Real-time vehicle maintenance monitoring management method based on Internet of Things technology

Abstract

The invention discloses a real-time vehicle maintenance monitoring management method based on the Internet of Things technology, and relates to the technical field of vehicle maintenance monitoring management, and the method comprises the following steps: firstly, measuring and recording the operation parameters of a battery pack according to a preset time interval; according to the real-time vehicle maintenance monitoring method based on the Internet of Things technology, the battery performance is finely monitored, and tiny degradation changes are recognized in time. The system collects key features, dynamically adjusts the detection frequency in combination with machine learning, and captures a potential degradation problem. Different from traditional monitoring, the system intelligently classifies battery performance changes, flexibly adjusts maintenance strategies, improves monitoring efficiency, reduces fault risks, prolongs the service life of the battery, optimizes maintenance management, and improves vehicle safety and reliability.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle maintenance monitoring and management, and particularly to a real-time vehicle maintenance monitoring and management method based on Internet of Things technology. Background Art

[0002] Real-time vehicle maintenance monitoring and management based on Internet of Things technology refers to connecting various sensors, control modules, and systems in an electric vehicle (EV) to a cloud platform or a local server through Internet of Things devices to achieve real-time monitoring and maintenance management of the health status of the electric vehicle. These sensors can collect various vehicle data in real time, such as battery voltage, temperature, charging status, operating status of the drive system, tire pressure, etc., and transmit them to the cloud for analysis and processing. Based on Internet of Things technology, the system can evaluate the health status of the battery, the efficiency of the driving system, and the operating status of other key components of the vehicle, identify potential faults in advance, and automatically adjust the vehicle's usage strategy or issue maintenance reminders when necessary, so as to achieve fault warning, optimize the maintenance plan, extend the vehicle's service life, and ensure the operation safety and reliability of the electric vehicle.

[0003] The existing technologies have the following deficiencies:

[0004] Most battery management systems (BMS) monitor parameters such as battery voltage, current, and temperature at fixed time intervals. This method can provide basic battery operation data for analysis and diagnosis. However, some individual batteries in the battery pack may gradually lose capacity or show slight performance degradation, but the voltage, current, and temperature of the overall battery pack are still within the normal range. These minor changes are difficult to detect through fixed-frequency monitoring because their impact on the overall system performance is not significant. If these problems are not discovered in time, serious consequences may occur. As the individual batteries degrade, the voltage imbalance problem inside the battery pack may intensify. Although the overall voltage of the battery pack is normal, the voltage difference between different individual batteries gradually increases, thus limiting the total capacity of the battery pack because the performance of the battery pack is usually determined by the individual battery with the lowest voltage. When the battery pack is unbalanced, the charging efficiency will decrease, and some batteries may be overcharged or over-discharged, thereby causing battery damage and even thermal runaway. Although the overall voltage of the battery pack is still normal, the battery management system may not be able to identify the inefficient state during the charging process in time, thus increasing the risk of battery damage.

[0005] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and thus it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The object of the present invention is to provide a real-time vehicle maintenance and monitoring management method based on Internet of Things technology. By introducing the real-time vehicle maintenance and monitoring management method of Internet of Things technology, it is possible to finely monitor the battery performance and timely identify minor degradation changes. This solution captures key features (such as polarity drift index, battery internal electric field change index), combines with a machine learning model to dynamically adjust the detection frequency, and effectively captures potential degradation problems of the battery pack. Different from traditional fixed-frequency monitoring, the system can intelligently classify battery performance changes, flexibly adjust the maintenance strategy according to the evaluation results, improve the monitoring efficiency, reduce the failure risk, extend the battery life, thereby optimizing the maintenance management, enhancing the vehicle safety and reliability, so as to solve the problems in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: A real-time vehicle maintenance and monitoring management method based on Internet of Things technology, comprising the following steps:

[0008] First, measure and record the operating parameters of the battery pack at a preset time interval.

[0009] Organize the obtained operating parameters of the battery pack into a systematic analysis set, and in the analysis set, identify and extract the key features reflecting the potential degradation of the single battery performance.

[0010] Deeply analyze the extracted key features under the detection window, and input the analyzed key features into a pre-trained machine learning model. Through the machine learning model, intelligently evaluate the input key features, and identify and output the abnormal change situation of the battery pack performance.

[0011] Based on the evaluation result of the machine learning model, classify the performance change situation of the battery pack into normal change and potential degradation.

[0012] For the battery pack with normal change, continue to monitor and collect data at the preset fixed time interval to ensure that the battery pack is always in good operating condition.

[0013] For the battery pack with potential degradation, according to the evaluation result of the machine learning model, through the fuzzy logic control algorithm, dynamically adjust the detection frequency, timely capture the abnormal change, and avoid the accumulation of minor anomalies into serious faults.

[0014] Preferably, in the analysis set, key features reflecting potential degradation of the single cell performance are identified and extracted. The extracted features include the change in the electrochemical balance between the positive and negative electrodes of the battery and the distribution of the internal electric field of the battery varying with time. Under the detection window, the change in the electrochemical balance between the positive and negative electrodes of the battery and the distribution of the internal electric field of the battery varying with time obtained are analyzed to generate a polarity drift index and an internal electric field change index of the battery respectively. The polarity drift index quantifies the slight change in the electrochemical balance between the positive and negative electrodes of the battery; the internal electric field change index of the battery quantifies the slight change in the distribution of the internal electric field of the battery with time.

[0015] Preferably, the polarity drift index and the internal electric field change index of the battery generated after analyzing the extracted key features are input into a pre-trained machine learning model. The battery performance change coefficient is generated through the machine learning model, and the abnormal change condition of the battery pack performance is identified and output through the battery performance change coefficient.

[0016] Preferably, the battery performance change coefficient generated after analyzing the extracted features is compared and analyzed with the preset battery performance change coefficient, and the performance change condition of the battery pack is divided. The division steps are as follows:

[0017] If the battery performance change coefficient is greater than or equal to the preset battery performance change coefficient reference threshold, the performance change of the battery pack is divided into potential degradation;

[0018] If the battery performance change coefficient is less than the preset battery performance change coefficient reference threshold, the performance change of the battery pack is divided into normal change.

[0019] Preferably, for the battery pack with potential degradation, according to the evaluation result of the machine learning model, through the fuzzy logic control algorithm, the detection frequency is dynamically adjusted to capture the abnormal change in time and avoid the accumulation of minor anomalies into serious faults. The specific steps are as follows:

[0020] Establish a fuzzy rule set based on the input and output. Input the battery performance change coefficient BPCC and the battery performance change coefficient reference threshold BPCC ref , and output the adjusted detection frequency. The calculation expression is as follows:

[0021]

[0022] f new = f base +Δf

[0023] , where f base is the preset detection frequency, r is the adjustment ratio factor, ω is the non-linear adjustment parameter, is the sensitivity parameter, Δf is the adjustment amount of the detection frequency, f newis the new detection frequency after adjustment;

[0024] Convert the input parameters of the battery performance change coefficient BPCC and the battery performance change coefficient reference threshold BPCC ref into fuzzy sets, represented by two fuzzy levels of "low" and "high", and define the membership function μ(BPCC) to represent the membership degree of each level;

[0025]

[0026] , where μ Low (BPCC) is the membership degree of low risk, x max is the maximum threshold of the battery performance change coefficient, and μ High (BPCC) is the membership degree of high risk;

[0027] Based on fuzzy rules, combine fuzzy sets and the fuzzy rule base for reasoning, and adjust the detection frequency adjustment amount. The calculation expression is as follows:

[0028] where ΔF is the adjusted detection frequency adjustment amount, and H is the amplification factor, which is used to further amplify the influence in the battery degradation state;

[0030] Through defuzzification, convert the result of fuzzy reasoning into a specific numerical output, that is, the final detection frequency. The expression is as follows:

[0031]

[0032] , where F new is the final detection frequency, ΔF Low is the detection frequency adjustment amount at low risk, and ΔF High is the detection frequency adjustment amount at high risk.

[0033] Preferably, under the detection window, analyze the electrochemical balance change between the positive and negative electrodes of the battery, and the specific steps to generate the polarity drift index are as follows:

[0034] Under the detection window, first, based on the detection frequency preset by the battery management system, collect the operating parameters of the battery pack at different time points. After denoising and normalizing the collected parameters, extract the key parameters. The extracted key parameters include the battery terminal voltage, current, and temperature, and calibrate the extracted battery terminal voltage as V battery (t), representing the voltage value at both ends of the battery at time point t, calibrate the current as I(t), representing the current at time point t, and calibrate the temperature as T(t), representing the temperature at time point t;

[0035] An electrochemical equilibrium change model is established to describe the dynamic process of the electrochemical reaction inside the battery. Based on the equations of the physical and chemical processes inside the battery, the voltage-time relationship of the battery is derived, and the model is constructed by combining the effects of current and temperature on the reaction rate inside the battery. The specific model expression is as follows:

[0036]

[0037] , where V diff (t) is the change in the voltage difference between the positive and negative electrodes of the battery, representing the voltage difference between the positive and negative electrodes of the battery at time point t. V ref is the reference voltage, and α 1 , α 2 and α 3 are all weight coefficients. β 1 is the electrochemical reaction rate adjustment coefficient, t 1 is the start time of the detection window;

[0038] Based on the change in the voltage difference between the positive and negative electrodes of the battery V diff (t), the polarity drift value is calculated. The calculation expression is as follows:

[0039]

[0040] , where PDV(t) is the polarity drift value, that is, the polarity drift value at time point t. V max is the maximum voltage of the battery, V min is the minimum voltage of the battery, e -λ·t is the time decay factor, representing the degree of decline in battery performance over time. e is the natural base, and λ is the electrochemical decay constant;

[0041] The polarity drift value PDV(t) is converted into a polarity drift index to quantify the potential risk of battery performance degradation. The calculation expression is as follows:

[0042]

[0043] , where PDI is the polarity drift index, n is the total number of time points, and γ 1 is the weight coefficient of the polarity drift value PDV(t), γ 2 is the time decay term coefficient, and γ 3 is the time decay constant.

[0044] Preferably, under the detection window, the specific steps for analyzing the distribution of the internal electric field of the battery over time and generating the internal electric field change index of the battery are as follows:

[0045] Under the detection window, obtain the electric field distribution data of the battery. The electric field distribution is the change of the local electric field intensity of the battery measured by the sensor over time and spatial position. The calculation expression of the electric field intensity is as follows: E(x,t) = E 0 +ΔE(x,t). In the formula, E(x,t) is the electric field intensity inside the battery at the spatial position x and the time point t, and E 0 is the reference value of the initial electric field, and ΔE(x,t) is the dynamic change of the electric field intensity;

[0046] Perform a differential analysis on the electric field distribution data inside the battery in the spatial and time dimensions, and quantify the inhomogeneity of the electric field by calculating the electric field gradient. The calculation expression is as follows:

[0047]

[0048] . In the formula, is the electric field gradient, representing the gradient of the electric field at the position x and the time point t, is the gradient of the electric field in the x direction of the spatial position, is the gradient of the electric field at the time point t;

[0049] Identify the small changes in the electric field intensity in space and time through the electric field gradient , so as to extract the inhomogeneity characteristics of the electric field inside the battery, that is, the spatial gradient value of the electric field change. The calculation expression is as follows:

[0050]

[0051] . In the formula, ΔE grad is the inhomogeneity characteristic of the electric field;

[0052] Based on the inhomogeneity characteristic of the electric field ΔE grad , further identify the time evolution characteristics of the electric field change through dynamic mode analysis. Through wavelet transform, extract the change modes of different frequency bands from the electric field change sequence. The calculation expression is as follows:

[0053]

[0054] . In the formula, EFCI dynamic (t) is the electric field change index inside the battery, used to quantify the dynamic change degree of the electric field inside the battery, and W k (ΔE grad (t)) is the wavelet transform result, which is the wavelet transform result of the inhomogeneity characteristic of the electric field ΔE grad at the kth frequency component. k is the index of the frequency component, representing the different frequency components obtained after decomposing the signal at different scales, and N is the total number of frequencies;

[0055] The electric field non-uniformity feature ΔE grad and the electric field change index inside the battery EFCI dynamic (t) are fused to generate the electric field change index inside the battery, and the calculation expression is as follows:

[0056] IEFVI = A·ΔE grad + B·EFCI dynamic (t)

[0057] , where IEFVI is the electric field change index inside the battery, and A and B are both weight coefficients, which respectively control the contributions of the electric field non-uniformity feature ΔE grad and the electric field change index inside the battery EFCI dynamic (t) in the total electric field change index.

[0058] In the above technical solution, the technical effects and advantages provided by the present invention are as follows:

[0059] By introducing a real-time vehicle maintenance and repair monitoring and management method based on Internet of Things technology, the present invention can achieve refined monitoring of battery performance. The traditional battery management system (BMS) relies on fixed time intervals for data collection and cannot identify the tiny performance degradation changes of individual batteries in a timely manner. This solution can dynamically evaluate the performance of the battery pack and accurately identify potential degradation problems by collecting and analyzing the key features of the battery in real time (such as the polarity drift index, the electric field change index inside the battery, etc.). In particular, through the application of a machine learning model, the detection frequency can be adjusted in real time, and monitoring can be carried out in a timely manner at the stage of tiny changes in battery performance, avoiding the neglect of slight degradation by traditional methods. Therefore, when the performance of the battery pack decreases, necessary maintenance measures can be taken in a timely manner, reducing the risk of failures caused by the accumulation of degradation. For example, for a battery pack detected with potential degradation, the system will dynamically adjust the detection frequency according to the model evaluation results, enabling the battery management system to capture anomalies that may be missed at the conventional monitoring frequency, identify potential failures at an early stage and prevent them, thereby extending the service life of the battery pack.

[0060] The present invention can not only effectively identify the slight changes in battery performance, but also intelligently classify battery packs based on the evaluation results, dividing the performance changes of battery packs into normal changes and potential degradations. For battery packs with normal changes, the maintenance system continues to monitor them at a preset fixed time interval to ensure that the batteries are always in the best state; while for battery packs with potential degradations, the system will intelligently and dynamically adjust the detection frequency to ensure the timely capture of minor anomalies. The mechanism of dynamically adjusting the monitoring frequency ensures that the battery management system can flexibly adapt to the state changes of battery packs without relying on the traditional fixed frequency, which greatly improves the flexibility and real-time performance of monitoring, and further reduces the potential risks caused by the failure to timely identify the decline in battery performance. Specifically, for battery packs with potential degradations, the adjusted high-frequency monitoring can monitor the health status of the batteries in real time and adjust the maintenance strategy according to the changes, ensuring that when there are signs of degradation in the battery pack, more accurate maintenance measures are taken in a timely manner to avoid premature damage or failure of the battery pack, optimizing the overall maintenance management and improving the safety and reliability of vehicle operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings.

[0062] Figure 1 It is a method flowchart of the real-time vehicle maintenance monitoring and management method based on the Internet of Things technology of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] Now, the exemplary embodiments will be described more comprehensively with reference to the accompanying drawings. However, the exemplary embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these exemplary embodiments are provided so that this disclosure will be more thorough and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art.

[0064] The present invention provides a real-time vehicle maintenance monitoring and management method based on the Internet of Things technology as shown in Figure 1 and includes the following steps:

[0065] First, measure and record the operating parameters of the battery pack at a preset time interval;

[0066] In a Battery Management System (BMS), the preset time interval refers to the acquisition frequency set during the system design phase based on historical data and experience. This time interval is typically determined according to the performance characteristics of the battery pack and the operating environment. For example, historical data can reveal the typical charge-discharge cycles of the battery, temperature change patterns, and voltage fluctuation rules. This data helps determine a reasonable monitoring period to ensure that the normal operating state and potential abnormal changes of the battery pack can be effectively captured. Setting a reasonable time interval can avoid the computational and storage burdens caused by over-collecting data while maintaining sufficient monitoring of the battery pack status.

[0067] Organize the obtained battery pack operating parameters into a systematic analysis set. In the analysis set, identify and extract the key features that reflect the potential degradation of the performance of individual cells.

[0068] The specific steps to organize the obtained battery pack operating parameters into a systematic analysis set include: First, collect various operating parameters of the battery pack within a preset time interval, such as voltage, current, temperature, charge-discharge times, SOC (State of Charge of the battery), SOH (State of Health of the battery), etc. Then, perform data cleaning on these raw data to remove invalid values, outliers, or noise data to ensure the accuracy of the analysis. Subsequently, organize and classify the processed data according to the time series to construct a complete analysis set. This analysis set includes not only the historical data of each individual cell and the battery pack but also the current state data and provides support for subsequent feature extraction and anomaly detection. During this process, it is necessary to ensure the consistency and integrity of the data so as to comprehensively reflect the health status and performance of the battery pack.

[0069] Deeply analyze the extracted key features under the detection window and input the analyzed key features into a pre-trained machine learning model. Through the machine learning model, intelligently evaluate the input key features to identify and output the abnormal changes in the performance of the battery pack.

[0070] In the analysis set, identify and extract the key features that reflect the potential degradation of the performance of individual cells. The extracted features include the change in the electrochemical balance between the positive and negative electrodes of the battery and the distribution of the internal electric field of the battery over time. Under the detection window, analyze the change in the electrochemical balance between the positive and negative electrodes of the battery and the distribution of the internal electric field of the battery over time obtained, and generate a polarity drift index and an internal electric field change index of the battery respectively. The polarity drift index quantifies the small change in the electrochemical balance between the positive and negative electrodes of the battery; the internal electric field change index quantifies the small change in the distribution of the internal electric field of the battery over time.

[0071] The polarity drift index and the battery internal electric field change index generated after analyzing the extracted key features are input into a pre-trained machine learning model. The battery performance change coefficient is generated through the machine learning model, and the abnormal change of the battery pack performance is identified and output through the battery performance change coefficient.

[0072] The pre-trained machine learning model refers to a model that has been trained, optimized, and verified with a large amount of historical data and can make predictions or classifications based on the input feature data (such as the polarity drift index and the battery internal electric field change index). In the battery management system, the core role of such a model is to identify potential abnormal or degradation problems through intelligent analysis of the battery pack performance changes. To build such a machine learning model, it is necessary to first collect a large amount of battery operation data, including but not limited to the battery voltage, current, temperature, and the above-mentioned key features (such as the polarity drift index, the battery internal electric field change index, etc.). Based on these data, appropriate machine learning algorithms (such as support vector machines, random forests, neural networks, etc.) are used to train the data and find the potential relationship between each input feature and the battery performance change. During the training process, the machine learning model will continuously adjust its internal parameters according to the data patterns to improve the prediction accuracy of the battery performance degradation. These models can identify the weak signals of single cells in the early degradation stage, providing a basis for timely diagnosis and maintenance.

[0073] The pre-trained machine learning model generates a battery performance change coefficient through the analysis of the input features. This coefficient reflects the performance change of the battery pack within a certain time window. When the battery performance change coefficient exceeds a certain threshold, the system will identify that the battery pack may have abnormal or potential degradation problems, and then trigger an early warning mechanism or take corresponding treatment measures. The advantage of the machine learning model is that it can dynamically process and evaluate a large amount of real-time data obtained from the battery management system, and make accurate predictions about the potential changes in battery performance based on the training results of historical data. This model can not only reflect the battery health status in real time, but also identify small abnormalities that cannot be detected by conventional monitoring methods, capture the early signals of battery degradation in a timely manner, and thus avoid battery damage or system failures caused by delayed detection. By continuously accumulating new battery operation data and optimizing the training model, the machine learning model can gradually improve its accuracy and robustness in identifying battery performance changes, providing a more intelligent and reliable solution for battery management.

[0074] A slight shift in the electrochemical balance between the positive and negative electrodes of a battery indicates that the current performance of the individual battery is in a potentially changing state. During the charging and discharging process of the battery, the electrochemical reactions at the positive and negative electrodes determine performance indicators such as the voltage, capacity, and internal resistance of the battery. When the electrochemical balance between the positive and negative electrodes of the battery shifts, it means that the reaction mechanism inside the battery begins to change slightly, which usually reflects the aging or non-uniformity of the materials inside the battery. Although these changes may not immediately manifest as significant abnormalities in the battery voltage, current, or temperature, they are early signals of battery degradation. Especially after long-term use, the polarization phenomenon of the battery may gradually increase, leading to a decrease in the electrochemical activity on the electrode surface, thereby affecting the charging and discharging efficiency and capacity retention ability of the battery. Even if the performance of the overall battery pack still falls within the normal operating range, this shift in the electrochemical balance may mean that there is slight damage or degradation inside the battery, which has not reached the failure threshold, but the potential trend of performance decline has already begun to form. Therefore, timely capturing and quantifying this shift is crucial for early identification of potential battery degradation problems and helps prevent sudden drops or failures in battery performance.

[0075] Under the detection window, the specific steps for analyzing the change in the electrochemical balance between the positive and negative electrodes of the battery and generating the polarization drift index are as follows:

[0076] Under the detection window, first, based on the detection frequency preset by the battery management system, collect the operating parameters of the battery pack at different time points. After denoising and normalizing the collected parameters, extract the key parameters. The extracted key parameters include the battery terminal voltage, current, and temperature. And calibrate the extracted battery terminal voltage as V battery (t), which represents the voltage value across the battery at time point t. Calibrate the current as I(t), which represents the current at time point t, and calibrate the temperature as T(t), which represents the temperature at time point t;

[0077] The purpose of extracting key parameters is to screen out features with a high degree of correlation with important indicators such as battery health status and performance degradation from a large amount of battery pack operation data, so as to be able to more accurately evaluate the state of the battery. Since the operation data collected by the battery management system often contains redundant information and noise, directly analyzing the original data may lead to high processing complexity and inaccurate results. Therefore, after denoising and normalizing the data, extracting the key parameters that reflect battery health, performance changes, and abnormal states can help the system focus on the most distinguishable and warning features, enhancing the diagnostic accuracy and real-time response ability of the model. This can not only improve the efficiency of the battery management system but also identify potential performance degradation or failures earlier, ensuring the stability and safety of the battery pack.

[0078] An electrochemical equilibrium change model is established to describe the dynamic process of the electrochemical reaction inside the battery. Based on the equations of the physical and chemical processes inside the battery, the voltage-time relationship of the battery is derived, and the model is constructed by combining the effects of current and temperature on the reaction rate inside the battery. The specific model expression is as follows:

[0079]

[0080] , where V diff (t) is the change in the voltage difference between the positive and negative electrodes of the battery, representing the voltage difference between the positive and negative electrodes of the battery at time point t. V ref is the reference voltage. The reference voltage is the baseline voltage in the model and is usually a reference value obtained from historical data or standard battery performance data. α 1 , α 2 and α 3 are all weight coefficients. α 1 controls the influence of current change on the battery performance and reflects the degree of influence of current change on the electrochemical reaction. α 2 controls the influence of temperature change on the battery performance and reflects the degree of influence of temperature change on the reaction rate. α 3 is used to adjust the baseline reaction rate of the battery, representing the electrochemical reaction rate of the battery under normal operating conditions. β 1 is the electrochemical reaction rate adjustment coefficient, used to adjust the change rate of the voltage difference between the positive and negative electrodes of the battery, reflecting the influence of the electrochemical reaction rate inside the battery on the change of the battery terminal voltage. t 1 is the start time of the detection window, representing the total accumulation of the electrochemical equilibrium change between the positive and negative electrodes of the battery from the start time of the detection window to time point t;

[0081] The electrochemical equilibrium change model is a mathematical model used to describe and predict how the charge exchange, ion migration, and other chemical processes between the positive and negative electrodes of the battery change over time, environment, and load conditions during the electrochemical reaction process inside the battery. By quantifying the electrochemical equilibrium state inside the battery, this model reveals the non-linear and dynamic characteristics of the electrochemical reaction during the charging and discharging process of the battery. Its function is to simulate the electrochemical process of the battery through precise mathematical expressions, help evaluate the health status of the battery, identify potential degradation and damage, predict performance degradation or faults in advance, so as to provide key health assessment basis for the battery management system (BMS), support precise battery management, optimize the charging strategy, and extend the battery life.

[0082] Based on the change in the voltage difference between the positive and negative electrodes of the battery V diff (t), the polarity drift value is calculated, and the calculation expression is as follows:

[0083]

[0084] , where PDV(t) is the polarization drift value, that is, the polarization drift value at time point t, V max is the maximum voltage of the battery, V min is the minimum voltage of the battery, e -λ·t is the time decay factor, indicating the degree of battery performance degradation over time. e is the natural base, and λ is the electrochemical decay constant, a constant representing the decay rate of the electrochemical reaction inside the battery, which describes the decay rate of the battery during long-term use;

[0085] The role of the polarization drift value PDV(t) is to quantify the dynamic changes in the electrochemical balance between the positive and negative electrodes of the battery, and then detect the potential performance degradation of the battery. By analyzing the changes in the voltage difference between the positive and negative electrodes of the battery, the polarization drift value can reveal the imbalance of the electrochemical reaction inside the battery, even if these changes are relatively small and do not immediately affect the overall battery performance. The larger this value, the more significant the change in the voltage difference between the positive and negative electrodes of the battery, which may indicate a potential decline or gradual degradation of the battery performance. The polarization drift value not only helps to identify minor faults inside the battery, but also provides key data support for the battery management system, so as to take appropriate preventive measures in advance to avoid serious problems caused by battery performance degradation, such as voltage imbalance, uneven charging, etc.

[0086] The polarization drift value PDV(t) is converted into a polarization drift index to quantify the potential risk of battery performance degradation. The calculation expression is as follows:

[0087]

[0088] , where PDI is the polarization drift index, n is the total number of time points, γ 1 is the weight coefficient of the polarization drift value PDV(t), used to determine the influence weight of the polarization drift value PDV(t) in the polarization drift index, γ 2 is the time decay term coefficient, used to adjust the contribution of the battery health state changing over time to the polarization drift index PDI, γ 3 is the time decay constant, used to describe the decay rate of the influence of time factors on the polarization drift value.

[0089] Under the detection window, the larger the performance value of the polarity drift index generated after analyzing the change in the electrochemical balance between the positive and negative electrodes of the battery, the more significant the change in the electrochemical balance between the positive and negative electrodes of the current single-cell battery, indicating that there is a slight deviation in the internal electrochemical reaction of the battery, thus indicating that the battery performance is in a potential change state. This drift is usually caused by factors such as degradation of internal battery materials, changes in the electrolyte, or attenuation of the electrode plate activity. It may not yet manifest as an obvious fault of the battery, but it reflects the trend of gradual degradation of the battery. On the contrary, if the polarity drift index is low or close to zero, it indicates that there is no significant deviation in the electrochemical balance of the battery, indicating that the performance of the battery has not changed abnormally and is still in a relatively stable working state.

[0090] The uneven distribution of the internal electric field of the battery over time indicates that the performance of the current single-cell battery is in a potential change state. This is because during the use of the battery, the internal chemical reactions and physical phenomena will cause changes in the electric field distribution. For example, as the charge-discharge cycle progresses, the electrolyte, material layer, and electrode surface inside the battery will age, degrade, or react unevenly, resulting in fluctuations in the electric field strength between different regions. This kind of fluctuation usually does not immediately cause an obvious decrease in the overall performance of the battery, but it indicates that subtle changes have occurred in the internal electrochemical process of the battery, which may lead to performance degradation in local areas. The unevenness of the internal electric field of the battery is often an early signal of internal chemical instability, local overheating, or degradation of electrode materials in the battery. Although these changes have little impact on the voltage, capacity, or temperature of the overall battery, they often imply potential performance degradation of the battery. The change in this uneven electric field, especially gradually manifested during long-term use, may be a precursor to battery failure. If not discovered and corresponding measures are not taken in time, it may lead to accelerated decrease in battery capacity, increased internal resistance, and even dangerous situations such as overcharging and over-discharging. Therefore, the change in the internal electric field distribution of the battery, especially the slight unevenness, has a potential warning effect and is an important indicator of the change in the battery health state.

[0091] Under the detection window, the specific steps for generating the internal electric field change index of the battery by analyzing the change in the internal electric field distribution of the battery over time are as follows:

[0092] Under the detection window, obtain the electric field distribution data of the battery. The electric field distribution is the change in the local electric field strength of the battery measured by the sensor over time and spatial position. The calculation expression of the electric field strength is as follows: E(x,t) = E 0 +ΔE(x,t), where E(x,t) is the electric field strength inside the battery at the spatial position x and time point t, E 0 is the reference value of the initial electric field, and ΔE(x,t) is the dynamic change of the electric field strength, which is the part of the internal electric field of the battery that changes with time t and spatial position x, representing the dynamic change part of the electric field of the battery during operation;

[0093] An electric field is a physical phenomenon generated by electric charges in the surrounding space. It describes the force exerted by electric charges in the surrounding space. The strength of the electric field at a certain point is the ratio between the force generated by the charges at that point and the unit charge, usually measured in volts per meter (V / m). The direction of the electric field refers to the direction of the force that a charge experiences in the electric field. A positive charge moves along the direction of the electric field, while a negative charge moves in the opposite direction. The existence of the electric field enables electric charges to interact with each other. Therefore, it is one of the important descriptions of the behavior of electric charges and is widely used in various electrical devices such as batteries, capacitors, and power transmission.

[0094] Perform a differential analysis on the electric field distribution data inside the battery in the spatial and temporal dimensions, and quantify the non-uniformity of the electric field by calculating the electric field gradient. The calculation expression is as follows:

[0095]

[0096] , where is the electric field gradient, representing the gradient of the electric field at position x and time point t is the gradient of the electric field in the x direction of the spatial position, is the gradient of the electric field at time point t;

[0097] Electric field gradient is a quantity that describes the variation of the electric field strength with space, representing the rate of change of the electric field at a certain position. It is a measure of the change of the electric field between different positions and is usually used to reflect the non-uniform distribution of the electric field. The existence of the electric field gradient means that the electric field changes rapidly in some regions, which may be due to the non-uniformity of the local charge distribution or the changes of certain electrochemical reactions inside the battery. When the electric field gradient is large, it usually indicates that there may be phenomena such as non-uniform performance, local overheating, or unstable chemical reactions inside the battery. In the monitoring of the battery, the electric field gradient can reveal the non-uniform state inside the battery, thus providing potential fault warning signals to help evaluate the health state of the battery and identify possible degradation problems in advance.

[0098] Identify the small changes in the electric field strength in space and time through the electric field gradient and extract the characteristics of the electric field non-uniformity inside the battery, that is, the spatial gradient value of the electric field change. The calculation expression is as follows:

[0099]

[0100] , where ΔE grad is the characteristic of the electric field non-uniformity;

[0101] Based on the characteristic of the electric field non-uniformity ΔE grad, further identify the time-evolution characteristics of the electric field change through dynamic mode analysis. Through wavelet transform, extract the change patterns in different frequency bands from the electric field change sequence, and the calculation expression is as follows:

[0102]

[0103] , where EFCI dynamic (t) is the electric field change index inside the battery, which is used to quantify the dynamic change degree of the electric field inside the battery, reflects the change pattern of the electric field distribution inside the battery over time, can capture the fluctuations of the non-uniform electric field distribution inside the battery, W k (ΔE grad (t)) is the result of wavelet transform, which is the wavelet transform result of the electric field inhomogeneity feature ΔE grad on the k-th frequency component. Wavelet transform is used to decompose the electric field gradient signal into components in different frequency ranges, which helps to capture the local changes of the electric field in time and frequency. Through wavelet transform, we can extract the frequency components of the electric field change, and then evaluate the dynamic changes of the battery performance. In the battery management system, the degradation and anomalies of the battery usually manifest as high-frequency components (the electric field fluctuates violently in a short time). k is the index of the frequency component, representing the different frequency components obtained after decomposing the signal at different scales. Each k corresponds to the information of the signal in a specific frequency band, and N is the total number of frequencies;

[0104] Fuse the electric field inhomogeneity feature ΔE grad and the electric field change index EFCI dynamic (t) inside the battery to generate the electric field change index inside the battery, and the calculation expression is as follows:

[0105] IEFVI = A·ΔE grad + B·EFCI dynamic (t)

[0106] , where IEFVI is the electric field change index inside the battery, and A and B are both weight coefficients, which respectively control the contributions of the electric field inhomogeneity feature ΔE grad and the electric field change index EFCI dynamic (t) to the total electric field change index.

[0107] Under the detection window, the larger the performance value of the internal electric field change index of the battery generated after analyzing the distribution of the internal electric field of the battery over time, the more uneven the distribution of the internal electric field of the battery over time, which in turn implies that the performance of the current single battery may be in a potentially changing state. The change in the internal electric field of the battery reflects the changes in the internal chemical reactions, physical phenomena, and material aging processes of the battery. When the battery undergoes uneven electric field changes during long-term operation, it is usually a precursor to local degradation, performance decline, or material degradation inside the battery. The internal electric field change index quantifies this non-uniformity, enabling the battery management system to promptly identify potential performance degradation problems. When the index value is high, it indicates that there is an uneven electric field distribution inside the battery, which may mean that the battery is about to enter the degradation period, and it is necessary to strengthen monitoring or adjust the operation strategy. Conversely, if the internal electric field change index of the battery is low, it indicates that the electric field distribution of the battery within the monitoring window is relatively stable and the performance has not changed.

[0108] The machine learning model is not limited here. Any machine learning model that can comprehensively analyze the polarity drift index PDI and the internal electric field change index IEFVI of the battery to generate the battery performance change coefficient BPCC can be used. To implement the technical solution of the present invention, the present invention provides a specific implementation method;

[0109] The formula for generating the battery performance change coefficient BPCC is as follows:

[0110]

[0111] , where m 1 , m 2 are respectively the preset proportionality coefficients of the polarity drift index PDI and the internal electric field change index IEFVI of the battery, and m 1 , m 2 are both greater than 0.

[0112] It can be seen from the calculation expression of the battery performance change coefficient that under the detection window, the larger the performance value of the polarity drift index generated after analyzing the electrochemical balance change between the positive and negative electrodes of the battery, and the larger the performance value of the internal electric field change index of the battery generated after analyzing the distribution of the internal electric field of the battery over time, the larger the performance value of the battery performance change coefficient generated after analyzing the extracted features under the detection window, indicating that the current battery pack has a greater probability of potential abnormal changes. Conversely, it indicates that the performance of the current battery pack has not changed.

[0113] The preset proportionality coefficient here refers to the parameter in the formula used to weigh and adjust the contribution degrees of the polarity drift index PDI and the internal electric field change index IEFVI to the battery performance change coefficient BPCC, that is, m 1 and m 2. Their values are preset according to different application scenarios, experimental data, or battery characteristics, reflecting the proportion of the importance of these two indicators in evaluating battery performance changes. Since the effects of polarity drift and electric field changes on battery performance may be different, the setting of the proportionality coefficient can make the calculation results more accurate, and the weight setting needs to satisfy m 1 、m 2 greater than 0 to ensure that the two indicators have a positive contribution to the battery performance change coefficient BPCC. The selection of such a proportionality coefficient is usually based on actual test data and adjusted through optimization algorithms or expert experience to achieve the best evaluation effect.

[0114] Based on the evaluation results of the machine learning model, classify the performance changes of the battery pack into normal changes and potential degradation;

[0115] Compare and analyze the battery performance change coefficient generated after analyzing the extracted features with the preset battery performance change coefficient to divide the performance changes of the battery pack. The division steps are as follows:

[0116] If the battery performance change coefficient is greater than or equal to the preset reference threshold of the battery performance change coefficient, classify the performance change of the battery pack as potential degradation;

[0117] If the battery performance change coefficient is less than the preset reference threshold of the battery performance change coefficient, classify the performance change of the battery pack as normal change;

[0118] Normal change means that under the current monitoring parameters, the battery pack operates as expected without obvious abnormal signs; potential degradation means that there is a small but gradually accumulating performance decline in the battery pack, which may cause more serious problems in the future.

[0119] For the battery pack with normal changes, continue to monitor and collect data at the preset fixed time interval to ensure that the battery pack is always in good operating condition;

[0120] Maintaining the monitoring method at a fixed time interval is of great significance for the battery pack with normal changes. Its main purpose is to ensure that when the battery pack is in a stable state, the system resources are optimized. Monitoring at a fixed frequency can avoid unnecessary calculation and storage overhead caused by frequent adjustment of the monitoring frequency, thereby improving system efficiency. At the same time, monitoring at a fixed frequency can capture any possible small changes when the battery pack is in a stable state, identify potential problems in a timely manner, and prevent them from developing into more serious failures. For the battery pack with normal changes, continuing to use the original monitoring frequency can not only balance resource usage and risk control, but also ensure that the battery pack is always in good operating condition, thus providing an effective maintenance plan for the battery management system (BMS) and improving battery life and system safety.

[0121] For a potentially degraded battery pack, based on the evaluation results of the machine learning model, the detection frequency is dynamically adjusted through a fuzzy logic control algorithm to promptly capture abnormal changes and prevent minor anomalies from accumulating into serious faults.

[0122] The specific steps for a potentially degraded battery pack to dynamically adjust the detection frequency based on the evaluation results of the machine learning model through a fuzzy logic control algorithm to promptly capture abnormal changes and prevent minor anomalies from accumulating into serious faults are as follows:

[0123] Establish a fuzzy rule set based on inputs and outputs. Input the battery performance change coefficient BPCC and the reference threshold of the battery performance change coefficient BPCC ref , and output the adjusted detection frequency. The calculation expression is as follows:

[0124]

[0125] f new = f base + Δf

[0126] , where f base is the preset detection frequency, r is the adjustment ratio factor, which refers to the coefficient that quantifies the adjustment amplitude of the detection frequency according to the battery performance change. The adjustment ratio factor determines the amplitude and direction of the frequency adjustment based on the change in the battery health state. ω is the non-linear adjustment parameter, is the sensitivity parameter, Δf is the adjustment amount of the detection frequency, and f new is the new adjusted detection frequency;

[0127] The non-linear adjustment parameter refers to the parameter that adjusts the system response through a non-linear function during the detection frequency adjustment process. Usually, these parameters reflect the non-linear relationship between the battery performance change and the detection frequency. As the battery degradation risk increases, the system's adjustment of the frequency is not just a simple linear change but through a non-linear function, making the frequency adjustment response more drastic when the battery performance deteriorates. The role of non-linear adjustment is to enable the system to respond more sensitively and promptly when the battery shows significant performance changes and prevent the performance degradation problem from being ignored.

[0128] The sensitivity parameter represents the response sensitivity of the system to the battery performance change. In the battery performance change detection, the sensitivity parameter determines the degree to which the system triggers the frequency adjustment. A higher sensitivity parameter means that when there is a slight anomaly in the battery performance change, the system will promptly adjust the detection frequency; a lower sensitivity parameter means that the system will only respond when there is a large change in the battery performance. The role of the sensitivity parameter is to optimize the system's response speed and sensitivity to effectively monitor at the early stage of minor changes or potential risks while avoiding overreaction.

[0129] Convert the input parameters of the battery performance change coefficient BPCC and the reference threshold of the battery performance change coefficient BPCC ref into fuzzy sets, represented by two fuzzy levels of "Low" and "High", and define the membership function μ(BPCC) to represent the membership degree of each level;

[0130]

[0131] , where μ Low (BPCC) is the membership degree of low risk, x max is the maximum threshold of the battery performance change coefficient, and μ High (BPCC) is the membership degree of high risk;

[0132] When the battery performance change coefficient BPCC is less than the reference threshold of the battery performance change coefficient BPCC ref , it indicates that the battery performance is normal, the membership degree is 1, indicating the lowest risk; when the battery performance change coefficient BPCC is equal to or greater than the reference threshold of the battery performance change coefficient BPCC ref , the risk gradually increases, and the membership degree decreases linearly with the increase of the battery performance change coefficient BPCC. When the battery performance change coefficient BPCC exceeds the maximum threshold of the battery performance change coefficient x max , the membership degree is 0, indicating that the degradation risk has reached the highest.

[0133] When the battery performance change coefficient BPCC is less than the reference threshold of the battery performance change coefficient BPCC ref , it indicates that the battery performance is normal, the membership degree is 0; when the battery performance change coefficient BPCC is equal to or exceeds the reference threshold of the battery performance change coefficient BPCC ref , it indicates that the degradation risk of the battery gradually increases, the membership degree increases linearly, and as the battery performance change coefficient BPCC continues to increase, it finally reaches 1. When the battery performance change coefficient BPCC reaches the maximum threshold of the battery performance change coefficient x max , the degradation risk is the highest.

[0134] Through fuzzyfication, the battery performance change coefficient BPCC is divided into different risk levels according to the membership degree. Combining with the actual battery health status, the priority of the detection frequency can be dynamically adjusted.

[0135] Based on fuzzy rules, reasoning is carried out by combining fuzzy sets and fuzzy rule bases to adjust the detection frequency adjustment amount, and the calculation expression is as follows:

[0136] Where ΔF is the adjusted detection frequency adjustment amount, and H is the amplification factor, which is used to further amplify the influence in the battery degradation state;

[0137] Through defuzzification, the result of fuzzy inference is converted into a specific numerical output, that is, the final detection frequency, and the expression is as follows:

[0138]

[0139] , where F new is the final detection frequency, ΔF Low is the detection frequency adjustment amount at low risk, and ΔF High is the detection frequency adjustment amount at high risk.

[0140] Defuzzification outputs a final detection frequency F by comprehensively considering the influence of low risk and high risk on the detection frequency. new . Finally, the new frequency is used to update the battery monitoring system to ensure higher sensitivity monitoring of potential degradation.

[0141] The function of this step is to dynamically adjust the monitoring frequency of the battery pack based on the evaluation results of the machine learning model, so as to achieve timely discovery and intervention of potential degradation problems, and avoid the gradual accumulation of small abnormal changes and the evolution into serious failures. Traditional fixed detection frequencies may not be able to capture those early and progressive battery performance degradations, especially when the conventional parameters such as the overall voltage, current, and temperature of the battery pack are normal. As the individual battery cells gradually degrade, the subtle changes in their internal electrochemical balance and electric field distribution are often difficult to detect in a timely manner in traditional monitoring. If these small abnormalities are not controlled, they may ultimately lead to battery pack imbalance, reduced charging efficiency, and even catastrophic failures such as thermal runaway.

[0142] By using a machine learning model to intelligently evaluate the battery health status, potential degradations of battery performance can be identified more precisely, and dynamically adjusting the detection frequency is an effective means to address this challenge. When the model identifies a potential degradation risk in a single battery cell of the battery pack, the detection frequency can be adjusted according to the degree and trend of degradation to increase the acquisition density of key features, thereby capturing abnormal changes in advance. For example, when the machine learning model detects slight fluctuations in the electrochemical balance or internal electric field distribution of some single battery cells, the detection frequency will be increased accordingly to more precisely track these changes.

[0143] This dynamically adjusted strategy can not only improve the accuracy of monitoring but also avoid the waste of resources caused by overly frequent monitoring. When formulating a maintenance plan, based on the machine learning evaluation results and the dynamically adjusted detection strategy, a more accurate maintenance cycle can be set. When potential degradation problems are detected, the system can take preventive measures in advance, such as optimizing the charging strategy and adjusting the operating load, thus effectively extending the service life of the battery pack and avoiding failures. Ultimately, this monitoring method based on intelligent evaluation and dynamic adjustment not only improves the reliability of the battery management system but also reduces the overall operating cost, ensuring the running safety and efficiency of electric vehicles.

[0144] Through the introduction of a real-time vehicle maintenance monitoring and management method based on Internet of Things technology, the present invention can achieve refined monitoring of battery performance. Traditional battery management systems (BMS) rely on fixed time intervals for data collection and cannot identify the tiny performance degradation changes of individual batteries in a timely manner. This solution can dynamically evaluate the performance of the battery pack and accurately identify potential degradation problems by collecting and analyzing the key features of the battery in real time (such as the polarity drift index, the internal electric field change index of the battery, etc.). Especially through the application of machine learning models, the detection frequency can be adjusted in real time, and monitoring can be carried out in a timely manner at the stage of tiny changes in battery performance, avoiding the neglect of minor degradation by traditional methods. Therefore, when the performance of the battery pack decreases, necessary maintenance measures can be taken in a timely manner, reducing the risk of failures caused by the accumulation of degradation. For example, for a battery pack detected with potential degradation, the system will dynamically adjust the detection frequency according to the model evaluation results, enabling the battery management system to capture anomalies that might be missed under the conventional monitoring frequency, identify potential failures at an early stage and prevent them, thus extending the service life of the battery pack.

[0145] The present invention can not only effectively identify the tiny changes in battery performance but also classify the battery packs intelligently based on the evaluation results, dividing the performance changes of the battery packs into normal changes and potential degradation. For the battery packs with normal changes, the maintenance system continues to monitor them at the preset fixed time intervals to ensure that the batteries are always in the best state; while for the battery packs with potential degradation, the system will intelligently adjust the detection frequency dynamically to ensure the timely capture of tiny anomalies. The mechanism of dynamically adjusting the monitoring frequency ensures that the battery management system can flexibly adapt to the state changes of the battery pack without relying on the traditional fixed frequency, which greatly improves the flexibility and real-time nature of monitoring, and further reduces the potential risks caused by the failure to identify the decrease in battery performance in a timely manner. Specifically, for the battery packs with potential degradation, the adjusted high-frequency monitoring can monitor the health status of the battery in real time and adjust the maintenance strategy according to the changes, ensuring that when there are signs of degradation in the battery pack, more accurate maintenance measures can be taken in a timely manner to avoid premature damage or failures of the battery pack, optimizing the overall maintenance management and improving the running safety and reliability of the vehicle.

[0146] The above only describes certain exemplary embodiments of the present invention by way of illustration. Without doubt, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A real-time vehicle maintenance monitoring and management method based on Internet of Things technology, characterized in that: The following steps are involved: First, the operating parameters of the battery pack are measured and recorded at pre-set time intervals; Organize the acquired battery pack operating parameters into a systematic analysis set, in which key features reflecting the potential degradation of single cell performance are identified and extracted; The extracted key features are deeply analyzed under the detection window, and the analyzed key features are input into the pre-trained machine learning model. The machine learning model is used to intelligently evaluate the input key features and identify abnormal changes in the output battery pack performance; Based on the evaluation results of the machine learning model, the performance changes of the battery pack are classified into normal changes and potential degradation; For battery packs with normal changes, continue to monitor and collect data at preset fixed time intervals to ensure that the battery pack is always in good operating condition; For potentially degraded battery packs, the detection frequency is dynamically adjusted based on the evaluation results of the machine learning model through a fuzzy logic control algorithm to capture abnormal changes in a timely manner and prevent minor anomalies from accumulating into serious faults.

2. The real-time vehicle maintenance monitoring and management method based on Internet of Things technology according to claim 1 is characterized by: In the analysis set, key features reflecting the potential degradation of single cell performance are identified and extracted. The extracted features include changes in the electrochemical balance between the positive and negative electrodes of the battery and the distribution of the electric field inside the battery over time. Under the detection window, the acquired changes in the electrochemical balance between the positive and negative electrodes of the battery and the distribution of the electric field inside the battery over time are analyzed to generate a polarity drift index and a battery internal electric field change index, respectively. The polarity drift index quantifies the slight changes in the electrochemical balance between the positive and negative electrodes of the battery; the battery internal electric field change index quantifies the slight changes in the distribution of the battery's internal electric field over time.

3. The real-time vehicle maintenance monitoring and management method based on Internet of Things technology according to claim 2 is characterized by: The polarity drift index and battery internal electric field change index generated after analyzing the extracted key features are input into the pre-learned machine learning model, and the battery performance variation coefficient is generated by the machine learning model. The battery performance variation coefficient is used to identify abnormal changes in the output battery pack performance.

4. The real-time vehicle maintenance monitoring and management method based on Internet of Things technology according to claim 3 is characterized by: The battery performance variation coefficient generated after analyzing the extracted features is compared with the preset battery performance variation coefficient to divide the performance variation of the battery pack. The division steps are as follows: If the battery performance variation coefficient is greater than or equal to a preset battery performance variation coefficient reference threshold, the battery pack performance variation is classified as potential degradation; If the battery performance variation coefficient is less than a preset battery performance variation coefficient reference threshold, the battery pack performance variation is classified as a normal variation.

5. The real-time vehicle maintenance monitoring and management method based on Internet of Things technology according to claim 4 is characterized by: For potentially degraded battery packs, the specific steps to dynamically adjust the detection frequency based on the evaluation results of the machine learning model and the fuzzy logic control algorithm to capture abnormal changes in a timely manner and avoid the accumulation of minor abnormalities into serious faults are as follows: Establish a fuzzy rule set based on input and output, input the battery performance change coefficient BPCC, the battery performance change coefficient reference threshold BPCC ref , output the adjusted detection frequency, the calculation expression is as follows: f new =f base +Δf In the formula, f base is the preset detection frequency, r is the adjustment scale factor, ω is the nonlinear adjustment parameter, is the sensitivity parameter, Δf is the adjustment amount of the detection frequency, and f new is the new adjusted detection frequency; The input parameters of battery performance variation coefficient BPCC and battery performance variation coefficient reference threshold BPCC ref It is converted into a fuzzy set, expressed in two fuzzy levels: "low" and "high", and the membership function μ(BPCC) is defined to represent the degree of membership of each level; In the formula, μ Low (BPCC) is the low risk membership, x max is the maximum threshold of battery performance variation coefficient, μ High (BPCC) is a high risk affiliation; Based on fuzzy rules, fuzzy sets and fuzzy rule bases are combined for reasoning to adjust the detection frequency adjustment amount. The calculation expression is as follows: Where ΔF is the adjusted detection frequency adjustment amount, and H is the amplification factor, which is used to further amplify the impact of battery degradation; Through defuzzification, the result of fuzzy reasoning is converted into a specific numerical output, that is, the final detection frequency, which is expressed as follows: In the formula, F new is the final detection frequency, ΔF Low is the detection frequency adjustment for low risk, ΔF High is the amount to adjust the detection frequency when the risk is high.

6. The real-time vehicle maintenance monitoring and management method based on Internet of Things technology according to claim 2 is characterized in that: Under the detection window, the specific steps for analyzing the electrochemical balance changes between the positive and negative electrodes of the battery and generating the polarity drift index are as follows: In the detection window, first, based on the detection frequency preset by the battery management system, the operating parameters of the battery pack at different time points are collected, and the key parameters are extracted after denoising and normalization. The extracted key parameters include battery terminal voltage, current and temperature, and the extracted battery terminal voltage is calibrated as V battery (t), represents the voltage value across the battery at time point t, the current is calibrated as I(t), represents the current at time point t, and the temperature is calibrated as T(t), represents the temperature at time point t; An electrochemical equilibrium change model is established to describe the dynamic process of electrochemical reactions in the battery. The voltage-time relationship of the battery is derived through equations based on the physical and chemical processes in the battery, and the model is constructed by combining the effects of current and temperature on the reaction rate in the battery. The specific model expression is as follows: Where V diff (t) is the change in the voltage difference between the positive and negative electrodes of the battery, which indicates the change in the voltage difference between the positive and negative electrodes of the battery at time point t, V ref is the reference voltage, α1, α2 and α3 are weight coefficients, β1 is the electrochemical reaction rate adjustment coefficient, and t1 is the detection window start time; The voltage difference between the positive and negative electrodes of the battery changes V diff (t), calculate the polarity drift value, the calculation expression is as follows: Where PDV(t) is the polarity drift value, that is, the polarity drift value at time point t, V max is the maximum battery voltage, V min is the minimum battery voltage, e -λ·t is the time decay factor, which indicates the degree of battery performance decay over time, e is the natural base, and λ is the electrochemical decay constant; The polarity drift value PDV(t) is converted into a polarity drift index to quantify the potential risk of battery performance degradation. The calculation expression is as follows: Where PDI is the polarity drift index, n is the total number of time points, γ1 is the polarity drift value PDV(t) weight coefficient, γ2 is the time decay term coefficient, and γ3 is the time decay constant.

7. The real-time vehicle maintenance monitoring and management method based on Internet of Things technology according to claim 2 is characterized in that: Under the detection window, the distribution of the internal electric field of the battery is analyzed over time, and the specific steps for generating the internal electric field change index of the battery are as follows: Under the detection window, obtain the electric field distribution data of the battery. The electric field distribution is the change of the local electric field strength of the battery measured by the sensor with time and spatial position. The electric field strength calculation expression is as follows: E(x, t) = E0 + ΔE(x, t), where E(x, t) is the electric field strength inside the battery at spatial position x and time point t, E0 is the reference value of the initial electric field, and ΔE(x, t) is the dynamic change of the electric field strength; The difference analysis of the electric field distribution data inside the battery in the spatial and temporal dimensions is performed, and the inhomogeneity of the electric field is quantified by calculating the electric field gradient. The calculation expression is as follows: In the formula, is the electric field gradient, which represents the gradient of the electric field at position x and time point t, is the gradient of the electric field in the x direction of the spatial position, is the gradient of the electric field at time t; Through the electric field gradient Identify the slight changes in the electric field strength in space and time, so as to extract the electric field inhomogeneity characteristics inside the battery, that is, the spatial gradient value of the electric field change. The calculation expression is as follows: In the formula, ΔE grad It is the characteristic of electric field inhomogeneity; Based on the electric field inhomogeneity characteristic ΔE grad , and further identify the time evolution characteristics of the electric field change through dynamic pattern analysis. Through wavelet transform, the change patterns of different frequency bands are extracted from the electric field change sequence. The calculation expression is as follows: In the formula, EFCI dynamic (t) is the battery internal electric field variation index, which is used to quantify the dynamic variation of the battery internal electric field. k (ΔE grad (t)) is the result of wavelet transform and is the electric field inhomogeneity characteristic ΔE grad The wavelet transform result on the kth frequency component, k is the index of the frequency component, indicating the different frequency components obtained after decomposing the signal at different scales, and N is the total number of frequencies; The electric field inhomogeneity characteristic ΔE grad and the battery internal electric field variation index EFCI dynamic (t) is fused to generate the internal electric field change index of the battery, which is calculated and expressed as follows: IEFVI=A·ΔE grad +B·EFCI dynamic (t) Where IEFVI is the internal electric field variation index of the battery, A and B are weight coefficients, which control the electric field inhomogeneity characteristics ΔE grad and the battery internal electric field variation index EFCI dynamic (t) Contribution to the total electric field variation index.

Images