MIT steam plant pipe cracking during hot-weld repair

Tom's recommendation thirty years ago to leave a crack (with a patch over it) on a winter-active steam line near 100 Memorial Drive, rather than continue trying to hot-weld it while moisture-driven hydrogen kept reopening the crack. Used to illustrate the rare case where the practical answer is "leave the crack."

I've told people leave cracks. Thirty years ago here at MIT, they had a steam pipe — they've got steam pipes all through the campus, and they generate steam over the generating plant. And they had a weeping leak in the steam plant over here near 100 Memorial Drive. They came to me and said, we've tried to weld it four or five times and it keeps cracking, because they were trying to hot-weld it while it still had steam in it, and you've got moisture coming out of the crack and guess what you get? Hydrogen, right. They were going to have to shut down the steam plant — most of campus uses heat from that — and it was the middle of the winter and they probably would have had freeze-ups in other places. They wanted to know how they could weld it. I said, well, why don't you just put a patch over it and then fix it in the summer, okay.