Computer-implemented method of using a non-transitory computer readable memory device with a pre programmed neural network and a trained neural network computer program product for obtaining a true borehole sigma and a true formation sigma

A computer-implemented method of using a non-transitory computer readable memory device with a pre-programmed neural network and a trained neural network computer program product both with three layers each, to perform the operations of obtaining a true borehole sigma and a true formation sigma, the operations of the computer-implemented method comprising:

• initializing a non-transitory computer readable memory device having a pre-programmed neural network on an above-surface processing system, wherein the above-surface processing system is coupled to a memory resource, an above-surface telemetry system, a communication bus, a multi-channel analyzer, and a computing system device;
• initializing a nuclear logging tool wireline or logging-while-drilling system having a deuterium-tritium (D-T) neutron generator or a deuterium-deuterium (D-D) neutron generator neutron source coupled with at least three dual-function radiation detectors, wherein each dual-function radiation detector is pre-programmed through a non-transitory computer-readable memory device using pulsed shape discrimination technique, a high-voltage supplier, a sub-surface telemetry system, an electronic instrument, and a sub-surface non-transitory computer readable memory device;
• pulsing the deuterium-tritium (D-T) neutron generator or the deuterium-deuterium (D-D) neutron generator neutron source, for at least two pulses;
• measuring neutrons and neutron-induced gamma rays after each of the at least two pulses from each of at least three dual-function radiation detectors;
• separating signals of the measured neutrons and neutron-induced gamma rays into thermal neutrons and neutron-induced capture gamma rays, using the at least three dual-function radiation detectors pre-programmed through the non-transitory computer-readable memory device using pulsed shape discrimination technique;
• sending the separated signals of thermal neutrons and neutron-induced capture gamma rays to the multi-channel analyzer using the sub-surface telemetry system and using the at least three dual-function radiation detectors pre-programmed through the non-transitory computer-readable memory device;
• generating a thermal neutron time-spectrum and a neutron-induced capture gamma rays time-spectrum for each of the three dual-function radiation detectors, wherein the thermal neutron time-spectrum essentially consists of a time-decay curve of thermal neutrons, and the neutron-induced capture gamma rays time-spectrum essentially consists of a time-decay curve of neutron-induced capture gamma rays, using the pre-programmed neural network on an above-surface processing system;
• computing curve fitting using two exponential decays of the time-decay curve of thermal neutrons from a time immediately after pulsing the deuterium-tritium (D-T) neutron generator or the deuterium-deuterium (D-D) neutron generator neutron source, using two exponential decays, using the pre-programmed neural network on an above-surface processing system;
• computing curve-fitting using two exponential decays of the time-decay curve of neutron-induced capture gamma rays from a time immediately after pulsing the deuterium-tritium (D-T) neutron generator or the deuterium-deuterium (D-D) neutron generator neutron source, using the pre-programmed neural network on an above-surface processing system;
• acquiring from the computed curve-fitting of the time-decay curve of thermal neutrons and from the computed curve-fitting of the time-decay curve of neutron-induced capture gamma rays, a neutron-induced apparent borehole time-decay constant, and an apparent formation time-decay constant for each of the at least three dual-function radiation detectors, using the pre-programmed neural network on an above-surface processing system;
• training a first input-layer of the non-transitory computer readable memory device having a pre-programmed neural network for using, as input, the acquired neutron-induced apparent borehole time-decay constant and the apparent formation time-decay constant;
• training three, second-processing hidden layers of the non-transitory computer readable memory device having a pre-programmed neural network with a weighted nonlinear regression algorithm and a weighted regression algorithm;
• training a third output-layer of the non-transitory computer readable memory device having a pre-programmed neural network to output an apparent borehole sigma using a weighted linear regression algorithm and to output an apparent formation sigma using a weighted regression algorithm, and then comparing said outputs of said third layer to known values of a true borehole sigma and a true formation sigma;
• repeating the operations of pulsing the deuterium-tritium (D-T) neutron generator or the deuterium-deuterium (D-D) neutron generator neutron source, measuring neutrons and neutron-induced gamma rays, separating signals of the measured neutrons and the measured neutron-induced gamma rays from each other, sending the separated thermal neutrons and neutron-induced capture gamma rays to the multi-channel analyzer, acquiring a neutron-induced apparent borehole time-decay constant and an apparent formation time-decay constant, training of the first layer, training of the three, second-processing hidden layers, and training of the third layer of the non-transitory computer readable memory device having a pre-programmed neural network, until a relative difference between an output of the third layer and the known values of a true borehole sigma and a true formation sigma are less than 1 percent;
• generating a trained neural network computer program product, having three layers, using the non-transitory computer-readable memory device having a pre-programmed neural network on an above-surface processing system;
• storing the generated trained neural network computer program product having three layers, on a memory resource on the above-surface processing system;
• replacing the pre-programmed neural network of the non-transitory computer-readable memory device on an above-surface processing system, with the stored trained neural network computer program product, having three layers;
• pulsing the deuterium-tritium (D-T) neutron generator or the deuterium-deuterium (D-D) neutron generator neutron source, for at least two pulses;
• measuring neutrons and neutron-induced gamma rays after each of the at least two pulses from each of at least three dual-function radiation detectors;
• separating signals of the measured neutrons and neutron-induced gamma rays into thermal neutrons and neutron-induced capture gamma rays, using the at least three dual-function radiation detectors pre-programmed through the non-transitory computer-readable memory device using pulsed shape discrimination technique;
• sending the separated thermal neutrons and neutron-induced capture gamma rays to the multi-channel analyzer using the sub-surface telemetry system and using the at least three dual-function radiation detectors pre-programmed through the non-transitory computer-readable memory device;
• generating a thermal neutron time-spectrum and a neutron-induced capture gamma rays time-spectrum for each of the three dual-function radiation detectors, wherein the thermal neutron time-spectrum essentially consists of a time-decay curve of thermal neutrons, and the neutron-induced capture gamma rays time-spectrum essentially consists of a time-decay curve of neutron-induced capture gamma rays, using the pre-programmed neural network on an above-surface processing system;
• computing curve fitting using two exponential decays of the time-decay curve of thermal neutrons from a time immediately after pulsing the deuterium-tritium (D-T) neutron generator or the deuterium-deuterium (D-D) neutron generator neutron source, using two exponential decays, using the non-transitory computer readable memory device having the trained neural network computer program product;
• computing curve-fitting using two exponential decays of the time-decay curve of neutron-induced capture gamma rays from a time immediately after pulsing the deuterium-tritium (D-T) neutron generator or the deuterium-deuterium (D-D) neutron generator neutron source, using the non-transitory computer readable memory device having the trained neural network computer program product;
• acquiring from the computed curve-fitting of the time-decay curve of thermal neutrons and from the computed curve-fitting of the time-decay curve of neutron-induced capture gamma rays, a neutron-induced apparent borehole time-decay constant, and an apparent formation time-decay constant for each of the at least three dual-function radiation detectors, using the non-transitory computer readable memory device having the trained neural network computer program product;
• inputting into a first input-layer of the non-transitory computer readable memory device having the trained neural network computer program product, the acquired neutron-induced apparent borehole time-decay constant and the acquired apparent formation time-decay constant for each of the at least three dual-function radiation detectors;
• processing in one of the three, second-processing hidden layers of the non-transitory computer readable memory device having the trained neural network computer program product, the acquired neutron-induced apparent borehole time-decay constant and the acquired apparent formation time-decay constant for each of the at least three dual-function radiation detectors using the weighted nonlinear regression algorithm;
• processing in a second of the three, second-processing hidden layers of the non-transitory computer readable memory device having the trained neural network computer program product, the acquired neutron-induced apparent borehole time-decay constant and the acquired apparent formation time-decay constant for each of the at least three dual-function radiation detectors using the weighted nonlinear regression algorithm;
• processing in a third of the three, second-processing hidden layers of the non-transitory computer readable memory device having the trained neural network computer program product the acquired neutron-induced apparent borehole time-decay constant and the acquired apparent formation time-decay constant for each of the at least three dual-function radiation detectors using the weighted linear regression algorithm;
• and
• computing in a third output-layer of the non-transitory computer readable memory device having the trained neural network computer program product, a true borehole sigma and a true formation sigma using the weighted linear regression algorithms;
• generating a true borehole sigma and a true formation sigma using the weighted linear regression algorithms of the trained third output-layer of the non-transitory computer readable memory device having the trained neural network computer program product.