I am trying to show that this is true using the binomial formula or some taylor expansion:
$\frac{1}{(1+\epsilon \sum\limits_{n=0}^\infty Z_n(t) \epsilon^n)^2} = 1 - 2Z_0\epsilon + (3Z_0^2-2Z_1)\epsilon^2+O(\epsilon^3)$
So far i tried doing it using the binomial formula but I can't get it to work:
$\frac{1}{(1+\epsilon \sum\limits_{n=0}^\infty Z_n(t) \epsilon^n)^2} = \frac{1}{1+(\epsilon \sum\limits_{n=0}^\infty Z_n(t) \epsilon^n)^2+2\epsilon \sum\limits_{n=0}^\infty Z_n(t) \epsilon^n} = \frac{1}{1+(\epsilon(Z_0\epsilon^0+Z_1\epsilon^1+Z_2\epsilon^2))^2+2\epsilon(Z_0\epsilon^0+Z_1\epsilon^1+Z_2\epsilon^2) + O(\epsilon^3)} = \frac{1}{\epsilon^2 Z_0^2+2\epsilon^2 Z_1+2\epsilon Z_0+1+O(\epsilon^3)} $
Consider first the denominator as $$A=\left(1+\epsilon \sum _{n=0}^p Z_n \epsilon ^n\right)^2$$ Expand it using the binomial theorem to get $$A=1+2 Z_0 \epsilon +\left(Z_0^2+2 Z_1\right) \epsilon ^2+2 (Z_0 Z_1+Z_2) \epsilon ^3+\left(Z_1^2+2 Z_0 Z_2+2 Z_3\right) \epsilon ^4+\cdots $$ Now, use the long division to get $$\frac 1 A=1-2 Z_0 \epsilon +\left(3 Z_0^2-2 Z_1\right) \epsilon ^2+\left(-4 Z_0^3+6 Z_1 Z_0-2 Z_2\right) \epsilon ^3+\cdots$$
Another way to do it (since you are not concerned by high order expansions) could be to write $$1=A \sum_{i=0}^k a_i \epsilon^i$$ Expand the product for a few terms and cancel the coefficient for each power of $\epsilon^i$. You will get the same result.