A sequential dynamic Bayesian network for pore pressure estimation with uncertainty quantification.

Pore-pressure estimation is an important part of oil-well drilling, since drilling into unexpected highly pressured fluids can be costly and dangerous. However, standard estimation methods rarely account for the many sources of uncertainty, or for the multivariate nature of the system. We propose the pore pressure sequential dynamic Bayesian network (PP SDBN) as an appropriate solution to both these issues. The PP SDBN models the relationships between quantities in the pore pressure system, such as pressures, porosity, lithology and wireline log data, using conditional probability distributions based on geophysical relationships to capture our uncertainty about these variables and the relationships between them. When wireline log data is given to the PP SDBN, the probability distributions are updated, providing an estimate of pore pressure along with a probabilistic measure of uncertainty that reflects the data acquired and our understanding of the system. This is the advantage of a Bayesian approach. Our...