A lithium ion battery energy storage system fault diagnosis method and system

Abstract

This invention relates to the field of lithium-ion battery energy storage technology, specifically disclosing a fault diagnosis method and system for lithium-ion battery energy storage systems. Through a multi-level monitoring logic of cluster-module-cell, the fault range is quickly narrowed down before focusing on individual cell diagnosis, avoiding indiscriminate full-scale calculations and achieving accurate fault location and analysis for large-scale energy storage systems. Furthermore, through the collaborative calculation of spatial and temporal information entropy, real-time location of sudden faults and early warning of gradual faults are simultaneously achieved, effectively solving the problems of strong adaptability to single faults and weak collaborative diagnosis capabilities for multiple faults in existing technologies.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of lithium ion battery energy storage, in particular to a lithium ion battery energy storage system fault diagnosis method and system. BACKGROUND

[0002] At present, lithium ion battery energy storage systems are widely used in energy consumption, grid peak regulation and other scenarios due to their high energy density, long cycle life and other advantages. However, in the long-term operation process, lithium ion batteries are prone to internal short circuit, lithium precipitation, capacity degradation and other faults, which seriously affect the safety and reliability of the energy storage system. Therefore, efficient and accurate fault diagnosis is the key to ensuring the stable operation of the energy storage system.

[0003] In the prior art, lithium ion battery energy storage system fault diagnosis technology is mostly based on information entropy and anomaly detection algorithm, among which some technologies use a hierarchical structure of cluster-module-cell to locate the fault area by calculating the spatial information entropy inconsistency, or implement fault diagnosis by combining time information entropy.

[0004] However, in the prior art, only the fault diagnosis of lithium ion batteries can be realized, and the fault evolution path cannot be traced, which is not conducive to the accurate disposal of maintenance personnel. SUMMARY

[0005] The purpose of the present application is to provide a lithium ion battery energy storage system fault diagnosis method and system, which aims to solve the technical problem in the prior art that only the fault diagnosis of lithium ion batteries can be realized, and the fault evolution path cannot be traced, which is not conducive to the accurate disposal of maintenance personnel.

[0006] To achieve the above purpose, a lithium ion battery energy storage system fault diagnosis method is adopted, which comprises the following steps: Collecting the electrical signals and thermal signals of the lithium ion battery energy storage system, the electrical signals being voltage, current and SOC, and the thermal signals being battery temperature and ambient temperature, synchronously acquiring the historical data of the whole life cycle of the lithium ion battery, the historical data including charge-discharge rate, cycle number and fault record; Based on the second-order RC equivalent circuit model and the lumped parameter thermal model, the battery digital twin is constructed by fusing the charge-discharge rate, cycle number and fault record, and the parameters of the battery digital twin are dynamically updated to adapt to the changes in working conditions through the deviation feedback of the real-time collected data of the battery and the simulation data of the battery digital twin; Monitoring the cluster-level spatial information entropy and module-level spatial information entropy inconsistency; Performing time information entropy fault diagnosis on all battery cells inside the battery module that exceed the threshold value; By performing deviation analysis between the information entropy data of the physical battery and the simulation data of the digital twin, the fault diagnosis results are corrected. The fault evolution path is traced back through the digital twin, and the fault type, fault location and fault development trend of the lithium-ion battery energy storage system are output.

[0007] The construction of the battery digital twin includes the following steps: Construct a second-order RC equivalent circuit model, the formula is: In the formula, For ohmic internal resistance, For loop current, , These are the polarization voltages of the two RC branches, respectively. This is the mapping function between open-circuit voltage and state of charge (SOC). Construct a lumped parameter thermal model, the formula is: In the formula, The equivalent heat capacity of the battery. For equivalent thermal resistance, For battery temperature, For ambient temperature, The rate at which the battery generates heat; Based on the mean square error of real-time electrical signals, thermal signals, and battery digital twin simulation data, the model parameters of the second-order RC equivalent circuit model and the lumped parameter thermal model are dynamically updated.

[0008] The monitoring of inconsistency between cluster-level spatial information entropy and module-level spatial information entropy includes the following steps: Calculate the first spatial information entropy of each battery cluster at the q-th acquisition time; The difference between the first spatial information entropy and the maximum spatial information entropy at the cluster level is the cluster-level spatial information entropy inconsistency. If the cluster-level spatial information entropy inconsistency exceeds the cluster-level threshold, a cluster-level alarm is triggered. After a cluster-level alarm is triggered, the second spatial information entropy of each battery module within that cluster is calculated. The difference between the second spatial information entropy and the module-level maximum spatial information entropy is used to obtain the module-level spatial information entropy inconsistency. If the module-level spatial information entropy inconsistency exceeds the module-level threshold, a module-level alarm is triggered.

[0009] Among them, for battery modules that exceed the threshold, the time information entropy fault diagnosis algorithm is executed on all battery cells inside them; if no battery module exceeds the threshold, 1 to 3 battery modules are randomly selected to execute the time information entropy fault diagnosis algorithm. The time-information entropy fault diagnosis algorithm includes the following steps: For each battery cell in the selected battery module, the voltage, current, and temperature of the most recent W acquisition times are extracted through a sliding window to form a time-series data sequence of the cell. For each individual cell, the time-series data sequence is divided into intervals and frequency statistics are performed to calculate the time information entropy of that individual cell. For any two battery cells within the same module, calculate the Pearson correlation coefficient of their time information entropy. Calculate the average Pearson correlation coefficient between each battery cell and all other battery cells in its battery module; The average Pearson correlation coefficient of each battery cell is compared with a preset threshold. If the average Pearson correlation coefficient is lower than the preset threshold, the battery cell is determined to be a faulty cell.

[0010] Among them, the first spatial information entropy The calculation formula is: In the formula, Let n be the first spatial information entropy of the k-th battery cluster at the q-th acquisition time; n is the battery cell number; m is the battery module number; k is the battery cluster number; and q is the acquisition time number. This represents the operational data of the k-th battery cluster, m-th module, and n-th cell at time q. This operational data includes any one of voltage, current, and temperature. N represents the total number of battery cells in a single module. M represents the total number of battery modules in a single battery cluster. X represents the operational data type identifier, which includes any one of voltage, current, and temperature. Cluster-level spatial information entropy inconsistency It is the difference between the first spatial information entropy and the maximum spatial information entropy; The second spatial information entropy The calculation formula is: In the formula, Let n be the second spatial information entropy of the k-th battery cluster and the m-th module at the q-th acquisition time; n is the battery cell number; m is the battery module number; k is the battery cluster number; and q is the acquisition time number. This represents the operational data of the k-th battery cluster, m-th module, and n-th cell at time q. This operational data includes any one of voltage, current, and temperature. N represents the total number of battery cells in a single module. M represents the total number of battery modules in a single battery cluster. X represents the operational data type identifier, which includes any one of voltage, current, and temperature. Module-level spatial information entropy inconsistency It is the difference between the second spatial information entropy and the maximum spatial information entropy.

[0011] Among them, the time information entropy The calculation formula is: In the formula, Let be the temporal entropy of the k-th battery cluster, the m-th module, and the n-th battery cell at the q-th acquisition time; n is the battery cell number; m is the battery module number; k is the battery cluster number; q is the acquisition time number; X is the running data type identifier, including any one of voltage, current, and temperature; i is the interval number; D is the vector dimension after the reconstructed single-cell running data. Let i be the eigenvalue of the interval after reconstructing the data of the k-th battery cluster, the m-th module, and the n-th cell at time q. for The proportion of the total sum of all interval eigenvalues ​​of this single entity at time q.

[0012] The step of performing deviation analysis between the information entropy data of the physical battery and the simulation data of the digital twin to correct the fault diagnosis results is as follows: In the formula, This is the deviation correction factor; The average Pearson correlation coefficient of the physical data; The average Pearson correlation coefficient of the digital twin simulation data.

[0013] The cluster-level threshold and the module-level threshold are obtained by statistical analysis of historical fault data, and the cluster-level threshold and the module-level threshold are dynamically adjusted according to the number of cycles of the battery cell.

[0014] The digital twin's fault evolution path tracing includes parameter shifts in the fault initiation stage, characteristic mutations in the development stage, and state degradation in the outbreak stage.

[0015] The present invention also provides a fault diagnosis system for a lithium-ion battery energy storage system, used to perform the fault diagnosis method for a lithium-ion battery energy storage system as described above, including: Data acquisition module: used to collect electrical signals, thermal signals, and historical operating data of lithium-ion battery energy storage systems; Digital twin modeling module: used to construct second-order RC equivalent circuit models and lumped parameter thermal models, and dynamically update model parameters; Hierarchical monitoring module: used to calculate the inconsistency of spatial information entropy between cluster level and module level, and trigger hierarchical alarms; Individual Unit Diagnostic Module: Used to execute the time information entropy fault diagnosis algorithm to identify faulty individual units; Fusion and tracing module: used to fuse and verify virtual and real data, and trace the evolution path of faults; Memory: Used to store program instructions for fault diagnosis; Processor: Used to invoke program instructions in the memory to execute the fault diagnosis method.

[0016] The present invention provides a fault diagnosis method and system for a lithium-ion battery energy storage system, which has the following beneficial effects: 1. By coordinating the calculation of spatial information entropy and temporal information entropy, the system can simultaneously achieve real-time location of sudden faults and early warning of gradual faults, effectively solving the problem that existing technologies have strong adaptability to single faults but weak collaborative diagnosis capabilities for multiple faults. 2. By combining digital twins to achieve virtual and real data fusion verification, diagnostic parameters can be dynamically adjusted through deviation correction coefficients, which can adapt to dynamic operating conditions such as voltage, current, SOC, battery temperature, ambient temperature, charge and discharge rate, cycle number, and fault records. 3. By tracing the fault evolution path through digital twins, diagnosis and source tracing are integrated, providing operation and maintenance personnel with accurate root cause analysis of faults, which solves the limitation of existing technologies that can only diagnose but not trace the source. 4. Through multi-level monitoring logic of cluster-module-individual, the fault range is quickly narrowed down first, and then the focus is on individual diagnosis, avoiding indiscriminate full calculation, and realizing accurate fault location and analysis of large-scale energy storage systems. Detailed Implementation

[0017] This invention provides a fault diagnosis method for a lithium-ion battery energy storage system, comprising the following steps: The system collects electrical and thermal signals from the lithium-ion battery energy storage system. The electrical signals include voltage, current, and SOC, while the thermal signals include battery temperature and ambient temperature. The system also simultaneously acquires historical data of the lithium-ion battery's entire life cycle, including charge / discharge rate, cycle count, and fault records. Based on the second-order RC equivalent circuit model and the lumped parameter thermal model, a battery digital twin is constructed by integrating charge and discharge rate, cycle number, and fault records. The battery digital twin parameters are dynamically updated to adapt to changes in operating conditions by feedback on the deviation between the real-time acquired data of the battery and the simulation data of the battery digital twin. Monitor the inconsistency between cluster-level spatial information entropy and module-level spatial information entropy; Perform time information entropy fault diagnosis on all individual cells within the battery module that exceed the threshold; By performing deviation analysis between the information entropy data of the physical battery and the simulation data of the digital twin, the fault diagnosis results are corrected. The fault evolution path is traced back through the digital twin, and the fault type, fault location and fault development trend of the lithium-ion battery energy storage system are output.

[0018] Furthermore, the construction of the battery digital twin includes the following steps: Construct a second-order RC equivalent circuit model, the formula is: In the formula, For ohmic internal resistance, For loop current, , These are the polarization voltages of the two RC branches, respectively. This is the mapping function between open-circuit voltage and state of charge (SOC). Construct a lumped parameter thermal model, the formula is: In the formula, The equivalent heat capacity of the battery. For equivalent thermal resistance, For battery temperature, For ambient temperature, The rate at which the battery generates heat; Based on the mean square error of real-time electrical signals, thermal signals, and battery digital twin simulation data, the model parameters of the second-order RC equivalent circuit model and the lumped parameter thermal model are dynamically updated.

[0019] Furthermore, the monitoring of inconsistency between cluster-level spatial information entropy and module-level spatial information entropy includes the following steps: Calculate the first spatial information entropy of each battery cluster at the q-th acquisition time; The difference between the first spatial information entropy and the maximum spatial information entropy at the cluster level is the cluster-level spatial information entropy inconsistency. If the cluster-level spatial information entropy inconsistency exceeds the cluster-level threshold, a cluster-level alarm is triggered. After a cluster-level alarm is triggered, the second spatial information entropy of each battery module within that cluster is calculated. The difference between the second spatial information entropy and the module-level maximum spatial information entropy is used to obtain the module-level spatial information entropy inconsistency. If the module-level spatial information entropy inconsistency exceeds the module-level threshold, a module-level alarm is triggered.

[0020] Furthermore, for battery modules that exceed the threshold, the time information entropy fault diagnosis algorithm is executed on all battery cells within them; if no battery module exceeds the threshold, 1 to 3 battery modules are randomly selected to execute the time information entropy fault diagnosis algorithm. The time-information entropy fault diagnosis algorithm includes the following steps: For each battery cell in the selected battery module, the voltage, current, and temperature of the most recent W acquisition times are extracted through a sliding window to form a time-series data sequence of the cell. For each individual cell, the time-series data sequence is divided into intervals and frequency statistics are performed to calculate the time information entropy of that individual cell. For any two battery cells within the same module, calculate the Pearson correlation coefficient of their time information entropy. Calculate the average Pearson correlation coefficient between each battery cell and all other battery cells in its battery module; The average Pearson correlation coefficient of each battery cell is compared with a preset threshold. If the average Pearson correlation coefficient is lower than the preset threshold, the battery cell is determined to be a faulty cell.

[0021] Furthermore, the first spatial information entropy The calculation formula is: In the formula, Let n be the first spatial information entropy of the k-th battery cluster at the q-th acquisition time; n is the battery cell number; m is the battery module number; k is the battery cluster number; and q is the acquisition time number. This represents the operational data of the k-th battery cluster, m-th module, and n-th cell at time q. This operational data includes any one of voltage, current, and temperature. N represents the total number of battery cells in a single module. M represents the total number of battery modules in a single battery cluster. X represents the operational data type identifier, which includes any one of voltage, current, and temperature. Cluster-level spatial information entropy inconsistency It is the difference between the first spatial information entropy and the maximum spatial information entropy; The second spatial information entropy The calculation formula is: In the formula, Let n be the second spatial information entropy of the k-th battery cluster and the m-th module at the q-th acquisition time; n is the battery cell number; m is the battery module number; k is the battery cluster number; and q is the acquisition time number. This represents the operational data of the k-th battery cluster, m-th module, and n-th cell at time q. This operational data includes any one of voltage, current, and temperature. N represents the total number of battery cells in a single module. M represents the total number of battery modules in a single battery cluster. X represents the operational data type identifier, which includes any one of voltage, current, and temperature. Module-level spatial information entropy inconsistency It is the difference between the second spatial information entropy and the maximum spatial information entropy.

[0022] Furthermore, the temporal information entropy The calculation formula is: In the formula, Let be the temporal entropy of the k-th battery cluster, the m-th module, and the n-th battery cell at the q-th acquisition time; n is the battery cell number; m is the battery module number; k is the battery cluster number; q is the acquisition time number; X is the running data type identifier, including any one of voltage, current, and temperature; i is the interval number; D is the vector dimension after the reconstructed single-cell running data. Let i be the eigenvalue of the interval after reconstructing the data of the k-th battery cluster, the m-th module, and the n-th cell at time q. for The proportion of the total sum of all interval eigenvalues ​​of this single entity at time q.

[0023] Furthermore, the deviation analysis is performed between the information entropy data of the physical battery and the simulation data of the digital twin to correct the fault diagnosis results. The deviation correction formula is as follows: In the formula, This is the deviation correction factor; The average Pearson correlation coefficient of the physical data; The average Pearson correlation coefficient of the digital twin simulation data.

[0024] Furthermore, the cluster-level threshold and the module-level threshold are obtained by statistical analysis of historical fault data, and the cluster-level threshold and the module-level threshold are dynamically adjusted according to the number of cycles of the battery cell.

[0025] Furthermore, the digital twin traces the fault evolution path, including parameter shifts in the fault initiation stage, characteristic mutations in the development stage, and state degradation in the outbreak stage.

[0026] The present invention also provides a fault diagnosis system for a lithium-ion battery energy storage system, used to perform the fault diagnosis method for a lithium-ion battery energy storage system as described above, including: Data acquisition module: used to collect electrical signals, thermal signals, and historical operating data of lithium-ion battery energy storage systems; Digital twin modeling module: used to construct second-order RC equivalent circuit models and lumped parameter thermal models, and dynamically update model parameters; Hierarchical monitoring module: used to calculate the inconsistency of spatial information entropy between cluster level and module level, and trigger hierarchical alarms; Individual Unit Diagnostic Module: Used to execute the time information entropy fault diagnosis algorithm to identify faulty individual units; Fusion and tracing module: used to fuse and verify virtual and real data, and trace the evolution path of faults; Memory: Used to store program instructions for fault diagnosis; Processor: Used to invoke program instructions in the memory to execute the fault diagnosis method.

[0027] In this embodiment, The second-order RC equivalent circuit model includes the ohmic internal resistance. Polarization resistance Polarization resistance Polarized capacitors Polarization resistance A state transition optimization algorithm is employed, with the mean square error between the output voltage and the measured voltage of the second-order RC equivalent circuit model as the objective function. In the formula, Here is the parameter vector; Q is the number of sampling points; The measured terminal voltage at time q; The terminal voltage is simulated using a second-order RC equivalent circuit model; the initial parameters are obtained by minimizing the objective function. , , , , Construct a second-order RC equivalent circuit model, the formula is: ; The lumped-parameter thermal model defines battery heat generation as including irreversible and reversible heat, with a total heat generation rate of: In the formula, Irreversible heat; It is a reversible heat; Let the open-circuit voltage change with temperature be -0.4 mV / ℃. The lumped-parameter thermal model formula is: In the formula, ; ; The electrical parameters at temperature T are: In the formula, , For reference temperature; =Parameters at 25℃; To activate energy ( The corresponding activation energy is 3000 J / mol); R is the gas constant (8.314 J / (mol·K)); based on the mean square error of the real-time acquired electrical signals, thermal signals and simulation data, the lumped parameter thermal model parameters are updated every 10 sampling periods.

[0028] In the first spatial information entropy, For the first probability, N=16 (number of individuals per module), M=20 (number of modules per cluster). The maximum spatial information entropy at the cluster level is: Cluster-level spatial information entropy inconsistency By analyzing 50 sets of historical fault data, a cluster-level threshold was set. ,when Trigger cluster-level alarms at any time; After a cluster-level alarm is triggered, the second spatial information entropy of each battery module within that cluster is calculated using the following formula: In the formula, The second probability; The maximum spatial information entropy at the module level is: Module-level spatial information entropy inconsistency Set module-level thresholds ,when Module-level alarms are triggered at certain times.

[0029] For modules that trigger module-level alarms, a time-information entropy fault diagnosis algorithm is executed on all internal battery cells. A sliding window is used to extract the voltage data of each battery cell at the most recent W=20 time interval, which is then reconstructed into a D=5 dimensional vector by dividing the data into intervals. The interval boundaries are: In the formula, , ; The formula for calculating time information entropy is: In the formula, These are the interval feature values ​​reconstructed from the single-unit running data; For interval probabilities; Calculate the temporal information entropy correlation coefficient between individual unit n and individual unit l within the same module: In the formula, For covariance; Standard deviation; Calculate the average Pearson correlation coefficient for each battery cell: Set a preset threshold ,when When this happens, the individual unit is determined to be a faulty unit.

[0030] Deviation correction factor The calculation formula is as follows:

[0031] In the formula, The information entropy matrix of the physical battery; The information entropy simulation matrix for a digital twin; It is the F-norm; The corrected average Pearson correlation coefficient was obtained through weighted fusion: In the formula, The average Pearson correlation coefficient value of the digital twin simulation data, after correction. If so, the battery cell is confirmed to be faulty.

[0032] By tracing the fault evolution path using a digital twin, taking an internal short-circuit fault as an example, the tracing process is as follows: During the initial stage of the fault (sampling cycles 50-80), the internal resistance of the battery cell drifted from 0.005Ω to 0.007Ω, and the correlation coefficient of time information entropy decreased from 0.95 to 0.9. During the development stage (sampling cycles 81-120), the ohmic internal resistance of the battery cell continued to drift to 0.01Ω, the temporal information entropy correlation coefficient continued to decrease to 0.85, and the module-level spatial information entropy inconsistency broke through 0.2. During the outbreak phase (after the 121st sampling period), the ohmic internal resistance of a single battery cell suddenly dropped to 0.02Ω, and the time information entropy correlation coefficient continued to decrease to 0.75, triggering a cluster-level alarm.

[0033] The output fault type is internal short circuit, the location is the 8th cell of the 12th module in the 3rd cluster, and the development trend is continuous deterioration, requiring immediate shutdown and maintenance.

[0034] The above description discloses only one preferred embodiment of the present invention, and should not be construed as limiting the scope of the present invention. Those skilled in the art will understand that all or part of the processes of the above embodiments can be implemented, and equivalent changes made in accordance with the claims of the present invention are still within the scope of the invention.

Claims

1. A fault diagnosis method for a lithium-ion battery energy storage system, characterized in that, Includes the following steps: The system collects electrical and thermal signals from the lithium-ion battery energy storage system. The electrical signals include voltage, current, and SOC, while the thermal signals include battery temperature and ambient temperature. The system also simultaneously acquires historical data of the lithium-ion battery's entire life cycle, including charge / discharge rate, cycle count, and fault records. Based on the second-order RC equivalent circuit model and the lumped parameter thermal model, a battery digital twin is constructed by integrating charge and discharge rate, cycle number, and fault records. The battery digital twin parameters are dynamically updated to adapt to changes in operating conditions by feedback on the deviation between the real-time acquired data of the battery and the simulation data of the battery digital twin. Monitor the inconsistency between cluster-level spatial information entropy and module-level spatial information entropy; Perform time information entropy fault diagnosis on all individual cells within the battery module that exceed the threshold; By performing deviation analysis between the information entropy data of the physical battery and the simulation data of the digital twin, the fault diagnosis results are corrected. The fault evolution path is traced back through the digital twin, and the fault type, fault location and fault development trend of the lithium-ion battery energy storage system are output.

2. The fault diagnosis method for a lithium-ion battery energy storage system as described in claim 1, wherein the construction of a battery digital twin is characterized in that, Includes the following steps: Construct a second-order RC equivalent circuit model, the formula is: In the formula, For ohmic internal resistance, For loop current, , These are the polarization voltages of the two RC branches, respectively. This is the mapping function between open-circuit voltage and state of charge (SOC). Construct a lumped parameter thermal model, the formula is: In the formula, The equivalent heat capacity of the battery. For equivalent thermal resistance, For battery temperature, For ambient temperature, The rate at which the battery generates heat; Based on the mean square error of real-time electrical signals, thermal signals, and battery digital twin simulation data, the model parameters of the second-order RC equivalent circuit model and the lumped parameter thermal model are dynamically updated.

3. The fault diagnosis method for a lithium-ion battery energy storage system as described in claim 1, wherein the monitoring of the inconsistency between cluster-level spatial information entropy and module-level spatial information entropy is characterized in that, Includes the following steps: Calculate the first spatial information entropy of each battery cluster at the q-th acquisition time; The difference between the first spatial information entropy and the maximum spatial information entropy at the cluster level is the cluster-level spatial information entropy inconsistency. If the cluster-level spatial information entropy inconsistency exceeds the cluster-level threshold, a cluster-level alarm is triggered. After a cluster-level alarm is triggered, the second spatial information entropy of each battery module within that cluster is calculated. The difference between the second spatial information entropy and the module-level maximum spatial information entropy is used to obtain the module-level spatial information entropy inconsistency. If the module-level spatial information entropy inconsistency exceeds the module-level threshold, a module-level alarm is triggered.

4. The fault diagnosis method for a lithium-ion battery energy storage system as described in claim 1, wherein performing time information entropy fault diagnosis on all individual battery cells within a battery module exceeding a threshold is characterized in that, If a battery module exceeds the threshold, the time information entropy fault diagnosis algorithm is executed on all its internal battery cells. If no battery module exceeds the threshold, 1 to 3 battery modules are randomly selected to execute the time information entropy fault diagnosis algorithm. The time-information entropy fault diagnosis algorithm includes the following steps: For each battery cell in the selected battery module, the voltage, current, and temperature of the most recent W acquisition times are extracted through a sliding window to form a time-series data sequence of the cell. For each individual cell, the time-series data sequence is divided into intervals and frequency statistics are performed to calculate the time information entropy of that individual cell. For any two battery cells within the same module, calculate the Pearson correlation coefficient of their time information entropy. Calculate the average Pearson correlation coefficient between each battery cell and all other battery cells in its battery module; The average Pearson correlation coefficient of each battery cell is compared with a preset threshold. If the average Pearson correlation coefficient is lower than the preset threshold, the battery cell is determined to be a faulty cell.

5. The fault diagnosis method for a lithium-ion battery energy storage system as described in claim 3, characterized in that, The first spatial information entropy The calculation formula is: In the formula, Let n be the first spatial information entropy of the k-th battery cluster at the q-th acquisition time; n is the battery cell number; m is the battery module number; k is the battery cluster number; and q is the acquisition time number. This represents the operational data of the k-th battery cluster, m-th module, and n-th cell at time q. This operational data includes any one of voltage, current, and temperature. N represents the total number of battery cells in a single module. M represents the total number of battery modules in a single battery cluster. X represents the operational data type identifier, which includes any one of voltage, current, and temperature. Cluster-level spatial information entropy inconsistency It is the difference between the first spatial information entropy and the maximum spatial information entropy; The second spatial information entropy The calculation formula is: In the formula, Let n be the second spatial information entropy of the k-th battery cluster and the m-th module at the q-th acquisition time; n is the battery cell number; m is the battery module number; k is the battery cluster number; and q is the acquisition time number. This represents the operational data of the k-th battery cluster, m-th module, and n-th cell at time q. This operational data includes any one of voltage, current, and temperature. N represents the total number of battery cells in a single module. M represents the total number of battery modules in a single battery cluster. X represents the operational data type identifier, which includes any one of voltage, current, and temperature. Module-level spatial information entropy inconsistency It is the difference between the second spatial information entropy and the maximum spatial information entropy.

6. The fault diagnosis method for a lithium-ion battery energy storage system as described in claim 4, characterized in that, The time information entropy The calculation formula is: In the formula, Let be the temporal entropy of the k-th battery cluster, the m-th module, and the n-th battery cell at the q-th acquisition time; n is the battery cell number; m is the battery module number; k is the battery cluster number; q is the acquisition time number; X is the running data type identifier, including any one of voltage, current, and temperature; i is the interval number; D is the vector dimension after the reconstructed single-cell running data. Let i be the eigenvalue of the interval after reconstructing the data of the k-th battery cluster, the m-th module, and the n-th cell at time q. for The proportion of the total sum of all interval eigenvalues ​​of this single entity at time q.

7. The fault diagnosis method for a lithium-ion battery energy storage system as described in claim 1, characterized in that, The deviation analysis is performed between the information entropy data of the physical battery and the simulation data of the digital twin to correct the fault diagnosis results. The deviation correction formula is as follows: In the formula, This is the deviation correction factor; The average Pearson correlation coefficient of the physical data; The average Pearson correlation coefficient of the digital twin simulation data.

8. The fault diagnosis method for a lithium-ion battery energy storage system as described in claim 3, characterized in that, The cluster-level threshold and the module-level threshold are obtained by statistical analysis of historical fault data, and the cluster-level threshold and the module-level threshold are dynamically adjusted according to the number of cycles of the battery cell.

9. The fault diagnosis method for a lithium-ion battery energy storage system as described in claim 1, characterized in that, The digital twin traces the fault evolution path, including parameter shifts in the fault initiation stage, feature mutations in the development stage, and state degradation in the outbreak stage.

10. A fault diagnosis system for a lithium-ion battery energy storage system, used to execute the fault diagnosis method for a lithium-ion battery energy storage system as described in any one of claims 1 to 9, characterized in that, include: Data acquisition module: used to collect electrical signals, thermal signals, and historical operating data of lithium-ion battery energy storage systems; Digital twin modeling module: used to construct second-order RC equivalent circuit models and lumped parameter thermal models, and dynamically update model parameters; Hierarchical monitoring module: used to calculate the inconsistency of spatial information entropy between cluster level and module level, and trigger hierarchical alarms; Individual Unit Diagnostic Module: Used to execute the time information entropy fault diagnosis algorithm to identify faulty individual units; Fusion and tracing module: used to fuse and verify virtual and real data, and trace the evolution path of faults; Memory: Used to store program instructions for fault diagnosis; Processor: Used to invoke program instructions in the memory to execute the fault diagnosis method.

Images