I understand that you should wait until weldment cool down for safety reason. Who give you the period 48 hours? Is this from project specification?

Bader,

Whether there as a delay necessary depends on the type of material.  Certain high strength steels are susceptible to delayed cracking caused by hydrogen in the material.  I have seen specifications mandate a minimum of 24 hours, 48 hours, or 7 days after heat treatment prior to MT.  Apparently there is a history of certain alloys being fine immediately after cooling, but cracking hours or days later due to hydrogen in the material.

The delay is the exeption rather than the rule.  Unless you suspect the material you are using is susceptible to delayed cracking, you should be fine to test as soon as cool.

JR97

I have to support JR97. So called "cold cracking" or "hydrogen cracking" is the factor in selecting the delay.

The reason for the delay before liquid penetrant or magnetic particle examination is to allow all the hydrogen from metal to migrate to microvoids in material and cause cracking. Usually, one would perform those examinations prior to PWHT to avoid costly and embarassing repairs. The reason for NDE after PWHT might be reheat cracking in Cr/Mo or shrinkage, which seems to be the case you are describing. Nozzles, being discontinuities in the shell with additional reinforcement of the neck, can develop cracks in the welds during PWHT. However, you can not disregard a possibility that the furnace atmosphere may contain hydrogen. That was a mouthfull...
To recap: if it is related to hydrogen, delay is usually a good idea. 24 hours should be enough to identify if there is a cold cracking problem. If such a problem is found, I sugest a second test after 48 hours to assess the complete problem.
In the case of shrinkage, if hydrogen is not suspected, no delay is necessary. You could read more in Welding Journal.

Putting Human Factor Back in Engineering