The Final Leak Test

The roof passed its first leak test with flying colors (see "Moment of truth") and its second, and its third. These last two were to find leaks in the top membrane, and were, like the first, accomplished by flooding the roof and letting the water stand for at least 24 hours.  But—how to tell if there’s a leak in the top membrane when we know there’s none in the bottom? A leak in the bottom membrane would result in water in the collection room (Oh No!), while a leak in the top membrane would only mean water between the membranes. Invisible? Yes. Undetectable? NO!

The roof was designed and built with a leak detection system specifically to pick up leaks in the top membrane BEFORE there’s a hole in the bottom membrane. 

Here’s how it works:

Between the two membranes is a screen that looks like ordinary window screening (see the screen in the photo below). A wire is attached to the screen, so that it can be electrified. Above the top membrane there are also wires—generally around the periphery, including around the drains—so that when the surface of the membrane is covered with water, and the wires are plugged in, there is a continuous electric field above the membrane. The electric fields of the screen below and the water above should be completely separate, unless there’s a hole allowing water to connect them. 

Here’s a photo of Adam (International Leak Detector) looking for leaks. The two poles are used to measure the current on the surface. Leaks cause a change in resistance, which is read on the meter around Adam’s neck. Should there be a hole, it can be pinpointed with the two poles.

The test took only about 20 minutes, as Adam walked back and forth. There was a short period of suspense when he stopped walking and poked around, but it turned out to be water running over the lip of the drain, not water running through a hole in the membrane. What a relief!     

To maintain the roof from now on, we’ll be hiring Adam’s employer now and then to re-run the test. They can do this through up to six feet (!) of material above the membrane, so long as the membrane surface is thoroughly wet. This way, we’ll never again discover a leak using the flood-the-collection-room detection system. 

Written by Dorothy Tuthill, Berry Center