Abstract: A neural network system includes a first neural network configured to predict a mean value output and epistemic uncertainty of the output given input data, and a second neural network configured to predict total uncertainty of the output of the first neural network. The second neural network is trained to predict total uncertainty of the output of the first neural network given the input data through a training process involving minimizing a cost function that involves differences between a predicted mean value of a geophysical property of a geological formation from the first neural network and a ground-truth value of the geophysical property of the geological formation. The neural network system further includes one or more processors configured to run a software module that determines aleatoric uncertainty of the output of the first neural network based on the epistemic uncertainty of the output and the total uncertainty of the output.

Abstract: A method and a non-transitory computer readable medium for performing a calculation in a neural network comprise: accepting a data set into the neural network; performing a calculations with the neural network using the data set, wherein the calculations use a loss function and provide an aleatoric and epistemic uncertainty that is correlated to a value; and displaying results of the calculations performed.

Abstract: Systems and methods for monitoring a crude oil blending process use nuclear magnetic resonance (NMR) sensors which investigate properties of a plurality of crude oil streams that are mixed together to form a crude oil blend. An NMR sensor is also used to investigate the properties of the crude oil blend. The investigated properties may include viscosity. Resulting determinations may be used to control the input streams so that the output stream meets desired criteria. Additional sensors such as spectroscopy sensors, viscometers, and densitometers may be used in conjunction with the NMR sensors.

Abstract: Downhole properties of a geological formation may be determined using nuclear magnetic resonance (NMR) measurements obtained by a moving tool. To do so, an interpretation of the NMR data obtained by the moving data may take into account a moving model, characterization, or calibration of the downhole NMR tool. Additionally or alternatively, a partial interpretation mask may exclude interpretation of certain areas of data (e.g., T1-T2 data points or diffusion-T2 data points) that are expected to be less likely to describe downhole materials of interest.

Abstract: Methods and systems are provided that combine NMR and IR spectroscopy measurements on a rock sample to determine data representing at least one property of the rock sample. In one embodiment, cuttings can be split into first and second lots. Results of an NMR measurement performed on the first lot of cuttings without cleaning can be analyzed to determine pore volume of the cuttings. Results of an IR spectroscopy measurement performed on the second lot of cuttings after solvent cleaning can be analyzed to determine matrix density of the cuttings. Porosity can be determined from the pore volume and matrix density of the cuttings. In another embodiment, combined NMR and IR spectroscopy measurements can be performed on an unprepared rock sample (without solvent cleaning) to characterize properties of kerogen in the rock sample and porosity. In another aspect, a method is provided that employs multi-nucleic NMR measurements to determine porosity.

Abstract: Methods and systems are provided that combine NMR and IR spectroscopy measurements on a rock sample to determine data representing at least one property of the rock sample. In one embodiment, cuttings can be split into first and second lots. Results of an NMR measurement performed on the first lot of cuttings without cleaning can be analyzed to determine pore volume of the cuttings. Results of an IR spectroscopy measurement performed on the second lot of cuttings after solvent cleaning can be analyzed to determine matrix density of the cuttings. Porosity can be determined from the pore volume and matrix density of the cuttings. In another embodiment, combined NMR and IR spectroscopy measurements can be performed on an unprepared rock sample (without solvent cleaning) to characterize properties of kerogen in the rock sample and porosity. In another aspect, a method is provided that employs multi-nucleic NMR measurements to determine porosity.

Abstract: Downhole fluid volumes of a geological formation may be estimated using nuclear magnetic resonance (NMR) measurements, even in organic shale reservoirs. Multi-dimensional NMR measurements, such as two-dimensional NMR measurements and/or, in some cases, one or more well-logging measurements relating to total organic carbon may be used to estimate downhole fluid volumes of hydrocarbons such as bitumen, light hydrocarbon, kerogen, and/or water. Having identified the fluid volumes in this manner or any other suitable manner from the NMR measurements, a reservoir producibility index (RPI) may be generated. The downhole fluid volumes and/or the RPI may be output on a well log to enable an operator to make operational and strategic decisions for well production.

Abstract: Nuclear magnetic resonance (NMR) relaxation and/or diffusion measurements are used to deduce fluid compositional information such as a chain-length distribution, which may then be used to predict the true boiling points (TBP) of a sample of a complex hydrocarbon fluid mixture, such as a crude oil. The NMR measurements may be considered a fast and portable proxy measurement in estimating fluid TBP distributions in lieu of distillation methods, or the simulated distillation by gas chromatography.

Abstract: Embodiments herein include a method for characterizing a rock formation sample. The method for characterizing a rock formation sample includes obtaining a plurality of data sets characterizing the rock formation sample. The method further includes training a neural network to generate a computational model. Moreover, the method additionally includes using the plurality of data sets as input to the computational model, wherein the computational model may be implemented by a processor that derives an estimate of permeability of the rock formation sample.

Abstract: Methods and systems for determining epistemic and aleatoric uncertainties using a dual neural network architecture are described.

Abstract: Nuclear magnetic resonance (NMR) relaxation and/or diffusion measurements are used to deduce fluid compositional information such as a chain-length distribution, which may then be used to predict the true boiling points (TBP) of a sample of a complex hydrocarbon fluid mixture, such as a crude oil. The NMR measurements may be considered a fast and portable proxy measurement in estimating fluid TBP distributions in lieu of distillation methods, or the simulated distillation by gas chromatography.

Abstract: Nuclear magnetic resonance (NMR) methods and apparatus are provided for investigating a sample utilizing NMR pulse sequences. In various embodiments, the NMR pulse sequences have a solid state portion and a line-narrowing portion. In other embodiments, the NMR pulse sequences have a first line-narrowing portion and a second line-narrowing portion where the sequences of the different portions are different. In yet other embodiments, the NMR pulse sequences have a T1 portion and a line-narrowing portion. Processing of detected signals permits determination of characteristics of the sample including, in some cases, a differentiation of multiple components of the sample.

Abstract: Methods and systems are provided that combine NMR and IR spectroscopy measurements on a rock sample to determine data representing at least one property of the rock sample. In one embodiment, cuttings can be split into first and second lots. Results of an NMR measurement performed on the first lot of cuttings without cleaning can be analyzed to determine pore volume of the cuttings. Results of an IR spectroscopy measurement performed on the second lot of cuttings after solvent cleaning can be analyzed to determine matrix density of the cuttings. Porosity can be determined from the pore volume and matrix density of the cuttings. In another embodiment, combined NMR and IR spectroscopy measurements can be performed on an unprepared rock sample (without solvent cleaning) to characterize properties of kerogen in the rock sample and porosity. In another aspect, a method is provided that employs multi-nucleic NMR measurements to determine porosity.

Abstract: Systems and methods for monitoring a crude oil blending process use nuclear magnetic resonance (NMR) sensors which investigate properties of a plurality of crude oil streams that are mixed together to form a crude oil blend. An NMR sensor is also used to investigate the properties of the crude oil blend. The investigated properties may include viscosity. Resulting determinations may be used to control the input streams so that the output stream meets desired criteria. Additional sensors such as spectroscopy sensors, viscometers, and densitometers may be used in conjunction with the NMR sensors.

Abstract: Nuclear magnetic resonance (NMR) methods and apparatus are provided for investigating a sample utilizing NMR pulse sequences having solid state and CPMG pulse sequence portions. Various embodiments of solid state pulse sequences may be utilized including two-dimensional (repetitive) line-narrowing sequences. The hydrogen content of a solid portion of the sample may be determined by using one or more echoes resulting from the solid state sequence portion of the pulse sequence to establish a total organic hydrogen content of the sample, and by using a CPMG echo train to establish a fluid organic hydrogen content, and by subtracting one from the other to obtain the hydrogen content of the sample's solid portion. Additionally, or alternatively, the T2 values obtained from the line-narrowing and CPMG pulse sequences can be compared by plotting to obtain information regarding a characteristic of the sample. The NMR pulse sequence may also include a T1 portion.

Abstract: Methods for improved interpretation of NMR data acquired from industrial samples by simultaneously detecting more than one resonant nucleus without removing the sample from the sensitive volume of the NMR magnet or radio frequency probe are disclosed. In other aspects, the present disclosure provides methods for robust imaging/analysis of spatial distribution of different fluids (e.g., 1H, 23Na, 19F) within a core or reservoir rock. NMR data may be interpreted in real-time during dynamic processes to enable rapid screening, e.g. of enhanced oil recovery techniques and products and/or to provide improved interpretation of well-logs. Measurements of resonant nuclei other than 1H may be performed in the laboratory or downhole with a NMR logging tool. In other aspects, the present disclosure describes a novel kernel function to extract values for underlying parameters that define relaxation time behavior of a quadrupolar nucleus.

Abstract: A computer-implemented method for reservoir volumetric estimation, a non-transitory computer-readable medium, and a computing system. The method may include running a molecular dynamics simulation of a fluid-rock model of a first reservoir system at a plurality of pressures. The fluid-rock model includes a fluid that is at least partially adsorbed in the first reservoir system at one or more pressures of the plurality of pressures. The method may also include calculating a plurality of isothermal density profiles of the fluid in the first reservoir system, in association with the plurality of pressures using a result of the molecular dynamics simulation. The method may further include determining a first gas accumulation of the fluid in the first reservoir system for the plurality of isothermal density profiles. The first gas accumulation is at least partially a function of a pore surface area of a sample of the first reservoir system.

Abstract: Downhole properties of a geological formation may be determined using nuclear magnetic resonance (NMR) measurements obtained by a moving tool. To do so, an interpretation of the NMR data obtained by the moving data may take into account a moving model, characterization, or calibration of the downhole NMR tool. Additionally or alternatively, a partial interpretation mask may exclude interpretation of certain areas of data (e.g., T1-T2 data points or diffusion-T2 data points) that are expected to be less likely to describe downhole materials of interest.

Abstract: Downhole fluid volumes of a geological formation may be estimated using nuclear magnetic resonance (NMR) measurements, even in organic shale reservoirs. Multi-dimensional NMR measurements, such as two-dimensional NMR measurements and/or, in some cases, one or more well-logging measurements relating to total organic carbon may be used to estimate downhole fluid volumes of hydrocarbons such as bitumen, light hydrocarbon, kerogen, and/or water. Having identified the fluid volumes in this manner or any other suitable manner from the NMR measurements, a reservoir producibility index (RPI) may be generated. The downhole fluid volumes and/or the RPI may be output on a well log to enable an operator to make operational and strategic decisions for well production.

Abstract: A method for testing an unconventional core sample is provided. The method involves loading the unconventional core sample into a sample holder and introducing fluid into the sample holder at an elevated pressure such that fluid is injected into the internal pore space of the unconventional core sample in order to resaturate the unconventional core sample with the fluid, wherein the fluid is selected from the group including a hydrocarbon fluid and a water-based formation fluid. An apparatus and a system used in combination with the method are also provided.