Hey there Mostafa Shabani! Look, when it comes to the `atlocal` command in COMSOL, you've stumbled upon a bit of a tricky one. Now, I'm not claiming to be the absolute guru of COMSOL, but let me throw in my two cents.

The `atlocal` command in COMSOL is used to evaluate variables at a specific geometric location. It sounds like you're trying to use coordinates but aren't getting the expected results. First things first, are you Mostafa Shabani sure your coordinates are set up correctly? Double-check those values, and don't forget about the unit consistency—COMSOL can be a stickler for that.

If you're still hitting a wall, consider the surrounding context of your simulation. Sometimes, issues with `atlocal` can stem from the mesh or the way your model is defined. Ensure your geometry and mesh are accurate.

And hey, reaching out to the COMSOL community or forums might be a good move. There's a wealth of knowledge out there, and sometimes a fresh set of eyes can spot what you might be missing. Good luck, and let me know if you need more wisdom!

atlocal acts directly on mesh elements, so the coordinates you enter are local. You can for example make a volume plot for a 3D acoustics setup with pressure p using the expression atlocal(1,1,1p). Can I ask what it is that you are trying to achieve here? Mostafa Shabani

René Christensen Thank you for your response, I wanted to solve an equation using the global equations in the semiconductor module without using nonlocal coupling. Also, I wanted to access the module variables in a random coordinate of the solution space-for example using global evaluation -.

I am not sure if atlocal is the way to go. If you by random coordinate mean a global coordinate, then I think it should done otherwise. atlocal uses local coordinates on the mesh element so it varies from 0-1 for each local coordinate.