Latest Discussions - COMSOL Forums

Simulation of Ecoflex 00-30

Wed, 19 Nov 2025 20:09:57 +0000

Hi, I am trying to simulate a solid and a hollow cylinder made of ecoflex 00-30. I have used different hyperelastic material model values i.e. Yeoh, Ogden, Mooney-Rivilin, etc. from different research papers, yet, I failed to converge the solution in COMSOL 6.3 & 6.1 (both version).

I don't know, what boundary conditions and solver settings are missing, the model is 3D and dimensions are in millimeters.

How to know which solver settings and boundary conditions are required to solve the simulation correctly?

Thanks,

PDE for Modeling Liquid Interfaces

Wed, 19 Nov 2025 17:33:38 +0000

Hello!

I would appreciate some help, as I am trying to solve a partial differential equation of a type I have not worked with before. This PDE models the boundary of a liquid as a radius, , which is a function of the 2D plate the liquid is "sitting" on, such that :

where and are constants. The plate is bounded at the ends by circular plates such that .

It is my understanding that this equation cannot be put into divergence form, which is required by the Coefficient Form PDE interface. As such, I am wondering if there is another COMSOL interface which is best equipped to solve this kind of PDE or if there is another way to still use Coefficient Form PDE.

Thank you in advance for any assistance.

Plate displacement and force due to halbach array

Wed, 19 Nov 2025 10:43:04 +0000

I am working on a problem, where the circular permanent magnets halbach array is rotating using external motors over a Aluminium plate. Due to the flux of time-varying magnetic field, a electromagnetic force is experienced by the halbach array, making it levitate. I am trying to simulate this situation, to calculate the forces acting and the steady-state levitation height of the halbach array. I have tried to simulate it in two ways, first case in which the halbach array is fixed and the aluminium plate rotating with the surrounding region is taken as deformable zone and account for the forces and displacement of plate, and second case in which the plate is rotating at a fixed location and the halbach array inside a deformable region with the forces and displacement of the array. Currently, I have taken the reference from a example problem from application libraries named "electrodynamic_levitation_device". My simulation is running but getting stuck at 4-5% progress. Kindly guide me, I have the attached the .mph files for the both cases.

Modeling the vibration of microbubbles in Thermoviscous Acoustics

Wed, 19 Nov 2025 09:19:15 +0000

Hello,

I am trying to simulate the acoustic streaming around a microbubble under acoustic actuation. To do so, I coupled Pressure Acoustics with Thermoviscous Acoustics and defined surface tension at the bubble interface. First, I did Eigen study and found the resonant frequency. Then, I did frequency domain study at the resonant frequency I found, and use the results of the frequency domain study in a stationary study for Laminar flow to see the acoustic streaming. The problem is that the simulation result doesn't match the experiments. When I check the mode shapes of the bubble interface in Eigen study and Frequency domain study, they are not the same, while I think they must be the same because they are at the same frequency.

Can somebody help me with this contradiction?

Best, Ali

Transport of Concentrated species & Laminar flow

Wed, 19 Nov 2025 02:44:16 +0000

Hello,

I am a Nuclear science and technology doctoral student and I wish to simulate hydrogen & Helium mixture separation by palladium membrane. (Tube shaped) Here is the question.

Case Overview: This study involves helium-hydrogen mixture transport through a pipeline lined with a palladium (Pd) membrane. Hydrogen is able to permeate through the membrane walls, while helium remains confined and flows only within the pipeline.

Boundary Conditions: Constant inlet mass flow rate, Constant outlet pressure, Uniform temperature along the entire pipeline

Model Setup: A coupled model: using the Transport of Concentrated Speciesand Laminar Flow interfaces. Strong coupling between mass transfer and fluid flow is achieved via the Reaction Engineering functionality.

However, the simulation yields unphysical results, as described below. Observed Anomaly: As the He–H₂ mixture flows through the pipeline, hydrogen permeates outward through the Pd membrane. The total mass flow rate of the mixture is expected to decrease along the flow path, and the mass flow rate of helium should remain constant across any cross-section of the pipeline. However, the current model shows a constant flow velocity along the pipeline, while the mass flow rate of helium increases toward the outlet. This results in a higher helium mass flow at the outlet than at the inlet—indicating a clear violation of mass conservation. Visually, it appears that the model artificially compensates for the loss of hydrogen by increasing the helium content.

Objective: Identify the root cause of the mass non-conservation and adjust the model such that the flow velocity decreases gradually along the pipeline, consistent with physical expectations.

It would be of a great help if anybody could guide me on these.

Thanks, Leopold Jiang

COMSOL Multiphysics® Version 6.4 is now available!

Tue, 18 Nov 2025 15:29:12 +0000

We're pleased to announce the release of COMSOL Multiphysics^® version 6.4.

Highlights include:

Enhanced solver performance with the new NVIDIA CUDA® direct sparse solver (cuDSS) for NVIDIA^® GPUs for single-physics and multiphysics simulation
Structural explicit dynamics including support for nonlinear materials, for simulating impacts, blasts, and drop tests
Granular Flow Module: a new product that can be used to simulate the motion and interaction of solid particles in bulk processes, for applications such as conveying, mixing, and additive manufacturing
Support for OpenAI API-compatible models within the Chatbot window
Improved workflows for geometry and meshing

See the complete list of new features and download version 6.4 here: https://www.comsol.com/release/6.4

Issue with wave optics simulation

Tue, 18 Nov 2025 12:38:01 +0000

Hi say i am creating a box(air domain). n=0,k=0 (Refractive index) .

Using two port1 and port2 at opposite face and on other face i am assigning periodic boundary condition. Now i am getting 100% reflection and 0 % transmission. But if i am doing the same thing with microwave frequency range (RF module) i am getting 100 % transmission but not reflection which is correct. But why i am getting unusual results for infrared.

PCM Material based metasurface Phase graph re-simulation

Tue, 18 Nov 2025 04:23:40 +0000

Hello everyone, I am a student working on a PCM-based metasurface for amplitude and Phase modulation in COMSOL in ewfd. I am re-simulating the Fractional crystallisation of GST's amplitude and Phase from the paper "Reconfigurable multifunctional metasurfaces employing hybrid phase-change plasmonic architecture".
I have almost obtained the amplitude modulation, but the problem is the phase modulation, which is totally opposite to it. I have used Arg function and mod function, but the result is still not the same. I would be thankful if someone has the solution for this in Comsol.

Modelling Inflation

Mon, 17 Nov 2025 15:29:38 +0000

Hello, I am hoping to model how an air bladder inflates and how pressure changes. I am trying to create a weight-bearing inflatable and need a way to model this process. I am super lost with where to start so any help or advice would be greatly appreciated.

Correction in the theory manual: Nonlocal Plasticity Equation

Mon, 17 Nov 2025 14:34:34 +0000

Dear Community and Moderators,

I think I have found a mistake in the COMSOL documentation of nonlocal plasticity: https://doc.comsol.com/6.0/doc/com.comsol.help.sme/sme_ug_theory.06.30.html#4121532

Please see the first strong form equation there which I have quoted below:

In the equation, I think the term should not appear as it makes the equations dimensionally inconsistent.

Please let me know if I am wrong. Happy to hear counter arguments and discussions.

Cheers, Ratul

Object stacking problem

Mon, 17 Nov 2025 08:30:05 +0000

For a single object, I applied a 'prescribed displacement' to achieve compression, but after stacking multiple objects axially, using 'prescribed displacement' only compresses the bottom object, while the objects above it do not change at all, causing the simulation to fail.

Parametric Scan

Sun, 16 Nov 2025 07:56:08 +0000

I want to perform a parametric scan on three parameters a, b, and c. a scans from 0 to 30 with a step of 1. After a reaches 30, b starts scanning from 0 to 30 with a step of 1. After b reaches 30, c starts scanning from 0 to 30 with a step of 1.

Making Kymograph plot?

Sat, 15 Nov 2025 15:44:33 +0000

I'm fairly new to Comsol and have difficulties with the following: I have a time-dependent simulation that gives me a concentration as a function of space and time. The geometry and simulation is axially symmetric. I'd like to make a kymograph plot where one axis is time and the other is r, with color coding of the concentration. How can I make this happen? Thank you!

Simulation Times Increasing with LiveLink for MATLAB

Thu, 13 Nov 2025 14:36:45 +0000

I've been running automated simulations with LiveLink for MATLAB, where each run makes small geometry changes, remeshes, and solves. I’ve noticed that while early simulations take ~45 minutes, after 20–30 runs, the simulation time increases a lot (more than twice), even though the problem size stays the same.

Has anyone else experienced this slowdown?

Are there best practices for preventing resource issues or memory leaks when using LiveLink for MATLAB for multiple runs? Is it helpful to restart the COMSOL server periodically, and if so, how do you safely restart it without physically clicking from the start menu?

Would appreciate any advice!

Modelling Radial Fans as Flow Control in a Turbulent Flow Component

Thu, 13 Nov 2025 09:02:58 +0000

I am trying to simulate the airflow using the k-epsilon model in a ventilation system that uses radial fans. There are resistances before and after the radial fan, so I cannot simulate inlet path and outlet path separately. These radial fans naturally induce a 90 Degree bend as opposed to axial fans. There is an interior boundary condition called "Interior fan", which seems to be able to simulate axial fans only. Any ideas How I could simulate a radial fan? If I simply put the interior fan next to the intake our exhaust of the radial fan, I would severely underestimate the airflow because then the centrifugal fan space acts as an extreme bend and I also run into convergence issues.

I was thinking about skipping the volume of the fan that spins entirely by connecting an inlet where the fan exhaust, and an outlet where the fan takes in air, with the space between them just not part of the simulation, and the airflow through the outlet (fan intake) is "teleported" to the inlet (fan exhaust). I have so far not been able to figure out how to do this and do not know if it is possible.

Best Regards, David

Undefined value found in the stiffness matrix.

Thu, 13 Nov 2025 06:42:39 +0000

I was simulating the Darcy flow-fracture module and introduced pressure-dependent permeability and porosity. Then, I performed calculations on a two-dimensional rectangular area containing several simple fractures. However, within less than three seconds of computation, the following error popped up:

Undefined value found.

Details: There is an undefined value in the stiffness matrix. There are 229721 equations, and NaN/Inf appears in the matrix rows for the variable comp1.p. At coordinates: (140,60), (139.079,59.1385), (139.586,58.4897), (138.92,58.3842), (137.275,58.76), ... Similarly, NaN/Inf appears in the matrix columns for the degrees of freedom.

Based on this error message, I attempted the following steps:

Adjusted the mesh density.

Handled variables that might result in zero in the denominator. For example, I rewrote "a/p" as "a/max(p, 1e-7[MPa])".

Checked the boundary conditions to ensure no conflicting Dirichlet boundary conditions were applied to two adjacent boundaries. In reality, the external boundaries of the model are all set to no flow, and a pressure condition is only applied at the midpoint of one boundary, using a step function to ensure a smooth transition of the pressure at that point.

After taking these measures, the error still persists. The error coordinates are concentrated in the upper-right and lower-right corners of the rectangle. How can I resolve this issue?

Skin Effect Not Visible in Electric Currents (EC) Frequency Domain Simulation

Tue, 11 Nov 2025 17:50:20 +0000

Hi everyone,

I'm trying to visualize the skin effect in a tungsten cylinder using the Electric Currents (EC) interface in the frequency domain. I’ve added a Terminal and a Ground, and set the excitation frequency to 1 MHz.

I also refined the mesh near the boundaries using Boundary Layers (to capture the skin depth), but the current density distribution still looks uniform it behaves as if it’s a DC simulation.

Could anyone explain:

How to correctly make the skin effect visible in EC (without switching to the Magnetic Fields or EM Waves interface)?
Whether this setup can actually capture the skin effect, or if the EC interface inherently behaves as a low-frequency (quasi-static) solver?

Thanks in advance!

Current in a custom racetrack coil

Tue, 11 Nov 2025 14:52:35 +0000

I have a racetrack coil that is imported from CAD and I am having trouble exciting current in it. I have no error messages but no flux at all.

MoS2 based Transistor

Tue, 11 Nov 2025 10:41:31 +0000

To the COMSOL Community and Experts,

I am writing to request assistance with a convergence issue I am encountering while using the COMSOL Semiconductor Module.

My objective is to replicate a simulation based on a publicly available model of an MoS2 + HfO2 transistor (which I will refer to as the "2nd file"). I have constructed my own project, "MoS2_transistor" (the "1st file"), by carefully transferring the physics, parameters, and mesh settings from the reference model.

A critical problem has arisen: while the reference model (2nd file) computes and converges successfully, my implementation (1st file) consistently fails to converge and terminates with an error.

Given that the reference file functions correctly, I suspect there may be a subtle discrepancy or setup error in my project file that I am overlooking.

I would be extremely grateful if an expert could offer guidance or insights to help diagnose and resolve this convergence failure.

Thank you for your time and consideration.

Sincerely,

2D Fracture Network Generation

Tue, 11 Nov 2025 10:02:42 +0000

I currently working on discrete fracture networks. I wanna analyze via Monte Carlo approach multiple DFN´s. I know that there is the integrated Add-in (Discrete Fracture Network) but it is for me not possible to process the simulation with different networks.

Maybe anyone has a nice idea or work around for that. :)

Greetings Dominik