Diffusion-Corrected Sigma Using Dual Exponential Fit for Time-Lapse Reservoir Monitoring


Shaikha Al-Turkey (Kuwait Oil Company) | Daniel Freile (Weatherford) | Larisa Tagarieva (Weatherford) | Mohamed Elyas (Weatherford) | Gregory Schmid (Weatherford)


IPTC - International Petroleum Technology Conference

Publication Date

January 13, 2020


International Petroleum Technology Conference, 13-15 January, Dhahran, Kingdom of Saudi Arabia

Paper ID



Pulsed-neutron capture (PNC) logs are commonly used to determine formation water saturation in cased-hole environments, often for time-lapse monitoring purposes. This paper describes a new diffusion-corrected sigma algorithm developed for a pulse neutron logging tool.

In southeast Kuwait, diffusion-corrected sigma log data was recorded in three wells using an array of four optimally spaced gamma ray detectors above a neutron generator. To calculate a diffusion-corrected sigma, an algorithm based on a dual exponential fit was applied to the time-decay spectrum of the near and far detectors. This calculation separates the formation and borehole decays. This approach provides an apparent formation sigma for the near and far detectors. The algorithm uses the near detector for final sigma, and a diffusion correction to the near sigma is determined by a function of a near-far sigma difference.

The diffusion-corrected sigma matched the expected results and provided a good statistical quality—even at high sigma values—because it is based on the near detector with its higher count rate, as demonstrated in the examples presented. Also, the formation sigma was independent of different borehole conditions in which the data was recorded. The final formation sigma results were compared to volumetric results from open-hole data (volume of shale, effective porosity and water saturation) and sigma calculated from open-hole volumetric using material balance. The PNC data recorded in the three wells allowed determination of the most recent oil-water contact (OWC) and update of water encroachment maps from the time-lapse monitoring. Comparing with previous sigma data recorded in these wells, it was concluded a normalization transform is not needed because R2 value of the linear regression is close to 1.

The diffusion-corrected sigma algorithm using dual exponential fit showed that this technique was able to extract independent values for borehole sigma and formation sigma for each detector and to perform an accurate diffusion correction. This algorithm will provide reliable sigma values regardless of the borehole conditions in which the data was recorded.