Abstract

Modeling Production and Prospectivity in the Austin Chalk to Optimize Well Placement, Productivity and Completion Design

Sean Boerner, Rohit Singh, and Ross Peebles
Global Geophysical Services

We identify remaining prospectivity in the Austin Chalk within an area of Wilson County, TX by integrating seismic inversion products, post-stack attributes, and publically available historical production data using multivariate statistical analysis. The resulting production prediction volume can then be used to optimize new well placement both laterally and vertically as well as allow for customized completions. By adopting these strategies the operator can avoid perforating and completing unproductive regions along the wellbore length resulting in significant cost savings with a negligible impact on production.

Using 57 post-stack seismic attributes and pre-stack inversion products, correlations can be generated between historical production metrics in the Austin Chalk and extracted seismic attribute values at the well bore. This approach typically identifies a number of attribute volumes that correlate reasonably well with production. Frequently the correlation coefficients for individual attributes would not be considered high enough to represent a sole or definitive indicator of production potential. For this reason we perform multivariate regression analysis on a portfolio of performance indicators combining multiple seismic attributes to produce production prediction volumes that are highly correlated to production. It is important to exercise caution when developing these production prediction models to minimize the number of attributes used in the model. This reduces the possibility of false correlations and of including redundant (co-linear) seismic attributes.

Using production data from 27 vertical wells, we generated several production prediction models using 4-6 variables drawn from post-stack and inversion volumes as well as two directional component angle attributes. These had correlation coefficients ranging from 0.933 to 0.963 with historical Austin Chalk production. We validated the preferred model with a blind test using 17 horizontal wells by extracting the production prediction from the model along the producing interval of the well bore. Through this method of testing we achieve a correlation coefficient of 0.86 between the predicted and actual horizontal well production.

AAPG Search and Discovery Article #90164©2013 AAPG Southwest Section Meeting, Fredericksburg, Texas, April 6-10, 2013