Hu Zhi, Dong Anping, Du Dafan, Sun Dongke, Wang Donghong, Zhu Guoliang, Shu Da, Sun Baode
{"title":"铸流质量场和速度场的点阵玻尔兹曼模型","authors":"Hu Zhi, Dong Anping, Du Dafan, Sun Dongke, Wang Donghong, Zhu Guoliang, Shu Da, Sun Baode","doi":"10.15761/ams.1000140","DOIUrl":null,"url":null,"abstract":"The Lattice Boltzmann Method (LBM)-D2Q9 model is used to simulate velocity development and mass transfer of flows in casting. To quantify the basic flows in casting, stable flows in planes and pipes are simulated, which confirmed the LBM-D2Q9 model’s validation and numerical stability. Solute diffusion and vortex development are also investigated using LBM-D2Q9 model. The results show that the LBM model is capable to describe the velocity and solution field, which in a good match with the analytical calculations. *Correspondence to: Dong Anping, Shanghai Key Lab of advanced Hightemperature Materials and Precision Forming, School of Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, China, Tel: +86 13817882779; E-mail: apdong@sjtu.edu.cn Received: June 28, 2018; Accepted: July 20, 2018; Published: July 23, 2018 Introduction The ongoing demanding of advanced aero engines, which possess high thrust and lightweight, have caused a tremendous application of the near net shape forming technology of complex thin-wall superalloy casts [1]. During the casting, the solidification sequence, temperature and solute concentration distribution are affected by the complexity of geometry shape and thinness of the cast wall. These’re bringing a challenge for cast perfect forming and metallurgical quality improvement. It has been found that counter-gravity casting with additional pressure is more capable for complex thin wall cast near net shape forming than regular gravity casting [2-3]. During the pressured counter-gravity casting, forming and solidification are experiencing forced convection and constrained space condition. The mechanisms of melt flow and crystallization and the relation of microflows between dendrites and porosity suppression and microstructure evolution are complicated and have been a top focused area in the solidification researches [3-5]. Lattice Boltzmann method (LBM) has been proved that is an effective and powerful method to gain a numerical solution of Navier-Stokes equation [6], compared to other traditional numerical solutions of the Navier-Stokes equation, like Lax-Wendroff, MacCormack or SIMPLE method. To reveal the solidification microstructure evolution of superalloy complex thin-wall casting under complex constrained space and forced convective condition, simulations of the mass and heat transfer and distribution in this complex constrained cast is needed to carry out to understand the solidification condition. In the first step, it’s our goal to verify the LBGK model for representing the basic thermo-flow in the casting. Lattice Boltzmann modeling In this work, Lattice Boltzmann Method (LBM) is adapted to simulate fluid flow, solute and heat transfer. The LBM is a discrete approximation of Boltzmann equation, based on gas kinetic theory. The BGK approximation, proposed by Bhatnagar, Gross and Krook who replaced the collision term J(ff1) by a single relaxation time Ωf [7], has been widely accepted and utilized to solve Boltzmann equation. The Lattice BGK (LBGK) evolution equation can be described as: ( ) 1 , ( , ) ( , ) ( , ) ( , ) eq i i i i i i f x e t t t f x t f x t f x t F x t f τ + ∆ + ∆ − = − + (1) where, fi(x,t) is the discrete-velocity distribution function, it describes the density of particle with velocity ci at position and time (x,t),ei represents the discrete velocity space {e1,e2,...ei},Δt is the time step, τf is the relaxation time, ( , ) eq i f x t is the discrete equilibrium distribution function, ( , ) i F x t is the force term caused by physical field. The LBM also can be used to simulate the solute transport and heat transfer drive by a different mechanism such as diffusion and convection. Similar to the LBM for fluid flow, the solute distribution function ( , ) i g x t σ can be expressed as follow, using the passive scalar model [8]. , 1 ( ) ( , ) ( , ) ( , ) ( , ) eq i i i i i i g x e t t g x t g x t g x t G x t g σ σ σ σ σ τ + ∆ + ∆ − = − − + (2) where σ represents solute, τg is the relaxation time for the solute field, , ( , ) eq i g x t σ is the equilibrium distribution function for the solute field, ( , ) i G x t σ is the solute source term. Zhi H (2018) Lattice boltzmann modeling for mass and velocity fields of casting flows Adv Mater Sci, 2018 doi: 10.15761/AMS.1000140 Volume 3(1): 2-6 Results and Discussion Stable flows in planes and pipes When the melt forming in plane or pipe, stable flows can be achieved when casts are large enough. In present work, we simulated a typical plane flows by means of LBM and verified the results compared with an analytical solution and numerical stability in different meshes. As shown in the Figure 1, alloy melt is forming between two planes with distance h, assumed two planes have infinite width and length and the melt is incompressible viscous fluid. The upper plane is a velocity boundary with velocity U and the bottom plate is fixed. In this circumstance, the governing equation and its analytical solution are: 0 0 u V x ∂ ∇ ⋅ = = + ∂ ( ) 2 2 0 0 d u U or u y y h dy h = = ≤ ≤ Using the LBGK-D2Q9 model, the streamwise velocity distribution of a stable plane flow is simulated as shown in the Figure 2. Reynolds number is set to 100 assuming there is a stable flow. Fluid density ρ is set to unity and upper velocity U is 0.1 and the computation area are meshed by 156×156, 206×206 and 256×256 respectively. The colored velocity distribution suggested that the developed plane flow velocity differs in layers. The dimensionless velocity profile at the position of the middle x-axis is compared with the analytical solution, shown in the Figure 3a. The LBM results in a good agreement with the analytical solution, suggesting LBM is a validated model for simulating basic stable flows. In the Figure 3b, the results suggested that LBM in three different mesh have similar numerical stability. In the Figure 4, the velocity profile u = u(y) evolved from a shapely curve to a diagonal line as the timestep increased, suggesting the flow developed from unstable to stable flow. LBM is capable to simulate the dynamic process fluid flow in plane. The LBM for temperature is calculated using internal energy distribution function model [9]. The internal energy distribution function hi(x,t) is coupled by velocity distribution function fi(x,t), which can be written as: 1 ( ) ( , ) ( , ) ( , ) ( , ) eq i i i i i i h h x e t t h x t h x t h x t H x t τ + ∆ + ∆ − = − − + (3) where τh is the relaxation time for temperature field, ( , ) eq i h x t is the equilibrium distribution function, Hi(x,t) is the temperature source term. The two-dimensional D2Q9 model is chosen as the present discrete velocity model. Velocity space is discretized into a square lattice including nine discrete velocities ei, as shown as: where x c t ∆ = ∆ is the lattice speed, Δx is the lattice space, Δt is the time step. Related macroscopic variables such as density ρ, velocity u, concentration Cσ and temperature T, can be calculated from the relevant distribution functions as listed: 1 , , , 2 i i i i i i i i i a f u f e t C g T h σ ρ ρ = = + ∆ = = ∑ ∑ ∑ ∑ (4) The equilibrium distribution functions, which is related to the discrete velocity model, are defined as:","PeriodicalId":408511,"journal":{"name":"Advances in Materials Sciences","volume":"11 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1900-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Lattice boltzmann modeling for mass and velocity fields of casting flows\",\"authors\":\"Hu Zhi, Dong Anping, Du Dafan, Sun Dongke, Wang Donghong, Zhu Guoliang, Shu Da, Sun Baode\",\"doi\":\"10.15761/ams.1000140\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The Lattice Boltzmann Method (LBM)-D2Q9 model is used to simulate velocity development and mass transfer of flows in casting. To quantify the basic flows in casting, stable flows in planes and pipes are simulated, which confirmed the LBM-D2Q9 model’s validation and numerical stability. Solute diffusion and vortex development are also investigated using LBM-D2Q9 model. The results show that the LBM model is capable to describe the velocity and solution field, which in a good match with the analytical calculations. *Correspondence to: Dong Anping, Shanghai Key Lab of advanced Hightemperature Materials and Precision Forming, School of Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, China, Tel: +86 13817882779; E-mail: apdong@sjtu.edu.cn Received: June 28, 2018; Accepted: July 20, 2018; Published: July 23, 2018 Introduction The ongoing demanding of advanced aero engines, which possess high thrust and lightweight, have caused a tremendous application of the near net shape forming technology of complex thin-wall superalloy casts [1]. During the casting, the solidification sequence, temperature and solute concentration distribution are affected by the complexity of geometry shape and thinness of the cast wall. These’re bringing a challenge for cast perfect forming and metallurgical quality improvement. It has been found that counter-gravity casting with additional pressure is more capable for complex thin wall cast near net shape forming than regular gravity casting [2-3]. During the pressured counter-gravity casting, forming and solidification are experiencing forced convection and constrained space condition. The mechanisms of melt flow and crystallization and the relation of microflows between dendrites and porosity suppression and microstructure evolution are complicated and have been a top focused area in the solidification researches [3-5]. Lattice Boltzmann method (LBM) has been proved that is an effective and powerful method to gain a numerical solution of Navier-Stokes equation [6], compared to other traditional numerical solutions of the Navier-Stokes equation, like Lax-Wendroff, MacCormack or SIMPLE method. To reveal the solidification microstructure evolution of superalloy complex thin-wall casting under complex constrained space and forced convective condition, simulations of the mass and heat transfer and distribution in this complex constrained cast is needed to carry out to understand the solidification condition. In the first step, it’s our goal to verify the LBGK model for representing the basic thermo-flow in the casting. Lattice Boltzmann modeling In this work, Lattice Boltzmann Method (LBM) is adapted to simulate fluid flow, solute and heat transfer. The LBM is a discrete approximation of Boltzmann equation, based on gas kinetic theory. The BGK approximation, proposed by Bhatnagar, Gross and Krook who replaced the collision term J(ff1) by a single relaxation time Ωf [7], has been widely accepted and utilized to solve Boltzmann equation. The Lattice BGK (LBGK) evolution equation can be described as: ( ) 1 , ( , ) ( , ) ( , ) ( , ) eq i i i i i i f x e t t t f x t f x t f x t F x t f τ + ∆ + ∆ − = − + (1) where, fi(x,t) is the discrete-velocity distribution function, it describes the density of particle with velocity ci at position and time (x,t),ei represents the discrete velocity space {e1,e2,...ei},Δt is the time step, τf is the relaxation time, ( , ) eq i f x t is the discrete equilibrium distribution function, ( , ) i F x t is the force term caused by physical field. The LBM also can be used to simulate the solute transport and heat transfer drive by a different mechanism such as diffusion and convection. Similar to the LBM for fluid flow, the solute distribution function ( , ) i g x t σ can be expressed as follow, using the passive scalar model [8]. , 1 ( ) ( , ) ( , ) ( , ) ( , ) eq i i i i i i g x e t t g x t g x t g x t G x t g σ σ σ σ σ τ + ∆ + ∆ − = − − + (2) where σ represents solute, τg is the relaxation time for the solute field, , ( , ) eq i g x t σ is the equilibrium distribution function for the solute field, ( , ) i G x t σ is the solute source term. Zhi H (2018) Lattice boltzmann modeling for mass and velocity fields of casting flows Adv Mater Sci, 2018 doi: 10.15761/AMS.1000140 Volume 3(1): 2-6 Results and Discussion Stable flows in planes and pipes When the melt forming in plane or pipe, stable flows can be achieved when casts are large enough. In present work, we simulated a typical plane flows by means of LBM and verified the results compared with an analytical solution and numerical stability in different meshes. As shown in the Figure 1, alloy melt is forming between two planes with distance h, assumed two planes have infinite width and length and the melt is incompressible viscous fluid. The upper plane is a velocity boundary with velocity U and the bottom plate is fixed. In this circumstance, the governing equation and its analytical solution are: 0 0 u V x ∂ ∇ ⋅ = = + ∂ ( ) 2 2 0 0 d u U or u y y h dy h = = ≤ ≤ Using the LBGK-D2Q9 model, the streamwise velocity distribution of a stable plane flow is simulated as shown in the Figure 2. Reynolds number is set to 100 assuming there is a stable flow. Fluid density ρ is set to unity and upper velocity U is 0.1 and the computation area are meshed by 156×156, 206×206 and 256×256 respectively. The colored velocity distribution suggested that the developed plane flow velocity differs in layers. The dimensionless velocity profile at the position of the middle x-axis is compared with the analytical solution, shown in the Figure 3a. The LBM results in a good agreement with the analytical solution, suggesting LBM is a validated model for simulating basic stable flows. In the Figure 3b, the results suggested that LBM in three different mesh have similar numerical stability. In the Figure 4, the velocity profile u = u(y) evolved from a shapely curve to a diagonal line as the timestep increased, suggesting the flow developed from unstable to stable flow. LBM is capable to simulate the dynamic process fluid flow in plane. The LBM for temperature is calculated using internal energy distribution function model [9]. The internal energy distribution function hi(x,t) is coupled by velocity distribution function fi(x,t), which can be written as: 1 ( ) ( , ) ( , ) ( , ) ( , ) eq i i i i i i h h x e t t h x t h x t h x t H x t τ + ∆ + ∆ − = − − + (3) where τh is the relaxation time for temperature field, ( , ) eq i h x t is the equilibrium distribution function, Hi(x,t) is the temperature source term. The two-dimensional D2Q9 model is chosen as the present discrete velocity model. Velocity space is discretized into a square lattice including nine discrete velocities ei, as shown as: where x c t ∆ = ∆ is the lattice speed, Δx is the lattice space, Δt is the time step. Related macroscopic variables such as density ρ, velocity u, concentration Cσ and temperature T, can be calculated from the relevant distribution functions as listed: 1 , , , 2 i i i i i i i i i a f u f e t C g T h σ ρ ρ = = + ∆ = = ∑ ∑ ∑ ∑ (4) The equilibrium distribution functions, which is related to the discrete velocity model, are defined as:\",\"PeriodicalId\":408511,\"journal\":{\"name\":\"Advances in Materials Sciences\",\"volume\":\"11 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"1900-01-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Advances in Materials Sciences\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.15761/ams.1000140\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Advances in Materials Sciences","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.15761/ams.1000140","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
在这种情况下,控制方程及其解析解为:0 0 u V x∂∇⋅= = +∂()2 2 0 0 d u u或u y y h dy h = =≤≤利用LBGK-D2Q9模型,模拟稳定平面流动的顺流速度分布,如图2所示。假设有稳定的流动,雷诺数设为100。流体密度ρ设为单位,上流速U为0.1,计算区域分别用156×156、206×206和256×256进行网格划分。彩色速度分布表明,发育的平面流速在层间不同。将中间x轴位置的无量纲速度剖面与解析解进行对比,如图3a所示。LBM模型与解析解吻合较好,表明LBM模型是模拟基本稳定流动的有效模型。在图3b中,结果表明三种不同网格下的LBM具有相似的数值稳定性。在图4中,随着时间步长的增加,速度剖面u = u(y)由一条形状曲线演变为一条对角线,表明流动由不稳定向稳定发展。LBM能够模拟工艺流体在平面内的动态流动。温度的LBM采用内能分布函数模型[9]计算。的内部能量分布函数嗨(x, t)是由速度分布函数耦合fi (x, t),也可以写成:1 ( ) ( , ) ( , ) ( , ) ( , ) 情商我我我我我我h e t t h x t h x t h x t h tτ+∆+∆−=−−+(3)τh是温度场的弛豫时间,(,)情商h x t是平衡分布函数,你好(x, t)是温度源项。本文选择二维D2Q9模型作为离散速度模型。将速度空间离散成包含9个离散速度ei的方形晶格,如下所示:其中x c t∆=∆为晶格速度,Δx为晶格空间,Δt为时间步长。相关宏观变量如密度ρ、速度u、浓度Cσ、温度T等,可由相关分布函数计算得到:1,,,2 i i i i i i i i i ia f f f e T C g Th σ ρ ρ = = +∆= =∑∑∑∑∑(4)与离散速度模型相关的平衡分布函数定义为:
Lattice boltzmann modeling for mass and velocity fields of casting flows
The Lattice Boltzmann Method (LBM)-D2Q9 model is used to simulate velocity development and mass transfer of flows in casting. To quantify the basic flows in casting, stable flows in planes and pipes are simulated, which confirmed the LBM-D2Q9 model’s validation and numerical stability. Solute diffusion and vortex development are also investigated using LBM-D2Q9 model. The results show that the LBM model is capable to describe the velocity and solution field, which in a good match with the analytical calculations. *Correspondence to: Dong Anping, Shanghai Key Lab of advanced Hightemperature Materials and Precision Forming, School of Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, China, Tel: +86 13817882779; E-mail: apdong@sjtu.edu.cn Received: June 28, 2018; Accepted: July 20, 2018; Published: July 23, 2018 Introduction The ongoing demanding of advanced aero engines, which possess high thrust and lightweight, have caused a tremendous application of the near net shape forming technology of complex thin-wall superalloy casts [1]. During the casting, the solidification sequence, temperature and solute concentration distribution are affected by the complexity of geometry shape and thinness of the cast wall. These’re bringing a challenge for cast perfect forming and metallurgical quality improvement. It has been found that counter-gravity casting with additional pressure is more capable for complex thin wall cast near net shape forming than regular gravity casting [2-3]. During the pressured counter-gravity casting, forming and solidification are experiencing forced convection and constrained space condition. The mechanisms of melt flow and crystallization and the relation of microflows between dendrites and porosity suppression and microstructure evolution are complicated and have been a top focused area in the solidification researches [3-5]. Lattice Boltzmann method (LBM) has been proved that is an effective and powerful method to gain a numerical solution of Navier-Stokes equation [6], compared to other traditional numerical solutions of the Navier-Stokes equation, like Lax-Wendroff, MacCormack or SIMPLE method. To reveal the solidification microstructure evolution of superalloy complex thin-wall casting under complex constrained space and forced convective condition, simulations of the mass and heat transfer and distribution in this complex constrained cast is needed to carry out to understand the solidification condition. In the first step, it’s our goal to verify the LBGK model for representing the basic thermo-flow in the casting. Lattice Boltzmann modeling In this work, Lattice Boltzmann Method (LBM) is adapted to simulate fluid flow, solute and heat transfer. The LBM is a discrete approximation of Boltzmann equation, based on gas kinetic theory. The BGK approximation, proposed by Bhatnagar, Gross and Krook who replaced the collision term J(ff1) by a single relaxation time Ωf [7], has been widely accepted and utilized to solve Boltzmann equation. The Lattice BGK (LBGK) evolution equation can be described as: ( ) 1 , ( , ) ( , ) ( , ) ( , ) eq i i i i i i f x e t t t f x t f x t f x t F x t f τ + ∆ + ∆ − = − + (1) where, fi(x,t) is the discrete-velocity distribution function, it describes the density of particle with velocity ci at position and time (x,t),ei represents the discrete velocity space {e1,e2,...ei},Δt is the time step, τf is the relaxation time, ( , ) eq i f x t is the discrete equilibrium distribution function, ( , ) i F x t is the force term caused by physical field. The LBM also can be used to simulate the solute transport and heat transfer drive by a different mechanism such as diffusion and convection. Similar to the LBM for fluid flow, the solute distribution function ( , ) i g x t σ can be expressed as follow, using the passive scalar model [8]. , 1 ( ) ( , ) ( , ) ( , ) ( , ) eq i i i i i i g x e t t g x t g x t g x t G x t g σ σ σ σ σ τ + ∆ + ∆ − = − − + (2) where σ represents solute, τg is the relaxation time for the solute field, , ( , ) eq i g x t σ is the equilibrium distribution function for the solute field, ( , ) i G x t σ is the solute source term. Zhi H (2018) Lattice boltzmann modeling for mass and velocity fields of casting flows Adv Mater Sci, 2018 doi: 10.15761/AMS.1000140 Volume 3(1): 2-6 Results and Discussion Stable flows in planes and pipes When the melt forming in plane or pipe, stable flows can be achieved when casts are large enough. In present work, we simulated a typical plane flows by means of LBM and verified the results compared with an analytical solution and numerical stability in different meshes. As shown in the Figure 1, alloy melt is forming between two planes with distance h, assumed two planes have infinite width and length and the melt is incompressible viscous fluid. The upper plane is a velocity boundary with velocity U and the bottom plate is fixed. In this circumstance, the governing equation and its analytical solution are: 0 0 u V x ∂ ∇ ⋅ = = + ∂ ( ) 2 2 0 0 d u U or u y y h dy h = = ≤ ≤ Using the LBGK-D2Q9 model, the streamwise velocity distribution of a stable plane flow is simulated as shown in the Figure 2. Reynolds number is set to 100 assuming there is a stable flow. Fluid density ρ is set to unity and upper velocity U is 0.1 and the computation area are meshed by 156×156, 206×206 and 256×256 respectively. The colored velocity distribution suggested that the developed plane flow velocity differs in layers. The dimensionless velocity profile at the position of the middle x-axis is compared with the analytical solution, shown in the Figure 3a. The LBM results in a good agreement with the analytical solution, suggesting LBM is a validated model for simulating basic stable flows. In the Figure 3b, the results suggested that LBM in three different mesh have similar numerical stability. In the Figure 4, the velocity profile u = u(y) evolved from a shapely curve to a diagonal line as the timestep increased, suggesting the flow developed from unstable to stable flow. LBM is capable to simulate the dynamic process fluid flow in plane. The LBM for temperature is calculated using internal energy distribution function model [9]. The internal energy distribution function hi(x,t) is coupled by velocity distribution function fi(x,t), which can be written as: 1 ( ) ( , ) ( , ) ( , ) ( , ) eq i i i i i i h h x e t t h x t h x t h x t H x t τ + ∆ + ∆ − = − − + (3) where τh is the relaxation time for temperature field, ( , ) eq i h x t is the equilibrium distribution function, Hi(x,t) is the temperature source term. The two-dimensional D2Q9 model is chosen as the present discrete velocity model. Velocity space is discretized into a square lattice including nine discrete velocities ei, as shown as: where x c t ∆ = ∆ is the lattice speed, Δx is the lattice space, Δt is the time step. Related macroscopic variables such as density ρ, velocity u, concentration Cσ and temperature T, can be calculated from the relevant distribution functions as listed: 1 , , , 2 i i i i i i i i i a f u f e t C g T h σ ρ ρ = = + ∆ = = ∑ ∑ ∑ ∑ (4) The equilibrium distribution functions, which is related to the discrete velocity model, are defined as: