Removing the offending pads will not provide any assurance the undesirable microstructures that may have resulted from the absence of preheat will be removed.
If the PWHT is carried out above the transformation temperature, the microstructure can be controlled with appropriate cooling. However, a PWHT conducted below the transformation temperature may not be sufficient to eliminate the undesirable microstructures. At best, a PWHT conducted below the critical is essentially a tempering operation. It is better than a sharp stick in the eye, but the end result may be less than what the customer bargained for.
As DekDee mentioned, subsurface cracks will not be detected by penetrant testing. The subsurface cracks, should they exist, will not be magically healed by PWHT. They may become an issue after the vessel is placed into service.
An engineered solution may be possible. Perhaps the contractor responsible for fabrication could qualify a WPS without preheat to demonstrate or to validate the welding conducted without preheat is capable of producing an HAZ with hardness values low enough that underbead cracking is not an issue.
Since the vessel is not code stamped, the absence of preheat is a contractual violation and not a code violation. The intent of the preheat is to mitigate the probability of underbead cracking. The use of preheat provides everyone with a warm fuzzy feeling that everything is good. It is no warrantee delayed cracking will not happen. Likewise, the absence of preheat is no assurance delayed cracking will occur. A demonstration that the welding procedure used, i.e., no preheat, produces an HAZ with a low hardness value may be all that is necessary to ease the concern that underbead cracking is a potential problem and the vessel can be completed and placed into service.
If there is an lingering concern there may be underbead cracking, the contractor may further ease the Owner's concerns by using UT with a straight beam transduced to detect or to verify the absence of underbead cracks. An underbead crack would most likely lie in a plain parallel to the surface of the rolled plate. As long as the shell thickness is thicker than the near field of the transducer, cracks should be detectable from the inside of the vessel as a reflected signal that shifts to the left of the back reflector thereby indicating a difference in "thickness." That difference in thickness would be indicative of an underbead crack. It could also be indicative of a lamination, but that would be ruled out if the reflectors are relatively small areas limited to the HAZ of the weld.
Best regards - Al
