# Simplification assumptions strange behavior

I am trying to cancel common factors under square roots in some expression with all positive real variables: $$\text{Simplify}\left[\frac{\left(3+\sqrt{6}\right) \sqrt{g} \text{nn}^4}{2 \sqrt{\text{nn}^7 P*T}},\text{nn}>0\land P>0\land T>0\right]$$

which returns the same thing $$\frac{\left(3+\sqrt{6}\right) \sqrt{g} \text{nn}^4}{2 \sqrt{\text{nn}^7 P T}}.$$

However, removing the assumptions for $P,T$ makes it work: $$\text{Simplify}\left[\frac{\left(3+\sqrt{6}\right) \sqrt{g} \text{nn}^4}{2 \sqrt{\text{nn}^7 P T}},\text{nn}>0\right]$$ $$\frac{\left(3+\sqrt{6}\right) \sqrt{g \text{nn}}}{2 \sqrt{P T}}.$$

Why does it happen? And a more important question, what is a better way to cancel common factors under roots with positive real variables in the more general case with larger expressions and many variables to cancel, so that I am sure no common factors remain?

Maybe you are looking for the function PowerExpand?

PowerExpand[((3 + Sqrt) nn^4 Sqrt[g])/(2 Sqrt[nn^7 P*T])]

(*   ((3+Sqrt) Sqrt[g] Sqrt[nn])/(2 Sqrt[P] Sqrt[T])  *)

• Good! It does the job. – xivaxy Oct 17 '14 at 17:26

It works for me if I turn the simplifications into a list:

{nn>0,P>0,T>0}.

• Simplify[((3 + Sqrt) nn^4 Sqrt[g])/( 2 Sqrt[nn^7 P*T]), {nn > 0, P > 0, T > 0}] still gives the same result for me. I am using Mathematica 10.0.0.0 – xivaxy Oct 17 '14 at 17:03
• Odd. What about using Refine instead of Simplify? Both work for me. – Lauren Pearce Oct 17 '14 at 17:14
• Refine[...,{nn > 0, P > 0, T > 0}] works nicely. So, finally I need to do either Simplify[Refine[...,assumptions]], or Simplify[PowerExpand[...],assumption] – xivaxy Oct 17 '14 at 17:28