Monday, May 24, 2010

Imaging battles Gulf oil disaster

Satellite imaging and paricle image velocimetry are two of the imaging techniques being deployed against the oil spill in the Gulf of Mexico. A May 22 article in the New York Times describes several of the techniques that researchers are using to try a get an accurate measurement of the oil spill.

One approach is described in more detail by one of the Times authors, Steve Wereley at Purdue University, in a PowerPoint presentation entitled “Oil Flow Rate Analysis – Deepwater Horizons Accident”. He predicts that the Deepwater Horizon Gulf of Mexico oil spill is more than 50 times worse than initial BP predictions.

Using an imaging technique called particle image velocimetry (PIV), Wereley analyzed video obtained from BP to compute the magnitude of oil flowing from the site. According to his presentation, Wereley estimates that between 56,000 and 84,000 barrels a day are currently pouring into the Gulf of Mexico. Doug Suttle, chief operating officer for BP, initially said he thinks the estimate of 1,000 barrels a day is accurate, although BP is now admitting they have underestimated the amount of oil leaking.

To obtain his figures, Wereley computed the average plume velocity of the oil using PIV techniques, multiplied this figure by the cross-sectional area to find the volume flow rate, and then converted this figure to barrels per day.

PIV is an optical method of fluid visualization. It is used to obtain instantaneous velocity measurements and related properties in fluids. By measuring features in the fluid, motion of these features is used to calculate velocity information of the flow being studied.

A live video of the oil leak, provided by BP over Ustream is available on - search: live oil spill cam.

All this imaging doesn't even take into account the dozen or so remote underwater vehicles that are now in operation near the sea bed, around the leak, streaming video back to a control center in Houston.

The oil spill is a disaster that maybe imaging and machine vision can help understand and moderate.

No comments:

Post a Comment