Induction heating of a steel rebar whitin mortar/concrete

Numerical methods and mathematical models of Elmer
Post Reply
aimar
Posts: 9
Joined: 12 Jul 2011, 13:33
Antispam: Yes

Induction heating of a steel rebar whitin mortar/concrete

Post by aimar »

Hi,
I am trying to simulate the axisymmetric induction heating of a rebar within mortar/concrete for multiple frequencies and rebar diameters (I attach my files). There is not need to be very accurate. I only want to check which range of temperatures can be obtained for each diameter and frequency.
I followed the crucible tutorial, but some results do not seem coherent.
I have created the geometry through a python file loaded in Salomé and exported as an .unv mesh (units in metres).
There are four bodies: the surrounding Air, the Coil, the Mortar and the Steel rebar.
The input data are the following:
- Angular Frequency: 50000 (Hz=1/s)
- Relative Permeability = 1 (for Air, Coil and Mortar)
- Electric Conductivity = 0 (for Air, Coil and Mortar)
- Relative Permeability = 100 (for Steel)
- Electric Conductivity = 6.99e6 (S/m = A/(V·m), for Steel)
- Desired Heating Power = 4400 (W)

One question: How can I foresee which will be the real power consumption. Each rebar diameter and frequency would couple differently with the magnetic field. If a machine can provide a maximum of 15kW, how do one guess how much is consuming really?

- Current Density = 103796.70 (A/m2, flowing though the Coil, obtained from the maximum power supply of the machine (15kW) and its work tension (230V), and divided for the area of the coil section [0.01 m diameter] multiplied by its 8 turns --> (15000W/230V)/((pi*0.01m*0.01m/4)*8)

So I assume that all the units are coherent. If I have introduce improper values, let me know please.

My problems:
- The Joule Heating reaches 8e8 (ºC, I suppose). Extremely high. Even for lower frequencies or desired heating powers.

- There are negative temperature values, also very high, -1.9e8. This issue was mentioned in viewtopic.php?f=3&t=4501, but I check that the crucible tutorial also raises negative temperature values. Is that correct?

Peter mentions there smaller timesteps but I am performing a steady state simulation as the crucible tutorial. It would be interesting a transient one, but first I must go step by step. He suggest adding some nice boundary layers, but my simulation should be very simple and the boundaries I think are enough far (0.5 m). Also he mentions higher order elements, but If I define those in Salomé mesh, Elmer shows a very weird mesh and I think it does not work properly with them. What can be the cause.

- I read in another thread (viewtopic.php?f=3&t=313) about adding manually the Magnetic Permeability of the steel, otherwise Elmer does not calculate it and assumes a value of 1, since the permeability of vacuum is not defined as a constant. Peter points out that it should be manually introduced in the .sif file as follows:
Magnetic Permeability = 12.57e5
but the results does not change. Is it not really involved in the solver?

Thank you in advance.

aimar
Attachments
elmer_rebar_induction.zip
(210.89 KiB) Downloaded 318 times
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Induction heating of a steel rebar whitin mortar/concrete

Post by kevinarden »

I would think if you want to heat the rebar with energy into the coil then you would have to conduct heat through the materials using the heat solver equation
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Induction heating of a steel rebar whitin mortar/concrete

Post by kevinarden »

Look at the elmerfem test inductionheating2 it uses heat solver to conduct temperature through the materials.
induction2.zip
(259.18 KiB) Downloaded 350 times
aimar
Posts: 9
Joined: 12 Jul 2011, 13:33
Antispam: Yes

Re: Induction heating of a steel rebar whitin mortar/concrete

Post by aimar »

Hi Kevin,
Thank you for the hint.
The idea is to heat only the steel rebar through induction heating. I though it was sufficient using MgDyn2DHarmonic solver and checking the Joule Heating. I also realize now that Joule Heating is not temperature (ºC) as I said, but power (W, W/m2 or W/m3?).
I do not care too much about the heat conductivity from the steel rebar to the mortar and that is why I dropped heat equation.
I will give a look to your suggestion.
Thank you very much.
aimar
aimar
Posts: 9
Joined: 12 Jul 2011, 13:33
Antispam: Yes

Re: Induction heating of a steel rebar whitin mortar/concrete

Post by aimar »

Hi,
I downloaded the elmerfem test inductionheating2, but it uses "StatMagSolve" that according to the Models Manual is obsolete.
Should I change it for MgDyn2DHarmonic?
And then I suppose that MagDyn2DPost is also necessary since it is there where I can activate "Calculate Joule heating".
And "ResultOutputSolve" is also called, but I am puzzled when I see "Exec Solver = never". Is it "never" executed?
Does MagDyn2DPost does the same task as ResultOutputSolve? I mean, save the results in vtu format?
Thank you.
aimar
aimar
Posts: 9
Joined: 12 Jul 2011, 13:33
Antispam: Yes

Re: Induction heating of a steel rebar whitin mortar/concrete

Post by aimar »

Hi all,
I could not make it. Even though I customize the induction2 example to my geometry and materials it does not converge. I suppose it should not be so different, but it fails.
Thank you anyway for pointing me the right way.
aimar
williams
Posts: 1
Joined: 12 Jul 2019, 16:45
Antispam: Yes

Re: Induction heating of a steel rebar whitin mortar/concrete

Post by williams »

Hi Aimar,
maybe give it another try with CENOS induction heating simulation software?
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Induction heating of a steel rebar whitin mortar/concrete

Post by raback »

Hi

The thing with BSolver is that it tries to compute the discontinuous Joule heating to continuous basis. You can also use MagnetocDynamicsCalcFields which can give the Joule losses also as DG fields. As the continuity is not enforced it maintains the positivity.

I think you should not divided the heat source by the number of turns.

I don't really know how the induction loop works. I could imagine that the current is not simple deduced from the P=UI. The frequency is not that of the wall socket either.

Attached is sif that generates the DG fields.

-Peter
Attachments
case.sif
(3.14 KiB) Downloaded 264 times
Post Reply