how to properly use 'solve' function properly?

hi, my question might sound weired. i have this equation:

x/L - 1 + (h+eta)/ho + A*log(((h+eta)/ho - A)/(1-A)) = 0

all parameters are known numbers except 'eta'. when i solve this equation in my TI calculator using 'nSolve' it works fine and i get a solution, but when im trying to solve it with matlab using 'eval(solve(f,eta))' im getting NaN as an answer.

what am i doing wrong? how to resolve this problem?

thank you.

Answer 1

[Walter Roberson]

Do not eval() a symbolic formula. If the formula contains symbolic variables then use subs() on it; if the formula does not contain symbolic variables then use double() to represent the number a double precision.

The general solution for the above is

-ho*(-1+A)*exp((-lambertw(-(-1+A)*exp(((1-A)*L-x)/(A*L))/A)*A*L+(1-A)*L-x)/(A*L))+A*ho-h

there are a number of circumstances under which the above could generate a NaN, such as if A or L is 0.

[Dany]

thank you i'll try that. if i may ask, how did you get the general solution?

i know that in some circumstances it could give a NaN but i was talking on the same input number in both cases, only in matlab i get NaN.

[Dany]

the general solution seems to work. but when i use the 'solve' function i keep getting an expretion with 'wrightOmega' in it and even when i use 'subs' i get an error saying double values cant be used in the 'wrightOmega' function.