Liquid receiver sizing and refrigerant charge in reversible chiller

Hello,
my first post here.

I was wondering whether some of you would share some experience in reversible chiller, concerning refrigerant charge optimization and liquid receiver sizing.

The optimum refrigerant charge in a chiller is one of the most discussed and colntroversial topic from my point of view. I've come to read many methods for estimating a refrigerant charge, but the only one companies apply (at my knowledge) are empirical methods.

Usually they hypothize a certain % of liquid in the condenser, a smaller % in the evaporator, 100% in the liquid line and then they moltiply for the refrigerant density, and le jeux sont fait.
Some more sophisticated method use some thermodynimic model to simulate the void fraction during bi-phase flow in some conditions. I tried to use this method but didn't succeed to estimate not even closely the refrigerant charge.

This problem is especially frustrating in reversible air to water chiller when refrigerant charge required during heating operation is lower than in cooling mode... a liquid receiver than is needed... but how big must it be? If the different of required charge is not calculated correctly, it may lead to condenser flooding if too small or very lower capacity if too big..

What's your experience?