--> --> Abstract: Determining petrophysical properties and gas content in the Barnett Shale using a log-based neural network solution, by Lee Utley; #90010 (2003).

Datapages, Inc.Print this page

Determining petrophysical properties and gas content in the Barnett Shale using a log-based neural network solution.


Lee Utley, Petroleum Consultant, Houston, Texas


The Barnett Shale in the Fort Worth Basin of Texas is an organic-rich black shale capable of producing large amounts of natural gas and natural gas liquids. Traditional log analysis methods have not yielded acceptable results when attempting to determine standard petrophysical properties. Therefore, log analysis alone is an impractical method of predicting production in the Barnett Shale. Production in the Barnett Shale is affected by several factors, only some of which may be measured or calculated using log data, making gas content a poor predictor of well performance. However, a neural network technique has been developed to successfully estimate reservoir potential that relies on log derived qualitative and quantitative parameters.

Log analysis in the complex lithology of the Barnett Shale is very difficult. The existence of several exotic minerals in the matrix along with significant amounts of organic material makes a algorithm-based solution virtually impossible. Using extensive core data, a neural network solution was developed to calibrate the logs to the needed petrophysical properties, and thus enable the foot-by-foot calculation of gas content of the Barnett Shale. Since any evaluation technique requires proper verification, examples will be shown to demonstrate the effectiveness of the calibration.

The logs required to perform the analysis are readily available on most wells in the Fort Worth Basin, making the solution a practical exploration/exploitation tool. Outputs from the analysis include porosity, total organic content, water saturation, lithology, and gas content, both in the sorbed and free states.

Lee Utley is an independent petrophysical consultant specializing in unconventional reservoirs. After spending 10 years with Halliburton, he worked for 5 years as a petrophysical engineer with Mitchell Energy in The Woodlands as the primary petrophysicist in the Barnett Shale and the Fort Worth Basin. He is also the Houston representative for Petroleum Software Technologies, providing technical support for their suite of resistivity and SP modeling software as well as neural network software. Lee has a Bachelors degree in Petroleum Engineering from Texas Tech University and a Masters degree in Business Administration from Southern Methodist University.

AAPG Search and Discovery Article #90010©2003 AAPG Southwest Section Meeting, Fort Worth, Texas, March 1-4, 2003