Reported by Barry Morgan from Accra, Ghana (West Africa)
for Upstream (January 27, 1997, Oslo, Norway)
OIL COMPANIES are now pouring money into massively parallel computers --- proof, says Emeagwali, that the technology has broken through initial scepticism.
His prize-winning work in 1989 enabled him to develop the first "pseudo-time" approach to reservoir modelling. It proved to the US establishment that "the use of only Dirichlet-type equations in the vicinity of petroleum production wells located near the boundary is better suited to avoid the coning problems caused by the high velocity of converging flows in the vicinity of wells".
The work goes to the heart of a debated that raged in
the 1930s over equations governing porous media
flow modelling, and which this doughty scientist now sees
good reason to revive.
Uprooting old models will, he says, be costly and difficult. "Reservoir models have taken hundreds of man years and tens of millions of dollars to construct and it will be impractical to discard. But they can be modified to incorporate the additional physics used by my new equations." He does not feel his ideas have been under-utilised by the industry and points out that his original goal was simply to demonstrate that massively parallel computers can be used to discover and recover additional oil and gas.
The next challenge, he says, will require a mix of technologies with faster computers that can perform one trillion calculations per second --- a feat, Emeagwali predicts, that will be achieved in less than two years.