Latest Discussions - COMSOL Forums

Problems in Lunar Drilling Mechanical Model

Tue, 21 Jan 2025 19:43:47 +0000

Hi,

I am implementing a mechanical model to characterize the interaction between a drill and lunar regolith. I am basing the model in a paper that has analytical and test results so I can validate it later. From the simulations I need to retrieve the reaction Force and Torque between the drill and the regolith, considering that the former rotates and penetrates with defined velocities.

Initially I wanted to have a time-dependent study, but after trying for a couple of months without getting any results, I decided to move to a stationary case to assess both the force and torque at different depths. After some time I could make it converge and give what seems to be good results.

However, COMSOL does not allow to include a prescribed velocity of the drill in the stationary case (which makes sense) but I need to find a way to include a condition that recreates this velocity. I have tried to include a prescribed deformation, but I am not getting the results I expected.

I would like to ask for some recommendation to include a condition that recreates this penetration speed in the stationary case or, if possible, some advice to make the time-dependent solver converge. I am attaching a Google Drive link to the model in case you need it.

Thank you so much in advance!

(Link: https://drive.google.com/file/d/1GaoCKnW7yVVwVJ1-vpFyl1yktSUpQzp9/view?usp=sharing )

Piezoresistive Pressure Sensor Voltage Readout

Tue, 21 Jan 2025 19:03:38 +0000

Hi everyone,

I drew a piezoresistive pressure sensor with 4 ressitors (R1,R2,R3,R4) which are piezoressistive material. I want to build also a readout circuit with wheatstone bridge to simulate voltage output when pressure applied.

However, I cannot connect an electircal cirrcuit to solid materials in COMSOL. with electrical circuit module, it only defines nodes. but I have no idea which nodes will corresponds to my resistors in 3D solid sketch? Which module should I use to do this analysis? Can anybody help please?

![https://hizliresim.com/qjsaz59]

How do you use scaling factors?

Sat, 18 Jan 2025 15:16:36 +0000

I am trying to add custom boundary conditions to the built-in Pressure Acoustics (frequency domain) interface using weak contributions.

I see that the weak expressions in the pressure acoustics node contain a scaling factor equal to 1/omega^2.

I don't really understand the effect of including this scaling factor and don't know if I need to include it in the weak contributions that I want to add.

Can anyone shed some light on what the scaling factor is for and how to know when to use it, or point me in the direction of some references where I can learn more?

Thanks

Field lines penetrating through conductors in 2D electrostatics simulation

Sat, 18 Jan 2025 08:11:40 +0000

Hi all,

I'm seeing unexpected behavior in a basic 2D electrostatics simulation. Setup:

Top electrode: Straight line with center gap (set to 100V)
Bottom electrode: Parabolic curve f(x) = ax² (set to Ground)
Air domain between electrodes

Issue: Electric field lines are penetrating through the conductor surfaces, particularly:

Going straight through the bottom curved electrode
Curving up beyond the top plate surface

This seems physically incorrect - field lines should terminate perpendicular to conductor surfaces. Am I missing a key boundary condition or material property to properly define these as perfect conductors?

Currently using:

Electrostatics physics interface
Ground condition boundary condition on bottom curve
Electric Potential boundary condition (100V) on top lines

I have attached a picture of the plot which shows this issue. The field lines are clearly penetrating through the top and bottom conductors.

Any insights on how to fix this unwanted field penetration would be appreciated!

The returned solution does not converge

Fri, 17 Jan 2025 03:28:41 +0000

Hello, I would like to ask for some advice. I am currently using the electromagnetic field physics field to build an antenna with a frequency set at 4 GHz. I want to observe the antenna gain of this antenna at 4 GHz, but during the calculation process, it gets stuck at 70% and the calculation becomes very slow. Eventually, it reports an error. I coarsened the mesh division, but the result is still the same. I also modified the relative tolerance in the solver, but the result remains the same. Could you please tell me what factors might be causing this? Error: -The steady-state solver 1 (sol1/s1) failed to find a solution. The linear iteration diverged. The returned solution does not converge. Not all parameter step sizes were returned. How to solve it ? Thank you very much!

Form union ruining resolution

Thu, 16 Jan 2025 19:09:08 +0000

I have an electrode around which I put an air cube to simulate the e-field. I noticed that the last "Form Union" in the geometry node makes the previously curvy electrode full of edges, distorting the simulation results. I tried to adjust the repair tolerance to the smallest available value without success. Is there another way to keep the resolution of the geometry?

Lithium Plating Model

Thu, 16 Jan 2025 13:19:30 +0000

I am trying to simulate lithium plating model for porous electrode. I am getting solution for stationary case but not getting solution for time dependent case. What could be the reason?

Unable to Couple Tds and Semi Modules to model a MOSFET device

Wed, 15 Jan 2025 22:11:41 +0000

Hey everyone,

I'm trying to model impurity diffusion (Dopant) into Si substrate using Transport of Diluted Species (Tds) and the MOSFET device from the obtained diffused impurity profile using the Semiconductor (Semi) modules. I am able to successfully model dopant diffusion using the Tds module, but unable to transfer this dopant profile to Semi module to assign doping regions under the source and drain parts of the MOSFET device.

Here's what I tried: Tried to insert the concentration variable (comp1.c) from Tds module into the Semi module under analytical doping section and got the following error:

-Undefined variable.

Variable: comp1.c
Geometry: geom1
Domain: 6

Is the error due to "comp1.c" being an internally generated variable which the Semi module cannot access? Is there any other way to transfer dopant diffusion profile from Tds module to Semi module? Any hints would be greatly helpful. Thanks!

How to Assign Random Numbers for Particle Release at Inlet in Particle Tracing?

Wed, 15 Jan 2025 08:11:06 +0000

Hi everyone,

I'm working on a Particle Tracing for Fluid Flow simulation, and I need to assign random numbers as the input for the number of particles per release at the inlet. My goal is to generate a different random value for each release interval.

Is there a way to achieve this directly in COMSOL, perhaps using expressions or built-in functions like rand()? If not, are there alternative approaches, such as defining variables or importing external data, to accomplish this?

Any advice or examples would be greatly appreciated!

Thanks in advance!

Unexpected discrepancy when solving a plate bending problem using weak form PDE module

Wed, 15 Jan 2025 02:39:22 +0000

Hi there,

I am trying to solve the deformation of a boundary clamped thin circular plate bearing uniform pressure. This problem possesses axisymmetry, but I also tried to solve it in 2D space (model 2D) besides in the 1D axisymmetric space (model 1D) for validation. Please note that only the Kirchhoff-Love plate model is employed in model 2D, model 1D, and the analytical solution.

The value of bending stiffness is assigned directly, instead of calculated through Young's modulus, so and Poisson ratio in this case are considered independent.

It seems there is something wrong with model 2D: i) the model is hard to converge; ii) even if it converged, the deflection would be larger than theoretical prediction and show dependence on Poisson ratio. To be specific, the center deflection predicted by model 2D monotonically decreases toward the analytical solution as Poisson ratio decreases from 0.5 to -1.

By comparison, model 1D works well. It agrees with the analytical solution regardless the value of Poisson ratio.

I would appreciate if you could give me a hint.

Best,

HC L

If we take a=2, q=1, D=1000, and an arbitrary nu, the center deflection is theoretically 2.5e-4.

details about model A (the picture is also attached below):

details about model B (the picture is also attached below):

Post was updated on 2025.01.16 to provide more details.

Unexpected Displacement Change in Gear and Rack Calculations

Mon, 13 Jan 2025 23:36:04 +0000

Hi everyone,

When analyzing a simple straight gear and rack model, I applied a fixed displacement to one of the gears to drive the entire system. The goal was to validate the correctness of the model.

However, I encountered Unexpected Sudden Displacement Changes, as shown in the figure below. For reference, a sketch of the system is also attached.

Please let me know if you need any additional information. I would greatly appreciate any guidance on resolving this issue. Thank you!!

How to make a uniform magnetic flux density

Mon, 13 Jan 2025 16:55:14 +0000

Hi all, I'm trying to model a magneto-mechanical interaction (my focus is on the structure) and I need to apply a uniform magnetic flux density in a cube of air (which is my domain) in the z direction. Is there a way to do it? I've tried different things (magnetic potential, magnetic field and others) but they don't work. Thank you for your help

How should I set up the MUMPS solver?

Mon, 13 Jan 2025 09:38:00 +0000

In my solution model, the degree of freedom to be solved is 26032722, and the physical memory of the workstation used is 512G. However, when using the mumps solver, it always utilizes virtual memory, and the usage of virtual memory exceeds the physical memory, even if there is a substantial amount of free physical memory. This greatly slows down the speed of computation. How should this issue be addressed?

Transformer oil electrode configuration model

Mon, 13 Jan 2025 05:47:10 +0000

Hello, I am working on a model involving transformer oil using the following physics interfaces in COMSOL:

Electrostatics - for simulating the electric field. Transport of Diluted Species - to analyze charge movement and density. Heat Transfer - to incorporate the thermal effects in the oil. Despite setting up the physics, I am unable to observe charge density or the progression of charge movement in the oil. I suspect the issue lies in assigning the appropriate parameters, boundary conditions, or coupling mechanisms between the modules.

Could you please assist me in configuring the model correctly? Specifically, I need guidance on:

Defining charge density and its initialization in the Transport of Diluted Species module. Proper coupling between Electrostatics and Transport of Diluted Species for space charge effects. Setting suitable material properties (e.g., permittivity, diffusion coefficients, reaction rates). Any tips for observing charge density changes in the oil. I have attached a picture for reference that I want to model like that. I look forward to your advice and suggestions to resolve this issue.

Best regards,

Guidance on Setting Up Ring Modulator Simulation (Semiconductor + Wave Optics Modules)

Sun, 12 Jan 2025 18:07:29 +0000

Hello everyone,

I am currently working on simulating a silicon microring modulator and would appreciate some guidance. My ultimate goal is to analyze the device's electrical and optical behavior, including carrier dynamics, electric field distribution, and the overlap between the optical mode and the depletion region. I have already gone through the PN_Junction 1D example and some tutorials on the Wave Optics Module, so I have a basic understanding of how these modules function.

Here’s what I have done and what I am trying to achieve:

What I Have: Simulation Goals:

Model a microring modulator with Z-shaped doping and calculate: The depletion region under reverse bias. The electric field distribution. Optical mode overlap with the depletion region. Use the results to estimate modulation efficiency (e.g., ). Background Preparation:

Completed the PN_Junction 1D example to understand carrier dynamics and doping profiles. Practiced using the Wave Optics Module to simulate mode profiles in waveguides. the paper i want to reproduce is : DOI: 10.1038/s41467-024-45301-3

My Current Plan: Geometry Setup:

Create a 2D model of a microring with a Z-shaped doping profile. Include p-doped, n-doped, and intrinsic regions with appropriate geometrical dimensions. Doping Profile:

Define doping levels for the p-type and n-type regions (e.g., Use the Analytic Doping Model or custom doping profiles. Electrical Simulation:

Apply reverse bias to simulate the depletion region and calculate the electric field distribution. Optical Simulation:

Use the Wave Optics Module to calculate the optical mode overlap with the depletion region. What I Need Help With: Combining Physics:

How should I combine the Semiconductor Module (for electrical analysis) with the Wave Optics Module (for optical mode simulation)? For example, how can I transfer the depletion region and electric field results to the optical simulation for overlap analysis? Modeling Z-shaped Doping:

What’s the best approach to define the Z-shaped doping profile? Should I use the Analytic Doping Model, or is there a better way to implement such geometry-specific doping? Boundary Conditions:

For the electrical part, what boundary conditions should I use for the metal contacts and the waveguide cladding? Are there any best practices for ensuring accurate results in depletion region modeling? Performance Tips:

Are there any specific meshing or solver settings I should pay attention to when dealing with a microring structure (e.g., fine meshing in the depletion region or specific solvers for coupled physics)?

Extracting data from Global 1

Sun, 12 Jan 2025 10:46:32 +0000

Hi everyone, I would like to extract the data from Global 1, which contains two line graphs (Voltage 1 and Voltage 2), into one file. I need the output to include three columns: one for the x-axis values and two for Voltage 1 and Voltage 2, respectively. Thank you for your assistance! Cordially,

Auxiliary data refresh

Sat, 11 Jan 2025 13:56:12 +0000

Hello everyone, the button to refresh all nodes doesn't work, is it normal ? I have to refresh in every single node.

Only me ?

Layer addition on additive manufacturing (SLM) process

Sat, 11 Jan 2025 10:38:07 +0000

I'm trying to model a SLM process on COMSOL and I'm finding some issues on simulating the layer addition. For an initial test, I have the substrate and 2 more layers (the laser travels 3 times). In order to simulate their addition, I created piecewise functions that are 1 when the layer is active and 10e-15 when it's innactive, and I multiplied the physical properties of the material by the function. See images below to make it clearer. The GIF shows a zoom on the layers to better see the laser pass. However, when I use a point probe to measure the temperature inside the layers, it increases even when the layer in question is not supposed to be active. The plot below show the temperature in 3 points. Laser pass time is 0.36 seconds.

For the boundary conditions, the bottom surface has a fixed temperature (ambient), the sides have convection, the top has radiation to the ambient and also convection (but the convection I could only select the last layer surface, the software doesn't let me select the others). But even if theses conditions are not really correct, I think that the points I presented should have the increase in temperature they are showing when the layer is inactive, as you can see I multiplied the material properties by the function.

Does anyone have an idea of what is wrong?!

LiveLink for MATLAB: mpheval evaluates data far scarcer than mesh

Sat, 11 Jan 2025 05:04:11 +0000

It seems that mpheval evaluates data with a density far scarcer than that of the actual FEA mesh. How to control the density on which mpheval evaluates? Same question for mphinterp.

Guidance on Adaptive Closed-Loop Stimulation Current Simulation

Fri, 10 Jan 2025 19:25:23 +0000

Hello,

I am working on a simulation project involving adaptive closed-loop stimulation in using LiveLink to integrate feedback from the system to control the terminal source (electrode stimulation). I would appreciate any insights or guidance from the community. Specifically, I am interested in simulating an electrode-based stimulation system (interpolated waveform function as a terminal) where the current is applied adaptively based on feedback from the system. The feedback could be related to electric fields, voltage changes, or other measurable parameters.

Key questions: * Can anyone provide insights on implementing adaptive feedback control in COMSOL, particularly how to modify the electrode stimulation dynamically in response to field data (e.g., based on measured voltage)? * Are there any known issues or common pitfalls when solving time-dependent simulations that involve feedback from the electric field or voltage measurements?

If anyone has successfully implemented adaptive control systems or closed-loop stimulation in COMSOL using LiveLink for MATLAB, I would love to hear about the methods or algorithms you used.