Question: What kind of family structure best protects children from child abuse?
Answer: Married biological parents.
This study (pages 5–25) shows that answer, by an enormous margin, compared to every other family structure including solo parents, other married parents (primarily remarried), single parents living with a partner, cohabiting parents, and no parents. The big gap is between the intact married biological family and every other family form.
To answer your other question: Same-sex couple parents are not a specific option studied (probably because the number of such parents is so small). Perhaps, alone of all the family structures studied, children living with same-sex couples do just as well as children in intact married families. Perhaps. But does this study, which is one of a hundred, give Ted Olson and David Boies pause at all, I wonder, in their quest to prove that science says same-sex unions are just the same as opposite-sex unions, even with regard to child well-being?