Storage battery health state change trend monitoring system and method

Abstract

The invention relates to the technical field of electrical measurement, in particular to a storage battery health state change trend monitoring system and method. The method comprises the following steps: carrying out a charge-discharge cycle test on a storage battery to obtain terminal voltage dynamic fluctuation data; carrying out alternating current impedance spectroscopy test on the storage battery to obtain real-time change data of internal resistance; monitoring charge and discharge capacity attenuation rate data of the storage battery; performing multi-dimensional analysis on the terminal voltage dynamic fluctuation data, the internal resistance real-time change data and the charge and discharge capacity attenuation rate data to generate storage battery characteristic data; and carrying out collaborative analysis on the characteristic data of the storage battery, and generating a collaborative analysis result by analyzing the synchronism of voltage fluctuation and internal resistance change and the influence of capacity attenuation on the voltage and internal resistance change. According to the invention, accurate monitoring and prediction of the health state of the storage battery are realized through the charge-discharge cycle test, the AC impedance spectroscopy test and the capacity attenuation monitoring technology, so that the reliability and safety of the storage battery are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electrical measurement, in particular to a battery health state change trend monitoring system and method. BACKGROUND

[0002] Traditional methods often only focus on a certain indicator of the battery, such as simply measuring the terminal voltage or internal resistance; however, the health state of the battery is determined by multiple factors, and it is difficult to fully reflect its actual performance by relying on a single indicator. For example, through terminal voltage measurement alone, it is not possible to accurately determine the effect of internal resistance changes on battery performance; through internal resistance measurement alone, it is not possible to understand the voltage fluctuation situation, and this single indicator monitoring method is prone to misjudgment and cannot accurately identify the true health state of the battery; in addition, traditional methods often only discover problems after the battery has a significant failure, lacking real-time monitoring and early warning functions for the health state of the battery. In the early stage of abnormal state of the battery, abnormal characteristics cannot be identified and recorded in time, leading to further expansion of the battery failure and affecting the stable operation of the entire battery system. SUMMARY

[0003] Therefore, it is necessary to provide a battery health state change trend monitoring system and method to solve at least one of the above technical problems.

[0004] To achieve the above-mentioned purpose, a battery health state change trend monitoring method, the method comprising the following steps:

[0005] Step S1: performing a charge-discharge cycle test on the battery to obtain terminal voltage dynamic fluctuation data; performing an alternating current impedance spectrum test on the battery to obtain internal resistance real-time change data; monitoring the charge-discharge capacity attenuation rate data of the battery;

[0006] Step S2: performing multi-dimensional analysis on the terminal voltage dynamic fluctuation data, the internal resistance real-time change data, and the charge-discharge capacity attenuation rate data to generate battery characteristic data; performing collaborative analysis on the battery characteristic data, generating collaborative analysis results by analyzing the synchronicity of voltage fluctuation and internal resistance change, and the influence of capacity attenuation on voltage and internal resistance change; mapping the collaborative analysis results to a preset battery health state model to form health state mapping data;

[0007] Step S3: identifying the health state characteristics of the battery based on the health state mapping data; if an abnormal pattern exists in the health state characteristic mapping data, marking the abnormal characteristics of the battery and recording them as battery abnormal information;

[0008] Step S4: predicting long-term variation trend of the battery health state based on the health state feature mapping data, obtaining a health prediction result; if the health prediction result shows that the battery health state will significantly decrease in a future period of time, generating a health state warning signal.

[0009] Preferably, step S1 comprises the following steps:

[0010] Step S11: in the charge-discharge cycle test, charging the battery at a preset constant current until the battery reaches a rated voltage;

[0011] Step S12: discharging at the same preset constant current until the battery voltage drops to a preset minimum voltage threshold; repeating the above charge-discharge process at least three times, recording the terminal voltage and time data in each charge-discharge process, and calculating the terminal voltage dynamic fluctuation data;

[0012] Step S13: in the alternating current impedance spectrum test, applying an alternating current signal with a frequency range of 0.01 Hz to 100 kHz, and measuring the impedance value of the battery at different frequencies;

[0013] Step S14: recording the curve of the impedance value changing with the frequency, and extracting the real-time resistance change data;

[0014] Step S15: in the charge-discharge capacity attenuation rate monitoring, recording the capacity value of each charge-discharge cycle, and calculating the capacity difference value of adjacent two charge-discharge cycles;

[0015] Step S16: determining the charge-discharge capacity attenuation rate data according to the ratio of the capacity difference value to the initial capacity.

[0016] Preferably, the multi-dimensional analysis of the terminal voltage dynamic fluctuation data, the real-time resistance change data and the charge-discharge capacity attenuation rate data in step S2 comprises:

[0017] segmenting the terminal voltage dynamic fluctuation data according to time sequence, calculating the average value and standard deviation of each segment of data, and generating a voltage fluctuation feature vector;

[0018] grouping the real-time resistance change data according to the frequency range, calculating the resistance average value and resistance change rate of each group of data, and generating a resistance feature vector;

[0019] grouping the charge-discharge capacity attenuation rate data according to the number of charge-discharge cycles, calculating the capacity attenuation percentage and capacity attenuation slope of each group of data, and generating a capacity attenuation feature vector;

[0020] normalizing and time-aligning the voltage fluctuation feature vector, the resistance feature vector and the capacity attenuation feature vector to generate the battery feature data.

[0021] Preferably, the step S2 of performing collaborative analysis on the battery characteristic data includes:

[0022] performing time series analysis on the voltage fluctuation feature vector to extract periodic and trend components of the voltage fluctuation, and generating voltage fluctuation periodic feature and trend feature;

[0023] performing frequency domain analysis on the internal resistance feature vector to extract high frequency and low frequency components of the internal resistance change, and generating internal resistance high frequency feature and low frequency feature;

[0024] performing nonlinear fitting on the capacity attenuation feature vector to extract acceleration and deceleration stage features of the capacity attenuation, and generating capacity attenuation acceleration feature and deceleration feature;

[0025] performing cross-correlation analysis on the voltage fluctuation periodic feature and the internal resistance high frequency feature to evaluate the synchronism of the periodic change of the voltage fluctuation and the high frequency change of the internal resistance.

[0026] Especially important is that the cross-correlation analysis on the voltage fluctuation periodic feature and the internal resistance high frequency feature to evaluate the synchronism of the periodic change of the voltage fluctuation and the high frequency change of the internal resistance includes:

[0027] aligning the voltage fluctuation periodic feature and the internal resistance high frequency feature;

[0028] determining the cross-correlation coefficient between the voltage fluctuation periodic feature and the internal resistance high frequency feature, and calculating the correlation of the voltage fluctuation and the high frequency change of the internal resistance in each period;

[0029] generating synchronism evaluation data according to the cross-correlation coefficient, wherein the synchronism evaluation data includes specific numerical values and charts of the synchronism evaluation.

[0030] Preferably, the step S2 of performing collaborative analysis on the battery characteristic data includes:

[0031] performing difference analysis on the capacity attenuation acceleration feature and the voltage fluctuation trend feature by calculating the difference values of the two at the same time point, generating a difference feature vector, and evaluating the influence of the capacity attenuation acceleration stage on the voltage fluctuation trend;

[0032] performing ratio analysis on the capacity attenuation deceleration feature and the internal resistance low frequency feature by calculating the ratio values of the two at the same time point, generating a ratio feature vector, and evaluating the influence of the capacity attenuation deceleration stage on the internal resistance low frequency change.

[0033] Preferably, the step S2 of mapping the collaborative analysis results to the preset battery health state model includes:

[0034] characterize the synchronism data of the voltage fluctuation and the internal resistance change in the synergistic analysis result;

[0035] characterize the influence data of the capacity attenuation on the voltage and the internal resistance change;

[0036] splice the characterized synchronism data and the capacity attenuation influence data to form a binary feature vector;

[0037] input the binary feature vector into a preset battery health state model to output health state mapping data.

[0038] Preferably, the identifying the health state feature of the battery based on the health state mapping data in step S3 comprises:

[0039] extracting the voltage fluctuation amplitude, the internal resistance change rate and the capacity attenuation percentage from the health state mapping data and recording as battery feature parameters;

[0040] comparing the battery feature parameters with preset health state feature indicators to obtain a health state comparison result;

[0041] determining the current health state feature of the battery according to the health state comparison result, including the health state level, the voltage stability, the internal resistance uniformity and the capacity retention rate, to judge whether the battery is in a normal health state.

[0042] Preferably, in step S3, if an abnormal pattern exists in the health state feature mapping data, the battery abnormal feature is marked and recorded as battery abnormal information, which comprises:

[0043] extracting the voltage fluctuation amplitude, the internal resistance change rate and the capacity attenuation percentage from the health state mapping data to form a detection feature value set;

[0044] performing threshold value judgment on each feature value of the detection feature value set to determine whether it exceeds a preset normal range;

[0045] if any one of the voltage fluctuation amplitude, the internal resistance change rate or the capacity attenuation percentage exceeds the preset normal range, it is marked as a potential abnormal feature;

[0046] performing neighborhood analysis on the potential abnormal feature to check whether the feature values of the data points around it are also abnormal to confirm the existence of the abnormal pattern;

[0047] if the existence of the abnormal pattern is confirmed, the feature value corresponding to the abnormal pattern is marked as a battery abnormal feature and recorded as battery abnormal information.

[0048] Preferably, the step S4 of predicting long-term variation trend of the battery health state based on the health state characteristic mapping data comprises:

[0049] performing time series analysis on the health state characteristic mapping data, calculating variation trends at multiple time points to obtain time series analysis results;

[0050] determining long-term variation trend of each characteristic value according to the time series analysis results, calculating characteristic variation rate of the health state characteristic mapping data value at the past 10 time points;

[0051] if the characteristic value variation rate is continuously positive, marking as health state rising trend;

[0052] if the characteristic value variation rate is continuously negative, marking as health state declining trend;

[0053] if the characteristic value variation rate fluctuates within a preset threshold range, marking as health state stable trend;

[0054] comprehensively analyzing the long-term variation trend of each characteristic value to predict long-term variation trend of the overall battery health state.

[0055] Especially important is that, in the step S4, if the health prediction result shows that the battery health state significantly declines in a future period of time, comprising:

[0056] extracting variation trend of the health state grade from the health prediction result;

[0057] setting a health state decline threshold, wherein the health state decline threshold represents a critical variation rate of the health state grade decline;

[0058] comparing the predicted variation rate of the health state grade with the health state decline threshold to determine whether there is a significant declining trend;

[0059] if the predicted variation rate of the health state grade is lower than the health state decline threshold, determining that the battery health state significantly declines in the future, and generating a health state warning signal;

[0060] setting the warning level according to the health state warning signal into a first-level warning, a second-level warning and a third-level warning.

[0061] The beneficial effects of the present application are: through the acquisition of terminal voltage dynamic fluctuation data by charge-discharge cycle test, the acquisition of internal resistance real-time change data by AC impedance spectrum test, and the monitoring of charge-discharge capacity attenuation rate data, the key performance indicators of the storage battery in the actual operation process can be comprehensively and accurately reflected. The acquisition of these data covers the performance of the storage battery under different working conditions, providing a solid foundation for subsequent analysis and ensuring the reliability of the monitoring results. The terminal voltage dynamic fluctuation data, internal resistance real-time change data and charge-discharge capacity attenuation rate data are subjected to multi-dimensional analysis to generate storage battery characteristic data, and further through collaborative analysis, the synchronicity of voltage fluctuation and internal resistance change, and the influence of capacity attenuation on voltage and internal resistance change are investigated. This multi-dimensional collaborative analysis method can comprehensively consider the interaction between various factors inside the storage battery, avoid misjudgment that may be caused by single index analysis, and more accurately identify the health state characteristics of the storage battery, thereby improving the accuracy and reliability of the diagnosis. Based on the health state mapping data, the current health state characteristics of the storage battery can be quickly identified, and when abnormal patterns are detected in the health state characteristic mapping data, the abnormal characteristics of the storage battery are timely marked and recorded as abnormal information of the storage battery. This function enables the early stage of abnormal state of the storage battery to be accurately identified and recorded, providing timely basis for subsequent maintenance and repair, helping to avoid further expansion of storage battery failure, and ensuring stable operation of the storage battery system. Based on the health state characteristic mapping data, the long-term change trend of the health state of the storage battery is predicted to obtain a health prediction result. When the health prediction result shows that the health state of the storage battery will significantly decrease in the future, a health state warning signal can be generated in time. This prediction function enables the user to understand the change trend of the health state of the storage battery in advance, and plan the maintenance, replacement and other work of the storage battery in advance, thereby avoiding system downtime or safety accidents caused by sudden failure of the storage battery, improving the reliability and safety of the storage battery system, and reducing the operation and maintenance cost.

[0062] The present specification also provides a storage battery health state change trend monitoring system for performing the storage battery health state change trend monitoring method as described above, which comprises:

[0063] a storage battery data acquisition module for performing charge-discharge cycle test on the storage battery to acquire terminal voltage dynamic fluctuation data, performing AC impedance spectrum test on the storage battery to acquire internal resistance real-time change data, and monitoring the charge-discharge capacity attenuation rate data of the storage battery;

[0064] a health state analysis module configured to perform multi-dimensional analysis on the terminal voltage dynamic fluctuation data, the internal resistance real-time change data, and the charge-discharge capacity attenuation rate data to generate battery characteristic data, perform collaborative analysis on the battery characteristic data, generate collaborative analysis results by analyzing the synchronism of voltage fluctuation and internal resistance change, and the influence of capacity attenuation on voltage and internal resistance change, and map the collaborative analysis results to a preset battery health state model to form health state mapping data;

[0065] a battery abnormality feature detection module configured to identify the health state features of the battery based on the health state mapping data, mark the battery abnormality features, and record the battery abnormality information if an abnormal pattern is detected in the health state feature mapping data;

[0066] a health state prediction module configured to predict the long-term change trend of the battery health state based on the health state feature mapping data to obtain a health prediction result, and generate a health state warning signal if the health prediction result shows that the battery health state will significantly decrease in the future.

[0067] The present application can accurately obtain the terminal voltage dynamic fluctuation data, the internal resistance real-time change data, and the charge-discharge capacity attenuation rate data through the data acquisition module, generate the health state mapping data through the multi-dimensional and collaborative analysis of the health state analysis module, identify and record the abnormal features through the battery abnormality feature detection module, and predict the long-term change trend and generate the warning signal through the health state prediction module, thereby realizing the comprehensive monitoring, accurate diagnosis, real-time warning, and trend prediction of the battery health state, effectively ensuring the stable operation of the battery system, reducing the operation and maintenance cost, and improving the system reliability and safety. BRIEF DESCRIPTION OF DRAWINGS

[0068] Fig. 1 FIG. 1 is a schematic diagram of the step flow of a battery health state change trend monitoring method according to the present application;

[0069] Fig. 2 FIG. 2 is a schematic diagram of a battery charge-discharge test application scenario according to the present application;

[0070] The implementation, functional features, and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0071] The technical method of the present application will be described clearly and completely below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the present application.

[0072] In addition, the accompanying drawings are included to provide a further understanding of the application, and are incorporated herein and constitute part of this application. The application can be better understood with reference to the drawings together with the description, of which:

[0073] It should be understood that, although the terms "first", "second" or the like can be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first element could be termed a second element, and, similarly, a second element could be termed a first element, without departing from the scope of the example embodiments. The term "and / or" as used herein encompasses any and all combinations of one or more of the associated associated items.

[0074] To achieve the above object, there is provided Figs. 1-2 A method for monitoring the trend of state of health of a storage battery, the method comprising the steps of:

[0075] Step S1: performing a charge-discharge cycle test on the storage battery to obtain dynamic fluctuation data of terminal voltage; performing an alternating current impedance spectrum test on the storage battery to obtain real-time change data of internal resistance; monitoring the charge-discharge capacity attenuation rate data of the storage battery;

[0076] In the embodiments of the present application, there is provided Fig. 2The battery 102 to be tested is subjected to a charge-discharge cycle test to obtain terminal voltage dynamic fluctuation data; the charge-discharge test interface 101 is connected to a charge-discharge test device, and the charge-discharge current is set to 0.2C times of the rated capacity of the battery, where C is the rated capacity unit of the battery in Ah. During the charge-discharge process, the terminal voltage value of the battery is recorded in real time at a sampling interval of 1 second, thereby obtaining the terminal voltage dynamic fluctuation data. The terminal voltage dynamic fluctuation data reflects the voltage change characteristics of the battery during the charge-discharge process, including the rising trend of the voltage in the charging stage and the falling trend of the voltage in the discharging stage. These data can reflect the dynamic process of the internal electrochemical reaction of the battery and the health status of the battery. At the same time, the battery is subjected to an alternating current impedance spectrum test to obtain internal resistance real-time change data. An electrochemical workstation is used as a test device, the battery is connected to the electrochemical workstation, and the frequency range of the alternating current excitation signal is set to 0.01 Hz to 100 kHz, and the amplitude is 10 mV. During the test, the electrochemical workstation applies an alternating current excitation signal to the battery, and measures the response of the battery to the excitation signal, and calculates the impedance value of the battery at different frequencies. The impedance value at each frequency point is recorded at a frequency interval of 0.1 Hz, thereby obtaining the internal resistance real-time change data. The internal resistance real-time change data can reflect the activity of the internal electrode material of the battery, the conductivity of the electrolyte, and the interface state between the electrode and the electrolyte, and is an important parameter for evaluating the health status of the battery. In addition, the charge-discharge capacity attenuation rate data of the battery is monitored. After each charge-discharge cycle is completed, the actual charge-discharge capacity of the battery is recorded. Taking the capacity at the initial charge-discharge cycle as the reference, the difference between the capacity of each subsequent cycle and the initial capacity is calculated, and then divided by the initial capacity to obtain the capacity attenuation percentage. The capacity attenuation percentage of each cycle is recorded according to the cycle number, thereby obtaining the charge-discharge capacity attenuation rate data. The charge-discharge capacity attenuation rate data reflects the capacity decline trend of the battery after multiple charge-discharge cycles, and is a key indicator for measuring the service life and health status of the battery. Through the above technical means and specific operation process, the terminal voltage dynamic fluctuation data, internal resistance real-time change data and charge-discharge capacity attenuation rate data of the battery can be comprehensively obtained.

[0077] Step S2: Perform multi-dimensional analysis on the terminal voltage dynamic fluctuation data, internal resistance real-time change data and charge-discharge capacity attenuation rate data to generate battery characteristic data; perform collaborative analysis on the battery characteristic data, generate collaborative analysis results by analyzing the synchronicity of voltage fluctuation and internal resistance change, and the influence of capacity attenuation on voltage and internal resistance change; map the collaborative analysis results to a preset battery health status model to form health status mapping data.

[0078] In the embodiment of the present application, the terminal voltage dynamic fluctuation data, the internal resistance real-time change data and the charge-discharge capacity attenuation rate data are subjected to multi-dimensional analysis to generate the battery characteristic data. The above three types of data are introduced into a multi-dimensional data analysis module by using a data analysis software. Analysis parameters are set, including a time window of 10 seconds for performing a sliding average processing on the terminal voltage dynamic fluctuation data to smooth the high-frequency noise in the voltage fluctuation; for the internal resistance real-time change data, a linear fitting method is adopted to calculate the change slope thereof in each frequency band, and the frequency band is divided into a range of 0.01 Hz to 100 Hz, with each 10 Hz as an interval; for the charge-discharge capacity attenuation rate data, the average attenuation rate thereof in every 10 charge-discharge cycles is calculated. Through the above processing, the terminal voltage fluctuation characteristic parameters, the internal resistance change characteristic parameters and the capacity attenuation characteristic parameters are extracted, and the three types of characteristic parameters are combined to generate the battery characteristic data. Subsequently, the battery characteristic data is subjected to a collaborative analysis. Through a correlation analysis algorithm, the correlation coefficient between the terminal voltage fluctuation characteristic parameters and the internal resistance change characteristic parameters is calculated to analyze the synchronism thereof. A correlation coefficient threshold of 0.7 is set, and when the correlation coefficient is greater than the threshold, it is determined that the voltage fluctuation and the internal resistance change have significant synchronism. Meanwhile, the influence of the charge-discharge capacity attenuation rate on the voltage fluctuation and the internal resistance change is analyzed, a regression model between the capacity attenuation rate and the voltage fluctuation amplitude and the internal resistance change slope is established through a regression analysis method, a regression coefficient is calculated, and the influence degree of the capacity attenuation on the voltage and internal resistance change is evaluated. According to the synchronism analysis result and the regression analysis result, a collaborative analysis result is generated, including the voltage and internal resistance synchronism determination result and the capacity attenuation influence evaluation result. Finally, the collaborative analysis result is mapped to a preset battery health state model to form the health state mapping data. According to the voltage and internal resistance synchronism determination result, the battery health state is divided into two categories of “synchronous good” and “synchronous abnormal” by using a preset mapping rule; according to the capacity attenuation influence evaluation result, the battery health state is divided into two categories of “capacity attenuation influence significant” and “capacity attenuation influence not significant”. The synchronism determination result and the capacity attenuation influence evaluation result in the collaborative analysis result are combined according to the preset rule to generate the health state mapping data, for example, when the voltage and internal resistance are synchronous and the capacity attenuation influence is not significant, it is mapped as “health state good”, and when the voltage and internal resistance are synchronous and the capacity attenuation influence is significant, it is mapped as “health state abnormal”. Through the above technical means and specific operation process, the whole process from multi-dimensional data analysis to collaborative analysis and then to health state mapping is completed, which provides accurate mapping data for the battery health state change trend monitoring.

[0079] Step S3: identifying the health state characteristics of the battery based on the health state mapping data; if an abnormal pattern exists in the health state characteristic mapping data, marking the abnormal characteristics of the battery and recording as the battery abnormal information;

[0080] In the embodiment of the present application, the health state feature of the battery is identified based on the health state mapping data. The health state mapping data is input into a preset pattern recognition system by using a pattern recognition technology. The system analyzes the health state mapping data by a pre-trained classification algorithm. The classification algorithm adopts a support vector machine (SVM) method, the kernel function of which is selected as a radial basis function (RBF), the penalty parameter C is set to 10, and the γ parameter of the kernel function is set to 0.1. The health state mapping data is classified by the SVM algorithm to identify the current health state feature of the battery, including three categories of "good health state", "general health state" and "abnormal health state". Subsequently, if an abnormal pattern exists in the health state feature mapping data, the abnormal feature of the battery is marked and recorded as the battery abnormal information. In the pattern recognition process, when the SVM algorithm determines that the health state is "abnormal health state", the specific features of the abnormal pattern are further analyzed. By feature extraction technology, the feature parameters related to the abnormality are extracted from the health state mapping data, including the abnormal amplitude of voltage fluctuation, the abnormal slope of internal resistance change and the abnormal rate of capacity attenuation. The threshold of the abnormal amplitude of voltage fluctuation is set to 200 millivolts, the threshold of the abnormal slope of internal resistance change is set to 0.1 ohm / hertz, and the threshold of the abnormal rate of capacity attenuation is set to 5% per 10 cycles. When the detected feature parameters exceed the above thresholds, the corresponding abnormal features are marked. For example, if the voltage fluctuation amplitude exceeds 200 millivolts, it is marked as "voltage fluctuation abnormality"; if the internal resistance change slope exceeds 0.1 ohm / hertz, it is marked as "internal resistance change abnormality"; if the capacity attenuation rate exceeds 5% per 10 cycles, it is marked as "capacity attenuation abnormality". The marked abnormal features and their corresponding values are recorded as the battery abnormal information, including the abnormal type, the abnormal parameter value and the detection time, etc. Through the above technical means and specific operation process, the identification of the health state feature of the battery, as well as the marking and recording of the abnormal features, are realized.

[0081] Step S4: predicting the long-term change trend of the health state of the battery based on the health state feature mapping data to obtain a health prediction result; if the health prediction result shows that the health state of the battery will significantly decrease in a future period of time, a health state warning signal is generated.

[0082] In this embodiment of the invention, the long-term trend of battery health status is predicted based on health status feature mapping data. Time series analysis is employed, and the health status feature mapping data is input into a time series prediction model. Specifically, an autoregressive moving average (ARMA) model is used for prediction, where the autoregressive order p is set to 3 and the moving average order q is set to 2. The health status feature mapping data is preprocessed, including data standardization and missing value imputation, to ensure data integrity and consistency. The health status feature mapping data is fitted using the ARMA model, model parameters are calculated, and the fitted model is used to predict the future health status of the battery. The prediction time range is set to the next 12 months, with a time step of 1 month, yielding a long-term trend prediction result for the health status, including the health status category for each future month (e.g., "good health status," "moderate health status," "abnormal health status") and the corresponding health status index values ​​(e.g., voltage fluctuation amplitude, internal resistance change slope, capacity decay rate, etc.). Subsequently, based on the health prediction results, it is determined whether the battery health status will significantly decline in the future. The criteria for determining a significant decline in battery health status are set as follows: within the next 12 months, the battery health status changes from "good" or "fair" to "abnormal," and the capacity decay rate exceeds 10% every 12 months. If the health prediction result meets the above criteria, a health status warning signal is generated. The warning signal includes the battery identifier, the warning time, the expected time of significant health status decline, and the main characteristic parameters leading to the decline in health status (such as abnormal voltage fluctuation amplitude, abnormal slope of internal resistance change, abnormal capacity decay rate, etc.). Through the above technical means and specific operational procedures, the long-term trend of battery health status is predicted based on health status feature mapping data, and a warning signal is generated when the prediction result shows a significant decline in health status, providing timely decision-making basis for battery maintenance and management.

[0083] As an example of the present invention, reference is made to Fig. 2 As shown, in this example, step S1 includes:

[0084] Step S11: In the charge-discharge cycle test, the battery is charged with a preset constant current until the battery reaches the rated voltage.

[0085] Step S12: Discharge the battery at the same preset constant current until the battery voltage drops to the preset minimum voltage threshold; repeat the above charging and discharging process at least three times, record the terminal voltage and time data during each charging and discharging process, and calculate the terminal voltage dynamic fluctuation data.

[0086] Step S13: In the alternating current impedance spectrum test, an alternating current signal with a frequency range of 0.01 Hz to 100 kHz is applied, and the impedance value of the storage battery at different frequencies is measured;

[0087] Step S14: Record the curve of the impedance value changing with the frequency, and extract the real-time change data of the internal resistance;

[0088] Step S15: In the charge-discharge capacity attenuation rate monitoring, the capacity value of each charge-discharge cycle is recorded, and the capacity difference value of adjacent two charge-discharge cycles is calculated;

[0089] Step S16: According to the ratio of the capacity difference value to the initial capacity, the charge-discharge capacity attenuation rate data is determined.

[0090] In the embodiment of the present application, a high-precision direct current power supply is used to charge the battery. The positive electrode of the battery is connected to the positive electrode of the direct current power supply, and the negative electrode is connected to the negative electrode of the direct current power supply. The output current of the direct current power supply is set to 0.1C times of the rated capacity of the battery, where C is the rated capacity of the battery, and the unit is Ah. Start the charging process and monitor the terminal voltage of the battery in real time. When the terminal voltage reaches the rated voltage of the battery, stop charging. During the charging process, the terminal voltage of the battery and the corresponding charging time are recorded at a sampling interval of 1 second. In step S12, the battery is discharged using the same constant current value as the charging process. The positive electrode of the battery is connected to the positive electrode of the electronic load, and the negative electrode is connected to the negative electrode of the electronic load. The current value of the electronic load is set to 0.1C times. Start the discharging process and monitor the terminal voltage of the battery in real time. When the terminal voltage drops to a preset minimum voltage threshold, stop discharging. The minimum voltage threshold is set according to the technical specifications of the battery, and is usually 30% of the rated voltage of the battery. During the discharging process, the terminal voltage of the battery and the corresponding discharging time are also recorded at a sampling interval of 1 second. In step S13, an alternating current impedance spectrum test is performed. An electrochemical workstation is used as the test equipment, and the battery is connected to the test port of the electrochemical workstation. The frequency range of the alternating current signal is set to 0.01 Hz to 100 kHz, and the amplitude of the alternating current signal is 10 mV. Start the test, and the electrochemical workstation applies an alternating current signal to the battery and measures the impedance value of the battery at different frequencies. The impedance value at each frequency point is recorded, including the real part and the imaginary part, with a step frequency of 0.1 Hz. In step S14, a curve of the impedance value versus frequency is drawn according to the recorded impedance value data. Through curve analysis, real-time resistance change data is extracted. The real-time resistance change data includes the impedance modulus value and the phase angle at different frequencies. The average value of the impedance modulus value at each frequency point is calculated as the real-time resistance value at that frequency point. At the same time, the change trend of the impedance modulus value in different frequency bands is recorded, such as the change slope in the low frequency band (0.01 Hz to 1 Hz) and the high frequency band (10 Hz to 100 kHz). In step S15, the charge-discharge capacity attenuation rate monitoring is performed. After each charge-discharge cycle, the actual charge-discharge capacity value of the battery is recorded. The capacity value is calculated by integration, that is, the product of the recorded current value and time is accumulated to obtain the capacity value during charging or discharging. The capacity value of each charge-discharge cycle is recorded, and the capacity difference value of adjacent two charge-discharge cycles is calculated. In step S16, the charge-discharge capacity attenuation rate data is determined according to the ratio of the capacity difference value to the initial capacity. The initial capacity is the capacity value at the first charge-discharge cycle. The formula for calculating the capacity attenuation rate is: capacity attenuation rate = (capacity difference value of adjacent two charge-discharge cycles / initial capacity) x 100%. The capacity attenuation rate of each cycle is recorded according to the cycle number to form the charge-discharge capacity attenuation rate data.

[0091] Preferably, the multi-dimensional analysis of the terminal voltage dynamic fluctuation data, the internal resistance real-time change data and the charge-discharge capacity attenuation rate data in step S2 comprises:

[0092] The terminal voltage dynamic fluctuation data is segmented according to time sequence, the average value and the standard deviation of each segment of data are calculated, and a voltage fluctuation feature vector is generated;

[0093] The internal resistance real-time change data is grouped according to the frequency range, the internal resistance average value and the internal resistance change rate of each group of data are calculated, and an internal resistance feature vector is generated;

[0094] The charge-discharge capacity attenuation rate data is grouped according to the number of charge-discharge cycles, the capacity attenuation percentage and the capacity attenuation slope of each group of data are calculated, and a capacity attenuation feature vector is generated;

[0095] The voltage fluctuation feature vector, the internal resistance feature vector and the capacity attenuation feature vector are normalized and time-aligned to generate the battery feature data.

[0096] In the embodiment of the present application, when processing the dynamic voltage fluctuation data, first segment it according to the time sequence. Set the time length of each segment to 10 seconds, and use data processing software to perform sliding window segmentation processing on the terminal voltage data, with a window step of 1 second. For each segment of data, calculate its average value and standard deviation. The average value is obtained by adding all the voltage values in each segment and dividing by the number of data points in the segment; the standard deviation is obtained by calculating the square root of the variance of each segment of data. Combine the average value and standard deviation of each segment into a two-dimensional vector to form a voltage fluctuation feature vector. For the internal resistance real-time change data, perform grouping processing according to the frequency range. Divide the frequency range of 0.01 Hz to 100 kHz into multiple intervals, with each interval being 10 Hz. Process the impedance values in each frequency interval to calculate the average internal resistance, which is the arithmetic mean of the impedance modulus values in each interval. At the same time, calculate the internal resistance change rate by calculating the ratio of the difference between the impedance modulus values of adjacent frequency points to the impedance modulus value of the initial frequency point. Combine the average internal resistance and internal resistance change rate of each frequency interval into a two-dimensional vector to form an internal resistance feature vector. When processing the charge and discharge capacity attenuation rate data, group it according to the number of charge and discharge cycles. Set every 10 charge and discharge cycles as a group, and calculate the capacity attenuation percentage of each group, which is the ratio of the capacity difference between the adjacent two cycles in each group to the initial capacity. At the same time, calculate the capacity attenuation slope by fitting the capacity attenuation data of each group using the linear regression method to obtain the slope value. Combine the capacity attenuation percentage and capacity attenuation slope of each group into a two-dimensional vector to form a capacity attenuation feature vector. Finally, normalize the voltage fluctuation feature vector, internal resistance feature vector, and capacity attenuation feature vector. Use the min-max normalization method to scale each component value of each feature vector to the range of 0 to 1. The specific operation is as follows: for each component in each feature vector, calculate its minimum value and maximum value in all data, then subtract the minimum value from the value of the component and divide by the difference between the maximum value and the minimum value. After normalization, align the time dimensions of the feature vectors to ensure that the time segments of the voltage fluctuation feature vector correspond to the frequency intervals of the internal resistance feature vector and the cycle numbers of the capacity attenuation feature vector. Through the above technical means and specific operation process, complete battery feature data is generated, providing a standardized data basis for subsequent health state analysis.

[0097] Preferably, the battery feature data is analyzed in step S2, and the synchronization of voltage fluctuation and internal resistance change, and the influence of capacity attenuation on voltage and internal resistance change are analyzed, including:

[0098] Perform time series analysis on the voltage fluctuation feature vector to extract the periodicity and trend components of the voltage fluctuation, and generate voltage fluctuation period features and trend features;

[0099] The internal resistance characteristic vector is analyzed in the frequency domain, high-frequency and low-frequency components of the internal resistance change are extracted, and internal resistance high-frequency characteristics and internal resistance low-frequency characteristics are generated;

[0100] The capacity attenuation characteristic vector is nonlinearly fitted, the acceleration and deceleration stage characteristics of the capacity attenuation are extracted, and capacity attenuation acceleration characteristics and capacity attenuation deceleration characteristics are generated;

[0101] Cross-correlation analysis is performed on the voltage fluctuation period characteristic and the internal resistance high-frequency characteristic to evaluate the synchronism of the periodic change of the voltage fluctuation and the high-frequency change of the internal resistance.

[0102] In the embodiment of the present application, when the voltage fluctuation characteristic vector is analyzed in time series, the fast Fourier transform (FFT) technique is used. The mean component in the voltage fluctuation characteristic vector is taken as the input data, and the sampling frequency is set to 10 Hz, that is, 10 data points are collected per second. The time series data is converted into frequency domain data by the FFT algorithm, the components in the frequency range of 0.01 Hz to 1 Hz are extracted as periodic components, and the components with a frequency lower than 0.01 Hz are extracted as trend components. The amplitude spectrum and phase spectrum of the periodic components are calculated to generate the voltage fluctuation period characteristic; the amplitude of the trend component is calculated to generate the voltage fluctuation trend characteristic. When the internal resistance characteristic vector is analyzed in the frequency domain, the band-pass filter technique is used. The internal resistance mean component in the internal resistance characteristic vector is taken as the input data, the cutoff frequency range of the high-frequency band-pass filter is set to 10 Hz to 100 kHz, and the cutoff frequency range of the low-frequency band-pass filter is set to 0.01 Hz to 1 Hz. The high-frequency and low-frequency components are extracted by the band-pass filter respectively. The amplitude and phase angle of the high-frequency component are calculated to generate the internal resistance high-frequency characteristic; the amplitude and phase angle of the low-frequency component are calculated to generate the internal resistance low-frequency characteristic. When the capacity attenuation characteristic vector is nonlinearly fitted, the polynomial fitting technique is used. The capacity attenuation percentage component in the capacity attenuation characteristic vector is taken as the input data, and the order of the fitting polynomial is set to 3. The relationship curve between the capacity attenuation percentage and the charge and discharge cycle number is fitted by the least squares method. According to the slope change of the fitting curve, the acceleration stage characteristic and the deceleration stage characteristic of the capacity attenuation are extracted. The acceleration stage characteristic is the part of the fitting curve with a slope greater than 0.05% per cycle, and the deceleration stage characteristic is the part of the fitting curve with a slope less than 0.05% per cycle. Finally, cross-correlation analysis is performed on the voltage fluctuation period characteristic and the internal resistance high-frequency characteristic. The cross-correlation function technique is used, and the amplitude spectrum of the voltage fluctuation period characteristic and the amplitude of the internal resistance high-frequency characteristic are taken as the input data. The time delay range is set to -10 seconds to 10 seconds, and the step is 1 second. The cross-correlation function value is calculated to evaluate the synchronism of the periodic change of the voltage fluctuation and the high-frequency change of the internal resistance. The synchronism is quantified by the maximum value of the cross-correlation function and the corresponding delay time. The closer the maximum value is to 1 and the closer the delay time is to 0, the stronger the synchronism.

[0103] Especially important is to cross-correlate the voltage fluctuation period feature and the internal resistance high-frequency feature, evaluate the synchronism of the periodic change of the voltage fluctuation and the high-frequency change of the internal resistance, including:

[0104] Aligning the voltage fluctuation period feature and the internal resistance high-frequency feature;

[0105] Determine the cross-correlation coefficient between the voltage fluctuation period feature and the internal resistance high-frequency feature, calculate the correlation of the voltage fluctuation and the high-frequency change of the internal resistance in each period;

[0106] According to the cross-correlation coefficient, generate synchronism evaluation data, wherein the synchronism evaluation data includes specific numerical values and charts of the synchronism evaluation.

[0107] In the embodiment of the application, when aligning the voltage fluctuation period feature and the internal resistance high-frequency feature, time series alignment technology is adopted. First, the time reference point of the voltage fluctuation period feature is determined, which is usually the starting point or peak point of the voltage fluctuation period. At the same time, the time reference point of the internal resistance high-frequency feature is determined, which is usually the peak point or zero-crossing point of the internal resistance high-frequency change. By calculating the time offset between the two features, the time axis of the internal resistance high-frequency feature is adjusted to align with the time reference of the voltage fluctuation period feature. During the alignment process, the interpolated method is used to time-interpolate the adjusted internal resistance high-frequency feature data to ensure that the two features have the same sampling points in time. When determining the cross-correlation coefficient between the voltage fluctuation period feature and the internal resistance high-frequency feature, cross-correlation analysis technology is adopted. The amplitude sequence of the aligned voltage fluctuation period feature and the amplitude sequence of the internal resistance high-frequency feature are used as input data. The time delay range is set to -10 seconds to 10 seconds, with a step of 0.1 seconds. By calculating the cross-correlation function, the correlation coefficient value at each time delay is obtained. For each voltage fluctuation period, the average correlation coefficient of the voltage fluctuation and the internal resistance high-frequency change in that period is calculated, which is the average value of the correlation coefficients at all time delay points in that period. The average correlation coefficient reflects the overall correlation of the voltage fluctuation and the internal resistance high-frequency change in that period. According to the cross-correlation coefficient, synchronism evaluation data is generated. The synchronism evaluation data includes specific numerical values and charts. The numerical part is the average cross-correlation coefficient in each voltage fluctuation period, as well as the maximum value, minimum value and mean value of the average cross-correlation coefficient in all periods. The chart part is the cross-correlation coefficient curve with time delay, as well as the column chart of the average cross-correlation coefficient in each voltage fluctuation period. Through these data and charts, the synchronism between the voltage fluctuation period feature and the internal resistance high-frequency feature is intuitively displayed.

[0108] Preferably, the battery feature data in step S2 is analyzed in coordination, and the synchronism of the voltage fluctuation and the internal resistance change, and the influence of capacity attenuation on the voltage and internal resistance change are further analyzed, including:

[0109] The capacity attenuation acceleration feature and the voltage fluctuation trend feature are differentially analyzed, a difference feature vector is generated by calculating the difference of the two at the same time point, to evaluate the influence of the capacity attenuation acceleration stage on the voltage fluctuation trend;

[0110] The capacity attenuation deceleration feature and the internal resistance low-frequency feature are ratio analyzed, a ratio feature vector is generated by calculating the ratio of the two at the same time point, to evaluate the influence of the capacity attenuation deceleration stage on the internal resistance low-frequency change.

[0111] In the embodiment of the present application, when performing differential analysis of the capacity attenuation acceleration feature and the voltage fluctuation trend feature, first, ensure that the two feature vectors are aligned in time. Linear interpolation method is used to perform time alignment processing on the capacity attenuation acceleration feature vector and the voltage fluctuation trend feature vector, so that they have corresponding feature values at the same time point. Then, for each time point, calculate the difference between the capacity attenuation acceleration feature value and the voltage fluctuation trend feature value. The result of differential analysis is a new feature vector, called difference feature vector, each element of which represents the influence degree of the capacity attenuation acceleration stage on the voltage fluctuation trend at the corresponding time point. When performing ratio analysis of the capacity attenuation deceleration feature and the internal resistance low-frequency feature, first, perform time alignment processing on the two feature vectors. The same linear interpolation method is used to ensure that the capacity attenuation deceleration feature vector and the internal resistance low-frequency feature vector have corresponding feature values at the same time point. Then, for each time point, calculate the ratio of the capacity attenuation deceleration feature value and the internal resistance low-frequency feature value. The result of ratio analysis is a new feature vector, called ratio feature vector, each element of which represents the influence degree of the capacity attenuation deceleration stage on the internal resistance low-frequency change at the corresponding time point. In the above two analysis processes, the precision of time alignment processing is set to millisecond level to ensure the accuracy of difference and ratio calculation. The generation process of difference feature vector and ratio feature vector is based on strict mathematical operation, without any subjective judgment or assumption. Through these technical means and specific operation process, the influence of the capacity attenuation acceleration stage on the voltage fluctuation trend and the influence of the capacity attenuation deceleration stage on the internal resistance low-frequency change can be quantitatively evaluated, providing important data support for further analysis of the state of health of the battery.

[0112] Preferably, the step S2 of mapping the collaborative analysis result to the preset battery state of health model comprises:

[0113] Feature coding is performed on the synchronicity data of voltage fluctuation and internal resistance change in the collaborative analysis result;

[0114] Feature coding is performed on the influence data of capacity attenuation on voltage and internal resistance change;

[0115] The coded synchronism data and the capacity attenuation influence data are spliced to form a binary feature vector;

[0116] The binary feature vector is input into a preset battery health state model to output health state mapping data.

[0117] In the embodiment of the present application, when the synchronism data of voltage fluctuation and internal resistance change in the collaborative analysis result is encoded, binary encoding technology is adopted. According to the specific value of the synchronism evaluation, the synchronism threshold is set to 0.7. When the synchronism evaluation value is greater than or equal to 0.7, it is encoded as 1, indicating that the voltage fluctuation and the internal resistance change have strong synchronism; when the synchronism evaluation value is less than 0.7, it is encoded as 0, indicating that the synchronism is weak. The synchronism evaluation value of each cycle is encoded according to the above rule to generate a synchronism feature encoding vector. When the influence data of capacity attenuation on voltage and internal resistance change is encoded, a similar binary encoding method is adopted. The capacity attenuation influence threshold is set to 0.1, indicating the influence degree of capacity attenuation on voltage or internal resistance change. For the influence data of capacity attenuation on voltage change, when the influence degree is greater than or equal to 0.1, it is encoded as 1; when it is less than 0.1, it is encoded as 0. For the influence data of capacity attenuation on internal resistance change, it is also encoded according to the above rule. The influence data of capacity attenuation on voltage change and internal resistance change are encoded respectively and combined into a capacity attenuation influence feature encoding vector. When the encoded synchronism data and the capacity attenuation influence data are spliced, vector splicing technology is adopted. The synchronism feature encoding vector and the capacity attenuation influence feature encoding vector are sequentially connected to form a complete binary feature vector. In the splicing process, the length of each feature encoding vector is ensured to be consistent, and the length is adjusted by padding or truncation, so that the spliced binary feature vector has a fixed length. Finally, the spliced binary feature vector is input into a preset battery health state model. The model outputs corresponding health state mapping data according to the input binary feature vector through a preset mapping rule. The health state mapping data includes the health state category of the battery, such as "good health state", "general health state" or "abnormal health state", and the related health state index value. The output health state mapping data is generated based on the preset mapping rule and does not involve any subjective judgment or assumption.

[0118] Preferably, the health state feature of the battery is identified based on the health state mapping data in step S3, including:

[0119] The voltage fluctuation amplitude, the internal resistance change rate and the capacity attenuation percentage are extracted from the health state mapping data and recorded as battery feature parameters;

[0120] The battery feature parameters are compared with the preset health state feature indicators to obtain a health state comparison result;

[0121] According to the health state comparison result, the current health state characteristics of the storage battery are determined, including health state level, voltage stability, internal resistance uniformity and capacity retention rate, so as to determine whether the storage battery is in normal health state.

[0122] In the embodiment of the present application, when the storage battery characteristic parameters are extracted from the health state mapping data, data analysis technology is adopted. Firstly, the voltage fluctuation amplitude field, the internal resistance change rate field and the capacity attenuation percentage field in the health state mapping data are located. The voltage fluctuation amplitude field records the maximum change range of the terminal voltage of the storage battery during the charging and discharging process, and the unit is volt (V); the internal resistance change rate field records the change percentage of the internal resistance of the storage battery relative to the initial internal resistance, and the unit is percentage (%); the capacity attenuation percentage field records the attenuation percentage of the current capacity of the storage battery relative to the initial capacity, and the unit is percentage (%). The values of these fields are extracted and recorded as the storage battery characteristic parameters. When the storage battery characteristic parameters are compared with the preset health state characteristic threshold, threshold comparison technology is adopted. The preset health state characteristic threshold includes the voltage fluctuation amplitude threshold, the internal resistance change rate threshold and the capacity attenuation percentage threshold. The voltage fluctuation amplitude threshold is set to 0.5V, the internal resistance change rate threshold is set to 10%, and the capacity attenuation percentage threshold is set to 20%. The extracted storage battery characteristic parameters are compared with these thresholds respectively, and the health state comparison result is generated. The comparison result includes the Boolean value (True or False) of whether each characteristic parameter exceeds the preset threshold. When the current health state characteristics of the storage battery are determined according to the health state comparison result, logical judgment technology is adopted. According to the comparison result, the health state level is defined: if all the characteristic parameters do not exceed the threshold, the health state level is "good"; if one of the characteristic parameters exceeds the threshold, the health state level is "general"; if multiple characteristic parameters exceed the threshold, the health state level is "abnormal". At the same time, according to whether the voltage fluctuation amplitude exceeds the threshold, the voltage stability is judged, which is "stable" if it does not exceed the threshold and "unstable" if it exceeds the threshold; according to whether the internal resistance change rate exceeds the threshold, the internal resistance uniformity is judged, which is "uniform" if it does not exceed the threshold and "non-uniform" if it exceeds the threshold; according to whether the capacity attenuation percentage exceeds the threshold, the capacity retention rate is judged, which is "well maintained" if it does not exceed the threshold and "poorly maintained" if it exceeds the threshold. Through the above technical means and specific operation process, whether the storage battery is in normal health state is finally determined.

[0123] Preferably, in step S3, if it is detected that there is an abnormal pattern in the health state characteristic mapping data, the abnormal characteristics of the storage battery are marked and recorded as the abnormal information of the storage battery, including:

[0124] The voltage fluctuation amplitude, the internal resistance change rate and the capacity attenuation percentage are extracted from the health state mapping data to form a set of detection characteristic values;

[0125] threshold judgment is performed on each feature value in the detection feature value set to determine whether it exceeds the preset normal range;

[0126] If any one of the feature values of the voltage fluctuation amplitude, the internal resistance change rate or the capacity attenuation percentage exceeds the preset normal range, it is marked as a potential abnormal feature;

[0127] neighborhood analysis is performed on the potential abnormal feature to check whether the feature values of the data points around it are also abnormal to confirm the existence of an abnormal pattern;

[0128] If the existence of an abnormal pattern is confirmed, the feature values corresponding to the abnormal pattern are marked as battery abnormal features and recorded as battery abnormal information.

[0129] In the embodiment of the present application, when extracting feature values from the health state mapping data, a data extraction technique is adopted. The fields of the voltage fluctuation amplitude, the internal resistance change rate and the capacity attenuation percentage in the health state mapping data are located, and their numerical values are extracted respectively to form a detection feature value set. The unit of the voltage fluctuation amplitude is volt (V), the unit of the internal resistance change rate is percentage (%), and the unit of the capacity attenuation percentage is percentage (%). When performing threshold judgment on each feature value in the detection feature value set, a threshold comparison technique is adopted. The preset normal range thresholds are as follows: the voltage fluctuation amplitude threshold is 0.5 V, the internal resistance change rate threshold is 10%, and the capacity attenuation percentage threshold is 20%. Each feature value is compared with the corresponding threshold to determine whether it exceeds the preset normal range. If a certain feature value exceeds the corresponding threshold, the feature value is marked as a potential abnormal feature. When performing neighborhood analysis on the potential abnormal feature, a sliding window technique is adopted. The size of the sliding window is set to 5 data points, and the data points around the potential abnormal feature are taken as the center to form a neighborhood containing 5 data points. The feature values of each data point in the neighborhood are calculated to check whether they also exceed the corresponding thresholds. If more than half of the data points in the neighborhood exceed the thresholds, the existence of an abnormal pattern is confirmed. If the existence of an abnormal pattern is confirmed, the feature values corresponding to the abnormal pattern are marked as battery abnormal features and recorded as battery abnormal information. The recorded battery abnormal information includes the abnormal feature type (such as voltage fluctuation amplitude abnormality, internal resistance change rate abnormality or capacity attenuation percentage abnormality), the specific numerical value of the abnormal feature, the timestamp of the abnormality occurrence and the neighborhood range of the abnormal pattern. Through the above technical means and specific operation process, the whole process from feature value extraction to abnormal pattern confirmation is completed, and accurate abnormal information record is provided for the monitoring of the health state of the battery.

[0130] Preferably, the step S4 of predicting the long-term change trend of the battery health state based on the health state feature mapping data comprises:

[0131] performing time series analysis on the health status feature mapping data, calculating the change trend of multiple time points, and obtaining a time series analysis result;

[0132] determining a long-term change trend of each feature value according to the time series analysis result, and calculating a feature change rate of the health status feature mapping data value at the past 10 time points;

[0133] if the feature value change rate is continuously positive, marking as a health status rising trend;

[0134] if the feature value change rate is continuously negative, marking as a health status falling trend;

[0135] if the feature value change rate fluctuates within a preset threshold range, marking as a health status stable trend;

[0136] comprehensively analyzing the long-term change trend of each feature value, and predicting a long-term change trend of the overall health status of the battery.

[0137] In the embodiment of the present application, when performing time series analysis on the health state feature mapping data, a differential technique is used to calculate the change trend of multiple time points. First, time series data of characteristic values such as voltage fluctuation amplitude, internal resistance change rate and capacity attenuation percentage are extracted from the health state feature mapping data. The time interval is set to 1 day, and the data of the past 30 days are selected as the analysis object. By calculating the difference between adjacent time points, the change amount of each characteristic value in each time interval is obtained, and these change amounts are arranged in chronological order to form the time series analysis result. According to the time series analysis result, the long-term change trend of each characteristic value is determined. The data of the past 10 time points are selected, and the change rate of each characteristic value is calculated. The specific operation is as follows: for each characteristic value, the difference between the characteristic value at the current time point and the characteristic value at the 10 time points before is calculated, and then the difference is divided by the characteristic value at the 10 time points before to obtain the characteristic change rate. The characteristic change rates of the voltage fluctuation amplitude, the internal resistance change rate and the capacity attenuation percentage are calculated respectively. According to the positive and negative situation of the characteristic value change rate, the health state trend is marked. The preset threshold range is set to ±5%. If the characteristic value change rate is continuously positive, that is, the change rate of the past 10 consecutive time points is greater than 0, the characteristic value is marked as a health state upward trend; if the characteristic value change rate is continuously negative, that is, the change rate of the past 10 consecutive time points is less than 0, the characteristic value is marked as a health state downward trend; if the characteristic value change rate fluctuates between -5% and 5%, the characteristic value is marked as a health state stable trend. The long-term change trend of each characteristic value is comprehensively analyzed to predict the long-term change trend of the overall health state of the battery. According to the importance of the voltage fluctuation amplitude, the internal resistance change rate and the capacity attenuation percentage to the health state of the battery, weight coefficients are respectively assigned. The weight coefficient of the voltage fluctuation amplitude is 0.4, the weight coefficient of the internal resistance change rate is 0.3, and the weight coefficient of the capacity attenuation percentage is 0.3. The long-term change trend marking result of each characteristic value is converted into a numerical value: the health state upward trend is +1, the health state downward trend is -1, and the health state stable trend is 0. The weighted average value is calculated, that is, the change trend numerical value of each characteristic value is multiplied by the corresponding weight coefficient and then added. According to the numerical value of the comprehensive change trend, the long-term change trend of the overall health state of the battery is predicted: if the comprehensive change trend is greater than 0, it is predicted that the overall health state of the battery presents an upward trend; if the comprehensive change trend is less than 0, it is predicted that the overall health state of the battery presents a downward trend; if the comprehensive change trend is equal to 0, it is predicted that the overall health state of the battery remains stable.

[0138] Especially important is that if the health prediction result shows that the battery health state will significantly decrease in the future period of time in step S4, the following steps are performed:

[0139] extracting the change trend of the health state level from the health prediction result;

[0140] setting a health state decline threshold, wherein the health state decline threshold represents a critical change rate of a health state level decline;

[0141] comparing the predicted change rate of the health state level with the health state decline threshold to determine whether there is a significant decline trend;

[0142] if the predicted change rate of the health state level is lower than the health state decline threshold, it is determined that the battery health state will significantly decline in the future, and a health state warning signal is generated;

[0143] According to the health state warning signal, the warning level setting is divided into first-level warning, second-level warning and third-level warning.

[0144] In this embodiment of the invention, data parsing technology is used to extract the trend of health status level changes from the health prediction results. The health status level field in the health prediction results is located. This field records the health status level of the battery at different future time points, typically divided into three levels: "Good," "Average," and "Abnormal." By analyzing the changes of these levels over time, the trend of health status level changes is extracted, specifically the change in health status level between adjacent time points. An empirical parameter setting technique is used when setting the health status decline threshold. The health status decline threshold represents the critical rate of change for the health status level to decline, and is set according to the actual application scenario of the battery and health status assessment standards. For example, setting the health status decline threshold to -0.2 means that when the predicted rate of change of the health status level is lower than -0.2, a significant downward trend in health status is considered. This threshold setting is based on statistical analysis of the battery health status change patterns and actual operating experience. A threshold comparison technique is used when comparing the predicted rate of change of the health status level with the health status decline threshold. The predicted rate of change of the health status level is calculated by comparing the health status level values ​​at adjacent time points to determine the rate of change. The calculated rate of change is compared with a preset health status decline threshold to determine if a significant downward trend exists. If the predicted rate of change for the health status level is lower than the health status decline threshold, it is determined that the battery health status will significantly decline in the future. When setting warning levels into Level 1, Level 2, and Level 3 based on health status warning signals, a tiered warning technology is employed. The warning level is determined based on the difference between the predicted rate of change for the health status level and the health status decline threshold. For example, if the predicted rate of change is lower than the health status decline threshold of -0.2 and the difference is greater than -0.3, it is set to Level 1; if the difference is between -0.3 and -0.5, it is set to Level 2; and if the difference is less than -0.5, it is set to Level 3. A Level 1 warning indicates a relatively slight decline in health status, a Level 2 warning indicates a relatively significant decline, and a Level 3 warning indicates a relatively severe decline. Through the above technical means and specific operational procedures, a tiered warning system for battery health status is achieved.

[0145] This specification also provides a battery health status change trend monitoring system for performing the battery health status change trend monitoring method described above. The battery health status change trend monitoring system includes:

[0146] The battery data acquisition module is used to perform charge-discharge cycle tests on the battery to obtain dynamic fluctuation data of the terminal voltage; to perform AC impedance spectrum tests on the battery to obtain real-time change data of internal resistance; and to monitor the charge-discharge capacity decay rate data of the battery.

[0147] The health state analysis module is configured to perform multi-dimensional analysis on the terminal voltage dynamic fluctuation data, the internal resistance real-time change data and the charge-discharge capacity attenuation rate data to generate battery characteristic data; perform collaborative analysis on the battery characteristic data, generate a collaborative analysis result by analyzing the synchronism of voltage fluctuation and internal resistance change and the influence of capacity attenuation on voltage and internal resistance change; and map the collaborative analysis result to a preset battery health state model to form health state mapping data.

[0148] The battery abnormality feature detection module is configured to identify the health state feature of the battery based on the health state mapping data; if an abnormal pattern exists in the health state feature mapping data, mark the battery abnormality feature and record it as battery abnormality information.

[0149] The health state prediction module is configured to predict the long-term change trend of the battery health state based on the health state feature mapping data to obtain a health prediction result; if the health prediction result shows that the battery health state will significantly decrease in a future period of time, generate a health state early warning signal.

[0150] Therefore, from any viewpoint, the embodiments should be considered as exemplary and non-limiting, the scope of the present application being defined by the appended claims and not by the above description, therefore all the variations falling within the meaning and the scope of the equivalent elements of the application file are intended to be encompassed within the present application.

[0151] The above description is merely one specific implementation of the application, which enables those skilled in the art to understand or implement the application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the application. Therefore, the present application will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for monitoring the trend of changes in the health status of a storage battery, characterized in that, Includes the following steps: Step S1: Perform charge-discharge cycle tests on the battery to obtain dynamic fluctuation data of the terminal voltage; perform AC impedance spectroscopy tests on the battery to obtain real-time change data of internal resistance; monitor the charge-discharge capacity decay rate data of the battery. Step S2: Perform multidimensional analysis on the terminal voltage dynamic fluctuation data, internal resistance real-time change data, and charge / discharge capacity decay rate data to generate battery characteristic data; perform collaborative analysis on the battery characteristic data, and generate collaborative analysis results by analyzing the synchronicity of voltage fluctuation and internal resistance change, as well as the impact of capacity decay on voltage and internal resistance changes. The collaborative analysis results are mapped to a pre-defined battery health status model to form health status mapping data; Step S3: Identify the health status characteristics of the battery based on the health status mapping data; If an abnormal pattern is detected in the health status feature mapping data, the abnormal battery feature is marked and recorded as battery abnormal information. Step S4: Predict the long-term trend of battery health status based on health status feature mapping data to obtain health prediction results; if the health prediction results show that the battery health status will decline significantly in the future, a health status warning signal will be generated.

2. The method for monitoring the trend of battery health status changes according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: In the charge-discharge cycle test, the battery is charged with a preset constant current until the battery reaches the rated voltage. Step S12: Discharge the battery at the same preset constant current until the battery voltage drops to the preset minimum voltage threshold; repeat the above charging and discharging process at least three times, record the terminal voltage and time data during each charging and discharging process, and calculate the terminal voltage dynamic fluctuation data. Step S13: In the AC impedance spectrum test, an AC signal with a frequency range of 0.01Hz to 100kHz is applied, and the impedance value of the battery at different frequencies is measured. Step S14: Record the curve of impedance value changing with frequency and extract the real-time change data of internal resistance; Step S15: In the charge and discharge capacity decay rate monitoring, record the capacity value of each charge and discharge cycle and calculate the capacity difference between two adjacent charge and discharge cycles. Step S16: Determine the charge / discharge capacity decay rate data based on the ratio of the capacity difference to the initial capacity.

3. The method for monitoring the trend of changes in the health status of a battery according to claim 1, characterized in that, Step S2 involves multidimensional analysis of the terminal voltage dynamic fluctuation data, internal resistance real-time change data, and charge / discharge capacity decay rate data, including: The dynamic fluctuation data of terminal voltage is segmented according to the time series, and the average value and standard deviation of each segment are calculated to generate a voltage fluctuation feature vector. The real-time internal resistance change data is grouped according to frequency range, and the average internal resistance and internal resistance change rate of each group are calculated to generate an internal resistance feature vector. The charge / discharge capacity decay rate data are grouped according to the number of charge / discharge cycles. The capacity decay percentage and capacity decay slope of each group are calculated to generate a capacity decay feature vector. The voltage fluctuation feature vector, internal resistance feature vector, and capacity decay feature vector are normalized and time-aligned to generate battery feature data.

4. The method for monitoring the trend of changes in the health status of a battery according to claim 1, characterized in that, Step S2 involves a collaborative analysis of battery characteristic data. This analysis examines the synchronicity between voltage fluctuations and internal resistance changes, as well as the impact of capacity decay on voltage and internal resistance changes. Time series analysis is performed on the voltage fluctuation feature vector to extract the periodic and trend components of the voltage fluctuation, generating periodic and trend features of the voltage fluctuation. Frequency domain analysis is performed on the internal resistance feature vector to extract the high-frequency and low-frequency components of the internal resistance variation, generating high-frequency and low-frequency features of the internal resistance. Nonlinear fitting is performed on the capacity decay feature vector to extract the acceleration and deceleration phase features of capacity decay, generating capacity decay acceleration and deceleration features. Cross-correlation analysis was performed on the periodic characteristics of voltage fluctuations and the high-frequency characteristics of internal resistance to assess the synchronicity between the periodic changes of voltage fluctuations and the high-frequency changes of internal resistance.

5. The method for monitoring the trend of battery health status changes according to claim 4, characterized in that, Step S2 involves a collaborative analysis of battery characteristic data. This analysis includes examining the synchronicity between voltage fluctuations and internal resistance changes, as well as the impact of capacity decay on voltage and internal resistance changes. Differential analysis is performed on the characteristics of accelerated capacity decay and the characteristics of voltage fluctuation trend. By calculating the difference between the two at the same time point, a differential feature vector is generated to evaluate the impact of the accelerated capacity decay stage on the voltage fluctuation trend. The ratio analysis of the capacity decay deceleration characteristics and the low-frequency internal resistance characteristics is performed. By calculating the ratio of the two at the same time point, a ratio feature vector is generated to evaluate the impact of the capacity decay deceleration stage on the low-frequency change of internal resistance.

6. The method for monitoring the trend of changes in the health status of a battery according to claim 1, characterized in that, Step S2, which maps the collaborative analysis results to a preset battery health status model, includes: The synchronization data of voltage fluctuations and internal resistance changes in the collaborative analysis results are feature-encoded; Feature encoding is performed on the data regarding the impact of capacity decay on voltage and internal resistance changes; The encoded synchronization data and the data affected by capacity decay are concatenated to form a binary feature vector; The binary feature vector is input into the preset battery health status model, and the health status mapping data is output.

7. The method for monitoring the trend of battery health status changes according to claim 1, characterized in that, Step S3, which identifies the health status characteristics of the battery based on health status mapping data, includes: The voltage fluctuation amplitude, internal resistance change rate and capacity decay percentage are extracted from the health status mapping data and recorded as battery characteristic parameters. The battery characteristic parameters are compared with the preset health status characteristic indicators to obtain the health status comparison results; The current health status characteristics of the battery are determined based on the health status comparison results, including health status level, voltage stability, internal resistance uniformity and capacity retention rate, in order to determine whether the battery is in a normal health state.

8. The method for monitoring the trend of changes in the health status of a battery according to claim 1, characterized in that, If an abnormal pattern is detected in the health status feature mapping data in step S3, the abnormal battery features are marked and recorded as battery abnormal information, including: Extract voltage fluctuation amplitude, internal resistance change rate and capacity decay percentage from health status mapping data to form a set of detection feature values; For each feature value in the detection feature value set, a threshold judgment is performed to determine whether it exceeds the preset normal range; If any of the following characteristics—voltage fluctuation amplitude, internal resistance change rate, or capacity decay percentage—exceeds the preset normal range, it is marked as a potential abnormal characteristic. Neighborhood analysis is performed on potential anomalous features to check whether the feature values ​​of surrounding data points also show anomalies, in order to confirm the existence of anomalous patterns. If an abnormal pattern is confirmed to exist, the feature value corresponding to the abnormal pattern is marked as a battery abnormal feature and recorded as battery abnormal information.

9. The method for monitoring the trend of changes in the health status of a battery according to claim 1, characterized in that, Step S4, which predicts the long-term trend of battery health status based on health status feature mapping data, includes: Time series analysis is performed on health status feature mapping data to calculate the changing trends at multiple time points and obtain the time series analysis results. Based on the time series analysis results, determine the long-term trend of each feature value and calculate the feature change rate of the health status feature mapping data value over the past 10 time points. If the rate of change of the eigenvalue remains positive, it is marked as an upward trend in health status; If the rate of change of the eigenvalue remains negative, it is marked as a downward trend in health status; If the rate of change of the feature value fluctuates within a preset threshold range, it is marked as a stable trend of health status. By comprehensively analyzing the long-term trends of each characteristic value, the long-term trend of the overall health status of the battery can be predicted.

10. A battery health status change trend monitoring system, characterized in that, For performing the battery health status change trend monitoring method as described in claim 1, the battery health status change trend monitoring system includes: The battery data acquisition module is used to perform charge-discharge cycle tests on the battery to obtain dynamic fluctuation data of the terminal voltage; to perform AC impedance spectrum tests on the battery to obtain real-time change data of internal resistance; and to monitor the charge-discharge capacity decay rate data of the battery. The health status analysis module is used to perform multi-dimensional analysis on the dynamic fluctuation data of terminal voltage, the real-time change data of internal resistance, and the charge and discharge capacity decay rate data to generate battery characteristic data; it performs collaborative analysis on the battery characteristic data, and generates collaborative analysis results by analyzing the synchronicity of voltage fluctuation and internal resistance change, as well as the impact of capacity decay on voltage and internal resistance changes; it maps the collaborative analysis results to a preset battery health status model to form health status mapping data. The battery abnormal feature detection module is used to identify the health status features of the battery based on health status mapping data; if an abnormal pattern is detected in the health status feature mapping data, the abnormal battery feature is marked and recorded as battery abnormal information. The health status prediction module is used to predict the long-term trend of battery health status based on health status feature mapping data, and obtain health prediction results; if the health prediction results show that the battery health status will decline significantly in the future, a health status warning signal will be generated.

Images