Heat Solver Simulation

PrateekshaPrateeksha Member Posts: 3

Hi,

1) I tried to follow following example for simulating my structure:

But I am seeing constant temperature profile for my structure. What might be the reason?

2) Can we define Graphene as thermal conductor as well as heat source simultaneously? 

3) Is there any way to calculate thermal boundary conductance (TBC)

at the interface of two materials?

4) In the following example few layers of Graphene is 50 nm thick:

https://support.lumerical.com/hc/en-us/articles/360042273534-Transient-heat-flow-in-a-graphene-coated-glass

What thickness should be taken for single layer Graphene? Since heat solver is not having any option of 2D Graphene.

5) How does heat solver work for periodic structures? Since, I am simulating only unit cell.


Thanks

Prateeksha

Comments

  • kghaffarikghaffari Posts: 45Ansys Employee

    Hi Prateeksha,

    1-     So similar to the example, you are using a uniform heat source, is that correct? I would recommend confirming that this is being applied as expected and the value is large enough to make a considerable impact in the temperature profile. You can use the “boundaries” results in the solver to check the total power dissipated (through substrate, convection, etc.) which should match the expected input power set in your source.

    2-     When applying uniform heat source to a volume, the volume itself will also be included in the simulation (unlike if thermal boundary condition is used); meaning the thermal conductance of the material should also be relevant.

    3-     My understanding is that the temperature drop related to TBC is due to nanoscale molecular dynamics at the interface, which is not simulated in HEAT. If you already have an estimate of its impact, as a workaround you can define a buffer layer (between the two materials) with a thermal conductance such that the temperature drop will match your expectations.

    4-     In the example the 50 nm thickness replicates the few-layer graphene (FLG) mentioned in the reference work. I would suggest testing the same approach by further reducing the thickness as needed. You might need to add local mesh constraints to ensure the smaller features are appropriately simulated.

    5-     The simulation boundaries in HEAT assume symmetric boundary by default. So if your periodic structure is also symmetric then you can simply simulate a unit cell without any need of defining any special boundary conditions. For instance, this approach has been used in our Interleaved PN modulator example.

    Kind regards,

  • PrateekshaPrateeksha Posts: 17Member

    Hi,

    Thank you for your reply!!

    I tried to follow your instructions, but unable to figure out the issue. Could you please check my simulation files? I want to simulate a periodic grating with periodicity of 800 nm and check the thermal profile using Graphene heater.

    In one case, I used 800 nm of simulation region (Reflector_unit_cell.ldev) and used power boundary condition to provide input power, then I am getting variation in temperature profile, but error in parametric variation.

    In other case (Reflector_edited_bigger.ldev), I expanded the simulation domain larger than unit cell period, but then it's giving a job error. Here, I used uniform heat source.

    Is there any difference in "power boundary condition" and "uniform heat source"?

    Kindly check my simulation files attached.




    Looking forward to your reply!!


    Thanks

    Prateeksha

  • PrateekshaPrateeksha Posts: 17Member

    Hi,

    Have you checked my simulation files? If not, kindly check it. I really need your feedback.


    Thanks in advance!!


    Thanks

    Prateeksha

  • PrateekshaPrateeksha Posts: 17Member

    Hi,

    1-     So similar to the example, you are using a uniform heat source, is that correct? I would recommend confirming that this is being applied as expected and the value is large enough to make a considerable impact in the temperature profile. You can use the “boundaries” results in the solver to check the total power dissipated (through substrate, convection, etc.) which should match the expected input power set in your source.- What value is considered large enough for "uniform heat source"? I can see so many results in boundary and unable to figure out the exact match with input power, how to do that?

    2-     When applying uniform heat source to a volume, the volume itself will also be included in the simulation (unlike if thermal boundary condition is used); meaning the thermal conductance of the material should also be relevant.-I gave uniform heat source to a volume (i.e. Graphene) but even after I change material to Si or Silica, then also thermal profile doesn't change. It's not considering the thermal properties of material. I cross verified with the Lumerical example also. Is it correct?

    3- I used "power boundary condition" in place of "uniform heat source" in the Lumerical example https://support.lumerical.com/hc/en-us/articles/360042833673-Thermally-tuned-waveguide-FDE- and it gives same temperature profile. But it doesn't show same behavior for my structure. In my structure heater is smaller than waveguide and placed just above the waveguide but in the given lumerical example, it's farther and larger than waveguide, please check the attached files. What is the difference in "power boundary condition" and "uniform heat source"?

    4 How the results are dependent on "norm length" in 2D simulations? It's varying with "norm length".

    5 I am simulating following structure. I attached ppt as well as lumerical file. Kindly check it.


    Kindly reply to my queries.

    Thanks

    Prateeksha

  • kghaffarikghaffari Posts: 45Ansys Employee

    Hi Prateeksha,

    We are not allowed no download attachments on the forum. That being said for your first inquiry about the error, please make sure that:

    1) All (heat) properties are set for your materials. If missing (set to 0) this could lead to job error.

    2) When sweeping, extreme powers could mean drastic temperature increase and difficult convergence. Please try using smaller power range in your sweep in the beginning to observe the peak temperatures in your system. At the same time you can try increasing the iteration limit in your advanced settings to help with the convergence.

    Kind regards,

    Khashayar

  • kghaffarikghaffari Posts: 45Ansys Employee

    Regarding your more recent inquiry:

    1. All the generated power will be dissipated through the boundaries. Since heat source is not a boundary it will not be listed but you can rely on the sum of all dissipated power to obtain the generated power. In the list, items with A correspond to area and P correspond to power; you can further discern the boundaries by their labels (e.g. convection, temperature, etc.) The power enough to create temperature increase will depend on the specifics of your project such as size, and material properties.
    2. The material is accounted for when using uniform heat source. If you are not seeing any difference when changing materials it might be that: a) lock mesh option in the solver is enabled, so the changes you make are not applied to the simulation b) again, power might be too small for the system so the change in peak temperature is not noticeable.
    3. The power boundary will be applied at the specified surface and unlike the uniform heat source the material within is not accounted for in the simulation. Like you mentioned this could mean different final results depending on the specifics of your design.
    4. When using uniform heat source, the total power specified is attributed to the volume object which depends on the norm length. You will notice that norm length is one of the settings for the uniform heat source. So if the norm length is changed without the total power being updated, I’d expect the results to be different.
    5. We are not allowed to download attachments on the forum, however the settings you share in the presentation seem correct.

    Kind regards,

    Khashayar 

Sign In or Register to comment.