Latest Discussions - COMSOL Forums Most recent forum discussions Tue, 26 Mar 2019 13:38:32 +0000 COMSOL Forum: Latest Discussions Scatterer on substrate in RF module <p>The below post is related to an <a href="//">archived discussion</a></p> <hr /> <p>Hi!</p> <p>I try to reproduce results from the "Scatterer on substrate" model using RF module in Comsol 5.0. I have set up the model and it converges in 5 iterations with BiCGStab. However, when it comes to plotting the resulting fields it is completely wrong (even when I change the particle geometry to something else - which is actually what I need). It is not even the question of gold properties (I import J&amp;C which is quite similar in Vis to Rakić's results). Please, help me with this. I attach the file.</p> <p>Cheers</p> <p>Radek</p> Tue, 26 Mar 2019 13:38:32 +0000 4.2019-03-26 13:38:32.230681 Vanadium Rdox flow battery <p>I am a new user of Comsol. I'm working on the modeling of Vanadium Rdox flow battery.</p> <p>Please how do incorporate the local mass transfer coefficient, and source terms for species and charge in my model. Which physics will I used, particularly to add local mass transfer coefficient?</p> Tue, 26 Mar 2019 11:27:33 +0000 4.2019-03-26 11:27:33.230673 Adoptation and error estimate <p>Hello dear, I am trying to model a high mach number flow and as a part of that I want to use adaptation and error estimates in the stationary solver for refinement of the mesh where it is needed, however I face with aproblem which I cannot understand. In "Adaptation and error estimates" section in "Stationary solver" I used "L2 norm of error squared" and for refinement method I used either "Longest" or "Regular", and for both of these choices I changed the "Element growth rate" from "1.7" to any number (Lower or Higher). But no matter I changed this number after solving and showing the refined mesh in the Mesh section, the element growth rate which used to refine is "1.7". I wonder why this number stay constant however I change it. Thanks in advance. Best regards, Saleh</p> Tue, 26 Mar 2019 08:55:19 +0000 4.2019-03-26 08:55:19.230671 How to model a sharply changing BC <p>Ok, so I have a laser source that turns off at a particular point in time. I need to model the laser, heating and heat distribution until it completely cools off. I can do this with multiple steps however it makes it impossible to do a parameter/auxiliary sweep (As far as I know). So I want to use a variable BC, but I think COMSOL sees high gradients and decides that it needs to turn it off sooner and in steps (Without telling me, I might add). Is there anyway to model this? I thought a shorter time-step might allow me to achieve this. However,the duration is too long to reduce timestep for the whole run and COMSOL simply ignores my variable timesteps (example below) and only uses the initial value of the function throughout the simulation.</p> <p>Timestep Function Example (ignore the fact that it is needlessly complicated):</p> <p>1<em>(0.25</em>(t[1/s]&lt;(Duration-20))+0.75<em>(t[1/s]&lt;(Duration-10)))+(0.05</em>(t[1/s]>=(Duration-10)))+0.95*(t[1/s]>(Duration+10))</p> <p>P.S. Maybe I should mention that by timestep I actually mean the "Study Settings>Times" section.</p> Tue, 26 Mar 2019 07:48:47 +0000 4.2019-03-26 07:48:47.230653 How to trace particles in a time dependent flow. <p>My simulation is <strong>time dependent</strong> (Sinusoidal Velocity Inlet, zero pressure outlet). First, I solve the laminar fluid, then I want to <strong>trace particles</strong> in this transient velocity field. But all particles just locate at the inlet without any motion.<br /> I also search some examples, but their fluid flow is steady. So I wonder whether comsol can solve the question like mine.<br /> Thanks</p> Tue, 26 Mar 2019 07:43:08 +0000 4.2019-03-26 07:43:08.230643 Extract faces (surfaces) of an imported 3D geometry <p>I imported a 3D geometry (STEP file) in COMSOL and want to use shell elements for the interested faces for eignefrequency analysis.</p> <p>Is there a command that allows me to extract specific faces of the 3D model? Or a command explodes the model.</p> <p>Thank you.</p> Tue, 26 Mar 2019 05:42:53 +0000 4.2019-03-26 05:42:53.230633 How to create a function or variable based on results from component 1? <p>Hi all,</p> <p>I am calculating the stress change due to several point loads in the subsurface designed mainly in component 1 (solid mechanics module). I am expecting to calculate some other parameters which are related to the stress tensor (sig11, sig12, sig13, etc) based on the results of those stress components from component 1. I have my own equations to accomplish the above calculation.</p> <p>My questions are:</p> <ol> <li><p>Should I create a new component 2 and include those own-designed equations for following on calculations or should I just define a variable in global definitions in order to use results from component 1 (comp1.sig11?)?</p></li> <li><p>I have multiple points designed in the geometry. For each point, I have a loading force. Instead of manually adding several point load in the physics. Can I possibly adding different loading forces for different points automatically from reading a file? I can easily do this in matlab but I am a beginner in comsol. So I am looking for useful suggestions!</p></li> </ol> <p>Thanks, Shanshan</p> Tue, 26 Mar 2019 03:55:28 +0000 4.2019-03-26 03:55:28.230632 Nonlinear solver did not converge. <p>Hello,</p> <p>I made a fully coupled model of two physics (Darcy's law and Heat transfer in porous media) for 2D horizontal geothermal heat exchanger (i.e., coupled heat and moisture transfer in soil under thermal gradient). Once I ran the model, error messages came up regarding convergence issues with time dependent solver (i.e. nonlinear solver did not converge). I studied many resources in COMSOL website and followed the tips that posted (e.g. change method to ease tolerance factor). I've also learned that coarse mesh may cause the convergence issue. I thus set extremely fine mesh, especially near the boundaries between domains. Even the model was running for overnight, I still got same error messeage. Does anyone know the solution to fix this issue? I am looking for a way that I can stay with my normal mesh and run the model without error.</p> Mon, 25 Mar 2019 21:29:11 +0000 4.2019-03-25 21:29:11.230622 Setting last time step as initial conditions of next thermal model run <p>Hello,</p> <p>I've been trying to create long time scale thermal models (2000yr+) by running them in 100 year segments and using the final time step to start the next 100 year run, saving each 100 year run in its own file. For some reason using the final time step of a 100 year run (with 1 day steps) produces the following error:</p> <p>There are 53046 void equations (empty rows in matrix) for the variable at coordinates: (64.0549,-0.0610404), (62.5617,-1.55478), (62.5627,-4.54318), (64.056,-59.9383), (62.5614,-58.4453), ... and similarly for the degrees of freedom (empty columns in matrix). Last time step is not converged.</p> <p>If I run the model for only 60 days it has no problem using the final time step for the next set of 60 days and I can repeat the process for as many times as I would like.</p> <p>Currently I have it structured to delete all but the final time step from the previous solution by using "combine solution" to save memory and speed it up, this is not an issue for the 60 day model.</p> <p>Does anybody have an idea as to how I can fix this? I've been having this issue for weeks.</p> <p>Thanks!</p> Mon, 25 Mar 2019 19:32:55 +0000 4.2019-03-25 19:32:55.230612 RFID_TAG <p>hello sir,I wanted to design a RFID Antenna tag in Comsol software. I need the proper guideline for this field so that i can design it</p> Mon, 25 Mar 2019 06:34:19 +0000 4.2019-03-25 06:34:19.230583 Current Source Question <p>Hello all, I'm trying to apply a current through 3 electrodes to my geometry, it works fine with ac sources, however when I try to change the current source to Pulse it crashes and gives the following error: Undefined variable. - Variable: t - Global scope Failed to evaluate operator. - Operator: if - Global scope Failed to evaluate expression. - Expression: if(t&lt;,0,comp1.cir.pulse((,,, Failed to evaluate variable. - Variable: comp1.cir.I1_isrc - Defined as: (if(t&lt;,0,comp1.cir.pulse((,,,<em>comp1.cir.I1.value)+comp1.cir.I1.offset Failed to evaluate variable. - Variable: comp1.cir.I1_i - Defined as: comp1.cir.I1_isrc Failed to evaluate variable. - Variable: comp1.cir.gnd1_p_i - Defined as: comp1.cir.I1_i Failed to evaluate variable. - Variable: comp1.cir.I1_p_i - Defined as: comp1.cir.gnd1_p_i Failed to evaluate variable. - Variable: comp1.cir.termI1_i - Defined as: -comp1.cir.I1_p_i Failed to evaluate variable. - Variable: - Defined as: comp1.cir.termI1_i Failed to evaluate expression. - Expression: d(</em>test(,{test@1}) Failed to evaluate Jacobian of expression. - Expression:*test(</p> Sun, 24 Mar 2019 11:47:21 +0000 4.2019-03-24 11:47:21.230542 Fluid damper simulation <p>I go through the fluid damper simulation .</p> <p>I need piaton motion for different time .</p> <p>But in simulation it cannot show me piston motion . It only shows the velocity change.</p> <p>So kindly provide me the detail how to apllied piston dispalcement for different time</p> Sat, 23 Mar 2019 22:00:21 +0000 4.2019-03-23 22:00:21.230531 DEP effect on GUVs <p>i have generated the Gaint unilamellar vesicles from DPhPc lipid rehydrated with sucrose of PH-7. i need to check the DEP force on GUVs of different diameters. is it possible to do it ? and how to find the DEP parameters too for simulating this ?</p> Sat, 23 Mar 2019 14:33:27 +0000 4.2019-03-23 14:33:27.230521 Segregated group 2 Failed to evaluate variable. - Variable: material.mur13 <p>I'm simulating magnetic field in a current carrying conductor. I have a 3D model and the following are the sub-modules included: Ampere's law Magnetic Insulation initial Values and External Current Density -> this is inmported from the Jouls Heating module.</p> <p>Just the jouls heating works fine. but i have issue when I include heating. I get the following error I get: Segregated group 2 Failed to evaluate variable. - Variable: material.mur13 - Geometry: 1 - Domain: 1 Failed to evaluate variable. - Variable: - Defined as: material.mur13 Failed to evaluate variable. - Variable: - Defined as: (((<em></em><em>(</em><em>(</em><em>(</em><em>(</em><em>(</em><em>(</em></p> <p>Failed to evaluate variable. - Variable: - Defined as: (((<em></em><em> Failed to evaluate expression. - Expression: d((</em>(<em>test(mod1.curlAx)</em>test(mod1.curlAy)<em>test(mod1.curlAz)))</em>(dvol),{test@1})</p> <p>any light on this will be much appericiated.</p> <p>Thank you!</p> Fri, 22 Mar 2019 15:56:29 +0000 4.2019-03-22 15:56:29.230501 Multiturn Coil Voltage Excitation: Problem with Coil Current Results <p>Hello everyone</p> <p>I'm trying to simulate a multiturn coil with an iron core, with a certain BH curve. Simulations run well, but for the same applied voltage, but different coil resistances, it always gives me the same current.</p> <p>I've tried to use the electrical circuit physics as in the E-Core Transformer example, but the results are the same.</p> <p>Can anyone help? How is the coil current calculated?</p> <p>Thanks in advance</p> Fri, 22 Mar 2019 11:21:47 +0000 4.2019-03-22 11:21:47.230471 Frequency Response for a clamped beam with many vibration absorbers. <p>Hello all,</p> <p>I have a thin plate modeled as 3D in solid mechanics module. On the top of this plate, I have places a mass-spring resonator (vibration absorber) (please see the attached figure) where I have modeled it in COMSOL using the lumped mechanical system. The helix spring is just for schematic where the actual 3D plate is shown in second figure. I coupled the two systems using the “External Source” from the lumped mechanical system and I exert the reactions from this external source to the solid 3D plate. The eigenfrequencies obtained from the eigenfrequency study matched perfectly with the in-house FE code. However, now I am trying to study frequency response to a harmonic force of an amplitude 1 while changing its frequency. Since I am applying the reaction forces from the external source of the lumped mechanical system to the 3D solid plate, how can I apply the harmonic force?</p> <p>I tried to add something like this:</p> <p>“-lms.E136_f2 + linper(1)”</p> <p>where “-lms.E136_f2” is the reaction force from the external source and “linper(1)” is my harmonic force but still I could not succeed. I will be really appreciated if I can find an answer.</p> <p>Thank you</p> <p>Abdullah</p> Thu, 21 Mar 2019 17:03:39 +0000 4.2019-03-21 17:03:39.230441 View Factor Computation (?!) <p>Has anyone succesfully computed view factors in a 2D geometry yet? I have tried to follow this blog:, but when trying this within a square cavity, I get either errors or zeros with the used function from the blog.</p> <p>Sharing your succesful model or steps for implementation would be appreciated a lot!</p> <p>Kind regards,</p> <p>Marvin</p> Thu, 21 Mar 2019 15:08:03 +0000 4.2019-03-21 15:08:03.230421 Regarding Rosette configuration <p>Hello all,</p> <pre><code> How to make strain mesurement using Rosette Configuration in Comsol, If known please let me know the procedure. </code></pre> Thu, 21 Mar 2019 14:21:17 +0000 4.2019-03-21 14:21:17.230411 Create inner surface by heat transfer <p>Dear Users,</p> <p>I need your help with a hopefully simple thing. I have two geometric bodies. The problem is that the surface between them dont exist so I can´t create a thin layer to simulate the heat transfer from them. Do you have an answer as to how it is possible to create an inner surface or just divide one surface into two parts?</p> <p>Thank you for your help!</p> Thu, 21 Mar 2019 11:16:34 +0000 4.2019-03-21 11:16:34.230401 Question regarding the calculation of area <p>Hi all,</p> <p>I have a 3D geometry and I was trying to get the boundary area of certain portion as an input for the further calculation.</p> <p>As suggested, there are two ways to get the area value: 1. Use measure under the geometry tag to get the boundary area. 2. Use intergration function for this boundary (such as expression intop(1)) to get the area.</p> <p>However, I found the values from these two method are a bit different. For example, for a circular plane, from integration method it got 9.7193E-12 [m^2], but from the measure methode it showed 1.0169E-11 [m^2]</p> <p>Why there is a difference? And which one should I trust?</p> <p>Thank you</p> Thu, 21 Mar 2019 09:26:19 +0000 4.2019-03-21 09:26:19.230392