UC Davis researchers are developing a method to detect leaks by flying over long stretches of hard-to-reach natural gas pipelines.
The system uses sensitive instruments and specially-designed software, and it could prove more efficient and more cost-effective than current methods.
On a recent cool, sunny morning at the Lincoln regional airport just north of Sacramento, UC Davis atmospheric scientist Stephen Conley demonstrated just how it works.
Sitting in his small two-seater plane, Conley opens a laptop computer that's hooked up to the measurement instruments and a GPS locator synched to a map of the pipeline. He points to yellow and red lines on the screen:
"So there's the pipeline, here's where we are. We want to fly over there," Conley says.
The computer screen also displays a box with a real-time spiky line graph showing methane concentrations in the air. The line is actively moving up and down.
"Basically, what we're doing is if we fly over an area where's there's a leak, we're going to see a spike in methane."
Intake tubes underneath the plane's wing are constantly sampling the air, and a special greenhouse gas instrument measures the concentration of certain gases.
But it's not as simple as flying around looking for methane "plumes." Conley and his research partner have written complex software that takes wind into account.
"We use all these studies of the boundary layer to figure out, given the winds and the vertical mixing of turbulence, where should the plane be, how far downwind, to give this plume time to rise to our altitude."
This calculation lets the pilot steer into the exact position where if there were a leak, the very sensitive equipment would pick it up.
Conley eases the plane into the air and heads west toward a PG&E pipeline that runs along the coastal range.
The methane measurement graph bounces up and down as the plane sails over some neatly groomed agricultural fields and ponds -- but when it reaches the pipeline the chart levels off. No spikes today.
But at this point that's not a guarantee there are no leaks.
Conley's UC Davis colleague Ian Faloona says they're tweaking the program's algorithm to account for how wind really behaves.
"So there's a certain element, that we don't know what the wind is exactly doing, but we want to improve our models of that behavior based on these measurements that we make."
The $200,000 research project is largely funded by PG&E.
Conley says he and his colleagues will continue to refine the software, and he hopes the open-source program will eventually be picked up and further improved.