# CE展开:Navier-Stokes方程 文档B站视频讲解: > [LBM求解Navier-Stokes方程的Chapman-Enskog展开(CE展开)](https://www.bilibili.com/video/BV1EuibBiEYv/) **红色为相对于对流扩散方程增加的部分关键内容** ## LBM求解Navier-Stokes方程模型介绍 > 演化方程: ```{math} :label: CE-ns-eq1 f_i(\boldsymbol{x}+\boldsymbol{e}_i\delta _t,t+\delta _t)-f_i(\boldsymbol{x},t)=-\frac{1}{\tau}\left[ f_i(\boldsymbol{x},t)-f_{i}^{\mathrm{eq}}(\boldsymbol{x},t) \right] ``` > 平衡态分布函数: ```{math} :label: CE-ns-eq2 f_{i}^{\mathrm{eq}}(\boldsymbol{x},t)=w_i \rho \left[ 1+\frac{\boldsymbol{e}_{\boldsymbol{i}}\cdot \boldsymbol{u}}{c_{s}^{2}} +\frac{\left( \boldsymbol{e}_{\boldsymbol{i}}\cdot \boldsymbol{u} \right) ^2}{2c_{s}^{4}}-\frac{\boldsymbol{u}\cdot \boldsymbol{u}}{2c_{s}^{2}} \right] ``` > 宏观量计算: ```{math} :label: CE-ns-eq3 \rho=\sum_i{f_i}, \quad \rho \boldsymbol{u}=\sum_i{\boldsymbol{e}_i f_i} ``` **注:由平衡态分布函数可以得到** ```{math} :label: CE-ns-eq0 \sum_i{f_{i}^{\mathrm{eq}}}=\rho, \quad \sum_i{\boldsymbol{e}_{\boldsymbol{i}}f_{i}^{\mathrm{eq}}}=\rho \boldsymbol{u}, \quad \sum_i{\boldsymbol{e}_{\boldsymbol{i}}\boldsymbol{e}_{\boldsymbol{i}}f_{i}^{\mathrm{eq}}}=\Pi^{(0)}= \rho c_{s}^{2}\mathbf{I}+\rho \boldsymbol{u}\boldsymbol{u} ``` 其中 $\mathbf{I}$ 为单位张量,记$p=\rho c_s^2$,最后会看到$p$即为流体的静压力。 ## CE展开的基本假设 $f_i$可以展开为如下的形式: ```{math} :label: CE-ns-eq4 f_i=f_{i}^{\mathrm{eq}}+\varepsilon f_{i}^{(1)}+\varepsilon ^2f_{i}^{(2)}+\cdots ``` 时间导数可以展开为如下的形式: ```{math} :label: CE-ns-eq5 \frac{\partial}{\partial t}=\varepsilon \frac{\partial}{\partial t_1}+\varepsilon ^2\frac{\partial}{\partial t_2}+\cdots ``` 空间导数则只有一阶展开: ```{math} :label: CE-ns-eq6 \nabla =\varepsilon \nabla _1 ``` ## CE展开 将式{eq}`CE-ns-eq1`左侧进行泰勒展开得到: ```{math} :label: CE-ns-eq7 \delta _t\boldsymbol{D}_{\boldsymbol{i}}f_i+\frac{\delta _{t}^{2}}{2}\left( \boldsymbol{D}_{\boldsymbol{i}} \right) ^2f_i=-\frac{1}{\tau}\left[ f_i\left( \boldsymbol{x},t \right) -f_{i}^{\mathrm{eq}}\left( \boldsymbol{x},t \right) \right] ``` 其中$\boldsymbol{D}_i=\left( \partial_t+\boldsymbol{e}_i \cdot \nabla \right)$,将式{eq}`CE-ns-eq4`,{eq}`CE-ns-eq5`和{eq}`CE-ns-eq6`代入上式得到: ```{math} :label: CE-ns-eq8 \left( \varepsilon \partial _{t_1}+\varepsilon ^2\partial _{t_2}+\varepsilon \boldsymbol{e}_{\boldsymbol{i}}\cdot \nabla _1 \right) \left( f_{i}^{\mathrm{eq}}+\varepsilon f_{i}^{\left( 1 \right)}+\varepsilon ^2f_{i}^{\left( 2 \right)} \right) \\ +\frac{\delta _t}{2}\left( \varepsilon \partial _{t_1}+\varepsilon ^2\partial _{t_2}+\varepsilon \boldsymbol{e}_{\boldsymbol{i}}\cdot \nabla _1 \right) ^2\left( f_{i}^{\mathrm{eq}}+\varepsilon f_{i}^{\left( 1 \right)}+\varepsilon ^2f_{i}^{\left( 2 \right)} \right) \\ =-\frac{1}{\tau \delta _t}\left( \varepsilon f_{i}^{\left( 1 \right)}+\varepsilon ^2f_{i}^{\left( 2 \right)} \right) ``` CE展开中的一个基本准则:对于一个等式,其$\varepsilon$的各阶也要分别相等。举例,对于一个等式: ```{math} :label: CE-ns-eq9 A\varepsilon +B\varepsilon ^2=C\varepsilon +D\varepsilon ^2 ``` 则其中暗含了:$A=C$,$B=D$。 基于此准则,我们整理了式{eq}`CE-ns-eq8`中$\varepsilon$的各阶项: > $\varepsilon^1$ ```{math} :label: CE-ns-eq10 \boldsymbol{D}_{1\boldsymbol{i}}f_{i}^{\mathrm{eq}}=\left( \partial _{t1}+\boldsymbol{e}_{\boldsymbol{i}}\cdot \nabla _1 \right) f_{i}^{\mathrm{eq}}=-\frac{1}{\tau \delta _t}f_{i}^{\left( 1 \right)} ``` > $\varepsilon^2$ ```{math} :label: CE-ns-eq11 \partial _{t2}f_{i}^{\mathrm{eq}}+\boldsymbol{D}_{1\boldsymbol{i}}f_{i}^{\left( 1 \right)}+\frac{\delta _t}{2}\left( \boldsymbol{D}_{1\boldsymbol{i}} \right) ^2f_{i}^{\mathrm{eq}}=-\frac{1}{\tau \delta _t}f_{i}^{\left( 2 \right)} ``` 根据式{eq}`CE-ns-eq10`可以将式{eq}`CE-ns-eq11`化简: ```{math} :label: CE-ns-eq12 \partial _{t_2}f_{i}^{\mathrm{eq}}+\boldsymbol{D}_{1\boldsymbol{i}}\left[ \left( 1-\frac{1}{2\tau} \right) f_{i}^{\left( 1 \right)} \right] =-\frac{1}{\tau \delta _t}f_{i}^{\left( 2 \right)} ``` 同时,结合式{eq}`CE-ns-eq3`和{eq}`CE-ns-eq4`,我们使用基本准则,可以得到**可解性条件**: ```{math} :label: CE-ns-eq13 \sum_i{f_{i}^{\left( n \right)}}=0, \quad {\color[RGB]{240, 0, 0} \sum_i{\boldsymbol{e}_i f_{i}^{\left( n \right)}}=0} \quad \forall n\geqslant 1 ``` 对式{eq}`CE-ns-eq10`求0阶矩得到: ```{math} :label: CE-ns-eq14 \frac{\partial \rho}{\partial t_1} +\nabla _1 \cdot \left( \rho \boldsymbol{u} \right) =0 ``` 对式{eq}`CE-ns-eq10`求1阶矩得到: ```{math} :label: CE-ns-eq15 {\color[RGB]{240, 0, 0} \frac{\partial \left( \rho \boldsymbol{u} \right)}{\partial t_1} +\nabla _1 \cdot \Pi^{(0)} = 0} ``` 即 Euler 方程。 对式{eq}`CE-ns-eq12`求0阶矩得到: ```{math} :label: CE-ns-eq16 \frac{\partial \rho}{\partial t_2} = 0 ``` (注:此处利用了可解性条件 $\sum f_i^{(1)}=0$ 和 $\sum f_i^{(2)}=0$) 对式{eq}`CE-ns-eq12`求1阶矩得到: ```{math} :label: CE-ns-eq17 {\color[RGB]{240, 0, 0} \frac{\partial (\rho \boldsymbol{u})}{\partial t_2} + \nabla_1 \cdot \left[ \left( 1 - \frac{1}{2\tau} \right) \sum_i \boldsymbol{e}_i \boldsymbol{e}_i f_i^{(1)} \right] = 0} ``` 为了计算 $\Pi^{(1)} = \sum \boldsymbol{e}_i \boldsymbol{e}_i f_i^{(1)}$,利用式{eq}`CE-ns-eq10`(为了推导方便,这部分切换为索引标识法): ```{math} :label: CE-ns-eq18 \Pi^{(1)}_{ij} = \sum_\alpha e_{\alpha,i} e_{\alpha,j} f_\alpha^{(1)} = -\tau \delta_t \sum_\alpha e_{\alpha,i} e_{\alpha,j} \left( \frac{\partial}{\partial t_1}+\boldsymbol{e}_{\alpha}\cdot \nabla _1 \right)f_\alpha ^{\mathrm{eq}} ``` 展开上式: ```{math} :label: CE-ns-eq18-1 \Pi^{(1)}_{ij} = -\tau \delta_t \left[ \frac{\partial}{\partial t_1} \left( \sum_\alpha e_{\alpha,i} e_{\alpha,j} f_\alpha^{\mathrm{eq}} \right) + \frac{\partial}{\partial x_{1,k}} \left( \sum_\alpha e_{\alpha,i} e_{\alpha,j} e_{\alpha,k} f_\alpha^{\mathrm{eq}} \right) \right] ``` 其中二阶矩 $\sum_\alpha e_{\alpha,i} e_{\alpha,j} f_\alpha^{\mathrm{eq}} = \Pi^{(0)}_{ij} = \rho u_i u_j + p \delta_{ij}$。 三阶矩 $\Lambda^{(0)}_{ijk} = \sum_\alpha e_{\alpha,i} e_{\alpha,j} e_{\alpha,k} f_\alpha^{\mathrm{eq}}$。 --- **$\Pi^{(1)}_{ij}$的详细推导如下:(很复杂,可跳过)** 1. **分析第二项(空间导数项)** 对于标准的 D2Q9 或 D3Q19 模型,利用平衡态分布函数的性质,三阶矩为: ```{math} :label: CE-ns-eq18-2 \Lambda^{(0)}_{ijk} = \rho c_s^2 \left( u_i \delta_{jk} + u_j \delta_{ik} + u_k \delta_{ij} \right) ``` 对其求导: ```{math} :label: CE-ns-eq18-3 \begin{aligned} \frac{\partial}{\partial x_{1,k}} \Lambda^{(0)}_{ijk} &= \frac{\partial}{\partial x_{1,k}} \left[ \rho c_s^2 \left( u_i \delta_{jk} + u_j \delta_{ik} + u_k \delta_{ij} \right) \right] \\ &= \frac{\partial (\rho c_s^2 u_i)}{\partial x_{1,j}} + \frac{\partial (\rho c_s^2 u_j)}{\partial x_{1,i}} + \frac{\partial (\rho c_s^2 u_k)}{\partial x_{1,k}} \delta_{ij} \\ &= c_s^2 \left( \rho \frac{\partial u_i}{\partial x_{1,j}} + u_i \frac{\partial \rho}{\partial x_{1,j}} \right) + c_s^2 \left( \rho \frac{\partial u_j}{\partial x_{1,i}} + u_j \frac{\partial \rho}{\partial x_{1,i}} \right) \\ &+ c_s^2 \delta_{ij} \frac{\partial (\rho u_k)}{\partial x_{1,k}} \end{aligned} ``` 2. **分析第一项(时间导数项)** ```{math} :label: CE-ns-eq18-4 \frac{\partial}{\partial t_1} \Pi^{(0)}_{ij} = \frac{\partial}{\partial t_1} (\rho u_i u_j + p \delta_{ij}) = \frac{\partial (\rho u_i u_j)}{\partial t_1} + c_s^2 \delta_{ij} \frac{\partial \rho}{\partial t_1} ``` 利用 $\varepsilon^1$ 阶的连续性方程(式{eq}`CE-ns-eq14`):$\frac{\partial \rho}{\partial t_1} = -\frac{\partial (\rho u_k)}{\partial x_{1,k}}$,代入 $p$ 的导数项: ```{math} :label: CE-ns-eq18-5 c_s^2 \delta_{ij} \frac{\partial \rho}{\partial t_1} = -c_s^2 \delta_{ij} \frac{\partial (\rho u_k)}{\partial x_{1,k}} ``` 对于 $\rho u_i u_j$ 的导数项,展开为: ```{math} :label: CE-ns-eq18-6 \frac{\partial (\rho u_i u_j)}{\partial t_1} = u_j \frac{\partial (\rho u_i)}{\partial t_1} + u_i \frac{\partial (\rho u_j)}{\partial t_1} - u_i u_j \frac{\partial \rho}{\partial t_1} ``` 利用 $\varepsilon^1$ 阶的动量方程(式{eq}`CE-ns-eq15`):$\frac{\partial (\rho u_i)}{\partial t_1} = -\frac{\partial (\rho u_i u_k + p \delta_{ik})}{\partial x_{1,k}}$,代入上式: ```{math} :label: CE-ns-eq18-7 \begin{aligned} &\frac{\partial (\rho u_i u_j)}{\partial t_1} \\ &= u_j \left[ -\frac{\partial (\rho u_i u_k)}{\partial x_{1,k}} - \frac{\partial p}{\partial x_{1,i}} \right] + u_i \left[ -\frac{\partial (\rho u_j u_k)}{\partial x_{1,k}} - \frac{\partial p}{\partial x_{1,j}} \right] + u_i u_j \frac{\partial (\rho u_k)}{\partial x_{1,k}} \\ &= -u_j \frac{\partial (\rho u_i u_k)}{\partial x_{1,k}} - c_s^2 u_j \frac{\partial \rho}{\partial x_{1,i}} - u_i \frac{\partial (\rho u_j u_k)}{\partial x_{1,k}} - c_s^2 u_i \frac{\partial \rho}{\partial x_{1,j}} + u_i u_j \frac{\partial (\rho u_k)}{\partial x_{1,k}} \end{aligned} ``` 进一步化简上式中的对流项导数: ```{math} :label: CE-ns-eq18-8 \begin{aligned} & -u_j \frac{\partial (\rho u_i u_k)}{\partial x_{1,k}} - u_i \frac{\partial (\rho u_j u_k)}{\partial x_{1,k}} + u_i u_j \frac{\partial (\rho u_k)}{\partial x_{1,k}} \\ &= -u_j \left( \rho u_k \frac{\partial u_i}{\partial x_{1,k}} + u_i \frac{\partial (\rho u_k)}{\partial x_{1,k}} \right) - u_i \left( \rho u_k \frac{\partial u_j}{\partial x_{1,k}} + u_j \frac{\partial (\rho u_k)}{\partial x_{1,k}} \right) \\ & + u_i u_j \frac{\partial (\rho u_k)}{\partial x_{1,k}} \\ &= -\rho u_j u_k \frac{\partial u_i}{\partial x_{1,k}} - \rho u_i u_k \frac{\partial u_j}{\partial x_{1,k}} - u_i u_j \frac{\partial (\rho u_k)}{\partial x_{1,k}} \\ &= -\frac{\partial (\rho u_i u_j u_k)}{\partial x_{1,k}} \end{aligned} ``` 所以: ```{math} :label: CE-ns-eq18-9 \frac{\partial (\rho u_i u_j)}{\partial t_1} = -c_s^2 u_j \frac{\partial \rho}{\partial x_{1,i}} - c_s^2 u_i \frac{\partial \rho}{\partial x_{1,j}} - \frac{\partial (\rho u_i u_j u_k)}{\partial x_{1,k}} ``` 3. **合并各项** 将式{eq}`CE-ns-eq18-3`、式{eq}`CE-ns-eq18-5`和式{eq}`CE-ns-eq18-9`代入式{eq}`CE-ns-eq18-1`: ```{math} :label: CE-ns-eq18-10 \begin{aligned} & \Pi^{(1)}_{ij} \\ &= -\tau \delta_t \left[ \left( -c_s^2 u_j \frac{\partial \rho}{\partial x_{1,i}} - c_s^2 u_i \frac{\partial \rho}{\partial x_{1,j}} - \frac{\partial (\rho u_i u_j u_k)}{\partial x_{1,k}} - c_s^2 \delta_{ij} \frac{\partial (\rho u_k)}{\partial x_{1,k}} \right) \right. \\ &\quad \left. + \left( \rho c_s^2 \frac{\partial u_i}{\partial x_{1,j}} + c_s^2 u_i \frac{\partial \rho}{\partial x_{1,j}} + \rho c_s^2 \frac{\partial u_j}{\partial x_{1,i}} + c_s^2 u_j \frac{\partial \rho}{\partial x_{1,i}} + c_s^2 \delta_{ij} \frac{\partial (\rho u_k)}{\partial x_{1,k}} \right) \right] \end{aligned} ``` 可以看到,包含 $\frac{\partial \rho}{\partial x}$ 的项和包含 $\delta_{ij}$ 的项都相互抵消了: ```{math} :label: CE-ns-eq18-11 \Pi^{(1)}_{ij} = -\tau \delta_t \left[ \rho c_s^2 \left( \frac{\partial u_i}{\partial x_{1,j}} + \frac{\partial u_j}{\partial x_{1,i}} \right) - \frac{\partial (\rho u_i u_j u_k)}{\partial x_{1,k}} \right] ``` 其中 $\frac{\partial (\rho u_i u_j u_k)}{\partial x_{1,k}}$ 是 $O(u^3)$ 的误差项,通常忽略不计。 --- 最终得到非平衡态应力张量: ```{math} :label: CE-ns-eq19 \Pi^{(1)}_{ij} \approx -\tau \delta_t \rho c_s^2 \left( \frac{\partial u_i}{\partial x_{1,j}} + \frac{\partial u_j}{\partial x_{1,i}} \right) ``` 即: ```{math} \Pi^{(1)} \approx -\tau \delta_t \rho c_s^2 (\nabla_1 \boldsymbol{u} + (\nabla_1 \boldsymbol{u})^T) ``` 同时考虑式{eq}`CE-ns-eq0`、{eq}`CE-ns-eq14`、{eq}`CE-ns-eq15`、{eq}`CE-ns-eq16`、{eq}`CE-ns-eq17`及{eq}`CE-ns-eq19`得到: ```{math} :label: CE-ns-eq20 \frac{\partial \rho}{\partial t} + \nabla \cdot (\rho \boldsymbol{u}) = 0 ``` ```{math} :label: CE-ns-eq21 {\color[RGB]{240, 0, 0} \frac{\partial (\rho \boldsymbol{u})}{\partial t} + \nabla \cdot (\rho \boldsymbol{u} \boldsymbol{u}) = -\nabla p + \nabla \cdot \left[ \rho \nu (\nabla \boldsymbol{u} + (\nabla \boldsymbol{u})^T) \right]} ``` 其中$p=\rho c_s^2$,粘度由下式给出: ```{math} :label: CE-ns-eq22 {\color[RGB]{240, 0, 0}\nu = c_s^2 \left( \tau - \frac{1}{2} \right) \delta_t} ``` --- **如果您觉得这个项目对您有帮助,可以考虑用以下方式支持我:** - ☕ 请我喝杯咖啡 > 如果条件允许,[欢迎捐赠支持](../支持与捐赠/支持与捐赠.md)! > 每一分都是对我莫大的鼓励,让我能投入更多时间维护和更新。 - 📄 引用我的文章 > 如果暂时不便捐赠,[适当引用我的文章](../我的文章/我的文章.md)也是极好的支持! > 您的引用能帮助这个工作获得更多关注,同样让我感到无比欣慰。 无论哪种方式,都是对我的巨大支持!🙏感谢您让开源世界更美好!✨