Re: RE: [Piping Design] Hydrotest pressure

On Nov 2, 2005, at 5:57 AM, John Luf wrote:

> As I understand the testing procedure is along this
> line. The pressure is raised slowly watching the
> volume being input.

Funny you should mention. I used to get involved in this kind of testing in a previous life. We didn't do it for pipelines, just for specialized pressure vessels. If it's done properly it's precise enough to determine highly localized yielding or even redistribution of weld residual stress, but if you get to the point where you can't increase the pressure with additional water influx, you've already distorted the vessel to the point where the deformation is great enough so it won't pass the 'no visible permanent distortion' part of the Code hydro-test requirements. This can happen well before burst if the vessel details and material are ductile. You also need to vent all the air from the system or the test won't be reliable, since the compression of air bubbles diddles the results. The curve of pressure plotted against influx will be a straight line, if the system is properly filled.

I mentioned a test I was running where the vessel failed from a pre-existing crack in a weld. We were keeping track of the water influx very carefully, and had taken serious steps to vent _all_ the air from the vessel, We'd just finished reading some strain gages and recording the water influx when the the vessel let go with a helluva bang. It had split pretty much from end to end and the resulting jet of water had removed much of a sheet metal building wall. No one was hurt, although I'd just come back in off a platform on top of the vessel (an 8 foot dia 25 foot long test tank) when it blew. If I'd been out there, I'd've probably been tossed off by the rupture. It had held pressure for 3 or 4 minutes before burst without a drop in pressure or any indication of yielding. It was a straight brittle fracture in a weld HAZ, which should have been caught (mea culpa) by routine pre-test radiography. It might even have been found visually. That's why I've been pounding the table about inspection. We'd had a couple of years of good service out of the vessel, and the test was supposed to be a routine maintenance thing--no point doing any NDT becasue everything has been fine up to now. Just like the damn stock market--everything is always fine until it isn't fine any more.
Christopher Wright P.E. |"They couldn't hit an elephant at chrisw@skypoint.com | this distance" (last words of Gen.

.......................................| John Sedgwick, Spotsylvania