updating boundary condition after one iteration and permanently

babakflame · June 22, 2017, 22:19

Dear Fellows

I want to implement a boundary condition in a somehow unique way.

I want to start a simulation with a high value called $\varphi_1$ .
Immediately, after 1 iteration, I want to put this relation as the boundary condition:
$\varphi = \varphi_1 + a \frac{d \varphi}{d n}$ .

Actually, that 1 iteraion is needed to calculate $\frac{d \varphi}{d n}$ in the high value $\varphi_1$ and then modifying boundary condition to a much smaller value. Here, the point is that $\frac{d \varphi}{d n}$ comes from that first iteration.

Basically, what I want to do is start a simulation with a high Dirichlet boundary condition and after first iteration, calculate surface normal gradient and find the new $\varphi$ based on modification formula and continue simulation with new Dirichlet boundary condition. The new Dirichlet boundary condition remains intact throughout the simulation.

So, in first iteration my boundary condition is $\varphi_1$ and for the rest of simulation is $\varphi$ .

Any clue how to implement it?

arjun · June 22, 2017, 22:58

I am trying to code it and most of it is done for now.

Here is my take, during the initialization of model, you first solve the poisson equation for this variable (Ue as you called it in other thread). After it is solved the field comes to be smooth. dUedn is no longer as sharp as it is in the start for cplus and cminus.

So basically in my model

init cplus, cminus from user.

using cplus and cminus, solve poisson equation for Ue to initialize it.

Iterations

1. Solve Ue poisson equation
2. Solve Cplus and Cminus transport.

This is my rough outline.

babakflame · June 22, 2017, 23:15

Dear Arjun,

Many thanks for your reply. However, this is a new case and thats the reason I started a new thread for it.

Here, I want to do the case with high a $\varphi$ called $\varphi_1$ just for 1 iteration and then modify the boundary with a new Dirichlet which comes from a modification formula.

Regards

June 22, 2017, 22:19	updating boundary condition after one iteration and permanently	#1
babakflame Senior Member Bobby Join Date: Oct 2012 Location: Michigan Posts: 454 Rep Power: 16	Dear Fellows I want to implement a boundary condition in a somehow unique way. I want to start a simulation with a high value called $\varphi_1$ . Immediately, after 1 iteration, I want to put this relation as the boundary condition: $\varphi = \varphi_1 + a \frac{d \varphi}{d n}$ . Actually, that 1 iteraion is needed to calculate $\frac{d \varphi}{d n}$ in the high value $\varphi_1$ and then modifying boundary condition to a much smaller value. Here, the point is that $\frac{d \varphi}{d n}$ comes from that first iteration. Basically, what I want to do is start a simulation with a high Dirichlet boundary condition and after first iteration, calculate surface normal gradient and find the new $\varphi$ based on modification formula and continue simulation with new Dirichlet boundary condition. The new Dirichlet boundary condition remains intact throughout the simulation. So, in first iteration my boundary condition is $\varphi_1$ and for the rest of simulation is $\varphi$ . Any clue how to implement it?

June 22, 2017, 22:58		#2
arjun Senior Member Arjun Join Date: Mar 2009 Location: Nurenberg, Germany Posts: 1,285 Rep Power: 34	I am trying to code it and most of it is done for now. Here is my take, during the initialization of model, you first solve the poisson equation for this variable (Ue as you called it in other thread). After it is solved the field comes to be smooth. dUedn is no longer as sharp as it is in the start for cplus and cminus. So basically in my model init cplus, cminus from user. using cplus and cminus, solve poisson equation for Ue to initialize it. Iterations 1. Solve Ue poisson equation 2. Solve Cplus and Cminus transport. This is my rough outline.

June 22, 2017, 23:15		#3
babakflame Senior Member Bobby Join Date: Oct 2012 Location: Michigan Posts: 454 Rep Power: 16	Dear Arjun, Many thanks for your reply. However, this is a new case and thats the reason I started a new thread for it. Here, I want to do the case with high a $\varphi$ called $\varphi_1$ just for 1 iteration and then modify the boundary with a new Dirichlet which comes from a modification formula. Regards