Abstract: System and methods for optimizing parameters for drilling operations are provided. Real-time data including values for input variables associated with a current stage of a drilling operation along a planned well path are acquired. A neural network model is trained to produce an objective function defining a response value for at least one operating variable of the drilling operation. The response value for the operating variable is estimated based on the objective function produced by the trained neural network model. Stochastic optimization is applied to the estimated response value so as to produce an optimized response value for the operating variable. Values of controllable parameters are estimated for a subsequent stage of the drilling operation, based on the optimized response value of the operating variable. The subsequent stage of the drilling operation is performed based on the estimated values of the controllable parameters.

Abstract: An apparatus used to generate forecasts from a high-dimensional parameter data space. The apparatus comprising a reservoir model and a flow simulator module. The reservoir model comprising a plurality input variables, output variables, and at least one algorithmic model. The input variables and output variables are generated by the flow simulator module and variables from a formation and reservoir properties database and a field production database. The flow simulator module generates the at least one algorithmic model and the output variables using at least one selected from a group comprising a full-physics flow simulator, proxy flow simulator for assisted history matching, and a proxy flow simulator for field development optimization. The full-physics flow simulator and the two proxy flow simulators generate the at least one algorithmic model using at least one selected from a group comprising the reservoir model, history matching input variables, and optimization input variables.

Abstract: An apparatus for generating forecasts from a high-dimensional parameter data space comprising a reservoir model, a model order reduction module, and an assisted history matching module. The reservoir model having input variables, output variables, and an algorithmic model. The input variables, output variables, and the algorithmic model are generated by a flow simulator module and from a formation and reservoir properties database and a field production database. The model order reduction module generates a subset of the original or transformed input variables. This subset has a reduced parameter space than that of the input variables. The subset is generated using a function decomposition and a design of experiments (sensitivity analysis) to reduce number of original variables and identify original or transformed input variables that can be used to approximate output variables.

Abstract: System and methods for simulating fluid flow during downhole operations are provided. Measurements of an operating variable at one or more locations within a formation are obtained from a downhole tool disposed in a wellbore within the formation during a current stage of a downhole operation being performed along the wellbore. The obtained measurements are applied as inputs to a hybrid model of the formation. The hybrid model includes physics-based and machine learning models that are coupled together within a simulation grid. Fluid flow within the formation is simulated, based on the inputs applied to the hybrid model. A response of the operating variable is estimated for a subsequent stage of the downhole operation along the wellbore, based on the simulation. Flow control parameters for the subsequent stage are determined based on the estimated response. The subsequent stage of the operation is performed according to the determined flow control parameters.

Abstract: A hydraulic fracturing flow simulation method includes identifying a discrete fracture network that partitions a reservoir into porous rock blocks. The method further includes determining a current network state that includes flow parameter values at discrete points arranged one-dimensionally along the fractures in said network. The method further includes constructing a set of linear equations for deriving a subsequent network state from the current state. The method further includes repeatedly solving the set of linear equations to obtain a sequence of subsequent network states, the sequence embodying a time-dependent spatial distribution of at least one flow parameter, said flow parameter comprising a proppant volume fraction. The method further includes displaying the time-dependent spatial distribution.

Abstract: Aspects of the present disclosure relate to projecting control parameters of equipment associated with forming a wellbore, stimulating the wellbore, or producing fluid from the wellbore. A system includes the equipment and a computing device. The computing device is operable to project a control parameter value of the equipment using an equipment control process, and to receive confirmation that the projected control parameter value is within an allowable operating range. The computing device is also operable to adjust the equipment control process based on the confirmation, and to control the equipment to operate at the projected control parameter value. Further, the computing device is operable to receive real-time data associated with the forming of the wellbore, the stimulating of the wellbore, or the producing fluid from the wellbore. Furthermore, the computing device is operable to adjust the equipment control process based on the real-time data.

Abstract: In some aspects, a one-dimensional proppant transport flow model represents flow of a proppant-fluid mixture in a subterranean region. The one-dimensional proppant transport flow model includes a fluid momentum conservation model and a proppant bed momentum conservation model that account for viscoelastic effects of the proppant-fluid mixture. The one-dimensional proppant transport flow model may also include a proppant momentum conservation model. In some cases the one-dimensional proppant transport flow model may account for any of settling and resuspension of a proppant bed and interphase momentum transfer between the proppant, proppant bed, and the fluid.

Abstract: Historical information about a significant input parameter is stored in a data analytics model of a hydrocarbon reservoir. A historical deep recursive neural network (RNN) model is built based on time-series production data from the hydrocarbon reservoir as a function of the significant input parameter in the data analytics model. The historical deep RNN neural network model is stored on a data storage device. An experiment using the historical deep neural network model is designed to predict the significant input parameter. The experiment is run to produce a significant experimental input parameter. The significant experimental input parameter is compared to the significant input parameter stored in the data analytics model to determine a difference. The data analytics model is adjusted to reduce the difference.

Abstract: An illustrative hydraulic fracturing simulation method includes: creating an initial mesh representation of a subterranean formation, the mesh including mesh nodes; determining one or more fracture paths in the formation; for each of the one or more fracture paths, displacing a subset of the mesh nodes into alignment with the fracture path; interpolating from displacements of the aligned mesh nodes to obtain displacements for each non-aligned mesh node in the mesh, thereby obtaining a deformed mesh representation of the formation; using the deformed mesh to construct a linear set of equations representing fracture creation and propagation caused by injection of a hydraulic fracturing fluid; deriving one or more fracture path extensions from the linear set of equations; and displaying the one or more fracture paths with the one or more fracture path extensions accurately representing the fracture propagation path. The interpolation may be performed using radial basis functions.

Abstract: A system and method for controlling a drilling tool inside a wellbore makes use of Bayesian optimization with range constraints. A computing device samples observed values for controllable drilling parameters such as weight-on-bit (WOB) and drill bit rotational speed in RPM and evaluates a selected drilling parameter such a rate-of-penetration (ROP) for the observed values using an objective function. Range constraints can be continuously learned by the computing device as the range constraints change. A Bayesian optimization, subject to the range constraints and the observed values, can produce an optimized value for the controllable drilling parameter to achieve a predicted value for the selected drilling parameter. The system can then control the drilling tool using the optimized value to achieve the predicted value for the selected drilling parameter.

Abstract: A system, for controlling well site operations, comprising a machine learning engine, a predictive engine, a node system stack, and a blockchain. The learning engine includes a machine learning algorithm, an algorithmically generated earth model, and control variables. The learning algorithm generates a trained data model using the algorithmically generated earth model. The predictive engine includes an Artificial Intelligence (AI) algorithm. The AI algorithm generates a trained AI algorithm using the trained data model and earth model variables using the trained AI algorithm. The system stack is communicable coupled to the predictive engine, the learning engine, the blockchain, sensors, and a machine controller. The blockchain having a genesis block and a plurality of subsequent blocks. Each subsequent block comprising a well site entry and a hash of a previous entry. The well site entry comprises transacted operation control variables.

Abstract: A system for managing well site operations comprising a well site operations module, a chain of blocks of a distributed network, and a sensor bank and control module. The operations module generates earth model variables using a physics model, well log variables or seismic variables, or both, and a trained AI/ML algorithmic model. The chain of blocks comprises a plurality of subsequent blocks. Each subsequent block comprises a well site entry and a hash value of a previous well site entry. A well site entry comprises transacted operation control variables. The well site operations module generates production operation control variables or development operation control variables from earth model variables. The well site entry can also include transacted earth model variables and sensor variables. The sensor bank and control module provides well log variables and the operations module couples control variables to the control module to control well site equipment.

Abstract: A system for managing well site operations, the system comprising executable partitions, predictive engines, node system stacks, and a blockchain. The predictive engines comprise an Artificial Intelligence (AI) algorithm to generate earth model variables using a physics model, well log data variables, and seismic data variables. The node system stacks are coupled to the blockchain, sensors, and machine controllers. Each node system stack comprises a Robot Operating System (ROS) based middleware controller, with each coupled to each partition, each node system stack, each predictive engine, and an AI process or processes. The blockchain comprises chained blocks of a distributed network. The distributed network comprises a genesis block and a plurality of subsequent blocks, each subsequent block comprising a well site entry and a hash value of a previous well site entry. The well site entry comprises operation control variables. The operation control variables are based on the earth model variables.

Abstract: A method for optimizing real time drilling with learning uses a multi-layer Deep Neural Network (DNN) built from input drilling data. A plurality of drilling parameter features is extracted using the DNN. A linear regression model is built based on the extracted plurality of drilling parameter features. The linear regression model is applied to predict one or more drilling parameters.

Abstract: A current value of at least one operational attribute of a current treatment stage of multiple treatment stages of a wellbore treatment operation of a current well in real time is determined. A determination is made of whether a statistics-based model criteria has been satisfied. In response to determining that the statistics-based model criteria is not satisfied, a response to the current stage of the wellbore treatment operation is predicted based on a physics-based model. In response to determining that the statistics-based model criteria is satisfied, the response to the current stage is predicted based on a statistics-based model. A next value of the at least one operational attribute for a next stage is selected based on the predicted response. Adjustment of the next stage of the wellbore treatment operation is initiated based on the next value of the at least one operational attribute.

Abstract: Using production data and a production flow record based on the production data, a deep neural network (DNN) is trained to model a proxy flow simulation of a reservoir. The proxy flow simulation of the reservoir is performed, using an ensemble Kalman filter (EnKF), based on the trained DNN. The EnKF assimilates new data through updating a current ensemble to obtain history matching by minimizing a difference between a predicted production output from the proxy flow simulation and measured production data from a field. Using the updated current ensemble, a second proxy flow simulation of the reservoir is performed based on the trained DNN. The assimilating and the performing are repeated while new data is available for assimilating. Predicted behavior of the reservoir is determined based on the proxy flow simulation of the reservoir. An indication of the predicted behavior is provided to facilitate production of fluids from the reservoir.

Abstract: A method for fracturing a formation is provided. Real-time fracturing data is acquired from a well bore during fracturing operation. The real-time fracturing data is processed using a recurrent neural network trained using historical data from analogous wells. A real-time response variable prediction is determined using the processed real-time fracturing data. Fracturing parameters for the fracturing operation are adjusted in real-time based on the real-time response variable prediction. The fracturing operation is performed using the fracturing parameters that were adjusted based on the real-time response variable prediction.

Abstract: System and methods for training neural network models for real-time flow simulations are provided. Input data is acquired. The input data includes values for a plurality of input parameters associated with a multiphase fluid flow. The multiphase fluid flow is simulated using a complex fluid dynamics (CFD) model, based on the acquired input data. The CFD model represents a three-dimensional (3D) domain for the simulation. An area of interest is selected within the 3D domain represented by the CFD model. A two-dimensional (2D) mesh of the selected area of interest is generated. The 2D mesh represents results of the simulation for the selected area of interest. A neural network is then trained based on the simulation results represented by the generated 2D mesh.

Abstract: A system for determining completion parameters for a wellbore includes a sensor and a computing device. The sensor can be positioned at a surface of a wellbore to detect data prior to finishing a completion stage for the wellbore. The computing device can receive the data, perform a history match for simulation and production using the sensor data and historical data, generate inferred data for completion parameters using the historical data identified during the history match, predict stimulated area and production by inputting the inferred data into a neural network model, determine completion parameters for the wellbore using Bayesian optimization on the stimulated area and production from the neural network model, profit maximization, and output the completion parameters for determining completion decisions for the wellbore.

Abstract: Aspects of the present disclosure relate to receiving data associated with a subterranean reservoir to be penetrated by a wellbore and training a neural network with both the data and a physics-based first principles model. The neural network is then used to make predictions regarding the properties of the subterranean reservoir, and these predictions are in turn used to determine one or more controllable parameters for equipment associated with a wellbore. The controllable parameters can then be used to control equipment for formation, stimulation, or production relative to the wellbore.