Реферат: Решение обратной задачи вихретокового контроля

begin

B[i]:=B[i]+B[k1];

for k:=1 to mm1 do A[i,k]:=A[i,k]+A[k1,k];

end;

for i:=mp2 to m21 do

begin

Sx[i]:=B[i-mp1];

Nb[i-mp1]:=i;

end;

for i:=1 to mp1 do Sx[i]:=0;

Sx[1]:=B[k1];

Sx[mp1+k1]:=0;

Nb[k1]:=1;

103:

for i:=2 to m21 do N0[i]:=0;

104:

for i:=m21 downto 2 do

if N0[i]=0 then n11:=i;

for k:=2 to m21 do

if ((A[k1,n11]N0[k])) then n11:=k;

if A[k1,n11]<=0 then goto 105;

iq:=0;

for i:=1 to m1 do

if i<>k1 then

begin

if A[i,n11]>0 then

begin

iq:=iq+1;

if iq=1 then

begin

Sx[n11]:=B[i]/A[i,n11]; ip:=i;

end

else

begin

if Sx[n11]>B[i]/A[i,n11] then

begin

Sx[n11]:=B[i]/A[i,n11]; ip:=i;

end;

end;

end

else

if iq=0 then

begin

N0[n11]:=n11;

goto 104;

end;

end;

Sx[Nb[ip]]:=0;

Nb[ip]:=n11;

B[ip]:=B[ip]/A[ip,n11];

apn:=A[ip,n11];

for k:=2 to m21 do A[ip,k]:=A[ip,k]/apn;

for i:=1 to m1 do

if i<>ip then

begin

ain:=A[i,n11];

B[i]:=-B[ip]*ain+B[i];

for j:=1 to m21 do A[i,j]:=-ain*A[ip,j]+A[i,j];

end;

for i:=1 to m1 do Sx[Nb[i]]:=B[i];

goto 103;

105:

for k:=1 to mCur do Sx[k+1]:=Sx[k+1]+Gr[2,k];

a1:=0;

a2:=1.;

dh:=a2-a1;

r:=0.618033;

tl:=a1+r*r*dh;

tp:=a1+r*dh;

j:=1;

108:

if j=1 then tt:=tl else tt:=tp;

106:

for i:=1 to mCur do Rg[i]:=Zt[i]+tt*(Sx[i+1]-Zt[i]);

getFunctional( 0 );

cv:=abs(Fh[1,1]);

if nFreqs>1 then

for k:=2 to nFreqs do

begin

cv1:=abs(Fh[1,k]);

if cv

end;

if (j=1) or (j=3) then cl:=cv

else cp:=cv;

if j=1 then

begin

j:=2;

goto 108;

end;

if dh

if cl>cp then

begin

a1:=tl; dh:=a2-a1; tl:=tp; tp:=a1+r*dh ; tt:=tp; cl:=cp; j:=4;

end

else

begin

a2:=tp; dh:=tp-a1; tp:=tl; tl:=a1+r*r*dh; tt:=tl; cp:=cl; j:=3;

end;

goto 106;

107:

if (iterI < iterImax)AND(NOT saveResults( nStab,iterI )) then goto 102;

end;

End.

Приложение 2 - Удельная электрическая проводимость материалов

Приведем сводку справочных данных согласно[7-9].

Приложение 4 - Abstract

The inverse eddy current problem can be described as the task of reconstructing an unknown distribution of electrical conductivity from eddy-current probe voltage measurements recorded as function of excitation frequency. Conductivity variation may be a result of surface processing with substances like hydrogen and carbon or surface heating.

Mathematical reasons and supporting software for inverse conductivity profiling were developed by us. Inverse problem was solved for layered plane and cylindrical conductors.

Because the inverse problem is nonlinear, we propose using an iterative algorithm which can be formalized as the minimization of an error functional related to the difference between the probe voltages theoretically predicted by the direct problem solving and the measured probe voltages.

Numerical results were obtained for some models of conductivity distribution. It was shown that inverse problem can be solved exactly in case of correct measurements. Good estimation of the true conductivity distribution takes place also for measurement noise about 2 percents but in case of 5 percent error results are worse.

57