I'd like to know if my demonstration is "alright". Sorry for any grammar mistakes, english is not my first language.
I have to show that "if A and B are equivalent formulas, then you can replace A with B without changing any resulting truth value for any possible valuation of a formula φ". For that, I used structural induction:
Let φ and B be formulas and A is a subformula of φ.
Base case: Let A be equivalent to any atomic formula p. In this case, for every valuation v, v(A) = v(p). That is, all rows of the truth table are identical, so we can replace A with p and keep all possible truth values of φ.
Case ¬: Assume A is equivalent to B. We know, by definition, that the negation operator inverts the truth value of a formula. That is, for every valuation v such that v(A) = v(B), we also have v(¬A) = v(¬B). Therefore, we can replace A with B in a formula and keep the truth values for every possible valuation of A.
Case p o q: By the premise, for every possible valuation v, v(A) = v(p o q), for o ∈ {∧, ∨, →}. So, for any possible truth value in a formula, we can keep it by replacing a with p o q.
I'd rather do the structural induction for $\varphi$ (it looks like you are doing it for $A$):
If $A=\varphi$, the claim follows immediately from the equivalence of $A$ and $B$ (as in your proof). So assume from now on that $A\ne \varphi$