How can the differential equation for a sphere in a room be solved?

AI Thread Summary
The discussion revolves around solving a differential equation related to the temperature change of a sphere illuminated by a constant intensity in a controlled environment. The equation is expressed as dT/dt = a - bT^4, where various methods, including separation of variables and partial fraction decomposition, are suggested for finding T(t). Participants provide insights into integrating the equation, leading to complex expressions involving arctan and logarithmic functions. Despite the attempts, there is a consensus that expressing T(t) directly as a function of t may not be feasible, but a relationship can be established. The conversation highlights the challenges of solving nonlinear differential equations and the potential for deriving special functions.
atta_bo-y
Messages
5
Reaction score
0
Hey guys,

I have a smallish problem with a differential equation.

Set up:
There is a sphere in a room; and it's being illuminated with intensity I from one side.
The room temperature (T_{environment}) is set to be constant.

[ latex ]
\Delta Q = Q_{in} - Q_{out}\\
= I\Delta t A - P_{out}\Delta t + P_{in}\Delta t\\
= I\Delta t A - \sigma A T^{4} \Delta t + \sigma A T^{4}_{environment} \Delta t\\
\Delta Q = cm\Delta T\\
\frac{\Delta T}{\Delta t} = I\Delta t \pi r^2 - \sigma 4 \pi r^2 T^{4} \Delta t + \sigma 4 \pi r^2 T^{4}_{environment} \Delta t
[ /latex ]

Hence we have a differential equation in the form of

!
[ latex ]
\frac{dT(t)}{dt} = a - b*T^4(t)
[ /latex ]
!

I have tried different methods... But none of them worked... (And wolframalpha can only solve for t :-( )

Thanks for your consideration ;-)
atta_bo-y
 
Physics news on Phys.org
Well, first don't put the spaces in the <latex> bracket. Then have you tried to separate variables and then a partial fraction decomposition ?>
 
\frac{dT(t)}{dt} = a - b*T^4(t)

Any tips for how to get T(t)??

<br /> t+c &amp;= \frac{-\log(bT(t)- a) + \log(a + bT(t)) + 2\arctan\left(\frac{bT(t)}{a}\right)}{4a^3b}<br />
<br /> (t+c)4a^3b &amp;= \log\left(\frac{a+bT(t)}{-a + bT(t)}\right) + 2\arctan\left(\frac{bT(t)}{a}\right)<br />
<br /> (t+c)2a^3b = \arctan\left(\tan\left(\frac{1}{2}\log\left(\frac{a+bT(t)}{-a + bT(t)}\right)\right)\right) + \arctan\left(\frac{bT(t)}{a}\right)<br />
<br /> (t+c)2a^3b = \arctan\left(\frac{\tan\left(\frac{1}{2}\log\left(\frac{a+bT(t)}{-a + bT(t)}\right)\right) + \frac{bT(t)}{a}}{1-\tan\left(\frac{1}{2}\log\left(\frac{a+bT(t)}{-a + bT(t)}\right)\right)\frac{bT(t)}{a}}\right)<br />

furthermore I can set

<br /> \frac{a+bT(t)}{-a+bT(t)} = 1 + \frac{2a}{bT(t) - a}<br />

Now this looks like a sisyphean direction to me... Please correct me if I'm wrong...

Thanks
 
Usually at this point you define a new special function named after yourself and put a box around it. :smile:
 
Easy ;)

T(t)=\left(\frac{a}{b}\right)^{\frac{1}{4}}
 
Well... if the differential equation youre trying to solve is:

\frac{dT(t)}{dt}=a-bT^4

Then, there you go: first, separate the variables:

\frac{dT}{a-bT^4}=dt

and integrate each side:

\int \frac{1}{a-bT^4}dT=t+C
(Equation 1)

The T integral is somewhat lenghty to calculate. For making the calculations a bit better-looking, write:

a=e^4 and b=f^4.

Then decompose the integrand as:

\frac{1}{a-bT^4} = \frac{1}{e^4-f^4 T^4} = \frac{1}{2e^2} \left[\frac{1}{e^2+f^2 T^2} + \frac{1}{e^2-f^2 T^2} \right]

So the integral becomes:

\int \frac{1}{a-bT^4}dT = \frac{1}{2e^2} \left[ \int \frac{1}{e^2+f^2 T^2} dT + \int \frac{1}{e^2-f^2 T^2} dT \right]

The first integral can be calculated using the substituition

\frac{fT}{e}=tan(x)

Which after some calculations leads to:

\int \frac{1}{e^2+f^2 T^2} dT = \frac{1}{ef} arctan \left( \frac{fT}{e} \right)

The second integral can be decomposed as:

\int \frac{1}{e^2-f^2 T^2} dT = \frac{1}{2e} \int \left[ \frac{1}{e-fT} + \frac{1}{e+fT} \right] dT

And integrating it we get:

\int \frac{1}{e^2-f^2 T^2} dT = \frac{1}{2ef} ln \left[ \frac{e+fT}{e-fT} \right]

So, the final result is:

\int \frac{1}{a-bT^4}dT =\frac{1}{2e^3 f}\left[ arctan\left(\frac{fT}{e}\right) + \frac{1}{2} ln\left(\frac{e+fT}{e-fT}\right)\right]

As shown in the equation called Equation 1, this is equal to t+C (C being an arbitrary constant), so the solution for the differential equation is:

\frac{1}{2e^3 f}\left[ arctan\left(\frac{fT}{e}\right) + \frac{1}{2} ln\left(\frac{e+fT}{e-fT}\right)\right] = t+C

So you see you can write t as a function of T(t), but i think there is no way to find a way to write T(t) as a function of t. Anyway, its the best we can do... i hope it be enough.

BTW, in the above equation you can remove the e's and f's by writing them in function of the original parameters a and b as:

e=\sqrt[4]{a} and f=\sqrt[4]{b},

but the equation you will get is even uglier:

\frac{1}{2 \sqrt[4]{a^3 b} } \left[ arctan\left(\sqrt[4]{ \frac{b}{a} } T\right) + \frac{1}{2} ln\left(\frac{ \sqrt[4]{a}+\sqrt[4]{b}T }{ \sqrt[4]{a}-\sqrt[4]{b}T }\right)\right] = t+C
 
Last edited:
Back
Top