Examples described herein provide a computer-implemented method that includes analyzing a first dataset by applying the first dataset to a first model to generate a first result. The method further includes analyzing a second dataset by applying the second dataset to a second model to generate a second result. The method further includes performing validation on the first model and the second model by comparing the first result to the second result. The method further includes, responsive to determining that the first result and the second result match, modifying an operational action of a surface assembly based on at least one of the first result or the second result. The method further includes, responsive to determining that the first result and the second result do not match, updating at least one of the first model or the second model.

A system and method for uncertainty estimation of reservoir parameters in unconventional reservoirs using a physics-guided convolutional neural network to generate a plurality of reservoir models, a data analysis step, and an uncertainty step is disclosed. The method is a computationally efficient method to estimate uncertainties in models of unconventional reservoirs.

A method, apparatus and system is provided for assessing risk for well completion, comprising: obtaining, using an input interface, a Below Rotary Table hours and a plurality of well-field parameters for one or more planned runs, determining, using at least one processor, one or more non-productive time values that correspond to the one or more planned runs based upon the well-field parameters, developing, using at least one processor, a non-productive time distribution and a Below Rotary Table distribution via one or more Monte Carlo trials; and outputting, using a graphic display, a risk transfer model results based on a total BRT hours from the Below Rotary Table and the non-productive time distribution produced from the one or more Monte Carlo trials.

The disclosure presents processes to improve the calibration of adjustable choke valves corresponding to a specific size of positive choke bean. Typically, manufacturers specify a position of the adjustable choke valve that corresponds to a specific choke bean size. Hydrocarbon fluid conditions and composition vary and subterranean formation characteristics vary which can lead to errors in the calibration. By comparing flow rate parameters of the hydrocarbon fluid flowing through the adjustable choke manifold and the positive choke manifold, errors in calibration can be detected and corrected. The factors involved with the hydrocarbon fluid and the error correction can be used to update a choke model. The choke model can then be used for future calibrations of the adjustable choke valve.

An estimated gas leak flow rate can be determined using a teaching set of concentration profiles, a regression model implemented by a machine-learning subsystem, and a subset of attributes measured within an environment. The teaching set of concentration profiles can include gas flow rates associated with relevant attributes. The regression model can be transformed into a gas leak flow regression model via the machine-learning subsystem using the teaching set. The subset of attributes measured within the environment can be applied to the gas leak flow regression model to determine other attributes absent from the subset of attributes and an estimated gas flow rate for the environment. A gas leak attenuation action can be performed in response to the estimated gas flow rate.

A method for seismic processing includes extracting, using a first machine learning model, one or more seismic features from seismic data representing a subsurface domain, receiving one or more well logs representing one or more subsurface properties in the subsurface domain, and predicting, using a second machine learning model, the one or more subsurface properties in the subsurface domain at a location that does not correspond to an existing well based on the seismic data, the one or more well logs, and the one or more seismic features that were extracted from the seismic data.

The disclosure presents processes that utilize collected resistivity data, for example, from an ultra-deep resistivity tool located downhole a borehole. In some aspects, each slice of resistivity data can generate multiple distribution curves that can be overlaid offset resistivity logs. In some aspects, an analysis can be performed to identify trends in the distribution curves that can be used to identify approximate locations of subterranean formation surfaces, shoulder beds, obstacles, proximate boreholes, and other borehole and geological characteristics. As the number of distribution curves generated increase, the confidence in the analysis also increases. In some aspects, the number of distribution curves can be twenty, one hundred, one hundred and one, or other counts of distribution curves. In some aspects, the resistivity data can be used to generate two or more synchronized view perspectives of a specific location along the borehole, where each view perspective uses the same focus area.

Systems and methods include a computer-implemented method for determining an integrated surface-downhole integrity score for a natural flow oil well. Wellness surface parameters of a natural flow oil well are determined. Wellness downhole parameters for the natural flow oil well are determined, including parameters indicating well integrity and tubing/casing conditions. An integrated surface-downhole integrity score is determined using the wellness surface parameters and the wellness downhole parameters. An alert is provided for presentation to an operator in a user interface. The alert is provided in response to the integrated surface-downhole integrity score exceeding a threshold.

This disclosure describes systems, methods, and apparatuses for preventing wireline sticking during hydraulic fracturing operations, the system comprising: a sensor coupled to a fracking wellhead, circulating fluid line, or standpipe of a well and configured to convert acoustic vibrations measured in fracking fluid in the wellhead, fluid line, or standpipe into an electrical signal in a time domain; a memory configured to store the electrical signal; a converter configured to access the electrical signal from the memory and convert the time domain electrical signal into a frequency domain spectrum; a machine-learning system configured to classify the current frequency domain spectrum as associated with increasing wireline friction, the machine-learning system trained on previous frequency domain spectra measured during previous wireline operations and previously classified by the machine-learning system; and a user interface configured to return an indication of the increasing wireline friction to an operator of the hydraulic fracturing operations.

Aspects of the present disclosure relate to earth modeling using machine learning. A method includes receiving detected data at a first depth point along a wellbore, providing at least a first subset of the detected data as first input values to a machine learning model, and receiving first output values from the machine learning model based on the first input values. The method includes receiving additional detected data at a second depth point along the wellbore, providing at least a second subset of the additional detected data as second input values to the machine learning model, and receiving second output values from the machine learning model based on the second input values. The method includes combining the first output values at the first depth point and the second output values at the second depth point to generate an updated model of the wellbore, the updated model comprising an earth model.