A numerical approach to the optimal control of thermally convective flows

Yongcun Song, Xiaoming Yuan and Hangrui Yue

arXiv: 2211.15302v1 · PDF · 28 November 2022
Categories: math.OC
Licence: CC BY 4.0 The site shows the text.

Abstract

The optimal control of thermally convective flows is usually modeled by an optimization problem with constraints of Boussinesq equations that consist of the Navier-Stokes equation and an advection-diffusion equation. This optimal control problem is challenging from both theoretical analysis and algorithmic design perspectives. For example, the nonlinearity and coupling of fluid flows and energy transports prevent direct applications of gradient type algorithms in practice. In this paper, we propose an efficient numerical method to solve this problem based on the operator splitting and optimization techniques. In particular, we employ the Marchuk-Yanenko method leveraged by the L^2-projection for the time discretization of the Boussinesq equations so that the Boussinesq equations are decomposed into some easier linear equations without any difficulty in deriving the corresponding adjoint system. Consequently, at each iteration, four easy linear advection-diffusion equations and two degenerated Stokes equations at each time step are needed to be solved for computing a gradient. Then, we apply the Bercovier-Pironneau finite element method for space discretization, and design a BFGS type algorithm for solving the fully discretized optimal control problem. We look into the structure of the problem, and design a meticulous strategy to seek step sizes for the BFGS efficiently. Efficiency of the numerical approach is promisingly validated by the results of some preliminary numerical experiments.

Items 9

Theorem

E3SR

r@thm:soln_state Label ‘thm:soln_state’ multiply defined

SIAM Label Error: There were multiply-defined labels Correct multiply-defined labels. r@thm:soln_stateequationSMpage Suppose that y_0\in Y satisfies y_{0}|_{\Gamma}=0, \theta_0\in L^2(\Omega) and v\in L^2(0,T; H^{-1/2}(\Gamma)). Then, the state equation (?)-(?) admits a unique weak solution (y,p,\theta)\in W_y(0,T)\times P\times W_\theta(0,T).

Theorem

X4YU

r@thm:oc Label ‘thm:oc’ multiply defined

SIAM Label Error: There were multiply-defined labels Correct multiply-defined labels. r@thm:ocequationSMpage If u\in L^2(\Sigma_c) is an optimal control of (?)-(?), then it satisfies the following optimality system

\@bsphack \protected@write\@auxout{} {\string \@ifundefined{r@oc} \relax {\gdef \relax { \@latex@warning@no@line{Label `oc' multiply defined}} \GenericError{ \space\space\space\@spaces\@spaces\@spaces }{ SIAM Label Error: There were multiply-defined labels }{ Correct multiply-defined labels. }{} } \global\@namedef{r@oc}{{\p@equationSM\arabic{equation}}{SM\arabic{page}}}} \@esphack DJ(u)=\alpha u+\zeta|_{\Sigma_c}=0,

with \zeta obtained from the solutions of the following fully coupled systems:

\begin{aligned}\@bsphack \protected@write\@auxout{} {\string \@ifundefined{r@state_equation11} \relax {\gdef \relax { \@latex@warning@no@line{Label `state_equation11' multiply defined}} \GenericError{ \space\space\space\@spaces\@spaces\@spaces }{ SIAM Label Error: There were multiply-defined labels }{ Correct multiply-defined labels. }{} } \global\@namedef{r@state_equation11}{{\p@equationSM\arabic{equation}}{SM\arabic{page}}}} \@esphack \left\{ \begin{aligned} &\frac{\partial y}{\partial t}+(y\cdot{\nabla})y-\nu_1 \Delta y+\nabla p=\theta \mathbf{{e}}_2\quad \text{in}\quad \Omega\times(0,T), \\ &\nabla\cdot y=0\quad \text{in}\quad \Omega\times(0,T),\\ &y=0\quad \text{on}\quad \Gamma\times(0,T),\\ &y(0)=y_0, \end{aligned} \right. \end{aligned}

\begin{aligned}\@bsphack \protected@write\@auxout{} {\string \@ifundefined{r@state_equation12} \relax {\gdef \relax { \@latex@warning@no@line{Label `state_equation12' multiply defined}} \GenericError{ \space\space\space\@spaces\@spaces\@spaces }{ SIAM Label Error: There were multiply-defined labels }{ Correct multiply-defined labels. }{} } \global\@namedef{r@state_equation12}{{\p@equationSM\arabic{equation}}{SM\arabic{page}}}} \@esphack \left\{ \begin{aligned} &\frac{\partial \theta}{\partial t}-\nu_2 \Delta\theta+y\cdot\nabla\theta=0\quad \text{in}\quad \Omega\times(0,T), \\ &\nu_2\frac{\partial\theta}{\partial \vec{n}}=u\chi_{\Sigma_c}\quad \text{on}\quad \Gamma\times(0,T),\\ &\theta(0)=\theta_0, \end{aligned} \right. \end{aligned}

and

\begin{aligned}\@bsphack \protected@write\@auxout{} {\string \@ifundefined{r@adjoint_equation1} \relax {\gdef \relax { \@latex@warning@no@line{Label `adjoint_equation1' multiply defined}} \GenericError{ \space\space\space\@spaces\@spaces\@spaces }{ SIAM Label Error: There were multiply-defined labels }{ Correct multiply-defined labels. }{} } \global\@namedef{r@adjoint_equation1}{{\p@equationSM\arabic{equation}}{SM\arabic{page}}}} \@esphack \left\{ \begin{aligned} &-\frac{\partial {z}}{\partial t}-\nu_1\Delta{z}+(\nabla y)^T{z}-(y\cdot \nabla){z}-\nabla \xi+\zeta\nabla\theta=y-y_d\quad \text{in}~\Omega\times(0,T), \\ &-\nabla\cdot {z}=0\quad\text{in}~\Omega\times(0,T), \\ &{z}=0\quad \text{on}~ \Gamma\times(0,T),\\ &{z}(T)=0, \end{aligned} \right. \end{aligned}

\begin{aligned}\@bsphack \protected@write\@auxout{} {\string \@ifundefined{r@adjoint_equation2} \relax {\gdef \relax { \@latex@warning@no@line{Label `adjoint_equation2' multiply defined}} \GenericError{ \space\space\space\@spaces\@spaces\@spaces }{ SIAM Label Error: There were multiply-defined labels }{ Correct multiply-defined labels. }{} } \global\@namedef{r@adjoint_equation2}{{\p@equationSM\arabic{equation}}{SM\arabic{page}}}} \@esphack \left\{ \begin{aligned} &-\frac{\partial \zeta}{\partial t}-\nu_2\Delta\zeta-y\cdot\nabla\zeta={z}\cdot \mathbf{{e}}_2\quad \text{in}\quad \Omega\times(0,T), \\ &\nu_2\frac{\partial\zeta}{\partial \vec{n}}=0\quad \text{on}\quad \Gamma\times(0,T),\\ &\zeta(T)=0.\\ \end{aligned} \right. \end{aligned}

Here, equations (Theorem) and (Theorem) are the adjoint equations corresponding to velocity and temperature, respectively. Equations (Theorem)-(Theorem) are the state equations defined in (?)-(?).

Used by Remark

Remark

6TX4

The optimality systems for the vorticity reduction case (?) are the same as those given by (Theorem)-(Theorem) but with y-y_d in (Theorem) replaced by \nabla\times(\nabla\times y).

Uses Theorem

Remark

SAWZ

Note that the Neumann boundary condition \bm{y}^ n\cdot\vec{n}=0 is used to guarantee the well-posedness of (?). Alternatively, one may consider replacing (?) by the following equation with Dirichlet boundary condition:

\begin{aligned}\@bsphack \protected@write\@auxout{} {\string \@ifundefined{r@deg_Stokes_y_2} \relax {\gdef \relax { \@latex@warning@no@line{Label `deg_Stokes_y_2' multiply defined}} \GenericError{ \space\space\space\@spaces\@spaces\@spaces }{ SIAM Label Error: There were multiply-defined labels }{ Correct multiply-defined labels. }{} } \global\@namedef{r@deg_Stokes_y_2}{{\p@equationSM\arabic{equation}}{SM\arabic{page}}}} \@esphack \left\{ \begin{aligned} &\frac{\bm{y}^{n}-\bm{\tilde{y}}^n}{\Delta t}+\nabla (\bm{p}^n-\bm{p}^{n-1})= 0\quad \text{in}\quad \Omega, \\ &\nabla\cdot \bm{y}^n=0 \quad \text{in}\quad \Omega, \\ &\bm{{y}}^n =0\quad \text{on}\quad \Gamma. \end{aligned} \right. \end{aligned}

As mentioned in [glowinski2003finite, Chapter 7], problem (Remark) is generally not well-posed since the boundary condition \bm{{y}}^n=0 is too demanding for a solution which does not have the H^1(\Omega)-regularity. Despite this fact, the finite element discretized analogue of (?) (see (?)) is well-posed. Moreover, compared with (?)–(?), numerical experiments show that more accurate approximate solutions can be always obtained by the scheme (?), (?) and (Remark).

Theorem

6Q1Q

If \bm{u}=\{\bm{u}^n\}_{n=1}^N\in{U}^{\Delta t} is an optimal control of the time-discretized problem (?)-(?), then the following optimality conditions hold

\alpha\bm{u}^n+\bm{\zeta}^n|_{\Gamma_c}=0, \quad\forall n=1,2,\cdots,N,

where (\bm{\zeta}^n)_{n=1}^N is the successive solution of the time-discretized state equations (?)-(?) and the corresponding adjoint equations (?)-(?).

Remark

A5QK

r@br1 Label ‘br1’ multiply defined

SIAM Label Error: There were multiply-defined labels Correct multiply-defined labels. r@br1equationSMpage Following [glowinski2003finite], implementing the Peaceman-Rachford splitting method [peaceman1955] to (?)-(?) yields the following time-discretized equations (with 0 < \beta < 1, 0 < \gamma < 1 and \beta + \gamma = 1): \bm{y}^0=y_{0}, \bm{\theta}^0=\theta_0, then for n=1,\cdots, N, solve

\begin{aligned}\@bsphack \protected@write\@auxout{} {\string \@ifundefined{r@prsm1} \relax {\gdef \relax { \@latex@warning@no@line{Label `prsm1' multiply defined}} \GenericError{ \space\space\space\@spaces\@spaces\@spaces }{ SIAM Label Error: There were multiply-defined labels }{ Correct multiply-defined labels. }{} } \global\@namedef{r@prsm1}{{\p@equationSM\arabic{equation}}{SM\arabic{page}}}} \@esphack \qquad \left\{ \begin{aligned} &\frac{\bm{\tilde{y}}^n-\bm{y}^{n-1}}{\Delta t/2}-\beta\nu_1 \Delta\bm{\tilde{y}}^n+\nabla \bm{\tilde{p}}^{n}= \bm{\theta}^{n-1}\bm{e}_2+\gamma\nu_1 \Delta\bm{y}^{n-1}-(\bm{y}^{n-1}\cdot\nabla)\bm{{y}}^{n-1}\quad \text{in}\quad \Omega, \\ &\nabla\cdot \bm{\tilde{y}}^n=0 \quad \text{in}\quad \Omega, \\ &\bm{\tilde{y}}^n=0\quad \text{on}\quad \Gamma,\\ \end{aligned} \right. \end{aligned}

\begin{aligned}\@bsphack \protected@write\@auxout{} {\string \@ifundefined{r@prsm2} \relax {\gdef \relax { \@latex@warning@no@line{Label `prsm2' multiply defined}} \GenericError{ \space\space\space\@spaces\@spaces\@spaces }{ SIAM Label Error: There were multiply-defined labels }{ Correct multiply-defined labels. }{} } \global\@namedef{r@prsm2}{{\p@equationSM\arabic{equation}}{SM\arabic{page}}}} \@esphack \qquad \left\{ \begin{aligned} &\frac{\bm{y}^{n}-\bm{\tilde{y}}^n}{\Delta t/2}-\gamma\nu_1 \Delta\bm{{y}}^n+(\bm{y}^{n}\cdot\nabla)\bm{{y}}^{n}= \bm{\theta}^{n-1}\bm{e}_2+\beta\nu_1 \Delta\bm{\tilde{y}}^n-\nabla\bm{\tilde{p}}^n\quad \text{in}\quad \Omega, \\ &\nabla\cdot \bm{y}^n=0 \quad \text{in}\quad \Omega, \\ &\bm{{y}}^n\cdot \vec{n}=0\quad \text{on}\quad \Gamma, \end{aligned} \right. \end{aligned}

and

\begin{aligned}\@bsphack \protected@write\@auxout{} {\string \@ifundefined{r@prsm3} \relax {\gdef \relax { \@latex@warning@no@line{Label `prsm3' multiply defined}} \GenericError{ \space\space\space\@spaces\@spaces\@spaces }{ SIAM Label Error: There were multiply-defined labels }{ Correct multiply-defined labels. }{} } \global\@namedef{r@prsm3}{{\p@equationSM\arabic{equation}}{SM\arabic{page}}}} \@esphack \qquad \left\{ \begin{aligned} &\frac{\bm{\theta}^n-\bm{\theta}^{n-1}}{\Delta t}-\nu_2 \Delta\bm{\theta}^n+\bm{y}^{n}\cdot\nabla\bm{\theta}^{n}= 0\quad \text{in}\quad \Omega, \\ &\nu_2\frac{\partial\bm{\theta}^n}{\partial{\vec{n}}}=\bm{v}^n\chi_{\Gamma_c}\quad \text{on}\quad \Gamma. \end{aligned} \right. \end{aligned}

We can see that solving the resulting subproblems (Remark)-(Remark) is not difficult. However, in a similar way as what we have done for deriving (?), we found that it is challenging to obtain the adjoint equations associated with the time-discretized equations (Remark)-(Remark). A particular reason is that some terms arise in different time intervals (e.g., \beta\nu_1 \Delta\bm{\tilde{y}}^n and \nabla \bm{\tilde{p}}^{n} arise in both [t^{n-1}, t^{n-1}+\Delta t/2] and [ t^{n-1}+\Delta t/2, t^n]), which makes their adjoint variables difficult to be determined. The same concerns apply also to the Douglas-Rachford splitting method [douglas1956] and the \theta-scheme [glowinski1985, glowinski1986]. By contrast, there is no coupled term between different time intervals in our proposed scheme (?)-(?), and thus no particular difficulty is introduced in deriving the corresponding adjoint equation as shown in (?)-(?).

Remark

WYT9

The fully discretized state equations corresponding to the scheme (?), (?) and (Remark) read as: for n=1,2,\cdots,N, find (\bm{y}_h^n,\tilde{\bm{y}}_h^n,\bm{p}_h^n,\bm{\theta}_h^n)\in [V_{0h}]^2\times [V_{0h}]^2\times P_H\times V_h, such that

\begin{aligned}\@bsphack \protected@write\@auxout{} {\string \@ifundefined{r@fully-discretized-state-variationalform0} \relax {\gdef \relax { \@latex@warning@no@line{Label `fully-discretized-state-variationalform0' multiply defined}} \GenericError{ \space\space\space\@spaces\@spaces\@spaces }{ SIAM Label Error: There were multiply-defined labels }{ Correct multiply-defined labels. }{} } \global\@namedef{r@fully-discretized-state-variationalform0}{{\p@equationSM\arabic{equation}}{SM\arabic{page}}}} \@esphack \left\{ \begin{aligned} &(\frac{\tilde{\bm{y}}_h^n-\bm{y}_h^{n-1}}{\Delta t}, \varphi_h)-b(\bm{p}_H^{n-1},{\varphi_h})+ a(\tilde{\bm{y}}_h^n, \varphi_h)+{c}(\bm{y}_h^{n-1}, \tilde{\bm{y}}_h^n, \varphi_h)=(\bm{\theta}_h^n \bm{e}_2, \varphi_h),\quad \forall \varphi_h\in [V_{0h}]^2 ,\\ &(\frac{\bm{\theta}_h^n-\bm{\theta}_h^{n-1}}{\Delta t},\psi_h)+d(\bm{\theta}_h^n,\psi)+{e}(\bm{y}_h^{n-1},\bm{\theta}_h^n, \psi_h)=(\bm{v}_h^{n}\chi_{\Gamma_c},\psi_h)_\Gamma,\quad \forall \psi_h\in V_h,\\ &(\frac{\bm{y}_h^{n}-\tilde{\bm{y}}_h^n}{\Delta t}, {\varphi_h})-b(\bm{p}_H^n-\bm{p}_H^{n-1},{\varphi_h})=0,\quad \forall {\varphi_h}\in [V_{0h}]^2,\\ &b(q_H,\bm{y}_h^{n})=0,\quad \forall q_H\in P_H,\\ &\bm{y}_h^{0}=y_{0h},\quad \bm{\theta}_h^{0}=\theta_{0h}. \end{aligned} \right. \end{aligned}

Example

Example 1

BK45

We consider the tracking-type optimal control of the Boussinesq equations (?)-(?) with the objective functional (?). The boundary and initial conditions are specified as

\begin{aligned} \begin{aligned} &y=0~ \text{on} ~\Gamma\times(0,T),\quad &y(0)=0.&\\ &\left\{\begin{aligned} &\theta=0&~ \text{on} ~\Gamma\slash\Gamma_c \times(0,T),\\ &\nu_2\frac{\partial\theta}{\partial \vec{n}}=v &~\text{on} ~\Gamma_c \times(0,T),\end{aligned}\right. \quad &\theta(0)=0. \end{aligned} \end{aligned}

Example

Example 2

4IOA

This example is extended from [IR98], which is motivated by the transport process in high-pressure chemical vapor deposition (CVD) reactors. The original example focuses on the steady case and we extend it to an unsteady example in (0, T) with T=15.

Source and licence

The text on this page is from arXiv:2211.15302v1 by Yongcun Song, Xiaoming Yuan and Hangrui Yue, under the licence CC BY 4.0. The corpus made these changes: it read the items out of the LaTeX source, it expanded the macros of the paper, it put each proof after its statement, and it converted the text to HTML. The formulas are the TeX of the paper, and KaTeX renders them. A formula or a reference can look different from the PDF. The page is published under CC BY 4.0.

The record of each item is in analysis/objects/2211/2211.15302.jsonl.