Lithium battery energy storage system fire-fighting ventilation and explosion venting safety assessment method based on multi-dimensional simulation

Abstract

The invention relates to a lithium battery energy storage system fire-fighting ventilation and explosion venting safety assessment method based on multi-dimensional simulation. The method comprises the following steps: constructing an energy storage system digital twinborn model fusing structure parameters, material attributes and environmental parameters; generating a multi-mode failure scene set covering multiple temperature domains and aging states through mode recognition; simulating and quantifying dynamic interaction of a temperature field, a flow field and a stress field in the thermal runaway evolution process based on thermal-fluid-solid multi-physics field coupling; constructing a space-time associated dynamic safety evaluation matrix, and combining fuzzy comprehensive evaluation and Monte Carlo sampling to generate risk quantitative indexes; and iteratively correcting parameters of the fire-fighting ventilation and explosion venting system through a multi-objective optimization algorithm to form a graded safety assessment conclusion. According to the method, the technical bottlenecks of environmental parameter splitting and single failure scene in a traditional method are broken through, the thermal runaway suppression efficiency is improved, the combustible gas concentration control error is reduced, and collaborative optimization of explosion venting pressure fluctuation suppression and ventilation response is realized through a closed-loop evaluation mechanism.

Description

Technical Field

[0001] The present invention relates to a safety assessment method, in particular to a fire ventilation and explosion venting safety assessment method for a lithium battery energy storage system based on multi-dimensional simulation. Background Art

[0002] Lithium battery energy storage systems are developing rapidly towards the megawatt-level containerized direction, but the domino effect caused by thermal runaway at the cell level has led to frequent fire and explosion accidents. Most existing simulation technologies use a one-way thermal-electric coupling model (such as the Joule Heating module of COMSOL Multiphysics), ignoring the two-phase flow impact caused by electrolyte vaporization and the bidirectional mechanical feedback of shell deformation, resulting in a prediction deviation of more than 30% in the diffusion path of thermal runaway gas. Research by the Sandia National Laboratories in the United States in 2023 showed that the traditional thermal runaway prediction model based on the Arrhenius equation has an early warning response time error of 8 - 12 seconds in module-level applications due to the lack of consideration of the lithium deposition autocatalytic effect of aged cells, and cannot meet the stringent requirements of the NFPA 855 standard for millisecond-level linkage of explosion venting systems.

[0003] There are significant limitations in the current failure scenario construction methods: on the one hand, the analysis of historical case data relies on manual feature extraction (such as the voltage sudden drop threshold method), making it difficult to capture the non-linear failure trajectories of multi-parameter coupling; on the other hand, the random scenarios generated by the traditional Monte Carlo method lack physical constraints, resulting in more than 40% of the amplified scenarios violating Fourier's law of heat conduction. For example, the static explosion venting area calculation formula (A = 0.07×Q^0.5) used in the Tesla Megapack system has an error of up to 45% in a low-temperature and high-humidity environment, exposing the adaptability defects of empirical formulas under complex boundary conditions.

[0004] There is a "data island" problem in the existing safety assessment system: the BMS operation data, multi-physical field simulation results, and fire protection system parameters lack dynamic coupling. The thermal runaway suppression test in the UL 9540A standard still uses a constant temperature boundary condition, without considering the spatio-temporal coupling effect of temperature gradient and air pressure fluctuation during actual operation, resulting in a virtual marking rate of up to 22% for the coverage rate of the sprinkler system. German certification cases in 2022 showed that the explosion venting systems evaluated by traditional methods had a failure rate of up to 17% in real thermal runaway due to pressure feedback delay, highlighting the difficulty of existing technologies in meeting the precise protection requirements of high-dynamic risk scenarios. Summary of the Invention

[0005] The object of the present invention is to overcome the defects of the above-mentioned existing technologies and provide a fire ventilation and explosion venting safety assessment method for a lithium battery energy storage system based on multi-dimensional simulation, breaking through the technical bottlenecks of fragmented environmental parameters and single failure scenarios in traditional methods, improving the thermal runaway suppression efficiency, reducing the control error of combustible gas concentration, and realizing the collaborative optimization of explosion venting pressure fluctuation suppression and ventilation response through a closed-loop assessment mechanism.

[0006] The object of the present invention can be achieved through the following technical solutions:

[0007] The present invention provides a fire ventilation and explosion venting safety assessment method for a lithium battery energy storage system based on multi-dimensional simulation, including the following steps:

[0008] S1. Construct a digital twin model of the energy storage system including structural parameters, material properties and environmental parameters;

[0009] S2. Based on the digital twin model, import historical failure case data and generate a multi-modal failure scenario set through pattern recognition;

[0010] S3. Conduct multi-physical field coupling simulation based on the failure scenario set and perform thermal-fluid-solid simultaneous calculation;

[0011] S4. Construct a dynamic safety assessment matrix according to the simulation output and generate a risk quantification index based on this matrix;

[0012] S5. Establish a fire ventilation and explosion venting system safety assessment model based on the risk quantification index and generate a safety assessment conclusion.

[0013] Further, in S1, it specifically includes the following sub-steps:

[0014] Obtain the spatial layout of the battery module and the geometric characteristics of the box structure through three-dimensional point cloud scanning, and establish a parametric geometric model;

[0015] Integrate the pyrolysis kinetic parameters of the battery cell materials, the melting temperature threshold of the separator and the electrolyte volatility data to construct a multi-level material property database;

[0016] Then deploy temperature and humidity sensors and gas concentration detectors to collect real-time environmental baseline data in the energy storage cabin;

[0017] Based on the multi-physical field coupling simulation platform, dynamically couple the structural topology data, material constitutive equations and environmental boundary conditions to obtain the digital twin model of the energy storage system.

[0018] Further, in S2, it specifically includes the following sub-steps:

[0019] Clean and standardize the format of multi-source heterogeneous data in historical failure cases, including the voltage / temperature time series curves of battery modules, the positioning information of thermal runaway trigger points, and the text records of operation and maintenance logs, and construct a structured failure feature matrix;

[0020] Secondly, inject the data into the SOFC state tracking module of the digital twin model through the API interface, and extract multi-dimensional failure-sensitive features by using the joint time-frequency domain analysis method;

[0021] Then, fuse the Pearson correlation coefficient and the dynamic time warping distance to divide the failure mode categories, and establish a fault tree ontology model for the intrinsic failure mode;

[0022] Then, amplify the failure trajectory in the parameter space through the generative adversarial network (GAN), and generate multi-modal failure scenarios covering the temperature range of -20°C to 60°C and different aging degrees by combining the environmental parameter perturbation factors.

[0023] Furthermore, in S3, it specifically includes the following sub-steps:

[0024] First, establish the weak form expression of the thermo-fluid-structure coupling control equations, and adopt the partitioned coupling strategy to realize the fully coupled iterative calculation of the temperature field, the electrolyte flow field, and the structural stress field;

[0025] Based on the parametric mesh division of the failure scenario set, use the dynamic mesh technology to handle the topological changes of the contact interface caused by the expansion of the battery cell, set the time step of the unsteady solver to 0.1 ms, and enable the explicit integration algorithm;

[0026] In the thermal field module, load the anisotropic thermal conductivity of the battery cell and the latent heat of decomposition model of the SEI film, and use the finite volume method to solve the energy conservation equation including Joule heat and convective heat transfer;

[0027] In the flow field module, activate the porous medium turbulence model to simulate the two-phase flow migration process caused by the vaporization of the electrolyte due to heat;

[0028] In the solid mechanics module, define the elastoplastic constitutive relationship of the current collector and the creep failure criterion of the separator, and capture the deformation of the shell by the ALE method; based on the failure scenario set, inject the temperature gradient and air pressure fluctuation boundary conditions, and use the Newton-Raphson iteration method to solve the simultaneous equations until the residual is less than the preset value.

[0029] Furthermore, in S4, it specifically includes the following sub-steps:

[0030] Integrate the multi-dimensional characteristic parameters output by the multi-physics field simulation, and construct a dynamic safety assessment matrix by dimensionality reduction through principal component analysis (PCA). The matrix dimension covers time series and spatial distribution;

[0031] Based on the Bayesian network, a probability graph model of the thermal runaway propagation path is established to calculate the safety margin index of each node and the entropy value of cascade failure risk;

[0032] The improved Analytic Hierarchy Process (AHP) is used to determine the dynamic weights of key indicators, and the BMS operation data is mapped to the evaluation matrix through a real-time OPC interface;

[0033] The fuzzy comprehensive evaluation method is used to calculate the system-level risk quantification index RQI = 0.7×(temperature anomaly / 280°C) + 0.3×(deformation coordination coefficient), and a risk probability cloud map is generated by combining Monte Carlo sampling;

[0034] Finally, through cross-validation, the evaluation results are compared with the UL 9540A standard value, a three-dimensional risk dashboard containing four-level warning thresholds is established, and the rolling update of the risk trajectory in the next 30s is realized based on the LSTM time series prediction module to ensure that the prediction error rate of the risk quantification index ≤ 5%.

[0035] Further, in S5, it specifically includes the following sub-steps:

[0036] According to the risk indicators, a comprehensive safety evaluation model of the fire ventilation and explosion relief system is constructed to quantitatively analyze the thermal runaway suppression effect and gas diffusion control ability;

[0037] Optimize the layout parameters and triggering conditions of the explosion relief device, and verify its ability to quickly control the pressure after thermal runaway through fire simulation;

[0038] Establish a dynamic response mechanism for the fire extinguishing system and ventilation equipment to test their collaborative working efficiency;

[0039] Based on the multi-dimensional performance verification results, a hierarchical safety conclusion covering explosion relief protection, temperature control, and heat spread blocking is formed.

[0040] Further, in S2, during the failure trajectory amplification process, by introducing the constraint condition of the thermal runaway gas diffusion rate, a physical information loss function of the generative adversarial network is constructed to generate a subset of failure scenarios that conform to Fourier's law of heat conduction and Fick's law of diffusion, ensuring that the spatio-temporal distributions of the temperature gradient and combustible gas concentration gradient in the amplified scenarios satisfy the actual physical laws.

[0041] Further, in S3, during the moving mesh technology processing, a contact pressure feedback mechanism based on the expansion coefficient of the battery cell is established. When the volume expansion rate of the battery cell exceeds 15%, a local mesh refinement strategy is automatically triggered, and a Johnson-Cook plasticity model related to the strain rate is introduced in the shell deformation calculation to realize the two-way coupling calculation of the battery cell expansion and shell deformation.

[0042] Further, in S4, when constructing the dynamic safety assessment matrix, a cross-dimensional feature fusion module based on a graph neural network is embedded to perform graph structure modeling on the spatial distribution features of the temperature field and the voltage time series features, extract potential correlation features of thermal runaway propagation between battery cells through node embedding technology, and construct a three-dimensional assessment matrix including topological relationships.

[0043] Further, in S5, in the test of the dynamic response mechanism, a closed-loop verification system including pressure wavefront detection and gas component analysis is constructed. After the explosion relief valve acts, data on the propagation path of the shock wave is collected in real time, and the matching degree between the explosion relief air flow field and the simulation result is verified through particle image velocimetry technology to dynamically correct the frequency conversion control curve of the ventilation system.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] Through the construction of a multi-dimensional fusion digital twin model and multi-physical field coupling simulation, the present invention realizes the refined simulation and dynamic risk assessment of the thermal runaway evolution process of the lithium battery energy storage system, effectively solving the problems of fragmented environmental parameters and single failure scenarios in traditional methods; based on the risk quantification index generated by the dynamic safety assessment matrix, combined with the multi-objective optimization algorithm and real-time feedback mechanism, the parameter configuration of the fire ventilation and explosion relief systems can be accurately matched, improving the thermal runaway suppression efficiency and reducing the control error of the combustible gas concentration. At the same time, through the hierarchical protection system and standardized certification, the active safety protection ability of the energy storage system is significantly improved, providing a systematic solution for the fire prevention and control of large-scale energy storage power stations. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a flowchart of the fire ventilation and explosion relief safety assessment method for a lithium battery energy storage system based on multi-dimensional simulation in the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. Features such as component models, material names, connection structures, control methods, algorithms, etc. that are not clearly described in the present technical solution are regarded as common technical features disclosed in the prior art.

[0048] Embodiment 1

[0049] In the fire ventilation and explosion relief safety assessment method for a lithium battery energy storage system based on multi-dimensional simulation in this embodiment, it includes 5 steps, and the overall process is shown in Figure 1 .

[0050] S1. Construct a digital twin model of the energy storage system including structural parameters, material properties, and environmental parameters;

[0051] Specifically, it includes the following sub-steps:

[0052] The spatial layout of the battery module and the geometric features of the box structure are acquired through 3D point cloud scanning, and a parametric geometric model is established;

[0053] Integrate the pyrolysis kinetic parameters of battery material, the melting temperature threshold of the separator and the volatility data of the electrolyte to build a multi-level material property database;

[0054] Then, temperature and humidity sensors and gas concentration detectors are deployed to collect environmental baseline data in the energy storage cabin in real time;

[0055] Based on the multi-physics field coupling simulation platform, the structural topology data, material constitutive equations and environmental boundary conditions are dynamically coupled to obtain a digital twin model of the energy storage system.

[0056] Step S1 of the present invention constructs a digital twin model through multi-dimensional data fusion and dynamic coupling modeling technology. The principle is as follows: first, the geometric features of the battery module are reversely reconstructed through three-dimensional point cloud scanning, and the mechanical structure features (such as bolt preload and module spacing) are converted into adjustable variables using parametric modeling technology to solve the problem of structural parameter distortion in traditional simulation; secondly, the material pyrolysis kinetic parameters (such as activation energy and pre-exponential factor in the Arrhenius equation) and the critical threshold of phase change (membrane melting temperature T_m, electrolyte flash point T_f) are integrated to construct a multi-level material property digital twin model. The database provides basic data of constitutive equations for multi-physics field simulation; then the environmental baseline data (temperature and humidity gradient ΔT / ΔRH, H2 / CO gas concentration) is collected in real time through a distributed sensor network to form a dynamic boundary condition input flow; finally, based on the multi-field coupling platform, the structural topology (finite element mesh), material properties (anisotropic thermal conductivity tensor K_ij) and environmental parameters (convective heat transfer coefficient h) are solved jointly, and the physical characteristics and real-time status of the digital twin model are synchronized through bidirectional data drive, establishing a high-fidelity virtual mapping space for subsequent failure simulation.

[0057] S2. Based on the digital twin model, historical failure case data is imported, and a multimodal failure scenario set is generated through pattern recognition;

[0058] It includes the following sub-steps:

[0059] Clean and format the multi-source heterogeneous data in historical failure cases, including battery module voltage / temperature time series curves, thermal runaway trigger point location information, and operation and maintenance log text records, to build a structured failure feature matrix;

[0060] Secondly, the data is injected into the SOFC state tracking module of the digital twin model through the API interface, and the multi-dimensional failure sensitive features are extracted using the time-frequency domain joint analysis method;

[0061] Then, fuse the Pearson correlation coefficient and the dynamic time warping distance to divide the failure mode categories, and establish a fault tree ontology model for the intrinsic failure mode;

[0062] Then, through the generative adversarial network (GAN), amplify the failure trajectories in the parameter space, and combine the environmental parameter perturbation factors to generate multi-modal failure scenarios covering the temperature range of -20°C to 60°C and different aging degrees.

[0063] In S2, during the failure trajectory amplification process, introduce the constraint condition of the thermal runaway gas diffusion rate to construct the physical information loss function of the generative adversarial network, and generate a subset of failure scenarios that conform to Fourier's law of heat conduction and Fick's law of diffusion, ensuring that the spatio-temporal distributions of the temperature gradient and the combustible gas concentration gradient in the amplified scenarios satisfy the actual physical laws.

[0064] In step S2 of the present invention, a failure scenario generation method that combines data-driven and physical constraints is used to realize multi-modal failure modeling. The principle is as follows: First, through data cleaning and standardization processing, align the voltage / temperature time-series signals, thermal runaway localization coordinates, and operation and maintenance texts in space-time to construct a failure feature matrix containing three-dimensional features of time-space-semantics, and solve the problem of multi-source heterogeneous data fusion; Second, embed the SOFC state tracking module in the digital twin model, and use time-frequency domain joint analysis (such as wavelet packet energy spectrum and Hilbert marginal spectrum) to extract the phase mutation features of voltage dips and the frequency domain resonance characteristics of gas products, revealing the energy transfer law during the failure process; Furthermore, quantify the failure correlation degree between battery cells through the Pearson correlation coefficient, and combine the dynamic time warping (DTW) distance to measure the failure propagation delay across modules, and construct a fault tree ontology model with causal reasoning ability; Finally, in the GAN network training, introduce the Fourier number (Fo) and Schmidt number (Sc) to construct the physical information loss function, and constrain the temperature gradient and gas concentration field to satisfy Fourier's heat conduction equation and Fick's law of diffusion, ensuring that the amplified failure scenarios conform to the actual physical evolution law in key parameters such as the thermal runaway gas diffusion rate (vd = α·ΔT / Δx), so as to generate a multi-modal failure scenario library that covers both extreme environments and retains the intrinsic physical relationships.

[0065] S3. Perform multi-physics field coupling simulation based on the failure scenario set, and execute the coupled calculation of heat-fluid-solid;

[0066] Specifically, it includes the following sub-steps:

[0067] First, establish the weak form expression of the coupled control equations of heat-fluid-solid, and adopt the partition coupling strategy to realize the fully coupled iterative calculation of the temperature field, electrolyte flow field, and structural stress field;

[0068] Based on the parametric meshing of the failure scenario set, the dynamic mesh technology is used to handle the topological changes of the contact interface caused by the expansion of the battery cell. The time step of the unsteady solver is set to 0.1 ms and the explicit integration algorithm is enabled;

[0069] In the thermal field module, the anisotropic thermal conductivity of the battery cell and the latent heat model of SEI film decomposition are loaded, and the finite volume method is used to solve the energy conservation equation including Joule heat and convective heat transfer;

[0070] In the fluid flow field module, the porous medium turbulence model is activated to simulate the two-phase flow migration process caused by the vaporization of the electrolyte due to heating;

[0071] In the solid mechanics module, the elastoplastic constitutive relationship of the current collector and the creep failure criterion of the separator are defined, and the ALE method is used to capture the deformation of the shell;

[0072] Based on the failure scenario set, the temperature gradient and air pressure fluctuation boundary conditions are injected, and the Newton-Raphson iteration method is used to solve the simultaneous equations until the residual is less than the preset value.

[0073] In S3, during the process of handling the dynamic mesh technology, a contact pressure feedback mechanism based on the expansion coefficient of the battery cell is established. When the volume expansion rate of the battery cell exceeds 15%, the local mesh encryption strategy is automatically triggered, and the Johnson-Cook plasticity model related to the strain rate is introduced in the shell deformation calculation to realize the two-way coupling calculation of the battery cell expansion and the shell deformation.

[0074] Step S3 of the present invention realizes the high-precision simulation of the thermal runaway evolution process through the two-way coupling and dynamic mesh adaptive technology. The principle is as follows: The partition coupling strategy is used to solve the temperature field (energy equation), fluid flow field (N-S equation) and solid mechanics field (constitutive equation) simultaneously. The two-way energy transfer path between fields is established through the weak form expression (such as Joule heat-induced temperature rise → electrolyte vaporization generates flow stress → shell deformation reacts on the thermal boundary conditions), breaking through the accuracy limitation of the traditional one-way coupling; In the dynamic mesh processing, the contact interface pressure distribution is quantified based on the expansion coefficient of the battery cell. When the local volume expansion rate δ_v≥15%, the unstructured mesh adaptive encryption is triggered, and the Johnson-Cook model is combined to describe the plastic deformation related to the strain rate of the shell to realize the two-way mechanical coupling of the battery cell-shell contact interface; At the same time, the anisotropic thermal conductivity tensor (k_θ = k_r + β(T - T_0)) is used to characterize the radial / axial differential heat transfer characteristics inside the battery, and the latent heat release term (Q = ΔH·dα / dt) is embedded in the SEI film decomposition model to accurately simulate the heat release cumulative effect of the thermal runaway chain reaction; Finally, the Newton iteration method is used to solve the nonlinear coupling equations on the parametric mesh, and the residual is converged to the order of 1e-5, ensuring that the spatio-temporal evolution of the temperature gradient field, two-phase flow field and stress field is physically consistent with the real failure process within the error range of ±8%.

[0075] S4. Construct a dynamic safety assessment matrix based on the simulation output, and generate risk quantification indicators based on this matrix;

[0076] Specifically, it includes the following sub-steps:

[0077] Integrate the multi-physical field simulation output of multi-dimensional characteristic parameters, and reduce the dimension through principal component analysis (PCA) to construct a dynamic safety assessment matrix, where the matrix dimension covers time series and spatial distribution;

[0078] Based on the Bayesian network, establish a probability graph model of the thermal runaway propagation path, and calculate the safety margin index and cascade failure risk entropy value of each node;

[0079] Adopt the improved analytic hierarchy process (AHP) to determine the dynamic weights of key indicators, and map the BMS operation data to the assessment matrix through the real-time OPC interface;

[0080] Use the fuzzy comprehensive evaluation method to calculate the system-level risk quantification indicator RQI = 0.7×(temperature anomaly degree / 280°C) + 0.3×(deformation coordination coefficient), and generate a risk probability cloud map by combining Monte Carlo sampling;

[0081] Finally, through cross-validation, compare the evaluation results with the UL 9540A standard value, establish a three-dimensional risk dashboard including 4-level warning thresholds, and realize the rolling update of the risk trajectory in the next 30s based on the LSTM time series prediction module to ensure that the prediction error rate of the risk quantification indicator ≤ 5%.

[0082] In S4, when constructing the dynamic safety assessment matrix, embed a cross-dimensional feature fusion module based on the graph neural network, perform graph structure modeling on the spatial distribution characteristics of the temperature field and the voltage time series characteristics, extract the potential correlation characteristics of thermal runaway propagation between battery cells through node embedding technology, and construct a three-dimensional assessment matrix including topological relationships.

[0083] In step S4 of the present invention, risk quantification assessment is achieved through multi-modal data fusion and dynamic weight optimization, and the principle is as follows: First, principal component analysis (PCA) is used to perform orthogonal transformation on the high-dimensional features output by multi-physical fields (such as temperature gradient extreme values, stress concentration coefficients), and principal components with a contribution degree ≥ 95% are extracted to construct a spatio-temporal correlation dynamic assessment matrix, solving the problems of feature redundancy and dimensionality disaster in traditional assessments; Secondly, a heat runaway propagation probability graph model is constructed based on a Bayesian network, the failure transfer probability between adjacent battery cells is quantified through a conditional probability table (CPT), and the single-point failure risk is quantified by combining the safety margin index SMI = 1 - (σ / σ_y), and the cascade failure uncertainty is characterized by Shannon entropy H; In the improved AHP, a sliding time window is introduced to dynamically adjust the weight factor, and the SoC and SOH data of the BMS are mapped as weight update trigger conditions through a real-time OPC interface to achieve adaptive optimization of evaluation indicators; Further, a graph neural network (GNN) is embedded to construct a cross-dimensional feature fusion module, and the spatial distribution of the temperature field (grid node temperature values) and the voltage time series fluctuation (time series slope) are modeled as a heterogeneous graph structure, and the potential associated edge weights of heat runaway propagation between battery cells are extracted through a graph attention mechanism (GAT) to construct a three-dimensional evaluation matrix integrating topological relationships; Finally, a fuzzy membership function is used to quantify the non-linear relationship between temperature abnormality and deformation coordination coefficient, a risk probability distribution cloud map is generated by combining Monte Carlo sampling, and the time series evolution law of the evaluation matrix is learned through an LSTM module to achieve rolling prediction of the risk trajectory and adaptive correction of the threshold, forming a closed-loop dynamic risk assessment system.

[0084] S5. Establish a safety assessment model for the fire ventilation and explosion relief system based on the risk quantification index, and generate a safety assessment conclusion.

[0085] Specifically, it includes the following sub-steps:

[0086] Construct a comprehensive safety assessment model for the fire ventilation and explosion relief system according to the risk indicators, and quantitatively analyze the heat runaway suppression effect and gas diffusion control ability;

[0087] Optimize the layout parameters and trigger conditions of the explosion relief device, and verify its ability to quickly control the pressure after heat runaway through fire simulation;

[0088] Establish a dynamic response mechanism for the fire extinguishing system and ventilation equipment, and test their collaborative working efficiency;

[0089] Based on the multi-dimensional performance verification results, form a hierarchical safety conclusion covering explosion relief protection, temperature control and heat spread blocking.

[0090] In the test of the dynamic response mechanism, a closed-loop verification system including pressure wavefront detection and gas component analysis is constructed. After the explosion relief valve acts, data on the propagation path of the shock wave is collected in real time, and the matching degree between the explosion relief air flow field and the simulation results is verified through particle image velocimetry technology, and the frequency conversion control curve of the ventilation system is dynamically corrected.

[0091] In step S5 of the present invention, an accurate generation of the safety assessment conclusion is achieved through multi-dimensional verification and a closed-loop feedback mechanism. The principle is as follows: The assessment model constructed based on risk quantification indicators performs multi-objective optimization on the thermal runaway suppression efficiency and the gas diffusion control coefficient. The influence of the layout parameters of the explosion relief valve on the pressure wave attenuation efficiency is verified through fire dynamics simulation. The correlation between the shock wave propagation path and the shell structure resonance frequency is analyzed by combining computational fluid dynamics (CFD), and the trigger pressure threshold is optimized to avoid the structural natural frequency band; in the dynamic response test, a high-speed pressure sensor array and a laser absorption spectrometer are deployed to construct a closed-loop verification system, and the pressure gradient of the explosion relief air flow field is captured in real time And the concentration distribution of combustible gases (H2 / CO), and the vorticity of the flow field is obtained through particle image velocimetry (PIV) technology And an Euler-Lagrange joint analysis is performed with the simulation results, and the frequency conversion control curve of the ventilation system is dynamically corrected (such as the fan speed-time function n(t) = f(ΔP, C_g)); finally, a three-dimensional evaluation system of the explosion relief protection efficiency index (EPI = ∫(P_peak - P_safe)dt), the temperature control response coefficient (K_T = ΔT_max / Δt), and the heat blocking aging (τ_block) is established based on the hierarchical safety conclusion. Through the convergence analysis of the experimental data and the standard limit values, a safety assessment conclusion with engineering guiding significance is generated.

[0092] The above description of the embodiments is to enable those of ordinary skill in the art to understand and use the invention. It is obvious that those skilled in the art can easily make various modifications to these embodiments and apply the general principles described herein to other embodiments without creative efforts. Therefore, the present invention is not limited to the above embodiments, and the improvements and modifications made by those skilled in the art without departing from the scope of the present invention should be within the protection scope of the present invention.

Claims

1. A fire ventilation and explosion relief safety assessment method for a lithium battery energy storage system based on multi-dimensional simulation, characterized in that, It includes the following steps: S1. Construct a digital twin model of the energy storage system including structural parameters, material properties, and environmental parameters; S2. Based on the digital twin model, import historical failure case data, and generate a multi-modal failure scenario set through pattern recognition; S3. Conduct multi-physical field coupling simulation based on the failure scenario set, and perform thermal-fluid-solid simultaneous calculation; S4. Construct a dynamic safety assessment matrix according to the simulation output, and generate a risk quantification index based on this matrix; S5. Establish a safety assessment model for the fire ventilation and explosion relief system based on the risk quantification index, and generate a safety assessment conclusion.

2. The fire ventilation and explosion relief safety assessment method for a lithium battery energy storage system based on multi-dimensional simulation according to claim 1, wherein In S1, it specifically includes the following sub-steps: Obtain the spatial layout of the battery module and the geometric characteristics of the box structure through 3D point cloud scanning, and establish a parametric geometric model; Integrate the pyrolysis kinetics parameters of the cell material, the melting temperature threshold of the separator, and the volatile data of the electrolyte to construct a multi-level material property database; Then deploy temperature and humidity sensors and gas concentration detectors to collect real-time environmental baseline data in the energy storage cabin; Based on the multi-physical field coupling simulation platform, dynamically couple the structural topology data, material constitutive equations, and environmental boundary conditions to obtain a digital twin model of the energy storage system.

3. A method for fire ventilation and explosion venting safety assessment of a lithium battery energy storage system based on multi-dimensional simulation according to claim 1, characterized in that In S2, it specifically includes the following sub-steps: Clean and standardize the format of multi-source heterogeneous data in historical failure cases, including the voltage / temperature time series curve of the battery module, the positioning information of the thermal runaway trigger point, and the text record of the operation and maintenance log, to construct a structured failure feature matrix; Secondly, inject the data into the SOFC state tracking module of the digital twin model through the API interface, and use the time-frequency domain joint analysis method to extract multi-dimensional failure sensitive features; Then fuse the Pearson correlation coefficient and the dynamic time warping distance to divide the failure mode categories, and establish a fault tree ontology model of the intrinsic failure mode; Then, through the generative adversarial network (GAN), amplify the failure trajectory in the parameter space, and combine the environmental parameter perturbation factor to generate multi-modal failure scenarios covering the temperature range of -20°C to 60°C and different aging degrees.

4. A method for fire ventilation and explosion relief safety assessment of a lithium battery energy storage system based on multi-dimensional simulation according to claim 1, characterized in that In S3, it specifically includes the following sub-steps: First, establish a weak form expression of the thermal-fluid-solid coupling control equations, and adopt a partitioned coupling strategy to realize the full-coupling iterative calculation of the temperature field, electrolyte flow field, and structural stress field; Based on the parameterized mesh division of the failure scenario set, use the dynamic mesh technology to process the topological changes of the contact interface caused by the swelling of the cell, set the time step of the unsteady solver to 0.1 ms, and enable the explicit integration algorithm; In the thermal field module, load the anisotropic thermal conductivity of the cell and the latent heat model of SEI film decomposition, and use the finite volume method to solve the energy conservation equation including Joule heat and convective heat transfer; In the flow field module, activate the porous media turbulence model to simulate the two-phase flow migration process caused by the vaporization of the electrolyte due to heat; In the solid mechanics module, define the elastic-plastic constitutive relationship of the current collector and the creep failure criterion of the separator, and capture the deformation of the shell through the ALE method; inject the temperature gradient and air pressure fluctuation boundary conditions based on the failure scenario set, and use the Newton-Raphson iteration method to solve the simultaneous equations until the residual is less than the preset value.

5. A method for fire ventilation and explosion venting safety assessment of a lithium battery energy storage system based on multi-dimensional simulation according to claim 1, characterized in that, In S4, it specifically includes the following sub-steps: Integrate multi-physical field simulation output multi-dimensional characteristic parameters, and construct a dynamic safety assessment matrix through dimensionality reduction by principal component analysis (PCA). The matrix dimensions cover time series and spatial distribution; Based on the Bayesian network, establish a probability graph model for the thermal runaway propagation path, and calculate the safety margin index and cascade failure risk entropy value of each node; Use the improved analytic hierarchy process (AHP) to determine the dynamic weights of key indicators, and map the BMS operation data to the assessment matrix through the real-time OPC interface; Use the fuzzy comprehensive evaluation method to calculate the system-level risk quantification index RQI = 0.7×(temperature anomaly degree / 280°C) + 0.3×(deformation coordination coefficient), and generate a risk probability cloud map by combining Monte Carlo sampling; Finally, compare the evaluation results with the UL 9540A standard value through cross-validation, establish a three-dimensional risk dashboard including 4-level warning thresholds, and realize the rolling update of the risk trajectory in the next 30s based on the LSTM time series prediction module to ensure that the prediction error rate of the risk quantification index ≤ 5%; 6. A method for fire ventilation and explosion relief safety assessment of a lithium battery energy storage system based on multi-dimensional simulation according to claim 1, characterized in that, In S5, it specifically includes the following sub-steps: Construct a comprehensive safety assessment model for the fire ventilation and explosion relief system according to the risk indicators, and quantitatively analyze the thermal runaway suppression effect and gas diffusion control ability; Optimize the layout parameters and triggering conditions of the explosion relief device, and verify its ability to quickly control the pressure after thermal runaway through fire simulation; Establish a dynamic response mechanism for the fire extinguishing system and ventilation equipment, and test their collaborative working efficiency; Based on the multi-dimensional performance verification results, form a hierarchical safety conclusion covering explosion relief protection, temperature control, and heat spread blockage.

7. A method for fire ventilation and explosion relief safety assessment of a lithium battery energy storage system based on multi-dimensional simulation according to claim 3, characterized in that, In S2, during the failure trajectory amplification process, introduce the constraint condition of the thermal runaway gas diffusion rate to construct the physical information loss function of the generative adversarial network, and generate a subset of failure scenarios that conform to Fourier's heat conduction law and Fick's diffusion law, ensuring that the spatio-temporal distributions of the temperature gradient and combustible gas concentration gradient in the amplified scenarios satisfy the actual physical laws.

8. A method for fire ventilation and explosion relief safety assessment of a lithium battery energy storage system based on multi-dimensional simulation according to claim 4, characterized in that In S3, during the moving mesh technology processing, establish a contact pressure feedback mechanism based on the cell expansion coefficient. When the cell volume expansion rate exceeds 15%, automatically trigger the local mesh encryption strategy, and introduce the strain rate-related Johnson-Cook plasticity model in the shell deformation calculation to realize the two-way coupling calculation of cell expansion and shell deformation.

9. A method for fire ventilation and explosion venting safety assessment of a lithium battery energy storage system based on multi-dimensional simulation according to claim 5, characterized in that In S4, when constructing the dynamic safety assessment matrix, embed a cross-dimensional feature fusion module based on the graph neural network, perform graph structure modeling on the spatial distribution features of the temperature field and the voltage time series features, extract the potential correlation features of thermal runaway propagation between cells through node embedding technology, and construct a three-dimensional assessment matrix including topological relationships.

10. A method for fire ventilation and explosion relief safety assessment of a lithium battery energy storage system based on multi-dimensional simulation according to claim 6, characterized in that, In S5, during the dynamic response mechanism test, construct a closed-loop verification system including pressure wavefront detection and gas component analysis. When the explosion relief valve acts, collect the shock wave propagation path data in real time, verify the matching degree of the explosion relief air flow field and the simulation results through particle image velocimetry technology, and dynamically correct the frequency conversion control curve of the ventilation system.

Images