piece
A compression coupling holds by squeezing. That is the whole of its argument.
On December 15, 1977, near 747 Massachusetts Street in Lawrence, Kansas, a section of polyethylene gas main came out of a metal coupling. Gas leaked. There was an explosion and a fire, and two people died.
The main was a hybrid. A 2-inch plastic pipe had been fed by the insertion renewal method inside an existing 3-inch metal main, so the new line lived in the bore of the old one. Where the plastic met an adjacent section of metal, a standard Dresser Style 90 compression coupling joined them: a sleeve, gaskets, and nuts drawn tight until the gaskets bit down around the pipe. It was a grip, not a weld. Friction had been made into infrastructure.
Plastic answers temperature in its own voice. The federal laboratory that later studied the failure wrote that the pipe pulled out of its coupling apparently as a result of thermal contraction of the plastic. Apparently is the word an investigator uses when the mechanism is visible and the proof is still in the machine.
So the machine was built. The tests were lettered A through H. Specimens were cut from three-foot sections of the main and fitted with a new coupling and new gaskets. The tests varied installation procedure, environmental exposure and pulling speed. A special fixture fastened one nut of the coupling to the crosshead of a testing machine. Then test A was run again with the failed pipe itself in the original coupling from the accident site. That test was called A-2. Later came A-3 and A-4.
I keep returning to A-2. The same joint was asked to let go a second time, on purpose, while someone wrote down how hard it had to be pulled.
What the tests measured was axial stress, the force along the length of the pipe needed to draw it out. It is a number for the moment a grip becomes a suggestion. The number was meant to feed a larger question: whether the rules for installing gas systems that combine plastic and metal components should be revised. A coupling is a treaty between two materials with different opinions about winter. A regulation is a treaty about the treaty.
What escapes first here is not the gas. It is the pipe. It withdraws from the sleeve in a retreat paid for in degrees, and the gas only follows the absence it leaves.
The report I hold tells me what was asked, with what fixture, and that the failed specimen was pulled again. It does not tell me the answer. Somewhere a gauge recorded how much force it took. The street had already taken that measurement once, without one.
Night in a small Kansas town in winter. In the foreground, an excavated trench in frozen earth exposes an old dark metal gas main. A pale polyethylene pipe has slid backward out of a bolted metal compression coupling, and a thin black gap of empty sleeve shows between them, with frost on the rubber gaskets and the nuts. In the midground, a tall steel tensile testing machine stands in the snowy street under an amber sodium streetlamp, its crosshead clamped to a coupling nut, slowly pulling a short pipe specimen. In the background are brick storefronts with dark windows and a lone figure in a heavy coat holding a clipboard, watching. Cold blue shadow, sodium glow, breath vapor, snow dust, and a hushed sense of aftermath.