In summary, the conversation discusses solving a nonlinear PDE using the finite difference method. There is no single discretization method that works well for all problems, and it is important to use methods specifically designed for nonlinear systems. The conversation also includes a specific nonlinear PDE and its boundary conditions, and the individual has attempted to use a finite difference method to solve it. However, it is unclear if this is the correct approach and they are seeking further guidance.
  • #1
Last-cloud
4
0
i want to solve a nonlinear PDE with finite difference method ,but using just discretization like in linear PDE , it will lead to nowhere , what's the right way to use FDM to solve nonlinear PDE or could someone provide me with book's titles or articles that can help me solving a nonlinear pdf using FDM
 
Physics news on Phys.org
  • #2
Last-cloud said:
i want to solve a nonlinear PDE with finite difference method ,but using just discretization like in linear PDE , it will lead to nowhere , what's the right way to use FDM to solve nonlinear PDE or could someone provide me with book's titles or articles that can help me solving a nonlinear pdf using FDM

There is no single finite difference discretization that works well for all problems. This is true for both linear and nonlinear PDEs. Most methods used to solve nonlinear systems are based off of methods that work for linear models of the nonlinear problem. And I'm not sure what you mean by we you say "discretization like a linear PDE, it will lead to nowhere." It is true that nonlinear PDE's can be substantially more difficult to analyze, and there are additional issues that can arise.

Do you have a particular problem in mind? Are you running into a specific issue?
 
  • #3
yes , i have a problem ;
\begin{equation}
m_{z}\ddot{w}+EIw'''-Tw''-f+c_{1}\dot{w}-EAv''w'-EAv'w''-\dfrac{3}{2}EA(w')^2w''=0
\end{equation}
\begin{equation}
m_{z}\ddot{v}+c_{2}\dot{v}-EAv''-EAw'w''=0
\end{equation}
the boundary conditions of the system :
\begin{equation}
w''(0,t)=w''(L,t)=w(0,t)=v(0,t)=0
\end{equation}
\begin{equation}
-EIw'''(L,t)+Tw'(L,t)+EAv'(L,t)w'(L,t)+\dfrac{1}{2}EA\left[ w'(L,t)\right] ^{3}=u_{T}(t)
\end{equation}
\begin{equation}
\dfrac{1}{2}EA[w'(L,t)]^{2}+EAv'(L,t)=u_{L}(t)
\end{equation}
\begin{equation}
w(x,0)=w'(x,0)=v(x,0)=v'(x,0)=0
\end{equation}
where
\begin{equation}
w'=\dfrac{\partial w(x,t)}{\partial x} \;\; and \;\; \dot{w}=\dfrac{\partial w(x,t)}{\partial t}
\end{equation}
what I've tried to do is:
\begin{equation}
\begin{split}
& m_{z}\left( \dfrac{w_{i}^{j+1}-2w_{i}^{j}+w_{i}^{j-1}}{k^{2}}\right)+ EI\left( \dfrac{w_{i+2}^{j}-2w_{i+1}^{j}+2w_{i-1}^{j}-w_{i-2}^{j}}{2h^{3}}\right)-T\left( \dfrac{w_{i+1}^{j}-2w_{i}^{j}+w_{i-1}^{j}}{h^{2}}\right)+c_{1}\left( \dfrac{w_{i}^{j+1}-w_{i}^{j}}{k}\right)-\\
& EA\left( \dfrac{v_{i+1}^{j}-2v_{i}^{j}+v_{i-1}^{j}}{h^{2}}\right)\left( \dfrac{w_{i+1}^{j}-w_{i}^{j}}{h}\right)-EA \left( \dfrac{v_{i+1}^{j}-v_{i}^{j}}{h}\right)\left( \dfrac{w_{i+1}^{j}-2w_{i}^{j}+w_{i-1}^{j}}{h^{2}}\right)-\\
& \dfrac{3}{2}EA\left( \dfrac{w_{i+1}^{j}-w_{i}^{j}}{h}\right)^{2} \left( \dfrac{w_{i+1}^{j}-2w_{i}^{j}+w_{i-1}^{j}}{h^{2}}\right)=f
\end{split}
\end{equation}
and
\begin{equation}
\begin{split}
& m_{z}\left( \dfrac{v_{i}^{j+1}-2v_{i}^{j}+v_{i}^{j-1}}{k^{2}}\right)+c_{2}\left( \dfrac{v_{i}^{j+1}-v_{i}^{j}}{k}\right)- EA\left( \dfrac{v_{i+1}^{j}-2v_{i}^{j}+v_{i-1}^{j}}{h^{2}}\right)-\\
& EA\left( \dfrac{w_{i+1}^{j}-w_{i}^{j}}{h}\right) \left( \dfrac{w_{i+1}^{j}-2w_{i}^{j}+w_{i-1}^{j}}{h^{2}}\right)=0
\end{split}
\end{equation}
thus
the 1st equation of the system:
\begin{equation}
\begin{split}
&\left(\dfrac{(m_{z}+kc_{1})w_{i}^{j+1}-(2+kc_{1})w_{i}^{j}+w_{i}^{j-1}}{k^{2}}\right)+ \left( \dfrac{EIw_{i+2}^{j}-2(EI+Th)w_{i+1}^{j}+4Thw_{i}^{j}+2(EI-Th)w_{i-1}^{j}-EIw_{i-2}^{j}}{2h^{3}}\right)-\\
& EA\left( \dfrac{v_{i+1}^{j}-2v_{i}^{j}+v_{i-1}^{j}}{h^{2}}\right)\left( \dfrac{w_{i+1}^{j}-w_{i}^{j}}{h}\right)-EA \left( \dfrac{v_{i+1}^{j}-v_{i}^{j}}{h}\right)\left( \dfrac{w_{i+1}^{j}-2w_{i}^{j}+w_{i-1}^{j}}{h^{2}}\right)-\\
& \dfrac{3}{2}EA\left( \dfrac{w_{i+1}^{j}-w_{i}^{j}}{h}\right)^{2} \left( \dfrac{w_{i+1}^{j}-2w_{i}^{j}+w_{i-1}^{j}}{h^{2}}\right)=f
\end{split}
\end{equation}
the 2nd equation of the system:
\begin{equation}
\begin{split}
& \left(\dfrac{(m_{z}+kc_{2})v_{i}^{j+1}-(2+kc_{2})v_{i}^{j}+v_{i}^{j-1}}{k^{2}}\right)- EA\left( \dfrac{v_{i+1}^{j}-2v_{i}^{j}+v_{i-1}^{j}}{h^{2}}\right) -\\
& EA\left( \dfrac{w_{i+1}^{j}-w_{i}^{j}}{h}\right) \left( \dfrac{w_{i+1}^{j}-2w_{i}^{j}+w_{i-1}^{j}}{h^{2}}\right)=0
\end{split}
\end{equation}
where h is delta x and k is delta t
what should i do next ??am i in the right path ?,, is this a good start or there is something else to do before using finite difference method.
thank you.
 
Last edited:

FAQ: Finite difference method nonlinear PDE

What is the Finite Difference Method (FDM) used for?

The Finite Difference Method is a numerical technique used to solve partial differential equations (PDEs). It involves approximating the derivative of a function at a specific point by using the function values at nearby points. This method is commonly used in a wide range of scientific and engineering applications, such as fluid dynamics, heat transfer, and structural mechanics.

What are the advantages of using FDM over analytical methods for solving nonlinear PDEs?

FDM is advantageous because it can handle complex geometries and boundary conditions that are difficult to solve analytically. It is also a relatively simple and intuitive method, making it easier to implement and understand. Additionally, FDM allows for the use of computers to solve large systems of equations, which would be nearly impossible to do by hand.

What are the limitations of FDM for solving nonlinear PDEs?

One limitation of FDM is that it requires a structured grid, meaning the points must be arranged in a regular pattern. This can be a challenge for problems with irregular or complex geometries. Additionally, FDM can be computationally expensive for problems with high dimensions or when using a fine grid to achieve accurate results.

How do you handle nonlinear terms in FDM for solving PDEs?

Nonlinear terms in the PDE can be handled by using an iterative process, such as the Newton-Raphson method, to solve the resulting nonlinear system of equations. This involves linearizing the nonlinear terms and solving the linearized system at each iteration until a specified level of accuracy is reached.

How do you choose the appropriate grid size for FDM?

The grid size for FDM should be chosen based on the desired level of accuracy and the complexity of the problem. A smaller grid size will result in a more accurate solution, but it will also increase the computational cost. Therefore, it is important to balance accuracy and efficiency when choosing the grid size. In general, it is recommended to perform grid refinement studies to determine the optimal grid size for a given problem.

Similar threads

Replies
36
Views
1K
Replies
1
Views
2K
Replies
2
Views
1K
Replies
5
Views
4K
Replies
2
Views
2K
Replies
3
Views
632
Replies
2
Views
3K
Back
Top