based on the mapping relationship between the adaptive fractal dimension and the battery health status.

[0058] It should be noted that while traditional fractal dimension calculations, such as the difference box dimension method, can capture texture complexity, they have a thermal intensity blind spot in mining battery monitoring scenarios, failing to distinguish between complex benign heat dissipation textures and dangerous aging characteristics. Therefore, this step introduces a thermal intensity factor to adaptively correct the traditional fractal dimension: based on geometric complexity, the dimension is weighted by the thermal intensity factor. A larger thermal intensity factor indicates the presence of high-gradient accumulated heat spots, and the corrected dimension should be significantly increased to provide a warning; conversely, a smaller thermal intensity factor indicates mainly gentle thermal accumulation, and the corrected dimension should be reduced to suppress false alarms.

[0059] First, the initial fractal dimension of the thermal image is calculated using the box-difference dimension algorithm. It should be noted that the box-difference dimension method is suitable as a foundational algorithm due to its high computational efficiency and sensitivity to image texture variations. During the calculation process, the grid scale... The choice of grid size is crucial. In a preferred embodiment, the grid size is preset to [value missing]. The thermal characteristics of mining batteries span multiple spatial scales: smaller scales such as... It can capture pixel-level point-like thermal anomalies formed by micro-short circuits in individual battery cells; while larger-scale anomalies such as This allows for the coverage of large-area thermal halos texture caused by the lag in heat dissipation from the explosion-proof casing. Using this multi-scale ensemble ensures that the algorithm fully covers all frequency band features, from microscopic aging points to macroscopic thermal distribution, avoiding the loss of crucial information due to a single scale.

[0060] Then, the initial fractal dimension is corrected based on the thermal intensity factor of the thermal image to obtain an adaptive fractal dimension. It should be noted that in order to integrate the information of the two dimensions of geometric complexity and thermal intensity factor into a single health indicator, this step adopts a multiplicative weighted logic. This logic can make the aging characteristics of high complexity and high thermal intensity more prominent through the synergistic effect of feature amplification and noise suppression, while weakening the benign interference of high complexity but low thermal intensity.

[0061] Based on the above logic, the adaptive fractal dimension of thermal images Satisfying the relation:

[0062] ;

[0063] in, It is the thermal intensity factor of the thermal image; It is the initial fractal dimension of the thermal image.

[0064] In this relationship, when a high-gradient aging hot spot is detected... The initial fractal dimension is amplified, causing its value to exceed the upper limit of the dimension of a normal battery; when a low-gradient background texture is detected... The initial fractal dimension is compressed, causing its value to return to a safe range. This mechanism ensures that the adaptive fractal dimension has strong selective sensitivity to aging features.

[0065] Finally, based on the mapping relationship between the adaptive fractal dimension and the battery health state, the battery life is predicted. It should be noted that the adaptive fractal dimension and the battery's physical aging state exhibit a clear monotonic correlation. Therefore, this step transforms the abstract dimensionality indicator into an intuitive percentage of lifespan by constructing a normalized mapping model.

[0066] Based on the above logic, the predicted battery lifespan is... Satisfying the relation:

[0067] ;

[0068] in, It is the adaptive fractal dimension of the thermal image; , These are the baseline dimension of the battery in its brand-new state and the dimension threshold in its end-of-life state; both can be preset based on historical accelerated aging test data of this battery model. In this embodiment, it is preferred to set... , This means that when the monitored adaptive fractal dimension increases from 2.1 to 2.8, the battery life is determined to be exhausted.

[0069] Furthermore, in order to achieve proactive operation and maintenance, a first-level early warning threshold is set. Level II warning threshold ,like Send relevant maintenance suggestions to the mine control room, indicating that the battery pack consistency has deteriorated; if This triggers a forced shutdown alarm and disconnects the charging circuit to prevent underground accidents caused by aging batteries.

[0070] like Figure 3 The figure shows a schematic diagram of the result after processing using the method provided by this invention. In the figure, the background area is a large area of ​​dark gray, which represents the low gradient background thermal accumulation effectively suppressed by the algorithm. A high-brightness dot-shaped spot is clearly shown in the center of the image. This spot is the aging hot spot located by the algorithm based on the high gradient and spatial aggregation characteristics. In addition, the figure is marked with "Final judgment SOH: 69.8%" and "Aging hot spot detected", which shows the estimated lifetime prediction value and qualitative diagnosis conclusion obtained based on the adaptive fractal dimension.

[0071] like Figure 4 The figure shows a comparison of the lifespan prediction differences before and after fractal dimension correction. The horizontal axis represents different algorithm strategies, and the vertical axis represents the battery health status percentage. The green bars on the left show that the traditional algorithm, due to interference from background thermal accumulation, ignores tiny hot spots and gives a false health score as high as 99.8%. The red bars on the right show that after correction by the thermal intensity factor, the present invention successfully lowers the predicted battery health status percentage to 69.8%, which is consistent with the true aging state. This significant difference proves the effectiveness of the present invention in eliminating background thermal accumulation interference and accurately capturing early aging characteristics.

[0072] The second aspect of this embodiment provides a life prediction system for explosion-proof lithium-ion batteries used in mining, such as... Figure 5 As shown, the life prediction system for explosion-proof lithium-ion batteries used in mining includes a memory and a processor. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a method for predicting the life of explosion-proof lithium-ion batteries used in mining according to the first aspect of the present invention is implemented.

[0073] The life prediction system for explosion-proof lithium-ion batteries used in mining also includes other components well known to those skilled in the art, such as communication buses and communication interfaces. Their settings and functions are known in the art and will not be described in detail here.

[0074] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as resistive random access memory (DRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (DRAM), high-bandwidth memory, hybrid memory cube, etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device.

[0075] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for predicting the lifespan of explosion-proof lithium-ion batteries used in mining, characterized in that, include: Collect thermal images of the battery surface during the constant voltage stage at the end of charging; The geometric thermal complexity of the thermal image is calculated based on the gray-level extreme value difference in the neighborhood of each pixel, including: The geometric thermal complexity of the thermal image is obtained by summing and averaging the gray-level extreme differences in the neighborhood of all pixels in the thermal image. The pixels with the largest gradient magnitude in the thermal image are designated as high-gradient pixels. The sum of the gradient magnitudes of all pixels in the thermal image is multiplied by the geometric thermal complexity of the thermal image to obtain the initial intensity value. The negative exponential function value of the distance distribution characteristics of the high-gradient pixels is multiplied by the initial intensity value to obtain the thermal intensity factor of the thermal image. The initial fractal dimension of the thermal image is calculated, and the initial fractal dimension is corrected according to the thermal intensity factor of the thermal image to obtain an adaptive fractal dimension. Based on the mapping relationship between the adaptive fractal dimension and the battery health status, the battery life is predicted.

2. The method for predicting the lifespan of explosion-proof lithium-ion batteries for mining according to claim 1, characterized in that, The thermal intensity factor of the thermal image Satisfying the relation: ； in, It is the geometric thermal complexity of the thermal image; , These are the width and height of the thermal image, respectively. , They are pixels The horizontal gradient magnitude and the vertical gradient magnitude; It is the one with the largest gradient amplitude in the thermal image. The average Euclidean distance between pixels It is the preset quantity; This is an adjustment factor for the value. It is a standard normalized function; It is a natural exponential function.

3. The method for predicting the lifespan of explosion-proof lithium-ion batteries for mining according to claim 1, characterized in that, The distance distribution characteristic of the high-gradient pixels is the average Euclidean distance between the high-gradient pixels.

4. The method for predicting the lifespan of explosion-proof lithium-ion batteries for mining according to claim 1, characterized in that, The calculation of the initial fractal dimension of the thermal image includes: The fractal dimension of thermal images is calculated using the box-difference dimension method at multiple preset grid scales.

5. The method for predicting the lifespan of explosion-proof lithium-ion batteries for mining according to claim 1, characterized in that, The step of correcting the initial fractal dimension based on the thermal intensity factor of the thermal image to obtain an adaptive fractal dimension includes: The adaptive fractal dimension is obtained by multiplying the fractal dimension of the thermal image by the thermal intensity factor of the thermal image.

6. The method for predicting the lifespan of explosion-proof lithium-ion batteries for mining according to claim 1, characterized in that, The prediction of battery life based on the mapping relationship between the adaptive fractal dimension and battery health status includes: The adaptive fractal dimension is normalized between a preset baseline dimension and a threshold dimension to obtain the predicted lifespan of the battery.

7. The method for predicting the lifespan of explosion-proof lithium-ion batteries for mining according to claim 1, characterized in that, The thermal image of the battery surface during the constant voltage stage at the end of charging includes: The charging state of the battery is monitored, and a thermal image is acquired when the battery is in the constant voltage charging stage at the end of the charging process.

8. A life prediction system for explosion-proof lithium-ion batteries used in mining, characterized in that, The mining explosion-proof lithium-ion battery life prediction system includes a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a mining explosion-proof lithium-ion battery life prediction method according to any one of claims 1-7 is implemented.

Images