[charlie]

it is difficult to make a recommendation based on the information. Electrofusion joints are relatively easy to test (0° looking for a laminar-type nonfusion). You have not explained what the failure mechanism is. You have indicated that only RANDOM PAUT is being done. But you have not indicated that it was the tested joints that failed or one or many of the untested joints. How long after being put into service did the failures occur? Until you can identify how the failures are occurring there is not much anyone can do to recommend a method to detect the flaw. Creep slow-crack growth in PE pipe may start from a very small initiation site. Perhaps you could carry out 100% PAUT (i.e. test all joints) and your researchers could provide you with more input before you start looking for other solution.