A power bank power monitoring management method, system and electronic device thereof

Abstract

The application discloses a power bank electric quantity monitoring management method and system and electronic equipment thereof, and the method comprises the following steps: collecting impedance data under initial stable working conditions of the power bank, and constructing an impedance characteristic control chart; determining abnormal jump events and trend abnormal events in the use process; taking covariance matrix normalization as a density matrix, calculating spectral entropy, relative entropy and fidelity, representing the disorder degree of the correlation structure between each frequency point of the impedance, the information amount deviating from the initial state and the structure overlap degree; accumulating the time series of spectral entropy, relative entropy and fidelity with the measurement number, and grading outputting the remaining cycle number or time interval estimation. Through the minimum eigenvalue analysis of the spectral entropy evolution curvature and the linear mapping stability matrix of the adjacent time density matrix, the joint cross-validation based on the three sequence evolution trajectories of the spectral entropy, the relative entropy and the fidelity realizes the estimation of the remaining cycle number and the remaining service life of the battery, and provides a clear replacement timing reference for the user.

Description

Technical Field

[0001] This invention relates to the field of power monitoring technology, and in particular to a power bank power monitoring and management method, system and electronic device. Background Technology

[0002] Power banks are popular as portable energy storage devices. Accurate monitoring of the battery health status of power banks is crucial for safe use and extending their service life. Currently, the power monitoring of power banks mainly relies on threshold comparison methods of conventional electrical parameters such as voltage and current. There is a lack of effective technical means for refined assessment of battery health status and prediction of remaining life.

[0003] In existing technologies, there are schemes that apply statistical process control methods to battery state monitoring. These schemes establish control charts based on collected battery impedance data for anomaly detection. Other schemes assess health status by extracting geometric features of electrochemical impedance spectroscopy and combining them with Mahalanobis distance, and by inputting entropy features from battery power consumption sequence data into an assessment model for health status identification. However, the aforementioned Mahalanobis distance-based assessment methods only use it as a scalar distance metric, failing to utilize the complete correlation structure information contained in the covariance matrix. Furthermore, these schemes all rely on fixed, non-adaptive preset thresholds for anomaly detection, and cannot effectively predict the remaining cycle life and remaining service life of the battery based on the data's own evolutionary patterns, causing inconvenience to users. Therefore, this paper proposes a power bank power monitoring and management method, system, and electronic equipment to solve the above problems. Summary of the Invention

[0004] The purpose of this invention is to provide a method, system and electronic device for monitoring and managing the power consumption of a power bank, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for monitoring and managing the power bank's charge level, comprising the following operations: S1. Under the initial stable operating conditions of the power bank, at least one set of impedance data is collected, and the mean vector and standard deviation are calculated to construct an impedance characteristic control chart. The center line and control limits of the control chart under the initial state are stored as a fixed reference system. After each impedance measurement, a sliding window mechanism is used to recalculate the mean vector and standard deviation of the data in the window with the most recent set of measurement values. The center line and control limits of the control chart are updated accordingly, so that the control chart forms a local dynamic reference under the current operating conditions with the mean vector and standard deviation of the data in the current window. S2, during the use of the power bank, the center line and control limit in the fixed reference system stored in the initial state are used as absolute judgment thresholds. When a new sample point exceeds the fixed control limit, it is judged as an abnormal jump event. The current center line after the sliding window is updated is used as the trajectory tracking object. When the center lines of at least three consecutive windows are continuously located on the same side relative to the initial center line and the difference between the center lines of adjacent windows increases continuously and remains positive and monotonically undecreasing, it is judged as a trend abnormal event. S3. Construct a high-dimensional feature vector using the mean vector and standard deviation at each measurement time. Define the metric tensor of the current Mahalanobis space using the covariance matrix of the feature vectors within a fixed-width time window. Normalize the covariance matrix into a density matrix with a trace of 1. The feature spectrum of the density matrix characterizes the correlation weight distribution between impedance frequencies. Calculate the spectral entropy of the density matrix and the relative entropy and fidelity between the current density matrix and the initial density matrix. S4, the cumulative spectral entropy, relative entropy and fidelity time series with the number of measurements, are extrapolated to their respective critical thresholds according to the evolution rate of each series. Multiple extrapolation results are cross-validated and output in a graded manner the remaining cycle number or time interval estimate with confidence level labels.

[0006] Preferably, in step S3, the normalized density matrix is ​​subjected to spectral decomposition and complete matrix element reconstruction to obtain all diagonal and off-diagonal element descriptions. Then, the fidelity between the current density matrix and the density matrix under the initial stable condition in S1 is calculated. When the fidelity approaches 1, the current power bank structural state is consistent with the healthy state under the initial stable condition. When the fidelity approaches 0, the current power bank structural state has completely collapsed. The rate of change and acceleration of fidelity with the number of measurements are calculated simultaneously. The fidelity, its decay rate, and acceleration constitute the structural descriptor of the healthy state, and the fidelity decay spectrum is output.

[0007] Preferably, step S3 performs a matrix logarithm operation on the density matrix, multiplies the resulting matrix with the density matrix, extracts the trace of the product matrix, and uses the negative value of the trace as the output value of the relative entropy. The relative entropy value at the current moment is subtracted by twice the relative entropy value at the previous moment and then added to the relative entropy value two moments ago. The result is the spectral entropy evolution curvature. A curvature approaching zero indicates that the entropy value changes at a constant rate, indicating a stable aging state. A curvature turning from zero to positive and continuously increasing indicates that the entropy value accelerates, indicating an accelerated degeneration state. At the same time, the current density matrix and the initial density matrix are respectively subjected to matrix logarithm operations and subtracted. The resulting matrix is ​​multiplied with the current density matrix, and the trace of the product matrix is ​​extracted. This trace is used as the check channel value of the relative entropy. When the relative entropy exceeds a preset threshold determined based on the entropy value distribution under the initial healthy state, it is determined to be a health state deviation event. When the spectral entropy evolution curvature is continuously positive and exceeds a preset number of times, it is determined to be an accelerated degeneration event. The health state classification signal is output by combining the two.

[0008] Preferably, in step S4, after accumulating the relative entropy change sequence with the number of measurements, the battery aging process is mapped as the process of the density matrix eigenvalue distribution evolving from an initial centralized ordered state to a completely disordered critical state. The critical entropy value of the completely disordered critical state is obtained by taking the natural logarithm of the eigendimensional dimension of the density matrix. Then, the entropy evolution rate is obtained by subtracting the entropy value of the previous time from the current entropy value. Then, the sequence of entropy evolution rate is subjected to a first-order difference operation to obtain the evolution acceleration. A second-order polynomial prediction model of the entropy value changing with the number of measurements is constructed using the current evolution rate and the evolution acceleration. The number of measurements equal to the critical entropy value is solved. The difference between the current number of measurements and the solved number of measurements is used as the point estimate of the remaining lifespan. At the same time, the standard deviation is calculated based on the fluctuation amplitude of the historical entropy value around the prediction trajectory. The upper and lower boundaries of the confidence interval of the remaining lifespan are constructed using a preset multiple of the calculated standard deviation. When the lower boundary of the confidence interval is lower than the preset safety warning line, a graded replacement reminder is output.

[0009] Preferably, in step S4, while accumulating the time series of relative entropy changing with the number of measurements, the time series of fidelity changing with the number of measurements is also accumulated simultaneously. The critical moment when the fidelity reaches zero is calculated using the historical drift trajectory of the fidelity decay rate. When the second-order difference value of the fidelity is continuously positive and the absolute value of the first-order difference value exceeds a preset multiple of the historical average, it is determined that the battery has entered the end of its lifespan.

[0010] Preferably, the remaining lifetime predicted by the relative entropy evolution trajectory and the remaining lifetime predicted by the fidelity decay trajectory are jointly cross-validated. When the difference between the two predicted remaining lifetime values ​​is within the preset allowable deviation range, a high-confidence prediction result is output and the weighted average of the two is used as the final output. When the difference between the two predicted remaining lifetime values ​​exceeds the preset allowable deviation range, a low-confidence flag is output and an encrypted sampling process is triggered for verification.

[0011] Preferably, in step S3, at each measurement time, the linear mapping transformation relationship between the density matrix at the current time and the density matrix at the next time is extracted. Then, a set of orthonormal bases is selected, and each pair of elements in the orthonormal bases is linearly mapped. The result is then tensor-producted with the corresponding pair of base elements. All tensor-product results are summed to obtain a matrix that characterizes the stability of the linear mapping. The obtained matrix is ​​then decomposed into eigenvalues, and the minimum value among all eigenvalues ​​is extracted. In the healthy state, the minimum eigenvalue is positive. During the aging process, the minimum eigenvalue continuously decreases. When the minimum eigenvalue drops below a preset threshold, it is determined that the degradation has entered the irreversible stage, and the complete eigenspectrum of the matrix is ​​output.

[0012] Preferably, while accumulating the relative entropy time series, the time series of the minimum eigenvalue of the matrix characterizing the stability of the mapping is simultaneously accumulated with the number of measurements. The number of measurements reaching zero is calculated based on the historical drift rate of the minimum eigenvalue. When the minimum eigenvalue enters the negative range, a judgment that the degradation has crossed the critical stability boundary is output, and an estimate of the remaining lifetime is given based on the degree of its negative value, wherein the larger the negative value, the shorter the remaining time.

[0013] A power bank power monitoring and management system, used to execute the aforementioned power bank power monitoring and management method, includes: Impedance data acquisition module is used to collect the impedance and related electrical parameters of the power bank and form a time series of electrical characteristic parameters; The benchmark modeling module is used to establish a healthy benchmark model based on impedance data in the initial state, set control limits, and store benchmark parameters. The real-time monitoring module is used to identify abnormal jump events and abnormal trend events; The information processing module receives the anomaly markers or trend markers and their corresponding impedance data sequences output by the real-time monitoring module. After normalizing the covariance matrix of the eigenvectors within a fixed-width time window into a density matrix, it calculates the spectral entropy of the density matrix, the relative entropy and fidelity between the current density matrix and the initial density matrix, as well as the intrinsic spectral features of the linear mapping transformation between density matrices at adjacent time points. The combined output of the spectral entropy, relative entropy, fidelity attenuation features and intrinsic spectral features is used as the structural descriptor of the health state and output to the lifetime prediction module. The lifetime prediction module receives the time series of the structural descriptor output by the information processing module, as well as the time series of the accumulated spectral entropy, relative entropy, and fidelity with the number of measurements. It then extrapolates each sequence to its respective critical threshold according to the evolution rate of each sequence, cross-validates multiple extrapolation results, and outputs graded estimates of the remaining lifetime interval with confidence labels and replacement reminders. Simultaneously, based on the sign and magnitude of the real part of the minimum eigenvalue of the mapping transformation between adjacent time points, it outputs a graded determination of irreversible degradation and feeds the determination result back to the central control module. The central control module coordinates the data flow and timing of each module and adjusts the sampling parameters of the real-time monitoring module based on the feedback signal from the life prediction module.

[0014] An electronic device includes one or more processors and a memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the power bank power monitoring and management method.

[0015] The technical effects and advantages of this invention are as follows: 1) This invention normalizes the covariance matrix into a density matrix and calculates its spectral entropy, relative entropy and fidelity. It takes the overall correlation structure between various electrical characteristic parameters of the battery as the analysis object and provides a panoramic characterization of the battery health state from three dimensions: the degree of disorder in the correlation weight distribution, the amount of information deviating from the initial healthy state and the degree of structural overlap. This is different from the existing technology that only uses Mahalanobis distance as a scalar distance measurement tool. It makes full use of the complete correlation structure information contained in the covariance matrix. (2) This invention establishes an adaptive judgment mechanism based entirely on the evolution characteristics of the data itself by analyzing the minimum eigenvalue of the stability matrix of the linear mapping of the density matrix at adjacent time points and the evolution curvature of the spectral entropy. The evolution curvature of the spectral entropy is used to identify the trend of accelerated degradation, and the minimum eigenvalue changes from positive to negative to determine that the battery has crossed the critical point of irreversible degradation. At the same time, the prediction of the remaining cycle number and remaining service life of the battery is realized based on the joint cross-validation of the evolution trajectory of the three sequences of spectral entropy, relative entropy and fidelity. This overcomes the defect of the existing technology that relies on a fixed preset threshold to only make anomaly judgments but cannot predict the remaining life. It provides users with a clear reference for replacement time and improves the convenience and safety of using power banks. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the operation of the power monitoring and management method of the present invention.

[0017] Figure 2 This is a block diagram of the power monitoring and management system of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1: The present invention provides as follows Figure 1 The method for monitoring and managing the power bank's battery level, as shown, includes the following operations: S1, Under the initial stable operating conditions of the power bank, collect at least one set of impedance data, and construct an impedance characteristic control chart after calculating the mean vector and standard deviation. The mean vector is calculated as follows: (1); in, This represents the average of at least a set of impedance eigenvectors collected, where each component corresponds to the arithmetic mean of impedance characteristics in each dimension, and serves as the center point of the health baseline state; n aThis indicates the number of impedance data sets collected under steady-state operating conditions; This represents the multidimensional impedance data obtained from the i-th measurement under steady-state conditions. The standard deviation is calculated as follows: (2); Where, n b This represents the total number of impedance eigenvectors involved in the calculation; The center line and control limits of the control chart in the initial state are stored as a fixed reference system. After each impedance measurement, a sliding window mechanism is used to recalculate the mean vector and standard deviation of the data within the window using the most recent set of measurements. Based on this, the center line and control limits of the control chart are updated, so that the control chart forms a local dynamic reference under the current operating conditions using the mean vector and standard deviation of the data within the current window. When constructing the control chart, the center line is set at the mean, the upper control limit is set at the mean plus three standard deviations, and the lower control limit is set at the mean minus three standard deviations, forming an impedance characteristic control chart. The center line being at the mean reflects the central trend of the power bank being in a controlled state under stable operating conditions, serving as the benchmark point for subsequent calculation of control limits and as a reference system for judging trend anomalies. After each impedance measurement, the newly measured impedance value is plotted on the impedance characteristic control chart as a sample point. A sliding window mechanism is used to recalculate the mean vector and standard deviation of the data within the window using the most recent set of measurements. Based on this, the center line and control limits of the control chart are updated, so that the control chart forms a local dynamic reference under the current operating conditions using the mean vector and standard deviation of the data within the current window. It is important to clarify that this local dynamic reference is only used to assess the short-term dispersion of impedance data within the current adjacent time period, serving as a background reference in trend determination, and is not used to replace the centerline and control limits in the initial state. Once set, the centerline and control limits in the initial state are permanently stored as fixed reference systems and remain unchanged throughout the power bank's entire lifespan. The sliding window's function is to track the slow drift trajectory of the battery impedance characteristics, amplifying long-term aging signals through the differential sequence between the centerlines of adjacent windows, rather than moving the judgment threshold in real time. Through hierarchical management of the initial fixed reference system and the sliding local dynamic reference, abnormal change judgment can focus on absolute deviations exceeding the tolerance of the initial healthy state, and trend anomaly judgment can focus on the cumulative aging degree relative to the initial baseline, avoiding missed detections due to the benchmark drifting synchronously with operating conditions.

[0020] S2, during the use of the power bank, the center line and control limit in the fixed reference system stored in the initial state are used as absolute judgment thresholds. When a new sample point exceeds the fixed control limit, it is judged as an abnormal jump event. The current center line after the sliding window is updated is used as the trajectory tracking object. When the center lines of at least three consecutive windows are continuously located on the same side relative to the initial center line and the difference between the center lines of adjacent windows increases continuously and remains positive and monotonically undecreasing, it is judged as a trend abnormal event. It should be noted that during the use of the power bank, two independent judgments are executed simultaneously after each impedance measurement. The first judgment uses the center line and upper and lower control limits in the fixed reference frame stored in the initial state as absolute thresholds. When a new sample point exceeds the fixed upper or lower control limit, it is judged as an abnormal jump event. Abnormal jump events reflect sudden and drastic changes in the internal state of the battery, including electrode breakage or poor contact, sudden decomposition or gas production of the electrolyte, or drastic transient changes in the external load. The second judgment uses the current center line after the sliding window is updated as the trajectory tracking object. When the center lines of at least three consecutive windows are continuously on the same side relative to the initial center line, and the difference between the center lines of adjacent windows remains continuously positive and monotonically undecreasing, it is judged as a trend anomaly event. Trend anomalies reflect the long-term cumulative slow aging of the battery, including the continuous increase in battery internal resistance, capacity decay caused by irreversible loss of active materials, and impedance rise caused by slow electrolyte consumption.

[0021] S3: Construct a high-dimensional feature vector using the mean vector and standard deviation at each measurement time. Define the metric tensor of the current Mahalanobis space using the covariance matrix of the feature vectors within a fixed-width time window. Normalize the covariance matrix into a density matrix with a trace of 1. The eigenspectrum of the density matrix characterizes the correlation weight distribution between impedance frequencies. Calculate the spectral entropy of the density matrix, as well as the relative entropy and fidelity between the current and initial density matrices. When calculating the relative entropy of the density matrix, perform a matrix logarithm operation on the density matrix, multiply the resulting matrix by the density matrix, and extract the trace of the product matrix, taking its negative value. If spectral decomposition is performed on the density matrix, the eigenvalues ​​of the density matrix are... Treating it as a probability distribution, the calculation is expressed as: (3); in, The density matrix is ​​obtained by normalizing the covariance matrix; Represents the density matrix The i-th eigenvalue, through the... Eigenvalue decomposition yields Fidelity is used to measure the degree of structural overlap between the current density matrix and the initial density matrix, and is calculated as follows: (4); in, This represents the initial density matrix.

[0022] Specifically, the normalized density matrix is ​​subjected to spectral decomposition and complete matrix element reconstruction to obtain descriptions of all diagonal and off-diagonal elements. Then, the fidelity between the current density matrix and the density matrix under the initial stable operating condition S1 is calculated. When the fidelity approaches 1, the current structural state of the power bank is consistent with the healthy state under the initial stable operating condition; when the fidelity approaches 0, the current structural state of the power bank has completely collapsed. The rate of change and acceleration of fidelity with the number of measurements are calculated simultaneously. The fidelity, its decay rate, and acceleration constitute the structural descriptor of the healthy state, and the fidelity decay spectrum is output.

[0023] In this scheme, the physical essence of the density matrix is: the correlation weight distribution matrix obtained by normalizing the covariance matrix of impedance data at each measurement time, whose matrix elements directly correspond to the coupling correlation between different frequency points of the impedance spectrum; spectral entropy is used to quantify the disorder of the current correlation structure; relative entropy is used to measure the information deviation between the current correlation structure and the initial healthy state; fidelity is used to characterize the degree of overlap between the two structures. All the above computational objects are derived from the real symmetric positive semi-definite matrix composed of measured impedance data. The technical purpose is to fully utilize the complete correlation structure information contained in the covariance matrix and avoid the information loss caused by traditional methods that only use scalar distance metrics.

[0024] Perform a matrix logarithm operation on the density matrix, multiply the resulting matrix by the density matrix, and extract the trace of the product matrix. Use the negative value of the trace as the output value of the relative entropy. Subtract twice the relative entropy value of the previous time step from the current relative entropy value, and add the relative entropy value from two time steps ago. The result is the spectral entropy evolution curvature. A curvature approaching zero indicates a uniform change in entropy value, representing a stable aging state. A curvature turning positive from zero and continuously increasing indicates an accelerated increase in entropy value, representing an accelerated degradation state. Simultaneously, perform a matrix logarithm operation on the current density matrix and subtract it. Multiply the resulting matrix by the current density matrix and extract the trace of the product matrix. Use this trace as the check channel value of the relative entropy. When the relative entropy exceeds a preset threshold determined based on the entropy value distribution under the initial healthy state, it is judged as a health state deviation event. When the spectral entropy evolution curvature is continuously positive and exceeds a preset number of times, it is judged as an accelerated degradation event. Combine the two to output a health state classification signal. The specific classification content is as follows: Level 1 signal: When the relative entropy does not exceed the preset threshold and the spectral entropy evolution curvature does not remain positive for more than the preset number of times, a Level 1 signal is output, indicating that the current correlation structure has not deviated significantly and there are no signs of accelerated degradation. The power bank is in a normal and uniform aging stage. Under the Level 1 signal, the normal monitoring frequency is maintained and the lifespan prediction is performed according to the normal cycle without triggering any additional operations. The secondary signal is output when the relative entropy exceeds a preset threshold but the spectral entropy evolution curvature does not remain positive for a preset number of times, or when the relative entropy does not exceed the preset threshold but the spectral entropy evolution curvature remains positive for a preset number of times. The secondary signal includes two different triggering scenarios: one where deviation has occurred but the trend has not yet accelerated, and the other where deviation has not yet occurred but the accelerating trend has been stably established. In the first scenario, it indicates that the power bank has entered a significant deviation state, but has not yet shown an accelerating deterioration trend and is still in a uniform deviation phase. In the second scenario, it indicates that the power bank is in a sub-health warning zone rapidly approaching the threshold boundary. Both scenarios suggest the need to increase the sampling period to obtain more dynamic data, triggering subsequent operations for detailed lifespan prediction. The Level 3 signal is output when the relative entropy exceeds a preset threshold and the spectral entropy evolution curvature remains positive for more than a preset number of times. The Level 3 signal indicates that two judgments simultaneously confirm that the degradation has entered an irreversible acceleration stage. The first judgment confirms that a significant deviation has occurred, and the second judgment confirms that the deviation is accelerating. Since the two judgment events are based on different data sources, the simultaneous triggering of both increases the confidence of the alarm and avoids false alarms from a single indicator under special operating conditions. The Level 3 signal triggers an emergency response, immediately triggering subsequent steps to perform an emergency estimation of the remaining lifespan and adopting a conservative estimate. At the same time, it triggers the S2 sliding window width to be halved to improve the sensitivity of subsequent data and outputs a reminder to replace the power bank as soon as possible.

[0025] S4, the time series of cumulative spectral entropy, relative entropy and fidelity with the number of measurements, are extrapolated to their respective critical thresholds according to the evolution rate of each series. Multiple extrapolation results are cross-validated and output in a hierarchical manner the remaining cycle number or time interval estimate with confidence level labels. After accumulating the sequence of relative entropy changes with the number of measurements, the battery aging process is mapped as the evolution of the density matrix eigenvalue distribution from an initial centralized ordered state to a completely disordered critical state. The critical entropy value of the completely disordered critical state is obtained by taking the natural logarithm of the eigendimensional dimension of the density matrix. Then, the entropy evolution rate is obtained by subtracting the entropy value of the previous time from the current entropy value. The sequence of entropy evolution rate is then subjected to a first-order difference operation to obtain the evolution acceleration. A second-order polynomial prediction model of entropy value changes with the number of measurements is constructed using the current evolution rate and the evolution acceleration. The number of measurements equal to the critical entropy value is calculated. The difference between the current number of measurements and the calculated number of measurements is used as the point estimate of the remaining lifespan. At the same time, the standard deviation is calculated based on the fluctuation amplitude of historical entropy values ​​around the predicted trajectory. The upper and lower boundaries of the confidence interval of the remaining lifespan are constructed using a preset multiple of the calculated standard deviation. When the lower boundary of the confidence interval is lower than the preset safety warning line, a graded replacement reminder is output. While accumulating the time series of relative entropy changes with the number of measurements, the time series of fidelity changes with the number of measurements is also accumulated simultaneously. The critical moment when the fidelity reaches zero is calculated using the historical drift trajectory of the fidelity decay rate. When the second-order difference of the fidelity is consistently positive and the absolute value of the first-order difference exceeds a preset multiple of the historical average, the battery is determined to have entered the end of its lifespan. The remaining lifespan predicted by the spectral entropy evolution trajectory, the remaining lifespan predicted by the relative entropy evolution trajectory, and the remaining lifespan predicted by the fidelity decay trajectory are jointly cross-validated. The relative deviations between each pair of the three predicted values ​​are calculated. When the relative deviations between each pair of the three predicted values ​​are all within the first preset allowable deviation range, a high-confidence prediction result is output, and the weighted average of the three is used as the final output. When the relative deviation between any pair exceeds the preset allowable deviation range, a low-confidence flag is output and an encrypted sampling process is triggered for verification. When the difference between the predicted remaining lifespan values ​​exceeds the preset allowable deviation range, a low-confidence flag is output and an encrypted sampling process is triggered for verification.

[0026] Furthermore, at each measurement time, the linear mapping transformation relationship between the density matrix at the current time and the density matrix at the next time is extracted. This mapping satisfies the constraints of complete positive definiteness and trace preservation. Then, a set of orthonormal bases is selected, and a linear mapping is performed on each pair of elements in the orthonormal bases. The result is then tensor-producted with the corresponding pair of basis elements. All tensor-product results are summed to obtain a matrix that characterizes the stability of the linear mapping. The obtained matrix is ​​then decomposed into eigenvalues, and the minimum value among all eigenvalues ​​is extracted. In the healthy state, the minimum eigenvalue is positive. During the aging process, the minimum eigenvalue continuously decreases. When the minimum eigenvalue drops below a preset threshold, it is determined that the degradation has entered the irreversible stage, and the complete eigenspectrum of the matrix is ​​output.

[0027] While accumulating the relative entropy time series, the time series of the minimum eigenvalue of the matrix characterizing the stability of the mapping is simultaneously accumulated with the number of measurements. The number of measurements reaching zero is calculated based on the historical drift rate of the minimum eigenvalue. When the minimum eigenvalue enters the negative range, the determination that the degradation has crossed the critical stability boundary is output, and the remaining lifetime is estimated based on the degree of its negative value. The larger the negative value, the shorter the remaining time.

[0028] Example 2: The present invention provides as follows Figure 2 The power bank power monitoring and management system shown is used to execute the power bank power monitoring and management method of Embodiment 1, including an impedance data acquisition module, a benchmark modeling module, a real-time monitoring module, an information processing module, a life prediction module, and a central control module; The impedance data acquisition module is used to collect the impedance and related electrical parameters of the power bank at a preset sampling frequency during the initial state and use of the power bank. The related electrical parameters include the battery terminal voltage, charging and discharging current and cell surface temperature, forming a time series of electrical characteristic parameters, and outputting them to the benchmark modeling module and the real-time monitoring module.

[0029] The benchmark modeling module receives stable operating condition impedance data output by the impedance data acquisition module in the initial state, calculates the mean vector and standard deviation, constructs an initial impedance characteristic control chart, and stores the center line and control limits in the initial state as a fixed reference system. At the same time, a sliding window mechanism is established to continuously calculate the local mean vector and local standard deviation to form a local dynamic reference. The fixed reference system and the local dynamic reference serve as the basis for subsequent abnormal jump event judgment and trend abnormal event judgment, respectively.

[0030] The real-time monitoring module receives new sample points from the impedance data acquisition module in real time during the use of the power bank. Using the centerline and control limits of a fixed reference system stored in the benchmark modeling module as absolute judgment thresholds, when a new sample point exceeds these fixed control limits, it is judged as an abnormal jump event and an anomaly marker is output. Simultaneously, using the current centerline updated by the sliding window in the benchmark modeling module as the trajectory tracking object, when the centerlines of at least three consecutive windows remain on the same side relative to the initial centerline, and the increasing difference between the centerlines of adjacent windows remains continuously positive and monotonically undecreasing, it is judged as a trend anomaly event and a trend marker is output. The anomaly marker and trend marker, along with the corresponding impedance data sequence, are output to the information processing module.

[0031] The information processing module receives anomaly or trend markers and their corresponding impedance data sequences output by the real-time monitoring module. After normalizing the density matrix to the covariance matrix of the eigenvectors within a fixed-width time window, it calculates the spectral entropy, relative entropy and fidelity between the current and initial density matrices, and the intrinsic spectral features of the linear mapping transformation between density matrices at adjacent time points. The combined output of spectral entropy, relative entropy, fidelity decay features, and intrinsic spectral features is then used as a structural descriptor of the health state and output to the lifetime prediction module. The lifetime prediction module receives the time series of the structural descriptor output by the information processing module, as well as the time series of accumulated spectral entropy, relative entropy, and fidelity with the number of measurements. It then extrapolates each sequence to its respective critical threshold according to the evolution rate of each sequence, cross-validates multiple sets of extrapolation results, and outputs graded estimates of the remaining lifetime interval with confidence labels and replacement reminders. At the same time, based on the sign and magnitude of the real part of the minimum eigenvalue of the mapping transformation between adjacent time points, it outputs the graded judgment of irreversible degradation and feeds the judgment result back to the central control module.

[0032] The central control module coordinates the data flow and timing of each module, and issues encrypted sampling commands to the impedance data acquisition module to adjust the sampling frequency based on the confidence level flags fed back by the lifetime prediction module, and issues sliding window width adjustment commands to the real-time monitoring module. Redundant and repetitive intermediate process descriptions have been removed to maintain concise and accurate language.

[0033] Example 3: The present invention provides an electronic device, including one or more processors and a memory, wherein the memory is used to store one or more programs; wherein, when one or more programs are executed by one or more processors, the one or more processors implement the power bank power monitoring and management method of Example 1; It also includes a battery cell, an impedance measurement circuit, a voltage and current sampling circuit, a charge and discharge control circuit, a wireless communication unit, and a display unit; The battery unit is used to store electrical energy and power external devices. The battery unit is selected from lithium-ion batteries or lithium polymer batteries. The rated capacity is determined according to the product specifications, including but not limited to 5000mAh, 10000mAh or 20000mAh. The positive and negative terminals of the battery unit are connected to the charging and discharging control circuit respectively.

[0034] The impedance measurement circuit is connected to the battery cell and the processor. It is used to collect the impedance data of the battery cell under the control of the processor and output it to the processor, providing a data basis for constructing high-dimensional feature vectors in step S3.

[0035] The voltage and current sampling circuit is connected to the battery cell and the processor. It is used to collect the terminal voltage and charging / discharging current of the battery cell and output them to the processor. By detecting the voltage difference across the resistor and combining it with the resistance value of the detection resistor, the current value flowing through the battery cell can be calculated.

[0036] The charge / discharge control circuit is connected to the battery cell, external charging interface, and external discharging interface, and is also connected to the processor to perform charge / discharge management operations according to the processor's control instructions.

[0037] The wireless communication unit is connected to the processor and is used to send the remaining cycle number or time interval estimate, grade replacement reminder and confidence level label information output by the processor to an external terminal device.

[0038] The display unit is connected to the processor and is used to display the remaining cycle number or time interval estimate, grade replacement reminder, and confidence level marking information output by the processor.

[0039] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for monitoring and managing the power level of a power bank, characterized in that, Includes the following operations: S1. Under the initial stable operating conditions of the power bank, at least one set of impedance data is collected, and the mean vector and standard deviation are calculated to construct an impedance characteristic control chart. The center line and control limits of the control chart under the initial state are stored as a fixed reference system. After each impedance measurement, a sliding window mechanism is used to recalculate the mean vector and standard deviation of the data in the window with the most recent set of measurement values. The center line and control limits of the control chart are updated accordingly, so that the control chart forms a local dynamic reference under the current operating conditions with the mean vector and standard deviation of the data in the current window. S2, during the use of the power bank, the center line and control limit in the fixed reference system stored in the initial state are used as absolute judgment thresholds. When a new sample point exceeds the fixed control limit, it is judged as an abnormal jump event. The current center line after the sliding window is updated is used as the trajectory tracking object. When the center lines of at least three consecutive windows are continuously located on the same side relative to the initial center line and the difference between the center lines of adjacent windows increases continuously and remains positive and monotonically undecreasing, it is judged as a trend abnormal event. S3. Construct a high-dimensional feature vector using the mean vector and standard deviation at each measurement time. Define the metric tensor of the current Mahalanobis space using the covariance matrix of the feature vectors within a fixed-width time window. Normalize the covariance matrix into a density matrix with a trace of 1. The feature spectrum of the density matrix characterizes the correlation weight distribution between impedance frequencies. Calculate the spectral entropy of the density matrix and the relative entropy and fidelity between the current density matrix and the initial density matrix. S4, the cumulative spectral entropy, relative entropy and fidelity time series with the number of measurements, are extrapolated to their respective critical thresholds according to the evolution rate of each series. Multiple extrapolation results are cross-validated and output in a graded manner the remaining cycle number or time interval estimate with confidence level labels.

2. The power bank power monitoring and management method according to claim 1, characterized in that, In step S3, the normalized density matrix is ​​subjected to spectral decomposition and complete matrix element reconstruction to obtain all diagonal and off-diagonal element descriptions. Then, the fidelity between the current density matrix and the density matrix under the initial stable condition in S1 is calculated. When the fidelity approaches 1, the current power bank structural state is consistent with the healthy state under the initial stable condition; when the fidelity approaches 0, the current power bank structural state has completely collapsed. The rate of change and acceleration of fidelity with the number of measurements are calculated simultaneously. The fidelity, its decay rate, and acceleration constitute a structural descriptor of the healthy state, and the fidelity decay spectrum is output.

3. The power bank power monitoring and management method according to claim 2, characterized in that, S3 performs a matrix logarithm operation on the density matrix, multiplies the resulting matrix with the density matrix, and extracts the trace of the product matrix. The negative value of the trace is used as the output value of the relative entropy. The relative entropy value at the current moment is subtracted by twice the relative entropy value at the previous moment and then added to the relative entropy value two moments ago. The result is the spectral entropy evolution curvature. A curvature approaching zero indicates that the entropy value changes at a constant rate, indicating a stable aging state. A curvature turning from zero to positive and continuously increasing indicates that the entropy value accelerates, indicating an accelerated degeneration state. At the same time, the current density matrix and the initial density matrix are respectively subjected to a matrix logarithm operation and then subtracted. The resulting matrix is ​​multiplied with the current density matrix, and the trace of the product matrix is ​​extracted. This trace is used as the check channel value of the relative entropy. When the relative entropy exceeds a preset threshold determined based on the entropy value distribution under the initial healthy state, it is judged as a health state deviation event. When the spectral entropy evolution curvature is continuously positive and exceeds a preset number of times, it is judged as an accelerated degeneration event. The health state classification signal is output by combining the two.

4. The power bank power monitoring and management method according to claim 1, characterized in that, In step S4, after accumulating the relative entropy change sequence with the number of measurements, the battery aging process is mapped as the evolution of the density matrix eigenvalue distribution from an initial centralized ordered state to a completely disordered critical state. The critical entropy value of the completely disordered critical state is obtained by taking the natural logarithm of the eigendimensional dimension of the density matrix. Then, the entropy evolution rate is obtained by subtracting the entropy value of the previous time from the current entropy value. The sequence of entropy evolution rate is then subjected to a first-order difference operation to obtain the evolution acceleration. A second-order polynomial prediction model of the entropy value changing with the number of measurements is constructed using the current evolution rate and the evolution acceleration. The number of measurements equal to the critical entropy value is calculated. The difference between the current number of measurements and the calculated number of measurements is used as the point estimate of the remaining lifespan. At the same time, the standard deviation is calculated based on the fluctuation amplitude of the historical entropy value around the predicted trajectory. The upper and lower boundaries of the confidence interval of the remaining lifespan are constructed using a preset multiple of the calculated standard deviation. When the lower boundary of the confidence interval is lower than the preset safety warning line, a graded replacement reminder is output.

5. The power bank power monitoring and management method according to claim 4, characterized in that, In step S4, while accumulating the time series of relative entropy changes with the number of measurements, the time series of fidelity changes with the number of measurements is also accumulated simultaneously. The critical moment when the fidelity returns to zero is calculated using the historical drift trajectory of the fidelity decay rate. When the second-order difference value of the fidelity is continuously positive and the absolute value of the first-order difference value exceeds a preset multiple of the historical average, it is determined that the battery has entered the end of its lifespan.

6. The power bank power monitoring and management method according to claim 5, characterized in that, The remaining lifetime predicted by the relative entropy evolution trajectory and the remaining lifetime predicted by the fidelity decay trajectory are jointly cross-validated. When the difference between the two predicted remaining lifetime values ​​is within the preset allowable deviation range, a high-confidence prediction result is output and the weighted average of the two is used as the final output. When the difference between the two predicted remaining lifetime values ​​exceeds the preset allowable deviation range, a low-confidence flag is output and an encrypted sampling process is triggered for verification.

7. The power bank power monitoring and management method according to claim 3, characterized in that, In step S3, at each measurement time, the linear mapping transformation relationship between the density matrix at the current time and the density matrix at the next time is extracted. Then, a set of orthonormal bases is selected, and each pair of elements in the orthonormal bases is linearly mapped. The result is then tensor-producted with the corresponding pair of base elements. All tensor-product results are summed to obtain a matrix that characterizes the stability of the linear mapping. The obtained matrix is ​​then decomposed into eigenvalues, and the minimum value among all eigenvalues ​​is extracted. In the healthy state, the minimum eigenvalue is positive. During the aging process, the minimum eigenvalue continuously decreases. When the minimum eigenvalue drops below a preset threshold, it is determined that the degradation has entered the irreversible stage, and the complete eigenspectrum of the matrix is ​​output.

8. The power bank power monitoring and management method according to claim 7, characterized in that, While accumulating the relative entropy time series, the time series of the minimum eigenvalue of the matrix characterizing the stability of the mapping is simultaneously accumulated with the number of measurements. The number of measurements reaching zero is calculated based on the historical drift rate of the minimum eigenvalue. When the minimum eigenvalue enters the negative range, the determination that the degradation has crossed the critical stability boundary is output, and the remaining lifetime is estimated based on the degree of its negative value. The larger the negative value, the shorter the remaining time.

9. A power bank power monitoring and management system, used to execute the power bank power monitoring and management method according to any one of claims 1-8, characterized in that, include: Impedance data acquisition module is used to collect the impedance and related electrical parameters of the power bank and form a time series of electrical characteristic parameters; The benchmark modeling module is used to establish a healthy benchmark model based on impedance data in the initial state, set control limits, and store benchmark parameters. The real-time monitoring module is used to identify abnormal jump events and abnormal trend events; The information processing module receives the anomaly markers or trend markers and their corresponding impedance data sequences output by the real-time monitoring module. After normalizing the covariance matrix of the eigenvectors within a fixed-width time window into a density matrix, it calculates the spectral entropy of the density matrix, the relative entropy and fidelity between the current density matrix and the initial density matrix, as well as the intrinsic spectral features of the linear mapping transformation between density matrices at adjacent time points. The combined output of the spectral entropy, relative entropy, fidelity attenuation features and intrinsic spectral features is used as the structural descriptor of the health state and output to the lifetime prediction module. The lifetime prediction module receives the time series of the structural descriptor output by the information processing module, as well as the time series of the accumulated spectral entropy, relative entropy, and fidelity with the number of measurements. It then extrapolates each sequence to its respective critical threshold according to the evolution rate of each sequence, cross-validates multiple extrapolation results, and outputs graded estimates of the remaining lifetime interval with confidence labels and replacement reminders. Simultaneously, based on the sign and magnitude of the real part of the minimum eigenvalue of the mapping transformation between adjacent time points, it outputs a graded determination of irreversible degradation and feeds the determination result back to the central control module. The central control module coordinates the data flow and timing of each module and adjusts the sampling parameters of the real-time monitoring module based on the feedback signal from the life prediction module.

10. An electronic device, characterized in that, It includes one or more processors and a memory, the memory being used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the power bank power monitoring and management method as described in any one of claims 1 to 8.

Images