1. Hi,
I am doing my PhD in Physics (electrical Physics) and I need to solve a transcendental equation (first time I met such a problem).
This equation comes from a physics article writen in 1974.

Here is the equation:

exp(E/kT) = s/B * (kT^2)/E

- k and s are some physcal constants (Boltzmann constant etc..)
- B is the temperature rate, given in "degre Kelvin/seconds" (known value)
- T is the temperature (in Kelvin)
- E is the energy

and from that I need to plot E versus T...?!
In the article, the author doesn't really say how he gets the graph but the solution/graph he shows is a linear line such as
E = CT - D
where C and D are some constants that he determines from the slope and the intercept with the y-axis.
So I would like to know how do I solve this equation and get this graph ? Do I need to use any program/algorythm..?
Marc

2.

3. Well, the way you wrote that equation is a little ambiguous. I read it as: exp(E/(k*t)) = s/B * ((k*T)^2)/E. (Correct me if I'm wrong.)

With that equation, you can substitute u=E/(k*T) and v=s/B * k*T and get exp(u) = v/u or u*exp(u) = v. This can be solved with the Lambert W function. After that, it's easy to solve for E.

4. sorry for the writing.
It is:

exp (E / (k*T)) = s/B * (k*T^2)/E

(T^2 in the right hand part)

So if I do as you recommand, u=E/(k*T) and v=s/B * k*T, I end up with:

v = k*u*exp(u) ( k is a physical constant)

I checked on the link you gave for the Lambert W function but I am not familiar with this type of calculation...
could you develop a bit more ? how to use this lambert function in order to solve the pb ?
thanks a lot for your help.

5. Well, if that's the case, then just rewrite v = s/B * T.

As stated on the Lambert W page, if u*exp(u) = v, then u = W(v). (It uses X and Y, but, same thing.) Undo the substitutions and solve for E and you get something like E = k*T*W(s/B * T). (You should double-check that though.) This certainly isn't linear, but if the range of temperatures is right, it might could be estimated by a linear function.

6. If you are feeling adventurous you can to a power series expansion of E in terms of T and see where that gets you. Else use the implicit function theorem and see what the first two approximations will be about some point that is in the region you are interested.

 Bookmarks
##### Bookmarks
 Posting Permissions
 You may not post new threads You may not post replies You may not post attachments You may not edit your posts   BB code is On Smilies are On [IMG] code is On [VIDEO] code is On HTML code is Off Trackbacks are Off Pingbacks are Off Refbacks are On Terms of Use Agreement