What makes you think the NER on the two vessels is twice the others? Have you done a test on them? What data do you have that suggests a problem? If data is circumstantial, you may need to test to verify there really is a problem.
Diagnosing a tank such as this can be difficult. Once you are sure there's an issue, (and not just a leaky PB valve or something like that such as EdStainless mentions) start with a check on vacuum level. Verify it is on the order of 10 micron or less during operation. If you detect high vacuum levels, see if you can determine what gas is in the annular space. Nitrogen obviously indicates an internal leak, air or water indicates an external leak.
One issue with perlite is that it can be compacted if the vessel undergoes numerous thermal cycles such as being emptied regularly or possibly just filled much more frequently. The problem is that perlite can be crushed by thermal cycles when the inner tank changes dimensions (shrink/expand/shrink/expand, etc...). The perlite is relatively fragile and can be crushed which then drops to the bottom, opening holes in the insulation on the top of the vessel. It can also get packed in around piping, causing stresses and cracking of the pipes.