The following error occurred converting from sym to double: Unable to convert expression containing symbolic variables into double array Apply' subs' function first to substitute values for variables.

Hello,

When I try the loop for ii (see attached mix file) I get this error:

The following error occurred converting from sym to double: Unable to convert expression containing symbolic variables into double array Apply'subs' function first to substitute values ​​for variables.

It has solution because I tried writing each ii and the respective equations. But I need the loop (obviously) because I want to introduce several values ​​of ii and unknows to the system.

PD: Not so important but, as you can see, inside of the equation eqll, eqll. eqlll ... are the unknowns CO, C1 .... DO, D1, ... which are inside of a series (but I) couldnt find a way, so I wrote the elements of the serie) Is there any way to optimize that?

[Walter Roberson]

No attached file.

[Sila Fundora]

clear
clc

% General input
h = 8; 
l = 2; 
d = 0.09; 
sections = 5;

% Data from previous calculations
z = [0,-1.6,-3.2,-4.8,-6.4,-8];
P = [-0.923072381390763 + 0.083645843312860i,-0.763379411108664 + 0.047842477328015i,-0.647844785539475 + 0.001886672095714i,-0.569785303973405 - 0.037689986193477i,-0.524685547937092 - 0.062861417563918i,-0.509936682892244 - 0.071256842113728i];
Q = [0.000000000000000 + 0.100765147467902i,0.000000000000000 + 0.093981142166511i,0.000000000000000 + 0.095329989814646i,0.000000000000000 + 0.100156105292950i,0.000000000000000 + 0.099088979609004i,0.000000000000000 + 0.096724494080166i];

points = sections + 1;

eqI = zeros (1,points);
eqII = zeros (1,points);
eqIII = zeros (1,points);

syms C0 C1 C2 C3 C4 C5 C6 C7 C8 D0 D1 D2 D3 D4 D5 D6 D7 D8 

for ii = 1:points
    
eqI(ii) = (1/h)*C0 + (2/h)*(C1*cos(1*pi*(z(1,ii)+h)/h)+ ...
    C2*cos(2*pi*(z(1,ii)+h)/h)+ ...
    C3*cos(3*pi*(z(1,ii)+h)/h)+ ...
    C4*cos(4*pi*(z(1,ii)+h)/h)+ ...
    C5*cos(5*pi*(z(1,ii)+h)/h)+ ...
    C6*cos(6*pi*(z(1,ii)+h)/h)+ ...
    C7*cos(7*pi*(z(1,ii)+h)/h)+ ...
    C8*cos(8*pi*(z(1,ii)+h)/h))-d;

eqII(ii) = (1/h)*(C0+D0*l) + (2/h)*(((C1*cosh(1*pi*l/h) + D1*sinh(1*pi*l/h))*cos(1*pi*(z(1,ii)+h)/h)) ...
    +((C2*cosh(2*pi*l/h) + D2*sinh(2*pi*l/h))*cos(2*pi*(z(1,ii)+h)/h)) ...
    +((C3*cosh(3*pi*l/h) + D3*sinh(3*pi*l/h))*cos(3*pi*(z(1,ii)+h)/h)) ...
    +((C4*cosh(4*pi*l/h) + D4*sinh(4*pi*l/h))*cos(4*pi*(z(1,ii)+h)/h)) ...
    +((C5*cosh(5*pi*l/h) + D5*sinh(5*pi*l/h))*cos(5*pi*(z(1,ii)+h)/h)) ...
    +((C6*cosh(6*pi*l/h) + D6*sinh(6*pi*l/h))*cos(6*pi*(z(1,ii)+h)/h)) ...
    +((C7*cosh(7*pi*l/h) + D7*sinh(7*pi*l/h))*cos(7*pi*(z(1,ii)+h)/h)) ...
    +((C8*cosh(8*pi*l/h) + D8*sinh(8*pi*l/h))*cos(8*pi*(z(1,ii)+h)/h))) - P(1,ii);

eqIII(ii) = (1/h)*(D0) + (2/h)*(((C1*(1*pi/h)*sinh(1*pi*l/h) + D1*(1*pi/h)*cosh(1*pi*l/h))*cos(1*pi*(z(1,ii)+h)/h)) ...
    + ((C2*(2*pi/h)*sinh(2*pi*l/h) + D2*(2*pi/h)*cosh(2*pi*l/h))*cos(2*pi*(z(1,ii)+h)/h)) ...
    + ((C3*(3*pi/h)*sinh(3*pi*l/h) + D3*(3*pi/h)*cosh(3*pi*l/h))*cos(3*pi*(z(1,ii)+h)/h)) ...
    + ((C4*(4*pi/h)*sinh(4*pi*l/h) + D4*(4*pi/h)*cosh(4*pi*l/h))*cos(4*pi*(z(1,ii)+h)/h)) ...
    + ((C5*(5*pi/h)*sinh(5*pi*l/h) + D5*(5*pi/h)*cosh(5*pi*l/h))*cos(5*pi*(z(1,ii)+h)/h)) ...
    + ((C6*(6*pi/h)*sinh(6*pi*l/h) + D6*(6*pi/h)*cosh(6*pi*l/h))*cos(6*pi*(z(1,ii)+h)/h)) ...
    + ((C7*(7*pi/h)*sinh(7*pi*l/h) + D7*(7*pi/h)*cosh(7*pi*l/h))*cos(7*pi*(z(1,ii)+h)/h)) ...
    + ((C8*(8*pi/h)*sinh(8*pi*l/h) + D8*(8*pi/h)*cosh(8*pi*l/h))*cos(8*pi*(z(1,ii)+h)/h)))- Q(1,ii);
end


eqI1 = eqI(1,1);
eqI2 = eqI(1,2);
eqI3 = eqI(1,3);
eqI4 = eqI(1,4);
eqI5 = eqI(1,5);
eqI6 = eqI(1,6);

eqII1 = eqII(1,1);
eqII2 = eqII(1,2);
eqII3 = eqII(1,3);
eqII4 = eqII(1,4);
eqII5 = eqII(1,5);
eqII6 = eqII(1,6);

eqIII1 = eqIII(1,1);
eqIII2 = eqIII(1,2);
eqIII3 = eqIII(1,3);
eqIII4 = eqIII(1,4);
eqIII5 = eqIII(1,5);
eqIII6 = eqIII(1,6);


result = vpasolve([eqI1,eqI2,eqI3,eqI4,eqI5,eqI6,eqII1,eqII2,eqII3,eqII4,eqII5,eqII6,eqIII1,eqIII2,eqIII3,eqIII4,eqIII5,eqIII6])

disp 'Cm = [C0 C1 C2 C3 C4 C5 C6 C7 C8]'
Cminitial = [result.C0 result.C1 result.C2 result.C3 result.C4 result.C5 result.C6 result.C7 result.C8];
Cm = double (Cminitial)

disp 'Dm = [D0 D1 D2 D3 D4 D5 D6 D7 D8]'
Dminitial = [result.D0 result.D1 result.D2 result.D3 result.D4 result.D5 result.D6 result.D7 result.D8];
Dm = double (Dminitial)

Answer 1

[Walter Roberson]

eqI = zeros (1,points, 'sym');
eqII = zeros (1,points, 'sym');
eqIII = zeros (1,points, 'sym');

[Sila Fundora]

OMG!!!! haha Thank you

Any idea to optimize writing the serie...?

[Walter Roberson]

C3*(3*pi/h)*sinh(3*pi*l/h) + D3*(3*pi/h)*cosh(3*pi*l/h))*cos(3*pi*(z(1,ii)+h)/h)

suggests

vector C
vector D
N = 1:8;
Npih = N.*pi.*h
sum( C .* Npih .* sinh(Npih/l) + D .* Npih .* cosh(Niph./l) .* cos(Npih.*(z(1,ii)+h), 2)

except I would probably take it further and z(1,:) instead of z(1,ii) so that the for ii loop could be avoided.

[Sila Fundora]

Hi, I introduced your suggestion (with some changes responding to my equations). However, the final equations are diffferent and also...the solution for the equations system is empty

clear

clc

% General input

h = 8;

l = 2;

d = 0.09;

sections = 5;

% Data from previous calculations

z = [0,-1.6,-3.2,-4.8,-6.4,-8];

P = [-0.923072381390763 + 0.083645843312860i,-0.763379411108664 + 0.047842477328015i,-0.647844785539475 + 0.001886672095714i,-0.569785303973405 - 0.037689986193477i,-0.524685547937092 - 0.062861417563918i,-0.509936682892244 - 0.071256842113728i];

Q = [0.000000000000000 + 0.100765147467902i,0.000000000000000 + 0.093981142166511i,0.000000000000000 + 0.095329989814646i,0.000000000000000 + 0.100156105292950i,0.000000000000000 + 0.099088979609004i,0.000000000000000 + 0.096724494080166i];

points = sections + 1;

eqI = zeros (1,points, 'sym');

eqII = zeros (1,points, 'sym');

eqIII = zeros (1,points, 'sym');

syms C0 C1 C2 C3 C4 C5 C6 C7 C8 D0 D1 D2 D3 D4 D5 D6 D7 D8

C = [C1 C2 C3 C4 C5 C6 C7 C8];

D = [D1 D2 D3 D4 D5 D6 D7 D8];

N = 1:8;

Npih = N.*pi./h;

for ii = 1:points

eqI(ii) = (1/h)*C0 + (2/h)*sum( C .*cos(Npih.*(z(1,ii)+h)), 2)-d;

% eqI(ii) = (1/h)*C0 + (2/h)*(C1*cos(1*pi*(z(1,ii)+h)/h)+...+C8*cos(8*pi*(z(1,ii)+h)/h))-d;

eqII(ii) = (1/h)*(C0+D0*l) + (2/h)*sum( C .* Npih .* cosh(Npih.*l) + D .*sinh(Npih.*l).* cos(Npih.*(z(1,ii)+h)), 2) - P(1,ii);

% eqII(ii) = (1/h)*(C0+D0*l) + (2/h)*(((C1*cosh(1*pi*l/h) + D1*sinh(1*pi*l/h))*cos(1*pi*(z(1,ii)+h)/h))+...+((C8*cosh(8*pi*l/h) + D8*sinh(8*pi*l/h))*cos(8*pi*(z(1,ii)+h)/h))) - P(1,ii);

eqIII(ii) = (1/h)*(D0) + (2/h)*sum( C .* Npih .* sinh(Npih.*l) + D .* Npih .* cosh(Npih.*l) .* cos(Npih.*(z(1,ii)+h)), 2)- Q(1,ii);

% eqIII(ii) = (1/h)*(D0) + (2/h)*(((C1*(1*pi/h)*sinh(1*pi*l/h) + D1*(1*pi/h)*cosh(1*pi*l/h))*cos(1*pi*(z(1,ii)+h)/h))+...+ ((C8*(8*pi/h)*sinh(8*pi*l/h) + D8*(8*pi/h)*cosh(8*pi*l/h))*cos(8*pi*(z(1,ii)+h)/h)))- Q(1,ii);

end

eqI1 = eqI(1,1);

eqI2 = eqI(1,2);

eqI3 = eqI(1,3);

eqI4 = eqI(1,4);

eqI5 = eqI(1,5);

eqI6 = eqI(1,6);

eqII1 = eqII(1,1);

eqII2 = eqII(1,2);

eqII3 = eqII(1,3);

eqII4 = eqII(1,4);

eqII5 = eqII(1,5);

eqII6 = eqII(1,6);

eqIII1 = eqIII(1,1);

eqIII2 = eqIII(1,2);

eqIII3 = eqIII(1,3);

eqIII4 = eqIII(1,4);

eqIII5 = eqIII(1,5);

eqIII6 = eqIII(1,6);

result = vpasolve([eqI1,eqI2,eqI3,eqI4,eqI5,eqI6,eqII1,eqII2,eqII3,eqII4,eqII5,eqII6,eqIII1,eqIII2,eqIII3,eqIII4,eqIII5,eqIII6]);

disp 'Cm = [C0 C1 C2 C3 C4 C5 C6 C7 C8]'

Cminitial = [result.C0 result.C1 result.C2 result.C3 result.C4 result.C5 result.C6 result.C7 result.C8];

Cm = double (Cminitial)

disp 'Dm = [D0 D1 D2 D3 D4 D5 D6 D7 D8]'

Dminitial = [result.D0 result.D1 result.D2 result.D3 result.D4 result.D5 result.D6 result.D7 result.D8];

Dm = double (Dminitial)

[Walter Roberson]

As the question is not about how to write the algorithm at all, and is instead about how to write the algorithm "more nicely", then you can

eqI(ii) = (1/h)*C0 + (2/h)*sum( C .*cos(Npih.*(z(1,ii)+h)), 2)-d;
eqI_old(ii) = (1/h)*C0 + (2/h)*(C1*cos(1*pi*(z(1,ii)+h)/h)+...+C8*cos(8*pi*(z(1,ii)+h)/h))-d;  %all of the old code

and compare compare the two, such as checking whether their difference is 0. If the two are different, then start debugging.