Geochemical logging is essential for identifying formation lithology, and its quantitative analysis of the composition of formation elements relies heavily on accurate response gamma spectra. Measuring the response spectra by experiment is complicated, but using the Monte Carlo method for simulations is a feasible approach. We have developed and verified a high-precision technique to establish a comprehensive and highly accurate database of response spectra for geochemical logging tools with isotope neutron sources.

We established a large cylindrical formation geometry using the MCNP6.1 code, including a geochemical logging tool with a Φ7.62 × 15 cm bismuth germanate detector and an Am-Be neutron source. A suitable neutron-induced gamma-ray nuclear data library was selected, and the method for obtaining the formation gamma-ray and capture response spectra was enhanced. The constructed database was then evaluated by analyzing the spectra of actual formations after ensuring that the simulated spectra were consistent with the measured response spectra and characteristic gamma rays.

We applied the weighted least-squares method to predict the element yields, achieving a maximum mean absolute error compared to the standard yields of less than 0.05. Meanwhile, the reconstructed spectra showed good agreement with the measured spectra, indicating the high accuracy of the spectra database generated by this technology. This can provide assistance for geochemical logging to determine the composition of formation elements, providing an essential foundation for geochemical logging’s application in oil exploration and uranium development.