Using 3D Seismic Attributes to Investigate Mud Losses in a Fluvial-Deltaic Sand: An Example From the Oliver Sandstone

Abstract

Drilling operations in an actively developed basin in Wyoming can experience large quantities of mud losses in a shallow zone along the intermediate (vertical) section of the wellbore. These losses can negatively impact well economics due to the unforeseen expenditures that come with the price of the mud itself and the extended rig time it takes to overcome the mud loss issues. A typical well in the area costs approximately 8 million dollars, of which, drilling cost accounts for 40-50%. Of the roughly $4 million associated with the drilling cost, rig time makes up ~20%, and rig rates range from $18,000 - $25,000 a day. It is common for severe losses to take multiple days of additional rig time and even minor mud loss issues can add ~12 - 24 hours. Since the depositional nature of fluvial deltaic sands produces moderate to high degrees of grain sorting which translate to moderate or high values of porosity and permeability, I hypothesized that regions in the Oliver sandstone exhibiting porosity and permeability may correlate with localized areas of mud loss. Unfortunately, many of the lost circulation wells do not have logs within the Oliver, prohibiting a direct mud loss versus log porosity correlation. In contrast, all of the wells are contained within a seismic data survey, providing an alternative correlation workflow. I computed a suite of seismic attributes which reveal subtle depositional and structural information about the area. There was no significant correlation of mud loss to structural attributes (e.g. similarity, curvature, and dip magnitude). The same suite of seismic attributes were analyzed using an Artificial Neural Network (ANN) algorithm to predict estimates of porosity within the region of interest. Although the resultant volume delineated laterally variable zones of low and high porosity in the Oliver Sandstone, the correlations between mud loss volumes and porosity were very low, concluding that porosity does not have a direct influence in the lost circulation.