How should I use the averaging approximation to find this?

  • Thread starter Thread starter Math100
  • Start date Start date
  • Tags Tags
    Approximation
Math100
Messages
813
Reaction score
229
Homework Statement
Consider the differential equation ## \ddot{x}+\epsilon h(x, \dot{x})+x=0, 0<\epsilon<<1 ##. Let ## x(t) ## and ## y(t)=\dot{x}(t) ## be expressed in terms of the polar coordinates ## a(t) ## and ## \theta(t) ## through ## x(t)=a(t)\cos\theta(t), y(t)=a(t)\sin\theta(t) ##. Note that ## a(t) ## and ## \theta(t) ## satisfy the following differential equations:
## \dot{a}=-\epsilon\sin\theta h(a\cos\theta, a\sin\theta) ##,
## \dot{\theta}=-1-\frac{\epsilon}{a}\cos\theta h(a\cos\theta, a\sin\theta) ##.

a) By letting ## \theta(t)=\psi(t)-t ##, transform Equations ## \dot{a}=-\epsilon\sin\theta h(a\cos\theta, a\sin\theta) ## and ## \dot{\theta}=-1-\frac{\epsilon}{a}\cos\theta h(a\cos\theta, a\sin\theta) ## into non-autonomous differential equations for ## a(t) ## and ## \psi(t) ##. Hence use an averaging theorem to show that provided that ## \epsilon ## is small enough, Equations ## \dot{a}=-\epsilon\sin\theta h(a\cos\theta, a\sin\theta) ## and ## \dot{\theta}=-1-\frac{\epsilon}{a}\cos\theta h(a\cos\theta, a\sin\theta) ## can be well approximated by ## \dot{a}=-\epsilon p_{0}(a), \dot{\theta}=-1-\frac{\epsilon r_{0}(a)}{a} ##, where ## p_{0}(a)=\frac{1}{2\pi}\int_{0}^{2\pi}\sin u h(a\cos u, a\sin u)du, r_{0}(a)=\frac{1}{2\pi}\int_{0}^{2\pi}\cos u h(a\cos u, a\sin u)du ##.

b) Now consider the equation ## \ddot{x}+\epsilon(x^4-b^4)\dot{x}+x=\epsilon\gamma x\dot{x}^4, 0<\epsilon<<1, b>0 ##, where ## b ## and ## \gamma ## are constants. Use the averaging approximation of part (a) to show that the system has a stable limit cycle with amplitude ## A ## and angular frequency ## \omega ## given approximately by ## A=8^{1/4}b, \omega=1-\frac{1}{2}\epsilon\gamma b^4 ##. (Hint: Note the following identity: ## cos^{4}\theta\sin^{2}\theta=\frac{1}{16}+\frac{1}{32}\cos 2\theta-\frac{1}{16}\cos 4\theta-\frac{1}{32}\cos 6\theta. ##)
Relevant Equations
Not given.
a) Proof:

Consider the equations ## \dot{a}=-\epsilon\sin\theta h(a\cos\theta, a\sin\theta) ## and ## \dot{\theta}=-1-\frac{\epsilon}{a}\cos\theta h(a\cos\theta, a\sin\theta) ##.
Let ## \theta(t)=\psi(t)-t ##.
Then ## \dot{\theta}(t)=\dot{\psi}(t)-1 ##.
By direct substitution of ## \dot{\theta}(t)=\dot{\psi}(t)-1 ##, we have ## \dot{\psi}(t)-1=-1-\frac{\epsilon}{a}\cos\theta h(a\cos\theta, a\sin\theta)\implies \dot{\psi}(t)=-\frac{\epsilon}{a}\cos(\psi(t)-t)\cdot h(a(t)\cos(\psi(t)-t), a(t)\sin(\psi(t)-t)) ##.
Hence, the non-autonomous differential equations for ## a(t) ## and ## \psi(t) ## are ## \dot{a}(t)=-\epsilon\sin(\psi(t)-t)\cdot h(a(t)\cos(\psi(t)-t), a(t)\sin(\psi(t)-t)) ## with ## \dot{\psi}(t)=-\frac{\epsilon}{a}\cos(\psi(t)-t)\cdot h(a(t)\cos(\psi(t)-t), a(t)\sin(\psi(t)-t)) ##.

b) Proof:

Consider the equation ## \ddot{x}+\epsilon(x^4-b^4)\dot{x}+x=\epsilon\gamma x\dot{x}^4, 0<\epsilon<<1, b>0 ##, where ## b ## and ## \gamma ## are constants.
By definition, the system ## \ddot{x}+\epsilon h(x, \dot{x})+x=0, 0<\epsilon<<1 ## has an approximately circular limit cycle given by the equations ## x=a_{0}\cos\omega t, y=\dot{x}=-a_{0}\sin\omega t, \omega=1 ##, where ## a_{0} ## satisfies the equation ## g(a)=0 ## such that ## g(a)=\epsilon a\int_{0}^{2\pi} h(a\cos t, -a\sin t)\sin t dt ##. The limit cycle is stable if ## g'(a_{0})<0 ## and unstable if ## g'(a_{0})>0 ##.
Note that the first-order approximation to the frequency ## \omega ## of the limit cycle equation is given by ## \omega=1+\frac{\epsilon}{2\pi a_{0}}\int_{0}^{2\pi} h(a_{0}\cos\theta, a_{0}\sin\theta)\cos\theta d\theta ##.
Then we have ## h(x, y)=h(x, \dot{x})=(x^4-b^4)\dot{x} ## where ## x=a\cos t ## and ## y=\dot{x}=-a\sin t ##.
This gives ## g(a)=\epsilon a\int_{0}^{2\pi} h(a\cos t, -a\sin t)\sin t dt=\epsilon a\int_{0}^{2\pi} (a^{4}\cos^{4} t-b^{4})(-a\sin t) dt=-\epsilon a^{2}\int_{0}^{2\pi}(a^{4}\cos^{4} t-b^{4})\sin t dt ##.
Observe that ## g(a)=-\epsilon a^{2}\int_{0}^{2\pi}(a^{4}\cos^{4}t-b^{4})\sin t dt\implies g(a)=-\epsilon a^{2}(\int_{0}^{2\pi} a^{4}\cos^{4} t\cdot\sin t dt)+\epsilon a^{2}\int_{0}^{2\pi} b^{4}\sin t dt\implies g(a)=-\epsilon a^{6}\int_{0}^{2\pi}\cos^{4} t\cdot\sin t dt+\epsilon a^{2}b^{4}\int_{0}^{2\pi}\sint dt\implies g(a)=-\epsilon a^{6}(-\frac{1}{5}+\frac{1}{5})+\epsilon a^{2}b^{4}(-1+1) ##.
Hence, ## g(a)=0 ##.

Up to here, I don't think the result I've got from part b) is correct, since ## g(a)=0 ##. But I want to know what exactly is the averaging approximation from part (a) that I should use/apply in part b) to obtain the given amplitude and the angular frequency ## \omega ##. Also for part (a), after I've got non-autonomous differential equations, how should I use/apply the averaging theorem? I have limited resource and couldn't find the exact definition for the averaging theorem from my book.
 
Physics news on Phys.org
What "averaging theorems" do you know?
 
pasmith said:
What "averaging theorems" do you know?
I think I've made mistakes in part b). Here's what I've revised so far:
Part b) Proof:

By definition, the system ## \ddot{x}+\epsilon h(x, \dot{x})+x=0, \lvert\epsilon\rvert<<1 ## has an approximately circular limit cycle given by the equations ## x=a_{0}\cos\omega t, y=\dot{x}=-a_{0}\sin\omega t, \omega=1 ##, where ## a_{0} ## satisfies the equation ## g(a)=0 ## such that ## g(a)=\epsilon a\int_{0}^{2\pi} h(a\cos t, -a\sin t)\sin t dt ##. The limit cycle is stable if ## g'(a_{0})<0 ## and unstable if ## g'(a_{0})>0 ##.
Consider the equation ## \ddot{x}+\epsilon(x^4-b^4)\dot{x}+x=\epsilon\gamma x\dot{x}^{4}, 0<\epsilon<<1, b>0 ##, where ## b ## and ## \gamma ## are constants.
Then we have ## h(x, y)=h(x, \dot{x})=(x^4-b^4)\dot{x}-\gamma x\dot{x}^{4} ##, because ## \ddot{x}+\epsilon(x^4-b^4)\dot{x}+x=\epsilon\gamma x\dot{x}^{4}\implies \ddot{x}+\epsilon(x^4-b^4)\dot{x}+x-\epsilon\gamma x\dot{x}^{4}=0\implies \ddot{x}+\epsilon[(x^4-b^4)\dot{x}-\gamma x\dot{x}^{4}]+x=0 ##.
Let ## x=a\cos t ## and ## y=\dot{x}=-a\sin t ##.
Note that ## h(x, y)=h(x, \dot{x})=(x^4-b^4)\dot{x}-\gamma x\dot{x}^{4}\implies h(x, y)=h(x, \dot{x})=(a^4\cos^{4} t-b^4)(-a\sin t)-\gamma(a\cos t)(-a\sin t)^{4} ##.
This gives ## h(x, y)=h(x, \dot{x})=ab^{4}\sin t-a^{5}\cos^{4} t\cdot\sin t-\gamma a^{5}\cos t\cdot\sin^{4} t ##.
Observe that ## g(a)=\epsilon a\int_{0}^{2\pi} h(a\cos t, -a\sin t)\sin t dt\implies g(a)=\epsilon a\int_{0}^{2\pi}(ab^{4}\sin t-a^{5}\cos^{4} t\cdot\sin t-\gamma a^{5}\cos t\cdot\sin^{4} t)\sin t dt\implies g(a)=\epsilon a^{2}\int_{0}^{2\pi}(b^{4}\sin^{2} t-a^{4}\cos^{4} t\cdot\sin^{2} t-\gamma a^{4}\cos t\cdot\sin^{5} t)dt\implies g(a)=\epsilon a^{2}[b^{4}\int_{0}^{2\pi}\sin^{2} t dt-a^{4}\int_{0}^{2\pi}\cos^{4} t\cdot\sin^{2} t dt-\gamma a^{4}\int_{0}^{2\pi}\cos t\cdot\sin^{5} t dt]\implies g(a)=\epsilon a^{2}[b^{4}(\pi-0)-a^{4}(\frac{24\pi}{192}-0)-\gamma a^{4}(0)] ##.
Hence, ## g(a)=\epsilon a^{2}(b^{4}\pi-\frac{24a^{4}\pi}{192})=0\implies g(a)=\epsilon a^{2}(b^{4}\pi-\frac{a^{4}\pi}{8})=0\implies g(a)=\epsilon a^{2}\pi(b^{4}-\frac{a^{4}}{8})=0 ##, where ## b^{4}-\frac{a^{4}}{8}=0 ## for ## a\neq 0 ##, so ## a^{4}=8b^{4}\implies a=\pm 2^{3/4}\cdot b ##.
Since there are limit cycles at ## a=\pm 2^{3/4}\cdot b ##, it follows that ## g'(a)=\epsilon\pi(2ab^{4}-\frac{3a^{5}}{4}) ##.

Up to here, what should I do for part b) in order to find the given amplitude ## A ## and the angular frequency ## \omega ##? As for part a) and the averaging theorem, I've done research and found out that the averaging theorem talks about the system with the following form: ## \dot{x}=\epsilon f(x, t, \epsilon), 0\leq\epsilon<<1 ## of a phase space variable ## x ##. The fast oscillation is given by ## f ## versus a slow drift of ## \dot{x} ##. The averaging method yields an autonomous dynamical system ## \dot{y}=\epsilon\frac{1}{T}\int_{0}^{T} f(y, s, 0)ds=: \epsilon\bar{f}(y) ##. But how should I apply this on part a)?
 
Last edited:
How does the unperturbed system in part (a) behave?
 
There are two things I don't understand about this problem. First, when finding the nth root of a number, there should in theory be n solutions. However, the formula produces n+1 roots. Here is how. The first root is simply ##\left(r\right)^{\left(\frac{1}{n}\right)}##. Then you multiply this first root by n additional expressions given by the formula, as you go through k=0,1,...n-1. So you end up with n+1 roots, which cannot be correct. Let me illustrate what I mean. For this...
Back
Top