Robert Koslover
Certified Consultant
Please login with a confirmed email address before reporting spam
Posted:
4 years ago
Mar 22, 2021, 3:10 p.m. EDT
Updated:
4 years ago
Mar 22, 2021, 3:13 p.m. EDT
If by "white space" you are referring to regions that are outside the computational volume, then "outside the computational volume" is exactly what those spaces are. The finite element computations take place inside the computational volume. Post-processing may or may not make predictions outside the computational volume, depending on the application. (For example, the compution of a radiated pattern from an antenna can be thought of as making a prediction outside the computational volume.) The various boundaries and boundary conditions impose... well, conditions upon the boundaries. If the physics in your problem requires the modeling of a spatial region that you choose to not include, then you will generally not get the right answer. In short, you need to specify an appropriate computational volume, and appropriate boundary conditions, such that what might "happen" in any region outside the computational volume is, in essence, irrelevant to the answer. Conversely, if such a region matters, then you had better include it in your computational space or (if you can) manage it by means of appropriate boundary conditions that ensure that the appropriate physics will be modeled correctly in the region you actually care about.
-------------------
Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
If by "white space" you are referring to regions that are *outside* the computational volume, then "outside the computational volume" is exactly what those spaces are. The finite element computations take place *inside* the computational volume. Post-processing may or may not make predictions outside the computational volume, depending on the application. (For example, the compution of a radiated pattern from an antenna can be thought of as making a prediction outside the computational volume.) The various boundaries and boundary conditions impose... well, *conditions* upon the *boundaries*. If the physics in your problem *requires* the modeling of a spatial region that you choose to *not* include, then you will generally not get the right answer. In short, you need to specify an appropriate computational volume, and appropriate boundary conditions, such that what might "happen" in any region *outside* the computational volume is, in essence, *irrelevant* to the answer. Conversely, if such a region *matters*, then you had better *include* it in your computational space or (if you can) manage it by means of appropriate boundary conditions that ensure that the appropriate physics will be modeled correctly in the region you actually care about.
Please login with a confirmed email address before reporting spam
Posted:
4 years ago
Apr 5, 2021, 12:43 p.m. EDT
Hi Robert,
Thank you for your reply!
Makes sense!
Hi Robert,
Thank you for your reply!
Makes sense!