IBM and oil company Shell are collaborating on a research project to use predictive analytics to identify possible oil and gas reserves.
Predictive analytics provides a form of business intelligence that uses number crunching and statistical analysis to predict future events. This helps a business to make a decision before something happens. At Shell, the technique will be used to predict likely sites for oil reserves.
Hans Potters, manager for reservoir surveillance technologies at Shell, said, "With more sophisticated analysis, Shell sees large potential for enhancing our understanding of existing oil and gas fields. We need to blend traditional and novel types of data into new information to better evaluate the various recovery processes in complex, geological settings, and we need to do this faster than we can at the moment. IBM's strength in computer modelling is now added to Shell's subsurface domain expertise. With this collaboration, we break new ground in this area, and we expect this will increase oil and natural gas production in the future."
Modern oil and gas exploration often involves massive amounts of data collection and number crunching. Geophysicists must examine time-lapse seismic data from subsurface rock formations; reservoir engineers receive well and laboratory data, and geophysicists receive information - sound waves - covering wide spaces between the wells. It can take several months to complete and involves measurements of production volumes, flow rates and pressures, The two companies are working on building mathematical models to speed up the process.
Over the past five years IBM has spent $10bn buying business intelligence companies including Cognos and SPSS and has 200 mathematicians at its Watson labs dedicated to predictive analytics.