Condition monitoring and prediction method of intelligent battery pack and related equipment

Abstract

The present invention relates to the field of battery management technology, and discloses a state monitoring and prediction method for an intelligent battery pack and related equipment. The method includes: obtaining the original battery pack data and historical operation data of multiple preset battery pack parameters in the target intelligent battery pack, and performing signal filtering and spectrum decomposition on the original battery pack data; performing abnormal data detection, abnormal correction, multi-modal feature decomposition and phase space mapping of the battery pack state on the initial battery pack feature data; extracting the corresponding state of charge dynamic characteristics in the battery pack state evolution characteristics, and based on the historical operation data, performing multi-layer mapping and threshold classification on the state evolution characteristics to obtain the state monitoring results; obtaining the state transfer matrix through state parameter coupling mapping and conversion probability calculation, and performing state evolution path calculation to output the prediction results. The present application realizes the correlation analysis between electrochemical parameters, and enhances the reliable evaluation and accurate prediction of the battery pack state.

Description

Technical Field

[0001] The present invention relates to the technical field of battery management, and in particular to a state monitoring and prediction method for an intelligent battery pack and related equipment. Background Art

[0002] In the field of smart battery pack technology, condition monitoring and life prediction are important links in product development and quality control, while real-time monitoring and condition evaluation of electrochemical parameters are the most critical and challenging stages in battery management. Accurately evaluating the operating status of battery packs is crucial for manufacturers, R&D institutions, and testing centers, as it directly affects product reliability assessment, performance optimization, and overall service life. Therefore, an effective condition monitoring method to evaluate these electrochemical parameters is essential to ensure the operating stability and safety of battery packs.

[0003] Nowadays, digital signal processing technology is used to analyze sampling data, establish a state assessment model to monitor the operating status, or try to apply artificial intelligence technology to the state prediction process to better evaluate battery performance and degradation trends. However, these methods still have challenges in integrating multidimensional electrochemical parameters, processing nonlinear characteristics, and adapting to dynamic loads. In addition, these monitoring methods often ignore some unique characteristics of battery packs, such as temperature gradients, internal resistance changes, and charge and discharge efficiency, which may have a significant impact on state assessment and life prediction. That is, the existing battery pack state monitoring methods ignore multidimensional electrochemical parameters and some unique characteristics of battery packs, resulting in low accuracy of the final state assessment and prediction results. Summary of the invention

[0004] The main purpose of the present invention is to solve the problem that the existing battery pack state monitoring method ignores the multi-dimensional electrochemical parameters and some unique characteristics of the battery pack, resulting in low accuracy of the final state evaluation and prediction results.

[0005] The first aspect of the present invention provides a state monitoring and prediction method for an intelligent battery pack, the state monitoring and prediction method for an intelligent battery pack comprising: obtaining original battery pack data and historical operation data of a plurality of preset battery pack parameters in a target intelligent battery pack, and performing signal filtering and spectrum decomposition on the original battery pack data to obtain initial battery pack feature data; performing abnormal data detection and abnormal correction on the initial battery pack feature data to obtain a standard battery pack feature sequence of the target intelligent battery pack, and performing multi-modal feature decomposition and phase space mapping of the battery pack state on the standard battery pack feature sequence to obtain battery pack state evolution characteristics; extracting The state of charge dynamic characteristics corresponding to the battery pack state evolution characteristics are obtained, and based on the historical operation data, the state of charge dynamic characteristics are subjected to multi-layer feature mapping and state threshold classification to obtain state monitoring results of multiple state monitoring parameters; the state monitoring results are subjected to coupling mapping of multiple state monitoring parameters and conversion probability calculation of the battery pack state to obtain a state space transfer feature matrix, and based on the historical operation data and preset battery pack operating parameters, the state space transfer feature matrix is ​​subjected to state evolution path calculation and associated confidence calculation of the corresponding battery pack life characteristics to generate a state prediction result of the target intelligent battery pack.

[0006] Optionally, in a first implementation manner of the first aspect of the present invention, the original battery pack data includes battery pack voltage data, battery pack current data, and battery pack temperature data, and the signal filtering and spectral decomposition of the original battery pack data to obtain initial battery pack characteristic data includes: performing multi-layer recursive decomposition of discrete wavelets on the battery pack voltage data to obtain multi-band voltage characteristic group data, and performing gain calculation on the voltage characteristic group data to obtain battery pack voltage attenuation characteristics; performing time-frequency domain calculation on the battery pack current data to obtain current characteristic distribution, and performing multi-scale decomposition on the current characteristic distribution to obtain battery pack current fluctuation characteristics; performing two-dimensional spatial gradient calculation on the battery pack temperature data to obtain temperature distribution matrix, and calculate the uniformity index of the temperature distribution matrix to obtain the temperature field characteristics of the battery pack; extract the impedance characteristics of the battery pack voltage attenuation characteristics and the battery pack current fluctuation characteristics to generate an internal resistance change curve, and perform charge and discharge cross calculation on the battery pack voltage attenuation characteristics and the battery pack current fluctuation characteristics to obtain the battery pack charge and discharge response characteristics, and perform impedance parameter matching on the battery pack charge and discharge response characteristics and the internal resistance change curve to obtain the battery pack impedance dynamic characteristics; perform thermoelectric coupling mapping on the battery pack impedance dynamic characteristics and the battery pack temperature field characteristics to obtain multi-physical field coupling characteristics, and construct a time series feature matrix for the multi-physical field coupling characteristics to generate initial battery pack feature data.

[0007] Optionally, in a second implementation of the first aspect of the present invention, the abnormal data detection and abnormal correction are performed on the initial battery pack characteristic data to obtain a standard battery pack characteristic sequence of the target intelligent battery pack, including: dividing the initial battery pack characteristic data into time windows to obtain segmented time series data, and performing extreme value deviation detection and interquartile range test on the corresponding voltage and current time series data segments in the segmented time series data to obtain abnormal point marking data, and performing thermal distribution test on the corresponding battery pack temperature field data in the segmented time series data to obtain temperature abnormal area marking; performing cubic segmented spline interpolation on the abnormal point marking data to obtain a continuous characteristic curve, and performing thermal distribution test on the temperature abnormal area marking. A local weighted smoothing calculation is performed on the domain to obtain the corrected temperature field distribution data, and residual detection and deviation test are performed on the continuous characteristic curve to obtain the voltage and current quality index, and a gradient consistency test is performed on the temperature field distribution data to obtain the temperature field quality index; multi-dimensional threshold features are extracted from the voltage and current quality index and the temperature field quality index to obtain the battery pack working state quality score, and based on the battery pack working state quality score, the voltage and current quality index and the temperature field quality index are graded and screened and data segmented to obtain the identification collection data segment; the identification collection data segment is time-series reconstructed to obtain the standard battery pack feature sequence of the target intelligent battery pack.

[0008] Optionally, in a third implementation of the first aspect of the present invention, the multi-modal feature decomposition of the standard battery pack feature sequence and the phase space mapping of the battery pack state are performed to obtain the battery pack state evolution characteristics, including: dividing the corresponding voltage and current feature data in the standard battery pack feature sequence into charge and discharge cycles to obtain the initial battery pack working mode, and classifying the corresponding temperature and impedance feature data in the initial battery pack working mode into multiple charge and discharge rates to obtain multi-dimensional battery pack modal components; extracting the time-frequency features of each of the battery pack modal components to obtain the time-frequency response characteristics of the target intelligent battery pack. The energy distribution of the time-frequency response characteristics is calculated to obtain a battery pack operating condition energy spectrum; the charge and discharge state characteristics of the corresponding voltage and current characteristic data in the battery pack operating condition energy spectrum are extracted to obtain a charge and discharge state characteristic; the corresponding temperature and impedance characteristic data in the battery pack operating condition energy spectrum are mapped to the working state to obtain a temperature impedance state characteristic; the charge and discharge state characteristics and the temperature impedance state characteristics are feature combined to obtain a battery pack multidimensional state sequence; the battery pack multidimensional state sequence is subjected to state space reconstruction of the battery pack state and evolution of the battery pack state to obtain a battery pack state evolution characteristic.

[0009] Optionally, in a fourth implementation method of the first aspect of the present invention, the state space reconstruction of the battery pack state and the evolution of the battery pack state of the multidimensional state sequence of the battery pack are performed to obtain the battery pack state evolution characteristics, including: performing working state mapping of the charge and discharge cycle of the multidimensional state sequence of the battery pack to obtain the battery pack working trajectory characteristics, and performing state space mapping of the corresponding voltage and current feature data in the battery pack working trajectory characteristics to obtain a voltage and current trajectory matrix, and performing state space mapping of the corresponding temperature and impedance feature data in the battery pack working trajectory characteristics to obtain a temperature impedance trajectory matrix; performing feature combination on the voltage and current trajectory matrix and the temperature impedance trajectory matrix to obtain a state trajectory matrix, and based on a preset plurality of charge and discharge rate parameters, extracting the battery pack capacity attenuation characteristics of the voltage and capacity attenuation corresponding to the state trajectory matrix; based on a preset battery pack charging temperature range, performing correlation calculation of the battery pack impedance and cycle life on the battery pack capacity attenuation characteristics to generate the battery pack state evolution characteristics.

[0010] Optionally, in a fifth implementation of the first aspect of the present invention, the state monitoring results include battery pack state of charge monitoring results, health state monitoring results and power state monitoring results, and the state of charge dynamic characteristics corresponding to the battery pack state evolution characteristics are extracted, and based on the historical operation data, the state of charge dynamic characteristics are subjected to multi-layer feature mapping and state threshold classification to obtain state monitoring results of multiple state monitoring parameters, including: performing time series decomposition on the voltage and current state data in the battery pack state evolution characteristics to obtain the state of charge dynamic characteristics, and based on a preset voltage difference threshold, calculating the cell voltage consistency of the state of charge dynamic characteristics to obtain cell voltage correlation parameters; based on a preset cell temperature difference range, performing correlation calculation of temperature distribution characteristics on the cell voltage correlation parameters to obtain inter-cell state correlation characteristics, and performing spatial position mapping on the inter-cell state correlation characteristics to obtain battery pack spatial distribution characteristics; The spatial distribution characteristics of the battery pack are combined in time series to obtain the spatiotemporal characteristics of the battery pack, and based on the preset multiple state monitoring parameters, the numerical distribution of the multi-state monitoring parameters of the spatiotemporal characteristics of the battery pack is calculated to obtain the multi-state monitoring probability distribution; the confidence level of each monitoring probability distribution is calculated to obtain the state estimation interval, and based on the corresponding historical operating condition data in the historical operation data, the state estimation interval is state classified to obtain the battery pack charge state monitoring result, health state monitoring result and power state monitoring result.

[0011] Optionally, in a sixth implementation method of the first aspect of the present invention, the state monitoring results are subjected to coupling mapping of multiple state monitoring parameters and conversion probability calculation of the battery pack state to obtain a state space transfer feature matrix, and based on the historical operating data and preset battery pack operating condition parameters, the state space transfer feature matrix is ​​subjected to state evolution path calculation and associated confidence calculation of corresponding battery pack life characteristics to generate a state prediction result of the target smart battery pack, including: performing correlation calculation of state monitoring parameters and parameter coupling calculation on the battery pack state of charge monitoring results, the health state monitoring results and the power state monitoring results to obtain state transfer characteristics, and based on a plurality of preset charge and discharge rate parameters, extracting the operating condition classification characteristics corresponding to the state transfer characteristics to construct an operating condition characteristic matrix; performing state transition calculation on the operating condition characteristic matrix The state space transfer feature matrix is ​​obtained by constructing the probability statistics of the shift and the state space, and based on the preset battery pack operating parameters, the path planning and multi-path combination of the battery pack state evolution are performed on the state space transfer feature matrix to obtain at least one initial battery pack evolution path; based on the preset battery pack capacity attenuation termination threshold, the life attenuation calculation and threshold interval classification are performed on the initial battery pack evolution path to obtain an initial life prediction path, and the state change interval calculation and mean calculation of the multi-state monitoring parameters are performed on the initial life prediction path to obtain a state prediction interval; based on the corresponding number of charge and discharge cycles in the historical operation data, the life feature mapping and confidence calculation are performed on the state prediction interval to obtain a state prediction confidence, and the state prediction confidence is weighted and weighted fused to generate a state prediction result of the target intelligent battery pack.

[0012] The second aspect of the present invention provides a state monitoring and prediction device for an intelligent battery pack, the state monitoring and prediction device for an intelligent battery pack comprising: a preprocessing module for acquiring original battery pack data and historical operation data of a plurality of preset battery pack parameters in a target intelligent battery pack, and performing signal filtering and spectrum decomposition on the original battery pack data to obtain initial battery pack characteristic data; a space mapping module for performing abnormal data detection and abnormal correction on the initial battery pack characteristic data to obtain a standard battery pack characteristic sequence of the target intelligent battery pack, and performing multi-modal characteristic decomposition and phase space mapping of the battery pack state on the standard battery pack characteristic sequence to obtain battery pack state evolution characteristics; a state A monitoring module is used to extract the corresponding state of charge dynamic characteristics from the battery pack state evolution characteristics, and based on the historical operation data, perform multi-layer feature mapping and state threshold classification on the state of charge dynamic characteristics to obtain state monitoring results of multiple state monitoring parameters; a state prediction module is used to perform coupling mapping of multiple state monitoring parameters and conversion probability calculation of the battery pack state on the state monitoring results to obtain a state space transfer feature matrix, and based on the historical operation data and preset battery pack operating parameters, calculate the state evolution path of the state space transfer feature matrix and the associated confidence calculation of the corresponding battery pack life characteristics to generate a state prediction result of the target intelligent battery pack.

[0013] The third aspect of the present invention provides a state monitoring and prediction device for an intelligent battery pack, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the state monitoring and prediction device of the intelligent battery pack executes the various steps of the above-mentioned state monitoring and prediction method for the intelligent battery pack.

[0014] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the various steps of the above-mentioned method for monitoring and predicting the state of an intelligent battery pack.

[0015] The above-mentioned state monitoring and prediction method and related equipment of the intelligent battery pack. In the embodiment of the present invention, the original battery pack data and historical operation data of multiple preset battery pack parameters in the target intelligent battery pack are obtained, and the original battery pack data are subjected to signal filtering and spectrum decomposition to obtain the initial battery pack feature data; the initial battery pack feature data are subjected to abnormal data detection and abnormal correction to obtain the standard battery pack feature sequence of the target intelligent battery pack, and the standard battery pack feature sequence is subjected to multi-modal feature decomposition and phase space mapping of the battery pack state to obtain the battery pack state evolution characteristics; the corresponding state of charge dynamic characteristics in the battery pack state evolution characteristics are extracted, and based on the historical operation data, the state of charge dynamic characteristics are subjected to multi-layer feature mapping and state threshold classification to obtain the state monitoring results of multiple state monitoring parameters; the state monitoring results are subjected to coupling mapping of multiple state monitoring parameters and conversion probability calculation of the battery pack state to obtain the state space transfer feature matrix, and based on the historical operation data and the preset battery pack operating parameters, the state space transfer feature matrix is ​​subjected to state evolution path calculation and associated confidence calculation of the corresponding battery pack life characteristics to generate the state prediction result of the target intelligent battery pack. Compared with the prior art, this application obtains basic feature data by multi-dimensional acquisition and feature extraction of the electrochemical parameters of the target battery pack, and performs signal filtering and spectrum decomposition on these data to obtain initial feature data; then, through abnormal detection and correction, a standardized feature sequence is obtained, and multimodal feature decomposition and state space mapping are performed to obtain state evolution characteristics; then, based on historical operation data, multi-layer mapping and threshold classification are performed on the state evolution characteristics to obtain state monitoring results; finally, through state parameter coupling mapping and conversion probability calculation, the state transfer matrix is ​​obtained, and the state evolution path calculation is performed to output the prediction results. Through hierarchical data processing and feature analysis, the problem of accurate evaluation of battery pack state monitoring and prediction is solved, especially in terms of voltage fluctuation, temperature gradient and internal resistance change, the dynamic characteristics and decay law of the battery pack are fully considered, and the prediction accuracy is effectively improved; and the multi-level feature decoupling and state mapping strategy are adopted to realize the correlation analysis between electrochemical parameters and enhance the reliability of state evaluation; in addition, through state space reconstruction and evolution prediction, the performance decay trend is accurately evaluated, thereby realizing efficient and accurate prediction of the battery pack state as a whole.

[0016] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0017] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of a first embodiment of a method for monitoring and predicting the state of an intelligent battery pack according to an embodiment of the present invention;

[0019] Figure 2 A schematic diagram of an embodiment of a state monitoring and prediction device for a smart battery pack according to an embodiment of the present invention;

[0020] Figure 3 Schematic diagram of an embodiment of a state monitoring and prediction device for a smart battery pack in an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] The terms "including" and "having" and any variations thereof mentioned in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device end including a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or device ends.

[0023] To facilitate understanding of this embodiment, the specific process of the embodiment of the present invention is described below. Figure 1 The first embodiment of the state monitoring and prediction method of the intelligent battery pack in the embodiment of the present invention includes:

[0024] 101. Acquire original battery pack data and historical operation data of multiple preset battery pack parameters in the target intelligent battery pack, and perform signal filtering and spectrum decomposition on the original battery pack data to obtain initial battery pack characteristic data;

[0025] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0026] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0027] In this embodiment, the preset battery pack parameters here may include but are not limited to electrical parameters (current, voltage, internal resistance and capacity), thermal parameters (temperature (surface temperature, internal temperature and ambient temperature), temperature gradient (temperature distribution uniformity and hot spot location), heat dissipation performance (thermal resistance and heat capacity), state parameters (charge state, health state and power state), usage parameters (number of charge and discharge cycles, charge and discharge rate (C rate), usage time and static time) and environmental parameters, etc.); the historical operation data here include various data generated during the historical operation of the smart battery pack such as historical operating condition data and number of charge and discharge cycles; the original battery pack data include battery pack voltage data, battery pack current data, battery pack temperature data, and the battery pack voltage data is subjected to multi-layer recursive decomposition of discrete wavelets to obtain multi-band voltage feature group data, and the voltage feature group data is gain calculated to obtain the battery pack voltage attenuation characteristics; the battery pack voltage data is subjected to multi-layer recursive decomposition of discrete wavelets to obtain multi-band voltage feature group data, and the voltage feature group data is gain calculated to obtain the battery pack voltage attenuation characteristics; the battery pack voltage data is subjected to multi-frequency recursive decomposition of discrete wavelets to obtain ... The battery pack current data is calculated in the time-frequency domain to obtain the current characteristic distribution, and the current characteristic distribution is decomposed in multiple scales to obtain the battery pack current fluctuation characteristics; the battery pack temperature data is calculated in two-dimensional space gradient to obtain the temperature distribution matrix, and the uniformity index of the temperature distribution matrix is ​​calculated to obtain the battery pack temperature field characteristics; the impedance characteristics of the battery pack voltage attenuation characteristics and the battery pack current fluctuation characteristics are extracted to generate the internal resistance change curve, and the charge and discharge cross calculation of the battery pack voltage attenuation characteristics and the battery pack current fluctuation characteristics is performed to obtain the battery pack charge and discharge response characteristics, and the impedance parameters of the battery pack charge and discharge response characteristics and the internal resistance change curve are matched to obtain the battery pack impedance dynamic characteristics; the battery pack impedance dynamic characteristics and the battery pack temperature field characteristics are thermoelectrically coupled mapped to obtain the multi-physical field coupling characteristics, and the time series characteristic matrix of the multi-physical field coupling characteristics is constructed to generate the initial battery pack characteristic data.

[0028] In practical applications, the basic parameter data of the smart battery pack is first collected through a pre-installed distributed sensor network. Among them, voltage data collection is realized through a high-precision voltage sampling module, and the sampling frequency is set to 100Hz, which can simultaneously obtain the total voltage of the battery pack and the single cell voltage; current data collection uses a Hall current sensor, which also records the charging and discharging current values ​​at a sampling frequency of 100Hz; temperature data is collected through a thermocouple array arranged on the surface of the battery pack, with a sampling frequency of 1Hz, and the surface temperature distribution data of the battery pack can be obtained. At the same time, the historical operation data of the smart battery group is collected and recorded, including the number of charge and discharge cycles, usage time, environmental parameters and other information; then the collected voltage data is recursively decomposed by discrete wavelet transform, and the voltage data is decomposed into 5 layers using db4 wavelet basis function to obtain voltage feature group data of different frequency bands (these feature group data reflect the change characteristics of voltage on different time scales), and the voltage attenuation characteristics of the battery pack are obtained by gain calculation of the voltage feature group data of each frequency band coefficient (reflecting the trend of battery capacity attenuation and performance degradation), and the short-time Fourier transform is used to calculate the time spectrum to decompose the current data in the time-frequency domain to obtain the current characteristic distribution, and then the wavelet packet decomposition is used for multi-scale analysis to obtain the battery pack current fluctuation characteristics (reflecting the dynamic response characteristics of the battery under different load conditions); and based on the two-dimensional spatial gradient calculation, the temperature data of the battery pack is differentially operated to obtain the temperature distribution matrix (wherein the matrix describes the spatial distribution characteristics of the surface temperature of the battery pack), and the uniformity index of the temperature field is calculated, including the maximum temperature difference, the average temperature and standard deviation, so as to obtain the temperature field characteristics of the battery pack to evaluate the thermal management status of the battery pack and the potential risk of thermal runaway; based on the obtained voltage decay characteristics and current fluctuation characteristics, the electrochemical impedance spectrum analysis is performed to extract the impedance characteristics, generate the internal resistance change curve, and the voltage and current data are cross-calculated in the charge and discharge process to calculate the charge and discharge efficiency, energy conversion efficiency and other parameters to obtain the charge and discharge response characteristics of the battery pack, and the battery pack charge and discharge response characteristics are matched with the internal resistance change curve by parameters, and the equivalent circuit model is used for fitting, and finally the impedance dynamic characteristics reflecting the change of the internal state of the battery are obtained; in order to comprehensively evaluate the working state of the battery pack, the impedance dynamic characteristics and temperature field characteristics are thermoelectrically coupled mapped, that is, by establishing a thermo-electric coupling model, the influence of temperature on internal resistance and the effect of internal resistance heating on temperature distribution are considered, so as to obtain the multi-physical field coupling characteristics, and the multi-physical field coupling coupling characteristics are organized in time series to construct a feature matrix, which contains the comprehensive feature information of the battery pack at multiple levels such as electrical, thermal and electrochemical to generate the initial battery pack feature data.

[0029] 102. Perform abnormal data detection and abnormal correction on the initial battery pack feature data to obtain a standard battery pack feature sequence of the target intelligent battery pack, and perform multi-modal feature decomposition and phase space mapping of the battery pack state on the standard battery pack feature sequence to obtain battery pack state evolution characteristics;

[0030] In this embodiment, the initial battery pack characteristic data is divided into time windows to obtain segmented time series data, and the corresponding voltage and current time series data segments in the segmented time series data are subjected to extreme value deviation detection and interquartile range test to obtain abnormal point marking data, and the corresponding battery pack temperature field data in the segmented time series data is subjected to thermal distribution test to obtain temperature abnormal area marking; the abnormal point marking data is subjected to cubic segmented spline interpolation to obtain a continuous characteristic curve, and the temperature abnormal area is subjected to local weighted smoothing calculation to obtain corrected temperature field distribution data, and the continuous characteristic curve is subjected to residual detection and deviation test to obtain voltage and current quality indicators, and the temperature field distribution data is subjected to gradient consistency test to obtain the temperature field. Quality indicators; extract multi-dimensional threshold features of voltage and current quality indicators and temperature field quality indicators to obtain the battery pack working state quality score, and based on the battery pack working state quality score, perform graded screening and data segment identification on the voltage and current quality indicators and temperature field quality indicators to obtain the identification collection data segment; perform time series reconstruction on the identification collection data segment to obtain the standard battery pack feature sequence of the target intelligent battery pack; divide the corresponding voltage and current feature data in the standard battery pack feature sequence into charge and discharge cycles to obtain the initial battery pack working mode, and classify the corresponding temperature and impedance feature data in the initial battery pack working mode into multiple charge and discharge rates to obtain the multi-dimensional battery pack modal component; perform The time-frequency characteristics are extracted to obtain the time-frequency response characteristics of the target intelligent battery pack, and the energy distribution of the time-frequency response characteristics is calculated to obtain the battery pack working condition energy spectrum; the charge and discharge state characteristics are extracted from the corresponding voltage and current characteristic data in the battery pack working condition energy spectrum to obtain the charge and discharge state characteristics, and the corresponding temperature and impedance characteristic data in the battery pack working condition energy spectrum are mapped to the working state to obtain the temperature impedance state characteristics, and the charge and discharge state characteristics and the temperature impedance state characteristics are combined to obtain the multi-dimensional state sequence of the battery pack; the state space of the battery pack state and the evolution of the battery pack state are performed on the multi-dimensional state sequence of the battery pack to obtain the battery pack state evolution characteristics, that is, the charge and discharge of the multi-dimensional state sequence of the battery pack is performed. The working state of the cycle is mapped to obtain the working trajectory characteristics of the battery pack, and the corresponding voltage and current characteristic data in the working trajectory characteristics of the battery pack are mapped in state space to obtain the voltage and current trajectory matrix, and the corresponding temperature and impedance characteristic data in the working trajectory characteristics of the battery pack are mapped in state space to obtain the temperature impedance trajectory matrix; the voltage and current trajectory matrix and the temperature impedance trajectory matrix are combined to obtain the state trajectory matrix, and based on the preset multiple charge and discharge rate parameters, the battery pack capacity decay characteristics corresponding to the voltage and capacity decay of the state trajectory matrix are extracted; based on the preset battery pack charging temperature range, the battery pack capacity decay characteristics are correlated with the battery pack impedance and cycle life to generate the battery pack state evolution characteristics. The charge and discharge rate parameter here refers to the charging or discharging rate of the smart battery pack per unit time.

[0031] In practical applications, the initial battery pack characteristic data is first divided into time windows, that is, by adopting the sliding window method, the window size is dynamically adjusted according to the sampling frequency of different parameters. A smaller time window (such as 10 seconds) is used for the voltage and current data sampled at a high frequency, and a larger time window (such as 60 seconds) is used for the temperature data sampled at a low frequency to ensure the time continuity and characteristic integrity of the data. In each time window, the voltage and current time series data are tested for extreme deviations. By calculating the mean and standard deviation of the data, outliers beyond the range of μ±3σ are identified. At the same time, the interquartile range method is used to calculate Q1 (25% quantile), Q3 (75% quantile) and IQR (Q3 - Q1). The values ​​beyond [Q1-1.5IQR, Q3+1.5IQR] as potential abnormal points, and for the temperature field data, calculate the spatial distribution characteristics of the temperature field, identify areas with abnormal temperature gradient or local overheating, and mark these areas as temperature abnormal areas; then use the cubic piecewise spline interpolation method, select normal data points before and after the abnormal points as control points, and construct the cubic spline function as follows:

[0032] ;

[0033] in, is the Hermite basis function, is the data point value, , , They are first-order, second-order, and third-order derivative values ​​(for example, in the process of charging at rated current 1C, the detected current mutation point is interpolated and corrected) to achieve smooth correction of abnormal data and obtain continuous characteristic curves. For temperature abnormality areas, a local weighted smoothing algorithm is used to reconstruct the temperature field of the abnormal area considering the influence of the surrounding normal temperature points to obtain the corrected temperature field distribution data. Then, the corrected characteristic curve is subjected to residual detection, the deviation between the correction value and the original data is calculated, the correction effect is evaluated, and the voltage and current quality indicators of the voltage and current data are determined through deviation test. At the same time, the temperature field distribution data is subjected to gradient consistency test to evaluate the uniformity and stability of the temperature field and obtain the temperature field quality indicator reflecting the quality of the temperature distribution. Then, feature extraction is performed by setting multidimensional thresholds, and these thresholds include data integrity thresholds, volatility thresholds, and consistency thresholds. Based on these multidimensional thresholds, the comprehensive quality score Q of the working state of the battery pack is calculated by multidimensional feature weighted fusion. The quality score formula is:

[0034] ;

[0035] in, is the weight of the i-th feature, is the normalized eigenvalue of the ith feature, β is the attenuation coefficient, is the characteristic standard deviation of the i-th feature, The maximum standard deviation of the feature (for example, voltage stability weight 0.4, temperature uniformity weight 0.3, current fluctuation weight 0.3, and the quality score is obtained by comprehensive calculation) is used to reflect the reliability and validity of the data; then, according to the quality score, the voltage and current quality indicators and the temperature field quality indicators are graded and screened, and the data segments are divided into high-quality segments, medium-quality segments and low-quality segments, and marked accordingly; then, the marked high-quality data segments are time-series reconstructed, and a standardized battery pack feature sequence (including complete feature information after anomaly detection and quality assessment) is generated through time alignment and feature fusion.

[0036] Secondly, the data is divided into charge and discharge cycles according to the changing trends of voltage and current. By identifying the characteristic points of the voltage rising section (charging process) and the falling section (discharging process), combined with the change of current direction, the complete charge and discharge cycle is determined (for example, for the charge and discharge process of a battery pack, when the charging voltage rises from 3.2V to 4.2V, and the charging current gradually decreases from 2C to 0.05C, the charging termination point can be identified). In each cycle, the working state is divided into fast charging, slow charging, standard charging, high current discharge, low current discharge and other different working modes according to the current size. At the same time, the temperature and impedance characteristic data of the corresponding time period are classified according to different charge and discharge rates (such as 0.5C, 1C, 2C, etc.), so as to obtain multi-dimensional modal components reflecting the performance characteristics of the battery pack under different working conditions; then the S transform method is used to extract the time-frequency features to obtain the time-frequency response characteristics of the battery pack under different working modes, and the S transform formula is:

[0037] ;

[0038] in, is a frequency-dependent Gaussian window function, f is the frequency, τ is the time offset, As the input signal, the time domain and frequency domain information can be obtained simultaneously (for example, the process of voltage fluctuation frequency changing from 0.1Hz to 1Hz during charging can be analyzed), so as to obtain the time-frequency response characteristics; and by calculating the energy distribution of each frequency band, including the power spectrum density and energy concentration of the signal, the working condition energy spectrum of the battery pack is obtained, which reflects the energy distribution characteristics of the battery pack under different working conditions; then, on the basis of the working condition energy spectrum, the charge and discharge state characteristics of the voltage and current data are extracted respectively, including charging efficiency, discharge depth, voltage platform characteristics, etc., and the working state mapping of the temperature and impedance data is performed at the same time to obtain the state characteristics reflecting the temperature distribution and internal resistance change, and these characteristics are combined to construct a multi-dimensional state sequence containing various aspects of the performance indicators of the battery pack; then, by mapping the working state during the charge and discharge cycle, the trajectory characteristics reflecting the working process of the battery pack are obtained, and the voltage and current data are mapped in the state space to obtain the trajectory matrix describing the voltage and current change law. Similarly, the temperature and impedance data are mapped in the state space to obtain the temperature impedance trajectory matrix reflecting the thermodynamic and electrochemical characteristics (wherein, these two trajectory matrices contain The performance change characteristics of the battery pack under different working conditions); then the voltage and current trajectory matrix and the temperature impedance trajectory matrix are feature combined to construct a complete state trajectory matrix to comprehensively reflect the overall working state of the battery pack, and based on the pre-set different charge and discharge rate parameters (such as multiple gears from 0.5C to 3C), the voltage change characteristics and capacity change characteristics are extracted from the state trajectory matrix to obtain the capacity decay characteristics reflecting the degree of battery pack performance degradation (and the capacity decay characteristics can be quantified by the coulomb counting method and the incremental capacity analysis method to evaluate the health status of the battery pack); then according to the preset battery pack charging temperature range (usually 0℃ to 45℃), the relationship between the capacity decay characteristics and the impedance change is analyzed, and the cycle life parameters under different temperature conditions are calculated, including the capacity retention rate, the internal resistance growth rate, etc., and by establishing an electrochemical-thermodynamic coupling model, the influence of temperature on battery performance and the influence of energy loss during charging and discharging on the temperature field are comprehensively considered, and finally the battery pack state evolution characteristics reflecting the law of battery pack state change are generated, wherein the characteristics not only contain the current working state information of the battery pack, but also contain the trend information of performance degradation.

[0039] 103. Extract the state of charge dynamic features corresponding to the battery pack state evolution features, and perform multi-layer feature mapping and state threshold classification on the state of charge dynamic features based on historical operation data to obtain state monitoring results of multiple state monitoring parameters;

[0040] In this embodiment, the state monitoring results here include the battery pack state of charge monitoring results, health state monitoring results and power state monitoring results; the voltage and current state data in the battery pack state evolution characteristics are time-series decomposed to obtain the state of charge dynamic characteristics, and based on the preset voltage difference threshold, the single cell voltage consistency is calculated for the state of charge dynamic characteristics to obtain the single cell voltage association parameters; based on the preset single cell temperature difference interval, the temperature distribution characteristics of the single cell voltage association parameters are associated with the calculation to obtain the inter-cell state association characteristics, and the inter-cell state association characteristics are spatially mapped to obtain the battery pack spatial distribution characteristics; the battery pack spatial distribution characteristics are combined in time series to obtain the battery pack spatiotemporal characteristics, and based on the preset multiple state monitoring parameters, the numerical distribution calculation of the multi-state monitoring parameters is performed on the battery pack spatiotemporal characteristics to obtain the multi-state monitoring probability distribution; the confidence level calculation is performed on each monitoring probability distribution to obtain the state estimation interval, and based on the corresponding historical operating condition data in the historical operation data, the state estimation interval is state-classified to obtain the battery pack state of charge monitoring results, health state monitoring results and power state monitoring results. The above-mentioned voltage difference threshold refers to the important parameters used to evaluate the voltage balance of each single cell in the battery pack (such as single cell voltage difference threshold, balance trigger threshold, safety protection threshold and working status threshold, etc.); the above-mentioned single cell temperature difference range refers to the maximum temperature difference range allowed between each single cell in the battery pack; the above-mentioned multiple status monitoring parameters include but are not limited to charge state, health state and power state, etc.

[0041] In practical applications, the empirical mode decomposition (EMD) method is used to decompose complex time series signals into multiple intrinsic mode functions, thereby extracting the dynamic characteristics that reflect the change of charge state during the charging and discharging process of the battery pack. Based on the preset voltage difference threshold (usually set to ±10mV), the entropy weight method is used to perform consistency analysis on the voltage of each single cell in the battery pack, and the standard deviation and coefficient of variation of the single cell voltage are calculated. The single cell voltage consistency calculation formula is:

[0042] , ;

[0043] Where: H is the consistency index, is the normalized voltage deviation of the ith monomer, is the voltage of the ith cell, is the average voltage, σ is the standard deviation, and n is the number of battery cells (for example, assuming that the battery pack has 100 cells and the voltage difference threshold is set to 50mV to calculate the consistency of the cell voltage distribution), so as to calculate the correlation parameters that characterize the voltage balance between cells. These parameters reflect the consistency of the working state of each cell in the battery pack. Then, based on the preset cell temperature difference range (generally controlled within 3-5°C), the temperature distribution characteristics of the cell voltage correlation parameters are analyzed, that is, by establishing a temperature-voltage mapping relationship, the influence coefficient of the temperature gradient on the voltage distribution is calculated, and the state correlation characteristics between cells are obtained. These correlation characteristics are spatially mapped according to the actual physical layout inside the battery pack to obtain the characteristics reflecting the spatial distribution of the overall working state of the battery pack (including the distribution law of voltage and temperature in the spatial dimension). Then, the spatial distribution characteristics are combined in time series to construct the spatiotemporal characteristics of the battery pack including the two dimensions of time and space, and based on the preset multiple state monitoring parameters, including the state of charge The key indicators such as state of charge (SOC), state of health (SOH) and state of power (SOP) are selected, and the probability density estimation method is used to perform multi-dimensional numerical distribution calculations on the spatiotemporal characteristics to obtain the monitoring probability distribution that reflects the distribution characteristics of each state parameter, so as to comprehensively reflect the performance status of the battery pack under different working conditions; then a 95% confidence interval is used to calculate the confidence level of each monitoring probability distribution to obtain the estimated range of each state parameter, and these state estimation intervals are classified and analyzed in combination with the operating condition information recorded in the historical operation data, including the performance data under different working conditions, and finally the three key state monitoring results of the battery pack are obtained: the state of charge monitoring result reflects the current available capacity level, the health state monitoring result characterizes the degree of capacity attenuation, and the power state monitoring result indicates the current power output capacity, so as to comprehensively evaluate the performance status of the battery pack and realize accurate monitoring of the battery pack status, especially in terms of voltage balance, temperature distribution uniformity and overall performance evaluation, providing comprehensive and accurate status evaluation results.

[0044] 104. The state monitoring results are subjected to coupling mapping of multi-state monitoring parameters and conversion probability calculation of battery pack states to obtain a state space transfer feature matrix. Based on historical operating data and preset battery pack operating parameters, the state space transfer feature matrix is ​​subjected to calculation of state evolution path and associated confidence calculation of corresponding battery pack life characteristics to generate a state prediction result for the target intelligent battery pack.

[0045] In this embodiment, the correlation of state monitoring parameters and the coupling of parameters are calculated for the battery pack state of charge monitoring results, health state monitoring results and power state monitoring results to obtain state transfer characteristics, and based on the preset multiple charge and discharge rate parameters, the operating condition classification characteristics corresponding to the state transfer characteristics are extracted to construct an operating condition characteristic matrix; the operating condition characteristic matrix is ​​subjected to state migration probability statistics and state space construction to obtain a state space transfer characteristic matrix, and based on the preset battery pack operating condition parameters, the state space transfer characteristic matrix is ​​subjected to path planning and multi-path combination of battery pack state evolution to obtain at least one initial battery pack evolution path; based on the preset battery pack capacity attenuation termination threshold, the initial battery pack evolution path is subjected to life attenuation calculation and threshold interval classification to obtain an initial life prediction path, and the initial life prediction path is subjected to state change interval calculation and mean calculation of multiple state monitoring parameters to obtain a state prediction interval; based on the corresponding number of charge and discharge cycles in the historical operation data, the state prediction interval is subjected to life feature mapping and confidence calculation to obtain a state prediction confidence, and the state prediction confidence is subjected to weight allocation and weighted fusion to generate a state prediction result of the target intelligent battery pack. The multiple charge and discharge rate parameters here refer to the set charge and discharge rates of the smart battery pack in the corresponding working state; the battery pack operating condition parameters here refer to a series of key parameters that describe the working environment and usage conditions of the battery pack (such as operating condition parameters (charging condition and discharging condition), cycle condition parameters, environmental condition parameters and safety condition parameters, etc.); the battery pack capacity attenuation termination threshold here refers to an important indicator for measuring the end of battery pack life, such as the state point when the remaining capacity of the battery pack drops to 80% of the rated capacity.

[0046] In practical applications, the Pearson correlation coefficient is used to calculate the correlation between the state monitoring parameters of the battery pack charge state monitoring results, health state monitoring results and power state monitoring results. At the same time, the mutual information theory is used to evaluate the coupling strength between the parameters, so as to obtain the state transition characteristics that reflect the mutual influence of the state parameters, so as to characterize the state change law of the battery pack under different working conditions, and based on the preset charge and discharge rate parameters (including different rates such as 0.5C, 1C, 2C, etc.), the state transition characteristics are classified into working conditions, and the characteristic performance under different charge and discharge conditions is obtained, so as to construct a working condition characteristic matrix containing multiple working conditions; and then the Markov chain model is used to calculate the migration between the states of the constructed working condition characteristic matrix. The state transition probability matrix is ​​established, and the characteristic vector reflecting the performance evolution of the battery pack is constructed in the state space to obtain the state space transfer characteristic matrix. Based on the preset battery pack operating parameters, including different temperature conditions, load characteristics, etc., the state space transfer characteristic matrix is ​​analyzed, and the state evolution path is designed through the dynamic programming algorithm. Considering multiple possible evolution directions, at least one initial evolution path reflecting the performance change trend of the battery pack is combined to form; then based on the capacity attenuation termination threshold of the battery pack (usually set to 80% of the rated capacity), the improved particle swarm optimization algorithm is used to analyze the life characteristics of the initial evolution path, and the capacity attenuation rate and remaining life under different operating conditions are calculated. The calculation formula is:

[0047] ;

[0048] Among them, v(t) is the velocity of the particle at time t, x(t) is the position of the particle at time t, is the local optimal solution of the particle, is the global optimal solution, ω is the inertia weight, c1, c2 are learning factors, is a random number (in the range [0,1]), β is the convergence adjustment coefficient, The distance between the particle position and the global optimal solution (for example: predicting the possible path of battery decay from the current health state = 90% to 80%), and classifying these decay characteristics according to different threshold intervals to obtain the initial life prediction path; and based on the preset battery pack capacity decay termination threshold, calculate the change range and average level of each state monitoring parameter at different time points, so as to obtain the state prediction interval containing prediction uncertainty; and then combine the number of charge and discharge cycles recorded in the historical operation data to map the life characteristics of the state prediction interval, use the Monte Carlo method to evaluate the reliability of the prediction results, calculate the confidence level of different prediction paths, and obtain the state prediction confidence. And for these prediction results with different confidence levels, assign weight coefficients based on their reliability levels, and fuse multiple prediction results through the weighted average method, and finally generate a state prediction result that reflects the future performance change trend of the battery pack. This prediction result not only contains the trend information of battery pack performance degradation, but also provides prediction uncertainty assessment, which provides an important reference for the maintenance and replacement decision of the battery pack, thereby realizing the transformation from state monitoring to performance prediction, especially in dealing with multi-parameter coupling relationships and prediction uncertainties, ensuring the reliability and practicality of the prediction results.

[0049] In the embodiment of the present invention, the electrochemical parameters of the target battery pack are collected and feature extracted in multiple dimensions to obtain basic feature data, and these data are subjected to signal filtering and spectrum decomposition to obtain initial feature data; then, a standardized feature sequence is obtained through abnormal detection and correction, and multimodal feature decomposition and state space mapping are performed to obtain state evolution features; then, based on historical operation data, multi-layer mapping and threshold classification are performed on the state evolution features to obtain state monitoring results; finally, the state parameter coupling mapping and conversion probability calculation are performed to obtain the state transfer matrix, and the state evolution path calculation is performed to output the prediction results. Through hierarchical data processing and feature analysis, the problem of accurate evaluation of battery pack state monitoring and prediction is solved, especially in terms of voltage fluctuation, temperature gradient and internal resistance change, the dynamic characteristics and decay law of the battery pack are fully considered, and the prediction accuracy is effectively improved; and the multi-level feature decoupling and state mapping strategy are adopted to realize the correlation analysis between electrochemical parameters and enhance the reliability of state evaluation; in addition, through state space reconstruction and evolution prediction, the performance decay trend is accurately evaluated, thereby realizing efficient and accurate prediction of the battery pack state as a whole.

[0050] The above describes the state monitoring and prediction method of the intelligent battery pack in the embodiment of the present invention. The following describes the state monitoring and prediction device of the intelligent battery pack in the embodiment of the present invention. Figure 2 , an embodiment of the state monitoring and prediction device of the intelligent battery pack in the embodiment of the present invention includes:

[0051] The preprocessing module 201 is used to obtain the original battery pack data and historical operation data of multiple preset battery pack parameters in the target intelligent battery pack, and perform signal filtering and spectrum decomposition on the original battery pack data to obtain initial battery pack characteristic data;

[0052] The space mapping module 202 is used to perform abnormal data detection and abnormal correction on the initial battery pack feature data to obtain a standard battery pack feature sequence of the target intelligent battery pack, and perform multi-modal feature decomposition and phase space mapping of the battery pack state on the standard battery pack feature sequence to obtain battery pack state evolution characteristics;

[0053] The state monitoring module 203 is used to extract the state of charge dynamic features corresponding to the battery pack state evolution features, and perform multi-layer feature mapping and state threshold classification on the state of charge dynamic features based on the historical operation data to obtain state monitoring results of multiple state monitoring parameters;

[0054] The state prediction module 204 is used to perform coupling mapping of multi-state monitoring parameters and conversion probability calculation of battery pack states on the state monitoring results to obtain a state space transfer feature matrix, and based on the historical operating data and preset battery pack operating parameters, calculate the state evolution path of the state space transfer feature matrix and the associated confidence calculation of the corresponding battery pack life characteristics to generate a state prediction result of the target intelligent battery pack.

[0055] In the embodiment of the present invention, the electrochemical parameters of the target battery pack are collected and feature extracted in multiple dimensions to obtain basic feature data, and these data are subjected to signal filtering and spectrum decomposition to obtain initial feature data; then, a standardized feature sequence is obtained through abnormal detection and correction, and multimodal feature decomposition and state space mapping are performed to obtain state evolution features; then, based on historical operation data, multi-layer mapping and threshold classification are performed on the state evolution features to obtain state monitoring results; finally, the state parameter coupling mapping and conversion probability calculation are performed to obtain the state transfer matrix, and the state evolution path calculation is performed to output the prediction results. Through hierarchical data processing and feature analysis, the problem of accurate evaluation of battery pack state monitoring and prediction is solved, especially in terms of voltage fluctuation, temperature gradient and internal resistance change, the dynamic characteristics and decay law of the battery pack are fully considered, and the prediction accuracy is effectively improved; and the multi-level feature decoupling and state mapping strategy are adopted to realize the correlation analysis between electrochemical parameters and enhance the reliability of state evaluation; in addition, through state space reconstruction and evolution prediction, the performance decay trend is accurately evaluated, thereby realizing efficient and accurate prediction of the battery pack state as a whole.

[0056] above Figure 2The state monitoring and prediction device of the intelligent battery pack in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The state monitoring and prediction device of the intelligent battery pack in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0057] Figure 3 1 is a schematic diagram of the structure of a state monitoring and prediction device for a smart battery pack provided by an embodiment of the present invention. The state monitoring and prediction device 300 for the smart battery pack may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 (for example, one or more mass storage devices) storing application programs 333 or data 332. Among them, the memory 320 and the storage medium 330 may be short-term storage or permanent storage. The program stored in the storage medium 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the state monitoring and prediction device 300 of the smart battery pack. Furthermore, the processor 310 may be configured to communicate with the storage medium 330 to execute a series of instruction operations in the storage medium 330 on the state monitoring and prediction device 300 of the smart battery pack.

[0058] The intelligent battery pack state monitoring and prediction device 300 may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input and output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. It will be understood by those skilled in the art that Figure 3 The structure of the smart battery pack condition monitoring and prediction device shown does not constitute a limitation on the smart battery pack condition monitoring and prediction device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0059] The present invention also provides a state monitoring and prediction device for an intelligent battery pack, wherein the computer device includes a memory and a processor, wherein the memory stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the processor executes each step of the state monitoring and prediction method for the intelligent battery pack in the above-mentioned embodiments.

[0060] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes each step of the method for monitoring and predicting the state of the intelligent battery pack.

[0061] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0062] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.

[0063] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0064] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring and predicting the state of an intelligent battery pack, characterized in that: The state monitoring and prediction method of the intelligent battery pack includes: Acquire original battery pack data and historical operation data of multiple preset battery pack parameters in the target intelligent battery pack, and perform signal filtering and spectrum decomposition on the original battery pack data to obtain initial battery pack characteristic data; The initial battery pack characteristic data is subjected to abnormal data detection and abnormal correction to obtain a standard battery pack characteristic sequence of the target intelligent battery pack, and the corresponding voltage and current characteristic data in the standard battery pack characteristic sequence are divided into charge and discharge cycles to obtain an initial battery pack working mode, and the corresponding temperature and impedance characteristic data in the initial battery pack working mode are classified into a variety of charge and discharge rates to obtain a multi-dimensional battery pack modal component; the time-frequency characteristics of each of the battery pack modal components are extracted to obtain the time-frequency response characteristics of the target intelligent battery pack, and the time-frequency response characteristics are energy-efficiently analyzed. The method comprises the following steps: calculating the quantity distribution of the battery pack and obtaining the battery pack working condition energy spectrum; extracting the charge and discharge state characteristics of the voltage and current characteristic data corresponding to the battery pack working condition energy spectrum to obtain the charge and discharge state characteristics, and mapping the working state of the temperature and impedance characteristic data corresponding to the battery pack working condition energy spectrum to obtain the temperature impedance state characteristics, and combining the charge and discharge state characteristics with the temperature impedance state characteristics to obtain a multi-dimensional state sequence of the battery pack; reconstructing the state space of the battery pack state and evolving the battery pack state on the multi-dimensional state sequence of the battery pack to obtain the battery pack state evolution characteristics; Extracting the state of charge dynamic features corresponding to the battery pack state evolution features, and performing multi-layer feature mapping and state threshold classification on the state of charge dynamic features based on the historical operation data to obtain state monitoring results of multiple state monitoring parameters; The state monitoring results are subjected to coupling mapping of multi-state monitoring parameters and conversion probability calculation of battery pack states to obtain a state space transfer feature matrix, and based on the historical operating data and preset battery pack operating parameters, the state space transfer feature matrix is ​​subjected to state evolution path calculation and associated confidence calculation of corresponding battery pack life characteristics to generate a state prediction result of the target intelligent battery pack.

2. The state monitoring and prediction method of the intelligent battery pack according to claim 1, characterized in that: The original battery pack data includes battery pack voltage data, battery pack current data, and battery pack temperature data. The signal filtering and spectrum decomposition of the original battery pack data to obtain initial battery pack characteristic data includes: Performing a multi-layer recursive decomposition of discrete wavelets on the battery pack voltage data to obtain multi-band voltage feature group data, and performing gain calculation on the voltage feature group data to obtain the battery pack voltage attenuation characteristics; Performing time-frequency domain calculation on the battery pack current data to obtain current characteristic distribution, and performing multi-scale decomposition on the current characteristic distribution to obtain battery pack current fluctuation characteristics; Performing two-dimensional spatial gradient calculation on the battery pack temperature data to obtain a temperature distribution matrix, and performing uniformity index calculation on the temperature distribution matrix to obtain a battery pack temperature field characteristic; Extracting impedance characteristics of the battery pack voltage attenuation characteristics and the battery pack current fluctuation characteristics to generate an internal resistance change curve, performing charge and discharge cross calculation on the battery pack voltage attenuation characteristics and the battery pack current fluctuation characteristics to obtain battery pack charge and discharge response characteristics, and performing impedance parameter matching on the battery pack charge and discharge response characteristics and the internal resistance change curve to obtain a battery pack impedance dynamic characteristic; The dynamic characteristics of the battery pack impedance and the temperature field characteristics of the battery pack are thermoelectrically coupled mapped to obtain multi-physical field coupling characteristics, and a time series characteristic matrix is ​​constructed for the multi-physical field coupling characteristics to generate initial battery pack characteristic data.

3. The state monitoring and prediction method of the intelligent battery pack according to claim 1, characterized in that: The abnormal data detection and abnormal correction of the initial battery pack characteristic data to obtain the standard battery pack characteristic sequence of the target smart battery pack includes: Dividing the initial battery pack characteristic data into time windows to obtain segmented time series data, performing extreme value deviation detection and interquartile range test on the corresponding voltage and current time series data segments in the segmented time series data to obtain abnormal point marking data, and performing thermal distribution test on the corresponding battery pack temperature field data in the segmented time series data to obtain temperature abnormal area marking; Performing cubic segmented spline interpolation on the abnormal point marking data to obtain a continuous characteristic curve, performing local weighted smoothing calculation on the temperature abnormality area to obtain corrected temperature field distribution data, performing residual detection and deviation inspection on the continuous characteristic curve to obtain voltage and current quality indicators, and performing gradient consistency inspection on the temperature field distribution data to obtain a temperature field quality indicator; Extracting multi-dimensional threshold features of the voltage and current quality indicators and the temperature field quality indicators to obtain a battery pack working state quality score, and based on the battery pack working state quality score, performing graded screening and data segment identification on the voltage and current quality indicators and the temperature field quality indicators to obtain an identified collection data segment; The identification collection data segment is time-sequence reconstructed to obtain a standard battery pack feature sequence of the target intelligent battery pack.

4. The state monitoring and prediction method of the intelligent battery pack according to claim 1, characterized in that: The state space reconstruction of the battery pack state and the evolution of the battery pack state are performed on the multi-dimensional state sequence of the battery pack to obtain the battery pack state evolution characteristics, including: Performing working state mapping of the charge and discharge cycle on the multi-dimensional state sequence of the battery pack to obtain the working trajectory characteristics of the battery pack, and performing state space mapping on the corresponding voltage and current characteristic data in the working trajectory characteristics of the battery pack to obtain a voltage and current trajectory matrix, and performing state space mapping on the corresponding temperature and impedance characteristic data in the working trajectory characteristics of the battery pack to obtain a temperature impedance trajectory matrix; Performing feature combination on the voltage-current trajectory matrix and the temperature-impedance trajectory matrix to obtain a state trajectory matrix, and extracting battery pack capacity decay characteristics corresponding to voltage and capacity decay of the state trajectory matrix based on a plurality of preset charge and discharge rate parameters; Based on the preset battery pack charging temperature range, the battery pack capacity attenuation characteristics are correlated with the battery pack impedance and the cycle life to generate the battery pack state evolution characteristics.

5. The state monitoring and prediction method of the intelligent battery pack according to claim 1, characterized in that: The state monitoring results include battery pack state of charge monitoring results, health state monitoring results and power state monitoring results. The state of charge dynamic features corresponding to the battery pack state evolution features are extracted, and based on the historical operation data, multi-layer feature mapping and state threshold classification are performed on the state of charge dynamic features to obtain state monitoring results of multiple state monitoring parameters, including: Performing time series decomposition on the voltage and current state data in the battery pack state evolution characteristics to obtain a state of charge dynamic characteristic, and calculating the cell voltage consistency of the state of charge dynamic characteristic based on a preset voltage difference threshold to obtain a cell voltage correlation parameter; Based on the preset cell temperature difference interval, the cell voltage correlation parameter is correlated with the temperature distribution characteristic to obtain the inter-cell state correlation characteristic, and the inter-cell state correlation characteristic is spatially mapped to obtain the battery pack spatial distribution characteristic; Combining the spatial distribution characteristics of the battery pack in time series to obtain the spatiotemporal characteristics of the battery pack, and performing numerical distribution calculation of multi-state monitoring parameters on the spatiotemporal characteristics of the battery pack based on a plurality of preset state monitoring parameters to obtain a multi-state monitoring probability distribution; The confidence level of each monitoring probability distribution is calculated to obtain a state estimation interval, and based on the corresponding historical operating condition data in the historical operation data, the state estimation interval is state classified to obtain the battery pack charge state monitoring result, health state monitoring result and power state monitoring result.

6. The state monitoring and prediction method of the intelligent battery pack according to claim 5, characterized in that: The state monitoring result is subjected to coupling mapping of multi-state monitoring parameters and conversion probability calculation of battery pack state to obtain a state space transfer feature matrix, and based on the historical operation data and preset battery pack operating condition parameters, the state space transfer feature matrix is ​​subjected to state evolution path calculation and associated confidence calculation of corresponding battery pack life characteristics to generate a state prediction result of the target intelligent battery pack, including: The state monitoring parameter correlation and parameter coupling are calculated for the battery pack state of charge monitoring result, the health state monitoring result and the power state monitoring result to obtain state transition characteristics, and based on a plurality of preset charge and discharge rate parameters, the operating condition classification characteristics corresponding to the state transition characteristics are extracted to construct an operating condition characteristic matrix; Performing state migration probability statistics and state space construction on the operating condition characteristic matrix to obtain a state space transfer characteristic matrix, and performing path planning and multi-path combination of battery pack state evolution on the state space transfer characteristic matrix based on preset battery pack operating condition parameters to obtain at least one initial battery pack evolution path; Based on a preset battery pack capacity attenuation termination threshold, life attenuation calculation and threshold interval classification are performed on the initial battery pack evolution path to obtain an initial life prediction path, and state change interval calculation and mean value calculation of multi-state monitoring parameters are performed on the initial life prediction path to obtain a state prediction interval; Based on the corresponding number of charge and discharge cycles in the historical operation data, life feature mapping and confidence calculation are performed on the state prediction interval to obtain state prediction confidence, and weight allocation and weighted fusion are performed on the state prediction confidence to generate a state prediction result of the target intelligent battery pack.

7. A state monitoring and prediction device for an intelligent battery pack, characterized in that: The state monitoring and prediction device of the intelligent battery pack includes: A preprocessing module, used to obtain raw battery pack data and historical operation data of multiple preset battery pack parameters in the target intelligent battery pack, and perform signal filtering and spectrum decomposition on the raw battery pack data to obtain initial battery pack characteristic data; A spatial mapping module is used to perform abnormal data detection and abnormal correction on the initial battery pack characteristic data to obtain a standard battery pack characteristic sequence of the target intelligent battery pack, and to divide the corresponding voltage and current characteristic data in the standard battery pack characteristic sequence into charge and discharge cycles to obtain an initial battery pack working mode, and to classify the corresponding temperature and impedance characteristic data in the initial battery pack working mode into multiple charge and discharge rates to obtain a multi-dimensional battery pack modal component; extract the time-frequency characteristics of each of the battery pack modal components to obtain the time-frequency response characteristics of the target intelligent battery pack, and classify the time-frequency response The energy distribution of the characteristics is calculated to obtain the battery pack working condition energy spectrum; the charge and discharge state characteristics are extracted from the corresponding voltage and current characteristic data in the battery pack working condition energy spectrum to obtain the charge and discharge state characteristics, and the temperature and impedance characteristic data corresponding to the battery pack working condition energy spectrum are mapped to the working state to obtain the temperature impedance state characteristics, and the charge and discharge state characteristics and the temperature impedance state characteristics are combined to obtain a multi-dimensional state sequence of the battery pack; the state space reconstruction of the battery pack state and the evolution of the battery pack state are performed on the multi-dimensional state sequence of the battery pack to obtain the battery pack state evolution characteristics; A state monitoring module, used for extracting the state of charge dynamic features corresponding to the battery pack state evolution features, and performing multi-layer feature mapping and state threshold classification on the state of charge dynamic features based on the historical operation data to obtain state monitoring results of multiple state monitoring parameters; A state prediction module is used to perform coupling mapping of multi-state monitoring parameters and conversion probability calculation of battery pack states on the state monitoring results to obtain a state space transfer feature matrix, and based on the historical operating data and preset battery pack operating parameters, calculate the state evolution path of the state space transfer feature matrix and the associated confidence calculation of the corresponding battery pack life characteristics to generate a state prediction result of the target intelligent battery pack.

8. A state monitoring and prediction device for an intelligent battery pack, characterized in that: The state monitoring and prediction device of the intelligent battery pack includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instructions in the memory to enable the state monitoring and prediction device of the intelligent battery pack to execute each step of the state monitoring and prediction method of the intelligent battery pack as described in any one of claims 1-6.

9. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the various steps of the state monitoring and prediction method of the intelligent battery pack as described in any one of claims 1 to 6 are implemented.

Images