A physics-informed attention-based neural network (PIANN) system, wherein the PIANN system is a computer system configured to implement a PIANN, the computer system comprising at least one processor and memory storing computer instructions, wherein, when the at least one processor executes the computer instructions, the PIANN system is trained to learn a solution or model for a partial differential equation (PDE) respecting one or more physical constraints, and wherein the PIANN includes a physics-informed neural network (PINN) implementing a deep neural network and a transition zone detector. According to at least some implementations, the PIANN implements a recurrent neural network (RNN).

A method and apparatus for forecasting oil production from an oil well in a geological formation includes receiving a plurality of sets of predicted geological data, for each of the plurality of sets of predicted geological data, determining a probability for the predicted geological data of the formation, iteratively selecting one of the plurality of sets of predicted geological data using Monte Carlo sampling based on the determined probabilities, assigning the selected set of predicted geological data to a cluster of historical data, and for each set of historical data of the cluster generating a predicted oil production rate as a function of time utilizing a machine learning based oil model, generating, based on the predicted oil production rates, a forecasted oil production rate, determining, based on the forecasted oil production rate, a preferred operating parameter for the well, and operating based on the preferred operating parameter.

The disclosure presents solutions for determining a casing wear parameter. Image collecting or capturing devices can be used to capture visual frames of a section of drilling pipe during a trip out operation. The visual frames can be oriented to how the drilling pipe was oriented within the borehole during a drilling operation. The visual frames can be analyzed for wear, e.g., surface changes, of the drilling pipe. The surface changes can be classified as to the type, depth, volume, length, shape, and other characteristics. The section of drilling pipe can be correlated to a depth range where the drilling pipe was located during drilling operations. The surface changes, with the depth range, can be correlated to an estimated casing wear to generate the casing wear parameter. An analysis of multiple sections of drilling pipe can be used to improve the locating of sections of casing where wear is likely.

The disclosure presents processes and methods for determining an overpull force for a stuck drill string in a borehole system. The fluid composition of a mud in the borehole at a specified depth can be broken down into a percentage of liquid and percentage of solids, as well as adjusting for material sag and settling factors. The fluid composition can be utilized to identify friction factors and drag in respective fluid composition zones. Each friction factor and drag can be summed to determine a total fluid drag on the drill string. In some aspects, the total fluid drag can be adjusted utilizing the relative positioning of casing collars and tool joints. The total fluid drag can be summed with the other force factors, such as a shear force and mechanical drag. The total drag can then be utilized as the overpull force applied to the stuck drill string.

A system can receive seismic data that can correlate to a subterranean formation. The system can derive a set of seismic attributes from the seismic data. The seismic attributes can include discontinuity-along-dip. The system can determine parameterized results by analyzing the seismic data and the seismic attributes using a deep learning neural network. The deep learning neural network can include a dilation module. The system can determine one or more fault probabilities of the subterranean formation using the parameterized results. The system can output the fault probabilities for use in a hydrocarbon exploration operation.

A method for producing a well includes receiving production information associated with wells within a field; deriving a field specific model from the production information; receiving production information associated with the well; projecting production changes associated with installing artificial lift at the well at a projected date, the projecting using a production analysis engine applied to the field specific model, the projecting including determining a set of artificial lift parameters; and installing the artificial lift at the well in accordance with the artificial lift parameters.

An apparatus and method for planning a drilling sequence. The solution relates also to a rock drilling rig and to a computer program product. The apparatus searches several possible drilling scenarios for drilling a given drilling pattern. Different actions required by the scenarios have given costs and the executed algorithm calculates total costs of found scenarios. Scenarios with the lowest costs are selected.

A method of forecasting well production in an accurate and computational cost-efficient manner implementing a hybrid, iterative approach which is computationally efficient, numerically stable, and improves the accuracy of results. The iterative method of forecasting well production can estimate a set of scaling factor, determining the average adjusted pressure in the matrix, conductive reservoir volume, and oriented hydraulic fracture, applying an algorithm to convert the average adjusted pressure to an actual average pressure, using actual average reservoir pressure to estimate a new set of scaling factor, estimating a relative error based upon the new scaling factor, determining if the relative error is within a user-defined tolerance, and performing the above steps again if relative error is not within the user-defined tolerance or storing the new scaling factors. The new scaling factors can be used to determine a production rate. A second approach implements statistical, data analytics, and pattern recognition techniques.

Aspects of the subject technology relate to systems and methods for optimizing production flow monitoring by utilizing data driven in-situ injection. Systems and methods are provided for receiving sensor data from at least one of a distributed fiber optic sensing line positioned along a wellbore and a plurality of subsurface and surface sensors, generating flow models based on the sensor data received from the at least one of the distributed fiber optic sensing line and the plurality of subsurface and surface sensors to optimize production flow, and generating flow profiles based on the flow models and the sensor data received from the at least one of the distributed fiber optic sensing line and the plurality of subsurface and surface sensors to adjust zonal inflow device.

Systems and methods facilitate wellbore planning and cleaning operations, and generally relate to use of a planning system that interfaces with a simulator. The planner and simulator interact, via the interface, to evaluate various wellbore plans and the effects of the plans on cuttings accumulation, cuttings transport, wellbore geometry, or other borehole conditions. The cost, risk, time, or other associated outcomes of the plans can be compared and evaluated and selected. The plan is then implemented. Planning can continue in substantially real-time to continue modifying the plan to provide a desired balance of outcomes.