Energy storage power station operation monitoring management system and method

By building a node health status evaluation model and differentiated power allocation, the problem of insufficient management of node aging in energy storage power stations is solved, and efficient, safe and stable energy storage system operation is achieved to adapt to the fluctuations in new energy generation.

CN120262703AActive Publication Date: 2025-07-04SHENZHEN HONCELL ENERGY CO LTD

Patent Information

Application Number
CN202510756851.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-06-09
Publication Date
2025-07-04
Estimated Expiration
2045-06-09

AI Technical Summary

Technical Problem

Traditional energy storage power station management systems cannot manage the aging differences in energy storage nodes in a refined manner, resulting in the expansion of performance differences between nodes, insufficient accuracy of health status assessment, and cannot accurately reflect the true aging degree, affecting the safe and stable operation of the system.

Method used

By obtaining voltage rebound timing data and charge state data, a node health status evaluation model is constructed, a node charge and discharge power allocation model is established, and power distribution is optimized by using the minimum power disturbance method to form a closed-loop optimization mechanism to achieve differentiated management.

Benefits of technology

It improves the energy conversion efficiency of the energy storage system, extends the life of node equipment, enhances the safety, stability and adaptability of the system, improves load balancing and response speed, reduces the risk of failure, and adapts to the fluctuations in new energy generation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120262703A_ABST
    Figure CN120262703A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of electric energy storage systems, in particular to an energy storage power station operation monitoring management system and method. The method comprises the following steps: obtaining voltage data of each energy storage node in a standing state, and carrying out periodic voltage sampling within a preset standing time, thereby obtaining voltage springback time sequence data; performing temperature correction processing on the voltage springback time sequence data so as to obtain standardized voltage springback characteristic data; acquiring charge state data of each energy storage node before and after operation; calculating a charge-discharge depth change value of each node according to the charge state data; and constructing a node health state evaluation model according to the standardized voltage rebound characteristic data and the charge and discharge depth change value, and obtaining node aging differentiation parameters. According to the method, the node charging and discharging power allocation model is constructed based on the node aging differentiation parameters, differentiation management of energy storage nodes in different states is achieved, and the problem that the aging degree between the nodes is further expanded in a traditional method is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of electrical energy storage systems, and particularly to an operation monitoring and management system and method for an energy storage power station. Background Art

[0002] Operation monitoring and management of an energy storage power station is a core technology to ensure the safe, efficient, and stable operation of an energy storage system. It aims to achieve the best performance and economic benefits of energy storage devices through real-time monitoring, intelligent scheduling, and scientific management. Its core architecture includes data acquisition, communication transmission, an energy management system (EMS), a power conversion system (PCS), a thermal management system, and a monitoring platform and cloud services. First, key parameters such as the voltage, current, SOC (state of charge), SOH (state of health), and temperature of the battery are collected in real time by sensors, and the data is transmitted to the monitoring system using a data collector and a communication network (such as industrial Ethernet, wireless communication). As the core of operation management, the energy management system dynamically adjusts the charge and discharge strategies of the energy storage system through functions such as real-time monitoring, status evaluation, fault diagnosis, and optimization scheduling to ensure efficiency and safety. For example, the EMS can formulate charge and discharge plans based on electricity price changes, load demand, and new energy generation conditions to achieve peak shaving and valley filling or participate in power market transactions. At the same time, the PCS is responsible for the energy conversion between the energy storage power station and the power grid or load, and improves the power quality by optimizing the power factor and quickly responding to grid fluctuations. The thermal management system prevents the battery from overheating or overcooling by monitoring the temperature in real time and controlling heat dissipation, thereby extending the equipment life. However, traditional energy storage power stations usually adopt a unified power distribution strategy, evenly distributing energy storage tasks to all energy storage nodes. This method ignores the different degrees of aging that different energy storage nodes will undergo during actual use, resulting in an increasing performance difference between the internal nodes of the battery pack. As the operation time extends, there will be obvious differences in capacity attenuation, internal resistance change, and response ability among energy storage nodes, and the traditional method cannot perform refined management for this difference. Previous energy storage power station management systems mainly evaluated the state of energy storage nodes through simple voltage monitoring, lacking in-depth analysis of an important parameter, namely the voltage rebound characteristic. The voltage rebound characteristic is an important indicator of the health state of battery nodes, but traditional methods often ignore the influence of factors such as temperature and charge and discharge history on the voltage rebound process, resulting in insufficient accuracy in health state assessment and inability to accurately reflect the true aging degree of each node. Summary of the Invention

[0003] Based on this, it is necessary for the present invention to provide an operation monitoring and management system and method for an energy storage power station to solve at least one of the above technical problems.

[0004] To achieve the above object, an operation monitoring and management method for an energy storage power station includes the following steps: Step S1: Obtain the voltage data of each energy storage node in the static state, and perform periodic voltage sampling within a preset static time to obtain voltage rebound time series data; perform temperature correction processing on the voltage rebound time series data to obtain standardized voltage rebound characteristic data; Step S2: Obtain the state of charge data of each energy storage node before and after operation; calculate the charge and discharge depth change values of each node according to the state of charge data; construct a node health state evaluation model based on the standardized voltage rebound characteristic data and the charge and discharge depth change values to obtain node aging differentiation parameters; Step S3: Determine the available capacity level of each node according to the node aging differentiation parameters; calculate the optimal working range of each node based on the available capacity level to obtain node operation ability classification data; establish a node charge and discharge power allocation model using the node operation ability classification data; Step S4: Obtain the overall power scheduling instruction of the energy storage power station; decompose the overall power scheduling instruction according to the node charge and discharge power allocation model to obtain a node differentiation power allocation plan; perform dynamic optimization on the node differentiation power allocation plan using the minimum power perturbation method to obtain an adaptive energy flow allocation instruction; Step S5: Send the adaptive energy flow allocation instruction to each energy storage node, and monitor the real-time operation state of each node after executing the adaptive energy flow allocation instruction; update the node aging differentiation parameters based on the real-time operation state to form a closed-loop optimized operation monitoring and management mechanism for the energy storage power station.

[0005] The present invention also provides an operation monitoring and management system for an energy storage power station, which is used to execute the above-mentioned operation monitoring and management method for the energy storage power station. The operation monitoring and management system for the energy storage power station includes: A voltage characteristic acquisition module, which is used to obtain the voltage data of each energy storage node in the static state, and perform periodic voltage sampling within a preset static time to obtain voltage rebound time series data; perform temperature correction processing on the voltage rebound time series data to obtain standardized voltage rebound characteristic data; A health state evaluation module, which is used to obtain the state of charge data of each energy storage node before and after operation; calculate the charge and discharge depth change values of each node according to the state of charge data; construct a node health state evaluation model based on the standardized voltage rebound characteristic data and the charge and discharge depth change values to obtain node aging differentiation parameters; A capacity classification and allocation module, which is used to determine the available capacity level of each node according to the node aging differentiation parameters; calculate the optimal working range of each node based on the available capacity level to obtain node operation ability classification data; establish a node charge and discharge power allocation model using the node operation ability classification data; The energy flow intelligent distribution module is used to obtain the overall power scheduling instruction of the energy storage power station; decompose and process the overall power scheduling instruction according to the node charge and discharge power allocation model to obtain a node-differentiated power allocation scheme; dynamically optimize the node-differentiated power allocation scheme by using the minimum power perturbation method to obtain an adaptive energy flow allocation instruction; The closed-loop monitoring and optimization module is used to send the adaptive energy flow allocation instruction to each energy storage node and monitor the real-time operation status of each node after executing the adaptive energy flow allocation instruction; update the node aging differentiation parameters based on the real-time operation status to form a closed-loop optimized operation monitoring and management mechanism for the energy storage power station.

[0006] Through refined power allocation and dynamic state monitoring, the present invention realizes the efficient, reliable and safe operation of the energy storage system. First, based on in-depth analysis of the node operation capacity and charge-discharge efficiency, the method can reasonably allocate the basic power weights and scheduling priorities of each energy storage node, ensuring that nodes in different performance states can all exert their maximum potential and avoiding resource waste and system imbalance caused by node performance differences. This scheduling interval design based on the node efficiency curve and the safe operating range not only improves the overall energy conversion efficiency of the system but also extends the service life of the node equipment, reducing losses and failure risks caused by overload or unreasonable operating states. Secondly, the dynamic adjustment mechanism combined with the real-time state of charge and temperature change of the node can flexibly adjust the power allocation during actual operation to adapt to the fluctuations of the node state and external environment changes. For example, according to the ratio of the state of charge to the boundary of the priority scheduling interval, a linear attenuation adjustment is implemented to effectively prevent performance degradation of the node caused by frequent entry and exit from the extreme operating range. At the same time, the introduction of the temperature correction coefficient can respond to node temperature anomalies in real time, reduce the power output, ensure the node operates within the safe temperature range, avoid the occurrence of thermal runaway, and further improve the system safety and stability. The comprehensive allocation of such multi-dimensional parameters significantly enhances the adaptability and self-protection ability of the energy storage power station. In addition, the analysis of power scheduling parameters and the application of the node power allocation model realize the preliminary power allocation of the energy storage node and the scientific determination of the constraint conditions, ensuring that the power output of each node is within its safe carrying range and avoiding overload risks. The adjustment based on the deviation between the state of charge of the node and the optimal operating range makes the power allocation scheme more balanced, improves the overall load balancing level of the system, reduces the pressure fluctuations of local nodes, and effectively prevents the accelerated decline of battery capacity and local overheating phenomena caused by unbalanced scheduling. At the same time, the temperature dynamic monitoring and risk coefficient adjustment mechanism enable the power allocation scheme to have the ability of adaptive adjustment, which can optimize the power output in a timely manner according to the real-time temperature change of the node, improving the operation flexibility and response speed of the system. The setting of the minimum step size for power disturbance and the generation of multiple groups of disturbance schemes enable the system to carefully analyze the sensitivity of the node to power changes, accurately identify the nodes with response hysteresis, and through priority adjustment and iterative optimization, achieve the optimal allocation of power response. This process ensures that the energy storage power station can quickly and stably adjust the output in the face of external power demand fluctuations, reducing system oscillation and response lag, and improving the accuracy and efficiency of scheduling. The generation and execution of the adaptive energy flow allocation instruction enable the energy storage nodes to work efficiently and cooperatively according to the optimal scheme, improving the dynamic regulation ability and load response speed of the entire energy storage system, and being applicable to scenarios with large fluctuations in new energy power generation, such as power smoothing regulation of wind power and photovoltaic power.Real-time operation status monitoring and online analysis, combined with high-frequency sampling and dynamic evaluation of voltage, current, temperature, and state of charge, can capture the performance changes of nodes in real time, promptly reflect key indicators such as power response delay, voltage fluctuation, and temperature change rate, and provide a scientific basis for system operation. The dynamic update of aging parameters based on a sliding time window ensures the timeliness and accuracy of node health status assessment, can effectively capture the trend of node performance degradation, and guide subsequent maintenance and scheduling decisions. Combining the closed-loop calibration of the health model, the accurate management of the node health status of the energy storage power station is realized, the failure risk is minimized, the service life of the equipment is extended, and the economy and stability of system operation are improved. Overall, this method realizes multi-dimensional, dynamic, and intelligent operation monitoring and power allocation of the energy storage power station. While meeting the external power dispatching requirements, it fully considers the internal state and aging conditions of the nodes, ensures the safe and stable operation of the system and efficient energy utilization. Its flexibility and self-adaptability enable the energy storage power station to cope with complex and changing power grid environments, effectively support the large-scale access of new energy and power grid peak shaving requirements, improve the reliability and economy of power grid operation, and provide solid technical support for the wide application of energy storage technologies. Brief Description of the Drawings

[0007] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments read in conjunction with the accompanying drawings: Figure 1 It is a schematic diagram of the step flow of a method for operating and monitoring an energy storage power station of the present invention; Figure 2 For Figure 1 a detailed step flow diagram of step S1 in Detailed Embodiments

[0008] The technical method of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0009] In addition, the accompanying drawings are only schematic diagrams of the present invention and are not necessarily drawn to scale. The same reference numerals in the drawings represent the same or similar parts, and thus their repeated description will be omitted. Some of the block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor methods and / or microcontroller methods.

[0010] It should be understood that although terms such as "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly, the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0011] To achieve the above object, please refer to Figures 1 to 2 , the present invention provides a method for monitoring and managing the operation of an energy storage power station, and the method includes the following steps: Step S1: Obtain the voltage data of each energy storage node in the static state, and perform periodic voltage sampling within a preset static time to obtain voltage rebound time series data; perform temperature correction processing on the voltage rebound time series data to obtain standardized voltage rebound characteristic data; In the static stage before the operation of the energy storage power station in the embodiment of the present invention, the control system first sends a control instruction to enter the static state to all energy storage nodes (such as a battery cluster composed of lithium iron phosphate battery cells), so that each node is stable for at least 30 minutes without external charge and discharge disturbances (this time is the preset static time, which can be set within the range of 15 minutes to 60 minutes according to the battery type). During the static period, the terminal voltage of each node is periodically collected at a frequency of 1 Hz for 5 minutes to obtain the time series data of the natural voltage rebound over time. The voltage rebound phenomenon refers to the physical process in which the voltage of the battery gradually rises due to the rebalancing of the internal ion concentration after the battery stops charging and discharging, and its curve shape is related to the internal health status of the battery. Subsequently, the corresponding temperature value is recorded by using the temperature sensor configured for the node, and the time series data is normalized according to the temperature. For example, taking the standard temperature of 25 °C as the reference, the voltage rebound curves at different temperatures are adjusted to the curves equivalent to 25 °C through the temperature correction coefficient, so as to remove the interference of the ambient temperature on the voltage rebound characteristics and obtain the standardized voltage rebound characteristic data of each node.

[0012] Step S2: Obtain the state of charge data of each energy storage node before and after operation; calculate the change value of the charge and discharge depth of each node according to the state of charge data; construct a node health status evaluation model according to the standardized voltage rebound characteristic data and the change value of the charge and discharge depth to obtain node aging differentiation parameters; When the embodiment of the present invention prepares to start the operation task of the energy storage power station, the initial state of charge data (State of Charge, abbreviated as SOC) of each node is collected, usually expressed as a percentage between 0% and 100%. Combining the termination SOC data collected again after the operation ends, the charge-discharge depth change value during this round of charge-discharge process is calculated through the difference between the two SOCs. To improve the evaluation accuracy, the SOC collection accuracy is set to ±0.5%. Then, the standardized voltage rebound characteristic data and the charge-discharge depth change value are jointly input into the constructed node health state evaluation model. This model is trained based on the decision tree regression algorithm and combined with multiple rounds of historical operation data, and can output the aging differentiation parameters of each energy storage node. This parameter comprehensively reflects the aging characteristics such as node capacity attenuation, internal resistance growth, and thermal stability decline, usually expressed as a decimal between 0 and 1. The larger the value, the higher the aging degree. The aging differentiation parameter will be used as the core input basis for subsequent energy management optimization.

[0013] Step S3: Determine the available capacity level of each node according to the node aging differentiation parameter; calculate the optimal working range of each node based on the available capacity level to obtain the node operation ability classification data; establish a node charge-discharge power allocation model using the node operation ability classification data; Based on the aging differentiation parameter, the embodiment of the present invention divides each node into three types of available capacity levels: high, medium, and low. For example, it is set that the aging differentiation parameter less than 0.3 is the high capacity level, between 0.3 and 0.7 is the medium capacity level, and higher than 0.7 is the low capacity level. Then, by combining the analysis of the typical safe working voltage range and temperature rise curve of each node's recent three operations, its optimal working range is determined. For example, the high capacity level node can be set to the SOC range of 10% to 90%, the medium level node is set to 15% to 85%, and the low capacity level node is restricted within 20% to 80%, thus forming the node operation ability classification data. Then, use the operation ability classification data to construct a node charge-discharge power allocation model. This model is designed through constraint optimization. Input the total power required by the power station and the node operation ability level data, and output a set of single-node power distribution sets that meet the node ability limitations and are achievable for subsequent power scheduling. This model has the ability to flexibly adjust and can automatically update its power upper and lower limits according to the node health changes.

[0014] Step S4: Obtain the overall power scheduling instruction of the energy storage power station; decompose the overall power scheduling instruction according to the node charge-discharge power allocation model to obtain the node differential power distribution plan; use the minimum power perturbation method to dynamically optimize the node differential power distribution plan to obtain the adaptive energy flow allocation instruction; When the power grid dispatching center issues an overall power dispatching instruction to the energy storage power station in an embodiment of the present invention (for example, in a daily load regulation scenario, issuing a control task of "output 10 MW of electricity for 1 hour"), the system calls the aforementioned charge and discharge power allocation model to decompose the dispatching task, and obtains a differential power allocation scheme that meets the available capacity and health level limits of each node. For example, the 10 MW task is allocated as node A outputs 2.5 MW, node B outputs 1.8 MW, node C outputs 0.7 MW, etc. Then, to further reduce the node load fluctuation, the minimum power perturbation method is applied for dynamic optimization. The core of this method is to minimize the change amount between the current allocated power and the power at the previous moment as the objective function, ensuring that the thermal and electrical stresses caused by the power change to the node are minimized, thereby prolonging the node life. After iterative optimization and solution, the final adaptive energy flow allocation instruction for controlling each node is obtained, and each instruction includes parameters such as target power, current direction, and lower limit of safe voltage.

[0015] Step S5: Send the adaptive energy flow allocation instruction to each energy storage node, and monitor the real-time operation status of each node after executing the adaptive energy flow allocation instruction; update the node aging differential parameters based on the real-time operation status, and form a closed-loop optimized operation monitoring and management mechanism for the energy storage power station.

[0016] In an embodiment of the present invention, the adaptive energy flow allocation instruction is sent to the battery management system (BMS) and the power conversion system (PCS) of the corresponding energy storage node through the substation control system. The instruction issuance frequency is set to be updated every 30 seconds to ensure that the system has the ability to respond to real-time fluctuations. Subsequently, continuously monitor the real-time operation status data of the node, including key indicators such as current, voltage, temperature, and SOC change rate. If it is found that an abnormality occurs during the operation of a certain node (such as a rapid increase in temperature or the output power deviates from the preset value by more than 5%), the system immediately records the abnormality and marks the node as an abnormally operating state node. After each complete power dispatching cycle, the aging differential parameters are updated according to the actual performance data of the nodes in this round, and the historical values are weighted adjusted in a sliding window manner to achieve a dynamic closed-loop for aging identification. The above process is cyclically executed in each operation cycle to realize a full-process closed-loop operation monitoring and management mechanism for the energy storage power station from node health monitoring, capacity assessment to dynamic optimized dispatching and feedback correction, which can significantly improve the system dispatching efficiency and node service life in a typical peak shaving and valley filling operation scenario.

[0017] Preferably, step S1 includes the following steps: Step S11: Determine the static time parameter according to the pre-acquired cell model characteristics of each energy storage node, so as to obtain the static time configuration data; In the initial stage of the deployment of the energy storage power station in the embodiments of the present invention, the system management module will preset the cell model information of each energy storage node. For example, node A uses the lithium iron phosphate model IFR-32700, and node B uses the ternary lithium NCM-622 type. The rest time parameter is obtained by querying the rest characteristic technical data sheet provided by the corresponding cell manufacturer. The rest time refers to the time required for the internal electrochemical reaction of the battery to tend to balance and the surface voltage to stabilize after the battery stops charging and discharging, which generally depends on its electrochemical system, cycle life state and temperature sensitivity. Taking lithium iron phosphate as an example, the recommended rest time is 30 minutes, while for the ternary lithium model, due to its high activity and fast recovery, it is generally 20 minutes. For the convenience of management, the system sets the rest time configuration data according to the cell model classification and stores the result in the operation management database in the form of a structured configuration table for subsequent rest control.

[0018] Step S12: Control the energy storage node to enter the rest mode according to the rest time configuration data, interrupt all charge and discharge operations, and obtain the node rest state confirmation signal; In the embodiments of the present invention, according to the rest time configuration data generated in the previous step, the system scheduling module issues a "enter the rest mode" control instruction to all energy storage nodes through the energy storage management system (EMS). This instruction contains specific rest duration parameters and forcibly interrupts all current charge and discharge tasks of the node. The interruption operation controls the current to zero through the PCS module and disconnects the relay connected to the bus, thereby achieving electrical isolation. In an actual scenario, to ensure the correct execution of the instruction, the system needs to monitor the BMS status return signal of each node. When the three conditions of "no current load, stable voltage curve, and control relay disconnected" returned by the node are all met, it is determined that the node has entered the rest state, and a node rest state confirmation signal is generated. The confirmation signals of all nodes are aggregated by the station control host and used as the trigger condition for starting the subsequent data acquisition process.

[0019] Step S13: After obtaining the node rest state confirmation signal, start the periodic voltage sampling process to record the voltage change at the end of each energy storage node based on the preset sampling frequency, and obtain the original voltage time series sampling data; In the embodiments of the present invention, when the system receives the rest state confirmation signals of all energy storage nodes, it immediately starts the periodic voltage sampling program. This program is controlled by the data acquisition unit (DAQ) through the CAN bus or Modbus interface to control the node BMS module to record the node terminal voltage in real time at a frequency of 1 Hz (i.e., once per second). The duration of the sampling process is the same as the corresponding rest time. For example, the lithium iron phosphate node samples for 30 minutes, and a total of 1800 voltage sampling points are obtained. The terminal voltage is the DC voltage value between the positive and negative electrodes of the battery and is a direct indicator for evaluating its rest recovery process. The collected original voltage data is automatically arranged according to the time stamp to form the original voltage time series sampling data and uploaded to the data processing module for standby.

[0020] Step S14: arranging and preliminarily screening the original voltage time series sampling data, removing abnormal data points, and forming voltage rebound time series data; In order to improve the validity of data, the system of the embodiment of the present invention pre-processes the original voltage time series sampling data. First, sort by timestamp to ensure that the data interval is consistent; then use the three-point sliding average method to smooth the curve to eliminate small measurement noise; then perform abnormal point elimination processing, mainly for voltage value mutations caused by sampling errors and BMS instantaneous power failure. The abnormal point identification method is based on the local anomaly factor (LOF) algorithm, which detects the discreteness of data points and previous and subsequent points to identify abnormal values ​​and automatically eliminate them. After the processing is completed, stable and continuous voltage rebound time series data is formed. This data reflects the curve change of the natural recovery of the voltage after the energy storage node stops working, usually showing an upward trend that is fast first and then slow, and is an important representation of health indicators such as battery internal resistance and electrolyte diffusion rate.

[0021] Step S15: performing temperature correction processing on the voltage rebound time series data to obtain standardized voltage rebound characteristic data.

[0022] After obtaining the voltage rebound timing data, the embodiment of the present invention introduces a temperature correction processing module into the system to eliminate the impact of ambient temperature differences. Because the battery voltage rebound speed is extremely sensitive to temperature, especially in a low temperature environment, the slowdown of ion diffusion will cause the rebound curve to lag, affecting the accuracy of the health assessment. The temperature correction method first calls the node temperature sensor to collect the average ambient temperature during the static period, and then compares the temperature with the standard temperature (usually 25°C), and calculates the voltage correction coefficient according to the temperature sensitivity function. The correction coefficient is obtained from a large number of experiments and reflects the average offset of the voltage response under a unit temperature change. The system adjusts the voltage value at each time point based on the coefficient, and finally generates standardized voltage rebound characteristic data equivalent to 25°C. This data is used as one of the input features of the subsequent health assessment model, which can significantly improve the accuracy and stability of the node aging status assessment.

[0023] It is particularly important that step S15 includes the following steps: Step S151: acquiring the surface and ambient temperature data of each energy storage node, and calculating the deviation value of the temperature from the standard test condition based on the surface and ambient temperature data of each energy storage node, thereby obtaining a temperature correction coefficient; Before the standardized voltage rebound characteristic data is generated in the embodiments of the present invention, the system needs to correct the temperature variable that affects the accuracy of the voltage rebound curve. For this purpose, two types of temperature sensors are integrated in the BMS sub-module installed at the energy storage node: one type is attached to the surface of the battery module housing to obtain the surface temperature data of the node in real time; the other type is installed on the inner wall of the battery compartment or the air-conditioning return air outlet to collect the ambient temperature where the node is located. The system periodically reads the data of these sensors and aggregates them by node. For example, during the static period of node A, its surface temperature is 28.5 °C and the ambient temperature is 27.2 °C. The system calculates the comprehensive working temperature of this node as the weighted average of its surface and ambient temperatures. The standard test condition is usually set at 25 °C. The system calculates the temperature deviation value for each node, that is, the current node temperature minus the standard temperature value. In order to quantify the influence of this deviation on the voltage rebound process, the system builds a temperature influence model, which is obtained by training a large amount of experimental data and can output the voltage rebound correction coefficient corresponding to the temperature deviation. For example, for a certain type of lithium iron phosphate battery cell, when the temperature increases by 1 °C, its rebound amplitude increases by an average of 2 mV. Then when the temperature deviation is +3.5 °C, its temperature correction coefficient is +7 mV. The correction coefficients of all nodes are stored as a temperature correction coefficient data set.

[0024] Step S152: Perform a correction calculation on the voltage rebound time series data according to the temperature correction coefficient to eliminate the interference of temperature on the voltage rebound process, and obtain a normalized voltage rebound curve; In the embodiments of the present invention, the original voltage rebound time series data is corrected point by point according to the temperature correction coefficient obtained in the previous step, and the correction process is executed by the data correction engine. The specific method is to add the reverse offset value of the temperature correction coefficient to the original voltage value sampled at each time point, so as to eliminate the influence of temperature on the voltage reading. For example, at a certain moment, the sampled voltage of node A is 3.314 V and the temperature correction coefficient is +7 mV. Then the system corrects the voltage value at this point to 3.307 V. This operation is carried out in the entire full range of the rebound curve to ensure that the voltage recovery process presents under the assumed standard temperature. It should be noted that the correction coefficient curve in different temperature ranges is not linear, and the system uses methods such as piecewise linear fitting or cubic spline interpolation to ensure the continuity and accuracy of the temperature correction coefficient. After the processing is completed, the generated normalized voltage rebound curve is a standard rebound curve that eliminates the temperature interference factor and can truly reflect the electrochemical recovery process of the node, and this curve will be used as the basis for feature extraction.

[0025] Step S153: Extract feature points from the normalized voltage rebound curve to obtain a set of voltage rebound characteristic parameters including the initial voltage value, the stable voltage value, and the rebound rate coefficient; In an embodiment of the present invention, for the normalized voltage rebound curve, the system calls the feature extraction module to extract key rebound feature parameters. First, the system identifies the initial voltage value corresponding to the starting point of the curve, which usually appears at 0 seconds when the static state starts and reflects the voltage level of the battery just after charging and discharging stops. Then, it identifies the stable section at the end of the curve, selects the time period within the last 60 seconds where the voltage change amplitude is less than 0.5 mV, and takes the average value of this section as the stable voltage value. Finally, it calculates the rebound rate coefficient, that is, the slope of the voltage rebound per unit time, usually approximated by the linear fitting slope within the first 300 seconds of the curve. Taking node A as an example, its initial voltage is 3.307 V, the stable voltage is 3.332 V, the voltage has risen by 0.021 V within 300 seconds, and the corresponding rate is 70 μV per second. The above three parameters together constitute the voltage rebound feature parameter set of the current static rebound feature of this node, expressed as a triple of "initial voltage value + stable voltage value + rebound rate", which is stored structurally for standardized processing and health modeling.

[0026] Step S154: Perform data standardization processing on the voltage rebound feature parameter set according to the preset cell characteristic reference library to obtain standardized voltage rebound feature data.

[0027] In an embodiment of the present invention, the system calls the cell characteristic reference library to perform standardization processing on the voltage rebound feature parameter set. The reference library is obtained through experimental tests by the system during the construction period and includes typical rebound feature reference values of different types of cells at each stage of the life cycle under standard working conditions. For example, for the lithium iron phosphate IFR-32700 cell, the initial voltage is recorded as 3.295 V, the stable voltage is 3.330 V, and the rebound rate is about 65 μV per second in the new cell state in the reference library. The system performs normalization processing by comparing the offset between the characteristic parameters of the current node and the reference value, and calculates the standardized voltage rebound feature data. The standardization method adopts the min-max normalization strategy, that is, each parameter is converted into a value between 0 and 1 according to its relative position within the reference range of this type, which is convenient for subsequent horizontal comparison between different cells and aging trend analysis. The finally formed standardized voltage rebound feature data can not only be used for the aging state assessment of a single node, but also support the construction of a whole-station operation health trend prediction model.

[0028] Preferably, step S2 includes the following steps: Step S21: Set the state of charge detection time window parameter, determine the state of charge acquisition frequency suitable for each energy storage node type, and obtain the state of charge monitoring configuration data; In the embodiments of the present invention, in order to accurately monitor the state of charge (SOC) during the charging and discharging processes of battery units of different types of energy storage nodes, the system first needs to set appropriate SOC detection time window parameters for each type of energy storage node. In the specific implementation process, the system configures different SOC sampling frequencies according to the cell chemistry type (such as lithium iron phosphate, ternary lithium), the cell capacity level (for example, 50Ah, 100Ah, etc.), and the location of the node (such as whether there is cooling and ventilation conditions). Taking the lithium iron phosphate 50Ah node as an example, this type of cell has the characteristics of a stable discharge platform and slow temperature rise. The system sets the SOC detection time window for it as "sampling once every 5 seconds, and the continuous sampling duration is 120 seconds". For the ternary lithium 80Ah node that is more prone to temperature mutations, it is set as "sampling once every 2 seconds, and the sampling window length is shortened to 60 seconds". The system tags all energy storage nodes through the node parameter initialization module, classifies and configures the sampling window parameters, and organizes the sampling interval, sampling period, and data format corresponding to each node into "SOC monitoring configuration data", which is stored in the operation monitoring configuration table for dynamic calling in subsequent steps.

[0029] Step S22: Collect the terminal voltage, current, and temperature at specific moments before and after the charging and discharging processes of the node according to the SOC monitoring configuration data to obtain the original data of the node's operating state before and after. In the embodiments of the present invention, according to the aforementioned SOC monitoring configuration data, the data acquisition unit is controlled in real time to trigger the acquisition of state data at specific time points set before and after the charging and discharging events of each node. Specifically, the system determines whether the node enters the charging or discharging stage through the start signal of the PCS (power conversion system). Within 30 seconds before the event and 60 seconds after the event, the battery terminal voltage, current, and the temperature inside the cabin are collected at the configured frequency. For example, for type A nodes, which are set to collect data once every 5 seconds within 30 seconds before the event and 90 seconds after the event, the system records the collected original data as a series of triplet sequences of "timestamp + voltage value + current value + temperature value". These data are uploaded to the data aggregation server in real time by the distributed acquisition controller, and metadata such as the current node ID, operation type (charging or discharging), time tag, and operating stage are marked, and an original dataset of the node's operating state before and after is aggregated as the input source for subsequent processing.

[0030] Step S23: Perform preliminary filtering and calibration on the original data of the node's operating state before and after to exclude the influence of measurement interference and transient fluctuations and obtain effective SOC data. In the embodiments of the present invention, in order to eliminate outliers and short-term fluctuation data caused by electromagnetic interference, communication delay, or sudden load switching, a state cleaning module is used to preliminarily filter and calibrate the original data of the node's operating state before and after. This module calls the combined strategy of "sliding median filtering + local range judgment". In each time window, a five-point sliding window is used to calculate the median to replace the original value, thereby suppressing spike interference. At the same time, a threshold is set. When the current difference between two adjacent sampling points exceeds 10% of the rated capacity and the voltage change exceeds 50 mV, the system determines that this section of data is disturbed and performs deletion processing. Taking node B as an example, at the initial stage of discharge, the current suddenly changes from 8 A to 42 A. The system identifies this as a transient charge-discharge switch and sets a "jump shielding mark" for this section of data, which is excluded from the results. The data after cleaning is converted into a continuous and stable voltage, current, and temperature sequence, and is segmented and classified according to the time period within the node cycle to form a structured "effective state of charge dataset".

[0031] Step S24: Calculate the change in the state of charge of each energy storage node within a complete operating cycle based on the effective state of charge data to obtain a preliminary evaluation value of the charge-discharge depth of the node; In the embodiments of the present invention, based on the aforementioned effective state of charge dataset, an integrated state of charge estimation method is adopted to calculate the change in SOC of each node within a complete operating cycle. This process first extracts the start and end times and the corresponding current data of a complete charge-discharge cycle of the node. By multiplying the current by the sampling period and integrating within the time window, the total charge or discharge amount of the node is approximately calculated. For example, during the discharge stage of node C from 08:00 to 10:15, the average current is 25 A and it lasts for 135 minutes. The system estimates the integrated charge amount of this section as 56.25 Ah, and combines its rated capacity to judge that its discharge depth is approximately 56%. At the same time, the system performs a difference analysis on the starting and ending voltages to verify whether the voltage change is consistent with the trend of the charge change. If they are consistent, it is confirmed that this SOC change is valid. The SOC change amounts of all nodes are uniformly sorted into a preliminary evaluation value of the charge-discharge depth of the node, marked as "node ID + time period + charge change value + discharge ratio".

[0032] Step S25: Correct the preliminary evaluation value of the charge-discharge depth of the node in combination with the actual operating duration of the node and the magnitude of the charge-discharge current to obtain the change value of the charge-discharge depth; In the embodiment of the present invention, considering the influence of current fluctuation and duration on the accuracy of state of charge (SOC) estimation during the actual operation process, a multi-dimensional correction model is introduced to dynamically adjust the preliminary evaluation value of the charge and discharge depth of the node. The correction model performs weighted adjustment on the SOC estimation based on the stability factor and operation time factor of the current load curve. For example, for type D nodes with a relatively short operation time and frequent current peaks, when the system detects that its discharge duration is only 28 minutes and the current fluctuation exceeds ±20%, it is determined as a short-cycle high-pulse load scenario. The correction model corrects the initially estimated discharge depth of 56% to 42%, reflecting the insufficient effective output power. This model configures weight coefficients according to different operation modes and calibrates through historical calibration data, and finally outputs "node ID + corrected charge and discharge depth change value" as the index of the true power change of the node and inputs it into the subsequent health model.

[0033] Step S26: Construct a node health state evaluation model based on the standardized voltage rebound characteristic data and the charge and discharge depth change value to obtain node aging differentiation parameters.

[0034] In the embodiment of the present invention, by integrating the standardized voltage rebound characteristic data and the charge and discharge depth change value, a multi-factor health state evaluation model is constructed to identify the level of node performance degradation. This model uses the random forest regression algorithm, with the triple of "initial rebound voltage, stable voltage, and rebound rate" and the "charge and discharge depth change value" as the input feature vectors, and outputs the health decay factor of a single node, reflecting the performance degradation rate of the node in the current cycle. For example, when the system identifies that the rebound rate of type E nodes in this cycle has decreased by 15% and the corrected value of the discharge depth remains above 80%, the model outputs the health state score of this node as 0.82 (with a full score of 1). To further highlight the aging difference of nodes inside the power station, the system performs standard deviation analysis on the scores of all nodes in the same operation batch, extracts the relative offset between each node and the mean value, and obtains the node aging differentiation parameters, which are used as the basis for load distribution or maintenance priority ranking on the dispatching side. Such parameters will also be regularly uploaded to the master station for station-level health trend modeling.

[0035] Preferably, step S26 includes the following steps: Step S261: Perform data fusion and matching according to the standardized voltage rebound characteristic data and the charge and discharge depth change value to construct a node operation behavior feature matrix; After obtaining the node normalized voltage rebound characteristic data and the charge-discharge depth change value in the embodiments of the present invention, the system uses the data fusion module to perform matching processing on these two types of data to construct a node operation behavior characteristic matrix with a unified structure and clear dimensions. This process includes three steps: First, the system binds the voltage rebound characteristics and the corresponding charge-discharge depths one by one according to the node ID and the time period label; Second, the voltage rebound characteristics are quantified into three indicators, including the rebound starting voltage (reflecting the starting point of the electrochemical reaction recovery), the rebound stable voltage (reflecting the final stable state), and the rebound time length (i.e., the time from power-off to voltage stabilization); Third, the above three voltage rebound parameters and the charge-discharge depth change value within the same period are integrated into a set of feature vectors to form a six-dimensional behavior characteristic record with the content of "node ID + time period + rebound starting voltage + rebound stable voltage + rebound time length + charge-discharge depth". The system stacks all records in the order of nodes and time to form a node operation behavior characteristic matrix with the dimension of "number of nodes × number of feature items × length of time series", providing a data basis for the next correlation analysis.

[0036] Step S262: Analyze the correlation between the voltage rebound characteristics and the charge-discharge depth based on the node operation behavior characteristic matrix, so as to establish a preliminary node health assessment model; After constructing the node operation behavior characteristic matrix in the embodiments of the present invention, the system enters the feature correlation analysis stage to explore the internal relationship between the voltage rebound behavior and the node charge-discharge depth. The system adopts a combination of Pearson correlation coefficient analysis and principal component dimensionality reduction method to extract the correlation intensity between variables as a whole. Specifically, the system traverses the behavior characteristic matrix node by node, and calculates the correlation coefficients of "rebound time length and charge-discharge depth", "rebound starting voltage and discharge depth", etc. on each node respectively, so as to judge whether there is an association trend between different parameters. For example, for node F, the system finds that there is an obvious positive correlation between its rebound time length and discharge depth (the correlation coefficient exceeds 0.85), indicating that the chemical reaction recovery process after deep discharge is more sluggish. Based on the above analysis, the system establishes a preliminary health assessment model, and uses a linear weighted model with "rebound time", "starting voltage offset" and "depth value" as independent variables to output the health score of the node in the current period. This score preliminarily represents the degree of deterioration of the electrochemical performance and provides a basis for subsequent model optimization.

[0037] Step S263: Calibrate the parameter weights in the preliminary node health assessment model through pre-obtained historical operation data, optimize the model response characteristics, so as to obtain a node health status assessment model; In the embodiments of the present invention, to improve the adaptability of the preliminary health assessment model under different nodes and different operating conditions, the system introduces historical operation data for parameter weight calibration to optimize the model response characteristics. Specifically, during implementation, the system extracts the complete charge and discharge behavior records and manual maintenance results (such as cell replacement records, voltage anomaly alarm data) of all nodes within the past 180 days from the station-level database, and uses them as the training set to input into the preliminary model, and adopts a genetic algorithm to optimize the weight of each feature item. For example, the system assigns an initial weight of 0.3 to the bounce time feature and 0.5 to the voltage starting offset feature, and continuously adjusts the weights through successive rounds of genetic evolution to minimize the error between the predicted health score and the actual failure probability on the validation set. The optimized model weight vector will be saved in the "Node Health Status Assessment Model" configuration table as the core call module for real-time health calculation. At the same time, the system also uses five-fold cross-validation to evaluate the generalization ability of the model to ensure its stable operation under different future cycles and different node conditions.

[0038] Step S264: Use the node health status assessment model to quantitatively analyze the current status of each node, output comparable health status difference indicators, so as to obtain node aging differential parameters.

[0039] After the embodiments of the present invention complete the optimization of the health status assessment model, it is officially applied to the real-time health status calculation of each operating node of the energy storage power station. During the specific process, the system regularly retrieves the behavior feature matrix of the latest cycle, and inputs the feature vector of each node into the optimized health status assessment model to output a health score value between 0 and 1. The closer the score is to 1, the better the node status. For example, the health score of node G in this cycle is 0.92, while the score of node H is only 0.63. The system automatically calculates the mean and standard deviation of the scores of the same batch of nodes and outputs a standardized difference indicator based on this, that is, the "node health score deviation value". All deviation values are uniformly sorted into "node aging differential parameters". This parameter is not only used to analyze the aging rate differences between nodes, but also feedback to the power station management system to prioritize the arrangement of aging nodes for maintenance or rotation. A typical scenario is to preferentially call nodes with high health scores during peak-valley electricity price scheduling to extend the cell life and improve the overall utilization efficiency of energy storage. Thus, this embodiment completes the comparable quantitative assessment of the node health status and ensures the dynamic closed-loop characteristic of the monitoring system.

[0040] Preferably, determining the available capacity level of each node according to the node aging differential parameters in step S3 includes: Perform standardization processing on the node aging differential parameters based on the 0-1 normalization method to eliminate the influence of different dimensions and obtain standardized aging parameters; Set the aging parameter threshold interval [0.15, 0.35, 0.55, 0.75] as the grading boundary point based on the standardized aging parameters, and perform segmented mapping to obtain the preliminary aging level classification result; Calculate the actual capacity loss rate based on the loss rate calculation formula according to the preliminary aging level classification result combined with the node rated capacity values of each energy storage node obtained in advance, where the loss rate calculation formula is loss rate = basic loss rate + aging level coefficient × attenuation acceleration factor; Dynamically correct the actual capacity loss rate. For nodes with an operating time exceeding 1000 hours, add a correction value of 0.8% to the original loss rate for every additional 500 hours to obtain the corrected capacity loss rate; Calculate the percentage of the actual available capacity of the node according to the corrected capacity loss rate, where the percentage of the actual available capacity of the node = 100% - corrected capacity loss rate; Set the capacity level division standard for the percentage of the actual available capacity of the node to form the node capacity grading standard; Rate and divide the percentage of the actual available capacity of each node based on the node capacity grading standard to obtain the node available capacity level data; Extract the nodes with adjacent level boundary values from the node available capacity level data, and perform secondary correction in combination with the stable voltage recovery rate in the voltage rebound characteristic data. When the recovery rate is lower than the threshold, lower the node level by one level to obtain the available capacity level.

[0041] After obtaining the "node aging differentiation parameter" of each energy storage node in the embodiments of the present invention, since this parameter is calculated based on the deviation value of the health score and has the problem of inconsistent numerical distribution among different nodes, the system first performs 0-1 normalization processing on this parameter to eliminate the influence of dimensions and make it convenient for subsequent hierarchical comparison. Specifically, when implementing, the system extracts the maximum aging difference value and the minimum aging difference value from all nodes, and adjusts the aging difference value of each node to a proportional numerical interval between 0 and 1 according to the normalization formula. For example, if the node difference value interval is from 0.12 to 0.85, the system maps the original difference value of 0.42 of node K to a normalized value of 0.459, realizing the comparison basis on a unified scale among different nodes. This result is defined as the "standardized aging parameter" and stored in the node evaluation feature table. After completing the calculation of the standardized aging parameter, the system sets the parameter threshold interval as [0.15, 0.35, 0.55, 0.75] according to the empirical aging level division as the five-level demarcation points, and performs piecewise mapping on the standardized aging parameters of all nodes to form the preliminary aging level classification result. The system defines the level division logic as follows: those with a standardized value less than or equal to 0.15 are classified as "Level 1 (new state)"; those greater than 0.15 and less than or equal to 0.35 are classified as "Level 2 (mild aging)"; those greater than 0.35 and less than or equal to 0.55 are classified as "Level 3 (moderate aging)"; those greater than 0.55 and less than or equal to 0.75 are "Level 4 (severe aging)"; those greater than 0.75 are "Level 5 (serious aging)". For example, the standardized aging value of node M is 0.62, and the system classifies it into Level 4 and writes this preliminary level into the initial evaluation table of the node aging level, forming the mapping relationship from the health state to the aging level. Based on the preliminary aging level result, the system further calculates the capacity loss situation of each energy storage node in its current state in combination with the "rated capacity value" of the energy storage node. For this purpose, the system introduces a calculation formula for the "actual capacity loss rate" based on the aging level, that is, assigns an "aging level coefficient" to each level, and introduces a "basic loss rate" of 5% and a "decay acceleration factor" of 3% to reflect the impact of aging on capacity decline. For example, the aging level coefficient of Level 2 is 0.6, and the corresponding capacity loss rate is 5% plus 0.6 multiplied by 3%, that is, the loss rate is 6.8%. Node N is at Level 3, and its corresponding coefficient is 1.1, then the loss rate is calculated as 8.3%. This calculation realizes the quantitative mapping between capacity decay and aging level, and preliminarily depicts the capacity performance degradation of each node caused by aging. Considering that some nodes have been running for a long time, their actual capacity loss is not only related to the aging level but also significantly affected by the cumulative running time, so the system introduces a dynamic correction mechanism to further adjust the "actual capacity loss rate". The specific rule is: for any node whose cumulative running time exceeds 1000 hours, for every additional 500 hours, an additional 0.8% is added to the original loss rate as the correction value.For example, if the current level of node P is 4, the corresponding initial loss rate is 10.25%, and its cumulative operating time is 2,100 hours, then two additional rounds of correction are required (i.e., an increase of 2 × 0.8%), and the loss rate after correction is 11.85%. This effectively improves the accuracy of the evaluation model's response to the capacity loss of nodes operating at high load for a long time, and avoids the deviation of ignoring the operating intensity due to a single evaluation of the aging level. After obtaining the corrected capacity loss rate of each node, the system further calculates the "actual available capacity percentage" of the node, which represents the remaining available capacity of the node compared to the rated capacity in the current state. The system processes it in the form of a percentage system, that is, subtracting the corrected capacity loss rate from 100% to obtain the final available capacity percentage. For example, if the corrected loss rate of node P is 11.85%, then its current available capacity is 88.15%. This indicator becomes the basic data for subsequent capacity level division, reflects the actual performance state of the schedulable capacity of the node, and is one of the important bases for formulating the energy storage resource scheduling strategy. To facilitate unified scheduling management and hierarchical maintenance, a node capacity grading standard is formulated, which is divided into five grade segments by percentage: those greater than or equal to 95% are "capacity grade A", those between 85% and 95% are "grade B", those between 75% and 85% are "grade C", those between 65% and 75% are "grade D", and those below 65% are "grade E". The system uses this standard as the basis for node capacity performance grading, and the "actual available capacity percentage" of all nodes will be archived under this standard. For example, if the calculation result of node Q is 93.6%, it is classified as capacity grade B; if the result of node R is 68.2%, it is classified as capacity grade D and marked as a capacity warning state. The system writes this grade into the capacity grade table and associates it with the aging grade table for unified management and analysis. To further improve the rating accuracy, after the preliminary capacity level division of all nodes is completed, a grade correction operation for boundary nodes is carried out. The specific method is as follows: The system screens out nodes within the range of ±1% of the boundary values (i.e., edge values such as 85%, 75%, 65%, etc.) from all nodes as boundary node samples, and then extracts the "stable voltage recovery rate" parameter in the "voltage rebound characteristic data" corresponding to its cycle. This parameter refers to the ratio of the final stable value of the rebound voltage to the discharge termination voltage, reflecting the voltage recovery ability of the node. The system sets the threshold to 92%, that is, a recovery rate lower than this value indicates a lag in the electrochemical reaction. If the recovery rate of the boundary node is lower than 92%, the system will downgrade the capacity level of this node by one level. For example, if node S is initially rated as grade B, but its recovery rate is 88.7%, the system will downgrade it to grade C, indicating that the voltage performance does not support the original grade capacity index. This mechanism can effectively avoid the scheduling risk caused by inaccurate judgment of a single capacity index and form a more reliable output result of the available capacity level.

[0042] Preferably, step S3 includes the following steps: Establish the mapping relationship between the capacity level and the state of charge (SOC) working range according to the available capacity level to obtain the preliminary SOC working range data; Collect the voltage-SOC curves of each node under standard charge and discharge conditions, select the area where the voltage change rate is less than 0.05V / 1%SOC as the voltage plateau area, and obtain the node voltage plateau interval data; Based on the node voltage plateau interval data and the preliminary SOC working range data, optimize and adjust the working range to ensure that the working range includes at least 75% of the voltage plateau area, and obtain the adjusted working interval data; Conduct constant power charge and discharge tests on each node, charge and discharge at a rate of 0.5C, and record the terminal voltage change rate at different states of charge to obtain the voltage response characteristic data; Calculate the internal resistance change trend of each node at different states of charge according to the voltage response characteristic data, and take the point where the internal resistance increases by 20% as the boundary to further optimize the working interval boundary, and obtain the internal resistance optimized working interval data; Calculate the intersection area of the adjusted working interval data and the internal resistance optimized working interval data as the safe working interval of the node, and divide it according to the available interval width: greater than 70% width is level one, 50%-70% is level two, 30%-50% is level three, less than 30% is level four, and obtain the node working interval level data; Based on the node working interval level data and the available capacity level data, establish a two-dimensional evaluation matrix to obtain the preliminary node operation ability level; Conduct a power response ability test on the preliminary node operation ability level to obtain the power response ability rating. Specifically, the power response ability test is to apply a pulse power of 0.8C within the safe working interval for 30 seconds and record the voltage drop amplitude. If the voltage drop is less than 0.15V, it is excellent; if it is 0.15V - 0.25V, it is good; if it is 0.25V - 0.35V, it is medium; if it is greater than 0.35V, it is poor; Determine the node operation ability classification according to the preliminary node operation ability level and the power response ability rating, and form the node operation ability classification data including three-dimensional characteristics of capacity state, working interval, and power response.

[0043] After the data of the available capacity level of the nodes is determined in the embodiments of the present invention, in order to achieve more accurate control of the state of charge (SOC) of the nodes, in this embodiment, a mapping relationship between the capacity level and the working range of the state of charge is established according to the available capacity level. In the specific operation process, the system presets the corresponding working ranges of the state of charge according to the obtained four-level node available capacity level data, which are the classification results of level 1 (available capacity greater than 85%), level 2 (70%-85%), level 3 (50%-70%), and level 4 (less than 50%). For example, for the nodes with the capacity level of level 1, the state of charge range is set to 25% to 90%, for level 2 it is set to 30% to 85%, for level 3 it is set to 35% to 80%, and for level 4 it is set to 40% to 75%. This mapping relationship is constructed through historical operation data and operation and maintenance experience, aiming to enable nodes with different capacity levels to operate efficiently within their remaining capacity capabilities, and at the same time avoid frequent charge and discharge or deep cycling due to insufficient capacity. Finally, the initial state of charge working range data of each node is formed, providing a reference for the subsequent adjustment of the voltage platform area. To further finely control the state of charge interval of each node, this embodiment collects the voltage and state of charge (SOC) curve data of each node under standard charge and discharge conditions, where the standard charge and discharge conditions refer to the test carried out under the experimental conditions of an ambient temperature of 25 degrees Celsius, a charging and discharging rate of 0.5C, and a voltage sampling frequency not lower than 1Hz. In the curve data, by statistically analyzing the voltage change amplitude corresponding to each 1% change point of the SOC, a continuous region where the voltage change rate is less than 0.05 volts per 1% SOC is selected, and this region is defined as the voltage platform area. The voltage platform area represents that the voltage changes stably within a specific state of charge section of the node, and this section is the most sensitive to the battery performance and the energy output is relatively stable. The extracted voltage platform interval data is used as an important basis for optimizing the state of charge working range and will be used later to adjust the initial state of charge range to ensure operation stability. After obtaining the initial state of charge working range data and the voltage platform interval data of each node, this embodiment optimizes and adjusts the working range. The adjustment strategy is to ensure that the final state of charge working range of each node needs to cover more than 75% of its voltage platform section, that is, to extend and cover outward with the voltage platform interval as the core area. When the initially set SOC working range cannot meet this coverage requirement, the upper and lower limits are reset according to the start and end positions of the platform interval. For example, if the voltage platform interval of a certain node is from SOC40% to SOC80%, and the initially set range is 35%-75%, then the upper limit of the range needs to be adjusted to 80% to meet the coverage ratio requirement. The adjusted working interval data will be used as the basis for subsequent performance tests to ensure that the test results are carried out within the stable working voltage area and improve the evaluation accuracy. After clarifying the adjusted state of charge working interval of each node, the embodiment conducts constant power charge and discharge tests on each node within this interval to obtain voltage response characteristic data.The specific test method is to charge from the lower limit to the upper limit of the SOC range at a 0.5C charging rate and discharge back to the lower limit at a 0.5C discharging rate. During this period, the voltage change curve is recorded with a 1-second sampling period. At different SOC points, the voltage change amount per unit time is extracted and divided by the SOC change rate ratio to calculate the voltage change rate. The voltage response characteristic data can reflect the voltage sensitivity of the node under different state of charge conditions and is used to judge the battery performance degradation characteristics, which is an important reference basis for subsequent internal resistance analysis and interval boundary adjustment. After obtaining the voltage response characteristic data, in this embodiment, the change trend of the battery internal resistance of the node under different SOCs is calculated and analyzed. By linking the voltage response characteristics with the current data under constant power charge and discharge conditions, the dynamic internal resistance values of the battery at different SOC points are calculated, and the internal resistance at the starting SOC point is used as the reference benchmark point. When the internal resistance value at a certain SOC point increases by more than 20% compared with the reference point, it is marked as the safety boundary point of the working interval. For example, if the internal resistance at the SOC = 75% point increases by more than 20% compared with the SOC = 35% point, then 75% is set as the upper limit boundary. By this method, some sections with too rapid internal resistance increase are further compressed, and the operating ranges with potential safety hazards are eliminated, and finally the working interval data with optimized internal resistance is obtained. Combining the adjusted working interval data and the internal resistance optimized working interval data described above, in this embodiment, the intersection area of the two is calculated and used as the safe working interval of the node. The intersection processing adopts the double-interval union method, that is, the overlapping part of the two intervals is taken as the effective interval to ensure operation under the dual restrictions of the safe voltage range and the stable internal resistance range. Subsequently, grading is carried out based on the width of this intersection interval (that is, the interval length divided by the total SOC range). A width greater than 70% is set as the first level, indicating a wide working interval and high flexibility; 50% - 70% is the second level, indicating that most areas are safe and controllable; 30% - 50% is the third level, with restricted operation; less than 30% is the fourth level, which requires key monitoring. The finally formed node working interval grade data will be used as an important indicator for operation ability evaluation. After obtaining the node working interval grade data and available capacity grade data, in this embodiment, a two-dimensional evaluation matrix model is established. The horizontal axis is the node working interval grade, and the vertical axis is the capacity grade. Each node can locate a grade coordinate point in this matrix, and a comprehensive score is obtained by setting a scoring rule. For example, the combination of the first-level capacity and the first-level working interval gets a full score of 10 points, and the combination of the fourth-level capacity and the fourth-level interval gets the lowest score of 2 points. The remaining combinations are respectively set with scoring rules in a weighted manner. For example, the combination with the second-level capacity and the third-level working interval gets a score of 6 points. This matrix model quantifies the comprehensive operation potential of the node into the preliminary operation ability grade, laying a data structure foundation for the subsequent response ability test. To further improve the node operation ability evaluation, in this embodiment, the power response ability test is carried out on each node. The test method is to apply a pulsed power at a 0.8C rate (the duration is 30 seconds) within the safe working interval of each node and record the voltage drop amplitude at the moment when the load is applied.If the voltage drop is less than 0.15 volts, it is judged as "excellent"; between 0.15 volts and 0.25 volts is "good"; between 0.25 volts and 0.35 volts is "medium"; greater than 0.35 volts is "poor". During the test process, the voltage change is monitored in real time through a high-frequency sampling device, with no less than 10 samples per second to ensure data accuracy. The test results are used to judge the dynamic response ability of the node in a high-power scenario, providing reliable data support for the correction of the operation ability level. After obtaining the preliminary operation ability level and power response ability rating of the node, in this embodiment, the final node operation ability classification is determined by integrating the information of the two. Among them, the operation ability classification adopts a three-dimensional scoring method, assigning certain weights to the capacity level, working range level, and power response rating respectively (such as 40% for capacity, 30% for working range, and 30% for response ability), and comprehensively scoring in combination with the actual working condition weight adjustment strategy. Finally, the nodes are divided into four levels: operation ability level one (high reliability and high responsiveness), level two (medium-high performance), level three (general performance), and level four (low performance requiring early warning), and structured node operation ability classification data is formed. This data will serve as the basic basis for the operation scheduling, maintenance plan, and abnormal prediction of the energy storage power station, improving the overall operation efficiency and safety of the energy storage system.

[0044] Preferably, establishing a node charge-discharge power allocation model using the node operation ability classification data in step S3 includes: Based on the node operation ability classification data, allocate basic power weight coefficients to each node to obtain a node basic power weight coefficient table; Collect the charge-discharge efficiency data of each node under different state of charge, conduct sampling tests at intervals of 10% state of charge, record the energy conversion efficiency at each sampling point, and obtain the node charge-discharge efficiency curve; Based on the node charge-discharge efficiency curve, calculate the optimal charge-discharge efficiency interval of each node, define the area with efficiency higher than 90% as the optimal working point, and obtain the node efficiency optimization interval data; Combine the node efficiency optimization interval data with the node safe working interval and calculate the intersection area of the two as the node priority scheduling interval to obtain the node priority scheduling interval data; Design the power allocation coefficient within the interval according to the node priority scheduling interval data; Monitor the current state of charge of the node in real time, calculate the distance ratio between the current state of charge and the boundary of the priority scheduling interval. When the distance ratio is less than 0.2, linearly attenuate the basic power weight coefficient to obtain the boundary adjustment coefficient; Monitor the temperature change state of the node, set the normal working temperature range as 15°C - 35°C. When the temperature exceeds this range, for every increase or decrease of 5°C, attenuate the basic power weight coefficient by 5% to obtain the temperature correction coefficient; A node charge and discharge power allocation model is established based on the power allocation coefficient within the interval, the boundary adjustment coefficient, and the temperature correction coefficient.

[0045] In the embodiment of the present invention in the operation monitoring and management method of the energy storage power station, based on the node operation ability grading data obtained above, a corresponding basic power weight coefficient is assigned to each energy storage node. In specific implementation, the above grading results are used to set the basic power weights according to a four-level distribution: the basic power weight coefficient assigned to the first-level nodes is 1.0, representing that the nodes have the optimal operation ability; the second-level nodes are 0.8, the third-level nodes are 0.6, and the fourth-level nodes are 0.4. This weight coefficient serves as the initial basis for dispatching and allocating the power of the power station, and is recorded to form a node basic power weight coefficient table for subsequent deployment calculations. The structure of this table includes node numbers, operation ability levels, and corresponding basic power weights, and is applicable to dynamically adjusting the node dispatching strategy. After establishing the basic power weights, to further improve the dispatching accuracy, the charge and discharge efficiency data of each node under different state of charge (SOC) are collected. With a sampling interval of every 10% SOC, 11 sampling points are set within the range of 0% - 100%, and standard charge and standard discharge operations are respectively performed. During the test process, the input and output electric energies are recorded, and the efficiency is calculated, that is, the percentage obtained by dividing the output electric energy by the input electric energy, to form the charge and discharge efficiency curve of each node. This test uses a 1C rate for charging and discharging, and ensures a constant temperature of 25°C to exclude the influence of temperature. The obtained data serves as the basic information for measuring the energy conversion performance of the node under different SOCs. Based on the above efficiency curve, the efficiency levels of each node in different SOC intervals are statistically analyzed, and the SOC range with an efficiency value higher than 90% is selected and defined as the "optimal charge and discharge efficiency interval", that is, the energy conversion of the node is the most sufficient and the loss is the smallest within this interval. In implementation, by traversing the efficiency curve, the SOC section with continuous efficiency higher than 90% is extracted and marked as the efficiency optimization interval data of the node. This process can be realized through tools such as Python to perform data fitting and screening, forming the optimal efficiency operation interval of each node between 0% and 100% SOC. To achieve more efficient power station operation dispatching, the intersection calculation is performed on the node efficiency optimization interval data and the node safe working interval data obtained previously to obtain the "node priority dispatching interval". This intersection area has both safety (such as stable voltage platform and reasonable internal resistance control) and high energy efficiency characteristics, and is suitable as the energy regulation window preferentially called during the operation of the power station. The intersection calculation can be completed through a data interval comparison algorithm. The priority dispatching interval is usually represented by SOC values. For example, the priority dispatching interval of node A is from 30% to 80%. After obtaining the priority dispatching interval, it is necessary to design the power deployment coefficient within the interval to refine the dispatching strategy. In implementation, the priority dispatching interval is further divided into three sub-sections: the front section (such as the first 20%), the core section (the middle 60%), and the tail section (the last 20%), and deployment coefficients of 0.8, 1.0, and 0.9 are respectively assigned to guide the dispatching system to preferentially dispatch this node to achieve high power output when the SOC is in the core section, and the front or tail section is slightly more conservative in dispatching.This coefficient is used to dynamically adjust the basic power weight to adapt to the scheduling response ability of the node within the preferred range. Subsequently, the system needs to monitor the current state of charge (SOC) of each node in real time and calculate the distance ratio between it and the boundary of the preferred scheduling range. The specific method is as follows: the difference between the current SOC value and the nearest boundary of the preferred scheduling range is divided by the width of the preferred scheduling range. For example, if the preferred range of a certain node is 30% - 80% and the current SOC is 32%, then the distance ratio is 2% / 50% = 0.04. If the distance ratio is less than 0.2, it indicates that the current state is close to the edge of the preferred area, and risk suppression processing should be carried out. Therefore, in this case, the basic power weight coefficient is linearly attenuated, with a maximum attenuation to 0.5 times. For example, if the initial coefficient is 1.0 and the distance ratio is 0.1, then the final boundary adjustment coefficient is 0.75. Further considering the influence of temperature factors on the performance of battery nodes, the operating temperature of each node is monitored in real time, and the normal operating range is set to 15°C to 35°C. When the temperature exceeds this range, for every 5°C higher or lower, a 5% attenuation is applied to the basic power weight coefficient. For example, if the current temperature of a node is 40°C, exceeding the upper limit by 5°C, then its temperature correction coefficient is 0.95. This correction is used to simulate the influence of high or low temperature on the battery reaction efficiency and life, and is a key safety parameter adjustment item in the operation monitoring of energy storage power stations. Finally, based on the above three weight factors: the power allocation coefficient within the range, the boundary adjustment coefficient, and the temperature correction coefficient, the product operation is performed on the current basic power weight coefficient of each node to form a real-time charging and discharging power allocation model for the node. In implementation, the system accesses the real-time operation data of all nodes through the scheduling control platform, automatically calls the above model to dynamically update the power allocation instructions of each node, and realizes the optimal configuration of the overall power output of the power station and the dynamic load response regulation. This model supports minute-level or finer-grained scheduling decisions and has good response effects and economy in scenarios such as peak-valley electricity price strategies and power grid load regulation.

[0046] Preferably, step S4 includes the following steps: Step S41: Obtain the overall power scheduling instruction of the energy storage power station, parse the instruction and extract the charge and discharge power demand value, the duration parameter, and the response time requirement to obtain the power scheduling parameter set; In the embodiment of the present invention, aiming at the real-time operation requirements of the energy storage power station, step S41 is first executed to obtain the overall power dispatch instruction issued by the superior dispatch platform or the grid operator. This instruction is usually accessed through the energy management system (EMS) or the dispatch communication interface, and includes the target power value of charge and discharge (for example, the dispatching requires the energy storage power station to output 5 MW of electricity), the duration parameter (such as 10 minutes), and the response time requirement (such as the response time is less than 30 seconds). The system analyzes the dispatch instruction to extract the key control parameters to form a power dispatch parameter set, which is used as the calculation basis for subsequent node power allocation. In practical applications, for example, during the grid frequency regulation or peak shaving process, the power station can receive multiple dispatch instructions with different strategies within 1 minute, and the system needs to quickly analyze and classify them into the database.

[0047] Step S42: Calculate the preliminary power distribution coefficients of each energy storage node according to the power dispatch parameter set in combination with the node charge and discharge power allocation model, and obtain a preliminary power distribution plan; In the embodiment of the present invention, the charge and discharge power allocation model of each previously established node is retrieved, and combined with the target power value and response time in the dispatch instruction, the preliminary power distribution coefficients of each energy storage node are calculated. The specific method is: the target power is distributed according to the comprehensive weight ratio composed of the node basic power weight multiplied by its interval power allocation coefficient, boundary adjustment coefficient and temperature correction coefficient. For example, the target output power is 5 MW, and the comprehensive allocation weight of node A is 0.25, then its preliminary allocated power is 1.25 MW, and the allocation results of all nodes form a preliminary power distribution plan. This plan is the initial estimation result based on the current state of the node and does not consider the maximum / minimum carrying capacity of the node.

[0048] Step S43: Calculate the upper and lower limits of the bearable power of each node based on the preliminary power distribution plan and the node operation ability classification data of each energy storage node, and obtain the node power constraint conditions; In the embodiment of the present invention, to ensure that the power distribution conforms to the physical limitations of the node, it is necessary to further calculate the maximum and minimum bearable powers in its current state based on the operation ability level data and rated capacity information of each energy storage node. For example, a certain node has a secondary operation ability, its rated power is 2 MW, and the power safety factor under the current SOC and temperature conditions is 0.85, then its maximum bearable power is 1.7 MW, and the minimum is -1.7 MW (charging). The system traverses all nodes to form a complete node power constraint condition table for subsequent correction processing to avoid problems such as overload, abnormal heating or power feedback during the power distribution process.

[0049] Step S44: Correct the preliminary power distribution plan according to the node power constraint conditions to ensure that the power values allocated to each node do not exceed its safe operating range, and obtain a safe corrected power distribution plan; According to the power constraint condition obtained in step S43, the present invention embodiment corrects the allocation values exceeding the upper and lower limits of a single node in the preliminary power allocation scheme. The specific method is as follows: For example, if the preliminary allocated power of node A is 1.9 MW, while its maximum load capacity is 1.7 MW, then its power is adjusted down to 1.7 MW, and the extra 0.2 MW is reallocated proportionally to other nodes with remaining load capacity. This process can be realized by a recursive algorithm to achieve dynamic iterative adjustment, and finally form a safe corrected power allocation scheme in which the power values of all nodes are within the safe range. This scheme is the basis for subsequent refined optimization.

[0050] Step S45: Calculate the interval deviation adjustment coefficient based on the deviation degree between the current state of charge (SOC) of each node and its optimal working interval, and optimize the safe corrected power allocation scheme to obtain a balanced optimized power allocation scheme; The present invention embodiment further calculates the deviation between the current state of charge (SOC) of each node and the central position of its aforementioned preferred scheduling interval. If the SOC of a node deviates greatly from the central region of the preferred area, its allocated power needs to be adjusted to guide it back to the central area. Here, the "interval deviation adjustment coefficient" is defined, which is calculated by dividing the SOC deviation value by the interval width. The greater the deviation, the lower the adjustment coefficient. For example, if the preferred area of node B is 30% - 80% and the current SOC is 85%, then its deviation is 5% and the adjustment coefficient is 0.9. This coefficient is used to increase or decrease the power allocation value of the corresponding node in the safe correction scheme to promote the balance of the node state of charge towards the optimal efficiency interval, and form a balanced optimized power allocation scheme.

[0051] Step S46: Monitor the real-time temperature change trend of each node. When the temperature change rate exceeds a preset threshold, calculate the temperature risk coefficient and dynamically adjust the balanced optimized power allocation scheme to obtain a temperature adaptive power allocation scheme; The present invention embodiment continuously monitors the operating temperature change trend of each node. If it is detected that the rate of temperature rise or fall exceeds a preset threshold (such as 1 °C / min), it is regarded as a temperature fluctuation risk, and the temperature risk coefficient is calculated. This coefficient is the reciprocal of the ratio of the temperature change rate to the threshold multiplied by a preset proportional decay factor. For example, if the temperature rise is 2 °C / min, then the temperature risk coefficient is 0.5. For nodes in a temperature risk state, their power allocation values need to be dynamically attenuated according to the risk coefficient to prevent cell swelling, life attenuation or thermal runaway caused by temperature changes. The finally formed temperature adaptive power allocation scheme has a certain degree of real-time performance and risk tolerance, and can realize the scheduling adaptation to the dynamic safety state of the nodes.

[0052] Step S47: Form a node-differentiated power allocation scheme based on the temperature adaptive power allocation scheme, including the power allocation values and execution order of each node; In the embodiment of the present invention, the temperature adaptive scheme is further refined into a node-differentiated power allocation scheme. In addition to including the specific allocated power values of each node, this scheme also stipulates the execution order and scheduling priority of the nodes. For example, under the condition of meeting the same allocation weight and status, nodes with the current SOC in the optimal middle section range, temperature in the normal temperature area, and high operating capacity are preferentially scheduled; while state boundary nodes or temperature rise nodes are executed later. This differentiated scheme is sent to the node controller through a sorted list or a scheduling task queue to guide each node to complete the scheduling task in sequence, which helps to form an orderly and reliable power response link.

[0053] Step S48: Dynamically optimize the node-differentiated power allocation scheme by using the minimum power perturbation method to obtain an adaptive energy flow allocation instruction.

[0054] The embodiment of the present invention introduces the "minimum power perturbation method" to dynamically optimize the differentiated power allocation scheme. This method means that on the premise of ensuring that the total target power remains unchanged, the overall load perturbation is minimized by fine-tuning the power values of individual nodes. The specific approach is that the system takes the current scheme as a benchmark, adjusts the power values of nodes one by one (such as ±0.05 MW), and then re-evaluates the power stability, total energy efficiency, and temperature risk indicators. The allocation result with the minimum perturbation is selected as the final execution instruction. This optimization process can be executed by using local search or gradient descent method, and finally an adaptive energy flow allocation instruction is generated and sent to the execution system of the energy storage power station to achieve real-time, efficient, and safe energy scheduling.

[0055] Particularly importantly, step S48 includes the following steps: Step S481: Set the minimum step size parameter of power perturbation, determine the minimum change amount of the power adjustment of each node, and obtain the perturbation step size configuration data; In the embodiment of the present invention, according to the adjustment accuracy and response ability of the power controllers of each node in the energy storage power station, the minimum step size parameter of power perturbation is set. This parameter is the minimum change amount when adjusting the power of the node each time, usually set to 1% to 3% of the rated power. For example, if the rated power of the node is 2 MW, the minimum perturbation step size is between 0.02 MW and 0.06 MW. By investigating the minimum adjustment unit and communication delay of the node equipment, it is determined that this perturbation step size ensures that the adjustment operation will neither be too frequent to cause an excessive burden on the control system nor be too large to affect the fine-grained control of power allocation, and finally a complete perturbation step size configuration data is formed for subsequent fine-tuning.

[0056] Step S482: Make a small adjustment to the node power allocation value based on the perturbation step size configuration data, generate multiple groups of perturbation schemes, and perform a comparison of the response effects before and after the perturbation to obtain the perturbation response effect data; In the embodiments of the present invention, based on the perturbation step configuration data, a slight adjustment is made to the previously determined node power allocation values. The power values of each node are respectively added and subtracted by the perturbation step to generate multiple sets of perturbation schemes. These schemes are executed through a simulation platform or real-time online testing in an actual energy storage control system, comparing key performance indicators such as the power output stability of the power station, energy efficiency indicators, temperature changes, and response time before and after the perturbation, and collecting and forming perturbation response effect data. In a specific test, for example, after adjusting node A by 0.04 MW, it is monitored whether the total power fluctuation is within the allowable range (±0.1 MW) and whether the system response is lower than the preset 100 ms, and these data are recorded for subsequent analysis.

[0057] Step S483: Calculate the power adjustment sensitivity coefficient based on the perturbation response effect data, evaluate the sensitivity of each node to power perturbation, and obtain a node power sensitivity table; In the embodiments of the present invention, using the perturbation response effect data, the power adjustment sensitivity coefficient is calculated for each node. This coefficient reflects the sensitivity of the node to power perturbation, that is, the ratio of the overall system response change after power adjustment. For example, if a fine adjustment of 0.05 MW of node B's power results in a system output fluctuation of 0.02 MW, the sensitivity coefficient is relatively low; otherwise, it is relatively high. By comparing the sensitivities of each node, the system forms a node power sensitivity table, which reveals which node power adjustments have a greater impact on the overall system and which nodes respond slowly or ineffectively, facilitating targeted optimization of the scheduling strategy.

[0058] Step S484: Based on the node power sensitivity table, identify the nodes with power response hysteresis in the system, and adjust the power allocation priorities of the nodes to obtain a priority-adjusted power allocation scheme; In the embodiments of the present invention, based on the node power sensitivity table, the system identifies the nodes with power response hysteresis, that is, the nodes that respond slowly or have little impact on power perturbation. These nodes are caused by hardware aging, controller performance differences, or communication delays. For these nodes, the system adjusts their power allocation priorities, preferentially reducing the scheduling weights of the hysteresis nodes and allocating more scheduling tasks to the nodes with high response sensitivity, forming a priority-adjusted power allocation scheme. For example, the power weight of the node with the lowest sensitivity is reduced by 20%, and the weights of the highly sensitive nodes are increased by the corresponding proportion to improve the overall scheduling efficiency and system response speed.

[0059] Step S485: Perform iterative optimization on the priority-adjusted power allocation scheme, and find the optimal point of the overall system power response through continuous multiple small perturbations to obtain an iteratively optimized power allocation scheme; The embodiments of the present invention perform iterative optimization based on a priority adjustment scheme. This process gradually approaches the optimal point of the overall power response of the system by continuously making small perturbations to the node power multiple times (such as adjusting the perturbation step up or down by 0.5 times to 1 time), and combining simulation to calculate the power response index of the system. After each iteration, the system compares the power fluctuation range, energy efficiency loss, and temperature safety index, and selects the optimal scheme as the benchmark for the next iteration until the adjustment result meets the preset convergence condition, such as terminating when the improvement of the power response index is less than 0.1%. This method effectively improves the robustness and response speed of the energy storage power station scheduling, and obtains an iterative optimized power distribution scheme.

[0060] Step S486: Convert the iterative optimized power distribution scheme into specific node power execution instructions, and form an adaptive energy flow distribution instruction including power values, execution timings, and response time requirements.

[0061] The embodiments of the present invention convert the iterative optimized power distribution scheme into specific execution instructions. Each instruction includes a node identifier, an allocated power value, an execution timing (for example, node A executes 2 seconds after the start of scheduling and lasts for 10 minutes), and a response time requirement (such as the response time is less than 200 ms). The instructions are sent to the power control units of each energy storage node through the control communication network of the energy storage power station to achieve precise power regulation and energy flow distribution. This adaptive energy flow distribution instruction ensures that the energy storage power station meets the scheduling requirements in a dynamic operating environment, while ensuring system safety and efficiency, and is applicable to specific application scenarios such as power grid peak shaving, emergency backup, and smooth output of new energy.

[0062] Preferably, step S5 includes the following steps: Step S51: Generate operation control commands for each energy storage node including specific charge and discharge power setting values and timing execution arrangements according to the adaptive energy flow distribution instruction, and send the operation control commands to each energy storage node to trigger the node to perform corresponding charge and discharge operations; The embodiments of the present invention convert the charge and discharge power setting values and specific execution timing parameters corresponding to each energy storage node into operation control commands according to the adaptive energy flow distribution instruction obtained in the previous step. During specific conversion, the control system refines the power value to the charge and discharge modules of a single node. For example, the charging power of node A is set to 1.2 megawatts, the charging start time is 5 seconds after the current scheduling time point, and the duration is 15 minutes to ensure precise matching of the scheduling requirements. The operation control commands are sent to each node controller one by one through the control network of the energy storage power station (such as industrial Ethernet or a dedicated communication protocol) to trigger the power electronic devices and control units inside the node to perform charge and discharge actions according to the instructions. This process includes confirming that the node receives the command and feedbacking the execution status to ensure that the instruction takes effect in a timely manner, and is applicable to actual scenarios such as power grid peak shaving and smooth output of new energy.

[0063] Step S52: Start the real-time operation status monitoring program of the energy storage nodes, collect the voltage, current, temperature, and state of charge of each node, and form a data stream of the real-time operation status of the nodes; In the embodiment of the present invention, the energy storage power station starts the real-time operation status monitoring program of the nodes, and uses a distributed sensor network to collect the key operation parameters of each node, including voltage, current, temperature, and state of charge (SOC). These data form a continuous data stream through high-frequency sampling (such as 10 times per second), which can reflect the power output and internal state of the nodes in real time. The monitoring program is deployed in the control center of the energy storage station, and industrial Internet of Things technology is used to ensure the real-time and integrity of the data, providing basic data support for subsequent operation status analysis. The application scenarios include the safety monitoring of energy storage devices and the evaluation of operation efficiency.

[0064] Step S53: Perform online analysis on the data stream of the real-time operation status of the nodes, calculate the power response delay time, voltage fluctuation amplitude, and temperature change rate of each node, and generate an evaluation report on the operation characteristics of the nodes; In the embodiment of the present invention, online analysis is performed on the data stream of the real-time operation status of the nodes collected. Through timestamp synchronization and signal processing technology, the power response delay time of each node is calculated, that is, the time required for the actual power output to reach the set value from receiving the charge and discharge instruction, usually required to be less than 200 milliseconds; at the same time, the voltage fluctuation amplitude is analyzed to reflect the smoothness of the node load change; and the temperature change rate is calculated to evaluate the heat dissipation and thermal stability of the node, ensuring that the temperature rise does not exceed 0.5 degrees Celsius per minute. Based on the above parameters, the system automatically generates an evaluation report on the operation characteristics of the nodes, providing detailed indicators of the node performance and potential abnormalities, which is convenient for operation and maintenance personnel to quickly locate problems and ensure the stable operation of the energy storage power station.

[0065] Step S54: Update the node aging differential parameters by using the sliding time window method according to the evaluation report on the operation characteristics of the nodes, dynamically evaluate the aging rate, and obtain the updated node aging differential parameters; In the embodiment of the present invention, based on the evaluation report on the operation characteristics, the system uses the sliding time window method to dynamically update the node aging differential parameters. The specific operation is to select the node operation data in the recent period (such as the past 24 hours), and count the change trend of the node performance indicators within the window, such as the increase in response delay, the intensification of voltage fluctuation, and the frequency of temperature anomalies, and infer the actual aging rate of the node in combination with the historical aging model. The characteristic of the sliding time window method is that the window moves forward continuously and the parameters are updated continuously to ensure the timeliness and accuracy of the aging evaluation. The resulting updated node aging differential parameters reflect the changes in the health status of different nodes during use, which is convenient for targeted maintenance and scheduling optimization.

[0066] Step S55: Calibrate and optimize the node health status evaluation model by using the node aging differentiation update parameters to form a closed-loop optimized energy storage power station operation monitoring and management mechanism.

[0067] In the embodiment of the present invention, the node health status evaluation model is calibrated and optimized by using the dynamically updated node aging differentiation parameters. The specific method is to input the latest aging parameters into the health model, adjust the performance decay weights and risk thresholds of each node in the model, and form an accurate health evaluation result. Through this closed-loop mechanism, the energy storage power station can feedback the node state changes in real time, dynamically adjust the operation strategy and maintenance plan, and realize the full life cycle management of the node health. This mechanism effectively improves the reliability and economy of the energy storage system and is applicable to the long-term operation monitoring and management of large-scale energy storage power stations.

[0068] The present invention also provides an energy storage power station operation monitoring and management system for executing the above-mentioned energy storage power station operation monitoring and management method. The energy storage power station operation monitoring and management system includes: A voltage characteristic acquisition module, configured to obtain the voltage data of each energy storage node in the static state, perform periodic voltage sampling within a preset static time to obtain voltage rebound time series data; perform temperature correction processing on the voltage rebound time series data to obtain standardized voltage rebound characteristic data; A health status evaluation module, configured to obtain the state of charge data of each energy storage node before and after operation; calculate the charge and discharge depth change values of each node according to the state of charge data; construct a node health status evaluation model according to the standardized voltage rebound characteristic data and the charge and discharge depth change values to obtain node aging differentiation parameters; A capacity grading and allocation module, configured to determine the available capacity level of each node according to the node aging differentiation parameters; calculate the optimal working range of each node based on the available capacity level to obtain node operation ability grading data; establish a node charge and discharge power allocation model by using the node operation ability grading data; An energy flow intelligent allocation module, configured to obtain the overall power scheduling instruction of the energy storage power station; decompose the overall power scheduling instruction according to the node charge and discharge power allocation model to obtain a node differentiation power allocation scheme; perform dynamic optimization on the node differentiation power allocation scheme by using the minimum power perturbation method to obtain an adaptive energy flow allocation instruction; A closed-loop monitoring and optimization module, configured to send the adaptive energy flow allocation instruction to each energy storage node, and monitor the real-time operation state of each node after executing the adaptive energy flow allocation instruction; update the node aging differentiation parameters based on the real-time operation state to form a closed-loop optimized energy storage power station operation monitoring and management mechanism.

[0069] Therefore, in any aspect, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the application documents are intended to be encompassed within the present invention.

[0070] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but rather to the broadest scope consistent with the principles and novel features invented herein.

Claims

1. A method for operating and monitoring a energy storage power station, characterized in that The following steps are involved: Step S1: obtaining voltage data of each energy storage node in a static state, and performing periodic voltage sampling within a preset static time, thereby obtaining voltage rebound time series data; Performing temperature correction on the voltage rebound time series data to obtain standardized voltage rebound characteristic data; Step S2: Obtain the state of charge data of each energy storage node before and after operation; calculate the charge and discharge depth change value of each node according to the state of charge data; construct a node health status assessment model according to the standardized voltage rebound characteristic data and the charge and discharge depth change value, and obtain the node aging differentiation parameter; Step S3: Determine the available capacity level of each node according to the node aging differentiation parameter; Calculate the optimal working range of each node based on the available capacity level to obtain the node operation capacity classification data; use the node operation capacity classification data to establish the node charging and discharging power allocation model; Step S4: obtaining the overall power dispatching instruction of the energy storage power station; decomposing the overall power dispatching instruction according to the node charging and discharging power allocation model to obtain the node differentiated power allocation scheme; dynamically optimizing the node differentiated power allocation scheme using the minimum power disturbance method to obtain the adaptive energy flow allocation instruction; Step S5: Send an adaptive energy flow allocation instruction to each energy storage node, and monitor the real-time operating status of each node after executing the adaptive energy flow allocation instruction; update the node aging differentiation parameters based on the real-time operating status to form a closed-loop optimized energy storage power station operation monitoring and management mechanism.

2. The operation monitoring and management method of the energy storage power station according to claim 1, wherein Step S1 includes the following steps: Step S11: determining a rest time parameter according to the pre-acquired cell model characteristics of each energy storage node, thereby obtaining rest time configuration data; Step S12: Control the energy storage node to enter the static mode according to the static time configuration data, interrupt all charging and discharging operations, and obtain a node static state confirmation signal; Step S13: after obtaining the node static state confirmation signal, start the periodic voltage sampling process to record the voltage change of each energy storage node based on the preset sampling frequency to obtain the original voltage time series sampling data; Step S14: arranging and preliminarily screening the original voltage time series sampling data, removing abnormal data points, and forming voltage rebound time series data; Step S15: performing temperature correction processing on the voltage rebound time series data to obtain standardized voltage rebound characteristic data.

3. The operation monitoring and management method of the energy storage power station according to claim 2, wherein Step S2 includes the following steps: Step S21: setting the state of charge detection time window parameters to be suitable for the state of charge acquisition frequency of each energy storage node type, and obtaining the state of charge monitoring configuration data; Step S22: collecting terminal voltage, current and temperature at specific times before and after the node charging and discharging process according to the state of charge monitoring configuration data to obtain original state data of the node before and after operation; Step S23: Preliminary filtering and calibration of the original data of the node state before and after operation to eliminate the influence of measurement interference and transient fluctuations, and obtain effective charge state data; Step S24: Calculate the change in state of charge of each energy storage node within a complete operation cycle based on the effective state of charge data to obtain a preliminary evaluation value of the node charge and discharge depth; Step S25: Correct the preliminary evaluation value of the charge and discharge depth of the node by combining the actual operation duration of the node and the magnitude of the charge and discharge current, so as to obtain the change value of the charge and discharge depth; Step S26: Construct a node health status evaluation model based on the standardized voltage rebound characteristic data and the change value of the charge and discharge depth, and obtain the node aging differentiation parameters.

4. The operation monitoring and management method of the energy storage power station according to claim 3, characterized in that Step S26 includes the following steps: Step S261: Perform data fusion and matching based on the standardized voltage rebound characteristic data and the change value of the charge and discharge depth to construct a node operation behavior characteristic matrix; Step S262: Analyze the correlation between the voltage rebound characteristic and the charge and discharge depth based on the node operation behavior characteristic matrix, so as to establish a preliminary node health evaluation model; Step S263: Calibrate the parameter weights in the preliminary node health evaluation model through pre-acquired historical operation data, optimize the model response characteristics, so as to obtain the node health status evaluation model; Step S264: Use the node health status evaluation model to quantitatively analyze the current status of each node, output comparable health status difference indicators, so as to obtain the node aging differentiation parameters.

5. The operation monitoring and management method of the energy storage power station according to claim 4, wherein, The determination of the available capacity level of each node according to the node aging differentiation parameters in Step S3 includes: Perform standardization processing on the node aging differentiation parameters using the 0-1 normalization method to eliminate the influence of different dimensions and obtain the standardized aging parameters; Set the aging parameter threshold interval [0.15, 0.35, 0.55, 0.75] as the classification boundary point based on the standardized aging parameters, and perform piecewise mapping to obtain the preliminary aging level classification result; Calculate the actual capacity loss rate based on the loss rate calculation formula according to the preliminary aging level classification result in combination with the pre-acquired node rated capacity value of each energy storage node, where the loss rate calculation formula is loss rate = basic loss rate + aging level coefficient × attenuation acceleration factor; Dynamically correct the actual capacity loss rate. For nodes with an operation time exceeding 1000 hours, add a correction value of 0.8% to the original loss rate for every additional 500 hours to obtain the corrected capacity loss rate; Calculate the actual available capacity percentage of the node according to the corrected capacity loss rate, where the actual available capacity percentage of the node = 100% - corrected capacity loss rate; set the capacity level division standard for the actual available capacity percentage of the node to form the node capacity classification standard; Rate and divide the actual available capacity percentage of each node based on the node capacity classification standard to obtain the node available capacity level data; Extract the nodes with adjacent level boundary values from the node available capacity level data, and perform secondary correction in combination with the stable voltage recovery rate in the voltage rebound characteristic data. When the recovery rate is lower than the threshold, lower the node level by one level to obtain the available capacity level.

6. The operation monitoring and management method of the energy storage power station according to claim 5, characterized in that, Step S3 includes the following steps: Establish a mapping relationship between the capacity level and the state of charge working range according to the available capacity level to obtain the preliminary state of charge working range data; Collect the voltage-state of charge curves of each node under standard charge and discharge conditions, and select the area where the voltage change rate is less than 0.05V / 1%SOC as the voltage platform area to obtain the node voltage platform interval data; Optimize and adjust the working range based on the node voltage platform interval data and the preliminary state of charge (SOC) working range data to ensure that the working range includes at least 75% of the voltage platform area, and obtain the adjusted working interval data; Conduct constant power charge and discharge tests on each node, charge and discharge at a rate of 0.5C, record the rate of change of the terminal voltage at different states of charge, and obtain the voltage response characteristic data; Calculate the internal resistance change trend of each node at different states of charge based on the voltage response characteristic data. Take the point where the internal resistance increases by 20% as the boundary to further optimize the boundary of the working interval and obtain the internal resistance optimized working interval data; Calculate the intersection area of the adjusted working interval data and the internal resistance optimized working interval data as the safe working interval of the node, and divide it according to the available interval width: width greater than 70% is level one, 50% - 70% is level two, 30% - 50% is level three, and less than 30% is level four, to obtain the node working interval level data; Establish a two-dimensional evaluation matrix based on the node working interval level data and the available capacity level data to obtain the preliminary node operation ability level; Conduct a power response ability test on the preliminary node operation ability level to obtain the power response ability rating. Specifically, the power response ability test is to apply a pulse power of 0.8C within the safe working interval for 30 seconds and record the voltage drop amplitude. A voltage drop less than 0.15V is excellent, 0.15V - 0.25V is good, 0.25V - 0.35V is medium, and greater than 0.35V is poor; Determine the node operation ability classification based on the preliminary node operation ability level and the power response ability rating, and form the node operation ability classification data including three-dimensional characteristics of capacity state, working interval, and power response; 7. The operation monitoring and management method of the energy storage power station according to claim 6, wherein The establishment of the node charge and discharge power allocation model using the node operation ability classification data in step S3 includes: Assign basic power weight coefficients to each node based on the node operation ability classification data to obtain the node basic power weight coefficient table; Collect the charge and discharge efficiency data of each node at different states of charge, conduct sampling tests at intervals of 10% state of charge, record the energy conversion efficiency at each sampling point, and obtain the node charge and discharge efficiency curve; Calculate the optimal charge and discharge efficiency interval of each node based on the node charge and discharge efficiency curve, and define the area with an efficiency higher than 90% as the optimal working point to obtain the node efficiency optimization interval data; Combine the node efficiency optimization interval data with the node safe working interval to calculate their intersection area as the node priority scheduling interval, and obtain the node priority scheduling interval data; design the power allocation coefficient within the interval according to the node priority scheduling interval data; Monitor the current state of charge of the node in real time, calculate the distance ratio between the current state of charge and the boundary of the priority scheduling interval. When the distance ratio is less than 0.2, linearly decay the basic power weight coefficient to obtain the boundary adjustment coefficient; Monitor the temperature change status of the nodes, set the normal operating temperature range to 15°C - 35°C. When the temperature exceeds this range, for every increase or decrease of 5°C, attenuate the basic power weight coefficient by 5% to obtain the temperature correction coefficient; establish a node charge and discharge power allocation model based on the power allocation coefficient within the interval, the boundary adjustment coefficient, and the temperature correction coefficient.

8. The operation monitoring and management method of the energy storage power station according to claim 7, characterized in that, Step S4 includes the following steps: Step S41: Obtain the overall power dispatch instruction of the energy storage power station, parse the instruction and extract the charge and discharge power demand value, the duration parameter, and the response time requirement to obtain the power dispatch parameter set; Step S42: Calculate the preliminary power distribution coefficient of each energy storage node according to the power dispatch parameter set in combination with the node charge and discharge power allocation model to obtain the preliminary power distribution plan; Step S43: Calculate the upper and lower limits of the power that each node can bear based on the preliminary power distribution plan and the node operation ability grading data of each energy storage node to obtain the node power constraint conditions; Step S44: Correct the preliminary power distribution plan according to the node power constraint conditions to ensure that the power values allocated to each node do not exceed its safe operating range to obtain the safe corrected power distribution plan; Step S45: Calculate the interval deviation adjustment coefficient based on the deviation degree between the current state of charge of each node and its optimal operating interval, and optimize the safe corrected power distribution plan to obtain the balanced optimized power distribution plan; Step S46: Monitor the real-time temperature change trend of each node. When the temperature change rate exceeds the preset threshold, calculate the temperature risk coefficient and dynamically adjust the balanced optimized power distribution plan to obtain the temperature adaptive power distribution plan; Step S47: Form a node differential power distribution plan based on the temperature adaptive power distribution plan, including the power distribution value and the execution order of each node; Step S48: Dynamically optimize the node differential power distribution plan using the minimum power perturbation method to obtain the adaptive energy flow distribution instruction.

9. The operation monitoring and management method of the energy storage power station according to claim 8, wherein, Step S5 includes the following steps: Step S51: Generate an operation control command for each energy storage node including the specific charge and discharge power setting value and the timing execution arrangement according to the adaptive energy flow distribution instruction, and send the operation control command to each energy storage node to trigger the node to perform the corresponding charge and discharge operation; Step S52: Start the real-time operation status monitoring program of the energy storage node, collect the voltage, current, temperature, and state of charge of each node to form a node real-time operation status data stream; Step S53: Perform an online analysis on the node real-time operation status data stream, calculate the power response delay time, voltage fluctuation amplitude, and temperature change rate of each node, and generate a node operation characteristic evaluation report; Step S54: Update the node aging differential parameter using the sliding time window method according to the node operation characteristic evaluation report, and dynamically evaluate the aging rate to obtain the node aging differential update parameter; Step S55: Calibrate and optimize the node health status evaluation model using the node aging differential update parameter to form a closed-loop optimized energy storage power station operation monitoring and management mechanism.

10. An energy storage power station operation monitoring and management system, characterized in that, For implementing the energy storage power station operation monitoring and management method as described in claim 1, the energy storage power station operation monitoring and management system includes: A voltage characteristic acquisition module, configured to obtain voltage data of each energy storage node in a static state, perform periodic voltage sampling within a preset static time, so as to obtain voltage rebound time series data; perform temperature correction processing on the voltage rebound time series data, so as to obtain standardized voltage rebound characteristic data; A health status evaluation module, configured to obtain the state of charge data of each energy storage node before and after operation; calculate the charge and discharge depth change values of each node according to the state of charge data; construct a node health status evaluation model based on the standardized voltage rebound characteristic data and the charge and discharge depth change values, and obtain node aging differentiation parameters; A capacity grading and allocation module, configured to determine the available capacity levels of each node according to the node aging differentiation parameters; calculate the optimal working intervals of each node based on the available capacity levels, so as to obtain node operation ability grading data; establish a node charge and discharge power allocation model using the node operation ability grading data; An energy flow intelligent allocation module, configured to obtain the overall power scheduling instruction of the energy storage power station; decompose the overall power scheduling instruction according to the node charge and discharge power allocation model, so as to obtain a node differentiation power allocation scheme; perform dynamic optimization on the node differentiation power allocation scheme using the minimum power perturbation method, so as to obtain an adaptive energy flow allocation instruction; A closed-loop monitoring and optimization module, configured to send the adaptive energy flow allocation instruction to each energy storage node, and monitor the real-time operation status of each node after executing the adaptive energy flow allocation instruction; update the node aging differentiation parameters based on the real-time operation status, and form a closed-loop optimized energy storage power station operation monitoring and management mechanism.

Citation Information

Patent Citations

  • Joint estimation method for state of health and state of charge of lithium battery based on machine learning

    CN112946499A

  • Distributed energy storage system management method and device, equipment and storage medium

    CN118399460A

  • Large-scale energy storage power station operation monitoring management system and management method

    CN119154513A

  • Battery energy balanced distribution and optimization method, device, equipment and storage medium

    CN119725825A

  • Storage battery management device and storage battery management method

    JP2020054214A

Cited By

  • Photovoltaic power generation control system and method thereof

    CN120528113A

  • Energy storage power station state monitoring method, system and server

    CN120824930A

  • Battery SOC real-time regulation and control method and device based on abnormal active safety protection

    CN120855578A

  • Frequency modulation method, device and equipment for multiple electrochemical energy storage power stations and medium

    CN120999685A

  • Group scheduling control method and system for PCS energy storage system

    CN121076907A