Abstract: Systems and methods include a computer-implemented method for well models. Historical well data for past operations of a well is received. An expanded minimum/maximum range of values for the historical well data is generated for parameters used in a well model. Object-oriented object instances are generated using the historical well data and the expanded range. Each object instance is an object that includes a specific combination of parameters within the expanded range. The well model is executed using the instances, generating result objects, each representing a result of the well model. A hybrid well model variant of the well model is generated using the object instances and result objects to model combinations of the object instances and the result objects. A data object encapsulating the object instances and result objects is passed by an object-oriented application to a machine learning model in an absence of database tables.

Abstract: A system and method for machine learning with physics-based models to predict multilateral well performance are provided. An exemplary method enables obtaining data associated with well completion, data associated with inflow control valves, and reservoir attributes of multilateral wells. Production scenarios are generated based on the data associated with well completion, the data associated with inflow control valves, and the reservoir attributes of the multilateral wells. The production scenarios are input into a physics-based model of the multilateral wells, and simulation data associated with the multilateral wells output from the physics-based model is obtained. A neural network based machine learning model is trained using the simulation data associated with the multilateral wells and target parameters associated with the multilateral wells, wherein the trained machine learning model is configured to predict multilateral well production parameters.

Abstract: Systems and methods include a computer-implemented method for well models. Historical well data for past operations of a well is received. An expanded minimum/maximum range of values for the historical well data is generated for parameters used in a well model. Object-oriented object instances are generated using the historical well data and the expanded range. Each object instance is an object that includes a specific combination of parameters within the expanded range. The well model is executed using the instances, generating result objects, each representing a result of the well model. A hybrid well model variant of the well model is generated using the object instances and result objects to model combinations of the object instances and the result objects. A data object encapsulating the object instances and result objects is passed by an object-oriented application to a machine learning model in an absence of database tables.

Abstract: Systems and methods include a computer-implemented method for generating synthetic corrosion logs. Processed corrosion log data is generated from historical corrosion logs of previously-drilled wells. A subset of the historical corrosion logs is selected, including selecting metal loss points to use as seed points for generating a corrosion model. The corrosion model is generated using the seed points, including using spatial interpolation to fill gaps between seed points. The corrosion model is validated by iteratively comparing seed logs and test logs to the corrosion model to ensure that the corrosion model fits the seed points within a threshold. A confidence interval is computed for each target location of a target well as a function of synthetic values associated with the seed points. A synthetic log is generated for the target well using the corrosion model, the target locations, and corresponding confidence intervals at each depth level of the target well.

Abstract: A computer system receives data obtained from multiple hydrocarbon wells. The data includes a first set of downhole temperature logs recorded before detection of one or more wellbore leaks in the multiple hydrocarbon wells. A second set of downhole temperature logs is recorded after detection of the one or more wellbore leaks. The computer system extracts multiple features from the data to generate an N-dimensional feature space. The computer system performs dimensionality reduction on the N-dimensional feature space to generate an M-dimensional feature space, wherein M is less than N. The computer system generates one or more machine learning models trained to determine the one or more wellbore leaks in the multiple hydrocarbon wells based on the M-dimensional feature space.