Method for determining an optimal value for a control of a drilling operation

专利摘要:
METHOD FOR DETERMINING AN IDEAL VALUE FOR A CONTROL OF A DRILLING OPERATION. A method for determining an optimal value for a control of a drilling operation is provided. Drill data from a drill operation is received. The drilling data includes a plurality of measured values for each of the plurality of drilling control variables during the drilling operation. An objective function model is determined using the received drilling data. The objective function model maximizes a penetration rate for the drilling operation. Measured drill data is received that includes the current drill data values for a different drill operation. An optimal value for a different drilling operation control is determined by running the determined objective function model with the measured drilling data that includes the current drilling data values for the different drilling operation as an input. The optimal value determined for the control of different drilling operation is produced.
公开号:BR112016001904B1
申请号:R112016001904-0
申请日:2014-09-19
公开日:2022-01-25
发明作者:Moray Laing；David Pope；Keith R. Holdaway；James Duarte
申请人:Sas Institute Inc；
IPC主号:

专利说明:

Background of the invention
[0001] Drilling holes of all types and sizes for various industries (water, natural gas, oil, construction, telecommunications, electric power, etc.) in various environments (land, frozen land, seabed, deep seabed) , etc.) can be a complex, expensive and risky process. summary
[0002] In an example embodiment, a method is provided for determining an optimal value for a control of a drilling operation. The drill data history from a previous drill operation is received. The drilling data history includes a plurality of measured values for each plurality of drilling control variables during the previous drilling operation. An objective function model is determined using the received drill data history and a neural network model. The objective function model maximizes a penetration rate for the previous drilling operation. The measured drill data that is received includes current drill data values for a different drill operation. Different drilling operation means that a drilling operation other than wellbore is different from a wellbore from the previous drilling operation. An optimal value for a different drill operation control is determined by running the objective function model with the measured drill data that includes the current drill data values for a different drill operation as an input. The optimal value determined for the control of the different drilling operation is produced (“output”). Additional drill data is received. The additional puncture data includes a second plurality of measured values for each plurality of puncture control variables. Additional drilling data is measured after producing the optimal value determined to control the different drilling operation. A second objective function model is determined using the additional drill data received. The second objective function model maximizes a penetration rate from the previous drilling operation. Measured second drill data is received including current second drill data values for different drill operations. A second ideal value is determined for the control of the different drilling operation by executing the second objective function model determined with the second measured drilling data that includes the second values of the current drilling data for the different drilling operation as an input . The second optimal value determined is produced for the control of the different drilling operation.
[0003] In another example embodiment, a computer-readable medium is provided having stored thereon computer-readable instructions which, when executed by a computing device, causes the computing device to execute the method for determining an optimal value for a control of a drilling operation.
[0004] In yet another exemplary embodiment, a computing device is provided. The computing device includes, but is not limited to, a processor and computer readable medium operatively coupled to the processor. The computer readable medium has instructions stored therein which, when executed by a computing device, cause the computing device to execute the method for determining an optimal value for a control of a drilling operation.
[0005] These and other embodiments may optionally include one or more of the following features. The objective function model can be determined by evaluating a plurality of neural network model configurations. Instructions for determining the objective function model may include: defining a training dataset as a first part of the received drill data history; defining a validation dataset as a second part of the received drilling data history; the definition of a first configuration of the neural network model; training a first neural network model using the defined training dataset based on the first configuration of the first defined neural network model; predicting model output data with the validation dataset defined as an input to the first trained neural network model, comparing the output data of the prediction model with the output data of the validation dataset, and determining a first validity rating for the first trained neural network model based on comparing the output data of the prediction model with the output data of the validation dataset. Instructions for determining the objective function model may also include: defining a second configuration of the neural network model, training a second neural network model using the defined training dataset based on the defined configuration of the second neural network model, predicting the output data of the second model with the validation dataset defined as an input to the second trained neural network model, comparing the output data of the second model of the prediction with the output data of the validation dataset, and determining a second validity rating for the second trained neural network model based on comparing the output data of the second prediction model with the output data of the validation dataset. The objective function model can be determined as the first trained neural network model or the second trained neural network model based on a comparison between the first determined validity rank and the second determined validity rank. A prediction model can be determined using the received drill data history, so that the prediction model can predict a value of a control variable input for the given objective function model. The computer-readable instructions can also cause the computing device to determine a current predicted value of the control variable input to the given objective function model, so that the given objective function model can still be executed with the current predicted value as one entrance. The prediction model can be determined using a decision tree model. The additional drilling data received may include a second plurality of measured values for each of the plurality of drilling control variables during the different drilling operation. The objective function model can also minimize a specific mechanical energy for the previous drilling operation. The objective function model can also optimize wellbore stability for the previous drilling operation. The objective function model can also optimize the wellbore stability of the previous drilling operation. The determined ideal value can be generated for a display device. The determined ideal value can be generated for a control device associated with the adjustment of a target value (“set point”) of the control. The computing device may be physically located on top of a drilling rig. In other embodiments, the computing device may be physically remote from the drilling rig.
[0006] Other main features of the described subject will become apparent to those skilled in the art after evaluating the drawings, the detailed description, and the appended claims below. Brief description of drawings
[0007] Illustrative embodiments of the subject matter described will hereinafter be described with reference to the accompanying drawings, with like numerals indicating like elements.
[0008] Figure 1 represents a block diagram of a drilling data collection system according to an illustrative embodiment;
[0009] Figure 2 represents a block diagram of a model defining device, according to an illustrative embodiment;
[00010] Figure 3 represents a flowchart illustrating examples of operations performed by the model defining device of figure 2, according to an illustrative embodiment;
[00011] Figure 4 represents an additional block diagram of an optimization device, according to an illustrative embodiment;
[00012] Figure 5 represents a flowchart illustrating examples of operations performed by the optimization device of figure 4, according to an illustrative embodiment;
[00013] Figure 6 represents calculations related to a penetration rate during a drilling operation, according to an illustrative embodiment;
[00014] Figure 7 represents a block diagram of a distributed processing system, according to an illustrative embodiment;
[00015] Fig. 8 represents a block diagram of a data stream ("event stream") processing device (ESP) of the distributed processing system of Fig. 7, according to an illustrative embodiment; and
[00016] Figure 9 represents a flowchart illustrating examples of operations performed by the ESP device of figure 8, according to an illustrative embodiment. Detailed Description
[00017] Referring to Figure 1, a block diagram of a drilling data collection system 100 is shown in accordance with an illustrative embodiment. The drilling data collection system 100 may include a plurality of drilling rigs 101, a network 110, and a data store 112. Few, different, and/or additional components may be incorporated into the drilling data collection system. drilling rig 100. For illustration, the plurality of drilling rigs 101 may include a first drilling rig 102, a second drilling rig 104, a third drilling rig 106, a fourth drilling rig 108. The plurality of drilling rigs 101 can include any number of drilling rigs. A drilling rig of the plurality of drilling rigs 101 can be active or inactive. The plurality of drilling rigs 101 can be configured to drill holes of any type and size for various industries (e.g. water, natural gas, oil, construction, telecommunications, electric power, etc.) in various environments (e.g. land , frozen land, seabed, deep seabed, etc.). The plurality of drilling platforms 101 can be distributed locally, regionally or globally.
[00018] Network 110 may include one or more networks of the same or different types. Network 110 may be any type or combination of wired and/or wireless public or private network, including a cellular network, a local area network, a wide area network such as the Internet, etc. Network 110 may additionally comprise subnets and consist of any number of devices. The plurality of drilling rigs 101 send communications over network 110 to data storage 112. The plurality of drilling rigs 101 may communicate using various transmission media which may be wired and/or wireless, as understood by the technicians on the subject.
[00019] Data store 112 stores drilling data from the plurality of drilling rigs 101 which includes a plurality of measured values for each of the plurality of drilling control variables (target variable or dependent for prediction modeling ) during a well or borehole drilling operation. The plurality of values may be measured for each of the plurality of puncture control variables at a plurality of points in time over a period of time. For example, the plurality of values may be measured for each of the plurality of drilling control variables hourly for a period of time throughout the year, although other measurement intervals and time periods may be employed.
[00020] Referring to Figure 2, a block diagram of a model defining device 200 is shown in accordance with an illustrative embodiment. The template definition device 200 may be located on a drilling rig of the plurality of drilling rigs 101 or remote from the plurality of drilling rigs 101. The template definition device 200 may include an input interface 202, a output interface 204, a communication interface 206, a computer readable medium 208, a processor 210, a model definition application 222, data storage 112, and an objective function model 224, and a prediction model 225. Few components, different and/or additional components can be incorporated into the 200 model definition device.
[00021] Input interface 202 provides an interface to receive user information for input to model defining device 200, as understood by those skilled in the art. The input interface 202 can interface with various input technologies, including, but not limited to, a keyboard 212, a mouse 214, a microphone 215, a display 216, a track ball, a numeric keypad, one or more buttons. , etc., to allow the user to enter information into a Model 200 defining device or to make selections presented in a user interface displayed on the display. The same interface may support both the input interface 202 and the output interface 204. For example, the display 216 comprising a touch-sensitive screen provides user input and presents the rating to the user. The model defining device 200 can have one or more input interfaces that use the same or a different input interface technology. The input interface technology may be further accessible by the model defining device 200 through the communication interface 206.
[00022] The output interface 204 provides an interface to information generated for analysis by a user of the model definition device 200 and/or for use by another application. For example, output interface 204 can interface with various output technologies, including, but not limited to, display 216, speaker 218, a printer 220, etc. The model definition device 200 can have one or more output interfaces that use the same or a different output interface technology. The output interface technology may additionally be accessible by the model defining device 200 via the communication interface 206.
[00023] The communication interface 206 provides an interface for receiving and transmitting data between devices using various protocols, transmission technologies, and means, as understood by those skilled in the art. Communication interface 206 may support communication using various transmission media which may be wired and/or wireless. The model defining device 200 may have one or more communication interfaces that use the same or a different communication interface technology. For example, the Model 200 definition device may support communication using an Ethernet port, a Bluetooth antenna, a telephone jack, a USB port, etc. Data and messages may be transferred between the model definition device 200 and/or distributed systems 232, one or more drilling operation sensors 226, and/or one or more drilling operation control parameters 228 of the plurality of drilling rigs 101 using the communication interface 206.
[00024] The computer readable medium 208 is an electronic holding or storage location for the information so that the information can be accessed by the processor 210, as understood by those skilled in the art. The computer readable medium 208 may include, but is not limited to, any type of random access memory (RAM), any type of read-only memory (ROM), any type of flash memory, etc. magnetic storage (e.g. hard disk, floppy disk, magnetic tapes,...), optical discs (e.g. compact disc (CD), digital versatile disk (DVD),...), smart cards, flash memory devices , etc. The model defining device 200 may have one or more computer readable media that use the same or a different technology of memory media. For example, computer readable media 208 may include different types of computer readable media that can be hierarchically organized to provide efficient access to stored data, as understood by those skilled in the art herein. As an example, a cache can be implemented in a smaller, faster memory that stores data copies of the most frequently/recently accessed key memory locations to reduce an access latency. The model definition device 200 may also have one or more drives that support loading from a memory medium, such as a CD, DVD, an external hard drive, etc. One or more external hard drives can additionally be connected to the model definition device 200 using the communication interface 206.
[00025] Processor 210 executes instructions as understood by those skilled in the art. Instructions can be executed by a computer for special purposes, logic circuits, or hardware circuits. Processor 210 may be implemented in hardware and/or firmware. Processor 210 executes an instruction, meaning that it performs/controls the operations designated by that instruction. The term “execution” is the process of executing an application or performing the operation designated by an instruction. Instructions can be written using one or more programming languages, scripting language, assembly language, etc. Processor 210 operatively couples with input interface 202, output interface 204, communication interface 206, and computer readable medium 208 for receiving, sending, and processing information. Processor 210 may retrieve a set of instructions from a permanent memory device and copy the instructions, in executable form, to a temporary memory device which is generally some form of RAM. Model defining device 200 may include a plurality of processors using the same or different processing technology.
[00026] Model definition application 222 performs operations associated with objective function model definition 224 and/or prediction model 225 for one or more drilling operations from data stored in data store 112. Some or all the operations described herein may be incorporated into the model definition application 222. The operations may be implemented using hardware, firmware, software, or any combination of these methods. Referring to the exemplary embodiment of Figure 2, the model definition application 222 is implemented in software (comprised of computer-readable instructions and/or computer executables) stored on computer-readable medium 208 and accessible by processor 210 for execution. of instructions that embody the operations of the template-defining application 222. The template-defining application 222 may be written using one or more programming languages, assembly languages, scripting languages, etc.
[00027] Model definition application 222 can be implemented as a web application. For example, the template definition application 222 can be configured to receive hypertext transport protocol (HTTP) responses and to send HTTP requests. HTTP responses can include web pages such as hypertext markup language (HTML) documents and linked objects generated in response to HTTP requests. Each web page can be identified by a uniform resource locator (URL) that includes the location or address of the computing device that contains the resource to be accessed in addition to the location of the resource on the computing device. The type of file or resource depends on the Internet application protocol, such as file transfer protocol, HTTP, H.323, and so on. The file accessed can be a plain text file, an image file, an audio file, a video file, an executable, a common port interface application, a Java applet, an extensible markup language (XML) file. ), or any other file type supported by HTTP.
[00028] Data store 112 may be stored on computer readable medium 208 or on one or more computing devices (eg, distributed systems 232) and accessed using communication interface 206. Data stored in data store 112 may be received from one or more drill operation sensors 226. Examples of sensors include pressure sensors, temperature sensors, position sensors, speed sensors, acceleration sensors, flow rate sensors, etc., which can be mounted to various components used as part of the drilling operation. For example, one or more drill operation sensors 226 may include surface sensors that measure a hook load, a fluid rate, a temperature and density in and out of the wellbore, a vertical pipe pressure, a torque at the surface, a drill pipe rotation speed, a penetration rate, a specific mechanical energy, etc., and downhole sensors that measure a drill bit rotation speed, fluid densities, downhole torque downhole, downhole vibration (axial, tangential, lateral), a weight applied to a drill bit, an annular pressure, a differential pressure, an azimuth, a slope, a severe downhole curve, a measured depth, a depth vertical, a downhole temperature, etc. Other data may include one or more drilling operation control parameters 228 which may control settings such as a mud motor speed to flow rate, a drill diameter, a predicted top formation, seismic data, meteorological, etc. Other data can be generated using physical models, such as an earth model, a meteorological model, a seismic model, a base hole arrangement model, a well plan model, an annular friction model, etc. In addition to sensor and control settings, predicted outputs, e.g. penetration rate, specific mechanical energy, hook load, inflow fluid rate, outflow fluid rate, pump pressure , surface torque, drill pipe rotation speed, annular pressure, annular friction pressure, annular temperature, equivalent circulation density, etc. can also be stored in the data store.
[00029] The plurality of values can be measured from the same drilling operation, from one or more neighboring drilling operations, from one or more drilling operations with similar geological characteristics, from any of a or more drilling operations, etc. For example, a drilling operation in an environment with similar permeability and porosity can be used. The plurality of values may result from the values of control variables chosen by an operator during a prior period of time over the same or different drilling operation.
[00030] Data stored in data store 112 may include any type of content represented in any computer-readable format, such as binary, alphanumeric, numeric, string, markup language, etc. Content may include textual information, graphic information, image information, audio information, numerical information, etc. which, in addition, can be encoded using various encoding techniques, as understood by those skilled in the art. The data store 112 may be stored using various formats as known to those skilled in the art, including a file system, a relational database, a table system, a structured query language database, etc. For example , the data store 112 may be stored in a storage space ("cube") distributed over a computer network, as understood by those skilled in the art. As another example, data store 112 may be stored on a multi-node Hadoop® cluster, as understood by those skilled in the art. ApacheTM Hadoop® is an open source software framework for distributed computing supported by the Apache Software Foundation. As another example, data storage 112 may be stored in a cloud of computers and accessed using cloud computing technologies, as understood by those skilled in the art. The SAS® LASRTM Analytical Server developed and supplied by SAS Institute Inc., of Cary, North Carolina, USA can be used as an analytical platform to allow multiple users to simultaneously access data stored in data warehouse 112.
[00031] If the data store 112 is distributed across distributed systems 232, a distributed processing system can be used. Through the illustration, the distributed processing system can be implemented using a multi-node Hadoop® cluster, using a computer network storing a data storage space, using SAS® LASRTM Analytical Server, using computer cloud, etc., as understood by experts in the field. For example, a distributed control device may coordinate access to distributed data storage 112 across distributed systems 232 when requested by model defining device 200. One or more components of the distributed processing system may support multitasking, as understood by those skilled in the art. on the subject. The components of the distributed processing system may be located in a single room or adjacent rooms, in a single facility, and/or may be geographically distributed from one another.
[00032] The data in the data store 112 may be purged to impute missing values, stable noise data, identify and eliminate outliers, and/or resolve inconsistencies, as understood by those skilled in the art. The data in the data store 112 may be transformed to normalize and aggregate the data, to unify data formats such as dates, and to convert nominal data types to numeric data types, as understood by those skilled in the art.
[00033] With reference to Figure 3, examples of operations associated with the Model 222 definition application are described. The model definition application 222 can be used to create the objective function model 224 and/or one or more predictive models 225 using the data stored in the data store 112. The objective function model 224 supports a determination of an ideal value for a control of a drilling operation using sensed data measured during the drilling operation by one or more sensors in the drilling operation 226 and/or using control settings for one or more of the drilling operation 228 control parameters of the drilling operation drilling. The prediction model 225 supports a determination of a predicted value for a variable control of the drilling operation using detected data measured during the drilling operation by one or more sensors in the drilling operation 226 and/or using readout control settings for a or more drilling operation control parameters 228 of the drilling operation.
[00034] Additionally, few or different operations can be performed depending on the embodiment. The order of presentation of operations in figure 3 is not intended to be limiting. Although some of the operational flows are presented in sequence, the various operations can be performed in several repetitions, at the same time (in parallel, for example, using wires), and/or in orders other than those illustrated. For example, a user may run the template definition application 222, which causes the presentation of a first user interface window, which may include a plurality of menus and selectors, such as drop-down menus, buttons, checkboxes, text, hyperlinks, etc., associated with the 222 template definition application, as understood by those skilled in the art. An indicator may indicate one or more user selections from a user interface, one or more data entries in a user interface data field, one or more data items read from the computer readable medium 208, or, otherwise defined with one or more default values, etc., which are received as an input by the Model 222 definition application.
[00035] In an operation 300, a first indicator of one or more types of predictive models, configurations, and one or more variables for prediction is received. Predictive models predict the values of one or more control variables (variables) in a dataset from the values of other variables in the dataset. For example, the first indicator indicates a name of a prediction model type and a control variable for the prediction using any prediction model type. One or more types of predictive models and configurations may be defined for a plurality of control variables to support the prediction of the plurality of control variables, independently, or in combination, as in the sequence, where a predicted control variable is a input to another prediction model for a control variable other than one or more control variables. A name of a prediction model type may be selectable for each of the plurality of control variables. Examples of predictive control variables include penetration rate, specific mechanical energy, hook load, inflow fluid rate, outflow fluid rate, pump pressure, surface torque, drill pipe rotation speed , annular pressure, annular temperature, annular friction pressure, equivalent circulation density, etc.
[00036] For illustration, the name of a prediction model type can be “Neural Network”, “Linear Regression”, “Nonlinear Regression”, “Support Vector Machine”, “Decision Tree”, “Partial Least Squares ”, “Gradient dynamization”, etc. A configuration identifies one or more initialization values based on the prediction model type. For example, when the prediction model type is indicated as “Neural Network”, a number of hidden layers, a number of nodes per layer, a propagation method, etc., can be identified by the first indicator. A plurality of configurations can be defined. For example, when the prediction model type is neural network, a range of numbers of hidden layers, a range of numbers of nodes per layer, etc., can be identified by the first indicator.
[00037] For further illustration, one or more control variables for prediction and data in data store 112 may be provided to SAS® Enterprise MinerTM for prediction modeling developed and provided by SAS Institute Inc., of Cary, North Carolina, USA. As an example, SAS® Enterprise MinerTM includes types of predictive models for neural networks (AutoNeural, DMNeural, Neural Network), decision trees (Decision Tree, Dynamic Gradient), regression models (Dmine Regression, Minimum Angle Regressions ( LARS), Regression), k-nearest neighbor models (Memory Based Reasoning (MBR)), a partial least squares model (Partial Least Squares), a support vector machine (Support Vector Machine), a set of models that are integrated to define a prediction model (Ensemble), etc.
[00038] The first indicator can be received by the Model 222 definition application after selection from a UI window or after input by a user in a UI window. A default value for the predictive model types and settings may additionally be stored, for example, on computer-readable medium 208. In an alternative embodiment, the predictive model types and settings and one or more control variables for prediction may not be selectable.
[00039] In an operation 302, a second data store indicator 112 is received. For example, the second indicator indicates a data storage location 112. As an example, the second indicator may be received by a model-defining application 222 after selection from a UI window or after input by a user in a UI window. In an alternative embodiment, storage data 112 may not be selectable. For example, a most recently created data store can be used automatically.
[00040] As discussed earlier, data storage 112 may be stored in a storage space distributed across a computer network, may be stored in a multi-node Hadoop® cluster distributed across one or more computers, may be stored in a file system distributed across one or more computers, in a relational database, in one or more tables, in a structured query language database, etc.
[00041] In an operation 304, the data stored in the data store 112 is explored and extracted to select the control variables (input, independent variables) significant for the determination of a prediction model for each one or more control variables for prediction. For example, in operation 304, data stored in data store 112 is reduced to obtain a minimal representation in size and volume, as well as to retain consistent variation and entropy for similar analytical results. Numerical data types may be discretized, as understood by those skilled in the art, to simplify analytical processing.
[00042] Examples of data mining techniques include factor analyses, principal component analyses, correlation analyses, etc., as understood by those skilled in the art. For illustration, SAS® Enterprise MinerTM, developed and supplied by SAS Institute Inc., of Cary, North Carolina, USA, includes nodes for exploring and selecting data or modifying control variables as input variables. Examples of nodes include transformation nodes, cluster nodes, association rules nodes, a variable selection node, a descriptive statistics node, a principal components node, etc.
[00043] For example, input variables with a high degree of correlation with respect to the prediction of each or more control variables for prediction can be selected. Examples of input variables include hook load, inflow fluid rate, outflow fluid rate, pump pressure, surface torque, rotational speed of a drill pipe, annular pressure, annular temperature, annular volume, etc. Input variables may additionally include contextual variables such as a lithology, a rheology of fluid properties, an applied electrical energy, an applied hydraulic energy, a type of bit, a downhole assembly design, a temperature , etc.. The input variables may additionally include a descriptive geomechanical model, a prognostic lithology, a well construction plan, etc.
[00044] In an operation 306, a third indicator is received for selecting training data for the prediction model from the data store 112. The third indicator may be received by a model definition application 222, for example , after selection from a UI window, or after input by a user in a UI window. The third indicator identifies a first piece of data stored in the data store 112 for use in training the prediction model. The third indicator may indicate a number of data points to include, a percentage of data points from the entire data store 112 to be included, etc. A subset may be created from the data store 112 by sampling. An example of a sampling algorithm is uniform sampling. Other random sampling algorithms may be used.
[00045] In an operation 308, a fourth indicator for selecting validation data for the prediction model from the data store 112 is received. The fourth indicator may be received by a model-defining application 222, for example, upon selection from a UI window or upon input by a user in a UI window. The fourth indicator identifies a second part of the data stored in the data store 112 for use in validating the prediction model. The fourth indicator may indicate a number of data points to include, a percentage of data points from the entire data store 112 to be included, etc. A subset may be created from the data store 112 by sampling. An example of a sampling algorithm is uniform sampling. Other random sampling algorithms may be used. The data points from the data store 112 selected for the validation data may be distinct from the data points from the data store 112 selected for the training data.
[00046] In an operation 310, a prediction model is selected based on the first indicator or based on a default model stored on computer readable medium 208. In an operation 312, the selected prediction model is initialized. In an operation 314, the initialized prediction model is trained using the selected training data, as indicated by the third indicator.
[00047] In an operation 316, the output data is predicted with the validation data, selected as indicated by the fourth indicator, as an input to the trained prediction model. In an operation 318, the predicted output data is compared to the actual output data included with the validation data. In a 320 operation, a validity rating is determined based on the comparison. In an operation 322, the determined validity rating is stored, for example, on computer readable medium 208 in association with an indicator of the selected prediction model.
[00048] In a 324 operation, a determination is made as to whether or not another prediction model exists for evaluation. When there is another prediction model for evaluation, processing continues in operation 310. When there is no other prediction model for evaluation, processing continues in operation 326. In operation 310, a next prediction model is selected based on the first indicator .
[00049] In operation 326, a best prediction model is selected for each one or more control variables for prediction. For example, the validity ratings stored for each iteration of operation 322 are compared, and the prediction model associated with the best validity rating is selected. The best validity rating can be a maximum or minimum value of the stored validity ratings for each iteration of operation 322. For example, if the validity rating is an incorrect rating rate, a minimum validity rating indicates the best model despite that if the validity rating is a correct rating rate, a maximum validity rating indicates the best model.
[00050] In an operation 328, the selected best prediction model is stored, for example, on computer readable medium 208. The selected prediction model can be stored in association with a specific drilling location, a specific drilling field, a specific type of drilling environment, etc. The selected prediction model is stored as prediction model 225. A different prediction model 225 can be defined for each one or more control variables for the prediction.
[00051] In an operation 330, a fifth indicator of one or more types of objective function models and settings is received. The fifth indicator may indicate that the objective function model maximizes penetration rate, minimizes specific mechanical energy, and/or optimizes wellbore stability. For example, the fifth indicator indicates a name of an objective function model type and one or more control variables to maximize/minimize/optimize. For illustration, the name can be “Neural Network”, “Linear Regression”, “Nonlinear Regression”, “Support Vector Machine”, “Decision Tree”, “Partial Least Squares”, “Gradient Dynamization”, etc. A configuration identifies one or more initialization values based on the model type of the objective function. For example, when the model type of the objective function is indicated as “Neural Network”, a number of hidden layers, a number of nodes per layer, a propagation method, etc., can be identified by the first indicator. A plurality of configurations can be defined. For example, when the model type of the objective function is neural network, a range of numbers of hidden layers, a range of numbers of nodes per layer, etc., can also be identified by the first indicator.
[00052] For example, SAS® Enterprise MinerTM includes types of objective function models for neural networks (AutoNeural, DMNeural, Neural Network), decision trees (Decision Tree, Gradient Dynamics), regression models (Dmine Regression, Regressions (LARS), Regression), k-nearest neighbor models (Memory Based Reasoning (MBR)), a partial least squares model (Partial Least Squares), a support vector machine (Support Vector Machine), a set of models that are integrated to define an objective function model (Ensemble), etc.
[00053] As an example, SAS® Enterprise MinerTM includes a PROCNEURAL neural network procedure that can configure, initialize, train, predict, and classify a neural network model. Input nodes can be input variables such as hook load, inflow fluid rate, outflow fluid rate, pump pressure, surface torque, speed of rotation of the drill pipe, annular pressure, drill type, applied electrical power, applied hydraulic power, and well construction plan. The outlet node(s) can be predicted target points for controlling the inflow fluid rate, outflow fluid rate, surface torque, and rotational speed. of the drill pipe to maximize penetration rate, minimize specific mechanical energy, and/or optimize wellbore stability. For illustration, the objective function model can be implemented as a neural network with two or three hidden nodes, a lookahead control adaptation, a supervised learning mode, and backpropagation to perform the sensitive analysis that determines how each input variable influences the exit node(s) to maximize penetration rate, minimize specific mechanical energy, and/or optimize well stability.
[00054] The fifth indicator can be received by the Model 222 definition application after selection from a UI window or after input by a user in a UI window. A default value for the objective function model types and settings may additionally be stored, for example, on computer-readable medium 208. In an alternative embodiment, the objective function model types and settings may not be selectable. Operations 302 to 328 are repeated with the predictive models replaced with the objective function models to select the best objective function model as opposed to a best prediction model. The best selected objective function model is stored as objective function model 224.
[00055] Cluster analysis of data can be used to stratify current results for the evaluation of the prediction model 225 and/or the objective function model 224 under working conditions. Cluster analyzes can be used to select variables in operation 304 and to select prediction model 225 and/or objective function model 224 in operation 326. The type of cluster analyzes can be selected based on the geological target and mechanical aspects of the wellbore plan. As an example, if drilling in the pre-salt layer, the type of clustering of the analysis model may be different than the one used during drilling in sand. Several neural networks can be used, so that one neural network provides inputs to another neural network.
[00056] For illustration, a prediction model can be selected in operation 324 for a plurality of solutions, such as specific mechanical energy, penetration rate, and downhole hydraulic stability. Prediction values for specific mechanical energy, penetration rate, and downhole hydraulic stability can be fed into objective function models that determine optimal values for surface control input variables such as head load. hook, the rate of fluid in the inflow, the rate of fluid in the outflow, the torque at the surface, the rotation speed of a drill pipe, and the annular pressure to achieve the specific mechanical energy of the prediction, the rate penetration, hydraulic stability in the wellbore, and equivalent circulation density.
[00057] Referring to Figure 4, a block diagram of an optimization device 400 is shown in accordance with an illustrative embodiment. The optimization device 400 may include a second input interface 402, a second output interface 404, a second communication interface 406, a second computer readable medium 408, a second processor 410, prediction model 225, objective function model 224, control data 414, sense data 412, and optimization model 416. Few components, different components, and/or additional ones can be incorporated into the optimization device 400.
[00058] After being selected, using the model definition device 200, the prediction model 225 and/or the objective function model 224 can be stored on the second computer-readable medium 408 and/or accessed by the optimization device 400 through of the second communication interface 406. The model definition device 200 and the optimization device 400 can be integrated in the same computing device. The model definition device 200 and the optimization device 400 can be different computing devices. Optimizing device 400 may be located on one drilling rig of the plurality of drilling rigs 101 or remote from the plurality of drilling rigs 101. Optimizing device 400 may be located on a different drilling rig of the plurality of rigs. hole 101 from which the data is stored in the data store 112. The data generated by the optimization device 400 can be stored in the data store 112 through the second communication interface 406.
[00059] The second input interface 402 provides the same or similar functionality as described with reference to the input interface 202 of the model defining device 200 to refer to the optimization device 400. The second output interface 404 provides the same or similar functionality as described with reference to output interface 204 of model defining device 200 to refer to optimizing device 400. Second communication interface 406 provides the same or similar functionality as described with reference to communication interface 206 of model defining device 200, although referring to optimizing device 400. Data and messages may be transferred between optimizing device 400 and drilling operation control(s) 228 and /or drill operation sensor(s) 226 using second communication interface 406. Data and messages ens can be transferred between the optimization device 400 and the drill operation control(s) 228 and/or drill operation sensor(s) 226 using the second input interface 402 and/or the second input interface. output 404. Second computer readable medium 408 provides the same or similar functionality as described with reference to computer readable medium 208 of model defining device 200, while referring to optimization device 400. Second processor 410 provides the same or similar functionality as described with reference to processor 210 of model defining device 200, while referring to optimizing device 400.
[00060] The optimization model 416 supports the determination of an optimal value for a control of the drilling operation using detected data 412 measured during the drilling operation and control data 414 generated during the drilling operation. Some or all of the operations described herein may be incorporated into the optimization model 416. The operations of the optimization model 416 may be implemented using hardware, firmware, software, or any combination of these methods. Referring to the exemplary embodiment of Figure 4, the optimization model 416 is implemented in software (comprised of computer-readable and/or computer-executable instructions) stored on second computer-readable medium 408 and accessible by second processor 410 for execution. of instructions that embody the operations of the optimization model 416. The optimization model 416 may be written using one or more programming languages, assembly languages, scripting languages, etc. The 416 optimization model can be implemented as a web application.
[00061] With reference to Figure 5, examples of operations associated with the optimization model 416 are described. Additionally, different or few operations may be performed depending on the embodiment. The order of presentation of operations in Figure 5 is not intended to be limiting. Although some of the operational flows are presented in sequence, the various operations can be performed in several repetitions, at the same time (in parallel, for example, using wires), and/or in orders other than those illustrated.
[00062] In a 500 operation, an initial penetration rate graph is defined for a selected drilling rig. For example, input variables describing the drilling conditions expected to be encountered during the drilling process are fed into the prediction model 225 and/or objective function model 224 to plot the initial penetration rate graph.
[00063] Referring to Figure 6, a penetration rate graph is shown according to an illustrative embodiment. The penetration rate graph can include a 600 planned penetration rate (ROP) curve, a 602 upper limit ROP curve, a 604 lower limit ROP curve, a 606 current ROP curve, a 608 ROP curve of the prediction, and a 610 curve of the ideal ROP. The planned ROP curve 600 can be defined by the input variables describing the expected drilling conditions to be encountered during the drilling process input to the 416 optimization model. The penetration rate graph shown in Figure 6 represents a final ROP after the end of drilling. The initial ROP graph can include Planned ROP Curve 600, Upper Limit ROP Curve 602, and Lower Limit ROP Curve 604.
[00064] Upper bound ROP curve 602 and lower bound ROP curve 604 can be calculated using bounds of statistical probability. For illustration, the upper limit ROP curve 602 and the lower limit ROP curve 604 can be determined using Shewhart formulas, as understood by those skilled in the art. The upper limit ROP curve 602 and the lower limit ROP curve 604 can be associated with what are traditionally known to those skilled in the art as the “Western Electric rules” created by Dr. Shewhart for creating alerts.
[00065] Current ROP curve 606 shows the current ROP that resulted during the drilling operation from start to finish. The prediction ROP curve 608 shows the prediction ROP during the drilling operation using a prediction model 225 that predicts the ROP. The prediction ROP curve 608 can be calculated using a prediction model determined for the ROP. The ideal ROP curve 610 shows a given ideal ROP using the objective function model 224 in the optimization model 416. The ideal ROP curve 610 can be calculated using an optimization model determined for the ROP.
[00066] Referring again to Figure 5, in an operation 502, control data 414 and sense data 412 are received. For example, control data 414 and sense data 412 associated with the input variables indicated in operation 304 are received at or near real time from the drill operation control(s) 228 and from the ) drilling operation sensor(s) 226 indicating and measuring, respectively, the variable values of the current control for the drilling operation.
[00067] In an operation 504, a target value, for one or more controls of the drilling operation, is determined by executing the selected objective function model with the control data 414 and sense data 412 received as an input. The target value determined for the control variables seeks to maximize penetration rate, minimize specific mechanical energy, and/or optimize wellbore stability. As an example, SAS®OR, developed and supplied by SAS Institute Inc. of Cary, North Carolina, USA, includes an OPTMODEL procedure that provides a framework for specifying and solving the objective function model 224. Examples of controls include those that control the rate of fluid in the inlet flow, rate of fluid in the outflow, the torque at the surface, and the rotation speed of the drill pipe for the drilling operation.
[00068] In a 506 operation, the target value determined for one or more controls of the drilling operation is produced. For example, the target value determined for one or more controls of the drilling operation can be output to display 216, loudspeaker 218, and/or printer 220 for analysis by a user. As yet another example, the setpoint determined for one or more controls of the drilling operation can be output to a control device associated with setting the setpoint of each control. When the determined setpoint is greater than a current setpoint for the control, the value for the control can be increased, although when the determined setpoint is less than a current setpoint for the control , the value for the control can be reduced.
[00069] In a 508 operation, the prediction ROP curve 608 and the ideal ROP curve 610 on the ROP graph are updated to reflect changes in the prediction and ideal values based on the target value determined for one or more drilling operation controls. On a 510 operation, the updated ROP graph is produced. For example, the updated ROP graph can be output to display 216 and/or printer 220 for analysis by a user.
[00070] In an operation 512, a determination is made as to whether an update to the objective function model 224 or to the prediction model 225, for any control variables, for the prediction has been performed. An indicator may be received indicating that an update to one or more of the models has been performed. For example, the objective function model 224 and/or prediction model 225 may be updated periodically, such as every second, minute, hour, day, week, month, year, etc. A timer can trigger the receipt of the indicator. A user can trigger the reception of the indicator. For example, a user can monitor drilling control variables, such as the updated ROP graph, and determine that an update is performed. If the determination is to perform an update, processing continues in a 514 operation to update one or more of the models. If the determination is not to perform an update, processing continues at operation 502 to proceed to processing control data 414 and sense data 412 as they are received in real time.
[00071] In operation 514, one or more models are updated, for example, by updating data stored in data store 112 and repeating one or more of operations 302 to 328 for the objective function model 224 and/or for the prediction model 225. For example, operations 310 to 328 can be repeated. Data stored in data store 112 in a previous iteration of operation 328, in addition to data measured and stored in data store 112 subsequent to the last iteration of operation 328, may be used to update one or more models.
[00072] Referring to Figure 7, a block diagram of a drilling system 700 is shown in accordance with an illustrative embodiment. Drilling system 700 may include first drilling rig 102, net 110, and model defining device 200. Few components, different components and/or additional components may be incorporated into drilling system 700. First drilling rig 102 may include drilling operation sensors 226, drilling operation control parameters 228 that generate control data 414, a platform control interface device 704, a local data aggregator 706, a stream processing device event data (ESP) 708, a display system 710, and a second network 712. The rig control interface device 704 can be configured to receive data from drill operation sensors 226 and drill operation control parameters. drilling 228. Incoming data may be aggregated in pre-existing platform aggregators, such as the local data aggregator 706, as and understood by technicians on the subject. The display system 710 provides displays relating to a current state of the first drilling rig 102. For example, the display system 710 can present the ROP graph of Figure 6 to operators of the first drilling rig 102 in addition to other values of the current control target value.
[00073] The second network 712 may include one or more networks of the same or different types. The second network 712 may be any type or combination of wired and/or wireless public or private network, including a cellular network, a local area network, a wide area network such as the Internet, etc. The second network 712 may additionally comprise subnets and be composed of any number of devices. Although connections through the second network 712 are not explicitly shown in the illustrative embodiment of Figure 7, one or more of the components of the drilling system 700 may communicate using the second network 712, which includes various transmission media that may be wired and /or wirelessly, as understood by those skilled in the art. One or more of the components of the drilling system 700 may be directly connected or integrated into one or more computing devices.
[00074] Referring to Figure 8, a block diagram of ESP device 708 is shown, according to an illustrative embodiment. The ESP device 708 may include a third input interface 800, a third output interface 802, a third communication interface 804, a third computer readable medium 806, a third processor 808, and an ESP application 810. different and/or additional components can be incorporated into the ESP 708 device.
[00075] The third input interface 800 provides the same or similar functionality as described with reference to the input interface 202 of the model defining device 200, while referring to the ESP device of the interface 708. The third output interface 802 provides the same or similar functionality as described with reference to output interface 204 of model defining device 200 while referring to ESP device 708. Third communication interface 804 provides the same or similar functionality as described with reference to the communication interface 206 of the model defining device 200, although referring to the ESP 708 device. Data and messages may be transferred between the ESP 708 device and the model defining device 200, the control interface device platform 704 and/or display system 710 using a second communication interface 804. The third computer readable medium 806 provides the same or similar functionality as described with reference to the computer readable medium 208 of the model definition device 200, while referring to the ESP device 708. The third processor 808 provides the same or similar functionality as described with reference to processor 210 of model defining device 200, while referring to ESP device 708.
[00076] The ESP 810 application performs operations associated with executing the 416 optimization model operations at or near real time. Some or all of the operations described here can be incorporated into the ESP 810 application. The operations can be implemented using hardware, firmware, software, or any combination of these methods. Referring to the exemplary embodiment of Figure 8, the ESP application 810 is implemented in software (comprised of computer-readable and/or computer-executable instructions) stored on the third computer-readable medium 806 and accessible by the second processor 808 for execution. of instructions that embody the operations of the ESP 810 application. The ESP 810 application can be written using one or more programming languages, assembly languages, scripting languages, etc. The ESP 810 application can be based on the Event Stream Processing Engine developed and supplied by SAS Institute Inc. from Cary, North Carolina, USA.
[00077] Referring to figure 9, examples of operations associated with the ESP 810 application are described. Additionally, fewer operations, or different operations may be performed depending on the embodiment. The order of presentation of operations in Figure 9 is not intended to be limiting. Although some of the operational flows are presented in sequence, the various operations can be performed in several repetitions, at the same time (in parallel, for example, using wires), and/or in orders other than those illustrated.
[00078] In an operation 900, an ESP application instance is demonstrated on the ESP 708 device. In an illustrative embodiment, an engine container is created, which demonstrates an ESP engine (ESPE). The components of an ESPE that run on the ESP 708 device can include one or more projects. A project can be described as a second-level container in an ESPE-managed model, where a segment group size for the project can be defined by a user. The tool container is the top-level container in a model that manages the resources of one or more projects. Each project, of one or more projects, can include one or more continuous queries, also referred to as a model. One or more continuous queries can include one or more source windows and one or more derived windows. For example, in an illustrative embodiment, there may be only one ESPE for each instance of the ESP 810 application. The ESPE may or may not be persistent.
[00079] Continuous query modeling involves defining window driven graphics for handling and transforming the event data stream. A continuous query can be described as a directed graph of source, relational, pattern matching, and procedural windows. One or more source windows and one or more derived windows represent the continuous execution of queries that generate updates to a result set of continuous as new blocks of flow events through the ESPE.
[00080] An event object can be described as a packet of data accessible as a set of fields, with at least one of the fields defined as a key or unique identifier (ID). The event object can be created using a variety of formats, including binary, alphanumeric, XML, etc. Each event object can include one or more fields designated as a primary ID for the event, so that ESPE can support the opcodes for the events, including insert, update, upsert, and delete. Upsert opcodes update the event if the key field already exists, otherwise the event is inserted. For illustration, an event object can be a compressed binary representation of a set of field values and include metadata and field data associated with an event. The metadata can include an opcode indicating whether the event represents an insert, update, delete, or upsert, a set of flags indicating whether the event is a partial, normal update, or a retention generated event from managing the retention policy. retention, and a set of strings (“timestamps”) in microseconds that can be used for latency measurements.
[00081] An event block object can be described as a grouping or package of event objects. A stream of event data can be described as a continuous stream of event block objects. A continuous query of one or more continuous queries transforms a stream of source event data, made from data stream of event block objects published in ESPE, into one or more stream of output event data using one or more windows source and one or more derived windows. A continuous query can also be considered data flow modeling.
[00082] One or more source windows are on top of the directed graph, and have no feed windows within them. Event data streams are published to one or more source windows, and from there, event data streams are routed to the next set of connected windows as defined by the drillhole model created. One or more derived windows are all instantiated windows that are not source windows and that have other dataflow event windows in them. One or more derived windows perform calculations or transformations on input event data streams. One or more derived windows transform event data streams based on window type (i.e. operators such as join, filter, compute, aggregate, copy, pattern match, procedural, join, etc.) and window definitions . As event data streams are published to ESPE, they are continually queried, and the resulting sets of windows derived from these queries are continually updated.
[00083] One or more continuous queries are instantiated by ESPE as a template. For illustration, one or more continuous queries can be defined to apply one or more of the optimization model operations 416 (e.g., operations 504 and 508 of Figure 5) within the ESPE to detected data 412 and/or control data 414 that are transmitted to the ESP device 708 and to the output of the determined target point(s) and the updated rate of the penetration graph to the visualization system 710 and/or the platform control interface device 704 To create a continuous query, input event frames that are keyed schemas that flow into one or more source windows are identified. Output event frames that are also keyed schemas that will be generated by one or more source windows and/or one or more derived windows are also identified. One or more source windows and one or more derived windows are created based on relational pattern matching, and procedural algorithms that transform input event data streams into output event data streams.
[00084] ESPE can analyze and process events in motion or “flow of event data”. As opposed to storing data and performing queries against the stored data, ESPE can store queries and data stream data through it to allow continuous analysis of data as it is received.
[00085] The publish/subscribe capability (pub/inscr) is initialized to the ESPE. In an illustrative embodiment, a pub/inscr capability is initialized for each project of one or more projects. To initialize and enable pub/inscr capability for the ESPE a port number is provided. Pub/inscr clients use the port number to establish pub/inscr connections to ESPE. One or more continuous queries instantiated by ESPE parse and process the inbound event data streams to form the outputs of the outbound event data streams to the event subscription device(s).
[00086] The pub/inscr application programming interface (API) can be described as a library that allows an event publisher such as platform control interface device 704, local data aggregator 706 and/or device template definition 200, can publish the event data stream to the ESPE or an event subscription, such as a visualization system 710 and control interface device 704, to subscribe to the event data stream from the ESPE. The pub/subscribe API provides cross-platform connectivity and sorting compatibility between the ESP 810 application and other network-connected applications. The pub/inscr API can include an ESP object support library so that the event editor or event subscriber can create or handle the events they send or receive, respectively. For example, the platform control interface device 704 can use the pub/inscr API to send a data stream of event blocks (event block objects) to the ESPE, and a visualization system 710 can use the API pub/inscr to receive a stream of event block data from ESPE.
[00087] In an operation 902, one or more event blocks are received by the ESPE that include control data 414 and/or sense data 412. An event block object containing one or more event objects is introduced in a window source from one or more source windows.
[00088] In a 904 operation, event blocks are processed through one or more 416 optimization model operations performed within the ESPE. In an operation 906, the second event blocks are sent to the visualization system 710. For example, the penetration rate graph may be updated and produced in one or more event blocks sent to the visualization system 710 for evaluation by an operator.
[00089] As another example, a control value can be calculated and generated in one or more event blocks and sent, in an operation 908, to the platform control interface device 704, which can control a change in a value target of a control of the control parameters of the drilling operation 228.
[00090] Similar to operation 512, in an operation 910 a determination is made as to whether an update is performed to the objective function model 224 or to the prediction model 225 for any control variables for prediction. If the determination is to perform an update, processing continues at an operation 912. If the determination is not to perform an update, processing continues at operation 902 to continue processing control data 414 and sense data 412 as they are received in real time.
[00091] In operation 912, the project is stopped. In an operation 914, the objective function model 224 and/or the prediction model 225 of the optimization model 416 are updated from the model definition device 200. In an operation 916 the project in the ESPE is restarted with the model definition. updated optimization 416 received from model defining device 200, and processing continues in operation to 902 to continue processing control data 414 and sense data 412 as they are received in real time.
[00092] The word "illustrative" is used herein to mean serving as an example, example, or illustration. Any aspect or design described herein as "illustrative" is not necessarily interpreted as preferred or advantageous over other aspects or designs. In addition, for the purposes of this description and unless otherwise specified, “one” or “an” means “one or more”. Still further, the use of "and" or "or" in the detailed description is intended to include "and/or" unless specifically stated otherwise. Illustrative embodiments may be implemented as a method, device or article of manufacture using standard programming and/or engineering techniques to produce software, firmware, hardware, or any combination thereof to control a computer to implement the described embodiments.
[00093] The above description of illustrative embodiments of the described matter has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the subject matter described to the precise form described, and modifications and variations are possible in light of the above teachings, or may be gained from practice of the subject matter described. The embodiments have been chosen and described in order to explain the principles of the subject matter described and as practical applications of the subject matter described to enable one skilled in the art to use the subject matter described in various embodiments and with various modifications, as appropriate to the particular use contemplated.

权利要求:
Claims (17)
[0001]
1. Method for determining an ideal value for a control of a drilling operation, characterized by the fact that it comprises: - receiving the history of drilling data from a drilling operation, the history of drilling data including a plurality of values measured for each of the plurality of drilling control variables during the previous drilling operation; - determine, through a computing device (208), an objective function model (224) using the history of the received drilling data and a neural network model, and the objective function model (224) maximizes a rate of penetration for the previous drilling operation; - receiving the measured drilling data that includes the current drilling data values for a different drilling operation; - determining, through a computing device (208), an ideal value for a different drilling operation control by executing the objective function model (224) determined with the measured drilling data that includes the values of the drilling data current for different drilling operation as an input; - producing, through a computing device (208), the ideal value determined for the control of the different drilling operation; - receiving the additional drilling data, the additional drilling data including a second plurality of measured values for each of the plurality of drilling control variables, the additional drilling data being measured after the production of the determined ideal value for controlling the drilling operation; - determining, through a computing device (208), a second model of the objective function (224) using the additional drilling data received, whereby the second model of the objective function (224) maximizes a penetration rate of the drilling operation previous; - receiving the second measured drilling data that includes the second values of the current drilling data for a different drilling operation; - determining, through a computing device (208), a second ideal value for controlling the different drilling operation by executing the second objective function model (224) determined with the second measured drilling data that includes the second values from the current drilling data to the different drilling operation as an input; and - producing, through a computing device (208), the second ideal value determined for the control of the different drilling operation.
[0002]
2. Method, according to claim 1, characterized in that the objective function model (224) is determined using a set of models that are integrated, and the neural network model is at least one of the set of models.
[0003]
3. Method according to claim 1, characterized in that a prediction model (225) is determined using the received drilling data history, and the prediction model predicts a value of a control variable input to the model of the objective function (224) determined.
[0004]
4. Method according to claim 3, characterized in that the computer-readable instructions additionally cause the computing device (208) to determine a current prediction value of the control variable input to determine the objective function model ( 224), whereby the objective function model (224) determined is additionally executed with the current prediction value (225) as an input.
[0005]
5. Method according to claim 3, characterized in that the prediction model (225) is determined using a decision tree model.
[0006]
6. Method according to claim 1, characterized in that the objective function model (224) additionally minimizes a specific mechanical energy for the previous drilling operation.
[0007]
7. Method according to claim 6, characterized in that the objective function model (224) additionally optimizes the wellbore stability of the previous drilling operation.
[0008]
8. Method according to claim 1, characterized in that the objective function model (224) is determined by evaluating a plurality of neural network model configurations.
[0009]
9. Method, according to claim 1, characterized in that the determination of the objective function model (224) comprises: - defining a training data set as a first part of the received drilling data history; - defining a validation dataset as a second part of the received drilling data history; - define a first configuration of the neural network model; - training a first neural network model using the defined training dataset based on the first defined neural network model configuration; - predict the model output data with the validation dataset defined as an input to the first trained neural network model; - comparing the output data of the prediction model (225) with the output data of the validation data set; and - determining a first validity rating for the first trained neural network model based on comparing the output data of the prediction model (225) with the validation data set of the output data.
[0010]
10. Method, according to claim 9, characterized in that the determination of the objective function model (224) additionally comprises: - defining a second configuration of the neural network model; - train a second neural network model using the defined training dataset based on the configuration of the second defined neural network model; - predict the output data of the second model with the validation dataset defined as an input to the second trained neural network model; - comparing the output data of the second prediction model (225) with the output data of the validation data set; and - determining a second validity rating for the second trained neural network model based on comparing the output data of the second prediction model (225) with the output data of the validation dataset.
[0011]
11. Method according to claim 10, characterized in that the objective function model (224) is determined as the first trained neural network model or the second trained neural network model based on a comparison between the first classification of determined validity and the second determined validity rating.
[0012]
A method according to claim 1, characterized in that the additional drilling data (225) received includes a third plurality of measured values for each of the plurality of drilling control variables during the different drilling operation.
[0013]
13. Method according to claim 1, characterized in that the objective function model (224) additionally optimizes the wellbore stability of the previous drilling operation.
[0014]
14. Method according to claim 1, characterized in that the determined ideal value is sent to a display device (216).
[0015]
15. Method according to claim 1, characterized in that the determined ideal value is sent to a control device associated with the adjustment of a control target point.
[0016]
16. Method according to claim 1, characterized in that the computing device (208) is physically located on a drilling rig.
[0017]
17. Method according to claim 1, characterized in that the computing device (208) is physically remote from a drilling rig.

类似技术:

---

公开号 | 公开日 | 专利标题

---

BR112016001904B1|2022-01-25|Method for determining an optimal value for a control of a drilling operation

---

US9163497B2|2015-10-20|Fluid flow back prediction

---

US20150354336A1|2015-12-10|Control parameter determination in steam assisted gravity drainage oil drilling

---

US10621500B2|2020-04-14|Completion design optimization using machine learning and big data solutions

---

US20200040719A1|2020-02-06|Machine-Learning Based Drilling Models for A New Well

---

US10495769B2|2019-12-03|System and method for reservoir simulation using on-demand data

---

Bello et al.2017|Next generation downhole big data platform for dynamic data-driven well and reservoir management

---

US10775531B2|2020-09-15|Big data point and vector model

---

CA3040926A1|2018-04-26|Improved stimulation using fiber-derived information and fracturing modeling

---

EP3743595A1|2020-12-02|Optimization of rate-of-penetration

---

US20210026030A1|2021-01-28|Geologic formation operations framework

---

WO2017059152A1|2017-04-06|Self-organizing swarm intelligent wells

---

US20200183044A1|2020-06-11|Computing System Assessment of Geological Similarity of Wells Employing Well-Log Data

---

CA2864265C|2016-07-26|Control parameter determination in steam assisted gravity drainage oil drilling

---

US20210294848A1|2021-09-23|Conversion of well operation information for cloud-native applications

---

US11131184B1|2021-09-28|Method and system for determining a drilling hazard condition using well logs

---

US20210224243A1|2021-07-22|Converting uni-temporal data to cloud based multi-temporal data

---

US20210042633A1|2021-02-11|Aggregation functions for nodes in ontological frameworks in representation learning for massive petroleum network systems

---

US20210389491A1|2021-12-16|Methods and systems for generating graph neural networks for reservoir grid models

---

US20210042634A1|2021-02-11|Representation learning in massive petroleum network systems

---

US20200334398A1|2020-10-22|Petro-Technical Global Fluid Identity Repository

---

EP3915069A1|2021-12-01|Rapid region wide production forecasting

同族专利:

---

公开号 | 公开日

---

US20150227841A1|2015-08-13|

---

HK1217744A1|2017-01-20|

---

US9085958B2|2015-07-21|

---

CN105473812A|2016-04-06|

---

GB2530945A|2016-04-06|

---

BR112016001904A2|2017-08-01|

---

CN105473812B|2017-07-28|

---

MX2016001660A|2016-10-07|

---

US20150081222A1|2015-03-19|

---

MX361510B|2018-12-07|

---

GB2530945B|2017-07-26|

---

CA2916762C|2016-07-26|

---

WO2015042347A1|2015-03-26|

---

CA2916762A1|2015-03-26|

---

NO339353B1|2016-12-05|

---

NO20160083A1|2016-01-14|

---

GB201522616D0|2016-02-03|

---

US10275715B2|2019-04-30|

引用文献:

---

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

---

---

US20090143923A1|2000-09-08|2009-06-04|Breed David S|Arrangement and Method for Monitoring Shipping Containers|

---

US5205166A|1991-08-07|1993-04-27|Schlumberger Technology Corporation|Method of detecting fluid influxes|

---

GB2279381B|1993-06-25|1996-08-21|Schlumberger Services Petrol|Method of warning of pipe sticking during drilling operations|

---

EP0718641B1|1994-12-12|2003-08-13|Baker Hughes Incorporated|Drilling system with downhole apparatus for transforming multiple downhole sensor measurements into parameters of interest and for causing the drilling direction to change in response thereto|

---

US5787986A|1995-03-29|1998-08-04|Halliburton Energy Services, Inc.|Control of particulate flowback in subterranean wells|

---

US6209643B1|1995-03-29|2001-04-03|Halliburton Energy Services, Inc.|Method of controlling particulate flowback in subterranean wells and introducing treatment chemicals|

---

GB9621871D0|1996-10-21|1996-12-11|Anadrill Int Sa|Alarm system for wellbore site|

---

US6234250B1|1999-07-23|2001-05-22|Halliburton Energy Services, Inc.|Real time wellbore pit volume monitoring system and method|

---

US6968909B2|2002-03-06|2005-11-29|Schlumberger Technology Corporation|Realtime control of a drilling system using the output from combination of an earth model and a drilling process model|

---

CA2493091C|2002-07-26|2008-12-30|Varco I/P, Inc.|Automated rig control management system|

---

US6820702B2|2002-08-27|2004-11-23|Noble Drilling Services Inc.|Automated method and system for recognizing well control events|

---

US6832650B2|2002-09-11|2004-12-21|Halliburton Energy Services, Inc.|Methods of reducing or preventing particulate flow-back in wells|

---

US7128167B2|2002-12-27|2006-10-31|Schlumberger Technology Corporation|System and method for rig state detection|

---

US7899657B2|2003-01-24|2011-03-01|Rockwell Automoation Technologies, Inc.|Modeling in-situ reservoirs with derivative constraints|

---

US6892813B2|2003-01-30|2005-05-17|Halliburton Energy Services, Inc.|Methods for preventing fracture proppant flowback|

---

GB2417792B|2003-03-31|2007-05-09|Baker Hughes Inc|Real-time drilling optimization based on mwd dynamic measurements|

---

US8209202B2|2005-04-29|2012-06-26|Landmark Graphics Corporation|Analysis of multiple assets in view of uncertainties|

---

EA012893B1|2005-08-19|2009-12-30|Эксонмобил Апстрим Рисерч Компани|Method and apparatus associated with stimulation treatments for wells|

---

US7621324B2|2006-03-30|2009-11-24|Don Atencio|Automated flowback and information system|

---

US7447611B2|2006-05-09|2008-11-04|Hsb Solomon Associates, Llc|Power generation performance analysis system and method|

---

US20080255892A1|2007-04-11|2008-10-16|The University Of Southern California|System and Method for Oil Production Forecasting and Optimization ina Model-Based Framework|

---

US7814989B2|2007-05-21|2010-10-19|Schlumberger Technology Corporation|System and method for performing a drilling operation in an oilfield|

---

US7966273B2|2007-07-27|2011-06-21|Schlumberger Technology Corporation|Predicting formation fluid property through downhole fluid analysis using artificial neural network|

---

US20090076873A1|2007-09-19|2009-03-19|General Electric Company|Method and system to improve engineered system decisions and transfer risk|

---

US7660673B2|2007-10-12|2010-02-09|Schlumberger Technology Corporation|Coarse wellsite analysis for field development planning|

---

US8244654B1|2007-10-22|2012-08-14|Healthways, Inc.|End of life predictive model|

---

US8121971B2|2007-10-30|2012-02-21|Bp Corporation North America Inc.|Intelligent drilling advisor|

---

US8417495B2|2007-11-07|2013-04-09|Baker Hughes Incorporated|Method of training neural network models and using same for drilling wellbores|

---

US8274399B2|2007-11-30|2012-09-25|Halliburton Energy Services Inc.|Method and system for predicting performance of a drilling system having multiple cutting structures|

---

BRPI0821118B1|2007-12-14|2018-11-06|Prad Research And Development Limited|method of completing a well, method of treating an underground formation intercepted by a well, using changeable additives, and method|

---

US8527248B2|2008-04-18|2013-09-03|Westerngeco L.L.C.|System and method for performing an adaptive drilling operation|

---

US7966166B2|2008-04-18|2011-06-21|Schlumberger Technology Corp.|Method for determining a set of net present values to influence the drilling of a wellbore and increase production|

---

US8256534B2|2008-05-02|2012-09-04|Baker Hughes Incorporated|Adaptive drilling control system|

---

GB2460096B|2008-06-27|2010-04-07|Wajid Rasheed|Expansion and calliper tool|

---

US20100076740A1|2008-09-08|2010-03-25|Schlumberger Technology Corporation|System and method for well test design and interpretation|

---

AU2009300240B2|2008-10-03|2013-02-21|Halliburton Energy Services, Inc.|Method and system for predicting performance of a drilling system|

---

US7861800B2|2008-10-08|2011-01-04|Schlumberger Technology Corp|Combining belief networks to generate expected outcomes|

---

MX2011003997A|2008-10-14|2011-08-12|Schlumberger Technology Bv|System and method for online automation.|

---

NO338750B1|2009-03-02|2016-10-17|Drilltronics Rig Systems As|Method and system for automated drilling process control|

---

US8170800B2|2009-03-16|2012-05-01|Verdande Technology As|Method and system for monitoring a drilling operation|

---

CA2767689C|2009-08-07|2018-01-02|Exxonmobil Upstream Research Company|Drilling advisory systems and methods based on at least two controllable drilling parameters|

---

WO2011016927A1|2009-08-07|2011-02-10|Exxonmobil Upstream Research Company|Drilling advisory systems and methods utilizing objective functions|

---

BR112012002907A2|2009-08-12|2016-04-05|Exxonmobil Upstream Res Co|method and system for optimizing field operational policy for a subsurface region, method for optimizing an optimization problem over time, and for decision support considering oil resource development, and, computer readable storage medium.|

---

US9317537B2|2009-09-25|2016-04-19|Adnan Fakeih|Database and method for evaluating data therefrom|

---

US8818779B2|2009-12-21|2014-08-26|Baker Hughes Incorporated|System and methods for real-time wellbore stability service|

---

US8453764B2|2010-02-01|2013-06-04|Aps Technology, Inc.|System and method for monitoring and controlling underground drilling|

---

US7998911B1|2010-03-02|2011-08-16|Oil Chem Technologies|Environmental friendly fracturing and stimulation composition and method of using the same|

---

US8646523B2|2010-03-15|2014-02-11|Baker Hughes Incorporated|Method and materials for proppant flow control with telescoping flow conduit technology|

---

US20120000658A1|2010-06-30|2012-01-05|Chevron U.S.A. Inc.|Method for in situ fluid assessment and optimization during wellbore displacement operations|

---

CN103038606B|2010-07-30|2016-03-30|埃森哲环球服务有限公司|Intelligent core engine|

---

US8733443B2|2010-12-21|2014-05-27|Saudi Arabian Oil Company|Inducing flowback of damaging mud-induced materials and debris to improve acid stimulation of long horizontal injection wells in tight carbonate formations|

---

CA2829378A1|2011-03-09|2012-09-13|Prad Research And Development Limited|Method for charcterizing subsurface formations using fluid pressure response during drilling operations|

---

US8695692B2|2011-07-29|2014-04-15|Baker Hughes Incorporated|Downhole condition alert system for a drill operator|

---

US9181792B2|2011-10-05|2015-11-10|Schlumberger Technology Corporation|Method for detecting and mitigating drilling inefficiencies|

---

MX2014003947A|2011-10-06|2014-08-21|Landmark Graphics Corp|Systems and methods for subsurface oil recovery optimization.|

---

US8752648B2|2011-11-02|2014-06-17|Landmark Graphics Corporation|Method and system for predicting a drill string stuck pipe event|

---

US8210283B1|2011-12-22|2012-07-03|Hunt Energy Enterprises, L.L.C.|System and method for surface steerable drilling|

---

US10385260B2|2012-01-12|2019-08-20|Ecolab Usa Inc.|Fracturing fluids including amine oxides as flowback aids|

---

US9103191B2|2012-03-02|2015-08-11|Schlumberger Technology Corporation|Master plan for dynamic phase machine automation system|

---

WO2013135288A1|2012-03-14|2013-09-19|Statoil Canada Limited|System and method for controlling the processing of oil sands|

---

US9701888B2|2012-03-27|2017-07-11|Ecolab Usa Inc.|Microemulsion flowback aid composition and method of using same|

---

US9255473B2|2012-05-07|2016-02-09|Halliburton Energy Services, Inc.|Methods and systems for real-time monitoring and processing of wellbore data|

---

US9784100B2|2012-06-01|2017-10-10|Baker Hughes Incorporated|Smart flowback alarm to detect kicks and losses|

---

US9022140B2|2012-10-31|2015-05-05|Resource Energy Solutions Inc.|Methods and systems for improved drilling operations using real-time and historical drilling data|

---

US9657558B2|2012-12-28|2017-05-23|Schlumberger Technology Corporation|Method for treating and measuring subterranean formations|

---

US9243486B2|2013-02-25|2016-01-26|Baker Hughes Incorporated|Apparatus and method for determining closure pressure from flowback measurements of a fractured formation|

---

US9085958B2|2013-09-19|2015-07-21|Sas Institute Inc.|Control variable determination to maximize a drilling rate of penetration|

---

US20150354336A1|2014-06-05|2015-12-10|Sas Institute Inc.|Control parameter determination in steam assisted gravity drainage oil drilling|

---

US20160017696A1|2014-07-18|2016-01-21|Sridhar Srinivasan|Determining One or More Parameters of a Well Completion Design Based on Drilling Data Corresponding to Variables of Mechanical Specific Energy|US9285794B2|2011-09-07|2016-03-15|Exxonmobil Upstream Research Company|Drilling advisory systems and methods with decision trees for learning and application modes|

---

US9734290B2|2011-12-16|2017-08-15|Neela SRINIVAS|System and method for evidence based differential analysis and incentives based healthcare policy|

---

US9085958B2|2013-09-19|2015-07-21|Sas Institute Inc.|Control variable determination to maximize a drilling rate of penetration|

---

US9163497B2|2013-10-22|2015-10-20|Sas Institute Inc.|Fluid flow back prediction|

---

US9579700B2|2014-05-30|2017-02-28|Iteris, Inc.|Measurement and modeling of salinity contamination of soil and soil-water systems from oil and gas production activities|

---

US20150354336A1|2014-06-05|2015-12-10|Sas Institute Inc.|Control parameter determination in steam assisted gravity drainage oil drilling|

---

US10221671B1|2014-07-25|2019-03-05|U.S. Department Of Energy|MSE based drilling optimization using neural network simulaton|

---

CA2971706A1|2015-03-05|2016-09-09|Halliburton Energy Services, Inc.|Method to optimize oilfield operations based on large and complex data sets|

---

CA2977943C|2015-04-01|2021-04-13|Landmark Graphics Corporation|Model generation for real-time rate of penetration prediction|

---

US10577894B1|2015-06-08|2020-03-03|DataInfoCom USA, Inc.|Systems and methods for analyzing resource production|

---

CN105302858B|2015-09-18|2019-02-05|北京国电通网络技术有限公司|A kind of the cross-node enquiring and optimizing method and system of distributed data base system|

---

US9593568B1|2015-10-09|2017-03-14|General Electric Company|System for estimating fatigue damage|

---

NO341053B1|2016-01-26|2017-08-14|Exebenus AS|A method for planning and executing real time automated decision support in oil and gas wells|

---

US10650046B2|2016-02-05|2020-05-12|Sas Institute Inc.|Many task computing with distributed file system|

---

US10095552B2|2016-02-05|2018-10-09|Sas Institute Inc.|Automated transfer of objects among federated areas|

---

US10795935B2|2016-02-05|2020-10-06|Sas Institute Inc.|Automated generation of job flow definitions|

---

US10338968B2|2016-02-05|2019-07-02|Sas Institute Inc.|Distributed neuromorphic processing performance accountability|

---

US10642896B2|2016-02-05|2020-05-05|Sas Institute Inc.|Handling of data sets during execution of task routines of multiple languages|

---

US10650045B2|2016-02-05|2020-05-12|Sas Institute Inc.|Staged training of neural networks for improved time series prediction performance|

---

US10352142B2|2016-09-26|2019-07-16|International Business Machines Corporation|Controlling operation of a stem-assisted gravity drainage oil well system by adjusting multiple time step controls|

---

US10570717B2|2016-09-26|2020-02-25|International Business Machines Corporation|Controlling operation of a steam-assisted gravity drainage oil well system utilizing continuous and discrete control parameters|

---

US10577907B2|2016-09-26|2020-03-03|International Business Machines Corporation|Multi-level modeling of steam assisted gravity drainage wells|

---

US10614378B2|2016-09-26|2020-04-07|International Business Machines Corporation|Cross-well allocation optimization in steam assisted gravity drainage wells|

---

US10378324B2|2016-09-26|2019-08-13|International Business Machines Corporation|Controlling operation of a steam-assisted gravity drainage oil well system by adjusting controls based on forecast emulsion production|

---

US10267130B2|2016-09-26|2019-04-23|International Business Machines Corporation|Controlling operation of a steam-assisted gravity drainage oil well system by adjusting controls to reduce model uncertainty|

---

EP3552125A4|2016-12-07|2020-08-19|Safekick Americas LLC|Automated model-based drilling|

---

CN106837295B|2017-01-25|2020-04-07|河南理工大学|Intelligent safe and efficient automatic drilling control system and control method|

---

EP3385497B1|2017-04-04|2019-12-25|VAREL EUROPE |Method of optimizing drilling operation using empirical data|

---

US10990882B2|2017-07-28|2021-04-27|International Business Machines Corporation|Stratigraphic layer identification from seismic and well data with stratigraphic knowledge base|

---

US20200190957A1|2017-08-18|2020-06-18|Landmark Graphics Corporation|Rate of penetration optimization for wellbores using machine learning|

---

AU2017428353A1|2017-08-21|2020-01-30|Landmark Graphics Corporation|Neural network models for real-time optimization of drilling parameters during drilling operations|

---

WO2019051435A1|2017-09-11|2019-03-14|Schlumberger Technology Corporation|Well planning system|

---

US10616121B2|2017-11-16|2020-04-07|Distech Controls Inc.|Inference server and environment control device for inferring an optimal wireless data transfer rate|

---

WO2019147297A1|2018-01-29|2019-08-01|Landmark Graphics Corporation|Controlling range constraints for real-time drilling|

---

CN110397402B|2018-04-23|2021-01-29|中国石油天然气股份有限公司|Drilling method and device|

---

CN108756848B|2018-05-21|2019-05-21|北京四利通控制技术股份有限公司|A kind of intelligent drilling control cloud platform and intelligent drilling control system|

---

TWI684080B|2018-06-21|2020-02-01|高聖精密機電股份有限公司|Smart adjustment system and method thereof|

---

US20210324723A1|2018-08-30|2021-10-21|Landmark Graphics Corporation|Automated Rate of Penetration Optimization for Drilling|

---

CN109839825A|2019-01-28|2019-06-04|华东交通大学|A kind of forecast Control Algorithm and system of Rare-Earth Extraction Process constituent content|

---

GB202200025D0|2019-08-09|2022-02-16|Landmark Graphics Corp|Model based preference learning and optimization systems and methods|

---

WO2021040829A1|2019-08-23|2021-03-04|Landmark Graphics Corporation|Active reinforcement learning for drilling optimization and automation|

---

US11055639B1|2020-04-28|2021-07-06|Sas Institute Inc.|Optimizing manufacturing processes using one or more machine learning models|

法律状态:
2018-11-06| B06F| Objections, documents and/or translations needed after an examination request according [chapter 6.6 patent gazette]|

---

2020-05-12| B06U| Preliminary requirement: requests with searches performed by other patent offices: procedure suspended [chapter 6.21 patent gazette]|

---

2021-11-16| B09A| Decision: intention to grant [chapter 9.1 patent gazette]|

---

2022-01-25| B16A| Patent or certificate of addition of invention granted [chapter 16.1 patent gazette]|Free format text: PRAZO DE VALIDADE: 20 (VINTE) ANOS CONTADOS A PARTIR DE 19/09/2014, OBSERVADAS AS CONDICOES LEGAIS. |

优先权:

---

申请号 | 申请日 | 专利标题

---

US201361879933P| true| 2013-09-19|2013-09-19|

---

US61/879,933|2013-09-19|

---

US14/490,189|US9085958B2|2013-09-19|2014-09-18|Control variable determination to maximize a drilling rate of penetration|

---

US14/490,189|2014-09-18|

---

PCT/US2014/056455|WO2015042347A1|2013-09-19|2014-09-19|Control variable determination to maximize a drilling rate of penetration|

---