A mine explosion-proof lithium ion battery state monitoring method and system

Abstract

The present application relates to the technical field of data processing, and more particularly to a mine-used explosion-proof lithium ion battery state monitoring method and system. The method comprises the steps of: obtaining the instantaneous health index of the battery in each monitoring period during operation, and obtaining the battery abnormal state factor according to the decline degree of the instantaneous health index; based on the cumulative value and trend efficiency weight of the battery abnormal state factor in the local monitoring window, the deterioration trend persistence of the current monitoring period is calculated; based on the increment of the deterioration trend persistence, combined with the variance of the battery abnormal state factor in the local monitoring window, the deterioration acceleration factor is calculated; the instantaneous health index is input into the pre-trained state classification model to obtain the basic risk value, the basic risk value is corrected by using the deterioration acceleration factor to obtain the final risk factor, and the final risk factor is compared with the preset risk factor threshold to determine whether the battery fails and issue a warning, and the accuracy of early fault warning is improved.

Description

Technical Field

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

[0002] Explosion-proof lithium-ion batteries for mining serve as the core power source for underground trackless rubber-wheeled vehicles, individual soldier equipment, and other mobile devices. Their operational safety and reliability are crucial to mine production efficiency and personnel safety. During long-term use, lithium-ion batteries can experience performance degradation due to electrochemical reactions, changes in internal structure, and external environmental influences, potentially leading to serious safety accidents such as thermal runaway. Therefore, real-time and accurate monitoring and early warning of battery health status are key to ensuring safe production in mines.

[0003] Currently, a mainstream technical approach is to integrate multiple battery operating parameters, such as voltage, current, and temperature, and comprehensively evaluate the battery status through signal processing and feature extraction. For example, by performing frequency domain analysis on each parameter, extracting its spectral features, and further analyzing the coupling correlation and linear correlation between different parameter features, an instantaneous health index that can comprehensively reflect the complex dynamics inside the battery can be constructed. Then, when judging the battery health status based on the instantaneous health index, a single instantaneous health index value is usually calculated in each monitoring cycle and directly input into a classification model to obtain risk factors, which are used to determine whether the battery is operating normally or has malfunctioned.

[0004] However, this method has limitations because its analysis process is isolated. It only focuses on the battery status in the current monitoring cycle. For some slow and gradual early-stage faults, such as the slow consumption of electrolyte, slight deactivation of active materials, or gradual aging of the separator, the instantaneous health indicators may be very weak, with small fluctuations within the normal range over a long period of time. In this case, the method completely ignores the dynamic trend information of the health indicator itself over time, and cannot effectively detect the continuous deterioration trend. This leads to a delay in the discovery of such faults, missing the best time for maintenance and intervention, and failing to achieve effective early prediction. Summary of the Invention

[0005] To address the technical problem that the above methods cannot identify and predict early-stage faults that develop slowly and gradually, this invention provides a method and system for monitoring the state of explosion-proof lithium-ion batteries used in mining.

[0006] In a first aspect, the present invention provides a method for monitoring the state of explosion-proof lithium-ion batteries used in mining, employing the following technical solution:

[0007] A method for monitoring the state of explosion-proof lithium-ion batteries used in mining includes the following steps:

[0008] The terminal voltage, charge / discharge current and surface temperature data sequences of the explosion-proof lithium-ion battery for mining are obtained and analyzed in each monitoring cycle to obtain the instantaneous health index for each monitoring cycle. The instantaneous health index represents the battery health level in each monitoring cycle.

[0009] A local monitoring window and a reference monitoring window are constructed based on the current monitoring cycle and several previous monitoring cycles; the battery abnormal state factor is obtained based on the degree of decrease in the mean of the instantaneous health indicators of the local monitoring window relative to the reference monitoring window; the persistence of the deterioration trend in the current monitoring cycle is obtained based on the cumulative value and trend efficiency weight of the battery abnormal state factor within the local monitoring window.

[0010] Calculate the increment of the deterioration trend persistence of the current monitoring period relative to the previous monitoring period, and obtain the deterioration acceleration factor by combining the variance of the battery abnormal state factor within the local monitoring window; input the instantaneous health index into the pre-trained state classification model to obtain the basic risk value of the current monitoring period, and use the deterioration acceleration factor to perform exponential correction on the basic risk value to obtain the final risk factor.

[0011] The final risk factor is compared with the preset risk factor threshold to determine whether the battery has malfunctioned and to issue a warning.

[0012] The innovation of this invention lies in analyzing the development of instantaneous health indicators to obtain a deterioration acceleration factor. This deterioration acceleration factor is then used to exponentially correct the basic risk value to obtain the final risk factor. This invention can keenly capture the acceleration of fault development and exponentially increase the final risk factor of early-stage hidden faults where the basic risk value is still within the normal range, thereby achieving early warning of early faults.

[0013] Preferably, the step of constructing a local monitoring window and a reference monitoring window based on the current monitoring cycle and several previous monitoring cycles includes:

[0014] The number of historical monitoring periods is preset to N. The window formed by the t-th monitoring period and the N previous monitoring periods is used as the local monitoring window of the t-th monitoring period. The (N+1)-th monitoring period before the t-th monitoring period is used as the target monitoring period. The window formed by the target monitoring period and the N previous monitoring periods is used as the reference monitoring window of the t-th monitoring period.

[0015] Preferably, obtaining the battery abnormal state factor includes:

[0016] ;

[0017] In the formula, The abnormal state factor of the battery in the t-th monitoring period; This represents the mean of instantaneous health indicators across all monitoring periods within the local monitoring window for the t-th monitoring period; The mean of instantaneous health indicators for all monitoring periods in the reference monitoring window for the t-th monitoring period; This represents the number of monitoring periods within the local monitoring window of the t-th monitoring period.

[0018] Preferably, obtaining the persistence of the deterioration trend in the current monitoring period includes:

[0019] ;

[0020] In the formula, This represents the persistence of the deteriorating trend during the t-th monitoring period; The battery abnormal state factor represents the i-th monitoring period within the local monitoring window of the t-th monitoring period; The number of monitoring periods in the local monitoring window of the t-th monitoring period; || represents the absolute value symbol; Represents the preset nonlinear adjustment coefficient; This represents the preset hyperparameters.

[0021] By constructing a persistent indicator of deteriorating trends and using trend efficiency weights to suppress oscillation noise, it is possible to effectively filter out instantaneous data fluctuations caused by road bumps, thereby significantly reducing the false alarm rate.

[0022] Preferably, obtaining the deterioration acceleration factor includes:

[0023] ;

[0024] In the formula, The deterioration acceleration factor represents the t-th monitoring period; Represents the increment of the deterioration trend in the t-th monitoring period; || represents the absolute value sign, in order to preserve the positive or negative sign of the numerator; The variance of the battery abnormal state factor represents the variance of all monitoring cycles within the local monitoring window of the t-th monitoring cycle. This represents the preset hyperparameters.

[0025] Preferably, the step of exponentially correcting the base risk value using a deterioration acceleration factor to obtain the final risk factor includes:

[0026] ;

[0027] In the formula, Represents the final risk factor for the t-th monitoring period; This represents the baseline risk value for the t-th monitoring period; This represents the preset risk gain coefficient; represents the deterioration acceleration factor in the t-th monitoring period; max() represents the maximum value function.

[0028] By introducing a deterioration acceleration factor to correct the basic risk value, the final risk factor of early-stage hidden faults with basic risk values ​​within the normal range can be increased exponentially, providing early warning.

[0029] Preferably, the step of comparing the final risk factor with a preset risk factor threshold to determine whether the battery has malfunctioned and to issue a warning includes:

[0030] A preset risk factor threshold T1 is set. When the final risk factor in the t-th monitoring period is greater than or equal to T1, the battery malfunctions in the t-th monitoring period, and the system issues an early warning.

[0031] By setting reasonable thresholds, fault diagnosis can be completed automatically, providing clear early warnings for maintenance personnel.

[0032] Preferably, obtaining the instantaneous health indicators for each monitoring period includes:

[0033] The Pearson correlation coefficient between the j-th parameter data sequence and the k-th parameter data sequence in the t-th monitoring period is denoted as the correlation between the j-th parameter data and the k-th parameter data in the t-th monitoring period. Based on the correlation between all parameter pairs in the t-th monitoring period, a linear correlation matrix for the t-th monitoring period is constructed, and the maximum eigenvalue of the linear correlation matrix is ​​extracted and denoted as the global linear correlation factor for the t-th monitoring period.

[0034] ;

[0035] In the formula, This represents the instantaneous health indicator during the t-th monitoring period; This represents the global dynamic coupling index in the t-th monitoring period; This represents the global linear correlation factor in the t-th monitoring period.

[0036] By converting time-domain signals into frequency-domain features and combining the linear relationships between various parameter data, it is possible to reveal internal battery faults at a deeper level and construct instantaneous health indicators for each monitoring cycle.

[0037] Preferably, the acquisition of the global dynamic coupling index in the t-th monitoring period includes:

[0038] The fast Fourier transform is used to transform the data sequence of each parameter in each monitoring period to obtain the spectrum of each parameter data in each monitoring period, and then the harmonic energy ratio of each parameter data in each monitoring period is obtained; the similarity of the harmonic energy ratio of the j-th parameter data and the k-th parameter data in the t-th monitoring period is obtained, and the mutual information of the spectrum of the j-th parameter data and the spectrum of the k-th parameter data in the t-th monitoring period is obtained. Based on the product of the similarity and the mutual information, the coupling correlation degree between the j-th parameter data and the k-th parameter data in the t-th monitoring period is obtained.

[0039] The average value of the coupling correlation between all two parameter data in the t-th monitoring period is used as the global dynamic coupling index in the t-th monitoring period.

[0040] Secondly, this invention provides a state monitoring system for explosion-proof lithium-ion batteries used in mining, employing the following technical solution:

[0041] A mining explosion-proof lithium-ion battery status monitoring system includes a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned mining explosion-proof lithium-ion battery status monitoring method is implemented.

[0042] By adopting the above technical solution, a computer program for monitoring the state of an explosion-proof lithium-ion battery used in mining is generated and stored in a memory so that it can be loaded and executed by a processor. This allows for the creation of a terminal device based on the memory and processor, making it convenient to use.

[0043] The present invention has the following technical effects: By analyzing the development of instantaneous health indicators, the present invention obtains the deterioration acceleration factor, and uses the deterioration acceleration factor to exponentially correct the basic risk value to obtain the final risk factor. It can keenly capture the acceleration of fault development and exponentially increase the final risk factor of early hidden faults where the basic risk value is still within the normal range, thereby achieving early warning of early faults. Furthermore, based on the incremental increase of the deterioration trend of the current monitoring cycle relative to the previous monitoring cycle, combined with the variance of the battery abnormal state factor within the local monitoring window, the deterioration acceleration factor is obtained, enabling the system to distinguish between normal uniform aging (low risk) and abnormal accelerated deterioration of early faults. Attached Figure Description

[0044] Figure 1 This is a flowchart of a method for monitoring the state of an explosion-proof lithium-ion battery for mining, according to an embodiment of the present invention.

[0045] Figure 2 A comparison chart of health status monitoring and early warning response for mining batteries. Detailed Implementation

[0046] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0047] This invention discloses a method for monitoring the state of explosion-proof lithium-ion batteries used in mining, referring to... Figure 1 This includes steps S1-S4:

[0048] S1: Obtain the instantaneous health indicators of the battery during each monitoring cycle in operation.

[0049] In this embodiment of the invention, every second is a sampling moment, and every five minutes is a monitoring cycle; the terminal voltage data, charge / discharge current data and surface temperature data of the mining explosion-proof lithium-ion battery during operation are obtained by using voltage sensor, current sensor and temperature sensor, to obtain the terminal voltage data sequence, charge / discharge current data sequence and surface temperature data sequence for each monitoring cycle.

[0050] It should be noted that this step requires the fusion of multiple battery parameter data, signal processing, and feature extraction to comprehensively evaluate the battery status. By first converting the battery's multiple parameter data sequences in each monitoring cycle to the frequency domain, the spectrum of each parameter data and the harmonic energy ratio that reflects the internal structure of the signal are extracted. Then, based on the similarity of the harmonic energy ratios of the parameter data and the spectral mutual information, a global dynamic coupling index is obtained. Next, the interaction between the parameter data is measured from the perspective of linear correlation. Finally, the global dynamic coupling index and the global linear correlation factor are combined to form an instantaneous health index, realizing a rapid and comprehensive assessment of the battery's health status.

[0051] In this embodiment of the invention, the fast Fourier transform is used to transform the j-th parameter data sequence in the t-th monitoring period to obtain the j-th parameter data spectrum in the t-th monitoring period, and the harmonic energy ratio of the j-th parameter data spectrum in the t-th monitoring period at the dominant frequency is extracted and denoted as the harmonic energy ratio of the j-th parameter data in the t-th monitoring period.

[0052] For the j-th parameter data and the k-th parameter data in the t-th monitoring period, the expression for calculating the coupling correlation between the j-th parameter data and the k-th parameter data is as follows:

[0053] ;

[0054] In the formula, This represents the degree of coupling between the j-th parameter data and the k-th parameter data in the t-th monitoring period; This represents the harmonic energy ratio of the j-th parameter data in the t-th monitoring period; This represents the harmonic energy ratio of the k-th parameter data in the t-th monitoring period; This indicates that the maximum value is retrieved. This represents the spectrum of the j-th parameter data during the t-th monitoring period; This represents the spectrum of the k-th parameter data during the t-th monitoring period; This indicates the acquisition of mutual information between two spectral distributions; The similarity of the harmonic energy ratios of the parameter data;

[0055] The mean of the coupling correlation between all pairwise parameter data in the t-th monitoring period is denoted as the global dynamic coupling index in the t-th monitoring period.

[0056] When an internal battery failure occurs, the spectrum of parameter data changes synchronously, mutual information increases, and the harmonic energy ratio tends to be consistent, leading to... The increase of leads to an increase in the global dynamic coupling index.

[0057] The Pearson correlation coefficient between the j-th parameter data sequence and the k-th parameter data sequence in the t-th monitoring period is denoted as the correlation between the j-th parameter data and the k-th parameter data in the t-th monitoring period. Based on the correlation between all pairwise parameter data in the t-th monitoring period, a linear correlation matrix for the t-th monitoring period is constructed, and the maximum eigenvalue of the linear correlation matrix is ​​extracted and denoted as the global linear correlation factor for the t-th monitoring period.

[0058] The maximum eigenvalue combines the correlation between all pairwise parameter data into a global value, directly quantifying the integrity of the linear causal chain inside the battery. Its value reflects whether the parameter data still maintains normal linear synchronization. When the battery has an internal fault, the parameter data will change synchronously. Therefore, the larger the value, the less healthy the battery is.

[0059] The formula for calculating the instantaneous health indicators in the t-th monitoring period is:

[0060] ;

[0061] In the formula, This represents the instantaneous health indicator during the t-th monitoring period; This represents the global dynamic coupling index in the t-th monitoring period; This represents the global linear correlation factor in the t-th monitoring period. The higher the value, the less healthy the battery is.

[0062] Similarly, instantaneous health indicators for each monitoring period are obtained.

[0063] S2: Construct a local monitoring window and a reference monitoring window based on the current monitoring period and several previous monitoring periods; calculate the degree of decrease in the mean of instantaneous health indicators of the local monitoring window relative to the reference monitoring window to obtain the battery abnormal state factor; calculate the persistence of the deterioration trend in the current monitoring period based on the cumulative value of the battery abnormal state factor and the trend efficiency weight within the local monitoring window.

[0064] It should be noted that, in order to overcome the limitations of single analysis of instantaneous health indicators, this invention analyzes the generated instantaneous health indicator sequence. It is known that for some slow and progressive early faults, such as slow consumption of electrolyte, slight deactivation of active materials, or gradual aging of the separator, the fault will cause the instantaneous health indicators of the battery to increase continuously in the continuous monitoring cycle. Therefore, based on this feature, the abnormal state factor of the battery in each monitoring cycle is obtained.

[0065] In this embodiment of the invention, the number of historical monitoring periods is preset to N=4, and the window formed by the t-th monitoring period and the N previous monitoring periods is used as the local monitoring window of the t-th monitoring period.

[0066] The N+1th monitoring period before the tth monitoring period is taken as the target monitoring period, and the window formed by the target monitoring period and the N previous monitoring periods is taken as the reference monitoring window for the tth monitoring period.

[0067] Obtain the battery abnormal state factor for the t-th monitoring period:

[0068] ;

[0069] In the formula, The abnormal state factor of the battery in the t-th monitoring period; This represents the mean of instantaneous health indicators across all monitoring periods within the local monitoring window for the t-th monitoring period; The mean of instantaneous health indicators for all monitoring periods in the reference monitoring window for the t-th monitoring period; This represents the number of monitoring periods in the local monitoring window of the t-th monitoring period; it should be noted that the number of monitoring periods in the local monitoring window of the t-th monitoring period is equal to the number of monitoring periods in the reference monitoring window; When the value is positive, it indicates that the instantaneous health indicator has increased within its local monitoring window compared to the reference monitoring window. The higher the value, the more significant the deterioration of battery health within a single monitoring cycle.

[0070] It should be noted that the battery abnormal state factor in a single monitoring cycle can reflect the degree of deterioration of battery health within that cycle, but it is extremely sensitive and easily affected by the working conditions in the mine (such as sudden starts and stops of locomotives and contact resistance jumps caused by road bumps), generating large positive and negative oscillation noise, which in turn causes the battery abnormal state factor to fluctuate wildly between positive and negative. Therefore, in order to avoid these oscillation noises triggering false alarms, this step constructs the persistence of the deterioration trend for each monitoring cycle. Its core objective is: only when the battery abnormal state factor is continuously positive for a period of time is it considered that the battery condition is deteriorating and accumulated. Furthermore, if the battery abnormal state factor fluctuates wildly between positive and negative for a period of time, it is considered noise and suppressed.

[0071] In this embodiment of the invention, the persistence of the deterioration trend in the t-th monitoring period is obtained:

[0072] ;

[0073] In the formula, This represents the persistence of the deteriorating trend during the t-th monitoring period; The battery abnormal state factor represents the i-th monitoring period within the local monitoring window of the t-th monitoring period; The number of monitoring periods in the local monitoring window of the t-th monitoring period; || represents the absolute value symbol; This represents a preset nonlinear adjustment coefficient. In this embodiment of the invention, the preset value is... Its purpose is to adjust the intensity of noise suppression; Representing preset hyperparameters, in this embodiment of the invention, the preset... This is used to prevent the denominator from being 0; Represents trend efficiency weights;

[0074] When the battery abnormal state factor remains positive for all monitoring cycles in the local monitoring window of the t-th monitoring cycle, The larger the value of , the more the denominator in the trend efficiency weight is almost equal to the numerator, and the value of the trend efficiency weight approaches 1, thus making The larger the value, the more likely the battery abnormal state factor is to remain positive for a period of time, and the battery abnormal state factor does not fluctuate drastically between positive and negative.

[0075] When the battery abnormal state factor fluctuates positively or negatively across all monitoring periods within the local monitoring window of the t-th monitoring period, the positive and negative values ​​will cancel each other out when added together. The value of approaches 0, and the numerator of the trend efficiency weight approaches 0 while the denominator is larger, causing the trend efficiency weight to approach 0, thus making The value approaches 0, indicating that the working conditions in the mine cause large-scale positive and negative oscillation noise, which in turn leads to violent positive and negative fluctuations in the battery abnormal state factor.

[0076] S3: Calculate the increment of the deterioration trend persistence of the current monitoring period relative to the previous monitoring period, and calculate the deterioration acceleration factor by combining the variance of the battery abnormal state factor within the local monitoring window; input the instantaneous health indicators into the pre-trained state classification model to obtain the basic risk value, and use the deterioration acceleration factor to exponentially correct the basic risk value to obtain the final risk factor.

[0077] It should be noted that while the persistence of the degradation trend can identify whether a battery is experiencing a continuous degradation trend, simply knowing that the battery is deteriorating is not enough. Batteries that deteriorate at a uniform rate reflect normal battery aging and have a relatively low risk. In other words, the difference in the persistence of the degradation trend of batteries that deteriorate at a uniform rate is small between each monitoring cycle. However, for some slow and gradual early-stage faults (such as slow consumption of electrolyte, slight deactivation of active materials, or gradual aging of the separator), the battery will deteriorate more rapidly. That is, the persistence of the degradation trend of the battery at each monitoring time is constantly increasing. Therefore, the goal of this step is to obtain the degradation acceleration factor for each monitoring cycle based on the growth of the persistence of the degradation trend.

[0078] In this embodiment of the invention, the value of the deterioration trend persistence in the t-th monitoring period minus the value of the deterioration trend persistence in the (t-1)-th monitoring period is used as the deterioration trend increment in the t-th monitoring period.

[0079] Obtain the deterioration acceleration factor for the t-th monitoring period:

[0080] ;

[0081] In the formula, The deterioration acceleration factor represents the t-th monitoring period; Represents the increment of the deterioration trend in the t-th monitoring period; || represents the absolute value sign, in order to preserve the positive or negative sign of the numerator; The variance of the battery abnormal state factor represents the variance of all monitoring cycles within the local monitoring window of the t-th monitoring cycle. Representing preset hyperparameters, in this embodiment of the invention, the preset... This is used to prevent the denominator from being 0;

[0082] If the increment of the deteriorating trend is positive and increases, then... The larger the value of , the faster the risk factor can be improved. Furthermore, the smaller the variance of the battery abnormal state factor across all monitoring periods within the local monitoring window of the t-th monitoring period, the more stable the trend of the battery abnormal state factor is over a period of time. The larger the value, the more likely the battery has experienced a stable, accelerated deterioration fault at the t-th monitoring cycle;

[0083] If the increment of the deteriorating trend is positive and increases, then... The larger the value of , the faster the risk level can be increased. Furthermore, a larger variance in the battery abnormal state factor across all monitoring periods within the local monitoring window of the t-th monitoring period indicates greater fluctuation in the battery abnormal state factor over a period of time. This suggests that a single increment may be a random fluctuation, thus suppressing . The value of .

[0084] If the increment of the deteriorating trend is negative... If the value is negative, the system will correctly determine that the risk is decreasing, rather than issuing an incorrect alarm.

[0085] It should be noted that when judging battery risk based on instantaneous health indicators, the instantaneous health indicators of each monitoring period are usually directly input into the trained model to obtain the basic risk value, and then the battery's current state category is determined based on the basic risk value, such as "normal" or "fault". However, this method only focuses on the battery state in the current monitoring period and is inaccurate for monitoring some slow and progressive early faults. Therefore, this invention increases the basic risk value based on the deterioration acceleration factor, thereby realizing early prediction and warning of faults.

[0086] In this embodiment of the invention, the instantaneous health index of the t-th monitoring period is input into a trained state classification model to obtain the basic risk value for that monitoring period; wherein the state classification model used in this embodiment is a support vector machine, and the method for training the state classification model is as follows:

[0087] A large number of instantaneous health indicators of mining explosion-proof lithium-ion batteries under normal operating conditions and under fault conditions are collected. Each instantaneous health indicator is manually labeled with a basic risk value; that is, if the mining explosion-proof lithium-ion battery is in normal operating condition, the basic risk value is set to 0.2; if the mining explosion-proof lithium-ion battery is in fault condition, the basic risk value is set to 0.8. This labeling result is recorded as the label of each instantaneous health indicator. A large number of instantaneous health indicators of mining explosion-proof lithium-ion batteries under normal operating conditions and under fault conditions, along with their corresponding labels, are used as a dataset. This dataset is used to train the state classification model. The specific training process is well-known in state classification models, and this embodiment will not elaborate on the specific training process.

[0088] Obtain the final risk factor for the t-th monitoring period:

[0089] ;

[0090] In the formula, Represents the final risk factor for the t-th monitoring period; This represents the baseline risk value for the t-th monitoring period; This represents a preset risk gain coefficient, used to control the amplification effect of the deterioration acceleration factor on the risk factor. In this embodiment of the invention, the preset... ; represents the deterioration acceleration factor in the t-th monitoring period; max() represents the maximum value function;

[0091] When deterioration accelerates factors When the value is zero or negative (no accelerated degradation). The value is close to 1, and the final risk level is mainly determined by the base risk value.

[0092] When deterioration accelerates factors When a significant positive value appears, This will grow exponentially, thus greatly amplifying the final risk factor. This design means that even if the battery is judged to be normal based on the baseline risk value, the deterioration acceleration factor will increase. It can also elevate the final risk factor to the battery failure level, thereby enabling early prediction and warning of early failures that develop slowly and gradually.

[0093] S4: Compare the final risk factor with the preset risk factor threshold. If the final risk factor exceeds the preset risk factor threshold, determine that the battery has malfunctioned and issue a warning.

[0094] In this embodiment of the invention, a preset risk factor threshold T1=0.75 is set. If the final risk factor of the t-th monitoring period is greater than or equal to T1, the battery malfunctions in the t-th monitoring period, and the system issues an early warning.

[0095] Figure 2 The comparison of the health status monitoring and early warning response of mining batteries using the present invention and traditional instantaneous monitoring methods shows that the gray dashed line reflects the trend of risk factors changing with the monitoring cycle under the traditional instantaneous monitoring method, while the red solid line reflects the trend of the final risk factors changing with the monitoring cycle in the embodiment of the present invention. At approximately the 60th monitoring cycle, the latent potential of the fault begins to emerge. The gray dashed line only shows a slow rise accompanied by violent fluctuations (underground interference factors), and does not reach the risk factor threshold for a long time. However, the red solid line (the present invention) quickly identifies the deterioration acceleration factor after the fault begins, and the curve rises exponentially, reaching the risk factor threshold first at approximately the 90th monitoring cycle. The gray dashed line does not reach the risk factor threshold until approximately the 180th monitoring cycle. This proves that the present invention can issue an early warning in the early stage of the fault (before the traditional method reacts), and the curve is smoother and has stronger anti-interference ability.

[0096] This invention also discloses a mining explosion-proof lithium-ion battery status monitoring system, including 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 status monitoring method provided by this invention is implemented.

[0097] The system also includes other components well-known to those skilled in the art, such as communication buses and communication interfaces, the setup and functions of which are known in the art and will not be described in detail here. In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0098] 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 monitoring the state of explosion-proof lithium-ion batteries used in mining, characterized in that, include: The terminal voltage, charge / discharge current and surface temperature data sequences of the explosion-proof lithium-ion battery for mining are obtained and analyzed in each monitoring cycle to obtain the instantaneous health index for each monitoring cycle. The instantaneous health index represents the battery health level in each monitoring cycle. Construct local monitoring windows and reference monitoring windows based on the current monitoring cycle and its previous monitoring cycles; The battery abnormal state factor is obtained based on the degree of decrease in the mean of instantaneous health indicators of the local monitoring window relative to the reference monitoring window. Based on the cumulative value of the abnormal battery state factor and the trend efficiency weight within the local monitoring window, the persistence of the deterioration trend in the current monitoring period is obtained. Calculate the increment of the deterioration trend persistence of the current monitoring period relative to the previous monitoring period, and obtain the deterioration acceleration factor by combining the variance of the battery abnormal state factor within the local monitoring window; input the instantaneous health index into the pre-trained state classification model to obtain the basic risk value of the current monitoring period, and use the deterioration acceleration factor to perform exponential correction on the basic risk value to obtain the final risk factor. The final risk factor is compared with the preset risk factor threshold to determine whether the battery has malfunctioned and to issue a warning.

2. The method for monitoring the state of an explosion-proof lithium-ion battery for mining according to claim 1, characterized in that, The construction of a local monitoring window and a reference monitoring window based on the current monitoring cycle and several previous monitoring cycles includes: The number of historical monitoring periods is preset to N. The window formed by the t-th monitoring period and the N previous monitoring periods is used as the local monitoring window of the t-th monitoring period. The (N+1)-th monitoring period before the t-th monitoring period is used as the target monitoring period. The window formed by the target monitoring period and the N previous monitoring periods is used as the reference monitoring window of the t-th monitoring period.

3. A method for monitoring the state of an explosion-proof lithium-ion battery for mining according to claim 1 or 2, characterized in that, The acquisition of battery abnormal state factors includes: ； In the formula, The abnormal state factor of the battery in the t-th monitoring period; This represents the mean of instantaneous health indicators across all monitoring periods within the local monitoring window for the t-th monitoring period; The mean of instantaneous health indicators for all monitoring periods in the reference monitoring window for the t-th monitoring period; This represents the number of monitoring periods within the local monitoring window of the t-th monitoring period.

4. The method for monitoring the state of an explosion-proof lithium-ion battery for mining according to claim 1, characterized in that, The acquisition of the persistence of the deterioration trend in the current monitoring period includes: ； In the formula, This represents the persistence of the deteriorating trend during the t-th monitoring period; The battery abnormal state factor represents the i-th monitoring period within the local monitoring window of the t-th monitoring period; The number of monitoring periods in the local monitoring window of the t-th monitoring period; || represents the absolute value symbol; Represents the preset nonlinear adjustment coefficient; This represents the preset hyperparameters.

5. The method for monitoring the state of an explosion-proof lithium-ion battery for mining according to claim 1, characterized in that, The acquisition of the deterioration acceleration factor includes: ； In the formula, The deterioration acceleration factor represents the t-th monitoring period; Represents the increment of the deterioration trend in the t-th monitoring period; || represents the absolute value sign, in order to preserve the positive or negative sign of the numerator; The variance of the battery abnormal state factor represents the variance of all monitoring cycles within the local monitoring window of the t-th monitoring cycle. This represents the preset hyperparameters.

6. The method for monitoring the state of an explosion-proof lithium-ion battery for mining according to claim 1, characterized in that, The process of exponentially correcting the base risk value using a deterioration acceleration factor to obtain the final risk factor includes: ； In the formula, The final risk factor represents the monitoring period t. This represents the baseline risk value for the t-th monitoring period; This represents the preset risk gain coefficient; represents the deterioration acceleration factor in the t-th monitoring period; max() represents the maximum value function.

7. The method for monitoring the state of an explosion-proof lithium-ion battery for mining according to claim 1, characterized in that, The step of comparing the final risk factor with a preset risk factor threshold to determine whether the battery has malfunctioned and to issue a warning includes: A preset risk factor threshold T1 is set. When the final risk factor in the t-th monitoring period is greater than or equal to T1, the battery malfunctions in the t-th monitoring period, and the system issues an early warning.

8. The method for monitoring the state of an explosion-proof lithium-ion battery for mining according to claim 1, characterized in that, The acquisition of instantaneous health indicators for each monitoring period includes: The Pearson correlation coefficient between the j-th parameter data sequence and the k-th parameter data sequence in the t-th monitoring period is denoted as the correlation between the j-th parameter data and the k-th parameter data in the t-th monitoring period. Based on the correlation between all parameter pairs in the t-th monitoring period, a linear correlation matrix for the t-th monitoring period is constructed, and the maximum eigenvalue of the linear correlation matrix is ​​extracted and denoted as the global linear correlation factor for the t-th monitoring period. ； In the formula, This represents the instantaneous health indicator during the t-th monitoring period; This represents the global dynamic coupling index in the t-th monitoring period; This represents the global linear correlation factor in the t-th monitoring period.

9. A method for monitoring the state of an explosion-proof lithium-ion battery for mining according to claim 8, characterized in that, The acquisition of the global dynamic coupling index in the t-th monitoring period includes: The fast Fourier transform is used to transform the data sequence of each parameter in each monitoring period to obtain the spectrum of each parameter data in each monitoring period, and then the harmonic energy ratio of each parameter data in each monitoring period is obtained; the similarity of the harmonic energy ratio of the j-th parameter data and the k-th parameter data in the t-th monitoring period is obtained, and the mutual information of the spectrum of the j-th parameter data and the spectrum of the k-th parameter data in the t-th monitoring period is obtained. Based on the product of the similarity and the mutual information, the coupling correlation degree between the j-th parameter data and the k-th parameter data in the t-th monitoring period is obtained. The average value of the coupling correlation between all two parameter data in the t-th monitoring period is used as the global dynamic coupling index in the t-th monitoring period.

10. A state monitoring system for explosion-proof lithium-ion batteries used in mining, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement a method for monitoring the state of an explosion-proof lithium-ion battery for mining as described in any one of claims 1-9.

Images