A lithium dendrite prediction and inhibition screening method of a multi-physical field coupling phase field model

Abstract

The application discloses a lithium dendrite prediction and inhibition screening method of a multi-physical field coupling phase field model, and comprises the following steps: establishing a coupling phase field model and a control equation group containing a phase field variable, an electric potential field, a lithium ion concentration field, a temperature field, a stress field and a flow field; calculating a thermal strain based on the temperature field and solving the stress field to obtain hydrostatic pressure; correcting a reaction rate in a Butler-Volmer reaction kinetics equation based on the hydrostatic pressure; solving the flow field by using a lattice Boltzmann method to obtain a flow velocity, feeding back the flow velocity to a Nernst-Planck equation to update a lithium ion concentration distribution; substituting the corrected reaction rate and the updated lithium ion concentration distribution into the coupling phase field model to solve and obtain a lithium dendrite morphology; changing a temperature gradient, a stacking pressure, a solute expansion coefficient and a gravity direction to obtain a plurality of dendrite morphology and field distribution results; extracting a dendrite morphology index to train a machine learning model, and screening a field condition for inhibiting lithium dendrite growth.

Description

Technical Field

[0001] This invention belongs to the field of lithium metal battery safety and multiphysics simulation technology, and particularly relates to a method for predicting and suppressing lithium dendrites in a multiphysics coupled phase field model. Background Technology

[0002] Lithium metal anodes possess high theoretical specific capacity and low electrochemical potential, making them important candidate anodes for high-energy-density batteries. However, during repeated charge-discharge cycles and rapid deposition, uncontrolled lithium dendrites can puncture the separator or solid electrolyte, leading to safety issues such as internal short circuits, capacity decay, and thermal runaway.

[0003] Phase-field methods can describe the lithium metal phase, electrolyte phase, and diffusion interface between them using continuous phase-field variables, thus avoiding explicit tracking of complex moving interfaces. They have been widely used to simulate lithium dendrite nucleation and morphology evolution. Existing phase-field models can typically describe the coupling relationship between phase-field variables, potential field, lithium ion concentration field, and dendrite morphology, providing an effective computational tool for predicting lithium dendrite growth morphology.

[0004] Existing documents disclose a method and system for predicting lithium dendrite growth morphology based on a nonlinear phase-field model. This method derives the governing equations for the phase-field variables, lithium-ion concentration field, and potential field based on the free energy variation law and reaction kinetics. The parameters and equations are then input into a finite difference program for transient solution, ultimately outputting the growth morphology of lithium dendrites on the negative electrode surface of a lithium battery during charging. This scheme can describe the lithium dendrite growth morphology using nonlinear continuous phase-field variables and, to some extent, reduces the computational complexity caused by solid-liquid interface tracking.

[0005] However, the aforementioned comparative documents primarily focus on establishing electrochemical phase-field models around phase-field variables, lithium-ion concentration fields, and potential fields. Their technical effectiveness mainly concentrates on nonlinear phase-field morphology prediction and reducing interface tracking complexity. They have not yet incorporated temperature, stress, and flow fields as computable coupled fields into the lithium dendrite prediction framework. Furthermore, they have not disclosed the technical solutions for feeding back temperature gradient-induced thermal strain and thermal stress, stacking pressure, and local hydrostatic pressure to the Butler-Volmer reaction kinetics via stress-coupled overpotential. Simultaneously, they have not disclosed the buoyancy-driven convection generated by the concentration gradient, solute expansion coefficient, and gravity direction, nor the technical solutions for feeding back to the lithium-ion mass transfer process via the Nernst-Planck convection term. Moreover, they have not combined high-throughput simulation data and machine learning models for rapid screening of dendrite suppression conditions. Therefore, these comparative documents fail to reveal the synergistic and competitive effects of thermal stress, stacking pressure, and buoyancy-driven convection on lithium dendrite tip growth, root stress concentration, and morphology smoothing.

[0006] Therefore, it is necessary to construct a multi-physics coupled phase-field model that can simultaneously describe temperature gradient, thermal stress, stacking pressure, solute expansion coefficient, gravity direction, and buoyancy-driven convection, and further combine high-throughput simulation and machine learning methods to achieve rapid prediction of lithium dendrite morphology indicators and screening of dendrite suppression field conditions. Summary of the Invention

[0007] To address the aforementioned technical problems, this invention proposes a method for predicting and suppressing lithium dendrites using a multi-physics coupled phase-field model, thereby resolving the issues present in the prior art.

[0008] To achieve the above objectives, this invention provides a method for lithium dendrite prediction and suppression screening using a multiphysics coupled phase-field model, comprising: A coupled phase-field model and governing equations were established, including phase field variables, electric potential field, lithium ion concentration field, temperature field, stress field and flow field. Thermal strain is calculated based on the temperature field, and hydrostatic pressure is obtained by solving the stress field based on the thermal strain. The reaction rate in the Butler-Volmer reaction kinetic equation was modified based on hydrostatic pressure; The flow field was solved using the lattice Boltzmann method to obtain the flow velocity, and the flow velocity was fed back into the Nernst-Planck equation to update the lithium ion concentration distribution. Substituting the corrected reaction rate and the updated lithium ion concentration distribution into the coupled phase-field model, the lithium dendrite morphology was obtained. Multiple sets of dendrite morphologies and field distribution results were obtained by changing the temperature gradient, stacking pressure, solute expansion coefficient and gravity direction; Dendrite morphology indices are extracted from dendrite morphology to train a machine learning model, and the trained machine learning model is used to screen field conditions that inhibit lithium dendrite growth.

[0009] Optionally, the coupled phase-field model is: In the formula, For phase field variables, It is a double-well function. The gradient energy coefficient, For stress-coupled overpotential, For local hydrostatic pressure, The molar volume of lithium metal. It is Faraday's constant. For phase field interpolation function, For interface migration rate, It is the electrochemical reaction constant. The charge transfer coefficient, To react to overpotential, denoted as dimensionless lithium-ion concentration, T as thermodynamic temperature in Kelvin, R as molar gas constant, and t as time.

[0010] Optionally, the set of governing equations includes an electric potential balance equation, a lithium ion concentration field governing equation, a heat conduction equation, and a mechanical equilibrium equation; the electric potential balance equation is established based on the principle of charge conservation, the lithium ion concentration field governing equation is established based on the Nernst-Planck equation, and the lithium ion concentration field governing equation includes diffusion terms, electromigration terms, convection terms, and reaction source terms.

[0011] Optionally, the process of correcting the reaction rate in the Butler-Volmer reaction kinetic equation based on hydrostatic pressure includes: The stress-coupled overpotential is calculated by dividing the product of local hydrostatic pressure and the molar volume of lithium metal by the Faraday constant. The stress-coupled overpotentials are superimposed onto the total overpotential; The reaction rate in the Butler-Volmer reaction kinetic equation is corrected by using the superimposed total overpotential.

[0012] Optionally, the process of obtaining the flow velocity by solving the flow field using the lattice Boltzmann method includes: The solute buoyancy term is calculated based on the lithium ion concentration difference, solute expansion coefficient, and gravity direction. Combine the solute buoyancy term with the interfacial resistance term to form a volume force term; The velocity field is obtained by processing the volume force terms using the lattice Boltzmann method.

[0013] Optionally, the process of feeding back the flow velocity to the Nernst-Planck equation to update the lithium-ion concentration distribution includes: inputting the flow velocity as a convection term into the Nernst-Planck equation and solving for the updated lithium-ion concentration distribution.

[0014] Optionally, the process of extracting dendrite morphology indicators from dendrite morphology to train a machine learning model includes: The maximum dendrite height and dendrite aspect ratio are extracted from the dendrite morphology and field distribution results as output labels. Temperature gradient, solute expansion coefficient, gravity direction, electrolyte Poisson's ratio and electrode Poisson's ratio are used as input features. The machine learning model is trained using the corresponding data of multiple sets of input features and output labels.

[0015] Optionally, the lattice Boltzmann method employs the D2Q9 velocity model to solve the electrolyte flow field through collision and migration processes.

[0016] Compared with the prior art, the present invention has the following advantages and technical effects: This invention establishes a multi-physics coupled phase-field model, incorporating phase field variables, electric potential field, lithium ion concentration field, temperature field, stress field, and flow field, achieving a unified description of lithium dendrite evolution under the coupling of thermo-mechanical-fluid-electrochemical fields. Based on the temperature field, thermal strain is calculated, and the hydrostatic pressure is obtained by introducing the stress field. This hydrostatic pressure is used to correct the reaction rate in the Butler-Volmer reaction kinetic equation, allowing for a quantitative feedback of the stress field's regulatory effect on dendrite growth. The flow field is solved using the lattice Boltzmann method, and the flow velocity is fed back to the Nernst-Planck equation to update the lithium ion concentration distribution, enabling a precise description of the influence of buoyancy-driven convection on ion mass transfer. By changing the temperature gradient, stacking pressure, solute expansion coefficient, and gravity direction, multiple sets of dendrite morphology and field distribution results are obtained. Combined with a machine learning model, rapid prediction of dendrite morphology indicators is achieved, providing an efficient technical means for screening multi-physics conditions to suppress lithium dendrite growth. Attached Figure Description

[0017] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart of the method for predicting and screening suppression conditions for lithium dendrite morphology using a multiphysics coupled phase-field model according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the lithium deposition structure according to an embodiment of the present invention. Detailed Implementation

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0020] Example 1 This invention aims to address the problem that existing single-physics-field phase-field models cannot uniformly describe the effects of temperature gradient, thermal stress, stacking pressure, solute expansion coefficient, and buoyancy-driven convection on lithium dendrite evolution, and provides a method for predicting lithium dendrite morphology and screening suppression conditions using a multi-physics-field coupled phase-field model.

[0021] The key to this invention lies in embedding thermal strain, stress-coupled overpotential, and buoyancy-driven convection as computable coupling terms into the lithium dendrite phase field model. Specifically, thermal strain couples the temperature gradient to the mechanical field, stress-coupled overpotential feeds back the stress field to the interfacial reaction kinetics, and buoyancy-driven convection feeds back the flow field velocity to the ion transport process.

[0022] To achieve the above objectives, this invention establishes a coupled set of control equations comprising phase field variables, electric potential field, lithium ion concentration field, temperature field, stress field, and flow field. Phase field variables... To distinguish between the lithium metal phase and the electrolyte phase, the electric potential field is solved by the charge conservation relation, the concentration field is established based on the Nernst-Planck equation, the temperature field is described by the heat conduction equation, the stress field is solved by the mechanical equilibrium equation, and the flow field is solved using the lattice Boltzmann method.

[0023] As shown in Figure 1, the multiphysics coupled phase-field model of this application is based on the electrochemical phase-field model, using the potential equation, lithium-ion transport equation, and phase-field equation as the fundamental evolution equations. Based on this, temperature-corrected diffusion coefficient, thermal strain, stress-coupled overpotential, buoyancy term, and flow velocity term are introduced to construct feedback paths for the thermal field, force field, and flow field on lithium dendrite evolution, respectively. Specifically, the thermal field influences the stress field and ion transport process through thermal strain and temperature-corrected diffusion coefficient; the force field generates stress-coupled overpotential and corrects interfacial reaction kinetics through hydrostatic pressure; and the flow field enters the lithium-ion transport equation through the buoyancy term and flow velocity term. The above calculation results further output dendrite morphology indicators and are used to construct a high-throughput simulation dataset and machine learning prediction model. Specifically, this includes: 1. Establishment of a multiphysics coupled phase-field model.

[0024] This embodiment first constructs a two-dimensional electrode-electrolyte interface computational domain. The computational domain includes the liquid electrolyte, lithium metal anode, lithium ions, and anions. Phase field variables are then utilized. Describing material phase: It is a lithium metal phase. It is an electrolyte phase. For diffusion interface.

[0025] An electrochemical-mechanical-thermal-fluid coupled phase-field model is established, incorporating phase-field variables, electric potential field, lithium-ion concentration field, temperature field, stress field, and flow field; the total Gibbs free energy functional expression is: This includes contributions from chemical potential, electrostatic energy, and interfacial energy.

[0026] The electric potential equation is derived from the principle of charge conservation, and the expression for the electric potential balance equation is: ; in, For phase field variables, For the first Concentration of each component For electric potential, For phase field interpolation function, and These are the free energy densities of the lithium metal phase and the electrolyte phase, respectively. It is a double-well function. The gradient energy coefficient; For effective conductivity, It is Faraday's constant. This represents the lithium metal site density.

[0027] The coupling relationships between the fields are as follows: the concentration field is described by the Nernst-Planck equations, which include diffusion, electromigration, and reaction source terms. The temperature field describes heat conduction and local temperature changes caused by heat sources; the stress field describes stacking pressure, thermal strain, and local stress caused by deposition; and the flow field describes buoyancy-driven convection generated by the concentration difference.

[0028] Thermo-mechanical coupling is achieved through thermal strain, which can be expressed as: in, For local temperature, For reference temperature, This is the effective coefficient of thermal expansion. The thermal strain is incorporated into the elastic strain decomposition relationship and updates the stress field.

[0029] The force-electrochemical coupling is achieved through a stress-coupled overpotential, which is: in, For local hydrostatic pressure, The molar volume of lithium metal. This is the Faraday constant. This term is superimposed on the total overpotential and corrects the Butler-Volmer reaction kinetics.

[0030] The fluid-electrochemical coupling is achieved through the Nernst-Planck equation containing a convection term: in, The velocity is obtained using the lattice Boltzmann method. The solute buoyancy term can be expressed as... It is used to describe the combined effect of concentration gradient, solute expansion coefficient and gravity direction on natural convection.

[0031] The phase-field equation in the electrochemical-mechanical-thermal-fluid coupled phase-field model is expressed as follows: In the formula, For phase field variables, It is a double-well function. The gradient energy coefficient, For stress-coupled overpotential, For local hydrostatic pressure, The molar volume of lithium metal. It is Faraday's constant. For phase field interpolation function, For interface migration rate, It is the electrochemical reaction constant. The charge transfer coefficient, To react to overpotential, denoted as dimensionless lithium-ion concentration, T as thermodynamic temperature in Kelvin, R as molar gas constant, and t as time, i.e., the sequence parameter changes with time.

[0032] 2. Thermal stress and buoyancy drive convection coupling.

[0033] In this embodiment, the temperature field is used to calculate the thermal strain, which is then used to obtain the stress field after entering the mechanical equilibrium solution process. The hydrostatic pressure is extracted from the stress field, and the stress-coupled overpotential is calculated to correct the interfacial reaction rate.

[0034] Simultaneously, the buoyancy term is calculated based on the lithium-ion concentration difference, solute expansion coefficient, and gravity direction, and the flow field volume force term is obtained by combining it with the interfacial resistance term. After obtaining the velocity field using the lattice Boltzmann method, the velocity is used as a convection feedback quantity input into the lithium-ion transport process.

[0035] 3. Numerical solution and topography output.

[0036] In this embodiment, the electrochemical-mechanical-thermal-fluid coupled phase-field model is solved transiently according to the following steps.

[0037] A1: Initialize the two-dimensional electrode-electrolyte interface computational domain, and set the computational domain size, grid spacing, time step, initial distribution of phase field variables, initial distribution of lithium ion concentration, potential boundary conditions, temperature boundary conditions, mechanical boundary conditions, and flow field boundary conditions.

[0038] A2: Calculate the interpolation function based on the current phase field variables, and update the effective diffusion coefficient, effective electrical conductivity, effective thermal conductivity, effective heat capacity, effective density, effective thermal expansion coefficient, and effective elastic parameters.

[0039] A3: Solve the potential control equation and the lithium ion concentration field control equation to obtain the current potential distribution, lithium ion concentration distribution, and interface reaction source term.

[0040] A4: Based on the temperature field, calculate the temperature gradient and thermal strain, and introduce the thermal strain and intrinsic strain caused by deposition into the mechanical equilibrium equation to solve the displacement field and stress field, and extract the local hydrostatic pressure and von Mises stress.

[0041] A5: Calculate the stress-coupled overpotential based on the local hydrostatic pressure, and then add this stress-coupled overpotential to the total overpotential to update the Butler-Volmer reaction overpotential term, thereby correcting the reaction rate at the lithium deposition interface. Figure 2 As shown.

[0042] A6: Calculate the solute buoyancy term based on the lithium ion concentration deviation, solute expansion coefficient, and gravity direction, and construct the force term in the lattice Boltzmann method by combining it with the interface resistance term; use the D2Q9 model to perform the collision and migration process to obtain the current velocity field.

[0043] A7: Feed the velocity field obtained in step A6 back to the Nernst-Planck equation containing the convection term to update the lithium-ion transport flux and concentration distribution.

[0044] A8: Based on the updated potential field, concentration field, temperature field, stress field, and flow field, solve the phase field evolution equation to obtain the lithium dendrite morphology at the next moment.

[0045] A9: Repeat steps A2 to A8 until the preset termination time is reached or the dendrite height reaches the preset threshold; output the maximum dendrite height, dendrite aspect ratio, concentration distribution, temperature distribution, hydrostatic pressure distribution, von Mises stress distribution, and flow velocity distribution.

[0046] 4. Database establishment and suppression condition filtering.

[0047] In this embodiment, by changing input parameters such as temperature gradient, stacking pressure, solute expansion coefficient, gravity direction, electrolyte Poisson's ratio, and lithium metal Poisson's ratio, dendrite morphology and field distribution results are obtained in batches.

[0048] The maximum dendrite height, dendrite aspect ratio, stress concentration index, and tip ion enrichment index are extracted from each sample and used as output labels for the prediction model; the corresponding field conditions are used as input features to train the machine learning model.

[0049] Based on the simulation results, dendrite growth zones were established to identify growth-promoting zones corresponding to negative temperature gradients and strong convection, as well as growth-inhibiting zones corresponding to positive temperature gradients, appropriate stacking pressure, and low convection intensity.

[0050] Compared with existing technologies, the present invention has the following advantages: First, it can describe the effects of electrochemical, mechanical, thermal and fluid fields on lithium dendrite evolution within a unified phase field framework; second, it can reveal the regulatory effects of thermal stress, stacking pressure and buoyancy-driven convection on dendrite tip growth, root stress concentration and morphology smoothing; third, it can reduce the cost of repeated simulations through high-throughput simulation and machine learning, and quickly screen field conditions that are conducive to dendrite suppression, such as positive temperature gradient, appropriate thermal stress, appropriate stacking pressure and low convection intensity.

[0051] Example 2 This invention provides a method for predicting and screening suppression conditions for lithium dendrite morphology based on a multi-physics coupled phase-field model. The method constructs an electrochemical-mechanical-thermal-fluidic coupled phase-field model incorporating phase-field variables, potential field, lithium-ion concentration field, temperature field, stress field, and flow field. Thermal strain generated by the temperature gradient is introduced into the stress balance equation, and the Butler-Volmer reaction kinetics are corrected by stress-coupled overpotential. Simultaneously, a convection term is introduced into the Nernst-Planck ion transport equation, and the lattice Boltzmann method is used to solve for the buoyancy-driven convection generated by the concentration gradient and solute expansion coefficient. The lattice Boltzmann method employs the D2Q9 velocity model, solving the electrolyte flow field through collision and migration processes, and feeding the obtained flow velocity back into the Nernst-Planck ion transport equation. By setting temperature gradient, stacking pressure, solute expansion coefficient, gravity direction, thermal stress intensity coefficient, and convection intensity coefficient, the morphology, maximum dendrite height, and aspect ratio of lithium dendrites are obtained. A machine learning model is then trained using high-throughput simulation data to screen for field conditions that inhibit dendrite growth. This invention can uniformly describe the synergistic and competitive effects of thermal stress, stacking pressure, and buoyancy-driven convection on lithium dendrite evolution, providing a predictive method for the safe operation and interface control of lithium metal batteries.

[0052] As a specific implementation method of this embodiment, the specific implementation process includes: Step 1: Construct the computational domain of the lithium metal battery electrode-electrolyte interface using phase field variables. Characterize the lithium metal phase, electrolyte phase, and diffusion interface between them.

[0053] Furthermore, in step 1, the phase field variable satisfy: Lithium metal electrode phase, Indicates the electrolyte phase. This represents the solid-liquid diffusion interface.

[0054] Step 2: Establish a multi-field coupled control equation system with phase field variables, electric potential field, lithium ion concentration field, temperature field, stress field, and flow field as state variables. The phase field variables are used to characterize the lithium metal phase, electrolyte phase, and diffusion interface. The electric potential field is used to characterize the interface charge conservation. The lithium ion concentration field is used to characterize ion transport under the combined effects of diffusion, electromigration, convection, and interface reaction. The temperature field, stress field, and flow field are used to characterize the effects of temperature gradient, mechanical constraints, and buoyancy-driven convection on dendrite evolution, respectively.

[0055] Step 3: Introduce the thermal strain generated by the temperature gradient and the external stacking pressure into the mechanical equilibrium solution process to obtain the stress field, hydrostatic pressure and von Mises stress distribution.

[0056] Furthermore, in step 3, the thermal strain By temperature Compared with reference temperature Calculate the difference between them, and satisfy the following conditions: .

[0057] Step 4: Introduce stress-coupled overpotential into the Butler-Volmer equation to adjust the reaction driving force of the lithium deposition interface.

[0058] The formula for calculating stress-coupled overpotential is: .

[0059] Furthermore, in step 4, the stress-coupled overpotential is determined by the hydrostatic pressure and the molar volume of lithium metal, and is superimposed on the total overpotential to correct the lithium deposition reaction rate.

[0060] Step 5: Characterized in that the lithium-ion transport process includes at least diffusion, electromigration, interfacial reaction source term and convection term, wherein the flow rate is provided by buoyancy-driven convection.

[0061] Furthermore, in step 5, the buoyancy-driven convection is driven by a volumetric force, which includes an interfacial resistance term and a solute buoyancy term caused by the concentration difference. The solute buoyancy term is related to the solute expansion coefficient, the direction of gravitational acceleration, and the concentration gradient.

[0062] Step 6: Set the temperature gradient, stacking pressure, solute expansion coefficient, gravity direction, thermal stress intensity coefficient and convection intensity coefficient, and obtain the lithium dendrite morphology, maximum dendrite height, dendrite aspect ratio, concentration distribution, temperature distribution, stress distribution and flow velocity distribution by transient solution.

[0063] Furthermore, in step 6, the temperature gradient includes a positive temperature gradient and a negative temperature gradient; the positive temperature gradient is used to enhance thermal stress and weaken buoyancy-driven convection, thereby suppressing the growth of lithium dendrite tips.

[0064] Furthermore, in step 6, the stacking pressure is used to reduce the overpotential driving force at the dendrite tip through stress-coupled overpotential, thereby promoting the transformation of lithium deposition from a sharp dendrite morphology to a smooth morphology.

[0065] Furthermore, in step 6, the solute expansion coefficient is a physical quantity used to characterize the diffusion ability of a solute in a solvent. A larger solute expansion coefficient enhances buoyancy-driven convection, while a smaller solute expansion coefficient reduces the intensity of directional mass transfer at the tip.

[0066] As shown in Figure 2, the computational domain includes the lithium metal electrode, the solid-liquid diffusion interface, and the electrolyte region. Phase field variables... Indicates the lithium metal electrode phase. Indicates the electrolyte phase. This represents the diffusion interface between lithium metal and the electrolyte. Lithium ions migrate in the electrolyte region and undergo reduction deposition at the interface; the morphology of lithium dendrites is characterized by the spatial evolution of phase field variables.

[0067] Step 7: Construct a high-throughput dendrite dataset based on simulation results under different field conditions, train a machine learning prediction model, and screen multi-physics field control conditions to suppress lithium dendrite growth based on the prediction results.

[0068] Furthermore, the input features of the machine learning prediction model include temperature gradient, solute expansion coefficient, gravity direction, electrolyte Poisson's ratio, and electrode Poisson's ratio, while the output labels include maximum dendrite height and dendrite aspect ratio.

[0069] Example 3 This embodiment also provides a lithium dendrite morphology prediction and suppression condition screening system based on a multiphysics coupled phase-field model, including: a model building module, a numerical solution module, a morphology index extraction module, a machine learning prediction module, and a mechanism diagram screening module; the model building module is used to establish an electrochemical-mechanical-thermal-fluid coupled phase-field model, the numerical solution module is used to transiently solve the phase-field variables, electric potential field, concentration field, temperature field, stress field, and flow field, and the machine learning prediction module is used to predict the maximum dendrite height and dendrite aspect ratio.

[0070] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described thereon.

[0071] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for predicting and suppressing lithium dendrites using a multiphysics coupled phase-field model, characterized in that, Includes the following steps: A coupled phase-field model and governing equations were established, including phase field variables, electric potential field, lithium ion concentration field, temperature field, stress field and flow field. Thermal strain is calculated based on the temperature field, and hydrostatic pressure is obtained by solving the stress field based on the thermal strain. The reaction rate in the Butler-Volmer reaction kinetic equation was modified based on hydrostatic pressure; The flow field was solved using the lattice Boltzmann method to obtain the flow velocity, and the flow velocity was fed back into the Nernst-Planck equation to update the lithium ion concentration distribution. Substituting the corrected reaction rate and the updated lithium ion concentration distribution into the coupled phase-field model, the lithium dendrite morphology was obtained. Multiple sets of dendrite morphologies and field distribution results were obtained by changing the temperature gradient, stacking pressure, solute expansion coefficient and gravity direction; Dendrite morphology indices are extracted from dendrite morphology to train a machine learning model, and the trained machine learning model is used to screen field conditions that inhibit lithium dendrite growth.

2. The method for lithium dendrite prediction and suppression screening using a multiphysics coupled phase-field model according to claim 1, characterized in that, The coupled phase-field model is as follows: In the formula, For phase field variables, It is a double-well function. The gradient energy coefficient, For stress-coupled overpotential, For local hydrostatic pressure, The molar volume of lithium metal. It is Faraday's constant. For phase field interpolation function, For interface migration rate, It is the electrochemical reaction constant. The charge transfer coefficient, To react to overpotential, denoted as dimensionless lithium-ion concentration, T as thermodynamic temperature in Kelvin, t as time, and R as molar gas constant.

3. The method for lithium dendrite prediction and suppression screening using a multiphysics coupled phase-field model according to claim 1, characterized in that, The governing equations include the potential balance equation, the lithium ion concentration field governing equation, the heat conduction equation, and the mechanical balance equation. The potential balance equation is based on the principle of charge conservation, and the lithium ion concentration field governing equation is based on the Nernst-Planck equation. The lithium ion concentration field governing equation includes diffusion terms, electromigration terms, convection terms, and reaction source terms.

4. The method for lithium dendrite prediction and suppression screening using a multiphysics coupled phase-field model according to claim 1, characterized in that, The process of correcting the reaction rate in the Butler-Volmer reaction kinetic equation based on hydrostatic pressure includes: The stress-coupled overpotential is calculated by dividing the product of local hydrostatic pressure and the molar volume of lithium metal by the Faraday constant. The stress-coupled overpotentials are superimposed onto the total overpotential; The reaction rate in the Butler-Volmer reaction kinetic equation is corrected by using the superimposed total overpotential.

5. The method for lithium dendrite prediction and suppression screening using a multiphysics coupled phase-field model according to claim 1, characterized in that, The process of obtaining the flow velocity by solving the flow field using the lattice Boltzmann method includes: The solute buoyancy term is calculated based on the lithium ion concentration difference, solute expansion coefficient, and gravity direction. Combine the solute buoyancy term with the interfacial resistance term to form a volume force term; The velocity field is obtained by processing the volume force terms using the lattice Boltzmann method.

6. The method for lithium dendrite prediction and suppression screening using a multiphysics coupled phase-field model according to claim 1, characterized in that, The process of feeding back the flow velocity into the Nernst-Planck equation to update the lithium-ion concentration distribution includes: inputting the flow velocity as a convection term into the Nernst-Planck equation, and solving for the updated lithium-ion concentration distribution.

7. The method for lithium dendrite prediction and suppression screening based on a multiphysics coupled phase-field model according to claim 1, characterized in that, The process of extracting dendrite morphology indicators from dendrite morphology to train a machine learning model includes: The maximum dendrite height and dendrite aspect ratio are extracted from the dendrite morphology and field distribution results as output labels. Temperature gradient, solute expansion coefficient, gravity direction, electrolyte Poisson's ratio and electrode Poisson's ratio are used as input features. The machine learning model is trained using the corresponding data of multiple sets of input features and output labels.

8. The method for lithium dendrite prediction and suppression screening using a multiphysics coupled phase-field model according to claim 1, characterized in that, The lattice Boltzmann method employs the D2Q9 velocity model to solve the electrolyte flow field through collision and migration processes.

Images