QUESTION: We are inputting the datapoints for B-H curve for Emag Analysis for Magnetisation curve (with H ranging from 0 to 150) and Demagnetisation curves (H coming down from 150 to 0). The typical values of B-H are given below for both the curves, wherein there are duplicate points as there are identical H values in both of these curves. So, when the below macro is run, the new values of B in the Demagnetisation curve are overwritten with the B values in the Magnetisation curve. Is there any way to avoid overwriting and to consider both the curves?
! Magnetisation curve
TB,BH,1 ! B-H CURVE FOR MATERIAL 1
TBPT,,0,0
TBPT,,20,.1 ! H AND B RESPECTIVELY
TBPT,,25,.125
TBPT,,50,.250
TBPT,,75,.35
TBPT,,100,.375
TBPT,,150,.425
!Demagnetisation curve
TBPT,,100,.4
TBPT,,75,.375
TBPT,,50,.35
TBPT,,25,.275
TBPT,,0,.15TBPLOT,BH,1
ANSWER:
ANSYS does not account for hysteresis in electromagnetic analyses. There is no way I know of to define two sets of BH data (one for magnetization, the other for demagnetization) for a given material.
You might be able to assign unique material property numbers to each individual iron element and monitor the field in a series of restarts performed in a do loop. When the magnitude of the field begins to decrease, switch to a demagnetization curve. Note that you will need a unique demagnetization curve for each element, because each element in your model will reach a unique level of saturation.
An easier way would be to use manufacturer's data for coreloss. Below is a response for a similar inquiry:
There are 2 components of coreloss that I'm aware of:
1) joule heating (losses due to induced eddy current)
2) hysteresis
We can calculate #1 directly, by including the VOLT degree of freedom
and defining electrical resistivity (RSVX) in eitherPLANE13/53 (2D) or
SOLID97/117 (3D) elements. We cannot directly calculate hysteresis loss
- the power associated with the area enclo
|
QUESTION: We are inputting the datapoints for B-H curve for Emag Analysis for Magnetisation curve (with H ranging from 0 to 150) and Demagnetisation curves (H coming down from 150 to 0). The typical values of B-H are given below for both the curves, wherein there are duplicate points as there are identical H values in both of these curves. So, when the below macro is run, the new values of B in the Demagnetisation curve are overwritten with the B values in the Magnetisation curve. Is there any way to avoid overwriting and to consider both the curves? ! Magnetisation curve TB,BH,1 ! B-H CURVE FOR MATERIAL 1 TBPT,,0,0 TBPT,,20,.1 ! H AND B RESPECTIVELY TBPT,,25,.125 TBPT,,50,.250 TBPT,,75,.35 TBPT,,100,.375 TBPT,,150,.425 !Demagnetisation curve TBPT,,100,.4 TBPT,,75,.375 TBPT,,50,.35 TBPT,,25,.275 TBPT,,0,.15TBPLOT,BH,1 ANSWER: ANSYS does not account for hysteresis in electromagnetic analyses. There is no way I know of to define two sets of BH data (one for magnetization, the other for demagnetization) for a given material. You might be able to assign unique material property numbers to each individual iron element and monitor the field in a series of restarts performed in a do loop. When the magnitude of the field begins to decrease, switch to a demagnetization curve. Note that you will need a unique demagnetization curve for each element, because each element in your model will reach a unique level of saturation. An easier way would be to use manufacturer`s data for coreloss. Below is a response for a similar inquiry: There are 2 components of coreloss that I'm aware of: 1) joule heating (losses due to induced eddy current) 2) hysteresis We can calculate #1 directly, by including the VOLT degree of freedom and defining electrical resistivity (RSVX) in eitherPLANE13/53 (2D) or SOLID97/117 (3D) elements. We cannot directly calculate hysteresis loss - the power associated with the area enclosed by a BH loop in the BH plane. Depending on how much detail one wishes to incorporate into evaluation of hysteresisloss, it should be possible (if somewhat inefficient) to use APDL in a transient simulation to calculate it. An alternative approach is given by Tony Patti and Mike Yaksh in "Duo Modeling Method to Implement Coreloss Impedance Calculations" - a paper in the 2000 ANSYS Conference Proceedings. It involves the use of a finite element model, a reduced order circuit model, and coreloss data supplied by the steel manufacturer. But it appears to be more practical, since I've never seen (nor do I know if it's ever possible to obtain) the material data one would need to implement the first approach in a transient. |
||
|