From the FRA’s web site.
Sec. 230.36 Hydrostatic testing of boilers.
(a) Time of test. The locomotive owner and/or operator shall
hydrostatically test every boiler at the following times:
(1) During the 1472 service day inspection, and at every annual
inspection thereafter;
(2) After making any alteration to the boiler;
(3) After installing a flush patch on an unstayed portion of the
boiler; and
(4) After any riveting on an unstayed portion of the boiler.
(b) Method of testing. The metal temperature of the boiler shall be
between 70 degrees Fahrenheit and 120 degrees Fahrenheit each time it is
subjected to any hydrostatic pressure. Hydrostatic testing required by
these rules shall be conducted at 25 percent above the MAWP.
(c) Internal inspection. An internal inspection of the boiler shall
be conducted following any hydrostatic test where the pressure exceeds
MAWP.
The requirement for a minimum temperature of 70 degrees is due to steel’s tendency to become more brittle at very low temperatures. The idea behind a hydro is to confirm the boiler’s strength without causing any unnecessary damage to the boiler from the hydro itself. It is possible that a part that would pass a 70 degree hydro with no problems could fail when subjected to the same pressure at 33 degrees. The maximum temperature of 120 degrees is for the safety of the people involved, so if there is a failure, nobody gets burned by hot water.
All locomotive boilers in the US are hydroed to 1.25 times MAWP, even though the standard in stationary service is (or at least used to be) 1.50 times MAWP. It is counterproductive to hydro a boiler to a higher pressure than is necessary, as the hydro itself is likely to cause damage to the boiler.
Pulling the dome lid and sending a man inside is always required so he can inspect the interior of the boiler for any damage caused by the hydro. This is the case even if the boiler had an internal inspection just the day before.
When working with locomotive boilers, the question during the hydro is not “does it leak”, but “where does it leak”, because they all do. In 27 years of earning my living repairing steam locomotives, I have only seen one boiler (and that one, one time only) that held pressure for more than a few seconds without the hydro pump feeding it. It is common for riveted joints, staybolts, and tubes that are all steam tight at operating temperature to leak slightly at room temperature. That combined with the dozens of valves attached to the boiler, each one a potential source of leaks, and sometimes you need a fairly good sized pump to keep up.