I suggest you make two tests under different unit system to justify the problem (Model can be a very simple one). If the parameters were correctly set up, the two simulation should provide the same temperature result. If not, unit conversion for some other parameters are missing, e.g. bolzmann constant and latent heat  are also dependent on unit system. On the other hand , if you got the same results for the two units, I guess the heat modeling may need to be refined.

@Hui Huang. Thank you for your response. As I mentioned in introduction, I cannot draw the part in M-Kg-N system because of its small dimensions. I mean, the millimeter size shape cannot be drawn easily in metric unit for some particular shapes. So, for this reason, I just have one follow my simulation in mm-Tonne-N system.

You mentioned about bolzmann constant or latent heat. Maybe it works. Can you explain about them more? 

I mean you can design another arbitrary model such as a cube with edges 0.5m, et al. Then define any heat flux, eg. on surface or in volume. The objective is just to make a comparison between results of two systems. The bolzmann constant will be used if you define radiation, latent heat will be considered if phase transformation occurs. Perhaps in your case, these may not have been activated. Hopefully you can solve the problem soon.

@Hui Huang: Actually, my modeling is radiation. So, I try to apply Bolzmann constant and let's see the results.

@hui Huang: I applied Bolzmann constant, but the results doesn't change at all! That's interesting!!

By the way, do you ever check my conversion units in the attachment?! I think, It should be related to them. 

Yeah, I checked your units, they seem correct. My experience is something like this. I did not change the unit of energy Joul but just convert m-Kg to mm-g, it works fine for my heat conduction analysis (the two produce the same temperature results). If possible, can you share the input file of your simplied heat model? Thank you.

@Hui Huang Thank you so much for your following up.

As you mentioned, I Just change m-kg to mm-Tonne (you mentioned mm-g wrongly), and I didn't change J to MJ or W to mW (units of energy). So, some parameters such as Specific Heat (j/kg.C to j/Tonne.C * e3), Conductivity (w/m.C to w/mm.C * e-3) and Heat Flux (w/m2 to w/mm2 *e-6) are changed. The final results of NT11 for the first 4 steps have not changed and they keep the last result with E3 exponentiation (For Ex. 7000 C temp.). Anyway, I'm running the program for all steps, now and I will inform you for the final obtained results. Also, I attached the CAE file in your inbox to be checked. Thank you again

Thank you, Mehrshad Mehrpouya, for your data and results. As you mensioned, the first 4 steps have not changed, the unit conversion would be no problem. And there could exist some heat modeling issues regarding the order of results. I am sorry my abaqus version is a lower one, so could you send me the inp.file only for loading your model.

@hui Huang of course. Check your inbox please for the both Inp.file and ODB.file.

As you know that the material becomes soft at high temperature, we need the data for Sy and E at these temperatures. Generally, the yield stress and young's modulus have the similar dependency with temperature (the values decrease by increasing temperature). You input the stress-strain curve at 20C correctly. The example data line which I gave is just a bilinear one (only two lines) for simplicity, as we are usually lack of the information at those high temperature. At the plasiticity data line, what you defined are yield stress (plastic strain=0.0), and tensile strength (plastic strain=0.1?). You can also define the yield stress only, which means no strain hardening. If you can find those exact values on literature, that would be the best; otherwise, you can deduce them from the tendency of elasticity. The publication below has the properties for carbon steel, I hope it is useful for you.

https://www.researchgate.net/publication/273898793_Effect_of_jig_constraint_position_and_pitch_on_welding_deformation?ev=prf_pub

or

http://www.sciencedirect.com/science/article/pii/S0924013615000655

Article Effect of jig constraint position and pitch on welding deformation

Thank you for your suggested paper and valuable information.

Actually, I understood your suggestion and the whole procedure to find a new values of tensile and yield stress by its relations with E modulus (in higher temp.). But, my problem was related to how I can obtain a new stress in a higher temp. when I have the modulus in this temp.? (How the relation can help me!) I mean, for ex.

E Modulus (GPa) .......Temperature (C)

       110                               20

       101                              100

While as you know, we are gonna define the plastic strain like below;

                                    Stress  Plastic Temp.

Yield Stress (MPa)         276       0.0       20

Tensile Stress (MPa)      345       0.1       20

                                        ?          ?        100

                                        ?          ?        100 

Anyway, I found the related diagram (check the attachment please) for the yield and tensile stresses in various temperatures. That's great (agree?). [1] but I still have a doubt to what I should consider for plastic strain in different temperatures?!

Yield Stress (MPa)         276     0.0    20

Tensile Stress (MPa)      345     0.1    20

        * New yield            180      ?     400

         *New tensile          200      ?     400 

Also, you know, these variations will have a influence on the deformation more than thermal variations. [2] Do you found any new thing which help to thermal problem?

I'm really thankful regards your following up in these days.

Regards,

Mehrshad

That's a great chart! Nice to see such a detailed figure. Generally, it is difficult to model the stress-strain curve use two or more data lines. Then let's face the case there is no further data for stress-strain at high temperature. In my expirence, it is better to use 0.05 plastic strain for tensile stress, as you see for a typical stress-strain relation, the strain hardening rate decreases when it reaches to the tensile strength. So using 0.1 or even larger value tends to underestimate the elastic strain, thus the obtained deformation may be smaller. That means, I will use the following format:

        * New yield             180            0    400

         *New tensile          200      0.05     400 

Regarding the thermal problem, I worry a bit about the laser spot size. One project I work recently is laser bending of aluminum. We have the similar conditions (plate thickness, power, speed), but the spot size is with radius 5mm as a circle, the maximum temperature is around 500C.

Another thing is that you heat model is not moving continuously, so the heat concentrates at the overlapped region. But this is not the major reason.

Thank you. Great.

I'm sure about the size of laser spot. The problem should be in another thing. I should try to find it. 

Also, it's the only way that I learned by someone to simulate this procedure! If you find any new thing, I'll be glad to know your suggestion.

Dear Mehrshad,

Thank you for your reply. I use user-subroutine DFLUX to move the heat source, in this case, the mesh can be made simpler. Attached is a sample for the fortran subroutine. It can be modified to fit your problem. But a corresponding fortran complier need to be installed with abaqus.

Regards,

Huang

@Hui Huang: That't great and I hope, it works on my simulation. I've never applied subroutine in my modeling. So, I think, I should learn to install the required compiler and apply the file. 

Thank you again

The all applicable units are available in the attachment's picture.