A new approach to field modelling


RSS Twitter YouTube Blogger Facebook Linkedin

Main > Support

Magnetic problems

Can magnetic fluxes from different coils be summed in a core?
Sure, please see transformer example. There are 3 coils there.

I am trying to find a program that will help me figure out how many windings are needed per coil on a 6 coil steel stator ring powered by 3-phase AC power to get a certain magnetic field strength, about 1.3T. Is there a way that QuickField could help me find out how many windings are needed per coil?
1. Set some current in the winding.
2. Solve the problem and adjust the current to get proper B value.
After you know the total current of the winding you can find the conductor size/number of turns required.

Is it possible to design some types of magnets using QuickField such as Maxwell coil, Helmholtz coil, Quadrupole, Tapered solenoids?
QuickField is suitable to simulate tapered solenoid, Helmholtz coils, Maxwell coils and complex solenoid systems. QuickField is suitable to simulate Quadrupole magnet.
Electromagnetic coils simulations with QuickField

I 'move' a 5mm metal piece. For each metal position I 'solve the problem'. Each set of results requires me to 'select local values' and record the voltage value on each input wire into an excel spreadsheet manually. Obviously this becomes labor intensive and I feel there must be a better way of doing this. Can this be automated?
You should try LabelMover.
QuickField.com > Product > Components > Parametric LabelMover

Please let me know about the possibility of analysis at 10MHz frequency by QuickField.
Depends on the device size. We don't take into account the displacement currents. The error may be neglected if the characteristics wavelength is much greater than the object size. Wavelength for 10 MHz is about 3e8 /1e7 = 30 meters. So, if it is smaller than 30 meters - why not?

I am wondering whether it is possible to purchase QuickField material libraries for the B-H characteristic of certain materials?
We do not produce materials and do not have databases for sale. All libraries included in freeware Student QuickField package are based on open sources and are for evaluation purposes only. On our website you may find some alternative libraries, provided by our users or 3-d parties - on their own terms. We recommend all our customers to use their own data or contact their material vendor for specific material properties.
QuickField.com > Download > Data libraries

Does QuickField handle 3D models? I am interested in modeling a simple dipole magnet and studying the field shape/dependence on the pole face shape
Magnetic packages of QuickField are currently for 2D problems only, but if the dipole magnet could be presented by axisymmetric model - it is also 2D.

Can QuickField specify arbitrary B-H data in a magnet? That is, can the data of the 1st quadrant and the 2nd quadrant be defined together as the hysteresis characteristic?
QuickField cannot take into account the hysteresis effect. But the BH curve with single branch may be used. You can define the horizontal shift of this curve by Hc value.

I would like to edit the second quadrant B-H curve of a permanent magnet. The problem arises when trying to enter negative values of coercive field H, it prompts a message "Value must be positive" Could you tell me how to edit negative values?
The position of BH-curve along H-axis is specified by coercive force. Please enter the value of coercive force first and then check the curve.

It produces wrong result sometimes. For example, a symmetric model does not yield symmetric result.
Finite element analysis is very sensitive to the quality of finite element mesh. If the mesh is bad, i.e. too coarse, or not smooth, or not enough dense in the places there the field varies - the results could be wrong, or the solution process may not converge at all.

Are there any means of getting the sense of the field? I can see absolute magnitude but would like to see + & - flux density and direction of flux.
To see the direction of flux and its magnitude switch on vector plot.

With the student version, I keep getting the message "none labeled block" when I try to solve the problem. What does this message mean? The model has no blocks.
QuickField calculates the field within some areas (blocks). The blocks are created automatically when you have area surrounded by connected edges. In problem there must be at least one labeled meshed block.

Should the real magnets normal curve or intrinsic curve be used when specifying the demagnetization curve in QuickField?
With QuickField you can only use normal curve (B vs. H). There is no way to define intrinsic curve (M vs. H).

QuickField calculates several magnetic forces. What is mine?
Magnetic field produces forces acting on the current carrying conductors and on the ferromagnetic bodies. The force acting on conductors is known as Lorentz force, while the Maxwell force incorporates both components. If the second (ferromagnetic) component is negligible or is not considered, we recommend calculating the electromagnetic force as Lorentz force. Its precision is less sensitive to the contour path, and you can simply select conductors via block selection to calculate the force. With Maxwell force, this method leads to very rough results, and you are recommended to avoid coinciding of your contour parts and material boundaries. Thus, yes, you should consider Maxwell force acting on ferromagnetic body as a total force.

Is it OK if I assume that the positive current is entering in the screen and the negative current is going out from the screen?
The negative current is going into the screen. You can check positive direction if stich on vector plots.

How do I determine the transformer parameters using QuickField, more specifically - the magnetizing reactance, do you know of any procedure that is used to accurately determine this?
The general rule is to perform the same experiments as you would do with real transformer. I believe you should perform open- and short- circuit tests.

I want to calculate the mechanical force on an iron sphere within a magnetic field. The results I get for the magnetic force are very different depending on the contour. Why is the force not calculated in two directions?
You should build the contour around the body (at some distance from body's boundary)
Radial force Fr is always zero in cylindrical problems.

What's the difference between conductor's connection in Series and in Parallel?
If you have several blocks with the same label you may specify that these blocks are connected in parallel. Thus you define the total current in all blocks. Each block will carry the part of the specified current. If you specify that blocks are connected in series then each block will carry the same specified current.

What is the difference in magnetic flux and flux linkage per one turn?
Magnetic flux is calculated as F = Integral B dS. If you select the block (or closed contour), then the S is NOT the cross-section of the block. It is the 3D surface of this block. For any closed contour, if there are no field sources inside the contour F = 0. You can also calculate flux for open contours (the flux through the surface). Flux linkage per one turn of the coil is calculated as W = 1/S *2PI*Integral rA dS, where S is the cross-section of the block. You cannot calculate this integral for open contour.

I am trying to model a cylinder within an induction heating coil however I keep receiving the R-coordinate must be greater than zero error.
In QuickField the axis of rotation is always horizontal. You should define only upper half of the model.

I noticed that LOCAL VALUES function could provide the values for a specific position. But I am wondering if it is possible to get a whole set of field data.
QuickField calculates the field data in nodes and then interpolates values inside triangles. It is possible to export the whole set of results on each finite element node - but usually this table is too large for any practical use. Much more efficient possibility - exporting the field results on the rectangular grid with user defined position and grid size. Both ways of field export may be performed in QuickField postprocessor using File > Export Field command.

How should I calculate inductance in case of two conductors?
There is special tool - Impedance wizard, but its use is not always intuitive. Instead, you may calculate the impedance manually. If the conductors are the sides of the same coil you should specify +i and -i in conductors. Then solve the problem and measure the flux. The inductance is derived as L = Flux / Current.

How do you model a coil with 1000 turns?
The coil consists of many individual turns made of thin wire. So there are no eddy currents in the coil and the current density is the same in each point of the coil. Then you can replace the coil with the only one block with g=0 and specify total current of the winding.
QuickField.com > Support > Glossary > Winding modeling

The characteristic of magnet is specified as (B,H) data: (0,-1300000), (1.15, -750000). Then, the computed value of mu is 1.66389. Was this value computed by what formula?
The formula is dB = mu * mu0 * dH, where mu0 = 4*PI*1e-7, dB= B2-B1, dH = H2 - H1, and mu is relative permeability. Mu = dB/dH / mu0

"not enough boundary conditions".
QuickField magnetic problem is solved in terms of vector magnetic potential. So you need to have some reference point with known potential. Usually the external boundary is considered to have zero potential.
(The simple analogy is calculating the temperature of the heater. To know the heater power is not enough. To calculate the heater temperature you should also specify ambient temperature).

I'm modelling motor with QuickField. The problem is that the rotor has a low flux density of 0.6 T compared to the stator of 1.6T. They are both of the same steel.
The flux density can be calculated as B = Flux / S, where S is a cross-section. The stator cross-section S is 2-3 times smaller than that of the rotor.

I would like to have the results of multiple problems plotted on a single graph, and I have not found an easy way of doing it. Is there any way to have multiple QuickField problems graphed on the same plot?
There is no way to plot graphs from several problems on the same plot within QuickField. But you may copy tables in some specialized program for data processing, e.g. Microsoft Excel and make the plots there.. You can copy the single row or the whole time table.

I need to know if QuickField investigates the penetration of radiation in the device I want to analyze and to what extent your demo version is limited, respectively what kind of problems it is able to solve.
QuickField uses linear finite elements. To approximate the fading field you need at least 3 layers of elements. So it may happen that the Student edition of QuickField (limited by 255 nodes) does not allow you to generate the finite element mesh dense enough for this kind of physical effect for the whole device.
QuickField.com > Support > Glossary > Skin effect

I select the block as a contour. Then I go to XY plot and I do not understand this graph, and the values of the table. What points are taken?
The points for the table and XY plot are taken along the block boundary. For the table you can specify how many points you need to have.