Real-time dynamic thermal management method and system for energy storage station battery pack

Abstract

The application discloses a kind of real-time dynamic thermal management methods of energy storage station battery pack, including obtaining the parameter data information of target energy storage station battery pack;Select single battery pack as basic unit and establish battery pack simulation model;Select normal working condition and accident working condition and carry out numerical simulation simulation;Data sparsification is handled and data regression model is trained;Real-time acquisition target energy storage station battery pack's working data information, the data information obtained is reconstructed to realize the dynamic visual expression of battery pack;According to working data information and data regression model, the real-time dynamic thermal management of target energy storage station battery pack is realized.The application also discloses a kind of system for realizing the real-time dynamic thermal management method of energy storage station battery pack.The reliability of the application is high, good and high efficiency.

Description

Technical Field

[0001] This invention belongs to the field of electrical automation, and specifically relates to a real-time dynamic thermal management method and system for battery packs in energy storage stations. Background Technology

[0002] With economic and technological development and the improvement of people's living standards, electricity has become an indispensable secondary energy source in people's production and daily life, bringing endless convenience. Therefore, ensuring the stable and reliable operation of electricity has become one of the most important tasks of the power system.

[0003] Currently, an increasing number of new energy power generation systems are being connected to the grid. The randomness of the output from these systems puts enormous pressure on the safe and stable operation of the power system. To address this change, power systems have begun deploying energy storage stations in recent years. These stations house numerous battery banks; by deploying additional energy storage stations, the power system can provide itself with additional adjustable loads, achieving peak shaving and valley filling functions, thereby significantly reducing the operational pressure on the power system itself. Therefore, dynamic thermal management of the battery banks in energy storage stations is of great significance to the power system.

[0004] Currently, traditional dynamic thermal management solutions for energy storage stations typically achieve dynamic thermal management through numerical simulation of key parameters within the battery pack. However, these numerical simulation methods are generally extremely complex, with very high computational complexity, and cannot achieve real-time thermal management. Furthermore, these numerical simulation methods also suffer from low reliability and poor accuracy. Summary of the Invention

[0005] One of the objectives of this invention is to provide a real-time dynamic thermal management method for energy storage station battery packs that is highly reliable, accurate, and efficient.

[0006] The second objective of this invention is to provide a system for implementing the real-time dynamic thermal management method for the battery pack of the energy storage station.

[0007] The real-time dynamic thermal management method for battery packs in energy storage stations provided by this invention includes the following steps:

[0008] S1. Obtain parameter data information of the target energy storage station's battery pack;

[0009] S2. Select a single battery pack as the basic unit, and establish a battery pack simulation model based on the data information obtained in step S1;

[0010] S3. Based on the battery pack simulation model obtained in step S2, select normal operating conditions and accident operating conditions, and perform corresponding numerical simulations.

[0011] S4. Based on the simulation results obtained in step S3, perform data sparsification processing and train the data regression model;

[0012] S5. Acquire the working data information of the target energy storage station battery pack in real time, and reconstruct the data based on the data information obtained in step S4 to realize the dynamic visualization of the battery pack;

[0013] S6. Based on the working data information obtained in step S5 and the data regression model obtained in step S4, realize the real-time dynamic thermal management of the target energy storage station battery pack.

[0014] Step S2 involves selecting a single battery pack as the basic unit and establishing a battery pack simulation model based on the data obtained in step S1. This specifically includes the following steps:

[0015] A single battery pack was selected as the basic unit.

[0016] Based on the data obtained in step S1, the geometric and physical models of the battery pack are established based on the cooling method, internal structure and physical properties of the resistor.

[0017] The geometric and physical models of the battery pack are simplified.

[0018] The reliability of the simplified model was verified through battery pack experimental results and grid independence assessment.

[0019] Step S3, based on the battery pack simulation model obtained in step S2, selects normal operating conditions and accident operating conditions, and performs corresponding numerical simulations, specifically including the following steps:

[0020] Based on the battery pack simulation model obtained in step S2, normal operating conditions and accident operating conditions are selected.

[0021] Based on the selected normal and accident operating conditions, determine the physical property parameters, heat generation of the heat source, boundary conditions, and initial conditions of the simulation model;

[0022] Before the experiment, under the experimental conditions, set up as many measurement points as possible for the battery pack;

[0023] Experiments were conducted to compare the detection parameters obtained through measurement points with the simulation results obtained from the simulation model, thereby correcting the simulation model.

[0024] Based on the internal characteristics of the battery pack, the battery pack is divided into zones;

[0025] Post-process the simulation results of each partition in the simulation model to obtain the mean and upper and lower limits of the physical quantities of each partition under the corresponding working conditions;

[0026] Using battery pack input parameters and time as independent variables, and the mean and upper and lower limits of physical quantities as dependent variables, regression analysis was performed using a regression model, and the obtained regression model was used to construct a standardized model library.

[0027] Step S4, which involves performing data sparsity processing based on the simulation results obtained in step S3 and training a data regression model, specifically includes the following steps:

[0028] The simulation results obtained in step S3 are then processed to be dimensionless.

[0029] Based on the sparsification algorithm, dimensionless data is sparsified to obtain a sparse matrix;

[0030] Using the input parameters of the battery pack as independent variables and the parameters at the non-zero sparse corresponding positions on the sparse matrix as dependent variables, a regression method is used to obtain the mapping model between the independent and dependent variables, thereby constructing a visual database.

[0031] Based on the obtained sparse matrix, monitoring points are arranged within a set range on the inner wall of the battery pack surface.

[0032] Based on the simulation results of numerical simulation, a regression model between the monitoring point data and the battery pack input variables is obtained by using the regression method, thereby constructing a management model library.

[0033] The aforementioned sparsification algorithm performs sparsification processing on dimensionless data to obtain a sparse matrix, specifically including the following steps:

[0034] Let the data signal be x, x∈R n R n A set of vectors representing all n-dimensional (vector data signals); the dictionary is D, D = [d1, d2, ..., dn]. L ]∈R n×L d L Let x be the Lth atom in the dictionary, and n be an n-dimensional vector (data signal); represent the data signal as x∈Dα, where α is a sparse matrix and α=[α1,α2,...,α L ] T ;

[0035] The sparse matrix α is calculated using the following steps:

[0036] A. Input the data signal x and the overcomplete dictionary D0, and input the sparsity K or the residual threshold ε;

[0037] B. Set initial values: The residual r0 is x, the index sets I0 and J are both empty sets; the number of iterations k is 0;

[0038] C. Calculate α k =D Tr k-1 and from the obtained α k Select the K largest elements; record the index of each element in set J. k Then, sort the indices in descending order of element value;

[0039] D. Find the subset And J0 satisfies: |α(i)|≤2|α(j)|, j∈J0, i≤j;

[0040] E. Merge the found subset J0 into the index set I. k-1 ;

[0041] F. Update the sparsity coefficient α k The value is Where |||| is the value of the determinant of the matrix. To use a sparse decomposition algorithm to find the decomposition coefficients of the data; R' is the coefficients that are only used in set I. k The coefficient vector containing all non-zero elements at index positions;

[0042] G. Update residual r k For r k =xD k α k ;

[0043] H. Increment the iteration count by 1 and make a judgment: if the iteration count reaches the set value, output the final sparse matrix; otherwise, return to step C to perform the next round of calculation.

[0044] Step S5 involves acquiring real-time operating data of the target energy storage station's battery pack and reconstructing the data based on the data obtained in step S4 to achieve dynamic visualization of the battery pack. Specifically, this includes the following steps:

[0045] Real-time acquisition of operational data information of the target energy storage station's battery pack;

[0046] Based on the parameters input from the battery pack in real time, the corresponding mapping model is obtained from the visualization database;

[0047] Data reconstruction is achieved using a fast signal recovery scheme;

[0048] Based on the parameters input from the battery pack in real time, the corresponding regression model and the mean and upper and lower limits of the physical quantities are obtained from the standardized model library;

[0049] Based on the mean and upper and lower limits of the acquired physical quantities, the reconstructed data is dimensionalized to achieve real-time dynamic visualization of the physical quantities.

[0050] The aforementioned data reconstruction using a fast signal recovery scheme specifically includes the following steps:

[0051] a. Determine the sparse matrix φ n×n To ensure the n-dimensional signal X n+1 The coefficient signal can be expressed in a sparse matrix coordinate basis.

[0052] b. Based on the random sampling matrix ψ used when compressing the signal. m×n The problem of compressed signal recovery is reduced to an optimization problem;

[0053] c. Solve using singular value decomposition to obtain... general solution form Substituting the objective function, the constrained optimization problem is transformed into an unconstrained optimization problem, expressed as:

[0054]

[0055] In the formula To find the minimum value of the objective function f(); f() is the expression for optimizing the objective function; The solution to the equation after singular value decomposition of the constraints;

[0056] d. Construct intermediate variable r as Where δ is a set small perturbation amount. To represent the degree of function f;

[0057] If r ≥ 0, then set the search direction d. k for If r < 0, then set the search direction d. k for Where I is the identity matrix;

[0058] e. Using the search direction from step d, solve for the objective function to obtain...

[0059] f. Using formulas The final recovery signal was obtained.

[0060] Step S6, which involves using the working data information obtained in step S5 and the data regression model obtained in step S4 to achieve real-time dynamic thermal management of the target energy storage station's battery pack, specifically includes the following steps:

[0061] Based on the work data information obtained in step S5, retrieve the regression model corresponding to the monitoring point from the management model library, and visualize the changing trends of the monitoring data and regression data.

[0062] Real-time calculation of the difference between monitoring data and regression data at each measuring point, and visualization of the trend of the difference over time;

[0063] When the difference exceeds the set threshold, corresponding fault diagnosis is performed to complete the real-time dynamic thermal management of the target energy storage station's battery pack.

[0064] This invention also provides a system for implementing the real-time dynamic thermal management method for the battery pack of the energy storage station, including a data acquisition module, a model building module, a numerical simulation module, a regression construction module, a data reconstruction module, and a dynamic thermal management module; the data acquisition module, model building module, numerical simulation module, regression construction module, data reconstruction module, and dynamic thermal management module are connected in series; the data acquisition module is used to acquire parameter data information of the target energy storage station battery pack and upload the data information to the model building module; the model building module is used to select a single battery pack as the basic unit based on the received data information, and build a battery pack simulation model based on the acquired data information, and upload the data information to the numerical simulation module; the numerical simulation module is used to perform real-time dynamic thermal management based on the received data information and the obtained data. The battery pack simulation model selects normal and accident operating conditions and performs corresponding numerical simulations, then uploads the data to the regression construction module. The regression construction module, based on the received data and simulation results, performs data sparsification processing, trains the data regression model, and uploads the data to the data reconstruction module. The data reconstruction module, based on the received data, acquires real-time operating data of the target energy storage station's battery pack, reconstructs the data to achieve dynamic visualization of the battery pack, and uploads the data to the dynamic thermal management module. The dynamic thermal management module, based on the received data, the acquired operating data, and the obtained data regression model, implements real-time dynamic thermal management of the target energy storage station's battery pack.

[0065] The real-time dynamic thermal management method and system for energy storage station battery packs provided by this invention advances the simulation process, obtains a mathematical model through regression, and directly utilizes the obtained mathematical model through data sparsification and reconstruction schemes to achieve real-time, efficient, and dynamic visualization of important physical quantities of the battery pack, and completes the corresponding dynamic thermal management of the battery pack; moreover, this invention has high reliability, good accuracy, and high efficiency. Attached Figure Description

[0066] Figure 1 This is a schematic diagram of the method flow of the present invention.

[0067] Figure 2 This is a schematic diagram of the functional modules of the system of the present invention. Detailed Implementation

[0068] like Figure 1The diagram shown is a flowchart of the method of the present invention: The real-time dynamic thermal management method for battery packs in energy storage stations provided by the present invention includes the following steps:

[0069] S1. Obtain parameter data information of the target energy storage station's battery pack;

[0070] S2. Select a single battery pack as the basic unit, and establish a battery pack simulation model based on the data obtained in step S1; specifically, this includes the following steps:

[0071] A single battery pack was selected as the basic unit.

[0072] Based on the data obtained in step S1, and considering the cooling method, internal structure of the resistor, and physical properties of the battery pack, a geometric and physical model of the battery pack is established to ensure mesh quality and convergence of simulation results. The simulation software can be selected according to specific conditions, as long as it meets the requirements.

[0073] The geometric and physical models of the battery pack are simplified.

[0074] The reliability of the simplified model was verified through battery pack experimental results and mesh independence assessment. The mesh structure can be selected according to the specific internal conditions of the battery pack. For regular structures, hexahedral meshes are recommended, while tetrahedral meshes can be considered for special structures.

[0075] S3. Based on the battery pack simulation model obtained in step S2, select normal operating conditions and accident operating conditions, and perform corresponding numerical simulations; specifically including the following steps:

[0076] Based on the battery pack simulation model obtained in step S2, normal operating conditions and accident operating conditions are selected; the purpose is to train a regression model of physical properties at typical locations as a function of input parameters and running time through numerical results.

[0077] Based on the selected normal and accident operating conditions, determine the physical property parameters, heat generation of the heat source, boundary conditions, and initial conditions of the simulation model. It is recommended to use the single-factor influence analysis method to arrange the experiment. If multi-factor analysis is used, attention should be paid to the rationality of the selection of factors and levels and the correct selection of data processing methods.

[0078] Considering that the placement of measurement points in actual operation will be limited by factors such as safety, lifespan, reliability, and environment, as many measurement points as possible should be placed on the battery pack before the experiment and under the experimental conditions to ensure the reliability of the verification results and the correction model. Through analysis of the internal structure of the battery pack, the physical variables that need to be monitored are selected, and monitoring points are placed in important locations.

[0079] Experiments were conducted to compare the detection parameters obtained through measurement points with the simulation results obtained from the simulation model, thereby correcting the simulation model.

[0080] Based on the internal characteristics of the battery pack, the battery pack is partitioned. The purpose is to save data sparsity and reconstruction time through parallel computing and improve the accuracy of reconstructed data. It is recommended to partition the battery pack according to its structural and material characteristics: air cooling and water cooling systems are the focus of battery pack thermal management and have relatively complex internal flow fields, so they can be further subdivided according to fluid channels.

[0081] Post-process the simulation results of each partition in the simulation model to obtain the mean and upper and lower limits of the physical quantities of each partition under the corresponding working conditions;

[0082] Using battery pack input parameters and time as independent variables, and the mean and upper / lower limits of physical quantities as dependent variables, a regression model is used for regression analysis, and the resulting regression models are used to construct a standardized model library. When the error of the multiple linear regression results is large, a multiple nonlinear regression method can be selected. Selection of physical properties: temperature, velocity, and pressure are selected for the flow field; temperature and deformation are selected for other regions.

[0083] The basic steps of multiple linear regression analysis are as follows:

[0084] Identify the independent and dependent variables, and preliminarily set the multiple linear regression equation based on their underlying mechanisms;

[0085] Estimate the parameters and determine the estimated multiple linear regression equation;

[0086] The significance of each item in the regression prediction model is tested using test statistics;

[0087] Choose statistical software that meets your needs, such as SPSS.

[0088] S4. Based on the simulation results obtained in step S3, perform data sparsity processing and train the data regression model; specifically, this includes the following steps:

[0089] The simulation results obtained in step S3 are then processed to be dimensionless.

[0090] Based on the sparsification algorithm, dimensionless data is sparsified to obtain a sparse matrix; the specific steps include the following:

[0091] The goal of sparse representation is to find a set of coefficients that represent the data, with as many zero elements as possible. The principle is to replace the basis functions with an overcomplete dictionary representation. The data elements in the dictionary are called atoms, and the data signal is represented by a linear combination of atoms.

[0092] Let the data signal be x, x∈R n R n A set of vectors representing all n-dimensional vectors (data signals); the dictionary is D, D = [d1, d2, ..., dn]. L ]∈Rn×L d L Let x be the Lth atom in the dictionary, and n be an n-dimensional vector (data signal); represent the data signal as x∈Dα, where α is a sparse matrix and α=[α1,α2,...,α L ] T ;

[0093] The sparse matrix α is calculated using the following steps:

[0094] A. Input the data signal x and the overcomplete dictionary D0, and input the sparsity K or the residual threshold ε;

[0095] B. Set initial values: The residual r0 is x, the index sets I0 and J are both empty sets; the number of iterations k is 0;

[0096] C. Calculate α k =D T r k-1 and from the obtained α k Select the K largest elements; record the index of each element in set J. k Then, sort the indices in descending order of element value;

[0097] D. Find the subset And J0 satisfies: |α(i)|≤2|α(j)|, j∈J0, i≤j;

[0098] E. Merge the found subset J0 into the index set I. k-1 ;

[0099] F. Update the sparsity coefficient α k The value is Where |||| is the value of the matrix determinant. To use a sparse decomposition algorithm to find the decomposition coefficients of the data; R' is the coefficients that are only used in set I. k The coefficient vector containing all non-zero elements at index positions;

[0100] G. Update residual r k For r k =xD k α k ;

[0101] H. Increment the iteration count by 1 and perform a judgment: if the iteration count reaches the set value, output the final sparse matrix; otherwise, return to step C for the next round of calculation.

[0102] Using the input parameters of the battery pack as independent variables and the parameters at the non-zero sparse corresponding positions on the sparse matrix as dependent variables, a regression method is used to obtain the mapping model between the independent and dependent variables, thereby constructing a visual database.

[0103] Based on the obtained sparse matrix, monitoring points are arranged within a set range on the inner wall of the battery pack surface. The basic principle is to select the surface location corresponding to the sparse data, that is, the location with a large gradient of the monitoring quantity, to arrange the measurement points. If there is no obvious gradient, it means that the measurement points on the surface are not sensitive to the arrangement location, and the arrangement can be made according to the actual situation.

[0104] Based on the simulation results of numerical simulation, a regression model between the monitoring point data and the battery pack input variables is obtained by using the regression method, thereby constructing a management model library;

[0105] S5. Acquire real-time operating data of the target energy storage station's battery pack, and reconstruct the data based on the data obtained in step S4 to achieve dynamic visualization of the battery pack; specifically including the following steps:

[0106] Real-time acquisition of operational data information of the target energy storage station's battery pack;

[0107] Based on the parameters input from the battery pack in real time, the corresponding mapping model is obtained from the visualization database;

[0108] Data reconstruction is achieved using a fast signal recovery scheme, which includes the following steps:

[0109] Reconstruction principle: By fitting the l0 norm, the problem of recovering compressed sensing signals is transformed into a function fitting problem; due to the nonlinearity of the fitting function, the process of recovering compressed sensing signals is transformed into a nonlinear optimization problem.

[0110] a. Determine the sparse matrix φ n×n To ensure the n-dimensional signal X n+1 The coefficient signal can be expressed in a sparse matrix coordinate basis.

[0111] b. Based on the random sampling matrix ψ used when compressing the signal. m×n The problem of compressed signal recovery is reduced to an optimization problem;

[0112] c. Solve using singular value decomposition to obtain... general solution form Substituting the objective function, the constrained optimization problem is transformed into an unconstrained optimization problem, expressed as:

[0113]

[0114] In the formula To find the minimum value of the objective function f(); f() is the expression for optimizing the objective function; The solution to the equation after singular value decomposition of the constraints;

[0115] d. Construct intermediate variable r as Where δ is a set small perturbation value, Let f be the gradient of the function f;

[0116] If r ≥ 0, then set the search direction d. k for If r < 0, then set the search direction d. k for Where I is the identity matrix;

[0117] e. Using the search direction from step d, solve for the objective function to obtain...

[0118] f. Using formulas The final recovery signal was obtained;

[0119] Based on the parameters input from the battery pack in real time, the corresponding regression model and the mean and upper and lower limits of the physical quantities are obtained from the standardized model library;

[0120] Based on the mean and upper and lower limits of the acquired physical quantities, the reconstructed data is processed to achieve real-time dynamic visualization of the physical quantities.

[0121] S6. Based on the working data information obtained in step S5 and the data regression model obtained in step S4, realize real-time dynamic thermal management of the target energy storage station's battery pack; specifically including the following steps:

[0122] Based on the work data information obtained in step S5, retrieve the regression model corresponding to the monitoring point from the management model library, and visualize the changing trends of the monitoring data and regression data.

[0123] Real-time calculation of the difference between monitoring data and regression data at each measuring point, and visualization of the trend of the difference over time;

[0124] When the difference exceeds a set threshold, corresponding fault diagnosis is performed to complete the real-time dynamic thermal management of the target energy storage station's battery pack. In specific implementation, fault diagnosis can be performed using a corresponding fault diagnosis model. A multi-classification algorithm is used to determine the fault type, activate the corresponding fault model, and visualize the changing trend of the measured point detection value and the model prediction value under the fault state over time to provide an early warning of the fault type. The fault development stage and possible development trend are also visualized. The choice of algorithm is not unique; any algorithm that meets the requirements is acceptable. The decision tree model is recommended.

[0125] like Figure 2The diagram shows the functional modules of the system of this invention: The system disclosed in this invention for implementing the real-time dynamic thermal management method for the battery pack of the energy storage station includes a data acquisition module, a model building module, a numerical simulation module, a regression construction module, a data reconstruction module, and a dynamic thermal management module; these modules are connected in series. The data acquisition module acquires parameter data information of the target energy storage station battery pack and uploads the data information to the model building module. The model building module selects a single battery pack as the basic unit based on the received data information, builds a battery pack simulation model based on the acquired data information, and uploads the data information to the numerical simulation module. The numerical simulation module is used to... Based on the obtained battery pack simulation model, normal and accident operating conditions are selected, and corresponding numerical simulations are performed. The data information is then uploaded to the regression construction module. The regression construction module performs data sparsification processing based on the received data information and the obtained simulation results, trains the data regression model, and uploads the data information to the data reconstruction module. The data reconstruction module acquires the working data information of the target energy storage station's battery pack in real time based on the received data information, reconstructs the data based on the obtained data information to achieve dynamic visualization of the battery pack, and uploads the data information to the dynamic thermal management module. The dynamic thermal management module implements real-time dynamic thermal management of the target energy storage station's battery pack based on the received data information, the acquired working data information, and the obtained data regression model.

Claims

1. A real-time dynamic thermal management method for battery packs in an energy storage station, comprising the following steps: S1. Obtain parameter data information of the target energy storage station's battery pack; S2. Select a single battery pack as the basic unit, and establish a battery pack simulation model based on the data obtained in step S1; S3. Based on the battery pack simulation model obtained in step S2, select normal operating conditions and accident operating conditions, and perform corresponding numerical simulations. S4. Based on the simulation results obtained in step S3, perform data sparsity processing and train the data regression model; specifically, this includes the following steps: The simulation results obtained in step S3 are then processed to be dimensionless. Based on the sparsification algorithm, dimensionless data is sparsified to obtain a sparse matrix; Using the input parameters of the battery pack as independent variables and the parameters at the non-zero sparse corresponding positions on the sparse matrix as dependent variables, a regression method is used to obtain the mapping model between the independent and dependent variables, thereby constructing a visual database. Based on the obtained sparse matrix, monitoring points are arranged within a set range on the inner wall of the battery pack surface. Based on the simulation results of numerical simulation, a regression model between the monitoring point data and the battery pack input variables is obtained by using the regression method, thereby constructing a management model library; The aforementioned sparsification algorithm performs sparsification processing on dimensionless data to obtain a sparse matrix, specifically including the following steps: Set the data signal as , , Let D be a set of vectors represented by all n-dimensional vectors; the dictionary is D. , For the Lth atom in the dictionary, Given an n-dimensional vector; the data signal is represented as ,in It is a sparse matrix and ; The sparse matrix is ​​calculated using the following steps. : A. Input data signal and super-complete dictionary And input the sparsity K or residual threshold. ; B. Set initial values: residuals Values index set Both J and J are empty sets; the number of iterations k is 0; C. Calculation and from the obtained Select the K largest elements; record the index of each element into the set. Then, sort the indices in descending order of element value; D. Find the subset and make satisfy: , , , ; E. The found subset Merge into index set ; F. Update sparsity coefficients The value is ;in, The value of the matrix determinant. To use sparse decomposition algorithms to obtain decomposition coefficients from the data; For only in the set The coefficient vector containing all non-zero elements at index positions; G. Update residuals for ; H. Increment the iteration count by 1 and make a judgment: if the iteration count reaches the set value, output the final sparse matrix; otherwise, return to step C for the next round of calculation. S5. Acquire the working data information of the target energy storage station battery pack in real time, and reconstruct the data based on the data information obtained in step S4 to realize the dynamic visualization of the battery pack; S6. Based on the working data information obtained in step S5 and the data regression model obtained in step S4, realize the real-time dynamic thermal management of the target energy storage station battery pack.

2. The real-time dynamic thermal management method for battery packs in energy storage stations according to claim 1, characterized in that... Step S2 involves selecting a single battery pack as the basic unit and establishing a battery pack simulation model based on the data obtained in step S1. This specifically includes the following steps: A single battery pack was selected as the basic unit. Based on the data obtained in step S1, the geometric and physical models of the battery pack are established based on the cooling method, internal structure and physical properties of the resistor. The geometric and physical models of the battery pack are simplified. The reliability of the simplified model was verified through battery pack experimental results and grid independence assessment.

3. The real-time dynamic thermal management method for battery packs in energy storage stations according to claim 2, characterized in that... Step S3, based on the battery pack simulation model obtained in step S2, selects normal operating conditions and accident operating conditions, and performs corresponding numerical simulations, specifically including the following steps: Based on the battery pack simulation model obtained in step S2, normal operating conditions and accident operating conditions are selected. Based on the selected normal and accident operating conditions, determine the physical property parameters, heat generation of the heat source, boundary conditions, and initial conditions of the simulation model; Before the experiment, under the experimental conditions, set up as many measurement points as possible for the battery pack; Experiments were conducted to compare the detection parameters obtained through measurement points with the simulation results obtained from the simulation model, thereby correcting the simulation model. Based on the internal characteristics of the battery pack, the battery pack is divided into zones; Post-process the simulation results of each partition in the simulation model to obtain the mean and upper and lower limits of the physical quantities of each partition under the corresponding working conditions; Using battery pack input parameters and time as independent variables, and the mean and upper and lower limits of physical quantities as dependent variables, regression analysis was performed using a regression model, and the obtained regression model was used to construct a standardized model library.

4. The real-time dynamic thermal management method for battery packs in energy storage stations according to claim 3, characterized in that... Step S5 involves acquiring real-time operating data of the target energy storage station's battery pack and reconstructing the data based on the data obtained in step S4 to achieve dynamic visualization of the battery pack. Specifically, this includes the following steps: Real-time acquisition of operational data information of the target energy storage station's battery pack; Based on the parameters input from the battery pack in real time, the corresponding mapping model is obtained from the visualization database; Data reconstruction is achieved using a fast signal recovery scheme; Based on the parameters input from the battery pack in real time, the corresponding regression model and the mean and upper and lower limits of the physical quantities are obtained from the standardized model library; Based on the mean and upper and lower limits of the acquired physical quantities, the reconstructed data is dimensionalized to achieve real-time dynamic visualization of the physical quantities.

5. The real-time dynamic thermal management method for battery packs in energy storage stations according to claim 4, characterized in that... The aforementioned data reconstruction using a fast signal recovery scheme specifically includes the following steps: a. Determine the sparse matrix To ensure n-dimensional signals The coefficient signal can be expressed in a sparse matrix coordinate basis. ; b. Based on the random sampling matrix used when compressing the signal. The problem of compressed signal recovery is reduced to an optimization problem; c. Solve using singular value decomposition to obtain... general solution form Substituting the objective function, the constrained optimization problem is transformed into an unconstrained optimization problem, expressed as: In the formula For the objective function Find the minimum value; To optimize the objective function expression; The solution to the equation after singular value decomposition of the constraints; d. Construct the intermediate variable r as ,in For the set small perturbation value, Let f be the gradient of the function f; like Then set the search direction. for ;like Then set the search direction. for ;in, It is the identity matrix; e. Using the search direction from step d, solve for the objective function to obtain... ; f. Using formulas The final recovery signal was obtained.

6. The real-time dynamic thermal management method for battery packs in energy storage stations according to claim 5, characterized in that... Step S6, which involves using the working data information obtained in step S5 and the data regression model obtained in step S4 to achieve real-time dynamic thermal management of the target energy storage station's battery pack, specifically includes the following steps: Based on the work data information obtained in step S5, retrieve the regression model corresponding to the monitoring point from the management model library, and visualize the changing trends of the monitoring data and regression data. Real-time calculation of the difference between monitoring data and regression data at each measuring point, and visualization of the trend of the difference over time; When the difference exceeds the set threshold, the corresponding fault diagnosis is performed.

7. A system for implementing the real-time dynamic thermal management method for battery packs in energy storage stations according to any one of claims 1 to 6, characterized in that... It includes a data acquisition module, a model building module, a numerical simulation module, a regression construction module, a data reconstruction module, and a dynamic thermal management module; these modules are connected in series. The data acquisition module is used to acquire parameter data of the target energy storage station's battery pack and upload the data to the model building module. The model building module is used to select a single battery pack as the basic unit based on the received data, build a battery pack simulation model based on the acquired data, and upload the data to the numerical simulation module. The numerical simulation module is used to select normal operating conditions and accident operating conditions based on the received data information and the obtained battery pack simulation model, and perform corresponding numerical simulations, and upload the data information to the regression construction module. The regression construction module is used to perform data sparsification processing based on the received data information and the obtained simulation results, train the data regression model, and upload the data information to the data reconstruction module. The data reconstruction module is used to acquire the working data information of the target energy storage station battery pack in real time based on the received data information, reconstruct the data based on the acquired data information to realize the dynamic visualization of the battery pack, and upload the data information to the dynamic thermal management module; the dynamic thermal management module is used to realize the real-time dynamic thermal management of the target energy storage station battery pack based on the received data information, the acquired working data information and the obtained data regression model.

Images