you can always derivitise the glycerol with compounds such as BSTFA.
BSTFA reacts with -OH groups however. But you could convert a sample to a triglyceride by reacting it with something as simple as acetic acid. This lowers the boiling point of glycerol and change how it is retained even on a C18 column. You will find that most non-polar organic columns will differentiate between a glyceride that has either 1, 2 or 3 acetyl ester groups.
I moved this post to an application area for better visibility .