There are several drilling challenges when operating in the deep-water Gulf of Mexico and in many other basins globally. One of the significant challenges related to drilling safe wells is to produce accurate pre-drill pore pressure estimates.
Pore pressure prediction in compacting shales is quantified by relating the porosity, using wireline data as a proxy, to the vertical effective stress. However, the pore pressure predicted in shales may be in error because of lithological variation in both the clay content and organic content, or due to temperature driven changes in the grain-to-grain contact relationship.
For example, elevated amounts of organic matter in thermally immature shales causes significant overprediction of pore pressure while temperature can have a range of effects on the rock properties such as cementation, dissolution and collapse of the rock structure and mineral transformation (e.g. smectite to illite). These all challenge the assumption that porosity can be directly related to pore pressure, which is the presumption of industry-standard pore pressure prediction techniques such as Eaton or Equivalent Depth method.
The solution to these challenges are to a) implement bulk density and sonic velocity log corrections to account for lithology variations (e.g. TOC) prior to performing any pore pressure prediction and b) to understand whether high temperatures have modified the original (shale) rock framework compromising the original effective stress-porosity relationship in the rock.
In this talk we will give an example of how to account for high TOC and how the pore pressure prediction is dramatically improved by using combination of rock physics models relating bulk density to total organic carbon (TOC), and P-wave velocity to both clay volume and bulk density in organic-rich and conventional shales, respectively.
We will also give examples on how geological pressure models may give and estimate of pore pressure at high temperature conditions where industry-standard techniques are challenged/may not work.