[Getdp] InterpolationLinear function

janvrot at infolink.com.br janvrot at infolink.com.br
Fri Sep 27 18:57:37 CEST 2002


Hello,

I'm using the InterpolationLinear function in my problem to model a
steel. To analyze the workings, I put it on the GetDP demo from the
distribution (test.*, using scalar potential).

In test.pro I made the following changes:

/* Start Non Linear Testing ------------------------------------*/
  //mu [ Core ]  = murCore * mu0 ;
/* -----------------------
1020 steel BH curve
B (Tesla) x H (A/m)
0.0000	0.0000
0.2000	159.1550
0.5200	238.7325
1.0000	557.0425
1.4200	1591.5500
1.6000	3183.1000
1.8036	7077.8066
2.1378	29076.3930
2.2800	79577.5000
--------------------*/
mu_steel_couples = {
    0.000, 0.0012566,
  159.155, 0.0012566,
  238.733, 0.0021782,
  557.043, 0.0017952,
 1591.550, 0.0008922,
 3183.100, 0.0005026,
 7077.800, 0.0002549,
29076.400, 0.0000735304,
79577.500, 0.0000286513
};

  fmucore[]=InterpolationLinear[$1]{List[mu_steel_couples]};
  mu [ Core ] = fmucore[Norm[$1]];

/* End Non Linear Testing ------------------------------------*/

As can be seen, I took for each couple the B value (Tesla) and B divided
by H (A/m). For the zero entry I repeated the second value. Is that ok?

Then, as we now have a nonlinear material, I used:

NL_NbrMaxIter = 60;
NL_Eps = 1.e-4;
NL_Relax = 1;

Resolution {
  { Name MagSta_phi ;
    System {
      { Name A ; NameOfFormulation MagSta_phi ; }
    }
    //Operation {
    //  Generate[A] ; Solve[A] ; SaveSolution[A] ;
    //}
    Operation {
      IterativeLoop[NL_NbrMaxIter, NL_Eps, NL_Relax] {
        GenerateJac[A]; SolveJac[A];
      }
      SaveSolution[A];
    }
  }
}

Convergence was immediate, but in my 3D problem convergence takes 30-40
iterations. Is that ok? 
Info      : Mean Error = 4.652e-11 after 2 Iterations

For point{-0.060, 0.020, 0.} in the core the resolution gives H = 805
A/m and B = 1.01 Tesla, were the B value for this H in the BxH curve
would be around 1.10 Tesla.

For point{0.060, 0.020, 0.} in the core the resolution gives H = 531 A/m
and B = 0.66 Tesla, were the B value for this H in the BxH curve would
be around 0.97 Tesla.

Should these results be considered consistent with the given function? I
noticed that increasing hc for the magnet things get worse. For an
adequate set of couples including high H pairs B goes far beyond the
saturation limit (doubles it).

Thanks in advance for any help.
-- 
Janvrot IVM
janvrot at infolink.com.br