To my opinion there is one limitation inherent with almost all wear laws. It is that no subsurface material structural evolution has ever been considered. Also adhesion wear between the deformationally and chemically modified surfaces has neither been considered. Both these phenomena change greatly the subsurface hardness as well as the real contact area so that initiallly hard materials easily deform plastically.
As for me, I think one limitation is the use of mean pressure in the equation. And the wear coefficient can vary during the wear process because of lubrication condition and the move of materials.
Archard equation W=k*Wn*s/H allows the calculation of wear volume, but one problem is the hardness H of the softer material at the sliding interfase. Since wear at the boundary/mixed or hydrodynamic regime only takes place in the few microns range, this hardness is not easy to measure, due to the fact that bulk hardness is the one that is involved using vickers and rockwell methods and usually during sliding the top surface suffers strain hardening. This is the reason why is preferable to use wear rate Wr=V/Wn*s
Where V is wear volume Wn is normal load and s sliding distance, being this equation independent of hardness.