Calculation method of heat transfer of battery eruption particle deposition based on CFD-DEM coupling

Abstract

The present invention discloses a method for calculating the deposition and heat transfer of battery-emitted particles based on CFD-DEM coupling. This method simulates and evaluates the impact of the deposition and heat transfer of high-temperature particles generated by thermal runaway in lithium-ion batteries on the safety of thermal runaway in other batteries, targeting the multiphase flow behavior of high-temperature flue gas and particles generated by thermal runaway in lithium-ion batteries. First, the composition, mass flow rate, temperature, and other information of the battery-emitted flue gas and particles are experimentally determined. Then, the CFD method is used to simulate the diffusion of flue gas, obtain the concentration and density of flue gas at various locations, and the DEM discrete element method is used to simulate the collision and deposition behavior between particles and between particles and walls. The particle information is then transferred to the CFD calculation, and particle-wall contact judgment is introduced. Finally, the coupling force and heat transfer between particles and flue gas are calculated, and the flow field and particle state information are updated. The method simulates the temperature changes after particle deposition, including between particles, flow field, and battery components, to assess the risk of thermal runaway accidents in the battery system.

Description

Technical Field

[0001] The present application relates to the field of physical field coupling simulation technology, and in particular to a method for calculating the deposition and heat transfer of battery eruption particles based on CFD-DEM coupling. Background Art

[0002] Currently, power batteries, particularly lithium-ion batteries, are susceptible to thermal runaway accidents due to external factors such as mechanical, thermal, and electrical abuse, seriously endangering the lives and property of users. Thermal runaway is typically characterized by fumes generated by various internal thermal runaway reactions that break through the battery safety valve and release a large amount of particulate matter. These high-temperature particulate matter deposits on the surface of the battery system, transferring heat to the battery components and heating the battery, accelerating the spread of thermal runaway and triggering thermal runaway in other batteries, leading to cascading accidents.

[0003] Battery eruption particles are characterized by high velocity, high temperature, and a high proportion of metal elements. Previous simulations of battery eruption products have focused primarily on smoke diffusion, with little consideration given to the impact of particles on battery thermal runaway. Furthermore, because particles are discrete media, they diffuse in the smoke after exiting the battery and settle as the smoke moves. Traditional CFD methods for simulating continuous media struggle to accurately simulate particle behavior, and are unable to account for collisions and contact heat transfer between particles or between particles and the walls of the battery system, making it impossible to assess the risk of battery thermal runaway accidents caused by particles. Summary of the Invention

[0004] Based on the current technical problems raised in the background technology, the present invention provides a calculation method for the deposition and heat transfer of battery eruption particles based on CFD-DEM coupling. This method simulates and evaluates the impact of the deposition and heat transfer of high-temperature particles ejected from a battery during thermal runaway on the safety of other batteries, thereby assisting in the safe design of battery systems. The specific technical solution is as follows:

[0005] In a discrete element program, a particle model is constructed, specifying the shape, size, mechanical / thermal properties of the battery-emitted particles, as well as a contact model describing the inter-particle forces and the particle injection method. In a CFD program, the flow field geometry of the battery system is established, and a computational grid is divided corresponding to the battery-emitted flue gas domain. The component types and mass fractions, mass flow rate, temperature, and flue gas flow domain boundary pressure of the battery thermal runaway flue gas are specified.

[0006] The overall control equation of the CFD-DEM coupling of particulate matter and flue gas is as follows:

[0007] The mass continuity equation for smoke diffusion, assuming there is no mass source in the smoke diffusion:

[0008]

[0009] where ρ f is the smoke density, ε f represents the smoke fraction of the fluid, is the velocity of the flue gas.

[0010] Momentum conservation equation for smoke diffusion:

[0011]

[0012] Where P is the flue gas pressure, is the stress tensor of the flue gas, is the interaction force between smoke and particles, and g is the acceleration due to gravity.

[0013] Energy equation for smoke diffusion:

[0014]

[0015] Among them, c f is the constant pressure specific heat capacity of flue gas, T f is the flue gas temperature, λ f is the thermal conductivity of flue gas, q P is the heat exchange between flue gas and particulate matter.

[0016] Construct the motion and force balance equations of a single particle based on Newton's laws:

[0017]

[0018] Among them, m p is the mass of the particulate matter, is the velocity of the particles, is the contact and collision force between particles and between particles and the wall, is the force exerted by smoke on particulate matter.

[0019] The heat transfer equation for a single particle based on energy conservation:

[0020]

[0021] Among them, Q c is the heat transfer due to contact and collision between particles and between particles and the wall, Q f It is the convection heat transfer between particles and flue gas.

[0022] In CFD-DEM coupling, the characteristics of a particle population are reflected by tracking the changes in velocity, position, and temperature of a large number of particles. The coupling between smoke and particles is reflected through interphase interaction forces and convective heat transfer.

[0023] Since the flue gas emitted by battery thermal runaway contains multiple components such as CO, CO2, H2, and CH4, the components and mass fractions of the flue gas can be measured by using a meteorological mass spectrometer after collecting the flue gas. In the coupled calculation, the physical properties of the flue gas in each fluid calculation grid are calculated using a weighted calculation based on the mass fraction of each component:

[0024]

[0025] In the above formula, is the physical parameters of the flue gas at various locations in the flow field (specifically, the density, specific heat capacity, thermal conductivity, and viscosity of the flue gas), m i is the mass fraction of each component in the flue gas at each location, are the physical properties parameters of each component.

[0026] In the flow field, because particles actually occupy a certain volume, each CFD fluid calculation grid does not only contain smoke. Therefore, it is necessary to update and calculate the smoke gas volume fraction in each grid based on the particle position:

[0027]

[0028] in, is the total volume of particles in the CFD calculation grid, V cell To calculate the volume of the grid.

[0029] The contact model in the discrete element DEM program is deeply developed, and the coordinates of the contact point between the particle and the wall and the normal force of the collision between the particle and the wall in the contact model are used as the inherent properties of the particle.

[0030] In the discrete element program, the data structure describing the particles is expanded to add custom attribute values ​​of the particles: the spatial coordinates of the contact point between the particle and the wall and the normal force between the particle and the battery wall;

[0031] Based on the existing contact model (such as the Hertz-Mindlin model), the new custom attribute is coupled with the contact model. During the particle collision detection step in the discrete element program, if a particle collides with a wall, the spatial coordinates of the contact point and the normal force are calculated based on the contact model and the particle's geometric relationship, and the custom attribute value is updated. If the particle does not collide with the wall at the current time step, the attribute value is not updated and the attribute value updated at the last collision is retained.

[0032] In the coupled discrete element method and computational fluid dynamics (CFD) program, the particle data structure class transmitted by the discrete element method to the CFD program is expanded accordingly. This particle data structure contains information such as the particle center coordinates, particle radius, particle temperature, the coordinates of the particle-wall contact point, and the normal force acting on the particle-wall collision. These values ​​are unique to each particle, and each particle's information is represented as a specific object instance of the particle data structure class. The information for all particles is consolidated into an array-like structure and written to computer memory. The CFD program accesses this memory to obtain information on all particles in the flue gas flow domain during the CFD-DEM coupled time step.

[0033] In a CFD program, a global coordinate index table is constructed. The purpose of the index table is to find the CFD calculation grid index containing the given spatial coordinates of a certain point. The specific construction process is as follows:

[0034] In CFD calculations, after constructing the geometric model of the flue gas flow domain and battery area in the battery system and dividing the calculation grids. Traverse all the calculation grids, extract the center coordinates of each calculation grid, the spatial coordinates of each vertex of the calculation grid, the properties of the calculation grid (flue gas flow domain or battery solid area), and the calculation grid index, and merge them into a global coordinate index table. Given any spatial coordinate position, traverse the global coordinate index table, calculate the distance between the spatial coordinate position and the grid center, filter out a series of calculation grids with the closest distance, and then substitute the spatial coordinate position into the vertex coordinates of the calculation grid to determine whether the spatial coordinate position is inside the calculation grid. If it is inside a certain calculation grid, it means that the particle is within the spatial range of the calculation grid. According to the grid index value, find the grid in the CFD program.

[0035] From the DEM program, the coordinates of the particle's contact point with the wall and the particle's center are obtained. An index table is used to locate the computational grid containing the particle. Each mesh face in the computational grid is then traversed to identify the wall with which the particle collided, and the temperature of the collision face mesh is determined. It should be noted that the particle simulation time step in the discrete element program differs from the flue gas simulation time step in CFD. The flue gas simulation time step is approximately 10-50 times the particle simulation time step. Because the particle contact point attribute values ​​are updated only upon collision with a wall, the coordinates of the particle-wall collision point in the current CFD calculation step may not necessarily reflect the current collision point in the flue gas; they may reflect the collision coordinates in the previous discrete element time step. Therefore, a secondary collision check is required in the CFD program. The distance between the particle center and the wall collision point is calculated. If the distance exceeds the particle radius, the particle has not collided with the wall at the current moment. If the distance is less than or equal to the particle radius, the particle is considered to have contacted the wall.

[0036] In the CFD program, the index table is used to find the computational grid containing the particle based on the coordinates of the collision point between the particle and the wall and the coordinates of the particle center, and the temperature of the collision wall grid is obtained. The contact heat transfer coefficient of the particle impacting the wall is then calculated based on the physical and mechanical parameters of the particle. The calculation of the contact heat transfer coefficient between particles and between particles and the wall is based on contact models, experimental measurements, and empirical models. There are many forms of contact heat transfer coefficient calculation model methods. In this invention, the most widely used contact heat transfer coefficient calculation model is used as an example:

[0037] Q c =h c ΔT p1p2 (8)

[0038]

[0039] Among them, h c In order to consider the heat transfer coefficient of the contact heat transfer area, ΔT p1p2 is the temperature difference between particles and between particles and the wall, k p1 and k p2 are the thermal conductivity of the two particles or the particle and the wall, F N is the normal force on the contact point, E * is the equivalent elastic modulus of the two particles or the particle and the wall in collision contact, r * is the geometric mean radius of the particles calculated by contact theory. All three can be derived from the contact model (Heartz Mindlin) model.

[0040] The heat transfer caused by particles hitting the wall is calculated using the convective heat transfer formula. Based on the principle of energy conservation, the heat transfer is converted into a volume heat source of the wall grid and loaded into the wall unit of the CFD calculation. The calculation equation for the volume heat source is:

[0041]

[0042] The heat transfer amount is then written into the particle data structure class object as the heat dissipation of the particle itself.

[0043] In the CFD program, the interphase force and convective heat transfer between the flue gas and the particles are converted into volume force source terms and volume heat source terms and loaded onto the computational grid containing the particle. The flue gas force and convective heat transfer on the particle are then written into the particle data structure of the particle and transmitted to the DEM calculation. The flow field CFD calculation is performed at a preset time step t f , discrete element calculation is based on the time step t p , update the state information of the flow field and particles, and perform two-phase coupling calculations according to preset steps.

[0044] Among them, the interaction between particulate matter and smoke The force of smoke on particulate matter in formula (4) It is a pair of interacting forces, whose components include drag, buoyancy, lift, thermophoresis, etc. Among them, drag is the force with the highest proportion. Taking the drag force exerted by the flue gas on the particle as an example, the basic form of drag is shown in the following formula:

[0045]

[0046] In the above formula, C d is the drag coefficient, determined by experimental methods and the drag model, It is the difference between the particle velocity and the flue gas velocity at the location of the particle, that is, the relative velocity between the particle and the flue gas. p is the projected area of ​​the particle in the direction of relative velocity.

[0047] The selection of the drag model is related to the flow state of the smoke and the volume fraction of the smoke in the space. Taking the Gidaspow drag model as an example, the calculation of the drag coefficient depends on the characteristic parameter Reynolds number Re that characterizes the flow field state. p And volume share:

[0048]

[0049] For a single particle, the force exerted by the smoke on the particle, represented by the drag force, is directly loaded into the momentum conservation equation of the particle, while the reaction force of the drag force on the smoke needs to be converted into a volume momentum source term:

[0050]

[0051] The contact and collision forces between particles and between particles and walls are calculated using a discrete element contact model. This contact model provides the normal and tangential components of the force. The normal force represents the depth of particle collision and contact and is a key factor in calculating the contact heat transfer coefficient. Within the DEM, the appropriate contact model is selected based on differences in particle properties, such as whether viscosity, spherical particles, and particle breakage are considered.

[0052] The calculation of heat transfer between particles and smoke is related to the Reynolds number of smoke diffusion. For example, in the case of low Reynolds numbers, the Ranz model is used. The specific calculation method can also be measured through experiments. There are many forms. This paper takes the Ranz model as an example:

[0053] Q f =h f (T p -T f ) (14)

[0054]

[0055] In the above formula, Nu i is the Nusselt number, d i is the characteristic length of smoke flow, k f is the thermal conductivity, Pr is the Prandtl number, Re i is the Reynolds number. f is the heat transfer coefficient.

[0056] Similarly, the coupled heat transfer between the flue gas and the particles is numerically equal, but with opposite signs. This means that if the flue gas transfers heat to the particles, the particles are heated by the flue gas, while the flue gas itself introduces a negative energy source. The particle temperature is updated using an iterative solution based on Equation (5). For CFD calculations, the particle heat transfer also needs to be converted into a body heat source:

[0057]

[0058] Repeat this process until the preset simulation time is met, and evaluate the thermal runaway safety of the battery system based on the calculation results of the global temperature field of the battery system.

[0059] This paper, based on a coupled CFD-DEM approach, simulates the multiphase flow behavior of high-temperature particulate matter ejected during battery thermal runaway, dispersing in the flue gas and then spreading through the battery system. This overcomes the inability of existing CFD methods to directly simulate the safety hazards of high-temperature particulate matter deposition to battery systems. Based on discrete element simulation, this paper eliminates the need for excessive particle simplification, selects and calibrates different contact models, and considers the influence of particle shape and physical properties on particle deposition and heat transfer.

[0060] Another beneficial effect of the present invention is its scalability. By adding more particle properties, the method can account for other forces acting on particles in the flow field, such as electromagnetic forces. Detailed particle deposition information, such as the average temperature, particle size, and mass of deposited particles, can also be obtained to consider other particle-induced battery thermal safety accident mechanisms. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the figures required for use in the embodiments of the present application or related technical descriptions. Obviously, the figures described below are only some embodiments of the present application. For ordinary technicians in this field, other related figures can be obtained based on these figures without paying any creative work.

[0062] Figure 1This is a calculation flow chart of a method for heat transfer of battery eruption particles based on CFD-DEM coupling according to the present invention;

[0063] Figure 2 Schematic diagram of the particle-wall contact collision detection principle of the present invention;

[0064] Figure 3 Schematic diagram of particle heat transfer and volume heat source conversion described in the present invention. DETAILED DESCRIPTION

[0065] In order to better illustrate the battery eruption particle deposition and heat transfer method based on CFD-DEM coupling described in the present invention, an implementation example of the present invention is given below using the specific discrete element program DEM and the CFD program Fluent.

[0066] The specific calculation method flow chart of the present invention is as follows Figure 1 As shown:

[0067] Initialize the flow field in Fluent, set the inlet boundary conditions based on the measured flue gas mass flow rate, composition, and temperature, set the initial pressure, temperature, and gas composition of the flow field within the battery system, establish a particle model in EDEM, and set the mass flow rate, temperature, incident position, generation location, and velocity of the inlet particles based on the measured results. Create a grid index table in Fluent;

[0068] Start flow field calculation in Fluent and obtain particle information from EDEM (particle temperature, particle center position, coordinates of the particle-wall contact point, particle radius, particle velocity, and normal force between the particle and the wall). Prepare to calculate the coupled heat transfer between the particle and the battery wall, and the heat transfer and coupled force between the particle and the flue gas.

[0069] Unlike conventional CFD-DEM coupling methods, this method requires special calculation of the coupled heat transfer between the particles and the battery wall to achieve conjugate heat transfer between the particle temperature and the battery surface. The particle center position and contact point position are used to locate the CFD calculation grid where the particle is located according to the index table. The flue gas density, velocity, and temperature information of the grid are obtained, and contact judgment is performed. The heat transfer between the particle and the wall, the coupled force and heat transfer between the particle and the flue gas are calculated, and these are loaded into Fluent as heat source / momentum source items. The flow field calculation results are updated and the data is saved and passed to EDEM as the basis for updating the particle temperature and force state.

[0070] Start EDEM discrete element calculation. EDEM obtains the resultant force of the flow field acting on the particles and the total heat transferred by the particles from Fluent. It performs iterative particle dynamics calculation according to equations (4) and (5), calculates the collision force between particles and between particles, and updates the temperature, velocity, and spatial position of the particles. It then performs contact judgment and updates the center coordinates, velocity, temperature, and contact point coordinates of the particles.

[0071] EDEM updates the discrete element according to the time step tp, while the flow field is updated and iterated according to the time step tf, which is 10 to 100 tp. After multiple EDEM calculations and multiple particle state updates, the particle state data is passed to Fluent, and the Fluent flow field calculation is restarted;

[0072] Repeat steps 2 to 5 above until the convergence conditions and the preset simulation time scale are met.

[0073] The overall control equation of the CFD-DEM coupling of particulate matter and flue gas is as follows:

[0074] The mass continuity equation for smoke diffusion, assuming there is no mass source in the smoke diffusion:

[0075]

[0076] where ρ f is the smoke density, ε f represents the smoke fraction of the fluid, is the velocity of the flue gas.

[0077] Momentum conservation equation for smoke diffusion:

[0078]

[0079] Where P is the flue gas pressure, is the stress tensor of the flue gas, is the interaction force between smoke and particles, and g is the acceleration due to gravity.

[0080] Energy equation for smoke diffusion:

[0081]

[0082] Among them, c f is the constant pressure specific heat capacity of flue gas, T f is the flue gas temperature, λ f is the thermal conductivity of flue gas, q P is the heat exchange between flue gas and particulate matter.

[0083] Construct the motion and force balance equations of a single particle based on Newton's laws:

[0084]

[0085] Among them, m p is the mass of the particulate matter, is the velocity of the particles, is the contact and collision force between particles and between particles and the wall, is the force exerted by smoke on particulate matter.

[0086] The heat transfer equation for a single particle based on energy conservation:

[0087]

[0088] Among them, Q c is the heat transfer due to contact and collision between particles and between particles and the wall, Q f It is the convection heat transfer between particles and flue gas.

[0089] In CFD-DEM coupling, the characteristics of a particle population are reflected by tracking the changes in velocity, position, and temperature of a large number of particles. The coupling between smoke and particles is reflected through interphase interaction forces and convective heat transfer.

[0090] Since the flue gas emitted by battery thermal runaway contains multiple components such as CO, CO2, H2, and CH4, the components and mass fractions of the flue gas can be measured by using a meteorological mass spectrometer after collecting the flue gas. In the coupled calculation, the physical properties of the flue gas in each fluid calculation grid are calculated using a weighted calculation based on the mass fraction of each component:

[0091]

[0092] In the above formula, is the physical parameters of the flue gas at various locations in the flow field (specifically, the density, specific heat capacity, thermal conductivity, and viscosity of the flue gas), m i is the mass fraction of each component in the flue gas at each location, are the physical properties parameters of each component.

[0093] In the flow field, because particles actually occupy a certain volume, each CFD fluid calculation grid does not only contain smoke. Therefore, it is necessary to update and calculate the smoke gas volume fraction in each grid based on the particle position:

[0094]

[0095] in, is the total volume of particles in the CFD calculation grid, V cell To calculate the volume of the grid.

[0096] The contact model in the discrete element DEM program is deeply developed, and the coordinates of the contact point between the particle and the wall and the normal force of the collision between the particle and the wall in the contact model are used as the inherent properties of the particle.

[0097] In the discrete element program, the data structure describing the particles is expanded to add custom attribute values ​​of the particles: the spatial coordinates of the contact point between the particle and the wall and the normal force between the particle and the battery wall;

[0098] Based on the existing contact model (such as the Hertz-Mindlin model), the new custom attribute is coupled with the contact model. During the particle collision detection step in the discrete element program, if a particle collides with a wall, the spatial coordinates of the contact point and the normal force are calculated based on the contact model and the particle's geometric relationship, and the custom attribute value is updated. If the particle does not collide with the wall at the current time step, the attribute value is not updated and the attribute value updated at the last collision is retained.

[0099] In the coupled discrete element method and computational fluid dynamics (CFD) program, the particle data structure class transmitted by the discrete element method to the CFD program is expanded accordingly. This particle data structure contains information such as the particle center coordinates, particle radius, particle temperature, the coordinates of the particle-wall contact point, and the normal force acting on the particle-wall collision. These values ​​are unique to each particle, and each particle's information is represented as a specific object instance of the particle data structure class. The information for all particles is consolidated into an array-like structure and written to computer memory. The CFD program accesses this memory to obtain information on all particles in the flue gas flow domain during the CFD-DEM coupled time step.

[0100] In a CFD program, a global coordinate index table is constructed. The purpose of the index table is to find the CFD calculation grid index containing the given spatial coordinates of a certain point. The specific construction process is as follows:

[0101] In CFD calculations, after constructing the geometric model of the flue gas flow domain and battery area in the battery system and dividing the calculation grids. Traverse all the calculation grids, extract the center coordinates of each calculation grid, the spatial coordinates of each vertex of the calculation grid, the properties of the calculation grid (flue gas flow domain or battery solid area), and the calculation grid index, and merge them into a global coordinate index table. Given any spatial coordinate position, traverse the global coordinate index table, calculate the distance between the spatial coordinate position and the grid center, filter out a series of calculation grids with the closest distance, and then substitute the spatial coordinate position into the vertex coordinates of the calculation grid to determine whether the spatial coordinate position is inside the calculation grid. If it is inside a certain calculation grid, it means that the particle is within the spatial range of the calculation grid. According to the grid index value, find the grid in the CFD program.

[0102] From the DEM program, obtain the coordinates of the contact point between the particle and the wall and the coordinates of the particle center, and use the index table to find the computational grid containing the particle. Then traverse each mesh surface in the computational grid, find the wall surface that collides with the particle, and obtain the temperature of the collision surface mesh. It should be noted that in the discrete element program, the simulation time step of the particle is inconsistent with the time step of the flue gas simulation in CFD. The time step of the flue gas simulation is about 10-50 times the time step of the particle simulation. Since the contact point attribute value of the particle is updated once the particle collides with the wall. In the Fluent flow field calculation, the steps of the method for the contact between the particle and the battery wall are as follows: Figure 2 As shown, the coordinates of the particle-wall collision point in the current CFD calculation step are not necessarily the coordinates of the particle-wall collision point in the current flue gas; they may be the coordinates of the particle-wall collision point in the previous discrete element time step. Therefore, in the CFD program, a secondary collision detection between the particle and the wall is required. The distance between the particle center and the coordinates of the wall collision point is calculated. If the distance exceeds the particle radius, it means that the particle has not collided with the wall at the current moment. If it is less than or equal to the particle radius, the particle is judged to be in contact with the wall.

[0103] In the CFD program, the index table is used to find the computational grid containing the particle based on the coordinates of the collision point between the particle and the wall and the coordinates of the particle center, and the temperature of the collision wall grid is obtained. The contact heat transfer coefficient of the particle impacting the wall is then calculated based on the physical and mechanical parameters of the particle. The calculation of the contact heat transfer coefficient between particles and between particles and the wall is based on contact models, experimental measurements, and empirical models. There are many forms of contact heat transfer coefficient calculation model methods. In this invention, the most widely used contact heat transfer coefficient calculation model is used as an example:

[0104] Q c =h c ΔT p1p2 (twenty four)

[0105]

[0106] Among them, h c In order to consider the heat transfer coefficient of the contact heat transfer area, ΔT p1p2 is the temperature difference between particles and between particles and the wall, k p1 and k p2 are the thermal conductivity of the two particles or the particle and the wall, F N is the normal force on the contact point, E * is the equivalent elastic modulus of the two particles or the particle and the wall in collision contact, r *is the geometric mean radius of the particles calculated by contact theory. All three can be derived from the contact model (Heartz Mindlin) model.

[0107] The heat transfer caused by particles hitting the wall is calculated using the convective heat transfer formula. Based on the principle of energy conservation, the heat transfer is converted into a volume heat source of the wall grid and loaded into the wall unit of the CFD calculation. The calculation equation for the volume heat source is:

[0108]

[0109] The heat transfer amount is then written into the particle data structure class object as the heat dissipation of the particle itself.

[0110] In the CFD program, the interphase force and convective heat transfer between the flue gas and the particles are converted into volume force source terms and volume heat source terms and loaded onto the computational grid containing the particle. The flue gas force and convective heat transfer on the particle are then written into the particle data structure of the particle and transmitted to the DEM calculation. The flow field CFD calculation is performed at a preset time step t f , discrete element calculation is based on the time step t p , update the state information of the flow field and particles, and perform two-phase coupling calculations according to preset steps.

[0111] Among them, the interaction between particulate matter and smoke The force of smoke on particulate matter in formula (4) It is a pair of interacting forces, whose components include drag, buoyancy, lift, thermophoresis, etc. Among them, drag is the force with the highest proportion. Taking the drag force exerted by the flue gas on the particle as an example, the basic form of drag is shown in the following formula:

[0112]

[0113] In the above formula, C d is the drag coefficient, determined by experimental methods and the drag model, It is the difference between the particle velocity and the flue gas velocity at the location of the particle, that is, the relative velocity between the particle and the flue gas. p is the projected area of ​​the particle in the direction of relative velocity.

[0114] The selection of the drag model is related to the flow state of the smoke and the volume fraction of the smoke in the space. Taking the Gidaspow drag model as an example, the calculation of the drag coefficient depends on the characteristic parameter Reynolds number Re that characterizes the flow field state. p And volume share:

[0115]

[0116] For a single particle, the force exerted by the smoke on the particle, represented by the drag force, is directly loaded into the momentum conservation equation of the particle, while the reaction force of the drag force on the smoke needs to be converted into a volume momentum source term:

[0117]

[0118] The contact and collision forces between particles and between particles and walls are calculated using a discrete element contact model. This contact model provides the normal and tangential components of the force. The normal force represents the depth of particle collision and contact and is a key factor in calculating the contact heat transfer coefficient. Within the DEM, the appropriate contact model is selected based on differences in particle properties, such as whether viscosity, spherical particles, and particle breakage are considered.

[0119] The calculation of heat transfer between particles and smoke is related to the Reynolds number of smoke diffusion. For example, in the case of low Reynolds numbers, the Ranz model is used. The specific calculation method can also be measured through experiments. There are many forms. This paper takes the Ranz model as an example:

[0120] Q f =h f (T p -T f ) (30)

[0121]

[0122] In the above formula, Nu i is the Nusselt number, d i is the characteristic length of smoke flow, k f is the thermal conductivity, Pr is the Prandtl number, Re i is the Reynolds number. f is the heat transfer coefficient.

[0123] Similarly, the coupled heat transfer between the flue gas and the particles is numerically equal, but with opposite signs. This means that if the flue gas transfers heat to the particles, the particles are heated by the flue gas, while the flue gas itself introduces a negative energy source. The particle temperature is updated using an iterative solution based on Equation (5). For CFD calculations, the particle heat transfer also needs to be converted into a body heat source:

[0124]

[0125] Since there are many forms of heat transfer in the flow field of the battery system, namely, convection heat transfer between the battery surface and the flue gas, contact heat transfer between the particles and the wall, and conjugate heat transfer between the particles and the flue gas, there are significant differences in the heat transfer forms. Therefore, in the present invention, the heat transfer of the particles and the volume heat source of the battery system need to be specially processed, such as Figure 3As shown, follow these steps:

[0126] After the particles contact and collide with the battery surface according to the above content, for the case where the particles collide with the battery surface, first find the calculation unit where the contact point is located in the index table. Since the contact unit found in the index table is not necessarily a solid unit, if the unit is judged to be a non-solid unit, all neighboring units of the unit are traversed to determine whether it is a battery surface unit. If so, it is the solid unit of the battery surface that the particle contacts, and the battery wall temperature where the contact point is located is obtained.

[0127] Since there is convective heat exchange between the battery surface and the high-temperature flue gas, the present invention here needs to divide the heat transfer between the particles and the battery surface by the wall unit volume of the battery surface according to the principle of conservation of energy, and use it as the volume heat source in the battery wall unit, rather than destroying the original convective heat transfer conditions between the battery surface and the flue gas, or loading the heat source into the fluid unit of the flue gas.

[0128] According to the center position of the particle, the index table is used to find the fluid calculation unit where the particle is located, and the coupled heat transfer between the particle and the flue gas is calculated, converted into a volume heat source and added to the fluid unit.

[0129] When particles deposit, the greater the amount of deposition, the greater the normal force between the lower particles and the battery surface, the larger the contact area, and the significant increase in heat transfer. The forces acting on the particles are updated through the discrete element DEM program, and the temperature, pressure, and density information of the calculation domain are obtained through the CFD computational fluid dynamics method, to comprehensively assess the risk of particle deposition and heat transfer causing thermal runaway of the battery.

Claims

1. A CFD-DEM coupled method for calculating the deposition and heat transfer of battery eruption particles. This method simulates and evaluates the impact of the deposition and heat transfer of high-temperature particles emitted by thermal runaway lithium-ion batteries on the safety of thermal runaway of other batteries. The method is characterized by: The following steps are included: Step 1: In a discrete element program, a particle model is constructed to set the shape, size, mechanical / thermal properties of the battery-emitted particles, as well as the contact model and particle injection method that describe the forces between the particles. In a computational fluid dynamics (CFD) program, the flow field geometry of the battery system is established, and the corresponding computational grid is divided for the battery-emitted flue gas domain. The component types and mass fractions, mass flow rate, temperature, and flue gas flow domain boundary pressure of the battery-emitted flue gas due to thermal runaway are set. Step 2: Deeply develop the contact model in the discrete element (DEM) program, using the coordinates of the particle-wall collision point and the normal force of the particle-wall collision in the contact model as the particle's own properties. Once a particle collides with a wall, the collision point and normal contact force attribute values ​​are updated. These attribute values, along with the particle's center coordinates, particle radius, and particle temperature, are then transmitted to the CFD calculation program. Step 3: In the CFD program, a global coordinate index table is constructed. The index table is used to find the CFD calculation grid index containing the coordinates of a given point in space. In the CFD program, the index table is used to find the calculation grid containing the particle based on the coordinates of the collision point between the particle and the wall and the coordinates of the particle center, and the temperature of the collision wall grid is obtained. The contact heat transfer coefficient of the particle impacting the wall is then calculated using the physical and mechanical parameters of the particle. The heat transfer of the particle impacting the wall is calculated using the convective heat transfer formula. Based on the principle of conservation of energy, the heat transfer is converted into a volume heat source of the wall grid and loaded into the wall unit of the CFD calculation. Step 4: Obtain the flow velocity and temperature of the fluid calculation grid containing the particles in the CFD program, and calculate the interphase force and heat exchange between the flue gas and the particles with the velocity and temperature of the particles. These are loaded on the CFD calculation grid and the resultant force acting on the discrete element, respectively, and the flow field motion state is calculated and updated. At the same time, the position, temperature, and velocity information of the particles are also updated. Realize the coupling between CFD and DEM; Step 5: Calculate the flow field CFD according to the preset time step , discrete element calculation is based on the time step , update the state information of the flow field and particulate matter, and perform two-phase coupling calculations according to the preset steps. Repeat this process until the preset simulation time is met, and evaluate the thermal runaway safety of the battery system based on the calculation results.

2. The method according to claim 1, characterized in that In step 2, the development steps of the particle contact model are: In the discrete element program, the data structure describing the particles is expanded to add custom attribute values ​​of the particles: the spatial coordinates of the contact point between the particle and the wall and the normal force between the particle and the battery wall; Based on the original contact model, the coupling between the custom properties of the particle and the contact model is realized. In the particle collision detection step in the discrete element program, if a particle is found to collide with the wall, the spatial coordinates of the contact point and the normal force of the collision between the particle and the wall are calculated based on the geometric relationship between the contact model and the particle, and the custom property value is updated. If the particle does not collide with the wall in the current time step, the custom property value is not updated and the custom property value updated at the last collision is maintained. In the discrete element and computational fluid dynamics coupling program, the particle data structure class transmitted by the discrete element to the computational fluid dynamics program is expanded accordingly. The information in the particle data structure class includes the particle center coordinates, particle radius, particle temperature, the coordinates of the contact point of the particle collision with the wall, and the normal force of the particle collision with the wall; the numerical value of each particle is different, and the information of each particle is a specific object instance of the particle data structure class; the information of all particulate matter is merged into a class array and written into the computer memory. The CFD program obtains the information of all particles in the flue gas flow domain in the CFD-DEM coupling time step by accessing the memory.

3. The method according to claim 2, characterized in that In step 3, the CFD program implements the following steps to achieve heat transfer from particles to the battery wall: In CFD calculations, after constructing the geometric model of the flue gas flow area and battery area in the battery system and dividing the computational grids, all computational grids are traversed to extract the center coordinates of each computational grid, the spatial coordinates of each vertex of the computational grid, the attributes of the computational grid, and the computational grid index, and merge them into a global coordinate index table. Given any spatial coordinate position, traverse the global coordinate index table, calculate the distance between the spatial coordinate position and the grid center, filter out a series of computational grids with the closest distance, and then substitute the spatial coordinate position into the vertex coordinates of the computational grid to determine whether the spatial coordinate position is inside the computational grid. If it is inside a computational grid, it indicates that the particle is within the spatial range of the computational grid. According to the grid index value, find the grid in the CFD program; From the DEM program, obtain the coordinates of the contact point between the particle and the wall and the coordinates of the particle center, and use the index table to find the computational grid containing the particle; since the flow field calculation time step is longer than the particle discrete element calculation time step, and the custom properties of the particle-wall collision are only updated when a collision occurs, while the particle center position is updated in real time, when the CFD calculation starts, the state information of the particle may not completely match the real-time particle state, and contact judgment is required; traverse each particle and calculate the distance between the coordinates of the particle center and the wall collision point. If the distance exceeds the particle radius, it means that the particle has not collided with the wall at the current moment. If it is less than or equal to the particle radius, it is judged that the particle is in contact with the wall; then traverse each grid surface in the computational grid, find the wall surface that collides with the particle, and obtain the temperature of the collision surface grid; The contact heat transfer coefficient of the particles hitting the wall is then calculated using the physical and mechanical parameters of the particles, and the heat transfer of the particles hitting the wall is calculated using the convective heat transfer formula. Based on the principle of conservation of energy, the heat transfer is converted into a volume heat source of the wall grid and loaded into the wall unit of the CFD calculation to complete the calculation of the heat transfer from the particles to the wall. The heat transfer value is then written into the particle data structure class object as the heat dissipation of the particle itself.

4. The method according to claim 1, wherein In step 4, in addition to the heat transfer caused by the deposition of particles on the wall, there is also convective heat transfer between the particles and the flue gas. According to the global coordinate index table, the flue gas flow rate and temperature of the calculation grid where the particle is located are found, and then the phase force and heat transfer between the flue gas and the particle are calculated in combination with the temperature and velocity of the particle. In the CFD program, the phase force and convective heat transfer between the flue gas and the particle are converted into volume force source terms and volume heat source terms and loaded on the calculation grid containing the particle. Then, the flue gas force and convective heat transfer on the particle are written into the particle data structure of the particle and transmitted to the DEM calculation. The flow field CFD calculation is performed according to the preset time step. , discrete element calculation is based on the time step , update the state information of the flow field and particles, and perform two-phase coupling calculations according to preset steps; This process is repeated until the preset simulation time is met, and the thermal runaway safety of the battery system is evaluated based on the calculation results of the global temperature field.

Images