Does flux consider coil reactance in the circuit?

Iñigo Izaguirre

Hello,

 

I was wondering if flux is considering the coil reactance in the electric circuit of a power transformer simulation. I want to calculate the short circuit voltage of a electric transformer but the simulation always give me lower values than the real measured short circuit voltage so I want to know if I have to define something else in the simulation and also if flux considers coil reactance when you define it as non meshed coil.

 

Thank you, 

Iñigo Izaguirre

SimonGuicheteau

Hello Inigo,

yes, the coil defined in the circuit context is only defined with a value of resistance but if it is linked with a coil region or with a non meshed coil, its inductance is determined by the 3D model.

If you are not seeing the results as expected, especially with non meshed coils, I would recommend checking the mesh around them. Sometimes, adding an extra box of air around these coils can allow to control the mesh in a better way to ensure the mesh is not too coarse around the coils.

Also for power transformers, the short circuit test means you will havre a lot of leakage around the core. So you need to make sure you are representing all the parts around the transformer and that the infinite box is not too close to the model otherwise you will miss some of the leakage field and get wrong results.

FYI there is an example of Power Transformer from the supervisor. You can check how the external frames are defined and the size of the infinite box as a reference. No need to solve it as it may take a few hours but just look at the physics set up.

Best regards

Iñigo Izaguirre

Hello Simon,

Thank you for your answer and help. I had already done the step of defining an air region in the non meshed coils and I think the infinite box is enough but the results are not correct yet. I am putting my model here if you want to take a look, I would be very grateful. Apart from that I am not able to open the example it does not appear any error but it is only like charging and does not finish.

The model is a 220KVA 320/400 Dyn11 three phase transformer and it should give near 15.4V of short circcuit voltage. At the moment it gives near 8.6V.

Thank you,

Iñigo

SimonGuicheteau

Hello,

The mesh seemed very light for the geometry you had. I cut the model to have only a quarter and save time. then I defined a manual mesh using small medium and large mesh points only.

I still don't get 15.6, but something around 10~11.

I suspect the coil resistance for the secondary is not exact as it was simply evaluated by Flux using the resistivity and the shape of the non meshed coil. If you have the exact value, please set it in the coil components in the cicuit and do not define a resistivity in the non meshed coil.

Best regards

Iñigo Izaguirre

Thank you for your time and help. I have checked the real resistance of the coils and didnt match with the model. Primary resistance is 0.00764 ohm and secondary resistance is 0.00597 ohm. In the model the primary resistance estimated is 0.00854 and in the secondary 0.00366. I put the additional resistance(0.00398 separated in the three caps of each coil) in the secondary coils but the short circuit voltage does not increase enough, in fact it has increase up to 11V. Do you see any solution to my problem? I will put the model attached

 

Thank you,

Iñigo

Iñigo Izaguirre

Hello 

I have made a mistake checking the resistance vallue of the coils. The values of the first simulation were okay (rp=0.00854, rs=0.00366). So the problem is not in the resistance value. Do you know where else could be?

SimonGuicheteau

I I were you, I would double check all the inputs including materials, number of turns, coil orientation and assignments.

Then try to look at various results to understand what is different than expected (currents, voltages, flux in coils...

Usually in transformers, if you supply one side only, the impedance of the whole device will affect the performance. If you know both currents, then try supplying both sides as this would force the correct currents whatever the impedance is. However the voltage needs to have the correct resistance

 

I noticed we have a 3 current source supply, we need to add an extra big resistance in parallel of the sources to avoid numerical issues. I fixed it but it dod not change the results so this is not the problem here.

Iñigo Izaguirre

Hello,

I have checked all the inputs and material properties but I haven't found any solution to my problem.

 

I have another question about the simulation you gave me the other day. I would like to know how you improve the mmesh and what steps I have to follow to make the same with geommetry (symmetry and those things)

 

Thank you,

Iñigo 

SimonGuicheteau

As my colleague Farid told you in another post, please try to create a separate thread for separate questions.

 

To give a short question about the mesh, a simple method we teach in trainings is to do the following steps :

- deactivate aided mesh

-assign mesh points to differente regions using the command Mesh> Assign mesh informatin > "assign mesh point to point"

             -assign MEDIUM to all points

             -assign LARGE to all points belonging to the infinite box

             -assign SMALL to airgaps or smaller details

             -Adjust mesh points value by playing the "mesh lines" command which goes fast, when ready, mesh the whole domain

             -add some relaxations on faces or volumes if necessary

 

Best regards

Iñigo Izaguirre

Thank you for your help. I dont understand very well how have you done to make the simulation in less time and a quicker mesh. I understand is a question related but if you want I can create a separated question.

 

Thanks 

SimonGuicheteau

Hello Inigo,

The model I sent you is only a quarter model using 2 symmetries, so I divide the model size by 4.

Then the mesh is assigned manually with meshpoints to dimensions that match the level of details so the mesher is able to mesh it faster. With bigger elements, it needs to iterate a few times to manage to fit in very large elements without having intersections all over.

SO the model is better leshed and solves faster as well

Iñigo Izaguirre

Thank you for your help. At the moment I could not fix the problem yet I have checked all the inputs but I did not find the problem. I dont know if you have any new idea of what could cause this inaccurate results.

 

Thank you,

Iñigo Izaguirre