Gas-plume imager scans refineries in field testing

A gas-plume imager developed by Sandia has left the lab and entered the field in the past year, undergoing refinery field tests from California to England.

The tests suggest that imaging technology is a viable alternative to hand-held sniffers currently used to detect leaks, says principal investigator Tom Kulp (8356). The Environmental Protection Agency is examining permitting use of the shoulder-carried imager as an alternate work practice for mandatory monitoring of refinery piping and valves. The petroleum industry sees gas imaging as a time- and cost-efficient alternative.

The imager (which resembles a large camcorder) works using the backscatter absorption gas imager (BAGI) method, in which a scene suspected of containing a leak is illuminated with infrared laser light. The laser light is reflected by the background of the scene and absorbed by the leaking gas. A camera in the device gives a visual display of the scene with the plume appearing as a darker region. The device was created by Tom and colleagues Karla Armstrong (8356), Dahv Kliner (8356), Ricky Sommers (8356), and Sal Birtola (8350).

With support from DOE’s Office of Fossil Energy and Energy Efficiency, the technology has been under development several years as a way to find leaks more quickly, to help protect the environment, and make monitoring refineries more efficient.

"Our biggest contribution has been to prove that a gas imager can be reduced to a format useable by a single operator under battery power," Tom says. "This has been accomplished using new nonlinear optical and fiber laser technologies."

The field trials started with a day at a refinery in nearby Martinez in January 2003, where the team established a method for carrying the imager around the piping and double-checked its performance. Then a longer field trial followed for four days in February in Beaumont, Texas. There, committee members from the American Petroleum Institute’s Smart Leak Detection and Repair Project organized tests observed by the EPA, oil companies, and potential commercialization partners. The tests compared the standard existing leak detection method using sniffers, called Method 21, and potential alternative technologies like BAGI. Also, leaks were entirely encapsulated and the flow rate measured.

Tom says the imager spotted 41 leaks and missed three whose flow rate was below 40 grams/hour, the current performance target stated by EPA. The EPA currently regulates leaks based on concentration, but is interested in detecting the presence of something with a particular flow rate.

"Our system images the plume so you get a visual sense of the flow rate," he says. If the approach is proven equivalent to existing technology, it could be approved for monitoring refinery leaks in addition to the already-approved Method 21. "The idea that there is a visual record is also useful," he says.

In August, the team traveled to a Chicago suburb for tests, funded by the state of Texas, that compared the laser imager to so-called "passive" imagers that sense thermal differences between the gas and its background without illuminating the scene with laser light.

Shortly after that, they traveled to Southampton, England, to show the technology to regulators and refinery representatives from England, Belgium, and the Netherlands. Jeff Siegell, chair of the Smart Leak Detection and Repair Project, organized the demonstration so European industry and government officials would be aware of other methods to potentially enforce fugitive emission laws.

Coming up this winter is another large Texas test, sponsored by Shell Oil, in which the technology is being evaluated for its usefulness in a refinery setting over six months.

Meanwhile, Sandia has been interacting with companies that would manufacture a commercial version of the device if it’s licensed and marketed. "I think there’s going to be a lot of development," Tom says. Even since the imager’s development began, cameras and lasers have advanced, he notes, and they are continuing to do so. "In five years, the technologies used will be completely different."