Latest Discussions - COMSOL Forums

Adding boundaries to connect with the Polarizing Beamsplitter

Thu, 23 Jul 2026 11:18:49 +0000

Hello,

I want to use the Polarizing Beamsplitter (PBS), (Application ID: 31041), within a larger system, so I am attempting to add an air boundary (just a rectangle with n=1) to the right of the PBS, but I am getting the result shown in this image link: imglink.cc/i/fkkT6h7BNT

As far as I can tell, in order to create a new domain I need to define the wavevectors in that domain and a phase shift to ensure phase continuity between domains. The wavevector would just be lambda/n_air and the phase shift I calculate using the ODEs that calculated all the prior phase shifts within the PBS, using the new wavevectors I've created. However, I am getting a result that clearly has some noisiness to it and indicates that something is not right. When I change the refractive index of the domain from 1 to 1.6586, the exact refractice index of the adjacent prism, everything works fine and results in the clean image shown: imglink.cc/cdn/ntutmapbAC.png

Changing it from 1 to 1.6585, just 0.0001 less than the refractive index of the adjacent prism, leads to a result almost identical to the first noisy result when n=1. From this I'm assuming that I must be doing something wrong with the wavevectors, as that is the only quantity that would change by a large amount as a result of a tiny change in refractive index. But I have no idea what could be wrong there. Following are some relevant information of the changes I made to the file in order to implement an air domain:

Refractive Index of 'Right Air Domain' = n_air = 1 Interface between right prism and right air domain = x8 = width of PBS divided by 2 k1Airx = k0n_air = k2Airy k1Airy = 0 = k2Airx k1Prismxx8+phi180-k1Airxx8-phiAirx k2Prismyx8+phi200-k2Airy*x8-phiAiry phiAirx and phiAiry are used in the definitions section for the right air domain as phi1 and phi2 respectively.

I've uploaded my attempt at this after clearing all built meshes. If there are any questions about how I implemented anything I'm more than happy to respond. Thank you for your time!

PFM-Brittle Material

Wed, 22 Jul 2026 14:22:29 +0000

I try to create a PFM for a 2D transitional model (Z-tetha and R-Z plane) of a brittle-ceramic pellet under thermal stress and it doesn't have notches. Using this model, I am eager to see the formation of different cracks over time. I have been following COMSOL examples I know so far, but I haven't been able to make them work. Do you have any extra examples or a guide to delve deeply PFM, considering the described model?

Thank you,

Best,

How to create a Model Method and the Method Call node through Livelink for MATLAB?

Mon, 20 Jul 2026 11:52:02 +0000

Hi there,

I need to create a series of mph files and upload them to a third-party cloud of HPC for calculation.

In each mph file, a two-step parameter sweeping study is going to be performed for critical strain. The 1st step is a coarse sweeping, whose results determines the interval of a subsequent refined sweeping referred to as 'the 2nd step'.

I have discussed the implementation with chatGPT, and it suggests to use a Model Method code to control the process. Firstly, use Model Method to control the coarse sweeping step, and stop it as soon as negative eigenvalues emerge. Then, the strains corresponding to the current step and the previous step are identified as the higher and lower ends of the interval for the refined sweep in the following study.

After building the model through GUI, I try to save it as a .m file as usual. However, the .m file doesn't contain the code about create the 'Model Method' and the 'Method Call' node.

Infinite element domain setup for an underground cable heat-transfer model

Sun, 19 Jul 2026 17:15:52 +0000

Background

I am modeling heat transfer from an underground cable using Heat Transfer in Solids. The cable is within in a finite hexahedral soil domain, with an outer soil region assigned as an Infinite Element Domain.

The top surface represents the physical ground surface and uses a Convective Heat Flux boundary condition. Therefore, I do not want the infinite-element region to extend upward.

I understand that a Cartesian infinite-element region should be divided into separate face, edge, and corner domains rather than treated as one hollow hexahedral domain. I am currently modifying the geometry accordingly.

I also understand that the infinite-element domains should preferably use a swept mesh. With my current geometry, physics-controlled meshing has difficulties because the top is open and some regions do not appear to have clear parallel source and destination faces. A free tetrahedral mesh could be generated, but the study then failed because COMSOL could not identify valid Cartesian regions in the infinite outer domain.

So my questions are:

1) Is a Cartesian infinite-element region appropriate for this model if it surrounds the finite soil only laterally, below, and possibly at the cable ends, while leaving the ground surface open?

2) Would a hemispherical finite and infinite soil geometry be more suitable than the rectangular Cartesian geometry?

3) Should the cable extend into the infinite-element domains at the front and back in case of both types of infinite domain (hemispherical and hexahedral)?

4) Must every infinite-element subdomain be swept from the finite-soil interface toward the outer boundary? For the top surface of the infinite domain I am getting an error saying that there is no corresponding surface face to get the sweeping done. (This issue did not appear when I created a free tetrahedral mesh in the infinite domain)

I have attached images of my current geometry (and the figures I have referred from your Learning Center post).

Thank you for the time taken to read my post.

RRAM TaOx Model I-V Curve Error

Sun, 19 Jul 2026 14:00:08 +0000

I am running into an error while trying to compute an I-V curve. I ran the Study 1 Computation but couldn't get the I-V curves like I did on the lab computer. I am including the integration I was doing for the current, the expression I used for current, and also the plot of current vs voltage below. Can you look at them and let me know how to fix the problem I'm running into?

Piezoelectric coupling matrix input formatting

Tue, 14 Jul 2026 01:09:44 +0000

I'm inputing the Voigt notation piezoelectric coupling matrix for strain-charge form into the "Material Contents" table. Can someone confirm that the formatting below is the proper notation?

My input: {{0,0,0,0,d15,0}, {0,0,0,d15,0,0}, {d31,d31,d33,0,0,0}}

where d15 and so on are replaced by the corresponding values in SI units.

Acoustic-Structure Interaction: Modeling Sound Transmission Loss (STL) of a Finite Cylindrical Shell

Wed, 08 Jul 2026 09:53:10 +0000

Hi everyone,

I am trying to model the Sound Transmission Loss (STL) through a finite-length, thin-walled circular cylindrical shell in COMSOL Multiphysics and validate it against an analytical paper.

Here are the specific details of the problem setup:

Geometry & Physics: A finite-length isotropic cylindrical shell modeled using Love's thin shell theory.

Acoustic Environment: The shell is excited from the outside by an oblique incident harmonic plane sound wave. The transmitted wave enters an interior anechoic cavity (non-reverberant / inward-traveling waves only).

Boundary Conditions: The cylinder is Simply Supported (S-S) at both circular ends (v=0, w=0, N_{xx}=0, M_{xx}=0).

I plan to use the Acoustic-Structure Interaction (Frequency Domain) interface with the Shell module for the structure and Pressure Acoustics for the fluid domains.

Could anyone guide me on the best practices for setting up:

The Background Pressure Field for an oblique incident plane wave acting on the outer cylindrical surface.

The Perfectly Matched Layer (PML) or radiation boundary conditions inside the cylinder to properly simulate the anechoic core without artificial reflections.

The exact constraints in the Shell interface to correctly replicate the Simply Supported conditions for a vibrating shell.

Any model templates, expressions, or advice on capturing the stiffness, resonance, and mass-controlled regions accurately would be highly appreciated!

Update on my current COMSOL setup and the issue I'm facing:

To give more context, here is what I have already done in my model:

Geometry: I have modeled a thin finite-length steel cylinder. Around this inner cylinder, I created a larger concentric cylinder domain to apply the Perfectly Matched Layer (PML).

Material & Domains: Since both ends of the inner finite cylinder are open and the paper assumes air inside and outside, I have assigned Air to the entire domain inside the larger cylinder (which automatically covers both the exterior fluid and the fluid inside the open inner cylinder).

Physics & Coupling: For the Acoustic-Structure Interaction, I have selected both the inner and outer curved surfaces of the steel cylinder.

Loading & BCs: I implemented the Simply Supported boundary conditions on the shell edges. To simulate the external oblique plane wave, I applied a Boundary Load as a pressure distribution directly on the outer surface of the inner cylinder.

The Issue: While the simulation runs and I am getting graphs, the results are not convincing and do not match the analytical curves from the paper. I suspect there might be an issue with how I am applying the oblique plane wave as a manual Boundary Load instead of using COMSOL's built-in Background Pressure Field or Incident Acoustic Fields. Could anyone please point out the flaws in this approach or suggest how to correctly set up the incident plane wave and acoustic-structure coupling for this specific open-ended case?

Please help me, this is the paper link I am trying to get results in COMSOL:

Golzari, M., & Jafari, A. A. (2019). "Sound transmission through truncated conical shells." Applied Acoustics, 156, 186-207.

Configuring Sequential Frequency Transient study for heat simulation of antenna

Tue, 07 Jul 2026 17:13:40 +0000

I am modelling a microantenna structure on 2 silver wires connected with a lumped port. Using the frequency domain, I have been able to vary for different frequencies, and study the resulting microwave field.

I am currently trying to perform heat simulations. I would like to study how the temperature of the antenna varies in time for different frequencies. Unlike the usual RF heating studies, the lumped port input is manually configured. I thought the Sequential Frequency Transient study would help with it. However, the simulations appear to be independent of each other. After the first attempt, i obtain the error,

Feature: Stationary Solver 1 (sol15/s1) Failed to find a solution. Matrix has zero on diagonal. Returned solution is not converged. Not all parameter steps returned.

The second attempt shows me that the simulations are not influencing each other. The temperature study is independant of the frequency.

I have tried inputing the heat source as the power losses emw.Qh, to no avail. I'd appreciate any help.

The only way i was able to account for this was assign the frequency changes to the parametric sweep

Thank you.

Stop and resume steady-state calculation

Sat, 04 Jul 2026 18:08:17 +0000

When I perform computational fluid dynamics (CFD) simulations using ANSYS Fluent, I can stop the computation at any time to inspect the real-time results, and then resume the calculation—even making adjustments to settings and parameters if needed—with the default starting point being the result state at the moment the calculation was stopped. Why is COMSOL unable to achieve this functionality? For example, when solving a steady-state process, clicking "Stop" in the progress bar does not necessarily return the real-time results, nor does it allow the calculation to be continued.

Add primitives to an imported geometry and mesh

Fri, 03 Jul 2026 10:20:15 +0000

does anybody know how to add points to an imported geometry and mesh without disturbing the mesh (the geometry is imported from gmsh)?

polarization dependent transmittance

Thu, 02 Jul 2026 09:32:11 +0000

I want to illuminate the rectangular slab from one end and taking the signal from another end, by this i want to measure the transmittance. The light i wanna illuminate is the linearly polarized light i wanna rotate the electric field vector of electromagnetic wave making an angle theta from X axis and measure the transmittance after each 10 degree.

Using MATLAB to control Topology Optimization module in COMSOL

Thu, 02 Jul 2026 09:18:00 +0000

**Help needed in implementing MATLAB script to maintain control over COMSOL Topology Optimization iterations **

Hello everyone,

As the title suggests, I want to use MATLAB Livelink with COMSOL to implement Topology Optimization. The concept is simple. For beta parameter between 1 and 1024, I want to control the number of iterations with three methods: 1 - Max iterations depending on beta value (this works) 2 - Ending iterations early based on some arbitary convergence criteria (this also works) 3- Reusing solution from previous beta as the initial solution for the next beta ( This does not work).

I was wondering if anyone has any idea on how this can be implemented without using the parametric sweep feature of COMSOL as that would disable controls 1 and 2 mentioned above.

How to caculate Young's Modulus of Magnetic Film

Mon, 29 Jun 2026 08:15:39 +0000

I want to obtain the variation curve of Young's modulus of the magnetic film in magnetic thin film bulk acoustic waves with the applied external magnetic field. Does anyone have relevant expertise on this?

How to calculate group velocity

Sun, 28 Jun 2026 02:04:55 +0000

For the case in "Generation of Lamb Waves for Nondestructive Inspection of Plate Specimens", the results of wave number-frequency and phase velocity-frequency were obtained. How to calculate the group velocity-frequency result?

Modeling of Non-Linear and Anisotropic Magnetic Materials

Thu, 18 Jun 2026 13:43:14 +0000

My question is very simple : How can I model a non-linear and anisotropic magnetic material to represent a soft core like Cobalt steel ? Also, is there any database available that includes anisotropic B-H curves for this kind of material ? Thank you in advance !

CFRP Modelling in COMSOL

Thu, 11 Jun 2026 11:36:14 +0000

I am working on a model that uses structural mechanics and acoustics. I need to replace a pre-existing material in a thick plate-like structure with CFRP (Carbon Fiber Reinforcement Polymer). Is it possible to do that or do I need to build the model from scratch for that ? What are the steps that I should be following for this ?

Beginner COMSOL RF parallel transmission line

Thu, 11 Jun 2026 11:26:35 +0000

I am trying to model a parallel transmission line in several different materials at 300MHz in the electromagnetic waves frequency domain modal analysis. I have followed the comsol guide (Finding the Impedance of a Parallel-Wire Transmission Line) and have yielded the same results for air, however, when i change the surrounding material domain from air to water i get strange responses in electric field and my current value drops significantly (to the order of e-9), resulting in a massive increase in impedance (calculated using Z = V/I). I am also seeing very strange electric field distributions (i have looked at multiple modes).

So far i have tried the following, but nothing leads to an expected impedance: -changing the domain size -changing the int_H size for -int_H(emw.Hxt1x+emw.Hyt1y) -changing between perfect magnetic conductor and diffusion outer boundary conditions. -changing the conductivity of water to 0

I have attached some images for reference of my model, one for air (electric field strongest between conductors) and one for water. What might be causing low current? Why are the electric field distributions for air and water so different?

COMSOL batch recovery option

Tue, 09 Jun 2026 05:56:08 +0000

Hi,

since the documentation of the comsol does not really say much about it, i wanted to ask if someone has already used the -recovery option for comsol batch command? If yes, how specifically?

I was thinking something like this:

comsol batch -recover -inputfile "inptufile_that_crashed.mph" -outputfile "output.mph"

Or should i set the .recovery as input? It would have been nice if the documentation would have shown an example usage of this option.

Best regards, Zoltan

Double emulsion simulation

Tue, 09 Jun 2026 05:27:34 +0000

I have done my single emulsion simulation now I am trying the double emulsion simulation in COMSOL Multiphysics I am not getting the double emulsion droplets microfluidics.

High residuals in solid mechanics frequency domain simulations

Mon, 08 Jun 2026 17:31:17 +0000

I am relatively new to solid mechanics simulations. I am working on simple 2D frequency domain simulations of simple solid resonators, as shown below. (the blue part is silicone, the white part is steel with custom stiffness and density values). I apply a harmonic pressure load on the top surface(edge in 2D), and measure the surface response. The bottom surface(edge) is free, the side ones are fixed. The model works fine with relative tolerance of 1e-6 when the thickness of the model is large (ex: 5mm ). The width of the model is 5mm. Now when i reduce the thickness to say 0.5 to 1mm, the model is not able to converge to even a relative tolerance of 1e-3. 1)Could anyone guide me, how to debug this issue? I tried playing with the mesh, refining it further, but that doesn't work. 2)On a side note, is the aspect ratio of cells being close to 1 a stricter requirement in solid mechanics FEM than that in standard CFD problems?