Abnormal battery box detection method and system for energy storage power station

Abstract

The present invention relates to the field of battery box detection technology, and in particular to a method and system for detecting abnormal battery boxes in energy storage power stations. The method comprises the following steps: obtaining battery box data and operating environment data of the energy storage power station, evaluating the initial performance of the battery box and generating operating status data of the energy storage power station accordingly; based on the operating status data, detecting the imbalance of the supporting configuration evolution, and then identifying the steady-state collapse of the battery electrothermal coupling, and judging the abnormal trend of the power quality accordingly; based on the abnormal trend of the power quality, identifying the attenuation trend of the coordinated operation of the power grid, and further detecting the short-term harmonic surge condition, and finally determining the attenuation level of the power grid stability; combining the attenuation condition of the power grid stability with the change of the battery state, identifying the degree of instability of the power grid frequency, and detecting the abnormal condition of the battery in the energy storage power station, and uploading the abnormal information to the cloud platform to realize early warning response; the present invention realizes more accurate identification of battery box abnormalities through the detection of battery box abnormalities.

Description

Technical Field

[0001] The present invention relates to the technical field of battery box detection, and in particular to a method and system for detecting abnormal battery boxes in an energy storage power station. Background Art

[0002] Energy storage power stations generally utilize lithium-ion battery systems for energy storage. The battery box, serving as the basic operating unit, contains multiple cells, conductive connections, temperature control components, and electrical protection devices, responsible for energy charge and discharge conversion and multi-node parallel operation. However, during long-term operation, due to factors such as frequent charge and discharge cycles, electrochemical material aging, ambient temperature fluctuations, loose electrical connections, and external vibration and impact, the battery box is prone to abnormal capacity decay, abnormal voltage and current fluctuations, overheating, cell swelling and leakage, or internal short circuits. If these anomalies are not detected and addressed promptly, they can lead to decreased system efficiency, increased energy loss, and even thermal runaway and fires, seriously threatening the safe operation of the energy storage power station. However, traditional energy storage power station anomaly detection methods suffer from inaccurate detection of abnormal trends in the power quality transmitted by the energy storage station and inaccurate detection of grid stability degradation conditions. Summary of the Invention

[0003] Based on this, it is necessary to provide a method and system for detecting abnormal battery boxes in an energy storage power station to solve at least one of the above technical problems.

[0004] To achieve the above objectives, a method for detecting abnormal battery boxes in energy storage power stations includes the following steps:

[0005] Step S1: Acquire battery box data of an energy storage power station; collect operating environment data of the energy storage power station based on the battery box data of the energy storage power station; evaluate the initial performance of the battery box of the energy storage power station based on the battery box data of the energy storage power station; evaluate the operating status data of the energy storage power station based on the initial performance of the battery box of the energy storage power station based on the battery box data of the energy storage power station and the operating environment data of the energy storage power station;

[0006] Step S2: detecting an imbalance in the evolution of the supporting configuration of the energy storage station based on the operating status data of the energy storage station; detecting a steady-state collapse of the battery electrothermal coupling based on the imbalance in the evolution of the supporting configuration of the energy storage station; and detecting an abnormal trend in the quality of power transmitted by the energy storage station based on the steady-state collapse of the battery electrothermal coupling.

[0007] Step S3: Identifying a power grid operation coordination attenuation trend based on the abnormal power quality trend transmitted by the energy storage station; detecting a short-term harmonic surge in the energy storage station's transmission based on the power grid operation coordination attenuation trend and the abnormal power quality trend transmitted by the energy storage station; and determining a power grid stability attenuation condition of the energy storage station based on the short-term harmonic surge in the energy storage station's transmission.

[0008] Step S4: Detect changes in the battery status of the energy storage station based on the attenuation of the stability of the energy storage station grid; identify the degree of grid frequency instability based on the changes in the battery status of the energy storage station; detect abnormal conditions of the energy storage station battery based on the degree of grid frequency instability and the changes in the battery status of the energy storage station, and upload them to the cloud platform for early warning.

[0009] The present invention achieves accurate identification of abnormal battery status evolution through layer-by-layer analysis from the basic performance of the battery box to the system power quality transmission process. Based on the acquisition of energy storage battery box data and operating environment data, this method ensures that the starting point of the analysis process has high credibility through initial performance modeling and operating status fusion evaluation; in the support configuration and electrothermal coupling state evolution detection, the dual stability changes of structure and thermal coupling are combined to enhance the recognition ability of complex physical interaction effects; for the power quality abnormal trend and short-term harmonic surge detection during the transmission process, a logical chain from data disturbance response to stability quantification is constructed, effectively improving the dynamic tracking ability of power grid stability attenuation; further, in the process of state change and frequency disturbance identification, a closed-loop analysis mechanism of linkage response between energy storage system and power grid is established, and the abnormal battery condition is jointly detected by combining frequency fluctuation and battery state evolution, and reported through the cloud platform to build a complete early warning system driven by data and real-time perception. The overall solution has the characteristics of wide coverage, clear data link and accurate abnormality tracing. Therefore, the present invention optimizes the abnormal battery boxes of traditional energy storage power stations, solving the problems of inaccurate detection of abnormal trends in the quality of power transmitted by the energy storage power station and inaccurate detection of the attenuation of the stability of the power grid of the energy storage power station by the abnormal battery boxes of traditional energy storage power stations, thereby improving the accuracy of detecting abnormal trends in the quality of power transmitted by the energy storage power station and the accuracy of detecting the attenuation of the stability of the power grid of the energy storage power station.

[0010] The present invention also provides an abnormal battery box detection system for an energy storage power station, which is used to execute the abnormal battery box detection method for an energy storage power station as described above. The abnormal battery box detection system for an energy storage power station comprises:

[0011] The energy storage power station operation status evaluation module is used to obtain the energy storage power station battery box data; collect the energy storage power station operation environment data based on the energy storage power station battery box data; evaluate the energy storage power station battery box initial performance based on the energy storage power station battery box data; and evaluate the energy storage power station operation status data based on the energy storage power station battery box initial performance based on the energy storage power station battery box data and the energy storage power station operation environment data;

[0012] The transmission power quality anomaly detection module is used to detect the imbalance of the supporting configuration evolution of the energy storage station based on the operating status data of the energy storage station; detect the steady-state collapse of the battery electrothermal coupling based on the imbalance of the supporting configuration evolution of the energy storage station; and detect the abnormal trend of the transmission power quality of the energy storage station based on the steady-state collapse of the battery electrothermal coupling;

[0013] The grid stability attenuation determination module is used to identify the grid operation coordination attenuation trend based on the abnormal power quality trend of the energy storage station transmission power; detect the short-term harmonic surge condition of the energy storage station transmission based on the grid operation coordination attenuation trend and the abnormal power quality trend of the energy storage station transmission power; and determine the grid stability attenuation condition of the energy storage station based on the short-term harmonic surge condition of the energy storage station transmission power;

[0014] The energy storage power station battery anomaly detection module is used to detect changes in the battery status of the energy storage power station based on the attenuation of the energy storage power station grid stability; identify the degree of grid frequency instability based on the changes in the battery status of the energy storage power station; detect abnormal conditions of the energy storage power station battery based on the degree of grid frequency instability and changes in the battery status of the energy storage power station, and upload them to the cloud platform for early warning.

[0015] The energy storage power station abnormal battery box detection system of the present invention can implement the abnormal battery box detection method of any energy storage power station of the present invention. It is used to combine the operation and signal transmission medium between various modules to complete the abnormal battery box detection method of the energy storage power station. The internal modules of the system cooperate with each other to accurately identify and provide real-time warnings of abnormal battery box conditions throughout the entire operation process of the energy storage power station, constructing a multi-source data-driven chain diagnosis system, and improving the stability monitoring and risk response capabilities of the energy storage system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic flow chart of the steps of a method for detecting abnormal battery boxes in an energy storage power station;

[0017] Figure 2 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0018] Figure 3 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0019] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0020] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. It is obvious that the embodiments described are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts are within the scope of protection of the present invention.

[0021] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0022] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0023] To achieve this, please refer to Figures 1 to 3 A method for detecting abnormal battery boxes in an energy storage power station comprises the following steps:

[0024] Step S1: Acquire battery box data of an energy storage power station; collect operating environment data of the energy storage power station based on the battery box data of the energy storage power station; evaluate the initial performance of the battery box of the energy storage power station based on the battery box data of the energy storage power station; evaluate the operating status data of the energy storage power station based on the initial performance of the battery box of the energy storage power station based on the battery box data of the energy storage power station and the operating environment data of the energy storage power station;

[0025] In an embodiment of the present invention, a multi-channel battery status detection system deployed within an energy storage power station acquires data from the power station's battery boxes. The multi-channel battery status detection system includes a multi-parameter synchronous data acquisition device that integrates a voltage acquisition module, a current sensing module, a temperature monitoring module, and a state of charge sampling module. This device uses distributed optical fiber temperature measurement units to collect real-time temperature data from different locations within each battery box. It employs a time-sharing current sampling circuit to obtain the instantaneous current values of each battery group during charging and discharging, and uses a high-frequency sampling ADC circuit to collect changes in the battery box's terminal voltage. The state of charge sampling module records the SOC curve. After completing the battery box data acquisition, environmental status sensing devices deployed near the battery boxes collect data on the energy storage power station's operating environment, including ambient temperature, relative humidity, external air pressure, dust concentration, and gas component concentrations (such as H₂, CO, and CO₂). This data is uploaded in real time by the integrated sensing unit to an on-site edge computing device for data preprocessing, including time synchronization, filtering and noise reduction, and outlier removal. Next, the system calls upon a locally configured performance evaluation threshold database and compares the acquired battery pack historical data (such as temperature rise rate, voltage fluctuation range, and capacity retention rate) with the calibrated performance curves one by one to assess the initial performance level of each battery pack within the current operating cycle. This evaluation is output as quantitative data, such as voltage stability score, current fluctuation index, and capacity consistency level. Based on the correspondence between this initial performance evaluation result and environmental data (such as daily temperature range and humidity fluctuation frequency), a multidimensional data cross-correlation function is used to construct a battery state-environmental stress impact map, thereby deriving the energy storage power station operating status data. This status data includes the battery pack performance degradation rate per unit time, the voltage-temperature coupling reaction index, and the environmentally induced abnormal growth trend indicator, which serves as direct input for the subsequent support structure state evolution analysis.

[0026] Step S2: detecting an imbalance in the evolution of the supporting configuration of the energy storage station based on the operating status data of the energy storage station; detecting a steady-state collapse of the battery electrothermal coupling based on the imbalance in the evolution of the supporting configuration of the energy storage station; and detecting an abnormal trend in the quality of power transmitted by the energy storage station based on the steady-state collapse of the battery electrothermal coupling.

[0027] In this embodiment of the present invention, after receiving the operating status data output in step S1, the imbalance of the support configuration evolution is detected based on the overall topology of the energy storage power station and the battery box deployment location relationship model. This process uses a structural stress-temperature difference distribution comparison method. By deploying stress sensor arrays at key nodes of the support structure (such as battery module brackets and support beam junctions), combined with historical data of ambient temperature changes, a thermal expansion and contraction impact curve of the support components is constructed to quantify the magnitude of the stress response change. The logic for determining support configuration imbalance is based on the following criteria: if the triaxial stress change rate in a certain area exceeds a set threshold (e.g., ±15% / h) and there is a discontinuity with the stress of the adjacent support structure (discontinuity greater than ±10%) for more than two consecutive time windows (each window is 5 minutes), the support configuration in that area is determined to be evolving unbalanced. Based on this, thermoelectric dual-response sensing units deployed on the battery box surface are used to monitor the coupled changes in the thermal power density and current density per unit volume generated by the batteries during operation. The stability of the electrothermal coupling is determined by the response delay between the thermal response on-time, the temperature rise slope, and the current input. If the delay exceeds the preset response window or the thermal output power curve shows a significant nonlinear mutation, it is defined as a steady-state collapse of the electrothermal coupling. Subsequently, based on the degree of electrothermal coupling collapse, the transmission impedance, voltage consistency index, and energy conversion efficiency between the single cells are calculated and compared within the preset upper and lower limits of the power factor. If the bus voltage echo signal frequently oscillates or the phase angle shifts by more than 5 degrees, it is determined that there is an abnormal trend in the transmitted power quality. This trend data is expressed as a relative stability factor and a harmonic anomaly integral value, and is output for use in subsequent steps.

[0028] Step S3: Identifying a power grid operation coordination attenuation trend based on the abnormal power quality trend transmitted by the energy storage station; detecting a short-term harmonic surge in the energy storage station's transmission based on the power grid operation coordination attenuation trend and the abnormal power quality trend transmitted by the energy storage station; and determining a power grid stability attenuation condition of the energy storage station based on the short-term harmonic surge in the energy storage station's transmission.

[0029] In this embodiment of the present invention, the abnormal power quality trend detected in step S2 is used as input data to identify trends in grid operation coordination degradation. This identification process relies on data acquisition devices at the boundary between the energy storage power station and the external grid. This device records the active and reactive power curves at the converter ports and acquires information on transient frequency, phase angle, and voltage changes at the interface. A correlation analysis is performed based on the lag time between the fluctuation period of the energy storage unit's power output and the degree of grid main frequency deviation. If the lag time exceeds 120ms and the grid response exhibits multi-peak oscillation (i.e., the feedback power curve is non-monotonically decreasing), this is considered a trend in grid operation coordination degradation. This trend is expressed in quantitative parameters such as the duration of the frequency disturbance and the grid load adjustment hysteresis rate. After identifying coordination degradation, the harmonic components within a short time window are decomposed in the frequency domain using a wavelet decomposition algorithm, combined with the previously output abnormal power quality trend data. Energy density changes in the low-frequency band (50Hz±5Hz) and high-frequency band (>100Hz) are compared. If the high-frequency harmonic energy surge rate exceeds a set value (e.g., 25% / min) within a certain time window, the energy storage power station is judged to have significantly increased the short-term harmonics transmitted during that time period. Based on these test results, the stability coordination factor between the energy storage power station and the main grid node is calculated (depending on the frequency synchronization error, voltage recovery rate, and phase angle recovery rate). If this factor remains below a set critical value (e.g., 0.75), grid stability is determined to be attenuated. This judgment result outputs stability attenuation level data, with classifications such as "slight fluctuation," "periodic oscillation," and "nonlinear imbalance," providing a basis for further evaluation of battery response changes.

[0030] Step S4: Detect changes in the battery status of the energy storage station based on the attenuation of the stability of the energy storage station grid; identify the degree of grid frequency instability based on the changes in the battery status of the energy storage station; detect abnormal conditions of the energy storage station battery based on the degree of grid frequency instability and the changes in the battery status of the energy storage station, and upload them to the cloud platform for early warning.

[0031] In this embodiment of the present invention, based on the grid stability degradation of the energy storage power station identified in step S3, the state changes of the energy storage power station's batteries under abnormal grid disturbances are further tracked. A multi-node, multi-cycle comparison method is used to record the SOC change rate, internal resistance change trend, and temperature change range of each battery compartment during periods of harmonic surges and coordination imbalances. This is then compared and analyzed in conjunction with an existing health threshold library. If a battery cell's SOC decrease rate exceeds 3% / min and its internal resistance increase rate exceeds 10% / min within 90 seconds after the grid disturbance, the battery state change is determined to be abnormal. This change is further used to infer the degree of grid frequency instability by fitting the battery response lag time and voltage rebound amplitude to the grid frequency fluctuation curve. The frequency fluctuation peaks, valleys, and fluctuation periods are calculated accordingly to form a frequency instability index, which is represented by the product of the frequency offset rate and the frequency recovery time. Based on this, a comprehensive assessment of whether the battery is in an abnormal state is made by combining the energy storage power station's current temperature distribution, voltage consistency index, and internal resistance anomaly index. If multiple battery boxes simultaneously experience rapid SOC drop, severe temperature rise, or voltage jump during the current cycle, the battery group in that area is determined to be an abnormal unit. The identification results of the abnormal unit are identified using an ID code and packaged together with the corresponding detection data, abnormality factor, determination timestamp, and geographic location code. This data is then uploaded to the cloud platform via the edge gateway device deployed at the station. After receiving the data, the cloud platform automatically triggers the early warning mechanism based on the set abnormal battery response level standard, generates a fault trend diagram, an abnormal battery list, and an early warning notice, and sends it to the operation and maintenance platform, completing a closed-loop process for detecting abnormal battery boxes throughout the entire cycle, tailored to the actual operating status of the energy storage power station.

[0032] Preferably, step S1 includes the following steps:

[0033] Step S11: setting the temperature measurement range of the temperature sensor to 0-200°C, the minimum temperature change to 0.05°C, and the temperature sampling frequency to 10Hz;

[0034] In this embodiment of the present invention, to ensure temperature data measurement accuracy and consistency in subsequent battery compartment performance evaluation, the temperature sensor used is an industrial-grade digital PT1000 sensor with a temperature measurement range of 0°C to 200°C. A resistor bridge circuit is configured and combined with a 24-bit analog-to-digital converter (ADC) to achieve a minimum temperature change detection accuracy of 0.05°C. These sensors are installed on the top, middle, and bottom inner walls of each battery compartment, capturing the temperature distribution gradient in three dimensions. The sampling frequency is uniformly set to 10Hz, and the central processing unit (using a high-performance STM32F429 microcontroller) periodically reads the ADC values and performs temperature calculations. All measured data is stored in real time in a local edge data processing module (equipped with 512MB DDR RAM and 4GB eMMC storage) and timestamped for subsequent operating environment analysis and initial performance evaluation.

[0035] Step S12: setting the humidity sensor's humidity measurement range to 0% to 100% RH, the minimum humidity change to 0.1% RH, and the humidity sampling frequency to 5 Hz;

[0036] In an embodiment of the present invention, the sensor used to detect humidity changes inside the battery box of an energy storage power station is a capacitive humidity sensor SHT35. This sensor has a humidity measurement range of 0% RH to 100% RH and can achieve a humidity change resolution of at least 0.1% RH. The sampling frequency is configured to 5Hz, and the same central control unit synchronizes the reading of the time series data with the temperature sampling, using the I²C bus for transmission to ensure data synchronization accuracy within the range of ±5ms. The sensors are arranged in the middle of the battery box and above the wiring area to capture environmental characteristics such as condensation risk and airtightness changes. The collected data is also stored in the edge computing node and aligned with the temperature sensor sampling timestamp, and is used as one of the input factors for the subsequent initial battery performance evaluation.

[0037] Step S13: Acquire battery box data of the energy storage power station;

[0038] In this embodiment of the present invention, the process for acquiring battery pack data in an energy storage power station includes multiple data channels: voltage, current, SOC (State of Charge), SOH (State of Health), cell temperature, charge and discharge cycles, and insulation resistance. Data uploaded in real time by the BMS (Battery Management System) is collected via a CAN bus interface. Data is read at a frequency of 10 frames per second, including the cell voltage (e.g., 3.65V ± 0.01V), corresponding current (-100A to +100A), and cell temperature (derived from a built-in thermistor) of each series-connected cell. After a CRC check, the data is received and categorized by the main control system for storage. All battery packs are assigned a one-to-one correspondence: number, location, and acquisition timestamp. This provides the data foundation for environmental information matching and state evolution in subsequent steps.

[0039] Step S14: using humidity sensors and temperature sensors to collect data from the battery boxes of the energy storage power station to obtain operating environment data of the energy storage power station;

[0040] In an embodiment of the present invention, a temperature and humidity sensor device deployed inside the battery box is used to synchronously collect operating environment data. A dual-channel data docking module is set in the central control system to process the temperature and humidity channels respectively, and a ring cache structure is used for real-time data caching. The cache area data is packaged and uploaded to the local environment data recording module every 10 seconds. For abnormal temperature gradient changes (more than 1.5°C / minute) or sharp fluctuations in humidity (more than 10%RH / minute), an alarm flag is set to indicate that the external environment has affected the battery performance. Such data is synchronized to the battery box data record item in a time series to form an "operating environment-performance" binding record, which is convenient for evaluating the environmental coupling impact in step S16.

[0041] Step S15: Collecting the cumulative operation time data of the energy storage power station based on the battery box data of the energy storage power station;

[0042] In an embodiment of the present invention, the cumulative operation time data is counted based on the battery box activation record and the operation log. Each battery box is built-in with a unique number when it is deployed at the factory. The first grid-connected time record (accurate to the second) is retrieved through the central platform and accumulated with the working period data recorded in the operation log. The collection cycle is set to 00:00 every day to perform a cumulative update of the total time, and the current working status (stationary, charging, discharging) is recorded in real time. For example, the battery box numbered "B01234567" has a cumulative operating time of 4867 hours since it was connected to the grid, and the current cumulative working time is 11.2 hours. Both are synchronously recorded in the database "battery_operation_log" as one of the reference bases for subsequent initial performance evaluation.

[0043] Step S16: Evaluate the initial performance of the energy storage power station battery box based on the energy storage power station battery box data;

[0044] In an embodiment of the present invention, the initial performance evaluation is performed in parallel using static parameter comparison and dynamic trend analysis. The benchmark reference values are taken from the standard performance parameter table in the technical specification provided by the manufacturer. Taking the single cell as the unit, the voltage deviation, temperature rise rate, current fluctuation, SOC / SOH stability and other indicators are calculated item by item. If the cell voltage fluctuates by more than 0.05V within 10 minutes and the SOC fluctuates by more than 5% per unit time, it is considered that there is an initial deviation in the performance. Taking the battery box numbered "B01234567" as an example, the initial evaluation shows that 4 of the cells have over-temperature deviations (exceeding the specified threshold by 4°C) and 2 cells have voltage drift. These abnormal data are marked and recorded as input variables for subsequent state analysis and are uniformly saved in the "initial_performance_record" database field.

[0045] Step S17: Evaluate the initial performance of the battery box of the energy storage station and the energy storage station operating status data based on the accumulated operating time of the energy storage station and the operating environment data of the energy storage station.

[0046] In this embodiment of the present invention, the operational status data assessment performs a joint analysis based on the multidimensional datasets obtained in steps S13 to S16. Environmental data (temperature, humidity) and operating time data are mapped to a time series linked table under the battery compartment number, and missing data is linearly filled using equal-interval interpolation. Subsequently, a rule engine is used to set thresholds for several evaluation indicators, such as the cumulative duration of ambient temperature exceeding 45°C for more than 100 hours and the humidity exceeding 80% RH for a long period of time. These are weighted and scored in combination with the operating time to form the Environmental Stress Index (ESI). This index is then integrated with indicators such as voltage fluctuation rate, current stability, and SOC fluctuation range to calculate the Operational Health Index (OHI), which is output as a numerical value. For example, for the battery compartment numbered "B01234567," the ESI is 0.72 and the OHI is 0.86. The system submits this state data to the "operation_state_database" for use in step S2 to detect imbalances in the support configuration evolution.

[0047] Preferably, step S16 includes the following steps:

[0048] Step S161: extracting internal arrangement structure data of the battery box based on the battery box data of the energy storage power station;

[0049] In an embodiment of the present invention, the battery box data of the energy storage power station includes physical structural drawings, structural assembly coding record sheets, and optical image acquisition data recorded during the installation phase. A high-resolution industrial camera (e.g., a 12-megapixel CMOS camera) is combined with a multi-angle robotic arm to perform internal structured light scanning to obtain an arrangement image of the battery cell modules inside the battery box. During the scanning process, each angle scanning data must be corrected for geometric distortion using calibration and correction software. Subsequently, an image segmentation algorithm is used to identify the positions of different battery cell modules in blocks, and the angle distribution between the edge lines and the construction lines extracted from the image is used to determine whether the arrangement is a regular matrix arrangement, a stacked arrangement, or an interlaced nested arrangement. The structural image is converted into a two-dimensional arrangement coordinate matrix and recorded in a JSON structure. The extracted internal arrangement structure data of the battery box includes the minimum center distance between cells, the total number of rows and columns of arrangement units, the density of structural units, etc.

[0050] Step S162: Identifying the internal conductive connection status of the battery box based on the internal arrangement structure data of the battery box;

[0051] In an embodiment of the present invention, based on the two-dimensional arrangement coordinate matrix extracted in step S161, the connection wire path and position are further obtained, and the temperature rise image of each connection conductive path during operation is collected by combining visible light images with an infrared thermal imager under low-power power-on conditions. The temperature rise area can accurately correspond to the position of the connection line segment. After superimposing the temperature rise image with the arrangement structure image, all conductive path segments are extracted through an image matching algorithm. The OCR algorithm is used to extract the specification parameters printed on the surface of the connection component (such as copper busbar width, connector type code) from the image, and the data comparison library is used to confirm whether it is a standard specification. Finally, a topological matrix containing the conductive connection mode between each battery cell and the adjacent battery cell is established. The elements in the matrix are represented by parameters such as conductive path length, cross-sectional area and material code, which constitute a complete conductive connection situation inside the battery box.

[0052] Step S163: Evaluate the regularity of the battery box busbar wiring based on the internal conductive connection status of the battery box and the internal arrangement structure data of the battery box;

[0053] In an embodiment of the present invention, the battery box arrangement structure extracted in step S161 and the conductive connection topology matrix established in step S162 are jointly analyzed. For each conductive connection path, the path length, direction offset angle and offset degree of the theoretical shortest path are calculated respectively; the connection node distribution density, the uniformity of the contact point spacing and the frequency of repeated crossing are further analyzed. If the average offset angle of the unit conductor is less than 3°, the repeated crossing rate is less than 10%, and the variance of the contact point density is less than 0.05, it is judged that the line regularity is good, otherwise it is poor. After all the evaluation values are aggregated, a line regularity score value is constructed. The value is output in percentage, representing the neatness, regularity and standardization of the entire battery box wire arrangement, and a battery box busbar wiring regularity data table is generated.

[0054] Step S164: Evaluate the heat dissipation capacity of the battery box based on the regularity of the battery box busbar wiring and the internal arrangement structure data of the battery box;

[0055] In this embodiment of the present invention, the busbar wiring regularity score obtained in step S163 and the cell arrangement structural parameters obtained in step S161 are used to comprehensively evaluate the patency of the heat dissipation channels and the integrity of the heat conduction paths. A three-dimensional thermal flow field simulation model is constructed using a CFD simulation tool (such as Ansys Fluent). The simulation inputs heat flux data for the cell and connecting conductor surfaces using an aerodynamic approach, and sets the thermal conductivity of the battery case material and the thermal resistance of the internal conductor material. The temperature evolution of the cells under constant power operation for 30 minutes is simulated to extract the maximum inter-cell thermal gradient, the average air velocity distribution in the air channels, and the area ratio of local hotspots. Statistical analysis of the simulation data is used to output the average heat dissipation capacity (in W / m²·K) and the heat dissipation nonuniformity index (defined as the maximum temperature difference divided by the average temperature difference) within the battery case, forming a data structure for the internal heat dissipation capacity of the battery case.

[0056] Step S165: Calculating the initial voltage deviation of the energy storage station battery according to the energy storage station battery box data;

[0057] In this embodiment of the present invention, the energy storage station battery box data contains the voltage values of all battery cells in the initial charge state, which are recorded once every 100ms by the voltage acquisition module. All voltage values are combined into a voltage data vector, and the maximum value V_max, minimum value V_min, and average value V_avg are calculated. The voltage deviation is further calculated as ΔV = V_max - Vmin, and the voltage standard deviation σ_v. The output initial voltage deviation of the energy storage station battery is composed of two parameters: the voltage difference ΔV (unit: mV) and the standard deviation σ_v (unit: mV). The larger the voltage deviation, the poorer the cell consistency, which in turn reflects the initial balance stability risk of the battery pack.

[0058] Step S166: Determine the initial operating stability of the battery box of the energy storage power station based on the internal heat dissipation capacity data of the battery box and the initial voltage deviation of the battery of the energy storage power station;

[0059] In this embodiment of the present invention, a logical correlation analysis is performed between the internal heat dissipation capacity data of the battery compartment generated in step S164 and the voltage deviation data in step S165 to determine whether there is localized heat accumulation in the heat dissipation capacity data (i.e., a heat dissipation unevenness index > 1.5). Secondly, a determination is made as to whether the voltage deviation ΔV is greater than a set threshold (e.g., 50mV). If both uneven heat dissipation and a large voltage difference exist, the initial operational stability level is determined to be low; if only one of these is abnormal, the level is medium; if both are normal, the level is high. Based on this judgment logic, an initial operational stability level value L_s (with values of high, medium, or low) is constructed, and the stability level boundary values are corrected in combination with ambient temperature and humidity data to prevent misjudgment of the initial operational stability level and its corresponding heat dissipation factor and voltage fluctuation factor due to sensor drift.

[0060] Step S167: Evaluate the initial performance of the energy storage power station battery box based on the initial operation stability of the energy storage battery box and the initial operation stability of the energy storage power station battery box.

[0061] In an embodiment of the present invention, the initial operational stability level of the battery box evaluated in step S166 is compared with the operational stability records of the entire battery box during the manufacturing process and the debugging stage, including data items such as the comparison records of battery cell expansion during historical operation, the battery balanced charge and discharge time difference, and the internal impedance distribution of the battery cell. A weighted evaluation of multiple indicators is performed using a decision rule tree method to determine a comprehensive score S_p. The scoring dimensions include heat dissipation capacity, structural consistency, voltage consistency, thermal stability level, etc. The score S_p>80 is set as excellent performance, 60-80 as medium, and <60 as poor. An initial performance evaluation result table of the battery box of the energy storage power station is generated, and a detailed result structure including the score value, grade classification, and the contribution ratio of each indicator is output. This data serves as the initial benchmark reference data for the subsequent battery box status change trend, and as a comparison reference for subsequent abnormal detection data.

[0062] Preferably, step S17 includes the following steps:

[0063] Step S171: collecting abnormal fluctuations in operating environment temperature and humidity based on the operating environment data of the energy storage power station;

[0064] In this embodiment of the present invention, industrial-grade temperature and humidity sensor modules are deployed in various locations within the energy storage power station's battery compartment environment. The deployed sensor equipment must provide 24-hour continuous monitoring capabilities, with temperature measurement accuracy of no less than ±0.1°C and humidity measurement accuracy of no less than ±1.5%RH. Sensors are installed near the top, bottom, left and right sides of the battery compartment, and in the external vents to capture temperature and humidity data covering key areas such as local hot spots, enclosed areas, and ventilated areas. Each sensor has a fixed sampling period of 10 seconds, and the collected data is archived uniformly by an edge data acquisition unit (equipped with A / D conversion and preliminary data caching). The collected temperature and humidity data is transmitted to the data preprocessing unit in the local controller for abnormal fluctuation identification. The identification rule is set as follows: if the temperature changes by more than ±3°C or the humidity changes by more than ±10%RH within 60 consecutive minutes, it is marked as an "abnormal fluctuation point"; if the number of abnormal fluctuation points in a certain time period accounts for more than 30% of the total number of points in that section, it is defined as "abnormal fluctuation of ambient temperature and humidity" in that section. The generated result is the "Abnormal Fluctuation Determination Data Table", which includes the time period number, abnormal fluctuation amplitude, frequency, and specific sensor location distribution.

[0065] Step S172: determining the sudden change of the operating environment's temperature and humidity during the day and night based on the abnormal fluctuation of the operating environment's temperature and humidity;

[0066] In this embodiment of the present invention, after completing the abnormal fluctuation determination data table in step S171, data from each time period is extracted and divided into diurnal cycles. Data is grouped according to the time of day, with each group corresponding to two 24-hour periods: daytime (06:00–18:00) and nighttime (18:00–06:00). The average temperature and humidity values within these two periods are compared, and the temperature difference (ΔT = |T_day - T_night|) and the humidity difference (ΔRH = |RH_day - RH_night|) are calculated. A sudden change is determined by setting the thresholds for ΔT exceeding 6°C and ΔRH exceeding 15%RH to identify a sudden change. If at least one of ΔT and ΔRH exceeds the threshold and occurs for more than three consecutive days during a given day, it constitutes a "diurnal temperature and humidity sudden change." The analysis results are summarized to generate a "diurnal temperature and humidity sudden change statistics table," which includes fields such as the sudden change time range, sudden change type (temperature, humidity, or both), duration, and sudden change amplitude range.

[0067] Step S173: Identifying the oxidation trend of the energy storage station structure based on the initial performance of the energy storage station battery box when the sudden change of temperature and humidity in the operating environment exceeds 30% during the day and night;

[0068] In an embodiment of the present invention, the proportion of mutation events in the total monitoring period is counted in the day and night mutation statistics table. If the proportion of mutation days exceeds 30%, the structural oxidation trend identification operation is performed. The high-incidence period of mutation (the continuous period in which the mutation proportion exceeds 30%) is used as the analysis window, and the oxidation trend is estimated by combining the relative humidity change in the ambient air and the material parameters of the battery box shell (such as aluminum alloy or steel structure). The specific method is: combining the relationship between ambient humidity and metal corrosion rate, by accumulating the daily oxidation growth rate data, the total oxidation trend level within the analysis window is obtained, and the data is normalized into a percentage value to generate structural oxidation trend data, such as 40%, 50%, 60%, etc., to output the "Structural Oxidation Trend Level Assessment Table", which is used to subsequently determine whether to enter the structural pressure test process.

[0069] Step S174: testing the pressure-bearing operation state of the energy storage station when the cumulative operation time of the energy storage station exceeds 48 hours and the oxidation trend of the energy storage station structure exceeds 50%;

[0070] In this embodiment of the present invention, if the structural oxidation tendency assessment results indicate an oxidation tendency value exceeding 50%, and the energy storage power station has been operating for more than 48 hours since startup (the startup time is recorded and accumulated by the operating time monitoring system), a pressure-bearing operating state test is performed. A pressure measurement assembly is installed at the connection between the battery box and the power station frame. It includes resistance strain gauges, vibration acceleration sensors, and thermal expansion displacement sensors installed at stress-bearing points on the box (such as the bottom support and top cover hinge points). Data from each measurement point is collected in real time and transmitted to the control unit. The data is then cross-compared with the factory-initial structural mechanical state data (stored in the equipment management system) using the three parameters of stress, temperature, and deformation. The differences are then compared with the factory-initial structural mechanical state data (stored in the equipment management system). If the strain value exceeds 80% of the design elastic limit, the vibration intensity increases by more than 50%, and the thermal expansion exceeds 1.5 mm, the pressure-bearing state is determined to be approaching a critical state, and a "Pressure-Bearing Operating State Report" is generated, which includes the structural deformation level, stress distribution diagram, and warning level indicators.

[0071] Step S175: estimating the overload condition of the energy storage power station according to the load-carrying operation status of the energy storage power station;

[0072] In an embodiment of the present invention, based on the pressure operation status report, the pressure level is mapped to the power and load level of the power station. Combined with the current load data of the energy storage power station (the power load curve recorded every 10 seconds is obtained from the power meter and current collector), a comparative analysis of the load trend and the pressure status is performed. The "overload threshold" is defined as: the corresponding load percentage when the structural strain exceeds the limit. For example, if the power corresponding to the strain critical value is 1000kW, when the current power exceeds this value, it is considered "operation overload". The measured actual power load data is compared with this threshold, the overload ratio is calculated, and the operation overload determination result is output.

[0073] Step S176: Evaluate the energy storage station operation status data based on the initial performance of the energy storage station battery box and the energy storage station structure oxidation trend.

[0074] In this embodiment of the present invention, the structural oxidation tendency percentage value obtained in step S173 and the operating overload determination result generated in step S175 are used as dual-variable inputs to assess the initial performance of the energy storage power station battery box. An operating status grading system is established, including: stable (oxidation tendency <30%, no overload), mild warning (oxidation tendency 30%-50%, mild overload), moderate risk (oxidation tendency 50%-70%, mild or severe overload), and high risk (oxidation tendency >70%, severe overload). Using the aforementioned logical rules, the corresponding operating status level is output by cross-combining the oxidation tendency level and the overload situation. The assessment results include the operating status level (e.g., moderate risk), influencing factors (structural oxidation tendency and operating load ratio), and the corresponding recommended maintenance time period. These results are summarized into an "operation status data report" to support subsequent monitoring, operation and maintenance, and early anomaly detection processes of the energy storage power station battery box.

[0075] Preferably, the energy storage support configuration evolution imbalance detection in step S2 includes:

[0076] Identify the degree of oxidation of the energy storage station connection structure based on the energy storage station operation status data;

[0077] In this embodiment of the present invention, energy storage power station operating status data includes, but is not limited to, real-time ambient humidity (unit: %RH) and temperature (unit: °C) acquired by the station's temperature and humidity acquisition unit; and electrical contact resistance (unit: mΩ) measured on the surface of energy storage connection terminals (including bolts, copper busbars, etc.). A four-wire resistance measurement module is installed at each connection terminal to periodically collect resistance values every five minutes and upload them, along with the current and voltage records at that location, to a local industrial control system. A hardware-implemented multi-channel data synchronization acquisition module processes this data in conjunction with metal surface temperature rise (obtained by a thermal imager) and changes in metal infrared reflectivity (obtained by an infrared spectrometer). Based on the increasing contact resistance trend of stainless steel or copper conductors during oxidation, a mapping table is established between the time series increase in contact resistance and oxidation level. Oxidation levels are categorized into four categories: mild (ΔR < 2 mΩ), moderate (2–5 mΩ), severe (> 5 mΩ), and extremely severe (> 10 mΩ). The system identifies the degree of oxidation in the connection structure by comparing these values within this table. This process does not rely on model inference, but is based on the measured physical resistance value and the trend of electrical performance changes during the metal oxidation process.

[0078] Determine the loosening trend of the energy storage station connection structure based on the oxidation degree of the energy storage station connection structure;

[0079] In an embodiment of the present invention, the three-axis acceleration sensor at the connection point is again called to collect the micro-vibration frequency (unit: Hz) and acceleration (unit: mg) generated at the joint during operation, wherein the acceleration sensor records dynamic response data within one second at a frequency of 500 Hz. Resistance mutation and vibration frequency abnormality together constitute loosening trend identification indicators. During implementation, connection points with an oxidation level of "medium" and above are key monitoring targets, and the situation where the number of resistance fluctuations per unit time (1 hour) exceeds 3 times and the acceleration amplitude fluctuation range exceeds ±25 mg is analyzed. This type of electrical connection structure is judged to have a loosening trend. This operation logic is based on the linkage change characteristics of the measured resistance and the dynamic response signal rather than empirical rules, and uses a fixed hardware signal threshold judgment method. All judgment results are stored in the non-volatile memory of the local PLC control system for subsequent use.

[0080] Estimate the micro-vibration of the energy storage station connection based on the loosening trend of the energy storage station connection structure;

[0081] In one embodiment of the present invention, a highly sensitive micro-vibration capture device (e.g., a combination of a MEMS tri-axis gyroscope and a high-frequency MEMS accelerometer) is placed on a connection structure identified as exhibiting a tendency to loosen. The device collects micro-vibration signals at a sampling rate of 1000 times per second. In this embodiment, the sensor quantifies the vibration intensity in the X, Y, and Z directions in μg. The frequency domain distribution of the vibration signal is analyzed using an FFT (Fast Fourier Transform) algorithm to extract the high-frequency vibration energy density (in dB / Hz) within the 1–100 Hz range. This density is used as a proxy for micro-vibration intensity, and the maximum and mean values are recorded. If the degree of oxidation of the connection structure is moderate or above and the vibration energy density exceeds a threshold (e.g., 20 dB / Hz), the micro-vibration level is estimated to be "active." This process does not involve model inference; all data processing is performed using deterministic algorithms and measurement threshold judgments.

[0082] Measure the degree of mechanical stress growth of the energy storage power station structure based on the micro-vibration of the energy storage power station connection;

[0083] In this embodiment of the present invention, to quantify the mechanical stress changes caused by micro-vibration, a strain gauge array (with bidirectional strain gauges deployed every 50 cm) is deployed on the support structure and connecting plates of the energy storage power station. A full-bridge electrical strain measurement system is used to collect the stress response generated at the corresponding locations under vibration excitation. The strain gauge module amplifies the strain signal and converts it into stress units (MPa), with a measurement accuracy of ±0.5%. The stress response waveform within each vibration cycle is collected, and its maximum and average amplitudes are calculated. The data is then compared and analyzed to determine the stress growth rate, i.e., dσ / dt (MPa / h). If the stress response growth rate at a structural location exceeds a set threshold (e.g., 1.5 MPa / h), the structural location is flagged as experiencing a continuous increase in mechanical stress. The stress measurement system and the micro-vibration analysis subsystem synchronize sampling using a unified timestamp to ensure accurate analysis.

[0084] Predict the micro-deformation of the energy storage power station structure based on the degree of mechanical stress growth of the energy storage power station structure and the micro-vibration of the energy storage power station connection;

[0085] In an embodiment of the present invention, the obtained structural stress growth data and the micro-vibration parameters in step S23 are used as input to call a structural displacement measurement unit (laser displacement meter or distributed fiber Bragg grating ranging system) to measure the micron-level displacement change of the structure per unit time. During the implementation process, each laser displacement meter records the maximum displacement value in the X and Y directions of a fixed support node, and a one-to-one mapping relationship is established between the displacement change (unit μm) and the stress growth rate and vibration frequency to determine whether micro-deformation characteristics appear. If the maximum displacement amplitude within a unit time (1 hour) continues to exceed 20 μm, combined with the known stress growth rate and vibration level, the system determines that the structure has a significant micro-deformation trend. All displacement data are completed using a physical measurement device and do not rely on algorithm estimation. The measuring device is fixedly installed in a preset interface in the structural shell, and the data is uploaded to the edge processing module via industrial Ethernet and stored in the database.

[0086] Detect the dynamic fatigue condition of the energy storage power station structure based on the micro-deformation condition of the energy storage power station structure and the degree of mechanical stress growth of the energy storage power station structure;

[0087] In an embodiment of the present invention, a stress-deformation history curve is constructed at the key nodes of the structure in combination with the stress growth rate and the deformation. A cycle number statistical method is used to record the number of daily stress fluctuation cycles (each cycle is defined as a loading-unloading-reloading process) and the maximum stress amplitude. Based on the fatigue limit of the material (different steels correspond to different SN curves, which are determined in the design stage), the actual number of stress cycle loadings is compared with the fatigue limit cycle threshold to determine the fatigue level: primary (<30% fatigue life), intermediate (30%-60%), and severe (above 60%). The fatigue state is labeled in real time and updated to the structural state database. Data processing does not involve abstract model reasoning, and is completed by directly comparing the material mechanics standard curve with the measured data.

[0088] The imbalance of the supporting configuration evolution of the energy storage power station is detected based on the dynamic fatigue condition and micro-deformation condition of the energy storage power station structure.

[0089] In an embodiment of the present invention, the fatigue level, micro-deformation amplitude and spatial distribution pattern of each structural support node are compared to construct a load-bearing state distribution diagram of each unit of the support structure. If the fatigue level of any node in the support system is severe, and the corresponding displacement amplitude is not less than 50μm, and the average displacement difference of the three connecting nodes around the node exceeds 30μm, the system determines that the current support configuration has evolved into an imbalance. The structural support configuration is no longer in the original design state, and exhibits stress concentration or uneven deformation distribution. The system records the number, spatial coordinates and deformation level of the support imbalance area to form a complete "support configuration evolution imbalance report" for subsequent maintenance strategy formulation and early fault warning. This conclusion comes from the cross-validation of physical data statistics, structural distribution displacement diagrams and fatigue curves, and does not involve any model algorithm or empirical rules.

[0090] Preferably, the battery electrothermal coupling steady-state disintegration condition detection in step S2 includes:

[0091] Identify abnormal extrusion inside batteries based on the imbalance of the supporting structure evolution of the energy storage power station;

[0092] In an embodiment of the present invention, based on the configuration monitoring data of the supporting structure of the energy storage power station box, a decoupling analysis of the configuration evolution trend is carried out. Strain gauge arrays and structural displacement gauges are installed at the bottom, side walls and key stress nodes of the energy storage unit. The strain gauges use high-precision foil resistance strain gauges with a response time of less than 10ms; the displacement gauges use laser triangulation displacement sensors with a measurement accuracy of 0.05mm. By scheduling the multi-time structural strain field and displacement data within the time period, the difference between the yield critical deformation amplitude and the current deformation of each supporting component is calculated, and the strain concentration area is used to identify the configuration evolution imbalance mode such as unilateral sinking of the support bracket, deformation distortion, loose connection, etc. If a sudden gradient displacement or abnormal strain gradient aggregation is identified in the support path of a local component, it can be confirmed that a local abnormal extrusion phenomenon caused by the redistribution of the support load has occurred inside the battery pack.

[0093] Measure the lateral force imbalance of the battery cell based on the abnormal extrusion inside the battery;

[0094] In an embodiment of the present invention, after confirming that abnormal extrusion has occurred, the change in the lateral contact pressure of the battery cell is obtained by the MEMS micro-pressure sensor array embedded in the side structure of the battery module. The sensor arrangement spacing is no more than 10mm to ensure that the accuracy covers the unit of a single battery cell. The pressure measurement data is compared with the mechanical equilibrium value of the nominal arrangement of the battery cell to calculate the deviation of the lateral pressure difference between each battery cell from the standard pressure gradient. The sensor data arranged inside the battery module is aggregated to the edge processing unit in real time through the serial I²C protocol, and the lateral pressure distribution field is constructed through the multi-point spatial interpolation method (non-model type), and its gradient direction and amplitude range are extracted. If the lateral pressure distribution has a single-direction asymmetric area of more than 30%, and the maximum gradient exceeds 200kPa / m, it is determined that the lateral force imbalance of the battery cell is serious.

[0095] Determine the horizontal displacement of the battery cell based on the degree of lateral force imbalance of the battery cell and the abnormal extrusion inside the battery;

[0096] In an embodiment of the present invention, the identified extrusion position coordinates are superimposed with the force distribution gradient map measured by S22. After determining the main force direction and the stress concentration point, the laser displacement probe is started to measure the horizontal displacement of the corresponding battery cell. The probe is deployed along the width direction of the module with an accuracy of not less than 0.01mm, and records the actual horizontal offset of the battery cell surface relative to the fixed reference plane of the battery box. Based on the measurement results, a battery cell sequence displacement distribution curve is established. Combined with the nominal arrangement spacing of the battery cells and the actual deviation range, parameters such as displacement peak value, deformation extension length, concentrated offset area, etc. are extracted to form a "battery cell horizontal displacement parameter matrix". If the horizontal displacement of any single battery cell exceeds the design allowable offset upper limit (usually 1.5mm), and is accompanied by more than three adjacent battery cells offset in the same direction, it is confirmed that the battery cell sequence has undergone horizontal displacement.

[0097] Measure the imbalance of insulation distance between battery cells based on the horizontal displacement of battery cells;

[0098] In an embodiment of the present invention, based on the obtained horizontal displacement parameter matrix of the battery cells, an ultrasonic pulse reflection module is deployed through interval measurement to perform reflection time scanning on the thickness of the insulating medium between the battery cells, and the ranging accuracy must reach 0.02mm. The middle, upper and lower points of each pair of adjacent battery cells are used as ranging sampling points to construct an insulation spacing profile. Based on the standard insulation distance (such as 2mm), the insulation spacing compression ratio at different points is calculated. The boundary of the area with a compression ratio exceeding 30% is judged. If the compression area is concentrated in a certain battery cell group and extends laterally for more than four battery cells, it can be determined that the insulation spacing between the battery cells is unbalanced to form an effective imbalance band. After jointly analyzing all the measured spacing data and the offset direction, the "insulation imbalance level judgment map" is output.

[0099] Determine the degradation of the electrical connection performance of the energy storage power station based on the imbalance of the insulation distance between battery cells;

[0100] In an embodiment of the present invention, the insulation imbalance level determination map output by this step matches the position of the conductive connecting piece between the corresponding battery cells. The local on-resistance of the connecting piece is measured using a laser conductivity tester, and the four-terminal method is used for precise measurement with an accuracy of not less than 1μΩ. The resistance of the connecting piece in the imbalanced area is compared with the nominal initial state resistance, and its resistance growth rate curve is extracted, and the transition impedance and abnormal contact heat value are evaluated at the same time. If the resistance growth rate in a certain area exceeds 200% and overlaps with the heat value rising area (>15%), it is considered that the electrical connection performance has been significantly degraded. A "connection performance degradation distribution map" is formed on the data, in which the spatial overlap rate of the high-resistance path concentration zone and the insulation imbalance zone is marked, which serves as the basis for subsequent thermal coupling instability deduction.

[0101] Measure the distortion of the heat dissipation path of the battery cells based on the imbalance of the insulation distance between the cells;

[0102] In an embodiment of the present invention, an embedded infrared thermal imaging unit is used to dynamically scan the thermal distribution on the surface of the battery module. The image frame rate is not less than 25 Hz, and the temperature resolution is better than 0.1°C. Combined with the insulation compression section obtained by S24, the extension direction and center offset of the heat diffusion area are extracted in a targeted manner. By constructing a heat flow path vector diagram (based on the first-order derivative of the temperature gradient) to invert the heat dissipation path morphology, the heat flux density distribution and isothermal line fitting are further used to detect whether it is offset, compressed or broken. If the original linear heat flow direction deflects by more than 30° in the compression area, or the width of the isothermal zone of the heat dissipation path decreases by more than 40%, it is determined that the heat dissipation path of the battery cell is severely distorted.

[0103] The steady-state collapse of the battery's electrothermal coupling is detected based on the distortion of the battery cell's heat dissipation path and the degradation of the electrical connection performance of the energy storage power station.

[0104] In this embodiment of the present invention, a coupled steady-state risk assessment grid is constructed based on the spatial overlap between areas where heat dissipation paths are obstructed and connection impedance increases. This grid incorporates the electrothermal combined power consumption density (the sum of thermal power consumption and conduction power consumption per unit area and per unit time) as an indicator. If the thermal power consumption density within a grid exceeds a threshold of 2W / cm² and is accompanied by an increase in conduction impedance, the coupled steady-state structure in that area is deemed unbalanced, and an "electrothermal coupled steady-state collapse map" is generated. By overlaying three cycles of time-series data, it is determined whether the collapse state of that area is in a continuously deteriorating trend. This allows identification of stability collapse units, providing a basis for invoking downstream intelligent alarm systems.

[0105] Preferably, the abnormal trend detection of power quality transmission of the energy storage power station in step S2 includes:

[0106] Detect the continuous growth of the battery's internal resistance to current transmission based on the battery's electrothermal coupling steady-state disintegration condition;

[0107] In this embodiment of the present invention, real-time current, voltage, and temperature data from each battery cell within an energy storage power station are collected. High-precision power quality monitoring devices are used to monitor the voltage drop changes associated with the battery's transmission current. Combined with the results of steady-state electrothermal coupling degradation testing, the dynamic internal resistance of the battery is calculated. Specific steps include: deploying current sampling sensors and voltage sampling modules to collect the current and voltage signals at the battery terminals, respectively. Synchronous sampling ensures temporal consistency of the data; digital filtering is then employed to eliminate noise interference and obtain stable voltage and current data. Internal resistance calculation is based on Ohm's law, using the ratio of the measured voltage drop to the current value to derive the battery's internal resistance. Continuous sampling over multiple time periods creates an internal resistance curve. Statistical analysis is then used to determine whether the internal resistance exhibits a sustained growth trend. A trend analysis algorithm is then used to identify the rate and duration of change in the internal resistance, generating quantitative parameters indicating the sustained internal resistance growth, such as the internal resistance growth rate and growth time window. This step is then connected to the next step, which uses the continuously increasing internal resistance parameters as input data to support the subsequent determination of the transmission current shunt imbalance.

[0108] Determine the battery transmission current shunt imbalance condition based on the continuous growth of the battery transmission current internal resistance;

[0109] In this embodiment of the present invention, based on the obtained battery internal resistance continuous growth parameter, a multi-point current distribution monitoring device is used to collect current data from each parallel branch in the energy storage battery pack. High-precision Hall-effect current sensors are installed on the transmission lines of each parallel branch of the battery pack to ensure real-time acquisition of current changes in each branch. Based on the internal resistance change data and current sampling values, the current ratio between each branch is calculated to quantify the degree of current shunt balance. Specifically, the root mean square deviation (RMSD) of the current distribution is used to calculate the deviation of the current in each branch and determine the degree of current shunt imbalance. Continuous internal resistance growth can cause the current in some branches to decrease, leading to current shunt imbalance. Time series analysis techniques are used to track the changing trend of the current shunt indicator and output a current shunt imbalance level parameter. The output parameter of this step serves as input for subsequent abnormal overcharge prediction, enabling gradual information transmission and in-depth analysis.

[0110] Predict abnormal overcharge of batteries in energy storage power stations based on the imbalance of battery transmission current shunt;

[0111] In an embodiment of the present invention, the current and voltage characteristics of the battery charging stage are analyzed by using anomaly detection technology in combination with the obtained current shunt balance parameters. By setting up a special battery charging monitor, the charging current curve and battery voltage response signal are continuously collected to form a charging status database. By comparing with the standard charging curve, abnormal deviation phenomena are detected. The specific operations include calculating the peak value of the charging current, the charging time, and the stability index of the voltage platform stage. The overcharging phenomenon of some batteries caused by current shunt balance will be reflected in the charging current fluctuation and voltage anomaly. The statistical threshold method is used to identify the abnormal current fluctuation points in the charging process, and the abnormal overcharging risk level is judged in combination with the current shunt balance level, and the abnormal overcharging probability parameter is quantified. The prediction results provide key input data for the next step of analyzing the collapse trend of battery active materials, completing the natural connection of the data flow.

[0112] Determine the damage trend of battery active materials based on abnormal overcharge conditions of batteries in energy storage power stations;

[0113] In an embodiment of the present invention, based on the predicted abnormal overcharge parameters, combined with battery chemical property detection equipment, active material status detection is implemented. The active materials of the sampled battery cells are quantitatively characterized by a laser scanning confocal microscope or an X-ray diffractometer to obtain information on changes in the material's structural integrity and surface morphology. Furthermore, combined with the battery's cyclic charge and discharge history data, the capacity loss rate and degree of structural damage of the active material are determined by chemical analysis methods. The abnormal overcharge parameters are correlated with the chemical analysis results to construct an active material collapse trend curve. This trend is measured in percentage of material capacity loss, reflecting the speed and cumulative degree of material degradation. This data serves as the basis for the subsequent evaluation of the degree of damage of the electrode interface synergistic mechanism, forming a complete technical chain from current anomaly to material collapse.

[0114] Test the degree of damage to the battery electrode interface synergy mechanism based on the collapse trend of battery active materials;

[0115] In an embodiment of the present invention, electrochemical impedance spectroscopy (EIS) technology is used to conduct in-depth testing of the synergistic mechanism of the battery electrode interface based on the active material collapse trend determined in step S2-4. By applying alternating current signals of different frequencies, the impedance response data of the battery electrode interface is collected to obtain key parameters such as the interface charge transfer impedance and the electrolyte diffusion impedance. The impedance spectrum data is compared and analyzed with the active material collapse trend to quantitatively describe the degree of damage to the synergistic mechanism of the electrode interface. Specifically, a complex impedance diagram and an equivalent circuit model fitting method are used to separate the impedance contribution of each component of the interface, calculate the proportion of reduced interfacial activity and the degree of ion migration obstruction. The degree of damage is expressed as a percentage increase in the interface impedance, providing important data support for the subsequent estimation of the restriction of internal ion diffusion channels, ensuring a closed technical loop from material collapse to interface damage.

[0116] Estimate the limitation of ion diffusion channels inside the battery based on the degree of damage of the battery electrode interface synergistic mechanism and the collapse trend of the battery active materials;

[0117] In an embodiment of the present invention, a multi-parameter coupled analysis method is used to estimate the restricted state of ion diffusion channels within the battery, combining acquired electrode interface damage parameters with material disintegration trend data. By comprehensively analyzing battery electrochemical test data, changes in the physical structure of the active materials, and changes in interface impedance, the degree of narrowing and obstruction of the ion diffusion path are determined. Specific steps include using a pulsed current method to determine the diffusion coefficient, observing internal microstructural changes using a scanning electron microscope (SEM), and quantitatively analyzing changes in the cross-sectional area and patency of the ion channels. A numerical interpolation method is used to construct a distribution map of the degree of restriction of the ion diffusion channels, and an ion diffusion obstruction index is output. This index serves as a direct input parameter for detecting abnormal power quality trends, effectively correlating internal and external structural changes with power quality anomalies.

[0118] The abnormal trend of power quality transmitted by energy storage power station is detected based on the restricted ion diffusion channels inside the battery.

[0119] In an embodiment of the present invention, the obtained ion diffusion barrier index is used as the core parameter, combined with the data of the overall power quality monitoring system of the energy storage power station, to detect abnormal trends in the transmission power quality. By monitoring the voltage waveform, harmonic content, power factor and transient current changes on the AC side, combined with the changes in the internal parameters of the energy storage battery pack, signal processing technology is used to extract abnormal features. Specifically, through Fourier transform and wavelet transform technology, the frequency components and change trends in the voltage and current signals are analyzed to identify the decrease in energy conversion efficiency and the deterioration of power quality caused by limited ion diffusion. The ion diffusion barrier index is combined with the power quality index to construct a multi-dimensional abnormal trend map, quantify the abnormal trend level, and output the abnormal trend parameter of the transmission power quality. This parameter provides an important reference for the overall abnormality detection system of the energy storage power station, realizing the technical closure from the physical changes inside the battery to the abnormal power quality.

[0120] Preferably, step S3 includes the following steps:

[0121] Step S31: detecting the voltage distortion of the energy storage power station according to the abnormal trend of the power quality transmitted by the energy storage power station;

[0122] In this embodiment of the present invention, voltage distortion in energy storage power stations is detected through real-time acquisition and spectrum analysis of voltage signals at the power station's output. High-precision voltage sensors are deployed at the power station's voltage output terminals to collect real-time time-domain data of the three-phase voltage. The sensor sampling frequency is set to no less than 10 kHz to ensure complete capture of high-frequency harmonic signals. The collected voltage signals undergo analog-to-digital conversion and are transmitted to a data acquisition system. Subsequently, the collected voltage signals are spectrally decomposed using a fast Fourier transform (FFT) algorithm to obtain the amplitude and phase information of each voltage frequency component. The total harmonic distortion (THD) of the voltage is calculated, defined as the ratio of the effective values of all harmonic components to the effective value of the fundamental wave. By setting the fundamental frequency (e.g., 50 Hz or 60 Hz) as a reference, the amplitude distribution of the higher harmonics is decomposed to determine the voltage waveform distortion. This process is performed using a digital signal processing chip or a dedicated power quality analyzer, ensuring accurate and real-time calculations. Calculate the voltage's total harmonic distortion (THD), defined as the ratio of the effective values of all harmonic components to the effective value of the fundamental wave. By setting the fundamental frequency (such as 50Hz or 60Hz) as a reference, the amplitude distribution of higher harmonics is decomposed to determine the voltage waveform distortion. This process relies on digital signal processing chips or dedicated power quality analyzers to ensure accurate and real-time calculations.

[0123] Step S32: Identifying the power grid operation coordination attenuation trend based on the transmission voltage distortion of the energy storage power station;

[0124] In this embodiment of the present invention, identifying grid coordination degradation trends involves further analyzing the voltage distortion data transmitted by the energy storage power station obtained in step S31 to construct a mapping relationship between voltage waveform distortion indicators and physical parameters related to grid coordination. Increased voltage waveform distortion typically reflects grid load imbalance, resonance, or increased voltage fluctuations, which in turn impacts grid synchronization and stable operation. Time series analysis methods are used to extract trend characteristics from the voltage distortion data, including trend line slope, fluctuation amplitude, and periodic variation. Combined with historical grid operation data, the correlation between the voltage distortion indicator and grid coordination indicators is calculated. Coordination indicators may include grid frequency stability, phase difference change rate, and synchronous generator speed deviation. These indicators are collected using synchronized phasor measurement units (PMUs) deployed at key grid nodes. A grid coordination degradation trend curve is generated. This trend curve reveals the ongoing impact of voltage distortion on grid synchronization performance and uses numerical calculation methods to predict coordination degradation over a certain period of time. This step provides a quantitative description of the dynamic relationship between voltage distortion and grid coordination, supporting subsequent grid stability assessments.

[0125] Step S33: Detecting a short-term harmonic surge in the energy storage power station transmission based on the power grid operation coordination attenuation trend and the energy storage power station transmission voltage distortion;

[0126] In this embodiment of the present invention, detecting short-term harmonic surges transmitted by energy storage power stations requires a comprehensive assessment based on both grid coordination attenuation trends and voltage distortion. Using the voltage harmonic spectrum information obtained in step S31, the harmonic amplitude change rate is continuously calculated over a short period of time, with particular attention paid to sudden changes in the values of higher-order harmonics (such as the 11th, 13th, and higher). A short-term harmonic surge manifests as a sharp increase in harmonic amplitude on a timescale of milliseconds to seconds. Combined with the grid coordination attenuation trend curve calculated in step S32, the grid conditions underlying the harmonic surge event are analyzed. As coordination attenuation intensifies, the system's ability to suppress harmonics decreases, making harmonic amplitudes more likely to rise rapidly. A multi-layer threshold mechanism is established to determine the starting point and peak duration of the harmonic surge. The threshold parameters are quantified based on historical grid operation data and historical harmonic records of the energy storage power station. During implementation, a digital signal processing unit performs sliding window FFT analysis on the real-time voltage signal, with a window length of 100 ms and a sliding step of 10 ms, to accurately capture dynamic harmonic changes. Harmonic surge data is transmitted to the central monitoring system via a high-speed data link. Combined with the coordination trend, statistical analysis methods are used to determine the severity and frequency of short-term harmonic surges, and the time mark and amplitude parameters of the surge events are obtained.

[0127] Step S34: determining the attenuation status of the energy storage power station grid stability based on the short-term harmonic surge status transmitted by the energy storage power station.

[0128] In this embodiment of the present invention, the stability degradation status of the energy storage power station grid is determined based on the short-term harmonic surge parameters detected in step S33. Key indicators of short-term harmonic surge events are extracted, including peak surge amplitude, surge duration, and event frequency. Real-time operational data from the energy storage power station grid, such as current harmonic content, voltage fluctuation amplitude, and frequency deviation, are combined to form comprehensive stability evaluation parameters. Frequency domain analysis techniques are used to match the harmonic surge frequency range with the grid oscillation mode to determine the impact of harmonic surges on the grid oscillation mode, with particular attention paid to the interaction between low-frequency oscillations (0.1 Hz to 2 Hz) and high-frequency harmonics. This data is collected using a high-precision power quality monitor to ensure parameter accuracy. The stability degradation status is quantified using a predefined comprehensive indicator function. This function combines the amplitude and frequency of harmonic surges, surge event density, and other grid operating indicators to generate a stability score. This score reflects the impact of the energy storage power station on grid stability. Stability degradation status data is uploaded to a monitoring platform for subsequent grid scheduling and energy storage power station maintenance decision-making.

[0129] It is particularly important that step S32 includes the following steps:

[0130] Step S321: predicting the damage of the grid waveform structure based on the voltage distortion of the energy storage power station;

[0131] In this embodiment of the present invention, an online power quality monitoring device deployed at an energy storage power station collects high-frequency voltage waveforms at the transmission busbar port. The sampling frequency is set to above 10 kHz, and the sampling period is an hourly window within the continuous operating cycle. The acquired raw voltage data is decomposed into harmonic components using a Fourier transform, and the total voltage distortion (THDv) index is calculated. When the THDv exceeds the 5% threshold, the system further calculates the amplitude contribution of the 3rd, 5th, and 7th harmonics and performs a multi-factor analysis based on the fundamental frequency drift (based on 50 Hz). Combining the data time series over a continuous sampling period of 48 hours, a sliding window correlation analysis is performed on the changing trends of THDv, harmonic content, and fundamental phase offset. If THDv continues to rise, even harmonics fluctuate violently, or the fundamental phase drifts by more than ±15°, the current transmission voltage is considered to have damaged the grid waveform structure. The degree of damage is quantified using a multi-parameter weighted value W (composed of THDv, even / odd harmonic ratio, and fundamental phase shift ratio) with a range of 0–1. A value exceeding 0.65 indicates structural damage.

[0132] Step S322: predicting the growth trend of line loss of the energy storage power station based on the damage to the grid waveform structure;

[0133] In this embodiment of the present invention, the power transmission loss calculation module is activated based on the grid waveform structural damage quantification value W obtained in step S321 and the three-phase output line resistance of the energy storage power station (calculated based on actual measurements of the wiring length and copper conductor cross-section within the station). The module reads the measured current data for each line during different time periods (every 15 minutes) and calculates the active and reactive power losses based on the loss coefficients of the resistive and non-resistive components under distortion conditions. The specific formula uses the I²R formula to calculate the active power loss under distortion conditions, and the additional losses caused by harmonics are adjusted using the harmonic loss multiplier factor recommended by IEC 61000. The module also compares the line loss rates for the same period with historical data. If, assuming the structural damage coefficient W is greater than 0.65, the daily average line loss rate increases by more than 12% compared to the three-day moving average baseline, the event is marked as a line loss growth trend event, and the loss growth amplitude ΔL is output in kWh / day. A 24-hour loss growth trend curve is generated as a basis for further analysis.

[0134] Step S323: Detecting the degree of grid error accumulation of the energy storage power station based on the growth trend of the energy storage power station line loss and the damage to the grid waveform structure;

[0135] In this embodiment of the present invention, based on the loss growth amplitude ΔL calculated in step S322, the grid waveform structure damage quantification value W obtained in step S321 is also introduced. Using these two key parameters as inputs and combined with the load in / out power difference data from the energy storage power station operation log (using the difference between the actual PCS charge / discharge power and the EMS scheduling set power as the error source), the system performs a retrospective check on error accumulation. Using an error integral analysis method, the real-time power deviations within each scheduling cycle are summed and accumulated to form an error accumulation value Eacc (unit: kWh). The growth rate of Eacc is analyzed by combining the changing trend of ΔL with the fluctuation of W. When ΔL exceeds 8 kWh / day and Eacc increases by more than 25% within 72 hours, the system marks the error accumulation rate as exceeding the limit and outputs the total Eacc value. This value, combined with the waveform distortion source location (based on the phase offset between the busbar and branch monitoring points), forms an error source impact factor matrix, which serves as the input for coordination assessment.

[0136] Step S324: Identify the attenuation trend of grid operation coordination based on the degree of grid error accumulation of the energy storage power station.

[0137] In this embodiment of the present invention, the error accumulation value Eacc calculated in step S323 is combined with the error source impact factor matrix and input into the grid coordination evaluation module. This evaluation module performs analysis based on four technical indicators: regional dispatch plan, load response, bidirectional energy regulation history, and energy storage charging and discharging target achievement rate. In this module, if Eacc exceeds a set threshold (e.g., 100 kWh) and a single node has an error source impact factor exceeding 30%, a local coordination breakdown is determined. The system also retrieves the dispatch response delay (the time difference between the actual response start and the EMS instruction issuance) and the feedback dispatch offset rate from the past 96 hours. A linear cumulative trend analysis method is used to extract overall operational trends. If the average response delay increases by more than 20% of the original value and the dispatch offset rate is greater than 10%, it is further determined that a significant system coordination degradation trend has emerged. The coordination degradation index Cdeg is output as a percentage and is used for subsequent analysis of energy storage battery anomaly identification and correlation with battery box operating stability, forming an upstream and downstream logical closed loop.

[0138] Preferably, step S4 includes the following steps:

[0139] Step S41: Assessing the degree of thermal runaway risk of the energy storage power station based on the stability decay status of the power grid of the energy storage power station;

[0140] In an embodiment of the present invention, after determining the grid stability degradation status of an energy storage power station, the thermal runaway risk faced by the energy storage power station is assessed based on this degradation data. Specifically, multi-channel voltage and current synchronous sampling equipment deployed at the grid-connected node of the energy storage power station acquires indicators such as frequency disturbance amplitude, voltage sag rate, short-term current reverse peak value, and power reverse fluctuation period during the stability degradation process. These time-domain and frequency-domain indicators are input into a hardware calculation unit built into a digital logic gate array. A Fourier transform analysis component simultaneously calculates the frequency concentration of short-period disturbances and the power oscillation phase deviation. Subsequently, using a pre-defined critical parameter library, overheat warning thresholds associated with stability degradation (e.g., a current continuous rise rate exceeding 0.8 A / ms, a voltage strobe amplitude exceeding 5% of the rated value, a temperature rise rate exceeding 2°C / min, etc.) are extracted. Combined with the cabin temperature rise trend output by the battery compartment thermistor array, the thermal runaway risk is comprehensively determined using Boolean logic relationships. During this process, five risk levels are set, ranging from "extremely low" to "extremely high". The output of the current thermal runaway risk level of the energy storage power station is the third level (medium to high), and the subsequent data is transmitted with a value of 3.

[0141] Step S42: Detecting changes in battery status of the energy storage power station based on the thermal runaway risk level of the energy storage power station;

[0142] In an embodiment of the present invention, after the thermal runaway risk level is determined, the state changes of each battery box in the energy storage power station are monitored in real time based on this level. In actual operation, the system calls a three-dimensional multi-point temperature detection module (including sensors on the top plate, side walls, and bottom of the box) installed inside each battery box, and uses an infrared thermal imaging module to obtain an image of the abnormal heat dissipation distribution outside the battery box. To ensure the integrity and timeliness of the detection data, full parameter acquisition and image update operations are performed every 10 seconds, and preliminary data preprocessing is performed by edge computing nodes, including outlier removal, differential moving average denoising, and block thermal gradient enhancement. Subsequently, based on the thermal runaway risk level value obtained in the previous step, the threshold is set and dynamically adjusted. For example, when the risk level is 3, a battery temperature change rate exceeding 1.5°C / min is marked as abnormal, and the distribution location of the abnormal points is recorded. It further combines the charge amplitude change amplitude, terminal voltage drop slope and other electrical parameter change trends output by the SOC (state of charge) sensor module and the terminal voltage stability monitor to determine whether the battery has attenuation-type, overcharge-type or thermal coupling-type abnormal behavior. The status of each battery box is divided into four categories: "stable", "slight fluctuation", "trend abnormality" and "critical abnormality", and the detection results of each battery box are recorded to form a sequence of state change results corresponding to a set of numbers.

[0143] Step S43: Identifying the degree of grid frequency instability based on changes in the battery status of the energy storage power station;

[0144] In this embodiment of the present invention, after determining battery status changes, the degree of grid frequency instability in the energy storage power station is determined based on the trend of these changes. During implementation, the "Trend Abnormal" and "Critical Abnormal" items in the battery status change result sequence are first filtered out, and the grid-connected frequency data for these batteries during their corresponding operating periods is extracted. Frequency data is collected with 0.1Hz accuracy by a high-precision frequency meter located at the converter output port. Frequency fluctuations are recorded once per second, continuously tracking data changes over the last 30 minutes. By setting a frequency fluctuation window (e.g., ±0.3Hz) and applying the three-point central difference method to extract the frequency change rate, it is possible to determine whether local frequency oscillation or frequency rebound occurs. Combined with the temperature rise and SOC imbalance of the corresponding battery compartment, the impact of this type of status change on the grid frequency is determined based on logical judgment criteria such as whether the simultaneously measured frequency disturbance period coincides with the battery status change time window and whether the overlap rate exceeds 60%. The number of frequency fluctuations exceeding the limit, the average fluctuation amplitude, and the frequency change rate are used as the main evaluation factors. After comprehensive calculation, the grid frequency instability degree classification data is output and expressed in a five-level integer system. In this embodiment, the evaluation result is level 4 (high frequency instability state).

[0145] Step S44: Detect abnormal conditions of the energy storage power station batteries based on the degree of grid frequency instability and changes in the status of the energy storage power station batteries, and upload them to the cloud platform for early warning.

[0146] In this embodiment of the present invention, after identifying the degree of grid frequency instability, combined with previously detected battery status changes, comprehensive detection and early warning upload of abnormal battery conditions in the energy storage power station are implemented. In actual operation, the battery status level and frequency instability level are jointly analyzed using pre-set logical combination rules. For example, when the battery status level is "critically abnormal" and the frequency instability level is 4 or above, a battery abnormality flagging event is triggered. Once the triggering conditions are met, the system immediately activates a multi-parameter verification module and retrieves the battery compartment's operating history for the past 24 hours, including continuous records of terminal voltage, terminal current, SOC change rate, internal impedance change curve, and infrared thermal spectrum. This multi-indicator linkage determines whether the condition is a short-term disturbance or a persistent abnormal trend. If the latter is the case, an abnormality confirmation is performed, and the abnormality event number, location identifier (battery compartment number), abnormality start time, duration, abnormality type, and key associated parameters are encapsulated in a JSON structure format. The data is then uploaded to a designated cloud platform using a cellular communication module (such as an NB-IoT or 5G module) deployed at the edge node of the control center. After the upload is completed, the cloud platform receives and triggers the back-end early warning processing program, marks the abnormal battery box as requiring maintenance, and displays real-time abnormal information through the cloud management interface.

[0147] It is particularly important that step S41 includes the following steps:

[0148] Step S411: Detecting the increasing trend of grid frequency fluctuations based on the grid stability attenuation status of the energy storage power station;

[0149] In this embodiment of the present invention, during the operation of the energy storage power station, a grid frequency acquisition module installed at the AC busbar collects frequency change signals in real time. This module consists of a high-precision phase-locked loop (PLL) unit, a voltage sampling circuit, and a digital signal processing unit. The sampling frequency is set to 2000 Hz. The filtering unit uses a 50 Hz bandpass filter to extract the main frequency fundamental signal and then calculate the grid frequency value every second using the zero-crossing method. The resulting frequency data, f(t), is structured and stored at a one-minute granularity. The grid stability assessment module utilizes data from the past 30 days, including reactive voltage offset, power fluctuation frequency, intermittent dispatch response failure records, and busbar voltage asymmetry, to construct a grid stability attenuation factor sequence, C_deg(t). The current stability attenuation level is calculated based on this factor. If the current C_deg(t) exceeds the threshold of 0.35 (indicating a significant decline in grid stability), the grid frequency fluctuation growth trend identification process is initiated. Select a sliding time window T of 1 hour, extract the maximum frequency f_max and minimum frequency f_min within this window, and calculate the fluctuation amplitude Δf = f_max − f_min. Continuously collect Δf for the last 12 hours to construct a time series Δf(t). Perform a first-order difference calculation on this series to analyze the rate of change of Δf over time. If the Δf growth rate is positive three times in a row, exceeds 0.15 Hz, and is higher than 1.5 times the average of the past seven days, it is determined that there is an increasing trend in grid frequency fluctuations. This trend is represented by the tier parameter Fv_tier, which is divided into four levels: 0 (no growth), 1 (slight growth), 2 (moderate growth), and 3 (severe growth), and serves as input for subsequent steps.

[0150] Step S412: predicting a frequent growth trend in the response of the energy storage power station based on a growth trend in the frequency fluctuation of the power grid;

[0151] In this embodiment of the present invention, based on the frequency fluctuation level Fv_tier obtained in step S411, the energy storage power station dispatch response record analysis phase begins. The energy storage system dispatch log is accessed to count the number of times the energy storage system started and stopped in response to grid frequency deviation within the last 24 hours, denoted as N_resp. Response records include trigger timestamps, start power, end power, and response duration. A program sets the time segmentation to hourly levels, generating a response count time series N_resp(t). A differential calculation is performed on N_resp(t) to obtain the hourly change in response count, ΔN_resp(t) = N_resp(t) − N_resp(t−1). If ΔN_resp(t) is greater than or equal to 2 for each of the last three hours, and the current hourly response count increases by more than 50% compared to the weekly average for the same time period, a significant response frequency increase is determined. A response frequency increase status parameter, R_flag, is output, taking a value of 0 (no significant increase) or 1 (increasing response frequency). In addition, the trend sequence of the response times is plotted as a trend curve R_trend(t), the slope of which is used to measure the response growth rate and is combined with Fv_tier in step S411 to be passed to the next step.

[0152] Step S413: Detecting the continuous power conversion of the energy storage power station based on the increasing trend of the power grid frequency fluctuation;

[0153] In this embodiment of the present invention, the Fv_tier and R_flag obtained in steps S411 and S412, respectively, are used as pre-judgment parameters. If Fv_tier ≥ 2 and R_flag = 1, the power conversion behavior analysis process begins. The energy storage system output power value P(t) recorded by the PCS control unit is retrieved, and the power change record within a continuous time period is obtained. The starting power P_start, ending power P_end, start time T_start, and end time T_end of each response operation are extracted. The duration of each power conversion operation ΔT = T_end − T_start and the power change amplitude ΔP = |P_end − P_start| are calculated. After counting the number of responses N_resp(t) within each hour, the average duration T_avg and average power amplitude P_avg within that hour are calculated. If T_avg ≥ 10 minutes, P_avg ≥ 150 kW, and the standard deviation of ΔP is less than 10%, the current power conversion is determined to be continuous and stable. Set the power conversion trend level parameter Pw_conv_tier, which is divided into four levels: 0 (no conversion), 1 (fluctuating conversion), 2 (medium-intensity stable conversion), and 3 (high-intensity, high-frequency continuous conversion). This parameter serves as a key input for the next step of thermal runaway risk assessment.

[0154] Step S414: Evaluate the risk level of thermal runaway of the energy storage power station based on the continuous power conversion of the energy storage power station and the frequent growth trend of the energy storage power station response.

[0155] In an embodiment of the present invention, the response frequent growth flag R_flag of step S412 and the power conversion level Pw_conv_tier of step S413 are combined to perform a multi-parameter cross-judgment on the thermal runaway risk of the energy storage battery box. The thermocouple temperature sensor installed on the surface of the battery module is called to collect the internal temperature T_in(t) and the shell temperature T_shell(t) of the battery box every 30 seconds. The temperature change trend is determined by analyzing the temperature rise rate dT / dt of T_in(t). When dT / dt ≥ 0.5℃ / min for 10 consecutive minutes and Pw_conv_tier ≥ 2, R_flag = 1, further analysis is performed to determine whether the cooling system is at its operating limit. The fan speed rpm_fan in the cooling system and the refrigerant flow rate v_coolant in the heat exchange pipe are called. If rpm_fan is lower than 80% of the rated speed and the refrigerant flow rate does not increase significantly, it is determined that the cooling response is lagging and the heat dissipation redundancy is insufficient. A comprehensive risk scoring function, H_risk, is constructed, with weights including the temperature rise rate (weight 0.4), power conversion level (weight 0.3), response frequency (weight 0.2), and cooling efficiency hysteresis (weight 0.1). The function outputs a thermal runaway risk index value ranging from 0 to 1. When H_risk ≥ 0.75, the system marks the current battery box as high-risk, adds the number to the "Abnormal Battery Box Risk List," activates the remote alarm system, and sends an early warning command to the maintenance platform via the SCADA system, enabling dynamic response processing.

[0156] The present invention also provides an abnormal battery box detection system for an energy storage power station, which is used to execute the abnormal battery box detection method for an energy storage power station as described above. The abnormal battery box detection system for an energy storage power station comprises:

[0157] The energy storage power station operation status evaluation module is used to obtain the energy storage power station battery box data; collect the energy storage power station operation environment data based on the energy storage power station battery box data; evaluate the energy storage power station battery box initial performance based on the energy storage power station battery box data; and evaluate the energy storage power station operation status data based on the energy storage power station battery box initial performance based on the energy storage power station battery box data and the energy storage power station operation environment data;

[0158] The transmission power quality anomaly detection module is used to detect the imbalance of the supporting configuration evolution of the energy storage station based on the operating status data of the energy storage station; detect the steady-state collapse of the battery electrothermal coupling based on the imbalance of the supporting configuration evolution of the energy storage station; and detect the abnormal trend of the transmission power quality of the energy storage station based on the steady-state collapse of the battery electrothermal coupling;

[0159] The grid stability attenuation determination module is used to identify the grid operation coordination attenuation trend based on the abnormal power quality trend of the energy storage station transmission power; detect the short-term harmonic surge condition of the energy storage station transmission based on the grid operation coordination attenuation trend and the abnormal power quality trend of the energy storage station transmission power; and determine the grid stability attenuation condition of the energy storage station based on the short-term harmonic surge condition of the energy storage station transmission power;

[0160] The energy storage power station battery anomaly detection module is used to detect changes in the battery status of the energy storage power station based on the attenuation of the energy storage power station grid stability; identify the degree of grid frequency instability based on the changes in the battery status of the energy storage power station; detect abnormal conditions of the energy storage power station battery based on the degree of grid frequency instability and changes in the battery status of the energy storage power station, and upload them to the cloud platform for early warning.

[0161] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting abnormal battery boxes in energy storage power stations, characterized in that: The following steps are involved: Step S1: Acquire battery box data of the energy storage power station; collect energy storage power station operating environment data for the energy storage power station battery box; evaluate the initial performance of the energy storage power station battery box based on the energy storage power station battery box data; evaluate the energy storage power station operating status data based on the initial performance of the energy storage power station battery box according to the energy storage power station battery box data and the energy storage power station operating environment data; Step S2: detecting an imbalance in the evolution of the supporting configuration of the energy storage station based on the operating status data of the energy storage station; detecting a steady-state collapse of the battery electrothermal coupling based on the imbalance in the evolution of the supporting configuration of the energy storage station; and detecting an abnormal trend in the quality of power transmitted by the energy storage station based on the steady-state collapse of the battery electrothermal coupling. Step S3: Identify the trend of grid operation coordination attenuation based on the abnormal trend of power quality transmitted by the energy storage power station; Detect the surge of short-term harmonics transmitted by energy storage power stations based on the attenuation trend of grid operation coordination and the abnormal trend of power quality transmitted by energy storage power stations; Determine the stability attenuation of the energy storage power station grid based on the short-term harmonic surge of the energy storage power station transmission; Step S4: Detect changes in the battery status of the energy storage station based on the attenuation of the stability of the energy storage station grid; identify the degree of grid frequency instability based on the changes in the battery status of the energy storage station; detect abnormal conditions of the energy storage station battery based on the degree of grid frequency instability and the changes in the battery status of the energy storage station, and upload them to the cloud platform for early warning.

2. The method for detecting abnormal battery boxes in an energy storage power station according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: setting the temperature measurement range of the temperature sensor to 0-200°C, the minimum temperature change to 0.05°C, and the temperature sampling frequency to 10Hz; Step S12: setting the humidity sensor's humidity measurement range to 0% to 100% RH, the minimum humidity change to 0.1% RH, and the humidity sampling frequency to 5 Hz; Step S13: Acquire battery box data of the energy storage power station; Step S14: using a humidity sensor and a temperature sensor to collect energy storage power station operating environment data from the energy storage power station battery box; Step S15: Collecting the cumulative operation time data of the energy storage power station based on the battery box data of the energy storage power station; Step S16: Evaluate the initial performance of the energy storage power station battery box based on the energy storage power station battery box data; Step S17: Evaluate the initial performance of the battery box of the energy storage station and the energy storage station operating status data based on the accumulated operating time data of the energy storage station and the operating environment data of the energy storage station.

3. The method for detecting abnormal battery boxes in an energy storage power station according to claim 2, characterized in that: Step S16 includes the following steps: Step S161: extracting internal arrangement structure data of the battery box based on the battery box data of the energy storage power station; Step S162: Identifying the internal conductive connection status of the battery box based on the internal arrangement structure data of the battery box; Step S163: Evaluate the regularity of the battery box busbar wiring based on the internal conductive connection status of the battery box and the internal arrangement structure data of the battery box; Step S164: Evaluate the heat dissipation capacity of the battery box based on the regularity of the battery box busbar wiring and the internal arrangement structure data of the battery box; Step S165: Calculating the initial voltage deviation of the energy storage station battery according to the energy storage station battery box data; Step S166: Determine the initial operating stability of the battery box of the energy storage power station based on the internal heat dissipation capacity data of the battery box and the initial voltage deviation of the battery of the energy storage power station; Step S167: Evaluate the initial performance of the energy storage power station battery box based on the initial operation stability of the energy storage power station battery box.

4. The method for detecting abnormal battery boxes in an energy storage power station according to claim 2, characterized in that: Step S17 includes the following steps: Step S171: collecting abnormal fluctuations in operating environment temperature and humidity based on the operating environment data of the energy storage power station; Step S172: determining the sudden change of the operating environment's temperature and humidity during the day and night based on the abnormal fluctuation of the operating environment's temperature and humidity; Step S173: Identifying the oxidation trend of the energy storage station structure based on the initial performance of the energy storage station battery box when the sudden change of temperature and humidity in the operating environment exceeds 30% during the day and night; Step S174: testing the pressure-bearing operation state of the energy storage station when the cumulative operation time of the energy storage station exceeds 48 hours and the oxidation trend of the energy storage station structure exceeds 50%; Step S175: estimating the overload condition of the energy storage power station according to the load-carrying operation status of the energy storage power station; Step S176: Evaluate the energy storage station operation status data based on the initial performance of the energy storage station battery box and the energy storage station structure oxidation trend.

5. The method for detecting abnormal battery boxes in an energy storage power station according to claim 1, characterized in that: The detection of imbalance of the support configuration evolution of the energy storage power station in step S2 includes: Identify the degree of oxidation of the energy storage station connection structure based on the energy storage station operation status data; Determine the loosening trend of the energy storage station connection structure based on the oxidation degree of the energy storage station connection structure; Estimate the micro-vibration of the energy storage station connection based on the loosening trend of the energy storage station connection structure; Measure the degree of mechanical stress growth of the energy storage power station structure based on the micro-vibration of the energy storage power station connection; Predict the micro-deformation of the energy storage power station structure based on the degree of mechanical stress growth of the energy storage power station structure and the micro-vibration of the energy storage power station connection; Detect the dynamic fatigue condition of the energy storage power station structure based on the micro-deformation condition of the energy storage power station structure and the degree of mechanical stress growth of the energy storage power station structure; The imbalance of the supporting configuration evolution of the energy storage power station is detected based on the dynamic fatigue condition and micro-deformation condition of the energy storage power station structure.

6. The method for detecting abnormal battery boxes in an energy storage power station according to claim 1, characterized in that: The battery electrothermal coupling steady-state disintegration condition detection in step S2 includes: Identify abnormal extrusion inside batteries based on the imbalance of the supporting structure evolution of the energy storage power station; Measure the lateral force imbalance of the battery cell based on the abnormal extrusion inside the battery; Determine the horizontal displacement of the battery cell based on the degree of lateral force imbalance of the battery cell and the abnormal extrusion inside the battery; Measure the imbalance of insulation distance between battery cells based on the horizontal displacement of battery cells; Determine the degradation of the electrical connection performance of the energy storage power station based on the imbalance of the insulation distance between battery cells; Measure the distortion of the heat dissipation path of the battery cells based on the imbalance of the insulation distance between the cells; The steady-state collapse of the battery's electrothermal coupling is detected based on the distortion of the battery cell's heat dissipation path and the degradation of the electrical connection performance of the energy storage power station.

7. The method for detecting abnormal battery boxes in an energy storage power station according to claim 1, characterized in that: The abnormal trend detection of power quality transmission from the energy storage station in step S2 includes: Detect the continuous growth of the battery's internal resistance to current transmission based on the battery's electrothermal coupling steady-state disintegration condition; Determine the battery transmission current shunt imbalance condition based on the continuous growth of the battery transmission current internal resistance; Predict abnormal overcharge of batteries in energy storage power stations based on the imbalance of battery transmission current shunt; Determine the damage trend of battery active materials based on abnormal overcharge conditions of batteries in energy storage power stations; Test the degree of damage to the battery electrode interface synergy mechanism based on the collapse trend of battery active materials; Estimate the limitation of ion diffusion channels inside the battery based on the degree of damage of the battery electrode interface synergistic mechanism and the collapse trend of the battery active materials; The abnormal trend of power quality transmitted by energy storage power station is detected based on the restricted ion diffusion channels inside the battery.

8. The method for detecting abnormal battery boxes in an energy storage power station according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: detecting the voltage distortion of the energy storage power station according to the abnormal trend of the power quality transmitted by the energy storage power station; Step S32: Identifying the power grid operation coordination attenuation trend based on the transmission voltage distortion of the energy storage power station; Step S33: Detecting a short-term harmonic surge in the energy storage power station transmission based on the power grid operation coordination attenuation trend and the energy storage power station transmission voltage distortion; Step S34: determining the attenuation status of the energy storage power station grid stability based on the short-term harmonic surge status transmitted by the energy storage power station.

9. The method for detecting abnormal battery boxes in an energy storage power station according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Assessing the risk of thermal runaway of the energy storage power station based on the stability decay status of the power grid of the energy storage power station; Step S42: Detecting changes in battery status of the energy storage power station based on the thermal runaway risk level of the energy storage power station; Step S43: Identifying the degree of grid frequency instability based on changes in the battery status of the energy storage power station; Step S44: Detect abnormal conditions of the energy storage power station batteries based on the degree of grid frequency instability and changes in the status of the energy storage power station batteries, and upload them to the cloud platform for early warning.

10. An abnormal battery box detection system for an energy storage power station, characterized in that: For executing the abnormal battery box detection method of the energy storage power station according to claim 1, the abnormal battery box detection system of the energy storage power station comprises: The energy storage power station operation status evaluation module is used to obtain the energy storage power station battery box data; collect the energy storage power station operation environment data for the energy storage power station battery box; evaluate the initial performance of the energy storage power station battery box based on the energy storage power station battery box data; and evaluate the energy storage power station operation status data based on the initial performance of the energy storage power station battery box according to the energy storage power station battery box data and the energy storage power station operation environment data; The transmission power quality anomaly detection module is used to detect the imbalance of the supporting configuration evolution of the energy storage station based on the operating status data of the energy storage station; detect the steady-state collapse of the battery electrothermal coupling based on the imbalance of the supporting configuration evolution of the energy storage station; and detect the abnormal trend of the transmission power quality of the energy storage station based on the steady-state collapse of the battery electrothermal coupling; The grid stability attenuation determination module is used to identify the grid operation coordination attenuation trend based on the abnormal power quality trend of the energy storage station transmission power; detect the short-term harmonic surge condition of the energy storage station transmission based on the grid operation coordination attenuation trend and the abnormal power quality trend of the energy storage station transmission power; and determine the grid stability attenuation condition of the energy storage station based on the short-term harmonic surge condition of the energy storage station transmission power; The energy storage power station battery anomaly detection module is used to detect changes in the battery status of the energy storage power station based on the attenuation of the energy storage power station grid stability; identify the degree of grid frequency instability based on the changes in the battery status of the energy storage power station; detect abnormal conditions of the energy storage power station battery based on the degree of grid frequency instability and changes in the battery status of the energy storage power station, and upload them to the cloud platform for early warning.

Images