Doesn't a restriction introduced while finding the general solution to a differential equation impact its interval of validity?

194 Views Asked by At

I've been working on this example in Boyce's Elementary Differential Equations and Boundary Value Problems book. The example shows how to find a solution to $$y''+4y=3\csc(t)$$ through the method of variation of parameters.

We know that the general solution of the homogeneous equation corresponding to the equation above, $y''+4y=0$, is $y_c(t)=c_1\cos(2t)+c_2\sin(2t)$. So the idea of the method is to replace $c_1$ and $c_2$ with the functions $u_1$ and $u_2$, respectively, and then try to determine these functions so that the expression $y=u_1\cos(2t)+u_2\sin(2t)$ is a solution to the non-homogeneous equation. At this point we would calculate the first and second derivatives of this last expression and substitute for $y$ in the non-homogeneous equation, what would lead us to one equation involving a combination of $u_1$ and $u_2$, and their derivatives. That allows us to impose a second condition so that we obtain two equations for two unknown functions $u_1$ and $u_2$.

In the example, they require that $u'_1\cos(2t)+u'_2\sin(2t)=0$. That condition, after we've done the calculations, leads us to the following system: $$\begin{cases} u'_1\cos(2t)+u'_2\sin(2t)=0 \\\\ -2u'_1\sin(2t)+2u'_2\cos(2t)=3\csc(t) \end{cases} $$

Now, the example goes on to solve this system beginning with $$u'_2=-u'_1\frac{\cos(2t)}{\sin(2t)}$$ Since we can't divide by zero, we have to suppose also that $\sin(2t)\ne0$, which implies $t\ne n\frac{\pi}{2}, n \in \mathbb{Z}$. But the book doesn't make this observation. It just moves on to find both $u'_1$ and $u'_2$ and, after integration, declares that $u_1=-3\sin(t)+c_1$ and $u_2=\frac{3}{2}\ln{|\csc(t)-\cot(t)|}+3\cos(t)+c_2$. These two last results, after some more manipulations, give us the general solution to our differential equation: $$y=3\sin(t)+\frac{3}{2}\ln{|\csc(t)-\cot(t)|}\sin(2t)+c_1\cos(2t)+c_2\sin(2t)$$

My question lies in that consideration not made by the book of $\sin(2t)$ not being equal to zero. I think that if we were solving a system of equations problem, not a differential equation one, taking into consideration only the system above, we would have to analyze both the case when $\sin(2t)\ne0$ and the case when $\sin(2t)=0$. Why doesn't the book take this restriction into account? Doesn't that impact at all the interval of validity of the general solution found at the end?

OBS: I'm taking my first course on differential equations, so I may be missing some key information or insight here.