how to use fsolve with more equation
4 views (last 30 days)
Show older comments
function F = fullfun(S_c,S_d,Ru,T,mu_s0,mu_w0,F,lambda_0,alpha,C_c,C_d,V_cp,r_am,r_cm,b_d,sigma_d,sigma_c,N_cp,h_r,sigma)
F = [T_s+(4*10^(-6)*S_d^(2))-(4*10^(-5)*S_d)-0.96;
T_w+(4*10^(-5)*S_c^(2))+(1.9*10^(-2)*S_c)-11.2;
L_s-(min(2*10^(-12)*S_d^(2)-3*10^(-10)*S_d+...
6*10^(-8),2*10^(-12)*S_c^(2)-3*10^(-10)*S_c+6*10^(-8)));
L_w-(5*10^(-4)*S_c^(-0.416));
mu_sc-(mu_s0+Ru*T*log(S_c/1000));
mu_sd-(mu_s0+Ru*T*log(S_d/1000));
mu_wc-(mu_w0+Ru*T*log((1000-S_c)/1000));
mu_wd-(mu_w0+Ru*T*log((1000-S_d)/1000));
E_m-(T_s/F*(mu_sc-mu_sd)+T_w/F*(mu_wc-mu_wd));
E_am-E_m;
E_cm-E_m;
lambda_c-lambda_0+alpha*sqrt(C_c);
lambda_d-lambda_0+alpha*sqrt(C_d);
lambda_r-((lambda_c + lambda_d)/2);
C_r-((C_c+C_d)/2);
k-(lambda_r * C_r);
I-((V_cp-E_am-E_cm)/(r_am+r_cm+b_d/(sigma_d*lambda_d*C_d)+(b_c/(sigma_c*lambda_c*C_c))+...
r_cm/N_cp+2*h_r/(sigma*k*N_cp)));
J_s-(T_s*I/F - L_s*(C_c-C_d));
J_w-(T_w*I/F + L_w*Ru*T*(C_c-C_d)*10^(-5)*2);
0 Comments
Answers (2)
Walter Roberson
on 21 Sep 2019
Bundle all of the variables into a single vector for the purposes of fsolve. You can unbundle them inside the function.
vars20init = vector of 20 initial values
sol = fsolve(@fullfun, vars20init)
function result = fullfun(vars20)
tcell = num2cell(vars20);
[S_c,S_d,Ru,T,mu_s0,mu_w0,F,lambda_0,alpha,C_c,C_d,V_cp,r_am,r_cm,b_d,sigma_d,sigma_c,N_cp,h_r,sigma] = tcell{:};
result = [T_s+(4*10^(-6)*S_d^(2))-(4*10^(-5)*S_d)-0.96;
T_w+(4*10^(-5)*S_c^(2))+(1.9*10^(-2)*S_c)-11.2;
L_s-(min(2*10^(-12)*S_d^(2)-3*10^(-10)*S_d+...
6*10^(-8),2*10^(-12)*S_c^(2)-3*10^(-10)*S_c+6*10^(-8)));
L_w-(5*10^(-4)*S_c^(-0.416));
mu_sc-(mu_s0+Ru*T*log(S_c/1000));
mu_sd-(mu_s0+Ru*T*log(S_d/1000));
mu_wc-(mu_w0+Ru*T*log((1000-S_c)/1000));
mu_wd-(mu_w0+Ru*T*log((1000-S_d)/1000));
E_m-(T_s/F*(mu_sc-mu_sd)+T_w/F*(mu_wc-mu_wd));
E_am-E_m;
E_cm-E_m;
lambda_c-lambda_0+alpha*sqrt(C_c);
lambda_d-lambda_0+alpha*sqrt(C_d);
lambda_r-((lambda_c + lambda_d)/2);
C_r-((C_c+C_d)/2);
k-(lambda_r * C_r);
I-((V_cp-E_am-E_cm)/(r_am+r_cm+b_d/(sigma_d*lambda_d*C_d)+(b_c/(sigma_c*lambda_c*C_c))+...
r_cm/N_cp+2*h_r/(sigma*k*N_cp)));
J_s-(T_s*I/F - L_s*(C_c-C_d));
J_w-(T_w*I/F + L_w*Ru*T*(C_c-C_d)*10^(-5)*2);]
end
I notice you use F as both input and output. That is unlikely to be a good idea when you are calling fsolve with the function.
2 Comments
Walter Roberson
on 22 Sep 2019
You use T_s in your function, but it is not a parameter to the function. Where do you want your code to find T_s from?
chirag patel
on 23 Sep 2019
3 Comments
Walter Roberson
on 23 Sep 2019
Okay, go ahead. Follow the same strategy as before: all inputs should go into a single vector, and when you get the vector output from the solver, break it up into meaningful variables. Inside the function to be solved, expect everything as a single input, and break the values out into meaningful variables (this is what the num2cell and {:} is for)
See Also
Categories
Find more on Signal Integrity Kits for Industry Standards in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!