thermal expansion?

Question

is the coefficient of linear expansion (alpha) constant for all temperature ranges? why or why not?

is the expansion of a metal a reversible process?

Answer 1

I have not seen anything to show that it is not a constant value (rate) for each metal within the temperature ranges it is commonly employed. I know, it is an evasive answer. But the tables I have seen have shown a constant , not a variable.
The ranges would not be particularly wide for the usual engineering operations involving structures and machines. Ranges are several hundred degrees. More for alloys of titanium, less for aluminum.
It would be based on the increased molecular activity which is shown to be directly and linearly temperature depenent , therefore so would be the resulting expansion. When the temperature reaches a point where weakening or softening begins, both the structural and mechanical application has been compromised and failure occurs. Engineering tries to keep materials out of those temperatures.

Shrinkage patterns are used in foundries. These are made to a different measurement scale to compensate for the larger volume of molten metal poured into the molds. For the cooled part to be the right final size, the pattern from which the mold cavity is formed is larger by that scale.

Thermal expansion and contraction are reversible and repeatable within the temperature ranges of the engineering applications.

But there is another effect called creep. This is the permanant deformation of metals under high stress at elevated temperatures. This occurs before failure and leaves a permanent change in the metal shape even after it has cooled. A metal within the design temperature and within design load will not creep. It will expand and contract, but that is reversable. Creep , like exceeding the stress-strain yield point is not.

By the way, the specific volume of water does change non-linearily with temperature and so its rate of thermal expansion is indeed a variable.