Energy storage container operation management and control system based on digital twinning

Abstract

This invention belongs to the field of energy storage system operation and management technology, and relates to an energy storage container operation and control system based on digital twins. It collects time-series data on the terminal voltage and shell temperature of each individual cell in the energy storage container under static conditions, inputs this data into a pre-constructed digital twin, and, following the physical consistency constraints within the cluster, iteratively identifies the self-discharge resistance and initial state of charge of each individual cell. The system drives the digital twin to simulate the entire cluster charging process, records the cumulative charging capacity value of each individual cell when the charging termination condition is triggered, locates the lagging cells that first limit the cluster charging depth through outlier analysis, calculates the capacity drag margin, and generates and executes pre-control commands. This invention uses digital twin technology to deeply identify the cell state during the static period and proactively simulate the charging process, helping to accurately locate and quantify the impact of lagging cells, thereby achieving preventative balancing and proactive maintenance of the energy storage container.

Description

Technical Field

[0001] This invention belongs to the field of energy storage system operation and management technology, and relates to an energy storage container operation and control system based on digital twins. Background Technology

[0002] As a unit in large-scale electrochemical energy storage systems, energy storage containers play a crucial role in peak shaving and valley filling in power systems, grid connection of renewable energy, and stable operation of microgrids. They typically integrate hundreds or thousands of individual battery cells, which are monitored and managed through a battery management system to ensure the safe and efficient operation of the entire battery cluster. The operational control of energy storage containers, especially the management of the consistency of their individual battery cells, is a key technological aspect in ensuring their performance and lifespan.

[0003] In existing technologies, the management of inconsistencies among individual battery cells within energy storage containers typically relies on monitoring the terminal voltage of each cell during or after charging and discharging. When a voltage deviation is detected between some cells and others, the battery management system activates an balancing circuit to correct the imbalance by discharging the higher-voltage cells or charging the lower-voltage cells. This voltage-based balancing strategy can mitigate the state differences between cells to some extent.

[0004] However, existing technical solutions have several shortcomings in handling inconsistencies within battery clusters: First, voltage threshold-based equalization methods are essentially passive responses, only triggered when inconsistencies have accumulated to the point of producing significant voltage differences, lacking the ability for early warning and preventative intervention.

[0005] Secondly, simple voltage information is insufficient to reveal the root cause of inconsistencies. For example, it is impossible to effectively distinguish whether the voltage difference is due to different initial states of charge or due to inherent performance degradation such as excessive self-discharge rate or increased internal resistance of individual cells. This means that balancing strategies may only address the symptoms, not the root cause.

[0006] Finally, existing management methods lack the ability to predict future operating conditions and cannot estimate the impact of inconsistencies on charging depth and available capacity before charging, resulting in a lack of foresight in the formulation of operating strategies. Summary of the Invention

[0007] In view of this, in order to solve the problems mentioned in the background technology, a digital twin-based energy storage container operation and control system is proposed.

[0008] The objective of this invention can be achieved through the following technical solution: This invention provides an energy storage container operation and control system based on digital twins, including: a data acquisition module, a digital twin identification module, a charging simulation module, a capacity analysis module, and a pre-control decision module.

[0009] The data acquisition module is connected to the digital twin identification module, the digital twin identification module is connected to the charging simulation module, the charging simulation module is connected to the capacity analysis module, and the capacity analysis module is connected to the pre-control decision module.

[0010] The data acquisition module acquires the time-series sampling values ​​of the terminal voltage and shell temperature of each individual cell after the energy storage container enters the static state. The terminal voltage and shell temperature values ​​are stored in a structured manner according to the individual cell identifier and the sampling time section to construct the static monitoring data tensor.

[0011] The digital twin identification module inputs the static monitoring data tensor into a pre-constructed digital twin and, following the physical consistency constraint within the cluster, it iteratively identifies the self-discharge resistance and initial state of charge of each individual cell, generating a self-discharge rate feature vector for the individual cells within the cluster. The digital twin is constructed based on an equivalent circuit model of the individual cell containing the self-discharge resistance parameter and a lumped parameter thermal model.

[0012] The charging simulation module, based on the self-discharge rate feature vector, drives the digital twin to simulate the charging process of the entire cluster under a predetermined charging strategy, records the cumulative charging capacity value of each individual cell when the charging termination condition is triggered, and summarizes and generates a sequence of the chargeable capacity of individual cells within the cluster.

[0013] The capacity analysis module performs outlier analysis on the chargeable capacity sequence to locate the lagging cells that first limit the cluster charging depth and calculates the capacity drag margin caused by the lagging cells.

[0014] The pre-control decision module generates and executes pre-control instructions based on the capacity drag margin. The pre-control instructions are used to trigger pre-balancing operations or proactive maintenance prompts for the lagging units.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) This invention constructs a digital twin to identify the self-discharge characteristics of each individual cell online after the energy storage container enters a static state. Compared with the traditional passive balancing strategy that relies on voltage thresholds, it can reveal the inherent performance differences between individual cells from the root, realizing the transformation from post-event remediation to pre-event prediction, thereby identifying potential lagging cells before the fault appears, which helps to improve the depth and foresight of energy storage system status assessment.

[0016] (2) By simulating the full-cluster charging process within an energy storage container under a predetermined charging strategy, this invention can quantify the charging potential of each individual cell and identify the bottleneck that first limits the depth of charging. This forward-looking assessment enables the system to anticipate the capacity loss that may be caused by inconsistency between individual cells before the actual charging operation begins, providing a quantitative target for subsequent optimization control and helping to maximize the actual usable capacity of the entire battery cluster.

[0017] (3) The present invention establishes an intelligent decision-making and pre-control execution mechanism based on capacity drag margin. It can autonomously decide on pre-balancing operation or proactive maintenance prompts based on the degree of impact of lagging cells on the overall cluster performance and the actual capability of online balancing circuits, avoid invalid or incomplete balancing actions, achieve precise allocation of operation and maintenance resources, and improve the automation and intelligence level of energy storage container operation and management.

[0018] (4) This invention introduces closed-loop feedback control based on digital twin pre-simulation in the pre-equalization operation. By comparing the actual voltage rise rate of lagging cells in real time during the equalization process, the equalization effect is dynamically monitored and the control strategy is adjusted in a timely manner, and even the parameters are re-identified when the model is mismatched. This ensures the accuracy and safety of the pre-equalization operation and guarantees that the intervention measures can accurately restore the lagging cells to a healthy state, thereby enhancing the long-term operational reliability of the energy storage container. Attached Figure Description

[0019] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the module connection of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see Figure 1 As shown, the present invention provides an energy storage container operation and control system based on digital twins, including: a data acquisition module, a digital twin identification module, a charging simulation module, a capacity analysis module, and a pre-control decision module.

[0023] The data acquisition module is connected to the digital twin identification module, the digital twin identification module is connected to the charging simulation module, the charging simulation module is connected to the capacity analysis module, and the capacity analysis module is connected to the pre-control decision module.

[0024] The data acquisition module acquires the time-series sampling values ​​of the terminal voltage and shell temperature of each individual cell after the energy storage container enters the static state. The terminal voltage and shell temperature values ​​are stored in a structured manner according to the individual cell identifier and the sampling time section to construct a static monitoring data tensor.

[0025] In this embodiment, the static state is determined in the following way: The energy management system of the energy storage container is pre-connected to obtain the output current monitoring value from the energy storage converter and the charging and discharging status flags of the battery management system in real time.

[0026] When the charging / discharging status flag is in a non-charging and non-discharging state, the continuous current monitoring logic is started: the output current value is continuously collected at a preset sampling period, a sliding data window is constructed and the arithmetic mean of the current sampling values ​​in the window is calculated. If the arithmetic mean is lower than the static current allowable threshold and the current range in the window is lower than the allowable fluctuation threshold, the current condition is determined to be met and a timer is started to record the duration of continuous satisfaction.

[0027] It should be noted that the static current allowable threshold is determined as follows: when the energy storage container is in an absolute zero current condition, the output value of the current sensor is collected, the static zero drift offset and the system noise standard deviation are statistically analyzed, and the static current allowable threshold is set as the sum of the static zero drift offset and the system noise standard deviation by a preset multiple, where the preset multiple is 3 to 6 times.

[0028] The permissible fluctuation threshold is determined as follows: under absolute zero current conditions, the distribution of current range within the sliding data window is statistically analyzed, and the upper quantile of the range distribution is taken as the permissible fluctuation threshold.

[0029] During the timing process, if the arithmetic mean or range of the current exceeds the corresponding threshold in any sampling period, or if the status flag is switched, the timer will be cleared and re-monitored.

[0030] When the timer reaches the polarization decay completion time, a static state trigger signal is output.

[0031] The polarization attenuation completion time is dynamically determined based on the polarization time constant in the single-unit equivalent circuit model. The specific determination process is as follows: according to the structural parameters of the single-unit equivalent circuit model in the digital twin, the polarization resistance and polarization capacitance of each polarization branch are obtained. The polarization time constant corresponding to each polarization branch is obtained by multiplying the polarization resistance and polarization capacitance. The largest polarization time constant is selected as the basic polarization attenuation time constant.

[0032] Based on the tolerance requirements for residual polarization voltage identified during static identification, a threshold percentage threshold is set for the polarization voltage to decay to its initial value, typically between 1% and 5% of the initial value.

[0033] According to the first-order exponential decay law, the time multiple required to decay to the stated percentage threshold is the negative of the natural logarithm of the percentage, i.e. In the formula The time multiple of the problem to be solved The percentage threshold that the polarization voltage needs to decay to its initial value.

[0034] Multiplying the basic polarization decay time constant by the time factor yields the polarization decay completion time.

[0035] The digital twin identification module inputs the static monitoring data tensor into a pre-constructed digital twin and, following the intra-cluster physical consistency constraint, it iteratively identifies the self-discharge resistance and initial state of charge of each individual cell, generating a self-discharge rate feature vector for the individual cells within the cluster. The digital twin is constructed based on an individual equivalent circuit model containing self-discharge resistance parameters and a lumped parameter thermal model.

[0036] Specifically, the digital twin is configured to take the estimated self-discharge resistance, estimated initial state of charge, and historical operating data of each individual cell as input, perform a simultaneous solution of the electrical state equation and the thermal state equation, and output the simulated terminal voltage, simulated casing temperature, and cumulative charging capacity of each individual cell at any time section.

[0037] A digital twin consists of the following components: (I) Single-cell equivalent circuit model: A second-order RC equivalent circuit architecture is adopted, with independent configuration for each single cell. The single-cell equivalent circuit model consists of the following circuit elements connected in series: Open-circuit voltage source: Its voltage value satisfies a pre-defined open-circuit voltage-state of charge mapping relationship with the state of charge of a single cell; Ohmic internal resistance: connected in series in the circuit to simulate the instantaneous voltage drop generated when the charging and discharging current flows through the conductive medium inside the battery; First polarization RC network: It is composed of a first polarization resistor and a first polarization capacitor connected in parallel, and is used to simulate the short time constant polarization effect caused by charge transfer and diffusion processes during electrochemical reactions; Second polarization RC network: It is composed of a second polarization resistor and a second polarization capacitor connected in parallel. It is used to simulate the long-time constant polarization effect caused by the diffusion process of lithium ions inside solid-phase active materials. Self-discharge resistance: connected in parallel across an open-circuit voltage source, its reciprocal represents the admittance characteristics of the leakage current path caused by internal side reactions during the resting period of a single cell.

[0038] The circuit elements are connected in series according to Kirchhoff's voltage law. The terminal voltage of the series circuit is equal to the open circuit voltage minus the voltage drop across each series element.

[0039] (ii) Lumped parameter thermal model: configured independently for each individual cell, consisting of the following thermal network elements: Lumped thermal capacity node: used to characterize the overall thermal energy storage capacity of a single battery cell. Its thermal capacity value is pre-calibrated based on the product of the mass and specific heat capacity of the single battery cell. Internal heat source port: Coupled with the single-cell equivalent circuit model, the input is the instantaneous heat generation power value of the single cell calculated by the single-cell equivalent circuit model. The instantaneous heat generation power value of the single cell is calculated from the terminal voltage simulation value and the self-discharge current according to Joule's law. External thermal boundary port: used to access the temperature simulation values ​​of adjacent individual cells and the ambient temperature sampling values ​​of the energy storage container; Equivalent thermal resistance network: It consists of multiple equivalent thermal resistances, which are connected to the lumped thermal capacity nodes and the external thermal boundary ports respectively, and are used to simulate the conduction resistance of heat on different heat transfer paths.

[0040] The above-mentioned thermal network elements establish a thermal balance equation based on the law of conservation of energy: the rate of temperature change of the lumped heat capacity node is equal to the ratio of the net heat flowing into the node per unit time to the heat capacity value.

[0041] The single-unit equivalent circuit model and the lumped-parameter thermal model are solved simultaneously in the following manner: The single-unit equivalent circuit model takes the calculated current and terminal voltage as input, substitutes them into the Joule heat generation formula of the relevant prior art to calculate the instantaneous heat generation power, and outputs it to the internal heat source port of the lumped parameter thermal model. The lumped parameter thermal model feeds back the calculated temperature simulation value to the single-unit equivalent circuit model to update the temperature-related parameters in the electrical sub-model, such as the temperature correction coefficient of the open-circuit voltage, the temperature correction coefficient of the ohmic internal resistance, the temperature correction coefficient of each polarization resistor, and the temperature correction coefficient of each polarization capacitor.

[0042] The execution process of a digital twin is as follows: Step 1: Initialization: At the initial moment after the energy storage container enters the static state, the sampled values ​​of the terminal voltage of each individual cell and the sampled values ​​of the shell temperature are obtained as initial conditions. Initial state of charge estimate, initial temperature estimate and self-discharge resistance estimate are set independently for each individual cell.

[0043] The initial state of charge estimate is calculated by reverse calculation based on the terminal voltage sample value and the open circuit voltage-state of charge mapping relationship. The initial temperature estimate is directly taken from the shell temperature sample value. The self-discharge resistance estimate is initially set to infinite (assuming no self-discharge), and then updated after the first iteration identification.

[0044] Step 2, Single-step recursive calculation: Set the integration time step For each individual cell, from the current moment Starting from the beginning, calculate backwards to the end. time.

[0045] Given the current time The state of the single-unit equivalent circuit model includes: the voltage across the first polarized capacitor. Voltage across the second polarized capacitor With the current input current (Stillness stage) The equivalent circuit model of the driving unit is used to recursively calculate the voltage based on Kirchhoff's voltage law and the differential equations of the RC network. Simulated terminal voltage value at time 1 and the voltage across each polarization capacitor , : Recursive formula for polarization capacitor voltage: ,in The first polarization time constant, For the first polarization resistor, This is the first polarization capacitor. This is a preset constant; the recursive formula for the voltage of the second polarized capacitor is analogous.

[0046] Terminal voltage calculation formula: ,in For ohmic internal resistance, The open-circuit voltage value at the current moment is represented by the current state of charge value. pass The mapping curve is obtained through interpolation.

[0047] Cumulative discharge capacity recursion .

[0048] Instantaneous heat generation power calculated based on the unit equivalent circuit model Combined with the current temperature estimate from the lumped parameter thermal model Substituting into the energy balance equation of the heat network, the calculation is performed recursively. Temperature estimate at time : Recursive derivation of the energy balance equation: ,in To centralize heat capacity, The heat dissipation power is calculated by dividing the difference between the current temperature estimate and the fixed temperature boundary by the equivalent thermal resistance.

[0049] Temperature estimate calculated using this step size According to the pre-calibrated temperature-parameter mapping relationship, the temperature correction coefficient of each parameter is calculated. The room temperature calibration value of each parameter is multiplied by the corresponding temperature correction coefficient to obtain the parameter value at the current temperature. The update range includes the temperature compensation value of the open circuit voltage-state of charge mapping curve, ohmic internal resistance, first polarization resistor, first polarization capacitor, second polarization resistor, and second polarization capacitor.

[0050] Step 3: Iterative Progression: Repeat Step 2, calculating sequentially using a series integration method. , ...until the preset simulation duration is completed or the terminal condition is triggered. During this process, the individual equivalent circuit model and the lumped parameter thermal model are updated alternately within each integration time step. The individual equivalent circuit model depends on the temperature parameter value of the previous moment, and the lumped parameter thermal model depends on the instantaneous heat generation power value of the previous moment, forming a coupled simultaneous solution.

[0051] It should be noted that the pre-calibrated open-circuit voltage-state-of-charge mapping relationship is a characteristic curve established by performing pulse charge-discharge tests on single cells of the same model, recording the open-circuit voltage values ​​corresponding to different state-of-charge points, and then performing polynomial fitting or interpolation processing.

[0052] The pre-calibrated temperature-parameter mapping relationship is established by repeatedly performing pulse charge-discharge tests at different ambient temperatures to obtain model parameter values ​​(including ohmic internal resistance, polarization resistance, polarization capacitance, etc.) at each temperature point, and then performing interpolation or fitting processing. Both mapping relationships were obtained by technicians through experiments and records before system development.

[0053] In this embodiment, the intra-cluster physical consistency constraint includes the following execution logic: During the alternating iterative identification process, the cluster terminal voltage of each battery cluster inside the energy storage container is collected in real time.

[0054] The cluster-end voltage measurement value is used as the sum of the simulated terminal voltage values ​​of each individual cell as a constraint condition. The simulated terminal voltage values ​​of each individual cell output in each iteration are normalized and adjusted so that the sum of the simulated terminal voltage values ​​of each individual cell after normalization is equal to the cluster-end voltage measurement value.

[0055] The normalization adjustment process is as follows: calculate the total deviation between the sum of the simulated terminal voltage values ​​of each individual cell in the current round and the measured terminal voltage value of the cluster, and distribute the total deviation to each individual cell according to the deviation ratio between the simulated terminal voltage value of each individual cell and the average voltage to obtain the normalized simulated terminal voltage value of each individual cell.

[0056] The spatial distribution of the temperature of each individual cell casing is input into the lumped parameter thermal network model in the digital twin. Specifically, in the lumped parameter thermal network model, adjacent individual cells are connected by equivalent thermal resistance. The expected temperature gradient between adjacent individual cells is calculated by multiplying the instantaneous heat generation power value and the equivalent thermal resistance value.

[0057] When the difference between the actual casing temperature of a single cell and the casing temperature of an adjacent single cell deviates from the expected temperature gradient by a preset value, it is determined that the estimated self-discharge resistance of that single cell has deviated, and the estimated self-discharge resistance of that single cell is forcibly corrected.

[0058] It should be noted that the preset amplitude is set based on the statistical fluctuation range of the temperature field under normal operating conditions of the energy storage container, and is twice the standard deviation of the temperature field.

[0059] The specific method of forced correction is as follows: Individual cells with deviations in their self-discharge resistance estimates are marked as target individual cells, and the correction is performed using the formula... Calculate the corrected self-discharge resistance estimate of the target single cell, where, This represents the corrected estimated self-discharge resistance value of the target individual battery cell. This represents the estimated self-discharge resistance of adjacent individual cells, indicated by the subscript. and These are used to distinguish different individual battery cells. Refers to the target individual battery cell that needs to be corrected. Refers to the individual battery cell adjacent to the target individual battery cell. This represents the equivalent thermal resistance between the target cell and its adjacent cells. This represents the difference between the actual casing temperature of the target cell and the casing temperature of the adjacent cell. This indicates the current open-circuit voltage value of the target individual cell.

[0060] like ,but ;like ,but .

[0061] It represents the equivalent conductance generated by the self-discharge behavior of adjacent individual cells when they are in a static state.

[0062] This represents the self-discharge conductance correction of the target cell relative to its adjacent cells.

[0063] This represents the equivalent conductance contribution due to the difference in heat generation caused by the self-discharge of the two cells. Taking its reciprocal yields the corrected estimate of the self-discharge resistance.

[0064] The formula as a whole is an analytical expression that calculates the self-discharge resistance of the target cell from the temperature deviation.

[0065] In this embodiment, the iterative identification of the self-discharge resistance and initial state of charge of each individual cell includes: In each iteration, the self-discharge resistance estimate of the current round is fixed, the simulated terminal voltage of each individual cell is compared with the real-time sampled value, the voltage residual sequence is calculated, and the initial state of charge estimate of each individual cell is updated with the goal of minimizing the sum of squares of the voltage residual sequence.

[0066] The specific update process is as follows: the rest period is divided into multiple sampling times, and for each sampling time, the difference between the simulated terminal voltage value and the sampled terminal voltage value is calculated and defined as the voltage residual.

[0067] Using the sum of squared voltage residuals at all sampling times during the resting period as the objective function, a functional relationship is established between the objective function and the initial state of charge to be identified. The simulated terminal voltage value is obtained recursively through the single-unit equivalent circuit model in the digital twin, with the initial state of charge as the initial condition.

[0068] A one-dimensional search method is employed to iterate through and solve for the initial state of charge (SOC) within a preset interval. The preset interval is determined based on the SOC range corresponding to the measured terminal voltage value in the open-circuit voltage-SOC mapping curve. In each search iteration, the current candidate initial SOC drives the digital twin to calculate the sequence of simulated terminal voltage values ​​for the entire resting period, and the objective function value is calculated.

[0069] The initial state of charge that minimizes the objective function value is determined by the interval contraction method and used as the estimated initial state of charge after this iteration.

[0070] b. Given the updated initial state of charge estimate, extract the measured voltage change trend of each individual cell after removing the initial polarization decay segment during the resting period. Compare the measured voltage change trend with the simulated voltage change trend output by the digital twin under the same conditions. Based on the deviation between the two, deduce the correction amount of the self-discharge resistance and output the updated self-discharge resistance estimate.

[0071] c. Repeat steps a and b until the relative change in the self-discharge resistance estimate between two adjacent iterations is lower than the preset convergence threshold, or the iteration count reaches the preset maximum iteration count. The preset maximum iteration count is 20 times for example. The self-discharge resistance estimate obtained when the iteration is terminated is taken as the final estimate. The final estimates of each individual cell are summarized to form the self-discharge rate feature vector.

[0072] It should be noted that the specific meaning of removing the initial polarization decay segment in step b is as follows: the data within the polarization decay completion time after the start of the resting period is defined as the initial polarization decay segment and removed. The terminal voltage data within the initial polarization decay segment is dominated by the polarization effect and cannot truly reflect the self-discharge behavior. After removal, the measured terminal voltage data sequence within the remaining resting period after the polarization decay completion time is extracted as valid data for subsequent trend analysis.

[0073] The specific method for calculating the self-discharge resistance correction in step b is as follows: The measured trend sequence of the terminal voltage after removing the initial polarization decay segment is denoted as... The simulated trend sequence of the terminal voltage output by the digital twin under the same conditions is denoted as... .

[0074] Calculate the deviation between the two sequences at each time step. The mean of the deviation sequence during the remaining static period is taken as the average voltage deviation. .

[0075] The formula for calculating the self-discharge resistance correction is: .

[0076] In the formula, This is the estimated value of the self-discharge resistance of the current wheel. This is the open-circuit voltage value at the current moment. This represents the total duration of the remaining resting period. The sensitivity coefficient for voltage deviation to resistance deviation is determined by the local slope of the open-circuit voltage-state-of-charge curve.

[0077] Set the updated self-discharge resistance estimate as... And restrict its value to a preset physical feasible range.

[0078] In step c, the relative change in the self-discharge resistance estimate between two adjacent iterations is defined as the ratio of the absolute difference between the self-discharge resistance estimates between two adjacent iterations to the self-discharge resistance estimate of the iteration that is earlier in the time sequence. The preset convergence threshold is a pre-set upper limit of the relative change, the size of which is set according to the accuracy requirements of self-discharge resistance identification and the parameter noise level of the digital twin model. An example value is 0.5%.

[0079] It should be further noted that this invention can distinguish between differences in initial state of charge (SOC) and differences in self-discharge rate (SCR). The principle behind this is that the difference in SOC is primarily reflected in the difference in terminal voltage at the start of the resting period, while the difference in SCR is reflected in the slope of voltage change over time. Alternating iterative identification utilizes the different characteristics of these two types of information on the time scale: step a calibrates the initial SOC by matching the absolute voltage level throughout the resting period; step b calibrates the SCR by matching the relative trend of voltage change. The orthogonality of these two methods in their optimization objectives ensures the uniqueness and physical interpretability of the identification results.

[0080] The charging simulation module, based on the self-discharge rate feature vector, drives the digital twin to simulate the charging process of the entire cluster under a predetermined charging strategy, records the cumulative charging capacity value corresponding to each individual cell when the charging termination condition is triggered, and summarizes and generates a sequence of the chargeable capacity of individual cells within the cluster.

[0081] In this embodiment, the driving digital twin simulates the charging process of the entire cluster under a predetermined charging strategy, including: reading the currently set charging current threshold and charging cutoff voltage threshold from the energy management system.

[0082] Using the charging current threshold as the constant current input for the charging process, the self-discharge rate characteristic vector, and the initial state of charge as the starting conditions, the digital twin is driven to recursively calculate the simulated terminal voltage, simulated casing temperature, and cumulative charging capacity of each individual cell at the end of each integral time step using a preset integral time step.

[0083] During the recursive process, it is determined in real time whether the simulated terminal voltage value of each individual cell reaches the charging cutoff voltage threshold, and whether the simulated casing temperature reaches the temperature safety protection threshold.

[0084] When the simulated terminal voltage of any single cell reaches the charging cutoff voltage threshold or the simulated casing temperature reaches the temperature safety protection threshold, the charging simulation of the single cell is immediately terminated, and the cumulative charging capacity value of the single cell at the termination time is recorded as the chargeable capacity.

[0085] After traversing all individual cells, the rechargeable capacity of each individual cell is summarized and sorted according to the individual cell identifier to generate a sequence of individual cell rechargeable capacities.

[0086] It should be noted that the charging cut-off voltage threshold is pre-configured according to the maximum allowable charging voltage value specified by the battery cell manufacturer, the temperature safety protection threshold is pre-configured according to the maximum allowable operating temperature defined in the battery specification, and the charging current threshold is limited by the maximum allowable charging current calibrated by the battery management system, specifically taking a value of 0.5 times the maximum allowable charging current.

[0087] The capacity analysis module performs outlier analysis on the chargeable capacity sequence to locate the lagging cells that first limit the cluster charging depth and calculates the capacity drag margin caused by the lagging cells.

[0088] In this embodiment, the outlier analysis of the fillable capacity sequence includes: Calculate the median and interquartile range of the fillable capacity sequence, and use the position of the median offset downward by a preset multiple of the interquartile range as the lower limit for outlier determination.

[0089] The rechargeable capacity of each individual cell is iterated sequentially, and individual cells that are below the outlier determination threshold are marked as candidate lagging cells.

[0090] The smallest chargeable cell in the candidate lagging cell set is selected as the lagging cell that first limits the cluster charging depth.

[0091] In this embodiment, the driving digital twin simulates the charging process of the entire cluster under a predetermined charging strategy, and also includes recording the trajectory of the simulated terminal voltage value of each individual cell as a function of simulation time to form a terminal voltage rise curve, which is used to authenticate lagging cells. The voltage inflection point of the terminal voltage rise curve of each individual cell at the end of the charging process is extracted to form a voltage inflection point sequence. Specifically, the end of the charging process refers to the charging stage from the starting position of the state of charge entering the steep rise range to the triggering of the charging termination condition. At the end of the charging process, the first derivative of the terminal voltage rise curve of each individual cell is calculated by the differential method. The simulation time corresponding to the maximum value of the first derivative or the zero crossing point of the second derivative is determined as the voltage inflection point of the individual cell, thereby forming a voltage inflection point sequence.

[0092] By comparing the voltage inflection point of each individual cell, at least one abnormal cell whose voltage inflection point is significantly earlier than that of other cells in the cluster can be identified.

[0093] The method for determining if the voltage inflection point is significantly earlier is as follows: The voltage inflection point times within the voltage inflection point sequence are reordered according to chronological order, and the median is selected as the reference inflection point time. The absolute time difference between the inflection point time of each individual cell and the reference inflection point time is calculated and recorded as the inflection point advance. A threshold for the inflection point advance is set, which is determined based on a fixed multiple (e.g., 10-50 times) of the integral time step of the digital twin simulation, with a typical value range of 30 to 120 seconds. When the voltage inflection point time of a certain individual cell is earlier than the reference inflection point time and the inflection point advance exceeds the threshold, the individual cell is determined to be an abnormal cell with a significantly earlier voltage inflection point time.

[0094] The index of the abnormal cell is cross-validated with the index of the smallest rechargeable cell identified from the rechargeable capacity sequence. The specific cross-validation process is as follows: when the index of an abnormal cell and the index of the smallest rechargeable cell identified from the rechargeable capacity sequence point to the same cell, the cross-validation is successful, the cell is finally identified as a lagging cell, and its cell index is output for subsequent pre-control decisions.

[0095] When the indexes of all abnormal cells and the index of the minimum chargeable capacity cell do not point to the same cell, all abnormal cells and the minimum chargeable capacity cell are marked as candidate lagging cells. The capacity drag margin and inflection point lead of each candidate lagging cell are calculated separately. The normalized capacity drag margin and inflection point lead are added together using the maximum-minimum normalization method, and the cell with the largest cumulative value is selected as the final lagging cell.

[0096] In this embodiment, the calculation of the capacity drag margin caused by lagging individual units includes: After removing the rechargeable capacity of lagging cells from the rechargeable capacity sequence, the average rechargeable capacity of the remaining cells is calculated as a reference rechargeable capacity.

[0097] The absolute capacity drag value is obtained by subtracting the rechargeable capacity of the lagging cells from the reference rechargeable capacity.

[0098] The ratio of absolute capacity drag to reference fillable capacity is used as the capacity drag margin caused by lagging cells.

[0099] The pre-control decision module generates and executes pre-control instructions based on the capacity drag margin. The pre-control instructions are used to trigger pre-balancing operations or proactive maintenance prompts for the lagging units.

[0100] In this embodiment, the step of generating and executing pre-control instructions based on capacity drag margin includes: The capacity drag margin is matched with the power compensation capability of the online equalization circuit, and the equalization current and equalization time required to complete the state of charge compensation of the lagging individual cells during the remaining rest period are calculated.

[0101] The remaining static time period is defined as the length of time from the current moment until the energy storage container exits the static state. It is obtained by subtracting the already elapsed static time period from the preset static plan duration of the energy management system to obtain an estimated value for the remaining static time period. When the preset static plan cannot be obtained, the historical average of the time interval between the most recent static state trigger and wake-up is used as the remaining static time period.

[0102] The calculation method for the equalization current and equalization duration is as follows: First, based on the capacity drag margin of the lagging cells... (unit: and nominal voltage Calculate the amount of electricity that needs to be compensated. Here, 3600 is used to convert electricity into joules or directly. The calculation depends on the energy metering method of the equalization circuit.

[0103] Secondly, read the maximum permissible balancing current of the online balancing circuit. Its main limitation is the equalization circuit hardware specifications, typically ranging from 0.5A to 2A, and the remaining time of the current resting period. .

[0104] Calculate the minimum required equilibrium time ,in For the preset energy conversion efficiency of the online equalization circuit, if Then set the equalization current. Balanced duration ;like Then set the equalization duration. And reverse the calculation of the required equalization current. If the reverse calculation result exceeds If the compensation cannot be completed within the remaining time period, an active maintenance prompt will be triggered.

[0105] If the equalization time is less than the remaining rest period, a pre-equalization activation command is generated, which enables the equalization circuit to provide targeted power replenishment to the lagging cells before the rest period ends.

[0106] If the equalization duration is greater than or equal to the remaining idle time, an active maintenance prompt instruction is generated, the lagging cell index and capacity drag margin are sent to the energy management system, the lagging cell is marked as pending maintenance, and the current pre-equalization operation is prohibited.

[0107] In this embodiment, after generating the pre-equalization activation instruction, the method further includes: The voltage rise rate of the lagging cells during the equalization process is monitored and compared in real time with the voltage rise reference curve pre-deduced by the digital twin to obtain the equalization deviation. The equalization deviation is defined as the relative error between the voltage rise rate at the current moment and the slope of the corresponding point of the voltage rise reference curve. Specifically, the absolute difference between the voltage rise rate at the current moment and the slope of the corresponding point of the voltage rise reference curve is used as the numerator, and the sum of the slope of the corresponding point of the voltage rise reference curve and a preset small positive number (to prevent division by zero) is used as the denominator, and the ratio is calculated.

[0108] When the balancing deviation exceeds the allowable deviation limit set based on the control accuracy of the balancing circuit, it is determined whether the energy storage container is still in a static state. Based on the determination result, the balancing duty cycle is adjusted or balancing is paused, and the self-discharge rate identification is retried to update the self-discharge rate feature vector. Specifically, under the premise that the energy storage container is still in a static state, if the balancing deviation exceeds the allowable deviation limit but does not exceed twice the allowable deviation limit, the balancing duty cycle is reduced proportionally according to the magnitude of the deviation. The ratio of the balancing deviation to the allowable deviation limit is multiplied by the current balancing duty cycle to obtain the duty cycle adjustment amount.

[0109] If the deviation exceeds twice the allowable limit, or if the energy storage container is not in a static state, the equalization operation is immediately suspended, the lagging cells are re-identified, and a new round of alternating iterative identification is triggered after the system enters the next static cycle. After the self-discharge rate identification is completed, the terminal voltage rise reference curve is regenerated based on the new self-discharge rate feature vector, and the equalization operation is resumed.

[0110] Once the equalization deviation returns to within the allowable deviation limit, the pre-equalization operation continues until the difference between the estimated state of charge of the lagging cell and the median state of charge within the cluster is less than the convergence tolerance set based on the state of charge estimation resolution. For a BMS that uses a combination of ampere-hour integration and open-circuit voltage calibration, the state of charge resolution is typically 0.5% to 2%, and the convergence tolerance is set to 1 to 2 times the state of charge estimation resolution, with a typical value of 1%.

[0111] In addition, if the balance deviation still fails to converge after two consecutive re-identifications within a single rest period, the pre-balancing operation will be terminated and an anomaly will be reported.

[0112] It should be noted that the pre-deduction method of the terminal voltage rise reference curve is as follows: while generating the pre-equalization activation command, the estimated value of the current state of charge of the lagging cell, the estimated value of the self-discharge resistance, and the set equalization current are used as inputs to drive the digital twin to simulate the rise trajectory of the terminal voltage of the lagging cell over time under the action of the equalization current in the same recursive way as the charging simulation, so as to obtain the terminal voltage rise reference curve.

[0113] The control accuracy of the equalization circuit refers to the current control error of the equalization circuit, and the corresponding allowable deviation limit can be set to 5% for example.

[0114] The online balancing circuit is a balancing execution unit configured in the energy storage container battery management system for performing targeted power compensation on each individual cell. It is electrically connected to the positive and negative busbars of the battery cluster and the tap nodes of each individual cell. After receiving the pre-balancing activation command from the pre-control decision module, it starts to perform targeted power compensation operation on the specified individual cell.

[0115] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A digital twin-based energy storage container operation and control system, characterized in that, include: The data acquisition module acquires the time-series sampling values ​​of the terminal voltage and shell temperature of each individual cell after the energy storage container enters the static state. The terminal voltage and shell temperature values ​​are stored in a structured manner according to the individual cell identifier and the sampling time section to construct the static monitoring data tensor. The digital twin identification module inputs the static monitoring data tensor into a pre-constructed digital twin and, following the intra-cluster physical consistency constraint, it iteratively identifies the self-discharge resistance and initial state of charge of each individual cell, generating a self-discharge rate feature vector for the individual cells within the cluster. The digital twin is constructed based on an individual equivalent circuit model containing self-discharge resistance parameters and a lumped parameter thermal model. The charging simulation module, based on the self-discharge rate feature vector, drives the digital twin to simulate the charging process of the entire cluster under a predetermined charging strategy, records the cumulative charging capacity value of each individual cell when the charging termination condition is triggered, and summarizes and generates a sequence of the chargeable capacity of individual cells within the cluster. The capacity analysis module performs outlier analysis on the chargeable capacity sequence to locate the lagging cells that first limit the cluster charging depth and calculates the capacity drag margin caused by the lagging cells. The pre-control decision module generates and executes pre-control instructions based on the capacity drag margin. The pre-control instructions are used to trigger pre-balancing operations or proactive maintenance prompts for the lagging units.

2. The energy storage container operation and control system based on digital twin according to claim 1, characterized in that, The static state is determined in the following way: The energy management system of the energy storage container is pre-connected to obtain the output current monitoring value from the energy storage converter and the charge and discharge status flag of the battery management system in real time. When the charging / discharging status flag is in a non-charging and non-discharging state, the continuous current monitoring logic is started: the output current value is continuously collected at a preset sampling period, a sliding data window is constructed and the arithmetic mean of the current sampling values ​​in the window is calculated. If the arithmetic mean is lower than the static current allowable threshold and the current range in the window is lower than the allowable fluctuation threshold, the current condition is determined to be met and a timer is started to record the duration of continuous satisfaction. During the timing process, if the arithmetic mean or range of the current exceeds the corresponding threshold in any sampling period, or if the status flag is switched, the timer will be cleared and re-monitored. When the timer reaches the polarization decay completion time, a static state trigger signal is output. The polarization attenuation completion time is dynamically determined based on the polarization time constant in the single-unit equivalent circuit model.

3. The energy storage container operation and control system based on digital twin according to claim 1, characterized in that, The intra-cluster physical consistency constraint includes the following execution logic: During the alternating iterative identification process, the cluster terminal voltage measurement value of each battery cluster in the energy storage container is collected in real time; The cluster-end voltage measurement value is used as the sum of the simulated terminal voltage values ​​of each individual cell as a constraint condition. The simulated terminal voltage values ​​of each individual cell output in each iteration are normalized and adjusted so that the sum of the simulated terminal voltage values ​​of each individual cell after normalization is equal to the cluster-end voltage measurement value. The spatial distribution of the shell temperature of each individual cell is input into the lumped parameter thermal network model in the digital twin. The expected temperature gradient between adjacent individual cells is calculated in real time. When the difference between the actual shell temperature of a certain individual cell and the shell temperature of the adjacent individual cells deviates from the expected temperature gradient by a preset value, it is determined that the estimated value of the self-discharge resistance of that individual cell has deviated, and the estimated value of the self-discharge resistance of that individual cell is forcibly corrected.

4. The energy storage container operation and control system based on digital twin according to claim 1, characterized in that, The iterative identification of the self-discharge resistance and initial state of charge of each individual cell includes: a. In each iteration, fix the self-discharge resistance estimate of the current round, compare the simulated terminal voltage of each individual cell with the real-time sampled value, calculate the voltage residual sequence, and update the initial state of charge estimate of each individual cell with the goal of minimizing the sum of squares of the voltage residual sequence. b. Given the updated initial state of charge estimate, extract the measured change trend of the terminal voltage of each individual cell after removing the initial polarization decay segment during the resting period. Compare the measured change trend of the terminal voltage with the simulated change trend of the terminal voltage output by the digital twin under the same conditions. Based on the deviation between the two, deduce the correction amount of the self-discharge resistance and output the updated self-discharge resistance estimate. c. Repeat steps a and b until the relative change in the self-discharge resistance estimate between two adjacent iterations is lower than the preset convergence threshold, or the number of iterations reaches the preset maximum number of iterations. Then terminate the iteration and take the self-discharge resistance estimate obtained at the time of termination as the final estimate. Summarize the final estimates of each individual cell to form the self-discharge rate feature vector.

5. The energy storage container operation and control system based on digital twin according to claim 1, characterized in that, The driving digital twin simulates the charging process of the entire cluster under a predetermined charging strategy, including: Read the currently set charging current threshold and charging cut-off voltage threshold from the energy management system; Using the charging current threshold as the constant current input for the charging process, the self-discharge rate characteristic vector and the initial state of charge as the starting conditions, the digital twin is driven to recursively calculate the simulated terminal voltage, simulated shell temperature and cumulative charging capacity of each individual cell at the end of each integral time step in a preset integral time step. During the recursive process, it is determined in real time whether the simulated terminal voltage value of each individual cell reaches the charging cut-off voltage threshold and whether the simulated casing temperature reaches the temperature safety protection threshold. When the simulated terminal voltage of any single cell reaches the charging cutoff voltage threshold or the simulated casing temperature reaches the temperature safety protection threshold, the charging simulation of the single cell is immediately terminated, and the cumulative charging capacity value of the single cell at the termination time is recorded as the chargeable capacity. After traversing all individual cells, the rechargeable capacity of each individual cell is summarized and sorted according to the individual cell identifier to generate a sequence of individual cell rechargeable capacities.

6. The energy storage container operation and control system based on digital twin according to claim 5, characterized in that, The outlier analysis of the fillable capacity sequence includes: Calculate the median and interquartile range of the fillable capacity sequence, and use the position of the median shifted downward by a preset multiple of the interquartile range as the lower limit for outlier determination; Iterate through the chargeable capacity of each individual cell in turn, and mark the individual cells that are below the outlier determination threshold as candidate lagging cells; The smallest chargeable cell in the candidate lagging cell set is selected as the lagging cell that first limits the cluster charging depth.

7. The energy storage container operation and control system based on digital twin according to claim 6, characterized in that, The driving digital twin simulates the charging process of the entire cluster under a predetermined charging strategy, and also includes recording the trajectory of the simulated terminal voltage value of each individual cell as a function of simulation time to form a terminal voltage rise curve, which is used to authenticate lagging cells. Extract the voltage rise curve of each individual cell at the voltage inflection point at the end of charging to form a voltage inflection point sequence; By comparing the voltage inflection point of each individual cell, at least one abnormal cell whose voltage inflection point is significantly earlier than that of other cells in the cluster can be identified. Cross-validate the index of the anomalous unit with the index of the smallest fillable unit identified from the fillable capacity sequence to confirm the identity of the lagging unit.

8. The energy storage container operation and control system based on digital twin according to claim 1, characterized in that, The calculation of the capacity drag margin caused by lagging individual units includes: After removing the rechargeable capacity of lagging cells from the rechargeable capacity sequence, the average rechargeable capacity of the remaining cells is calculated as a reference rechargeable capacity. The absolute capacity drag value is obtained by subtracting the rechargeable capacity of the lagging cells from the reference rechargeable capacity. The ratio of absolute capacity drag to reference fillable capacity is used as the capacity drag margin caused by lagging cells.

9. The energy storage container operation and control system based on digital twin according to claim 1, characterized in that, The generation and execution of pre-control instructions based on capacity drag margin includes: The capacity drag margin is matched with the power compensation capability of the online equalization circuit, and the equalization current and equalization time required to complete the compensation of the state of charge of the lagging individual cells during the remaining rest period are calculated. If the equalization time is less than the remaining rest period, a pre-equalization activation command is generated, so that the equalization circuit can perform targeted power replenishment on the lagging cells before the rest period ends. If the equalization duration is greater than or equal to the remaining idle time, an active maintenance prompt instruction is generated, the lagging cell index and capacity drag margin are sent to the energy management system, the lagging cell is marked as pending maintenance, and the current pre-equalization operation is prohibited.

10. A digital twin-based energy storage container operation and control system according to claim 9, characterized in that, After generating the pre-equalization activation instruction, the following is also included: The rate of voltage rise of the lagging cells during the equalization process is monitored and compared in real time with the reference curve of voltage rise pre-simulated by the digital twin to obtain the equalization deviation. When the balancing deviation exceeds the allowable deviation limit set based on the control accuracy of the balancing circuit, it is determined whether the energy storage container is still in a static state. Based on the determination result, the balancing duty cycle is adjusted or balancing is paused, and the self-discharge rate identification is retried to update the self-discharge rate feature vector. Once the equilibrium deviation returns to within the allowable deviation limit, the pre-equilibrium operation continues until the difference between the estimated state of charge of the lagging individual and the median state of charge within the cluster is less than the convergence tolerance set based on the resolution of the state of charge estimation.

Images