Industrial energy storage system operation state simulation method and system based on digital twinning

Abstract

This application provides a method and system for simulating the operational status of an industrial energy storage system based on digital twins. In this method, the digital twin model includes a first type of model and a second type of model, with each battery cell corresponding to a simulation node. During the time-domain recursive simulation, the local simulation error of each node is first determined, and then its error tolerance is dynamically adjusted according to the local operational status of each node. For nodes using the first type of model, when the local simulation error exceeds the upper limit, the system switches to the second type of model. For nodes using the second type of model, when the local simulation error is below the lower limit and the local simulation errors of all adjacent nodes have not exceeded their respective upper limits, the system reverts to the first type of model. Then, the system proceeds to the next simulation step and outputs the simulation results. Using this method, the computational overhead of the digital twin model of the industrial energy storage system can be dynamically reduced while ensuring simulation accuracy, and frequent model switching due to local disturbances can be avoided.

Description

Technical Field

[0001] This application relates to the field of digital twin technology, and more specifically, to a method and system for simulating the operating status of an industrial energy storage system based on digital twins. Background Technology

[0002] Digital twin technology has been widely used in the monitoring and simulation analysis of the operating status of industrial energy storage systems. By constructing a virtual model corresponding to the physical battery pack, key parameters such as the state of charge, temperature and terminal voltage of each battery cell can be simulated in real time, providing decision support for the safe operation and maintenance and optimized scheduling of energy storage systems. However, energy storage systems usually contain hundreds or thousands of battery cells. If a high-precision electrochemical model is used to simulate each cell in real time, the computational overhead will be extremely large, making it difficult to meet the application requirements of online operation.

[0003] In existing technologies, some solutions reduce computational burden by simplifying models, such as using equivalent circuit models for all battery cells. While these methods are computationally efficient, they struggle to accurately capture local anomalies, such as accelerated cell aging or precursors to thermal runaway. Other solutions attempt to switch between simplified and high-precision models based on error thresholds, but these typically rely solely on the current error of a single node, neglecting the electrical coupling and state consistency between adjacent battery cells. In industrial energy storage systems, local disturbances in one battery cell often affect surrounding cells through series and parallel connections, causing temporary fluctuations in the simulation errors of these cells. Existing fixed threshold or single-node judgment methods are prone to repeated model switching due to such local disturbances, failing to effectively reduce computational overhead and introducing additional switching overhead and state synchronization burden, thus weakening the stability and real-time performance of digital twin simulations. Therefore, dynamically reducing the computational overhead of digital twin models for industrial energy storage systems while ensuring simulation accuracy, and avoiding frequent model switching due to local disturbances, has become a challenge for the industry. Summary of the Invention

[0004] This application provides a simulation method and system for the operation status of an industrial energy storage system based on digital twins. It can dynamically reduce the computational overhead of the digital twin model of the industrial energy storage system while ensuring simulation accuracy, and avoid frequent model switching caused by local disturbances.

[0005] Firstly, this application provides a simulation method for the operational status of an industrial energy storage system based on digital twins, applied to a digital twin model of the industrial energy storage system. The digital twin model includes a first type of model and a second type of model, with the second type of model having higher simulation accuracy than the first type of model. The method includes the following steps: Construct the digital twin model, in which each battery cell corresponds to one simulation node; During the time-domain recursive simulation, the state of each simulation node is updated, and the predicted value of the measurable physical quantity is calculated. The local simulation error of the simulation node is determined based on the deviation between the predicted value and the measured value. Based on the local operating status of each simulation node, the error tolerance corresponding to that node is dynamically adjusted, and the error tolerance includes an upper limit and a lower limit. For nodes currently using the first type of model, when the local simulation error exceeds the corresponding error upper limit, the nodes are switched to the second type of model for state update starting from the next simulation step. For a node currently using the second type of model, when the local simulation error is lower than the corresponding lower limit of error and the local simulation errors of all its adjacent nodes do not exceed their respective upper limits of error, it is restored to the first type of model for state update from the next simulation step. Based on the updated node status and model configuration at the current step size, proceed to the next simulation step size iteration and output the simulation results of the industrial energy storage system's operating status.

[0006] In this embodiment, all simulation nodes initially use the first type of model to describe their operating state.

[0007] In this embodiment, constructing a digital twin model specifically includes: Obtain the battery cell layout diagram and electrical connection diagram of the industrial energy storage system; The spatial location and electrical properties of each battery cell are determined based on the battery cell layout diagram. According to the series and parallel connections specified in the electrical connection diagram, each battery cell is mapped to an independent simulation node, and a battery type label is set for each simulation node; Based on the battery type label and the spatial location, configure the parameters and state update interfaces of the corresponding first type model and second type model for each simulation node to complete the construction of the digital twin model.

[0008] In this embodiment, determining the local simulation error of the simulation node based on the deviation between the predicted value and the measured value specifically includes: Read the predicted output state vector of the simulation node at the current moment from the digital twin model; The measured state vector at the same moment is collected by sensors deployed on the corresponding battery cells; The predicted state vector is compared with the measured state vector component by component, and the absolute deviation of each component is calculated. The absolute deviations of all components are fused to obtain the local simulation error of the simulation node.

[0009] In this embodiment, dynamically adjusting the error tolerance corresponding to each simulation node based on its local operating state specifically includes: Obtain the current state of charge, temperature, and cumulative number of cycles for the simulation node; The health status decay factor is determined based on the cumulative number of cycles. Traverse all adjacent simulation nodes, calculate the absolute value of the difference in the state of charge between the node and each of its adjacent nodes, and take the maximum value as the maximum state deviation. The temperature adjustment coefficient is determined based on the deviation between the current temperature and the rated operating temperature; The health state decay factor, the maximum state deviation, and the temperature regulation coefficient are fused to obtain a comprehensive regulation factor; The error tolerance of the current step size is obtained by multiplying the preset upper and lower limits of the reference error by the comprehensive adjustment factor. The error tolerance includes the upper and lower limits of the error.

[0010] In this embodiment, the health state decay factor, the maximum state deviation, and the temperature regulation coefficient are fused to obtain a comprehensive regulation factor, which specifically includes: Multiply the health status decay factor by the temperature regulation coefficient to obtain the basic regulation factor; The maximum state deviation is mapped to the 0 to 1 interval using an S-shaped function and then 1 is added to obtain the normalized divergence factor. Dividing the basic adjustment factor by the normalized divergence factor yields the comprehensive adjustment factor.

[0011] In this embodiment, for a node currently using the first type of model, when the local simulation error exceeds the corresponding error upper limit, switching it to the second type of model for state update from the next simulation step specifically includes: Identify the target node that is currently using the first type of model and whose local simulation error exceeds the corresponding error upper limit; Marking this target node requires enabling the second type of model in the next simulation step; In the state update interface, the model pointer of the target node is redirected from the first type of model to the second type of model, while the current state variable of the node is retained as the initial condition of the second type of model.

[0012] Secondly, this application provides a digital twin-based simulation system for the operational status of an industrial energy storage system, the system comprising: The model building module is used to build digital twin models, in which each battery cell corresponds to a simulation node. The digital twin models include a first type of model and a second type of model. The error calculation module is used to update the state of each simulation node during the time-domain recursive simulation process, calculate the predicted value of the measurable physical quantity, and determine the local simulation error of the simulation node based on the deviation between the predicted value and the measured value. The tolerance adjustment module is used to dynamically adjust the error tolerance corresponding to each simulation node according to the local operating status of each node. The error tolerance includes an upper limit and a lower limit. The switching control module is used to switch a node currently using the first type of model to the second type of model for state update when the local simulation error exceeds the corresponding error limit, starting from the next simulation step. The switching control module is also used to update the state of a node currently using the second type of model when the local simulation error is lower than the corresponding lower error limit and the local simulation errors of all its adjacent nodes do not exceed their respective corresponding upper error limits, starting from the next simulation step. The simulation output module is used to advance to the next simulation step based on the updated node status and model configuration of the current step size, and output the simulation results of the industrial energy storage system's operating status.

[0013] Thirdly, this application provides a computer device, the computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described digital twin-based industrial energy storage system operation state simulation method.

[0014] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described simulation method for the operating status of an industrial energy storage system based on digital twins.

[0015] The technical solutions provided by the embodiments disclosed in this application have the following beneficial effects: The core idea of ​​this application is to first build a two-layer digital twin model containing two types of simulation models: one is a simplified model that is fast in calculation but relatively low in accuracy, and the other is a detailed model that is high in accuracy but has a large computational load. During the simulation, the error tolerance is dynamically adjusted according to the actual state of each battery cell, and the model is switched based on the consistency conditions of adjacent cells, so as to find a balance between the overall simulation accuracy and computational cost. Specifically, a comprehensive adjustment factor is derived based on the cumulative cycle count of a battery cell, its current temperature, and the difference in state of charge (SOC) between it and its neighboring cells. This factor is used to correct the upper and lower limits of the error for each cell in real time. For cells with more severe aging, the error tolerance is dynamically adjusted appropriately to facilitate timely switching to a high-precision model for detailed monitoring. For cells with more abnormal temperatures or poorer SOC consistency, the error tolerance is tightened appropriately, forcing a switch to a high-precision model for detailed monitoring and avoiding frequent switching. When the SOC difference between adjacent cells is too large, the tolerance is tightened, forcing a switch to a high-precision model for detailed monitoring. This allows limited computing resources to be used where they are most needed. During model switching, if the local error of a cell that was previously using a simplified model exceeds the dynamic error limit, it is immediately switched to a detailed model. For cells currently using a detailed model, they are only allowed to switch back to the simplified model when their local error is below the lower limit and the errors of all adjacent cells do not exceed their respective upper limits. This adjacency condition utilizes the close coupling of battery cells in the energy storage system, effectively avoiding repeated switching caused by local disturbances and reducing unnecessary computational burden. Ultimately, by repeatedly performing state updates, error calculations, tolerance adjustments, and model switching, the entire digital twin model can automatically adapt to the real-time changes of the energy storage system. While ensuring the accuracy of the simulation results, it significantly reduces the overall computational load and eliminates model jitter caused by local fluctuations. Attached Figure Description

[0016] Figure 1 This is an exemplary flowchart of a digital twin-based industrial energy storage system operation status simulation method according to some embodiments of this application; Figure 2 This is a flowchart illustrating the determination of local simulation errors according to some embodiments of this application; Figure 3 This is a schematic diagram of the structure of an industrial energy storage system operation status simulation system based on digital twin, according to some embodiments of this application; Figure 4 This is a schematic diagram of the structure of a computer device for implementing a digital twin-based industrial energy storage system operation status simulation method, according to some embodiments of this application. Detailed Implementation

[0017] To better understand the technical solution of this application, the technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] refer to Figure 1 The figure is an exemplary flowchart of a digital twin-based industrial energy storage system operation status simulation method according to some embodiments of this application. The digital twin-based industrial energy storage system operation status simulation method mainly includes the following steps: In step 101, the digital twin model is constructed, wherein each battery cell corresponds to a simulation node.

[0019] It should be noted that the first type of model in this application is a simplified model with low computational overhead, while the second type of model is a detailed model with high simulation accuracy. Specifically: the first type of model can be a single-particle model based on equivalent circuits or a simplified electrothermal coupling model. For example, a first-order RC equivalent circuit can be used to describe the terminal voltage response of a single battery cell, ignoring the internal temperature gradient of the battery and using only a lumped heat capacity to characterize the average temperature. This type of model has fast computation speed and is suitable for real-time simulation of large-scale parallel energy storage systems. The second type of model can be a quasi-two-dimensional model based on electrochemical mechanisms or a thermo-electrochemical-aging multiphysics coupling model. For example, a pseudo-two-dimensional electrochemical model combined with a three-dimensional thermal field distribution can be used to characterize in detail the lithium-ion concentration distribution, potential distribution, and heat generation and dissipation processes in the solid and liquid phases inside the battery. This type of model can accurately reproduce the dynamic behavior of a single battery cell, but the computational load is large. By integrating the above two types of models with different fidelity in the same digital twin model and dynamically switching them according to the local operating state of the simulation nodes, the overall computational burden can be effectively reduced while ensuring the simulation accuracy of key areas.

[0020] In this embodiment, the construction of a digital twin model can be achieved through the following steps: Obtain the battery cell layout diagram and electrical connection diagram of the industrial energy storage system; The spatial location and electrical properties of each battery cell are determined based on the battery cell layout diagram. According to the series and parallel connections specified in the electrical connection diagram, each battery cell is mapped to an independent simulation node, and a battery type label is set for each simulation node; Based on the battery type label and the spatial location, configure the parameters and state update interfaces of the corresponding first type model and second type model for each simulation node to complete the construction of the digital twin model.

[0021] It should be noted that the simulation node in this application is an independently runnable simulation unit used to represent a physical battery cell in the digital twin model and to perform state update, error calculation and model switching logic; the state update interface is a function call specification used to receive the state variable at the previous time step and output the state variable at the current time step.

[0022] In practical applications, firstly, those skilled in the art can export battery cell layout diagrams from the computer-aided design database of industrial energy storage systems and obtain electrical connection diagrams from manufacturing execution systems or electrical design drawings. The battery cell layout diagram records the installation position and relative spacing of each battery cell in two-dimensional or three-dimensional coordinates, while the electrical connection diagram clearly defines the series and parallel branch relationships between battery cells using circuit symbols and wiring methods. Secondly, edge detection and contour extraction algorithms can be used to analyze the battery cell layout diagram, identify the geometric contour of each battery cell, and calculate the horizontal, vertical, and height coordinates of its geometric center point in a preset coordinate system. This is something those skilled in the art cannot easily do without... The implementation, which requires no creative effort, will not be elaborated here. The coordinates of the center point are then used as the spatial location of the battery cell. Simultaneously, the nominal voltage, nominal capacity, internal resistance, and positive / negative polarity corresponding to each battery cell are read from the electrical connection diagram, and these parameters are summarized as the electrical properties of the battery cell. Then, following the series and parallel connection paths specified in the electrical connection diagram, an object-oriented programming technique is used to instantiate a simulation node object for each battery cell. This object contains an independent computation thread and state storage space, and a battery type label in string form is generated based on the chemical system category and specifications of the battery cell. This label is written to the simulation node object in metadata form, and then... Using tags as keywords, a matching query is performed in a pre-built model parameter library to retrieve the first and second types of model parameter sets corresponding to the tags. The first type of model parameter set includes ohmic internal resistance, polarization capacitance, and polarization resistance, while the second type of model parameter set includes solid-phase diffusion coefficient, liquid-phase diffusion coefficient, and reaction rate constant. Then, based on the spatial coordinates of the simulation node, the distance between the node's location and the energy storage system's cooling device or heat source is determined: if the node is adjacent to the cooling device's outlet, the heat exchange coefficient is increased by 20%; if the node is adjacent to the heat source, the heat exchange coefficient is decreased by 15%; otherwise, the heat exchange coefficient remains unchanged. The corrected heat exchange coefficient replaces the original heat exchange coefficient in the first and second type model parameter sets. Then, the corrected first and second type model parameter sets are written into the model parameter area of ​​the simulation node object, respectively. A unified state update interface is configured for the simulation node object. This interface defines the data structure of the input state vector and the data structure of the output state vector. The input state vector includes the current state of charge, the current temperature, and the cumulative number of cycles. The output state vector includes the updated state of charge, the updated temperature, and the predicted terminal voltage. Finally, all simulation node objects that have completed parameter configuration and interface configuration are gathered together as the completed digital twin model.

[0023] It should be noted that all simulation nodes initially use the first type of model to describe their operating state.

[0024] In step 102, during the time-domain recursive simulation, the state of each simulation node is updated, and the predicted value of the measurable physical quantity is calculated. The local simulation error of the simulation node is determined based on the deviation between the predicted value and the measured value.

[0025] It should be noted that, in the time-domain recursive simulation process, updating the state of each simulation node means: according to the preset simulation step time interval, starting from the initial moment, within each step, based on the model type currently used by the simulation node (i.e., the first type simplified model or the second type detailed model) and the input excitation at the current moment, such as the charging and discharging current of a battery cell, calling the state update interface configured for that node, calculating the new state variables at the current moment based on the state variables of the previous moment, namely the state of charge, temperature, and cumulative cycle count, the new state variables include the updated state of charge, the updated temperature, and the predicted terminal voltage, and writing the calculated new state variables into the state storage area of ​​that node to complete the dynamic evolution process of that node from one moment to the next.

[0026] In this embodiment, reference Figure 2 As shown in the figure, this is a flowchart illustrating the determination of local simulation errors according to some embodiments of this application. In this embodiment, the determination of local simulation errors of simulation nodes based on the deviation between predicted and measured values ​​can be achieved through the following steps: In step 1021, the predicted output state vector of the simulation node at the current moment is read from the digital twin model; In step 1022, the measured state vector at the same moment is collected by sensors deployed on the corresponding battery cells; In step 1023, the predicted state vector is compared with the measured state vector component by component, and the absolute deviation of each component is calculated. In step 1024, the absolute deviations of all components are fused to obtain the local simulation error of the simulation node.

[0027] The local simulation error in this application is an indicator used to measure the overall deviation between the simulation output of a single simulation node and the physical measured value at a certain moment.

[0028] In practical application, firstly, the predicted output state vector of the simulation node at the current moment is read from the digital twin model. Specifically, at the end of each simulation step, each simulation node in the digital twin model writes the predicted terminal voltage, surface temperature, and state of charge generated by this state update into the output buffer of that node. Then, by calling the data reading interface of the digital twin model, using the unique identifier of the simulation node as an index, the above three predicted values ​​are read sequentially from the output buffer in a fixed order of terminal voltage, surface temperature, and state of charge, and organized into a one-dimensional array, which is the predicted output state vector. Secondly, it is deployed on the corresponding battery... Sensors on each cell acquire measured state vectors at the same time. Specifically, in a physical industrial energy storage system operating synchronously with a digital twin model, each battery cell surface is equipped with a voltage sensor, a temperature sensor, and a current-integral-based state-of-charge (SOC) monitoring module. All sensors can be triggered simultaneously via an industrial Ethernet bus using a precise timing protocol to obtain the measured terminal voltage, surface temperature, and SOC values ​​for that battery cell. These values ​​are then organized into a one-dimensional array in the same order as the predicted output state vector; this one-dimensional array is the measured state vector. Next, the predicted output state vector is compared component by component with the measured state vector, and the absolute deviation of each component is calculated. Specifically, the... The absolute deviation of the terminal voltage is obtained by subtracting the predicted and measured terminal voltage values. Similarly, the absolute deviation of the surface temperature is obtained by subtracting the predicted and measured surface temperature values. The absolute deviation of the state of charge (SOC) is obtained by subtracting the predicted and measured SOC values. Finally, the absolute deviations of all components are fused to obtain the local simulation error of the simulation node. Specifically, before fusion, the absolute deviations of each component need to be normalized. This can be done by obtaining the maximum allowable fluctuation range of the terminal voltage, surface temperature, and SOC under normal operating conditions as their respective normalization reference values. The maximum allowable fluctuation range of the terminal voltage is... The range is 10% of the rated voltage of the battery cell, the maximum allowable fluctuation range of surface temperature is 20 degrees Celsius, and the maximum allowable fluctuation range of state of charge is 100%. The dimensionless normalized deviation of the terminal voltage is obtained by dividing the absolute deviation of the terminal voltage by the normalized reference value of the terminal voltage; the dimensionless normalized deviation of the surface temperature is obtained by dividing the absolute deviation of the surface temperature by the normalized reference value of the surface temperature; and the dimensionless normalized deviation of the state of charge is obtained by dividing the absolute deviation of the state of charge by the normalized reference value of the state of charge. Furthermore, in this embodiment, a weighted fusion method is used, setting the weight of the absolute deviation of the terminal voltage to 0.4, the weight of the absolute deviation of the surface temperature to 0.3, and the weight of the absolute deviation of the state of charge to 0.3. Calculate the product of each absolute deviation and its weight, then sum the products. Use the sum as the local simulation error of that simulation node at the current moment. This local simulation error is used to determine whether to switch the model type used by that simulation node.

[0029] In step 103, the error tolerance corresponding to each simulation node is dynamically adjusted according to the local operating state of each simulation node. The error tolerance includes an upper limit and a lower limit.

[0030] In this embodiment, dynamically adjusting the error tolerance corresponding to each simulation node based on its local operating state specifically includes: Obtain the current state of charge, temperature, and cumulative number of cycles for the simulation node; The health status decay factor is determined based on the cumulative number of cycles. Traverse all adjacent simulation nodes, calculate the absolute value of the difference in the state of charge between the node and each of its adjacent nodes, and take the maximum value as the maximum state deviation. The temperature adjustment coefficient is determined based on the deviation between the current temperature and the rated operating temperature; The health state decay factor, the maximum state deviation, and the temperature regulation coefficient are fused to obtain a comprehensive regulation factor; The error tolerance of the current step size is obtained by multiplying the preset upper and lower limits of the reference error by the comprehensive adjustment factor. The error tolerance includes the upper and lower limits of the error.

[0031] In this application, the cumulative cycle count refers to the total number of complete charge-discharge cycles completed by a single battery cell from the start of its use to the present; the health state decay factor is an indicator that measures the amplification requirements of the simulation error tolerance based on the degree of battery cycle aging; the maximum state deviation is an indicator that measures the degree of inconsistency in the state of charge within the battery pack; the temperature adjustment coefficient is an indicator that measures the impact of the battery's thermal state deviation from the safe range on the simulation accuracy requirements; and the comprehensive adjustment factor is a fusion coefficient used to simultaneously reflect the degree of battery aging, the degree of inconsistency in the state of charge, and the degree of thermal state deviation, and to uniformly adjust the upper and lower limits of the error tolerance accordingly.

[0032] In practical applications, the current state of charge, temperature, and cumulative number of cycles of the simulation node are obtained in the following way: In the state storage area of ​​each simulation node in the digital twin model, the state of charge, temperature, and cumulative number of cycles are maintained. At the start of the current simulation step, the current values ​​of the above three variables are directly read through the data reading interface of the node, and the read floating-point values ​​are used as the current state of charge, current temperature, and cumulative number of cycles completed of the simulation node, respectively.

[0033] In practical applications, the health status decay factor is determined based on the cumulative number of cycles in the following manner: The mapping relationship between the health status decay factor and the cumulative number of cycles is pre-calibrated through battery cycle aging experiments. This mapping relationship is expressed using a piecewise linear function: when the cumulative number of cycles is less than 500, the health status decay factor is 1.0; when the cumulative number of cycles is greater than or equal to 500 and less than 1500, the health status decay factor linearly decreases from 1.0 to 0.7; when the cumulative number of cycles is greater than or equal to 1500, the health status decay factor is 0.5. The obtained cumulative number of cycles is then substituted into the above piecewise linear function for calculation, and the calculated value is used as the current health status decay factor for the simulation node.

[0034] In practical applications, all adjacent simulation nodes are traversed, and the absolute value of the difference between the state of charge (SOC) of the current simulation node and each of its adjacent nodes is calculated. The maximum value is taken as the maximum SOC deviation. This is achieved in the following way: Based on the electrical connection topology established during the construction of the digital twin model, a list of adjacent nodes for each simulation node is pre-stored. An adjacent node refers to another simulation node that is directly connected in series or parallel with the current simulation node in terms of electrical connection. Each node in the list of adjacent nodes is traversed, and the current SOC of the current simulation node is read. The SOC of the adjacent node is subtracted from the current SOC of the simulation node, and the absolute value is taken to obtain the SOC of the adjacent node. The absolute value of the state of charge difference is used to traverse all values. The maximum value is then identified as the maximum state deviation. It should be further explained that the advantage of considering adjacent simulation nodes in calculating the maximum state deviation in this application is that in industrial energy storage systems, there is a direct electrical coupling between battery cells connected in series or parallel. A deviation in the state of charge of any cell from its adjacent cells will disrupt the overall consistency of the battery pack, leading to increased electrochemical polarization differences, increased local internal resistance, and unbalanced charging and discharging, potentially resulting in safety risks such as overcharging, over-discharging, or thermal runaway. By introducing the maximum value of the state of charge difference between adjacent nodes as the basis for adjusting the error tolerance, the weakest cell node in the battery pack can be identified first. The error tolerance for this node can be proactively reduced, i.e., the upper limit of the error is narrowed, making it easier for the local simulation error to exceed the narrowed upper limit, thereby triggering a switch to the second type of high-precision model. This mechanism allows computational resources to adaptively focus on the highest-risk area, achieving early capture and refined simulation of potentially faulty cells without significantly increasing the overall computational burden, effectively avoiding system-level simulation distortion caused by the gradual propagation of anomalies from adjacent nodes.

[0035] In practical applications, the temperature adjustment coefficient is determined based on the deviation between the current temperature and the rated operating temperature using the following method: The rated operating temperature of the battery cell, specified in the industrial energy storage system design parameters (25 degrees Celsius), is read. The absolute value of the difference between the current temperature and 25 degrees Celsius is calculated to obtain the absolute value of the temperature deviation. Then, a piecewise constant function is used to determine the temperature adjustment coefficient: if the absolute value of the temperature deviation is less than or equal to 5 degrees Celsius, the temperature adjustment coefficient is 1.0; if the absolute value of the temperature deviation is greater than 5 degrees Celsius and less than or equal to 15 degrees Celsius, the temperature adjustment coefficient is 0.9; if the absolute value of the temperature deviation is greater than 15 degrees Celsius, the temperature adjustment coefficient is 0.7. The value determined according to the above rules is used as the temperature adjustment coefficient.

[0036] In practical applications, the error tolerance of the current step size can be obtained by multiplying the preset upper and lower limits of the reference error by the comprehensive adjustment factor, respectively. This can be achieved as follows: Before the simulation begins, a set of fixed values ​​is preset according to the accuracy requirements of the terminal voltage measurement of the industrial energy storage system. The upper limit of the reference error is set to 0.05 (dimensionless), and the lower limit of the reference error is set to 0.01 (dimensionless). The upper limit of the reference error is multiplied by the comprehensive adjustment factor, and the product is used as the upper limit of the current step size; the lower limit of the reference error is multiplied by the comprehensive adjustment factor, and the product is used as the lower limit of the current step size. The above upper and lower limits of the error are used as the error tolerance of the current simulation step size to determine whether the model type should be switched at this node.

[0037] Preferably, in some embodiments, the health state decay factor, the maximum state deviation, and the temperature regulation coefficient are fused to obtain a comprehensive regulation factor, which can be achieved by the following steps: Multiply the health status decay factor by the temperature regulation coefficient to obtain the basic regulation factor; The maximum state deviation is mapped to the 0 to 1 interval using an S-shaped function and then 1 is added to obtain the normalized divergence factor. Dividing the basic adjustment factor by the normalized divergence factor yields the comprehensive adjustment factor.

[0038] In practical applications, firstly, the health state degradation factor is multiplied by the temperature adjustment coefficient. This involves directly multiplying the dimensionless coefficient describing the degree of battery aging with the dimensionless coefficient describing the impact of thermal state deviation to obtain the basic adjustment factor. This multiplication comprehensively reflects the amplification requirements of aging and abnormal high or low temperature conditions on error tolerance. Secondly, the maximum state deviation is normalized. Specifically, the maximum state deviation can be used as an independent variable input into the logistic function formula. This formula is equal to 1 divided by 1 plus the natural constant raised to the power of the negative maximum state deviation. The calculated result lies in the open interval between 0 and 1. Then, 1 is added to this result to obtain a normalized value between 1 and 2. The bifurcation factor ensures that even when the maximum state deviation reaches an extreme value, the normalized bifurcation factor will not exhibit drastic changes. Finally, the basic adjustment factor is divided by the normalized bifurcation factor, and the quotient is used as the comprehensive adjustment factor. Through the above division operation, when the maximum state deviation is small, the normalized bifurcation factor approaches 1, and the comprehensive adjustment factor is mainly determined by the health state decay factor and the temperature adjustment coefficient, with the error tolerance adjusted accordingly. When the maximum state deviation is large, the normalized bifurcation factor is greater than 1, and the comprehensive adjustment factor is reduced, thereby actively reducing the error tolerance and making it easier for local simulation errors to exceed the error upper limit, triggering a switch to the second type of high-precision model.

[0039] In step 104, for a node currently using the first type of model, when the local simulation error exceeds the corresponding error upper limit, it is switched to the second type of model for state update starting from the next simulation step.

[0040] In this embodiment, for a node currently using the first type of model, when the local simulation error exceeds the corresponding error upper limit, switching it to the second type of model for state update from the next simulation step can be achieved using the following steps: Identify the target node that is currently using the first type of model and whose local simulation error exceeds the corresponding error upper limit; Marking this target node requires enabling the second type of model in the next simulation step; In the state update interface, the model pointer of the target node is redirected from the first type of model to the second type of model, while the current state variable of the node is retained as the initial condition of the second type of model.

[0041] In practical applications, at the end of the current simulation step, all simulation nodes in the digital twin model are traversed, and target nodes whose current model type is identified as the first type and whose local simulation error is greater than the upper limit of the current step error are selected. The upper limit of the error is determined in real time by the aforementioned dynamic adjustment step based on the comprehensive adjustment factor multiplied by the benchmark error upper limit. For each target node that meets the conditions, a Boolean model switching flag is set inside the node and set to true to indicate that the node needs to use the second type of model for state update in the next simulation step. During the initialization phase before the start of the next simulation step, the model switching flag of each simulation node is checked in turn. For nodes with a true flag, the pointer of the model object stored inside the interface is changed from pointing to the address of the first type of model instance to pointing to the address of the second type of model instance pre-loaded in memory by calling its state update interface. This process can be implemented using the pointer redirection technique or strategy pattern in object-oriented programming. At the same time, the state variables such as the state of charge, temperature and cumulative number of cycles obtained by the node at the end of the current simulation step are directly copied and assigned to the initial state buffer of the second type of model as the initial condition for the second type of model to start the simulation calculation. After completing the above model pointer redirection and initial condition assignment, the model switching flag is restored to false. The node will then automatically use the second type of model to perform subsequent state update calculations from the current simulation step without re-instantiating the node or reloading the model parameters.

[0042] In step 105, for a node currently using the second type of model, when the local simulation error is lower than the corresponding lower error limit and the local simulation errors of all its adjacent nodes do not exceed their respective corresponding upper error limits, it is restored to the first type of model for state update from the next simulation step.

[0043] In this embodiment, for a node currently using the second type of model, when the local simulation error is lower than the corresponding lower error limit, and the local simulation errors of all its adjacent nodes do not exceed their respective corresponding upper error limits, the following steps can be used to restore it to the first type of model for state update from the next simulation step: Identify candidate nodes that are currently using the second type of model and whose local simulation error is lower than the corresponding error lower limit; Traverse all neighboring nodes of the candidate node and determine whether the local simulation error of each neighboring node does not exceed its corresponding error upper limit; If all adjacent nodes meet the conditions, then mark the candidate node as being restored to the first type model in the next simulation step; In the state update interface, the model pointer of the candidate node is redirected from the second type of model to the first type of model, and the running parameters of the first type of model are reinitialized based on the current state.

[0044] In practical application, at the end of the current simulation step, all simulation nodes in the digital twin model identified as the second type of model are traversed. Nodes whose local simulation errors are less than the lower limit of the current step error are marked as candidate nodes. This lower limit is determined in real time by the aforementioned dynamic adjustment step based on the product of the comprehensive adjustment factor and the baseline error lower limit. For each candidate node, the model type identifier and the local simulation error of each adjacent node at its current step are read sequentially from the list of adjacent nodes in the electrical connection topology pre-stored during the construction of the digital twin model. The local simulation error of each adjacent node is compared with the upper limit of its own current step error. This upper limit is also determined in real time by the dynamic adjustment step based on the product of the comprehensive adjustment factor and the upper limit of the respective baseline error of each adjacent node. Only when the local simulation errors of all adjacent nodes of the candidate node do not exceed their respective upper limits of current error is a Boolean model recovery flag set inside the candidate node and the flag is set to true. During the initialization phase before the next simulation step, the model recovery flag of each simulation node is checked sequentially. For nodes with a true flag, the pointer to the model object stored inside the interface is changed from pointing to the address of the second type of model instance to pointing to the address of the first type of model instance by calling its state update interface. This pointer redirection process is implemented using the well-known direct assignment in object-oriented programming or the strategy substitution technique in the strategy pattern. At the same time, the state variables such as the state of charge, temperature and cumulative number of cycles saved by the node at the end of the current step are retrieved. Since the first type of model is a simplified model, its required set of state variables is a subset of the second type of model. Therefore, only the state variables common to both are selected and assigned to the corresponding initial state parameters of the first type of model, while the detailed state variables unique to the second type of model are directly discarded. After completing the pointer redirection and initial condition assignment, the model recovery flag is cleared, and the node will automatically use the first type of model for state update from the next simulation step.

[0045] In step 106, based on the updated node status and model configuration of the current step size, the next simulation step size is recursively calculated, and the simulation results of the industrial energy storage system's operating status are output.

[0046] In this embodiment, based on the updated node states and model configurations at the current step size, proceeding to the next simulation step size and outputting the simulation results of the industrial energy storage system's operating status can be achieved through the following steps: Save the updated state variables of the current step size simulation node to the state space of the digital twin model; Mark the model type used by each node in the current step as the initial configuration for the next simulation step; Advance the time pointer by one simulation step, and repeat the steps of state update, error calculation, error tolerance adjustment and model switching until the preset simulation termination time is reached; After each simulation step, the state information of all nodes is collected to generate and output the simulation results of the industrial energy storage system's operating status.

[0047] In practical applications, after all nodes in the current simulation step have completed state updates, error calculations, error tolerance adjustments, and model switching, the state variables such as the predicted state of charge, temperature, and terminal voltage obtained from each simulation node are written into the state space of the digital twin model in the form of a structure array. This state space is implemented using a shared memory area in the computer's memory for easy retrieval in subsequent simulation steps. Simultaneously, the actual model type identifier used by each node in the current step is recorded (i.e., first-type model or second-type model), and this identifier is stored in the node's configuration storage area as the initial model configuration for the next simulation step. Then, the simulation... The true time pointer advances forward by one simulation step, which adds a preset fixed time interval to the current simulation time. This time interval is pre-calibrated based on the dynamic response speed of the industrial energy storage system, typically ranging from 0.1 to 1 second. Further, it checks whether the updated simulation time has reached a preset simulation termination time. This termination time is set according to the operational status monitoring requirements, such as 24 hours or the duration of a complete charge-discharge cycle. If the termination time has not been reached, the process returns to the steps of state update, error calculation, error tolerance adjustment, and model switching, forming a recursive loop. If the termination time has been reached, the recursion stops. After each simulation step, the current state information of all simulation nodes is collected from the state space of the digital twin model, including the state of charge, temperature, predicted terminal voltage, and the currently used model type for each node. These are then arranged chronologically and by node number to generate the operational status simulation results of the industrial energy storage system. These results are output to a display terminal or stored in a database in the form of a data table or time-series curve. Those skilled in the art can directly implement this process by combining the above steps with conventional time-domain recursive simulation procedures and data processing methods.

[0048] On the other hand, in some embodiments, this application provides an industrial energy storage system operation status simulation system based on digital twins, with reference to... Figure 3 The figure is a schematic diagram of the structure of a digital twin-based industrial energy storage system operation status simulation system according to some embodiments of this application. The digital twin-based industrial energy storage system operation status simulation system includes: a model building module 301, an error calculation module 302, a tolerance adjustment module 303, a switching control module 304, and a simulation output module 305, which are described below: The model building module 301 in this application is used to build a digital twin model, wherein each battery cell corresponds to a simulation node, and the digital twin model includes a first type of model and a second type of model; The error calculation module 302 in this application is used to update the state of each simulation node during the time-domain recursive simulation process, calculate the predicted value of the measurable physical quantity, and determine the local simulation error of the simulation node based on the deviation between the predicted value and the measured value. The tolerance adjustment module 303 in this application is used to dynamically adjust the error tolerance corresponding to each simulation node according to the local operating state of each node. The error tolerance includes an upper limit and a lower limit. The switching control module 304 in this application is used to switch a node currently using the first type of model to the second type of model for state update when the local simulation error exceeds the corresponding error upper limit, starting from the next simulation step. The switching control module 304 is also used to update the state of a node currently using the second type of model when the local simulation error is lower than the corresponding lower limit of error and the local simulation errors of all its adjacent nodes do not exceed their respective corresponding upper limits of error, and restore it to the first type of model from the next simulation step. The simulation output module 305 in this application is used to advance to the next simulation step based on the updated node status and model configuration of the current step size, and output the simulation results of the operating status of the industrial energy storage system.

[0049] In addition, this application also provides a computer device, which includes a memory and a processor. The memory stores code, and the processor is configured to acquire the code and execute the above-described digital twin-based industrial energy storage system operation state simulation method.

[0050] In some embodiments, reference Figure 4 The figure is a schematic diagram of the structure of a computer device for implementing a digital twin-based industrial energy storage system operation status simulation method according to some embodiments of this application. The digital twin-based industrial energy storage system operation status simulation method in the above embodiments can be implemented through... Figure 4 The computer device shown is used to implement this, and the computer device 400 includes at least one processor 401, a communication bus 402, a memory 403, and at least one communication interface 404.

[0051] Processor 401 can be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0052] The communication bus 402 can be used to transmit information between the aforementioned components.

[0053] The memory 403 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disks or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 403 may exist independently and be connected to the processor 401 via the communication bus 402. The memory 403 may also be integrated with the processor 401.

[0054] The memory 403 stores program code for executing the scheme of this application, and its execution is controlled by the processor 401. The processor 401 executes the program code stored in the memory 403. The program code may include one or more software modules. In the above embodiments, the industrial energy storage system operation status simulation method based on digital twins can be implemented by the processor 401 and one or more software modules in the program code in the memory 403.

[0055] Communication interface 404 uses any transceiver-like device to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area networks (WLAN), etc.

[0056] In a specific implementation, as one example, a computer device may include multiple processors, each of which may be a single-core (single-CPU) processor or a multi-core (multi-CPU) processor. Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0057] The aforementioned computer device can be a general-purpose computer device or a special-purpose computer device. In specific implementations, the computer device can be a desktop computer, a portable computer, a network server, a handheld digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. This application does not limit the type of computer device.

[0058] In addition, this application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described simulation method for the operating status of an industrial energy storage system based on digital twins.

[0059] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0060] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A simulation method for the operational status of an industrial energy storage system based on digital twins, applied to a digital twin model of an industrial energy storage system, wherein the digital twin model includes a first type of model and a second type of model, the simulation accuracy of the second type of model being higher than that of the first type of model, characterized in that, Includes the following steps: Construct the digital twin model, in which each battery cell corresponds to one simulation node; During the time-domain recursive simulation, the state of each simulation node is updated, and the predicted value of the measurable physical quantity is calculated. The local simulation error of the simulation node is determined based on the deviation between the predicted value and the measured value. Based on the local operating status of each simulation node, the error tolerance corresponding to that node is dynamically adjusted, and the error tolerance includes an upper limit and a lower limit. For nodes currently using the first type of model, when the local simulation error exceeds the corresponding error upper limit, the nodes are switched to the second type of model for state update starting from the next simulation step. For a node currently using the second type of model, when the local simulation error is lower than the corresponding lower limit of error and the local simulation errors of all its adjacent nodes do not exceed their respective upper limits of error, it is restored to the first type of model for state update from the next simulation step. Based on the updated node status and model configuration at the current step size, proceed to the next simulation step size iteration and output the simulation results of the industrial energy storage system's operating status.

2. The method according to claim 1, characterized in that, All simulation nodes initially use the first type of model to describe their operating state.

3. The method as described in claim 1, characterized in that, The construction of a digital twin model specifically includes: Obtain the battery cell layout diagram and electrical connection diagram of the industrial energy storage system; The spatial location and electrical properties of each battery cell are determined based on the battery cell layout diagram. According to the series and parallel connections specified in the electrical connection diagram, each battery cell is mapped to an independent simulation node, and a battery type label is set for each simulation node; Based on the battery type label and the spatial location, configure the parameters and state update interfaces of the corresponding first type model and second type model for each simulation node to complete the construction of the digital twin model.

4. The method as described in claim 1, characterized in that, The local simulation error of the simulation node is determined based on the deviation between the predicted and measured values, specifically including: Read the predicted output state vector of the simulation node at the current moment from the digital twin model; The measured state vector at the same moment is collected by sensors deployed on the corresponding battery cells; The predicted state vector is compared with the measured state vector component by component, and the absolute deviation of each component is calculated. The absolute deviations of all components are fused to obtain the local simulation error of the simulation node.

5. The method as described in claim 1, characterized in that, Based on the local operating status of each simulation node, the error tolerance corresponding to that node is dynamically adjusted, specifically including: Obtain the current state of charge, temperature, and cumulative number of cycles for the simulation node; The health status decay factor is determined based on the cumulative number of cycles. Traverse all adjacent simulation nodes, calculate the absolute value of the difference in the state of charge between the node and each of its adjacent nodes, and take the maximum value as the maximum state deviation. The temperature adjustment coefficient is determined based on the deviation between the current temperature and the rated operating temperature; The health state decay factor, the maximum state deviation, and the temperature regulation coefficient are fused to obtain a comprehensive regulation factor; The error tolerance of the current step size is obtained by multiplying the preset upper and lower limits of the reference error by the comprehensive adjustment factor. The error tolerance includes the upper and lower limits of the error.

6. The method as described in claim 5, characterized in that, The comprehensive regulation factor obtained by fusing the health state decay factor, the maximum state deviation, and the temperature regulation coefficient specifically includes: Multiply the health status decay factor by the temperature regulation coefficient to obtain the basic regulation factor; The maximum state deviation is mapped to the 0 to 1 interval using an S-shaped function and then 1 is added to obtain the normalized divergence factor. Dividing the basic adjustment factor by the normalized divergence factor yields the comprehensive adjustment factor.

7. The method as described in claim 1, characterized in that, For nodes currently using the first type of model, when the local simulation error exceeds the corresponding error upper limit, switching to the second type of model for state update from the next simulation step specifically includes: Identify the target node that is currently using the first type of model and whose local simulation error exceeds the corresponding error upper limit; Marking this target node requires enabling the second type of model in the next simulation step; In the state update interface, the model pointer of the target node is redirected from the first type of model to the second type of model, while the current state variable of the node is retained as the initial condition of the second type of model.

8. A simulation system for the operational status of an industrial energy storage system based on digital twins, wherein the operational status simulation is performed using the method described in any one of claims 1 to 7, characterized in that, The system includes: The model building module is used to build digital twin models, in which each battery cell corresponds to a simulation node. The digital twin models include a first type of model and a second type of model. The error calculation module is used to update the state of each simulation node during the time-domain recursive simulation process, calculate the predicted value of the measurable physical quantity, and determine the local simulation error of the simulation node based on the deviation between the predicted value and the measured value. The tolerance adjustment module is used to dynamically adjust the error tolerance corresponding to each simulation node according to the local operating status of each node. The error tolerance includes an upper limit and a lower limit. The switching control module is used to switch a node currently using the first type of model to the second type of model for state update when the local simulation error exceeds the corresponding error limit, starting from the next simulation step. The switching control module is also used to update the state of a node currently using the second type of model when the local simulation error is lower than the corresponding lower error limit and the local simulation errors of all its adjacent nodes do not exceed their respective corresponding upper error limits, by restoring it to the first type of model from the next simulation step. The simulation output module is used to advance to the next simulation step based on the updated node status and model configuration of the current step size, and output the simulation results of the industrial energy storage system's operating status.

9. A computer device comprising a memory and a processor, the memory storing code, characterized in that, The processor is configured to acquire the code and execute the digital twin-based industrial energy storage system operation state simulation method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the digital twin-based industrial energy storage system operation status simulation method as described in any one of claims 1 to 7.

Images