How to get closer to experimental values by simulation ?
Hello, we use Altair Flux 2020 in order to simulate our BLDC motor. Recently, i wanted to compare the experimental values with the simulation. I got really different values. I join the graph of Torque/Speed and Torque/Current. The curve of Torque/speed is not to far of the exmperimental values and I may already have an idea of the solution. But for the other curve It's to far of the reality. Do you have any advice to resetting the model ? Do you need further information to answer me ?
Thanks in advance,
Théo Briquet
Answers
-
Hello Theo;
The outputs from Flux are linked to the quality of the inputs, so please check that you are using the right:
- Material characteristics B(H) curves and magnet orientation.
- Winding schem and number of turns +number of parallel paths + resistance
-The right supply
-The right control angle
-The corresponding rotor Speed
-And the measurements conditions - Temperature (to adapt the Br of the magnets and coil resistance).
- The machine is it skewed or not.
I hope this will help you
Best regards
2 -
Farid zidat_20516 said:
Hello Theo;
The outputs from Flux are linked to the quality of the inputs, so please check that you are using the right:
- Material characteristics B(H) curves and magnet orientation.
- Winding schem and number of turns +number of parallel paths + resistance
-The right supply
-The right control angle
-The corresponding rotor Speed
-And the measurements conditions - Temperature (to adapt the Br of the magnets and coil resistance).
- The machine is it skewed or not.
I hope this will help you
Best regards
Thanks for yours answers,
I have checked all these inputs and I have further questions,
I saw on the Flux documentation that the electric disposition of the supply of coils is specific for a three phase in star geometry, I understood that I need to delete one of the power supplies ?
What do you mean by “the right control angle”, is it the initial position of the rotor ?
Otherwise I think that the inputs are correct.
Best regards,
0 -
Theo Briquet said:
Thanks for yours answers,
I have checked all these inputs and I have further questions,
I saw on the Flux documentation that the electric disposition of the supply of coils is specific for a three phase in star geometry, I understood that I need to delete one of the power supplies ?
What do you mean by “the right control angle”, is it the initial position of the rotor ?
Otherwise I think that the inputs are correct.
Best regards,
You are welcome !
for the electric circuit, you can use only 2 current sources only if the you have 3 phase balanced system (so at the neutral node, the I1+I2+I3 =0) otherwise, you have to use 3 current sources + 4th branche with only high resistance.
the control angle is the angle between the current supply and the back emf, that means you control the position of the rotor (rotor magnetic field) to stator magnetic field in a way to maximize the torque.
Can you check using 3D computation if you have the same results like with Flux 2D? just to be sure that the border effect are do not have a big impact on your machine performances.
two tests should give results close to measurements (between 5% to 10% difference)
1- The back EMF
2- Supply with measured current, and impose the measured speed.
If you don't have coherent results, that means that there is something wrong in your model.
Best regards
0 -
Farid zidat_20516 said:
You are welcome !
for the electric circuit, you can use only 2 current sources only if the you have 3 phase balanced system (so at the neutral node, the I1+I2+I3 =0) otherwise, you have to use 3 current sources + 4th branche with only high resistance.
the control angle is the angle between the current supply and the back emf, that means you control the position of the rotor (rotor magnetic field) to stator magnetic field in a way to maximize the torque.
Can you check using 3D computation if you have the same results like with Flux 2D? just to be sure that the border effect are do not have a big impact on your machine performances.
two tests should give results close to measurements (between 5% to 10% difference)
1- The back EMF
2- Supply with measured current, and impose the measured speed.
If you don't have coherent results, that means that there is something wrong in your model.
Best regards
Bonjour Mr Zidat,
J'ai appris que vous étiez français, donc j'en profites et cela facilitera surement la conversation
Je vais essayer de faire la simulation 3D dans les prochains jours.
Avant ça j'ai essayé de changer le circuit électrique en suivant vos indications mais les résultats ne sont pas ceux que j'espérais. J'ai remarqué que vous parliez de source de "courant" alors que mon circuit est alimenté en source de tension (trapézoïdale 0V-12V) . Est-ce que cela fait-il une différence ? Comment pourrais-je passer à une source de courant du même type ?
Cordialement,
0 -
Bonjour M. Briquet;
Oui en effet, mais pour faire profiter un grand nombre d'utilisateurs sur le forum, nous allons continuer en Anglais.
So for the current sources, you can directly use them from the circuit editor, as you did with voltage sources.
If you use voltage sources, so you can use 3 sources in star connection. With Voltage sources, again you have to well take account about your control angle, and the impedance also the number of turns, the end winding inductance, the phase resistance, the material characteristics.
P.S: You have access to the french support team if you want. So you can continue to use the forum or use the local french support.
1 -
Farid zidat_20516 said:
Bonjour M. Briquet;
Oui en effet, mais pour faire profiter un grand nombre d'utilisateurs sur le forum, nous allons continuer en Anglais.
So for the current sources, you can directly use them from the circuit editor, as you did with voltage sources.
If you use voltage sources, so you can use 3 sources in star connection. With Voltage sources, again you have to well take account about your control angle, and the impedance also the number of turns, the end winding inductance, the phase resistance, the material characteristics.
P.S: You have access to the french support team if you want. So you can continue to use the forum or use the local french support.
Hello Mr Zidat,
Oui je comprend, pas de problème.
I verify severals times all the project and I think I haven't forgotten anything. Maybe I reach the limits of the software and Flux 2020 can't give better results...
We install "FluxMotor 2020", there isn't our geometry but we may have better results.
I don't find the local french support, do I need to "create a case"?
Best regards,
0 -
Bonjour M. Briquet;
If it is an electromagnetic device, i could not see any limitation to Flux to achieve the desired performances with good accuracy.
If you have made the test in 3D, and you still have a big gap compared to the measurements, we should have a look on your project.
Please create the request as shown bellow, instead of using Ask a question, use create a case.
If i remeber well, your motor can not in FluxMotor, you can only make 3 slots + 2 poles or 3 slots + 4 poles. In addition to that, in FluxMotor, the supply is supposed to be sine wave, so only the fundamental is taken intoaccunt. Anyway, you can try those configuration and see what is the behavior of this machine.
Best regards
1
Altair Employee