Natural Gamma Ray Tool Response Discrepancies: Not Always Due to the Calibration!

Abstract

It is not uncommon to note discrepancies in natural gamma ray (NGR) data from different designs of gamma ray (GR) tools run in the same well. These discrepancies are often blamed on inaccurate primary or secondary calibrations of the tools; however, it can be shown that in many cases, the different readings are inherent in the NGR measurement itself. Because the primary calibration is performed at the University of Houston (UH) in a reference formation with a specific mixture of thorium, uranium, and potassium, a change in the relative elemental concentrations results in a change of the energy spectrum of the gamma rays that reach the tool’s detector. The tool response to the different energies depends on a large number of factors; e.g., thickness and type of material traversed by the gamma rays, dimensions and type of the GR detector, as well as differences in threshold settings of the acquisition electronics. The standard environmental corrections for NGR tools correct for borehole effects (mud weight and composition, borehole size, and tool position in the borehole), so that the tool response to a given formation is independent of the borehole environment. In this paper, we present a detailed Monte-Carlo modeling study quantifying the impact of the tool and detector design on the sensitivity to the three predominant naturally occurring radioactive materials (Th, U, and K). Also, the modeling study includes the impact the thickness and type of the materials have that is traversed by the gamma rays before reaching the detector. The modeling is benchmarked against the measured response in formations with a known Th, U, and K concentration, such as the UH Th and U pits, or internal calibration facilities. The modeling results show that there can be significant discrepancies in the response of different tools to the same formation if the relative Th, U, and K concentrations deviate considerably from those in the UH reference pit. This condition will be illustrated by field cases with comparisons of the total NGR measurement and the spectral NGR measurement, which show that tool-to-tool differences can be attributed to a change of the composite energy spectrum of the emitters from the formation. Finally, it can be shown that the tool calibration, as well as the interpretation of tool discrepancies, can be improved by using not only a 200-gAPI reference but also an independent Th, U, and K calibration associated with specific response functions for each tool design.