MACHINE LEARNING BASED CLOSED LOOP FRACTURING MODELS

Description

TECHNICAL FIELD

Some implementations relate to subsurface wells. More specifically, some implementations relate to controlling hydraulic-fracturing-related operations in the well.

BACKGROUND

Hydraulic fracturing (commonly known as “fracking”) is a well-stimulation technique used in the oil and gas industry to enhance the extraction of hydrocarbons from low-permeability rock formations (such as shale). The process may involve injecting a high-pressure mixture of water, sand, and chemicals into a wellbore to create small fractures in the rock. These fractures may significantly increase production rates by enabling oil and gas to flow more freely into the well. The sand (referred to as “proppant”) may create a continuous flow of hydrocarbons by keeping fractures open after the pressure is released.

Some fracking operations may involve multiple wells and various sensors that measure conditions in and around the wells. These measurements may be used by processes that model and control fracking operations in the wells.

BRIEF DESCRIPTION OF THE DRAWINGS

Implementations of the disclosure may be better understood by referencing the accompanying drawings.

FIG. 1 is a diagram illustrating associations between measurements and wells.

FIG. 2 is a diagram illustrating a process for tuning a physics-based fracture model.

FIG. 3 is a diagram illustrating a process for using a tuned physics-based fracture model to predict targets.

FIG. 4 is a diagram illustrating operations for training and utilizing a machine-learning model to control subsurface operations.

FIG. 5 is a diagram showing operations for closed loop control of fracking operations.

FIG. 6 is a diagram illustrating operations for modeling and remediating hydraulic fracturing operations.

FIG. 7 is a diagram illustrating an example artificial neural network included in some implementations.

FIG. 8 is a block diagram illustrating a computer system that may be utilized with some implementations.

FIG. 9 is an illustration depicting an example multi-well system, according to some implementations.

DESCRIPTION OF IMPLEMENTATIONS

The description that follows may include example systems, methods, techniques, and program flows that embody implementations of the disclosure. However, this disclosure may be practiced without these specific details. For clarity, some well-known instruction instances, protocols, structures, and techniques may not be shown in detail.

Overview

The technological development of both surface and downhole installed sensor-based systems has provided a significant uplift in how oil and gas operators monitor hydraulic fracture stimulations. These sensor systems may include, but are not limited to, pressure and temperature gauges, accelerometers, geophones, hydrophones, and distributed acoustic sensing acquisitions.

With the increase in sensor data recorded and stored for hydraulic fracture stimulations across North American unconventional basins, there may be a desire to utilize the derived answer products to optimize the completions in both a real-time and pseudo real-time manner. Some implementations develop machine-learning model workflows, in a regression and classification sense, and generate recommender systems that propose changes based on ML responses to well stimulation parameters at the surface. The learnings derived from the machine-learning models may enable focused extraction of physics-based fracture modeling insights to provide the recommender system with hybrid optimized stimulation changes to ensure the most consistent frac on a stage-by-stage basis.

Some implementations generate models based on far field and near field stimulation effectiveness and provide feedback/training of the models. In some implementations, the models are classification and anomaly detection models configured to detect key temporal features which require remediation efforts in the data with respect to the previously trained models. The models may then utilize a recommender system (such as a digital twin) which passes an optimized completion strategy to the automated fracturing process. As more field data (such as answer product metrics) are collected with remediation strategies applied, more model training and refinement will be implemented so the remediation strategies may become increasingly effective (such as by avoiding non-uniqueness in strategies).

EXAMPLE IMPLEMENTATIONS

Closed Loop Hydraulic Fracking System for Controlling Near Field and Far Field Completion

Some implementations may measure fracture characteristics using one or more far field sensors. For example, some implementations may perform and/or utilize strain measurements from offset/observation wells using fiber optics sensors. Some implementations may perform and/or utilize seismic measurements from offset/observation wells using fiber optic sensors and/or geophones. Some implementations may perform and/or utilize surface seismic measurements. Some implementations may perform and/or utilize pressure measurements from offset wells or a parent wells. Some implementations may perform and/or utilize tiltmeter measurements from tiltmeters and/or other sensors. Any suitable type of measurements and/or sensors may be used in conjunction with the implementations described herein.

Various measurement technologies may be utilized in some implementations. Some technologies (such as fiber, tiltmeters, and others) may measure spatial distribution of the signal over time whereas other technologies (such as pressure sensing technologies) may represent one value per observation well. If there are multiple nearby wells, a spatial distribution on pressure measurements may be constructed. These measurements may be used for determining some fracture characteristics of interest such as length of the fractures, spatial dimension of the subsurface region being treated, and others. In some implementations, a pressure response measured from an offset well may be utilized together with a poro-elastic model to obtain fracture model. Similarly, strain measurements may be utilized to obtain fracture dimensions using fracture propagation models. When multiple measurements are available more information may be inferred. For example, some implementations may utilize pressure responses from an offset well and strain measurements to infer length, height, and width of the fracture and the separation between the fractures. Moreover, some implementations may utilize multiple measurements to constrain a solution. These implementations may additionally depend on the other inputs such as total volume of the fluid pumped, maximum number of clusters available, observed treatment pressure, well geometry, and others.

Measurements also may be used to detect an event such as a fracture hit, fracture reaching depleted zone, fracture crossing a location, and other events. Some implementations may detect events by looking at changes in trends such as changes of slope. Alternatively, when the time history of these measurements is available a rate can also be determined (such as the rate of fracture propagation or the volume of fluid lost to the depleted zone). In addition to measuring fracturing characteristics (as described above), some implementations also may measure characteristics in one or more treatment wells using sensors in or about the one or more treatment wells.

The measurements may include in-well pressure measurements (such as surface/bottom hole or anywhere along the well), outside well pressure measurements (such as those using outside facing pressure sensors), and others. The sensors also may include fiber (such as DAS/DTS), cameras, vision sensors, and others. Any other type of measurement may be used. Some of the measured signals may be used to derive secondary features. For example, acoustic measurement may be used to compute uniformity index-measure of flow distribution, dominant clusters active clusters, and more. Pressure measurements may provide total number of active perforation holes, total number of active clusters, flow resistance, near well conductivity, and more. Outside facing pressure sensors can detect any stage-to-stage interaction (such as due to cement or plug loss) or any poro-elastic response.

After determining measurements that may be used to indicate aspects of one or more wells, some implementations may determine a set of features that may be utilized by a fracture model. The features may be based on one or more of the measurements. The features may include operational parameters such as flow rate, pressure, proppant (such as time history and cumulative), wellbore geometry, cluster configuration, far-field measurements from far field sensors, and more. In some implementations, the features may include temporally variant features and may include and/or be based on one or more of rolling statistical measures, derivatives, cumulative summations, lag features, Fourier information, wavelet information, entropy, and amplitude ratios. Additionally, each far field sensor may be associated with an attribute indicating the far field sensor's association with a parent well or offset well. For example, a pressure measurement may indicate a parent well or offset well for which the pressure measurement was performed. Fiber measurements may include a similar indication. For other surface measurements (such as micro-seismic or tiltmeter), there may be an indicator indicating a distance between the sensor and a parent well and/or offset well. Thus, in general, each measurement may carry two attributes-distance to the parent well and distance to the offset well. If the measurement is performed on a parent well, the distance to the parent well would be zero and distance to the offset well would be the distance to the nearest offset well (if such an offset well exits). The feature list may be further augmented by any derived quantities such as fracture length, propagation status, events (time series of true or false indicators) such as frac hits, and more. In some implementations, these derived features carry parent/offset association directly from the sensor involved in computing them. If multiple measurements are involved, then a weighted sum of the distance factor may be computed (such as by computing the average distance). Additionally, all the far field measurements and/or derived quantities may include distance to the treatment well.

FIG. 1 is a diagram illustrating associations between measurements and wells. FIG. 1 shows a treatment well 102, an offset well 104, and a parent well 106, and a sensor 108. FIG. 1 shows three measurements −x_T, x₀, and x_p. Each of these measurements may be associated with a tuple. In the tuple, the first entry may be the distance to the treatment well 102. In the tuple, the second entry may be the distance to the offset well 104, and the third entry may be the distance to the parent well 106. Any other system of distance representation can be utilized.

After determining the set of features, some implementations may define a target function. The target function may involve one or more near well objectives and one or more far field objectives. The near well objectives may include achieving a higher number of active perforation holes, improving fluid distribution, preventing screen-out, maintaining the pressure under a threshold, and more. The far field objectives may include achieving certain fracture dimensions, minimizing fluid loss to the depleted zone, and more. A weighted combination of these objectives may be provided as a target to achieve. The weighting function may indicate the importance of the objective function. For example, avoiding screen-out may have the highest priority as this can potentially stop the fracking operation. Thus, the weighting function may cause some implementations to achieve the highest production (far field objective) while eliminating the screen-out scenario (near field objective). As another example, some implementations may achieve uniform flow distribution (near well objective) while not reaching the depletion zone (far field objective).

Some implementations may create a fracture model based on the measurements, features, target function, and other aspects described herein. Some implementations create a physics-based fracture model. However, any other physics-based models may be used. In some implementations, the unknown parameters include flow distribution parameters (such as discharge coefficient), geomechanical properties (such as Youngs modulus and poisons ratio), and others. The unknown parameters may be tweaked until the fracture model's estimate matches the measurement. The set of unknown variables may vary based on the available measurements. For example, if fiber measurements are available, the active number of clusters and their flowrates will be known. Otherwise, these may be considered as unknown parameters. At that point, the geometry from the fracture model represents the generated fracture geometry.

FIG. 2 is a diagram illustrating a process for tuning a physics-based fracture model. The process 200 may include a fracture model 206 receiving known parameters 202 such as wellbore geometry, treatment data, and/or other known parameters. The fracture model 206 also may receive unknown parameters 204 such as discharge coefficient, geomechanical/reservoir properties, and other unknown parameters. The fracture model 206 may produce predicted signals based on the known and unknown parameters. The fracture model 206 also may compare the predicted signals 210 with measured signals 208. If the predicted signals 210 do not match the measured signals 208, values for the unknown parameters 204 may be adjusted.

After tuning the fracture model (such as by the process described with reference to FIG. 2), the tuned fracture model may be utilized to predict a desirable target. For example, if production is the target, the fracture geometry and parameters affecting the conductivity from the model can be provided to a reservoir model and thus the production can be estimated. If the estimated production does not match the target, the tuned model may be utilized to adjust inputs until the predicted results are within acceptable limits. These new inputs may be executed in the field to achieve the target. On the other hand, if the objective is to avoid the frac hit event, the inputs may be adjusted to avoid a frac-hit. Then, the plan may be executed. If the objective function involves multiple constraints (such as maximizing the production without causing the screen-out), both the tuned fracture model and a reservoir model may be combined together to deliver optimal corrections.

FIG. 3 is a diagram illustrating a process for using a tuned physics-based fracture model to predict targets. The process 300 may include a tuned fracture model 306 receiving inputs 302 such as wellbore geometry, treatment data, and/or other known parameters. The tuned fracture model 306 also may receive tuned parameters 304 (such as discharge coefficient, geomechanical/reservoir properties, and other unknown parameters). The tuned fracture model 306 may produce predicted outputs 310 based on the inputs and tuned parameters. The tuned fracture model 306 may compare the predicted outputs 310 with target values 308. If the predicted outputs 310 do not match the target values 308, values for the inputs 302 may be adjusted.

Some implementations may utilize a machine-learning model. In some cases, the machine-learning model may be used instead of a physics-based model. The machine-learning model may utilize the features described herein. The machine-learning model also may acquire objectives (as described herein) such as goals to accomplish and current values for them.

FIG. 4 is a diagram illustrating operations for training and utilizing a machine-learning model to control subsurface operations. In FIG. 4, the machine-learning model 402 may be configured with configuration data 404 including features and objectives. The machine-learning model 402 also may be configured to take actions 406 to achieve the objectives. The machine-learning model 402 may be trained with historical data such as any of the data described herein.

Any of the models described herein may be utilized in a process for closed loop control of fracking operations. FIG. 5 is a diagram showing operations for closed loop control of fracking operations. In FIG. 5, the process 500 begins at block 502 by starting a well treatment. At block 504, the process measures one or more values of one or more objective variables such as by using the sensors and/or a combination of models. For example, the current objective could be a fracture length not reaching a certain location. As another example, the objective could be keeping the volume lost to a depletion zone below a certain value. In these examples, current objective variables (F_i) (see also FIG. 4) may represent fracture length, or the volume lost to the depleted zone inferred using the far field measurements coupled with some models.

At block 506, the process compares one or more current objective variables with one or more target variables 507. At block 508, if the current variable(s) is/are below a threshold, the process continues at block 514. If the current variable(s) is/are not below the threshold, the process continues at block 510. At block 510, a machine-learning model or tuned physics-based model may apply actions in the well (block 512). The machine-learning model or tuned physics-based model may be configured with the current features and current and target objective variables.

At block 514, the process may continue with the fracking treatment.

Some implementations may perform the following workflow. Operation 1: Identify the available near and far field measurements. Some implementations may include a table of near and far field measurement technologies with a value of 1 or 0, where 1 indicates it is available and 0 indicates it is unavailable. There may not be any far field measurement. There may not be any near field measurements. Some implementations may present a user interface where the user can manually enter the available measurements or may automatically derive what is available or a combination of both.

Operation 2: Collect the near field and far objectives and their weights. Generate objective vs needed measurement mapping. Table 1 shows an example measurement mapping. From the mapping, determine all possible objectives that can be accomplished by knowing the available measurements from operation 1.

	TABLE 1
	Objective	Measurement
	Stage	Water hammer
	efficiency
	Uniformity	Treatment well fiber
	index
	Fracture	Offset well pressure and/or offset well
	dimension	fiber and/or surface microseismic
		and/or tiltmeter, surface pressure
	Volume	Parent well pressure and/or surface
	lost to	micro-seismic and/or tilt-meter
	depletion
	Prevent	Treatment pressure or treatment
	screen-out	well fiber
	Minimize	Treatment pressure
	horsepower
	Prevent	Offset well fiber and/or offset well
	well	pressure and/or surface micro-seismic
	interaction	and/or tilt-meter
	. . .	. . .

From the available measurements of operation 1, determine which near and far field objectives are achievable. For example, if there is no measurement available on the parent well any far field objectives related to the depletion zones are removed. Similarly, if there is no water hammer or fiber data available on the treatment well, the flow distribution related objectives from the near field are dropped. Determine the weight (or ranking) for the possible objectives from the user for the achievable objectives. If there is more than one sensor available to achieve a given objective, the model may use extra information to improve the fidelity of solution by constraining the solution.

Operation 3: From the available measurements and required objectives to meet, determine adequate physics based or ML based model.

Operation 4: Using material constraint (such as total volume of fluid to be pumped, total proppant mass, total available chemicals, available stage length, perf shots, and others) (the constraints could be per well or per pad) provide a baseline design.

Operation 5: Using the selected model and available measurements, perform closed loop control for attainable near field and far field objectives during the treatment or between the treatments to enforce adherence to the objectives.

In some implementations, a user can also infer from operation 2 a minimal number of measurements needed to achieve the objective or most economical means of obtaining the objective. For example, a user may decide to go with surface microseismic, instead of installing pressure sensor on parent well, and fiber and/or pressure sensor on offset well if he wants to control fracture dimension and volume lost to depletion.

Downstream Closed Loop Fracture Model

By employing a cascading workflow of trained machine learning models, some implementations automatically optimize hydraulic fracturing processes which lead to near-wellbore and far field uniformity. Some implementations extract defining temporal features of hydraulic fracturing processes. With the extraction of defining temporal features, some implementations identify inefficiencies in the hydraulic fracturing process and determine mitigation plans via a fracture model. After the mitigation strategies are defined, a separate classification machine learning model may be generated to probabilistically identify the inefficiencies in real-time and call a digital twin model of the fracture model to enable a set of treatment adjustments to mitigate the inefficiencies. Some implementations refine and update this process as more data becomes available and as more field trials are performed using the mitigation techniques defined by the fracture model.

In some implementations, the extraction of features which are critical to inefficiencies in the hydraulic fracturing process are determined via a supervised learning process. The supervised learning process may contain several time variant and global features related to surface pumping parameters. The global features may include spread pressure, rate, proppant concentration, elapsed stage timing, perforation diameter, and total number of perforations. After determining the controlling features and their respective effects on near and far field uniformity, some implementations detect and classify the features in a real-time/pseudo real-time manner to apply remediation efforts.

The generation of data for classification model training may come from clusters defined during a supervised learning regression analysis, and each population will be augmented using forward modelling processes. By implementing a supervised learning classification model that operates with either time series or tabular data (such as a transformer, convolutional neural network, gradient booster, and others), the model may build representations of the time series or tabular data and generate probabilistic predictions of their existence in real-time.

After the real-time classification model generates a positive prediction of a class existence, a trigger may be initiated which pushes the class specific information to a digital twin model to receive a tuned remediation procedure. The remediation procedures may be messaged to the automated fracturing processing, which will in turn implement them. The remediation may be logged both in time of occurrence and the specifics of the recommended fracturing changes. By logging a variety of remediation procedures and recording the sensor-based answer products, the supervised learning model (which extracts features and generates corresponding remediation procedures) can be tuned with constant training.

FIG. 6 is a diagram illustrating operations for modeling and remediating hydraulic fracturing operations. At block 601, some implementations acquire a database of clustered temporal features with largest importance and inverted remediation strategies for use as inputs for machine-learning classification. Some implementations may generate the database from which the features are determined and/or by which any one or more of the models described herein are trained. In some implementations, such data may be generated by only physics-based models, only by AI models, and/or by hybrid AI and physics-based models. In some implementations, such data is only field-acquired data such as surface pumping spread data.

At block 602, some implementations define machine-learning model architectures which utilize time variant input features and/or discrete features which emphasize the largest derived feature importance is (transformers, gradient boosting, and others). The features may relate to real-time pumping data. The features may include spread pressure, rate, proppant concentration, elapsed stage timing, perforation diameter, total number of perforations, and others.

At block 603, some implementations generate probabilistic classification models which detect various clustered feature presences in real-time pumping data. Time and class of detection may be extracted.

At block 604, some implementations utilize class and time of extracted features from the classification model as inputs to the trained digital twin to produce recommended remediation strategies (such as hydraulic fracturing operation to be performed in or about one or more wells).

At block 605, some implementations pass remediation strategies to an automated fracking controller for implementation in the fracturing process.

At block 606, some implementations utilize real-time/pseudo real-time in situ answer products to observe effects of remediation strategies in both/either near and far field.

At block 607, some implementations refine feature extraction and remediation recommendations based on newly recorded answer product measurements. Some implementations also update the digital twin for optimal implementation strategy (such as by reducing non-uniqueness and implementation strategy). In some implementations, this entails retraining of the classification model to attend to the newly refined features of interest in the continual training and deployment cycle (as the flow 600 loops in a plurality of iterations).

In summary, by utilizing machine learning temporal feature extraction methodologies, key time series metrics can be understood from a physical/data-driven perspective and in many cases been together based on the sensor data answer product responses. The clusters of extracted timeseries responses may be used to generate an augmented supervised learning training data set on which timeseries classification models will be created. The real-time classification model may seek to observe any key phenomenon in the sensor response and positive identification of a particular class may enable the call to the generated recommender system (digital twin) to provide applicable remediation strategies. After the strategies have been implemented, the timing and strategies may be logged and the corresponding sensor data answer products may be observed. Based on the response from answer products, the recommender system model (and the precursor feature extraction model) may be updated. This in turn may call for a retraining of the classification model to attend to the newly refined features of interest in the continual training and deployment cycle.

Example Computer Systems

Some implementations include machine-learning models. Machine-learning modelling may include machine-implemented algorithms that learn from and make predictions (or decisions) based on input data. The input data may be collected from a variety of sources. The input data may include time series, image, text, and more. The input data may be cleaned via a suite of signal processing routines before engineered features are extracted and tailored to the problem at hand.

In some implementations, the machine-learning models may be trained with input data that spans the entire solution space, thereby creating a generalized model that produces accurate results based on input data that was not used in training. Input data used for training machine-learning models may include field data, synthetic data, artificial intelligence (AI) generated field data, and more. Field Data may include data that was recorded by sensors of any suitable type, whether they be installed at the surface or subsurface. The field data may be labeled in terms of the current time state (stage number, well name, etc.) or may be unlabeled depending on whether supervised or unsupervised learning is applied, respectively. The field data may either be raw sensor data or have a variety of signal processing applied prior to any feature engineering processes or machine learning model training.

Synthetic Data may be generated by physics-based models that produce pseudo sensor data to be used in machine-learning model training. This data may be utilized in generating a suite of physical scenarios, where the machine-learning modelling may be targeted at speeding up the inversion process (digital twin development). AI-generated field data may have been developed using means such as generative AI and other similar techniques. Recorded sensor field data may be input into one or more generative AI models whose output is synthetically derived field data (also referred to as AI-generated field data). The synthetically derived field data may be built in massive quantities and used to train subsequent machine learning models.

AI generated hybrid field-synthetic data may be developed using means such as generative AI and other similar techniques. Recorded sensor field data may include inputs into one or more generative AI models whose output is hybrid field-synthetic data. An underlying physics model (included in the generative AI model) may restrict the space into which the generative AI model can generate synthetic field data. By restricting the space, the generative AI model may produce more realistic and constrained data (hybrid field-synthetic data). The hybrid field-synthetic data may be built in massive quantities and used to train subsequent machine learning models.

One or more of models described herein may include an artificial neural network. FIG. 7 is a diagram illustrating an example artificial neural network included in some implementations. In FIG. 7, a fracking module 700 includes the artificial neural network (ANN) 702. The ANN may be any suitable type of neural network. The ANN 702 may include a plurality of neurons 704. The ANN 702 also may include an input layer having any suitable number of neurons 704 (supporting any suitable number of features). The input layer may intake features indicating aspects related to hydraulic fracturing (as described herein). The ANN 702 also may include an output layer that may predict a target (as described herein) and perform any of the operations described herein. The output layer may include any suitable number of neurons.

FIG. 8 is a block diagram illustrating a computer system that may be utilized with some implementations. In FIG. 8, the computer system 800 may include one or more processors 802 connected to a system bus 804. The system bus 804 may be connected to memory 808 and a network interface 805. The memory 808 may include any suitable memory random access memory (RAM), non-volatile memory (e.g., magnetic memory device), and/or any device for storing information and instructions executable by the processor(s) 802. The network interface 805 may provide connectivity to any suitable network, such as a wired network, wireless network, satellite network, etc.

The computer system 800 may include additional peripheral devices. For example, the computer system 800 may include multiple external multiple processors. In some implementations, any of the components may be integrated or subdivided.

The computer system 800 also may include the fracking module 700. The fracking module 700 may implement any one or more of the methods and operations described herein. The fracking module 700 may include an ANN 702 or other logic for performing machine-learning and classification operations described herein. In some implementations, the computer system 800 may be included in a well system (such as the well system described with reference to FIG. 10) and may cooperate with other components and/or systems to perform the functionality described herein.

The computer system 800 also may include a sensor controller 812 configured to perform operations for capturing sensor data and processing the sensor data (as described herein). The sensor controller 812 may transmit sensor data to the fracking module 700 or any other component in or external to the computer system 800.

The computer system 800 also may include a fracking controller 810 configured to perform operations for controlling hydraulic fracturing in a well. The fracking controller 810 may respond to output (such as predictions, classifications, or other outputs) from the fracking module 700.

Although the components are shown separately, any of the components of the computer system 800 may be further combined or subdivided. For example, the fracking module 800 and fracking controller 810 may be combined into a single component or subdivided into three or more components. Any component of the computer system 800 may be implemented as hardware, firmware, and/or machine-readable media including computer-executable instructions for performing the operations described herein. For example, some implementations include one or more non-transitory machine-readable media including computer-executable instructions including program code configured to perform functionality described herein. Machine-readable media includes any mechanism that provides (e.g., stores and/or transmits) information in a form readable by a machine (e.g., a computer system). For example, tangible machine-readable media includes read only memory (ROM), random access memory (RAM), magnetic disk storage media, optical storage media, flash memory machines, etc. Machine-readable media also includes any media suitable for transmitting software over a network.

Example Environment

The computer system 800 may be part of a larger system for drilling and fracturing well. FIG. 9 is an illustration depicting an example multi-well system, according to some implementations. In particular, FIG. 9 is a schematic of a multi-well system 900 that includes a wellbore 902 and a wellbore 908 in a subsurface formation 901. The wellbore 902 includes casing 906 and a number of perforations 990A-990H being made in the casing 906 at different depths to allow reservoir fluids (i.e., oil, water, and gas) from the subsurface formation 901 to flow into the wellbore 902. Similarly, the wellbore 908 includes casing 910 and a number of perforations 980A-980H being made in the casing 910 to allow reservoir fluids (i.e., oil, water, and gas) from the subsurface formation 901 to flow into the wellbore 908. During hydraulic fracturing operations of the wellbores 902 908, fracturing fluid, with or without sand, may be pumped into the subsurface formation 901, via the perforations 990A-990H and perforations 980A-980H, to hydraulically fracture the rock such that reservoir fluid may flow into the wellbore 902, 908, respectfully.

In some implementations, one or more sensors may be positioned in a wellbore to obtain measurements while an offset well is being hydraulically fractured. For example, the wellbore 902 may include a fiber optic cable 920 to obtain strain measurements, temperature measurements, derived pressure measurements (from strain measurements), etc. of the subsurface formation 901 while the wellbore 908 is being hydraulically fractured. The fiber optic cable 920 may extend from the wellhead 914 on the surface 911 to the subsurface along the wellbore 902. The fiber optic cable 920 may be cemented in place in the annular space between the casing 906 of the wellbores 902 and the subsurface formation 901. The fiber optic cable 920 may be clamped to the outside of the casing 906 during deployment and protected by centralizers and cross coupling clamps. The fiber optic cables 920 may be included with coiled tubing, wireline, loose fiber using coiled tubing, or gravity deployed fiber coils that unwind the fiber as the coils are moved in the wellbore 902. The fiber optic cable 920 also may be deployed with pumped down coils and/or self-propelled containers. Additional deployment options for the fiber optic cable 920 may include coil tubing and wireline deployed coils where the fiber optic cables 920 are anchored at the toe of the wellbore. In such implementations the fiber optic cable 920 may be deployed when the wireline or coiled tubing is removed from the well. The fiber optic cable 920 may house one or more optical fibers, and the optical fibers may be single mode fibers, multi-mode fibers, or a combination of single mode and multi-mode optical fibers. The distribution of sensors shown in FIG. 9 is for example purposes only. Any suitable sensor deployment may be used.

The fiber optic cable 920 may be used for distributed sensing where acoustic, vibration, strain, and temperature measurements may be collected downhole in the wellbores 902. The measurements may be collected at various positions distributed along the fiber optic cable 920. For example, data may be collected every 1-3 ft along the full length of the fiber optic cable 920 downhole along the horizontal section of the wellbore. Fiber optic interrogation unit 922 of the wellbore 902 may be located on the surface 911 of the multi-well system 900. The fiber optic interrogation units 922 may be directly coupled to the fiber optic cables 920. Alternatively, the fiber optic interrogation units 922 may be coupled to a fiber stretcher module, wherein the fiber stretcher module is coupled to the fiber optic cable 920. The fiber optic interrogation unit 922 may receive measurement values taken and/or transmitted along the length of the fiber optic cable 920 such as acoustic, temperature, strain, etc. The fiber optic interrogation unit 922 may be electrically connected to a digitizer to convert optically transmitted measurements into digitized measurements.

The fiber optic interrogation unit 922 may operate using various sensing principles including but not limited to amplitude-based sensing systems like DTS, DAS, Low Frequency Distributed Acoustic Sensing (LFDAS), Distributed Vibration Sensing (DVS), and Distributed Strain Sensing (DSS). For example, the DTS system may be based on Raman and/or Brillouin scattering. A DAS system may be a phase sensing-based system based on interferometric sensing using homodyne or heterodyne techniques where the system may sense phase or intensity changes due to constructive or destructive interference. The DAS system may also be based on Rayleigh scattering and in particular coherent Rayleigh scattering. A DSS system may be a strain sensing system using dynamic strain measurements based on interferometric sensors or static strain sensing measurements using Brillouin scattering. DAS systems based on Rayleigh scattering may also be used to detect dynamic strain events. Temperature effects may in some cases be subtracted from both static and/or dynamic strain events, and temperature profiles may be measured using Raman based systems and/or Brillouin based systems capable of differentiating between strain and temperature, and/or any other optical and/or electronic temperature sensors, and/or any other optical and/or electronic temperature sensors, and/or estimated thermal events.

In some implementations, the fiber optic interrogation unit 922 may measure changes in optical fiber properties between two points in an optical fiber at any given point, and these two measurement points move along the optical sensing fiber as light travels along the optical fiber. Changes in optical properties may be induced by strain, vibration, acoustic signals and/or temperature as a result of the fluid flow. Phase and intensity based interferometric sensing systems are sensitive to temperature and mechanical, as well as acoustically induced, vibrations. DAS data may be converted from time series data to frequency domain data using Fast Fourier Transforms (FFT) and other transforms, like wavelet transforms, also may be used to generate different representations of the data. Various frequency ranges may be used for different purposes and low frequency signal changes may be attributed to formation strain changes or fluid movement and other frequency ranges may be indicative of fluid movement. Various techniques may be applied to generate indicators of events related to the generation and/or expansion of shear induced fracture fields during hydraulic fracturing operations. Although FIG. 9 depicts the fiber optic cable 920 in the wellbore 902, a fiber optic cable 920 may also be positioned in the wellbore 908 to obtain measurements when the wellbore 902 is hydraulically fractured.

The wellbore 902 may also include pressure sensors, such as externally ported pressure sensors 930, 932, to measure the formation pressure while the offset wellbore 908 is hydraulically fractured. Although FIG. 9 depicts the externally ported pressure sensors 930, 932 at the heel and toe of the wellbore 902, respectively, the externally ported pressure sensors 930, 932 may be positioned at any suitable location in the wellbore 902. Although FIG. 9 depicts the externally ported pressure sensors 930, 932 external to the casing 906 of the wellbore 902, externally ported pressure sensors 930, 932 may also be positioned in the wellbore 908 to obtain measurements when the wellbore 902 is hydraulically fractured.

During the hydraulic fracturing operations of wellbore 902 and/or wellbore 908, shear induced fracturing fields may be generated and/or dilated. For example, the shear induced fracturing fields comprising Mode 2 and/or Mode 3 failures may form between clusters of a stage, between stages of a wellbore, between clusters and/or stages of offset wellbores, etc. In some implementations, the fiber optic cable 920 and/or the externally ported pressure sensors 930, 932 may obtain measurements of the subsurface formation 901 to detect and/or monitor the subsurface formation 901 and the shear induced fracture fields.

A computer 970 may be communicatively coupled to the fiber optic interrogation units 922, externally ported pressure sensors 930, 932, and other sensors in the multi-well system 900. The computer 980 may include a signal processor to perform various signal processing operations on signals captured by the fiber optic interrogation units 922, externally ported pressure sensors 930, 932, and/or other components of the multi-well system 900. The computer 970 may have one or more processors and a memory device to analyze the measurements and graphically represent analysis results on a display device. The computer 970 may include one or more of the components described with reference to FIG. 8. Although FIG. 9 depicts a system with multiple wellbores, embodiments described herein may also be applicable to other systems such as a single well system, multiple pads, etc.

FIGS. 1-9 and the operations described herein are examples meant to aid in understanding example implementations and should not be used to limit the potential implementations or limit the scope of the claims. None of the implementations described herein may be performed exclusively in the human mind nor exclusively using pencil and paper. None of the implementations described herein may be performed without computer components such as those described herein. Some implementations may perform additional operations, fewer operations, operations in parallel or in a different order, and some operations differently. Some implementations may perform the operations with different components. Irrespective of section headings of this disclosure, any one or more aspects described herein may be combined with any one or more other aspects described herein.

As used herein, a phrase referring to “at least one of” a list of items refers to any combination of those items, including single members. As an example, “at least one of: a, b, or c” is intended to cover: a, b, c, a-b, a-c, b-c, and a-b-c.

The various illustrative logics, logical blocks, modules, circuits, and algorithm processes described in connection with the implementations disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. The interchangeability of hardware and software has been described in terms of functionality, and illustrated in the various illustrative components, blocks, modules, circuits and processes described throughout. Whether such functionality is implemented in hardware or software depends upon the particular application and design constraints imposed on the overall system.

The hardware and data processing apparatus used to implement the various illustrative logics, logical blocks, modules and circuits described in connection with the implementations disclosed herein may be implemented or performed with a general purpose single- or multi-chip processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general-purpose processor may be a microprocessor or any conventional processor, controller, microcontroller, or state machine. A processor also may be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration. In some implementations, particular processes and methods may be performed by circuitry that is specific to a given function.

In one or more implementations, the functions described may be implemented in hardware, digital electronic circuitry, computer software, firmware, including the structures disclosed in this specification and their structural equivalents thereof, or in any combination thereof. Implementations of the subject matter described in this specification also may be implemented as one or more computer programs, such as one or more modules of computer program instructions stored on a computer storage media for execution by, or to control the operation of, a computing device.

If implemented in software, the functions may be stored on or transmitted over as one or more instructions or code on a computer-readable medium. The processes of a method or algorithm disclosed herein may be implemented in a processor-executable instructions which may reside on a computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that may be enabled to transfer a computer program from one place to another. Storage media may be any available media that may be accessed by a computer. By way of example, and not limitation, such computer-readable media may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that may be used to store desired program code in the form of instructions or data structures and that may be accessed by a computer. Also, any connection may be properly termed a computer-readable medium. Disk and disc, as used herein, includes compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk, and Blu-Ray™ disc where disks usually reproduce data magnetically, while discs reproduce data optically with lasers. Combinations also may be included within the scope of computer-readable media. Additionally, the operations of a method or algorithm may reside as one or any combination or set of codes and instructions on a machine readable medium and computer-readable medium, which may be incorporated into a computer program product.

Various modifications to the implementations described in this disclosure may be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other implementations without departing from the spirit or scope of this disclosure. Thus, the claims are not intended to be limited to the implementations shown herein but are to be accorded the widest scope consistent with this disclosure, the principles and the novel features disclosed herein.

Certain features that are described in this specification in the context of separate implementations also may be implemented in combination in a single implementation. Conversely, various features that are described in the context of a single implementation also may be implemented in multiple implementations separately or in any suitable subcombination. Moreover, although features may be described as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination may in some cases be excised from the combination, and the claimed combination may be directed to a subcombination or variation of a subcombination.

Similarly, while operations are depicted in the drawings in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. Further, the drawings may schematically depict one more example process in the form of a flow diagram. However, some operations may be omitted and/or other operations that are not depicted may be incorporated in the example processes that are schematically illustrated. For example, one or more additional operations may be performed before, after, simultaneously, or between any of the illustrated operations. In certain circumstances, multitasking and parallel processing may be advantageous. Moreover, the separation of various system components in the implementations described should not be understood as requiring such separation in all implementations, and the described program components and systems may be integrated together in a single software product or packaged into multiple software products. Additionally, other implementations are within the scope of the following claims. In some cases, the actions recited in the claims may be performed in a different order and still achieve desirable results.

EXAMPLE CLAUSES

Some implementations may include the following clauses.

Clause 1: A method comprising: determining objectives for a treatment well based on measurements that indicate conditions in the treatment well or one or more other wells; determining a model based on the objectives and the measurements; performing the measurements via one or more sensors; controlling, via the model, one or more operations in the treatment well to achieve the objectives based on the performed measurements.

Clause 2: The method of clause 1, wherein the objectives indicate operational goals to be achieved in the treatment well.

Clause 3: The method of any one or more of clauses 1-2, wherein the measurements each include one or more of a water hammer in the treatment well, fiber optic measurements in the treatment well, fiber optic measurements in an offset well, pressure measurements in the treatment well, pressure measurements in the offset well, the operation further including: generating temporally variant features for the model based on the measurements, wherein the temporally variant features include one or more of rolling statistical measures, derivatives, cumulative summations, lag features, Fourier information, wavelet information, entropy, and amplitude ratios.

Clause 4: The method of any one or more of clauses 1-3 further comprising: determining a minimal number of measurements needed to achieve at least one of the objectives; and installing the sensors in the treatment well and/or the one or more other wells

Clause 5: The method of any one or more of clauses 1-4, wherein at least one of the objectives includes both a near field objective having a first weight and a far field objective having a second weight, wherein the near field objective is based on the measurements from the treatment well and the far field objective is based on the measurements from an offset well or multiple offset wells.

Clause 6: The method of any one or more of clauses 1-5 further including generating, via one or more machine-learning models, the measurements that indicate conditions in the treatment well or one or more other wells.

Clause 7: The method of any one or more of clauses 1-6, wherein the model is based on physical constraints including one or more of total fluid volume to be pumped in the treatment well, total mass of proppant to be used in the first well, total chemicals to be used in the first well, perforation shots to be used in the first well, and stage length of the first well.

Clause 8: A method comprising: determining features in surface pumping data from a hydraulic fracturing process; generating probabilistic classification models configured to detect clustered features in the surface pumping data; generating, by a trained digital twin of the hydraulic fracturing process, remediation operations for the hydraulic fracturing process based on the clustered features; and transmitting the remediation operations to a fracturing process controller.

Clause 9: The method of clause 8 further comprising: receiving, from one or more answer products configured for real-time monitoring of the hydraulic fracturing process, fracturing process information indicating effects of the remediation operations; and retraining the probabilistic classification models based on the fracturing process information.

Clause 10: The method of any one or more of clauses 8-9 further comprising: updating the digital twin based on the fracturing process information.

Clause 11: The method of any one or more of clauses 8-10, wherein the surface pumping data includes one or more of spread pressure, rate, proppant concentration, elapsed stage timing, perforation diameter, and total number of perforations made during the hydraulic fracturing process.

Clause 12: The method of any one or more of clauses 8-11, wherein the remediation operations include operations for modifying a physical aspect of the hydraulic fracturing process.

Clause 13: The method of any one or more of clauses 8-12, wherein the remediation operations modify one or more of temperature, chemical composition, and pressure in a well.

Clause 14: The method of any one or more of clauses 8-13 further comprising: generating training data with which the features are generated, wherein the training data are generated via a machine-learning model and/or a physics-based model.

Clause 15: One or more tangible machine-readable mediums including instructions configured for execution on one or more processors, the instructions comprising: instructions to determine objectives for a treatment well based on measurements that indicate conditions in the treatment well or one or more other wells; instructions to determine a model based on the objectives and the measurements; instructions to perform the measurements via one or more sensors; instructions to control, via the model, one or more operations in the treatment well to achieve the objectives based on the performed measurements.

Clause 16: The one or more tangible machine-readable mediums of clause 15, wherein the objectives indicate operational goals to be achieved in the treatment well.

Clause 17: The one or more tangible machine-readable mediums of any one or more of clauses 15-16, wherein the measurements each include one or more of a water hammer in the treatment well, fiber optic measurements in the treatment well, fiber optic measurements in an offset well, pressure measurements in the treatment well, and pressure measurements in the offset well.

Clause 18: The one or more tangible machine-readable mediums of any one or more of clauses 15-17 further comprising: instructions to determine a minimal number of measurements needed to achieve at least one of the objectives; and instructions to install the sensors in the treatment well and/or the one or more other wells.

Clause 19: The one or more tangible machine-readable mediums of any one or more of clauses 15-18, wherein at least one of the objectives includes both a near field objective having a first weight and a far field objective having a second weight, wherein the near field objective is based on the measurements from the treatment well and the far field objective is based on the measurements from an offset well.

Clause 20: The one or more tangible machine-readable mediums of any one or more of clauses 15-19, wherein the controlling attains the near field objective and the far field objective during a treatment of the treatment well.