Energy Policy Act of 2005 (Ultra-deepwater and Unconventional Resources Program)
Predicting Porosity and Saturations from Mud Logs and Drilling Rates Using Artificial Intelligence with Focus on a Horizontal Well
Correlations Company, Socorro, NM
Armstrong Energy Corporation
Read and Stevens, Inc.
Harvey E. Yates Company
New Mexico Bureau of Geology & Mineral Resources
The project objective is to predict porosity and saturations through a horizontal section from mud logs and drilling information obtained from drilling the lateral. Artificial intelligence consisting of fuzzy logic and neural networks will be used to correlate historic mud log and drilling information from nearby wells with known gamma ray logs, compensated neutron logs, litho density logs, and dual laterlogs. The correlations will then be used to predict porosity and saturations from mud logs and drilling information obtained from drilling a 5000-ft lateral in the Bone Springs formation. The predictions will be tested with gamma ray logs, compensated neutron logs, litho density temperature logs, and dual laterlogs run on coiled tubing through the horizontal section. The combina-tion of rate of penetration, weight on bit, and rotation is an indicator of porosity. The strength of the relationship between these measurements and known values of compensated neutron and litho density temperature logs will be evaluated using a fuzzy logic technique known as fuzzy ranking. In a similar manner, relationships will be established between the mud log data consisting of drill rate, cutting descriptions (porosity), florescence, and cut, and incorporated with gas units C1 through C5 and known dual later log values. The drilling and mud log measurements with the strongest relationships to log values of porosity and saturation will be correlated using a neural network. Optimum neural network architectures will be generated according to the number of data points available. Predictions made with the optimized neural networks will be compared to the log porosity and satura-tion values.
Due to costs many horizontal wells are drilled without conventional logging suites to guide the completion process. Successful Bone Springs zone correlations resulting from this project could be used to provide pseudo-logs, which will greatly benefit small companies operating in the Bone Springs. The methodology could certainly be applied to formations other than the Bone Springs. The potential exists for a set of hard-wired equations specific to a formation that could be incorporated into existing mud logging programs.
Key deliverables will be in the form of pseudo-porosities and -saturations through the horizontal section of a Bone Springs well. The methodology to generate these values will be placed in the public domain through presentations, technical papers, and the Internet.
Principal Investigator: William Weiss
Project Duration: 1 year