Magnetic force field in a conformable postprocessor

Several months ago, a potential customer asked me about using Magnum to calculate forces on small iron particles in a magnetic separator. The force on unsaturated particles in a fluid with relative magnetic permeability μr = 1 is given in SI units by the expression:

Fm = (1/μo) *(μr – 1)*Vp*B*gradB.

Here μo is the magnetic permeability of vacuum, μr is the relative magnetic permeability of the iron, Vp is the volume of the particle, B is the local magnetic flux density and gradB is its gradient. In its present form, MagView can certainly determine a numerical value for the force on a single particle at a given position. Instead, the customer wanted spatial maps of the force field: that is, plots of B*gradB.

Adding support for a special plot quantity in MagView would have involved a great deal of work (especially considering that the gradient of the magnetic flux density involves second derivatives of the Magnum potential functions). The customer’s application was in a gray area: between a unique application and something that should be a standard feature of the code. After I explained that a significant development cost would be involved, the customer continued his software search. Perhaps he found the perfect program.

In my heart I envision our programs as all things to all people, so a memory lingers in my mind when one falls short. From experience, I know that, given time, a solution usually turns up, hopefully a simple and elegant one. In this case, the solution is FE Builder (http://fieldp.com/myblog/2009/rpn/), a new software package that we will release October 1 . It consists of Geometer/Metamesh and the fully-conformable postprocessor FEVision. FE Builder is aimed at advanced developers of finite-element programs for any application. The package covers all input/output functions, allowing the developer to concentrate on the physical solution. Although the target audience is quite small, I felt the effort involved in creating FEVision would be justified by spin-offs.

FEVision resolves the B*gradB problem. More important, it provides a path to address any specialized user analysis without the need to cobble features on our core programs. FEVision will be included with all packages in the release of the AMaze 3.0 programs on February 1, 2010. Here how it works for the B*gradB problem:

1) Create a standard Magnum solution for the flux density B in a magnetic separator.

2) Set MagView to export the quantity |B|. Generate a matrix file covering the region of interest for the force map. (Note that the text format of MagView 3.0 matrix files has been modified so that they can serve as input files to FEVision.)

3) Define the desired calculated quantities in an FEVision configuration file (e.g., bgradb.cfg). Noting that |B| is stored quantity number one, the definitions would look like this:

INTERPOLATION
BGradB = &Q[1] &gradmag[1] *
END
VECTOR
Gradient = &gradx[1];&grady[1];&gradz[1]
END

4) Run FEVision and load the configuration file. (The setting is preserved in the registry, so this step is performed only once).

5. Load the matrix file and go. The program provides the following capabilities:

  • Create plots of B*gradB in any of the 2D or 3D styles (contour lines, filled contours,…).
  • Do quantitative point calculations or interactive scans.
  • Make arrow plots of gradB or plot trajectories of particles in the force field.

The key result is that the program users can find any specialized information they want with a few minutes of preparation. Most important, the vastly-expanded range of capabilities is available at no cost beyond the standard program price.

You must be logged in to post a comment.