Quantitative analysis of neutron captured γ spectra using a sequential quadratic programming method

High-precision, quantitative analysis is a key element in ensuring that neutron captured γ spectral logging can be used to accurately distinguish rock minerals and rock types from complex reservoirs. We solve a nonlinear optimization problem by implementing a sequential quadratic programming method into a quantitative spectral analysis. To quantitatively test our method, we build three kinds of formation models with different porosities: a sandstone model, a mixed model of sandstone and limestone, and mixed model of sandstone, limestone, and ferric oxide. Combining results from Monte Carlo simulations, we analyze quantitatively the mixed spectra of the models using a sequential quadratic programming method and compare the results to those obtained with the weighted least-squares method. The results indicate that the sequential quadratic programming method can ensure high-precision, quantitative spectral analysis and that its spectral analysis precision is slightly higher than that of the weighted least-squares method. The sequential quadratic programming method has higher ability of adapting to complex lithology formation and can be used to analyze quantitatively the neutron γ spectrum in complex lithology formations.