Paul -
I think there have been a few of these types of incidents but not many seem to be well documented. They appear to be connected with the accumulation of hydrogen gas in the bore of the pipe above the lower explosive limit. The souce of the gas varies. On some nuclear stations the hydrogen can result from the radiolysis of water, but for conventional plant simple corrosion can be a cause.
One similar incident on a site in my organisation was put down to corroison in a hollow sealed pipe support and grinding caused a quick flash-off. The cause has not been explained in detail but sampling revealed concentrations of hydrogen >10% of the lower explosion limit in many supports with some containign concentrations of >100% of the LEL (>4% hydrogen).
Unfortuntaly I don't have any further details and your information is not sufficient to enable a further discussion but the fact that you mention that it is enclosed suggests that it may be a similar mechanism.
Hope this is useful
Andy
We can probably point you in the right direction if you supply the following details:
pipe wall thickness
steel grade or specification or properties (particularly toughness, if known)
temperature of the pipe at the time of failure
what was inside the pipe and at what pressure
confirm that it was penetrated by a drill bit and not by a torch cut