I remember asking this exact question to my metalwork teacher at school. He told me off for being a smart arse, most likely because he couldn't ask the question. There is a lot of info on this subject out on the internet, the TLDR answer is the hole does get bigger.

There are doubtless videos, showing the ‘ball and ring’ thermal expansion school experiments, on the net?

As SOD points out, shrink fitting of parts could not happen if the hole ‘shrunk’!

Invar, a metal alloy, has a nearly zero coefficient of expansion.

SOD

One of them is antimony used in metal type metal to make sure letters are crisp

Same as water, which expands when it freezes. In its solid or liquid state it expands on heating, it's just the transition that is odd. Lucky though, otherwise ponds etc would freeze from the bottom up – but unlucky too 'cos Titanic hit a floating iceberg.

Hi, if you think about it, being that most things expand equally in all directions, unless otherwise restrained, think of the metal bar in a vice being heated up and then just drops out when it cools without opening the vice, then take the amount that the metal would expand towards the hole compared to the circumference of the hole, then it stands to reason that the inside of the hole will expand a greater distance and will have to become bigger. Of course, this is assuming that the whole piece is heated uniformly, because if it was only heated around the hole, it would be restrained by the cold metal surrounding it and the expansion would be greater sideways and the hole may even stay more or less the same size.

As for water does it really expand when it freezes? water becomes ice when it freezes, so maybe it can be considered no longer water. Ice of course is less dense because it expands and is lighter than water for the same volume, hence it floats on water. Change of state and all that.

Regards Nick.

The alloy invar still expands when heated, but at a lower rate than other metals.

Roy

Is your pendulum a grid iron type? If correctly proportioned, a grid iron pendulum should remain the same length irrespective of temperature.

Which is why Harrison used brass and steel (possibly iron ) elements in his grid iron pendulum

On Google, Colunbia metals state that invar has a near zero rate of thermal expansion.

So a straight pendulum made of invar will have coefficient of thermal expansion that is low.but not zero or negative.

Howard

Roy is correct that Invar needs to e in the right state of heat treatment to have near zero expansion. Chuck it away and replace it with carbon fibre, readily available as bar or (more convenient) tube I've got a bit in the rafters for my clock, which currently has a mild steel pendulum. It gains in winter and loses in summer, over a year it's pretty good!