Відео

I don't understand your question. Can you clarify or elaborate?

So your core is made of laminations that are separated by an insulator? I can think of two options: (1) I would suggest splitting the core into sections (using planes) to represent the laminations. Then insert 'surface current density' on the insulation planes and specify the current to be zero in the direction perpendicular to the laminations. (2) Define a custom material where the electrical conductivity of the core in the direction perpendicular to the laminations is close to zero. This simplification resembles the effect of the laminations wherein the eddy current cannot flow between laminations. Option 2 might be easier. Give it a try!

@@JafScience HI, thanks for your suggestion, actually the ansitropic material is the electrical conductivity matrix with zero on the direction perpendicular to the lamination. But, so far, I have tried several ways for do it (that's why I mentioned neubert paper) but the results are kinda bad, it's like the core is just a solid block . The other one maybe would work but It seems heavily in computational time and I don't think my pc would handle it. However, I would give it a try with few lamination to see what happen. Thanks for your suggestions!

You're right, you may significantly increase the number of mesh elements if you divide the geometry. Try using swept elements. When you say the results are bad do you mean electricity can still flow perpendicular to the laminations? Can you try putting a high electrical conductivity along that direction to see if the results look worse or the same?

@@JafScience when I say bad I mean like the skin effect on the window get worse, see when I simule a high eddy current scenario (open circuit) the flux gets spikes only on the window of the core and the rest of it its low flux. Thats the same effect that a solid core has when high conductivity is applied to them. However I will see what happen if I try the opposite and try to understand the phenomena

@@angelmachuca4696 Without seeing the set-up and results it's difficult for me to assess but the main purpose of the laminations is to reduce the flow of eddy current across the core. Not sure how that'll affect the skin effect. My knowledge in electrical machines in limited so I could be missing something.

If you're using the same physics in the 2D and 3D models, then the 3D model symmetry condition could be incorrect. Why are you using 3D if the 2D model works? Are you trying something more complex that cannot be reduced to 2D?

I thought of the following workaround: plastic strain= total strain (from results) - elastic strain. Elastic strain = yield stress divided young modulus.

@@JafScience Thank you so much! I want to get plastic strain results like Von mises stress in results section. but i couldn't

@@modrentechnologies9334 Von mises stress is not strain. You can plot total von-Mises stress and if you subtract yield stress you will get plastic stress.

Unfortunately I don't have the original source from which I extracted the information. But you can look up yield and tensile strength of specific aluminium alloys. if you get an estimate of the linear plastic modulus then you can use a best-fit curve to estimate the saturation exponent.

Okay, thank you​@@JafScience

You can message on linkedin.

In the video it was assumed that there is no air convection beneath the chip so no heat flux was added there. You may assume there is some heat transfer to other components through conduction. But using the same convective heat flux would not be realistic.

@@JafScience Yes that makes sense, thank you for replying!

There isn't a direct option for adding a switch but you may do it indirectly. For example, you can add a resistor in a circuit whose resistance changes from low (close circuit) to very high (open circuit). This resistance can be defined by a function (such as step) in the time domain.

Did you use an air volume? If all steps in the video are followed you should get the same results.

@@JafScience Sorry. I'm new to COMSOL ,I don't understand the air volume and which step would use the air volume

@@Nami_Jhon Air volume is just a box around the components and we define air properties for it. This would allows the magnetic field to spread in space.

@@JafScience Thank you very much!!!! I have completed the simulation and found the issue. The version of COMSOL I am using has some differences in settings compared to the version you used. Once I made the necessary adjustments, it worked.🥳🥳thank you for your help

Oh that happens sometimes. Good job on resolving it! I'm glad it worked out.

What are the questions you want to ask?

@@JafScience I am trying to simulate a biomechanical model for a seated human body.. I defin all the fixed joints and rigid domains and parametres.. but when i want to simulate the stress and strain I face some problems in comsol..if you can provide me with your contact to present or send the file to you.. i'll appreciaite that a lot

Can you send me a message on my twitter account?

@@JafScience yes sir send your account

Did you apply magnetic continuity boundary condition?

If it means, magnetic flux conservation, the answer is yes. There isn't magnetic continuity in COMSOL, only continuity in the pair boundary.

Yes the latter is what I meant. Because the nodes move with respect to each other, we need a boundary condition to compute the magnetic field between them. That's what continuity does.

In which component are you doing the swept mesh? Is it mapped or triangular? Can you try making it smaller?

@@JafScience It was mapped for the coil+ water.

Try triangular and smaller mesh on the face being swept.

@@JafScience Thank you.

I'm not aware of such capability. I never looked into it.

No it's not uploaded. You can create it from scratch.

For modelling Antennas I believe you need to use the RF module.

@@JafScience Will the RF module work with this geometry? I am trying to design a multi-turn loop antenna using a 200kHz bandwidth in COMSOL.

It can work with any geometry as long as the physics is defined properly.

@@JafScience If I use the magnetic field interface, will I be able to obtain a frequency plot?

@@JafScience when modelling a multi-turn loop antenna, I will have to use the magnetic field (mf) interface since I would have to adjust the amount of turns.

You're welcome! Glad you found it helpful!

Comsol has a simpler interface and a more efficient solver (takes less time to solve using the same resources as ansys). It's especially good with multiphysics applications since you can connect any number of different physics using the same interface (ansys is poor in this aspect). As for concrete models, I've never used them so I don't know but they have a broad library so I'd imagine the stuff you're looking for are included somewhere within the structural analysis module.

It's a revolution of the 2D diagram because the model is axisymmetric.

It looks like there's rigid body motion. Is the geometry constrained properly? Did you follow all steps for the contacts/physics/solver?

Thanks for your answer. I followed the same steps you did for the boundary load, contact, and physics, but I used different materials. The indenter, with a 2 mm radius, is made of PLA (polylactic acid) with a density of 1240 kg/m³, Young's modulus of 1639 MPa, and Poisson's ratio of 0.35. The sample, which is 5 mm tall and 10 mm wide, is made of NBR (nitrile butadiene rubber) with a density of 1000 kg/m³, Young's modulus of 4 MPa, and Poisson's ratio of 0.5.@@JafScience

Could it it be that the force is too much for these materials?

In this test, I used a maximum of 100 Newtons. And again the same error occurred.@@JafScience

Should I set "t1" in the parameters as shown in the error? Because in the "Definitions" part -->"Boundary System 1" --> "Coordinate names" - t1 (first), to(second) and n(third) are mentioned. Maybe that's why the error occurs. - Feature: Time-Dependent Solver 1 (sol1/t1) Undefined value found. ....@@JafScience

I don't currently have access to Comsol, but I can do so when possible.

I can't remember but learning involves experimentation with the software. You can also search for Comsol tutorials. They're available for free.

Thank you sir

Circuit voltage fixes the voltage and calculates current using Ohm's law. Circuit current fixes the current and calculates the voltage accordingly. In high frequency AC applications, the two options can give different results because the effective resistance of the current carrier changes with frequency (skin effect). In such cases, Circuit current preserves total electrical energy, Circuit voltage does not (energy is lost).

Thank you sir

What aspect of modelling do you need help with?

@@JafScience I have built a magnet pack that will go inside of oil and gas pipelines to pick up ferrous material. I am wondering if we could model the magnet pack to ensure I'm building it to pick up the optimum amount of ferrous material.

If you want to model small fragments of ferrous material, I'm not sure how to do that. If you want to study other related variables like how wide the magnetic field is and how strong it is in the region of interest, then that can be done.

Do you wan the magnet itself to be tilted or the magnetization inside the magnet is at an angle? If the latter is what you want then you can find the flux component along each axis and enter them. It should work.

@JafScience thanks for the reply. It is the first option. I want to take the magnet (which has its magnetization fixed along one of it's own axes), and then rotate the entire magnet.

@@ppxtb To rotate the magnet itself, go to 'transformations' under Geometry and select 'Rotate'. Then right click on Definitions, go to Coordinate Systems, and insert Base Vector System. This will allow you to define a new cartesian coordinate system that is oriented with the magnet. When specifying magnetic flux in the physics, choose the new coordinate system and let the magnetic flux direction be along the axis of the magnet.

It's true that you're selecting a 2D surface. But that surface is revolved to create a volume. So the density times volume gives mass in kg.

@@JafScience Got it. Thanks for your reply!