Automated well control based on detected fracturing-driven disturbances

Abstract

This invention provides a method for controlling the operation of a probe well located near an excitation well undergoing hydraulic fracturing, which can generate fracturing-driven interference (FDI) events on the probe well. The method includes: providing an FDI intervention system comprising a computer-implemented predictive model for determining the risk of an FDI event occurring during the hydraulic fracturing operation; calculating a risk-weighted FDI event cost affecting production from the probe well; and calculating the defensive intervention implementation cost of applying defensive intervention to the probe well to mitigate the harm from the FDI event. The method includes calculating a cost comparison based on a comparison between the defensive intervention implementation cost and the risk-weighted FDI event cost. The method concludes with automatic control of the probe well operation using the FDI intervention system based on the cost comparison.

Description

[0001] Related applications

[0002] This application claims priority to U.S. Patent Application Serial No. 17 / 149,706, filed January 14, 2021, entitled “Automatic Well Control Based on Detection of Fracture Driven Interference,” the disclosure of which is incorporated herein by reference. Technical Field

[0003] This invention relates generally to the field of oil and gas production, and more specifically, but not in a limiting way, to a system and method for automatically adjusting the operation of a probing well based on actual or predicted fracturing-driven disturbance (FDI) events in a nearby induced well. Background Technology

[0004] Drilling boreholes or wellbores in subsurface geological structures containing hydrocarbon reservoirs is the process of extracting hydrocarbons. Typically, the first set of wellbores is located in areas believed to define the boundaries of the reservoir block, or in areas of interest to the operator within the reservoir block. These existing, or “mother,” wellbores usually have horizontal components extending into the reservoir. A second set of wellbores may be drilled alongside the mother wellbore to increase hydrocarbon production and fully utilize the reservoir asset. The second set of wellbores may be referred to as infill or “daughter” wellbores. The term “exploratory well” generally refers to an existing well located near an “excitement” well that is being drilled or undergoing completion services (such as hydraulic fracturing).

[0005] Hydraulic fracturing can be used to improve hydrocarbon recovery from activated infill wells. A "fracturing shock" is a form of fracturing-driven disturbance (FDI) that occurs during completion when an infill (activated) well is connected to an existing (exploration) well. Fracturing shocks can negatively or positively impact production from existing wells. In some cases, pressure connectivity between adjacent wellbores will lead to increased pressure in the passive well, where fracturing fluid and proppant are lost from the activated well that underwent hydraulic fracturing. This can result in reduced production from the passive or exploration well due to the increased presence of sand and proppant in the well, or reduced production from the activated well due to ineffective enhancement.

[0006] To minimize the risk of adverse effects within the probe well, operators typically shut it down while the infill well is being hydraulically fracturing. Shutting down the probe well restricts the flow of fluids and proppant from the infill well. In other cases, operators may deploy defensive measures to the probe well to further mitigate the risk of adverse effects from FDI events. Defensive measures may include injecting fluid into the probe well to increase pressure within the well, thereby preventing proppant and high-pressure fracturing fluid from flowing into the infill well. In either case, deploying defensive measures or shutting down the well results in downtime and lost or delayed production.

[0007] The causes and impacts of FDI events are not yet fully understood. Operators tend to apply specific strategies for good protection, leading to negative economic consequences in terms of production delays and excessive intervention costs. Therefore, there is a need for an improved well management system that facilitates and automates the decision-making and deployment of interventions in probing wells. This implementation addresses these and other shortcomings in the existing technology. Summary of the Invention

[0008] In one aspect, the present invention provides a method for controlling the operation of a probe well located near an excitation well undergoing hydraulic fracturing operation that could generate fracturing-driven interference (FDI) events on the probe well. The method aims to optimize the economic recovery of hydrocarbons from both the excitation and probe wells. The method includes the step of providing an FDI intervention system comprising a computer-implemented predictive model for determining the risk of an FDI event occurring during the hydraulic fracturing operation. The method further includes the steps of: calculating a risk-weighted FDI event cost affecting production from the probe well, and calculating a defensive intervention implementation cost of applying a defensive intervention to the probe well to mitigate the harm from the FDI event. The method also includes the step of calculating a cost comparison based on a comparison of the defensive intervention implementation cost and the risk-weighted FDI event cost. The method concludes with the step of automatically controlling the operation of the probe well using the FDI intervention system based on the cost comparison.

[0009] In another aspect, an exemplary embodiment includes a method for controlling the operation of a probe well located near an excitation well undergoing a hydraulic fracturing operation capable of generating a fracturing-driven disturbance (FDI) event on the probe well, wherein the method aims to optimize the economic recovery of hydrocarbons from both the excitation well and the probe well. The method begins with the step of providing an FDI intervention system including a computer-implemented predictive model for determining the risk of an FDI event occurring during the hydraulic fracturing operation. Next, the method includes the steps of: calculating a risk-weighted FDI event cost affecting production from the probe well, and calculating a defensive intervention implementation cost of applying a defensive intervention to the probe well to mitigate the harm from the FDI event. Next, the method includes the step of calculating a cost comparison based on a comparison of the defensive intervention implementation cost and the risk-weighted FDI event cost. The method ends with the step of automatically controlling the operation of the probe well: if the calculated cost comparison determines that the defensive intervention implementation cost is less than the risk-weighted FDI event cost, then the defensive intervention is applied to the probe well.

[0010] In other embodiments, exemplary implementations include an FDI intervention system for automatically controlling the operation of an exploration well located near an excitation well undergoing hydraulic fracturing, which can generate fracturing-driven interference (FDI) events in the exploration well. The FDI intervention system includes: multiple pressure sensors configured to monitor pressure in the excitation and exploration wells; multiple automated controls configured to adjust the operation of the exploration well; a well intervention mechanism connected to the exploration well; and an analysis module including a predictive model for determining the risk of an FDI event, representing an FDI event occurring between the excitation and exploration wells. The analysis module is configured to automatically control the multiple automated controls, in part, based on the FDI event risk. Attached Figure Description

[0011] Figure 1 It is a depiction of a series of wells connected to the FDI intervention system.

[0012] Figure 2 This is a diagram illustrating an overview of the process for identifying and applying optimized well intervention strategies.

[0013] Figure 3 It is a flowchart of the process for developing an integrated predictive model to assess the risks, outcomes, and impacts of defensive interventions in FDI events.

[0014] Figure 4 This is a process flowchart for an automated method of controlling exploration wells.

[0015] Figure 5 This is a flowchart of the process used to automatically apply defensive interventions on exploration wells. Detailed Implementation

[0016] According to an exemplary implementation, Figure 1 An automated fracturing-driven interference (FDI) intervention system 100 is shown, deployed to optimize production from one or more probing wells 102 located near an excitation well 104. The excitation well 104 is undergoing hydraulic fracturing operations, while one or more probing wells 102 have been completed. As shown, the excitation well 104 is a second infill well (which can be, for example, a parent well and an earlier infill well) located between probing wells 102a and 102b. The excitation well 104 and the probing wells 102 extend from a common well site 106. Figure 1 The indication is that a fracturing impact (“FDI event”) occurred between the excitation well 104 and the exploration well 102b, and two fracturing impacts occurred between the excitation well 104 and the exploration well 102a.

[0017] Will realize, Figure 1The wells depicted herein are merely examples of how the FDI intervention system 100 can be deployed, and the systems and methods of the exemplary implementation will find utility in other arrangements of dense-network drilling. For example, the FDI intervention system 100 can be used to proactively monitor hydraulic fracturing operations performed simultaneously on multiple excitation wells 102. As used herein, the term “well” collectively refers to exploratory wells 102a, 102b and excitation well 104.

[0018] Each well includes one or more pressure sensors 108 that measure pressure at specific locations or areas within the well. For example... Figure 1 As shown, each well is divided into multiple stages for hydraulic fracturing and production operations. Each well also includes automation controls 110. Automation controls 110 may include control valves, throttles, and other equipment that can be activated to close, open, and manage the well. For example, automation controls 110 on probe well 102 may be remotely activated to close probe well 102 or to position probe well 102 in fluid communication with well intervention mechanism 112. Well intervention mechanism 112 may include pressurized injection fluids such as supercritical carbon dioxide, nitrogen, steam, hydrocarbon fluids (including crude oil fluids, diesel, wellhead gas, and natural gas), water, and brine, as well as treatment and production enhancement chemicals. In other embodiments, well intervention mechanism 112 includes equipment and materials for performing “repeated fracturing” operations on probe well 102, wherein pressurized hydraulic fracturing fluid and proppant are injected into probe well 102.

[0019] Pressure sensor 108 is configured to report the measured pressure to a computer-implemented analysis module 114 on a continuous or periodic basis. This analysis module also contains a database of field data. Figure 1 In the exemplary embodiment depicted, the analysis module 114 is configured to be accessed via a cloud computing network to one or more remote computers. A local communication system 116 can be used to collect and transmit raw data between the pressure sensor 108 and the automation control 110, as well as the analysis module 114, using commercially available telecommunications networks and protocols (e.g., the ModBus protocol). In other embodiments, some or all of the pressure sensor 108 and the automation control 110 are directly connected to the remote analysis module 114 via a direct network connection without interfering with the location communication system 116.

[0020] The hydraulic fracturing equipment 118 is positioned near the activation well 104 and controlled by a control station 120. In many applications, the control station 120 serves as a "fracturing operation monitoring vehicle," providing operators with control and real-time information regarding the hydraulic fracturing operation. Several performance criteria can be adjusted by the control station 120, including, for example, the composition of the fracturing fluid and slurry, the type and amount of sand or proppant injected into the activation well 104, and the pumping pressure and flow rate achieved during the hydraulic fracturing operation. Each of these criteria is referred to herein as an "operational variable" relevant to the activation hydraulic fracturing operation. The control station 120 is also connected directly or via a local telecommunications system 116 to an analysis module 114.

[0021] Although analysis module 114 is in Figure 1 While depicted as a cloud computing resource, in other implementations, the analytics module 114 is locally located near the well and control station 120. Positioning the analytics module 114 near the well reduces the latency between the time spent measuring real-time data and the time the analytics module 114 processes that data. Conversely, positioning the analytics module 114 in the cloud or at a remote location allows for the use of more powerful computing systems. In yet another implementation, some processing is performed using a local computer configured with an "edge-based" architecture near the well, while the balance of processing occurs at a remote location.

[0022] One or more workstations 122 are connected to the analysis module 114 via a local direct connection or a secure network connection. Workstations 122 are configured to run a computer-implemented FDI intervention program that provides users with real-time information generated by the analysis module 114. Workstations 122 may be located in different locations. In some embodiments, some workstations 122 are located remotely from the well, while others are located near the well in the control station 120 or as part of a local edge-based computing system. As used herein, the term "workstation" includes personal computers, thin client computers, mobile phones, tablet computers, and other portable electronic computing devices.

[0023] As used herein, the term "FDI intervention system 100" refers to a collection of at least two or more of the following components: pressure sensor 108, automation controls 110, well intervention mechanism 112, control station 120, analysis module 114, workstation 122, and any intervention data network, such as local telecommunications system 116. It should be understood that the FDI intervention system 100 may include additional sensors and controls in or near the activation well 104 and the exploration well 102. Such additional sensors may include, for example, microseismic sensors, temperature sensors, proppant or fluid tracer detectors, acoustic sensors, and sensors located in other downhole equipment in the lift, completion, or well. Data measurement signals provided by such additional sensors are transmitted directly or via an intermediate data network to the analysis module 114.

[0024] As described below, the FDI intervention system 100 is typically configured to monitor hydraulic fracturing operations on an excitation well 104, determine the likelihood of an FDI event occurring between the excitation well 104 and one or more exploration wells 102, develop one or more defensive intervention protocols designed to protect potentially affected exploration wells 102, compare the relative economic impact of deploying and not deploying defensive intervention protocols, and then, based on the determination of which option presents the lowest total risk of adverse economic impact, control the operation of the excitation well 104 and exploration wells 102 according to the selected well control protocol. In an exemplary embodiment, the FDI intervention system 100 is configured to automatically perform this comparative analysis in real time and implement the selected well control protocol on the exploration well 102 without requiring direct human guidance.

[0025] Defensive intervention protocols include, but are not limited to, injecting pressurized injection fluids into probe well 102 (e.g., supercritical carbon dioxide, nitrogen, wellhead gas, natural gas, steam, water, and brine), injecting well treatment and production enhancement chemicals into probe well 102 (e.g., surfactants, soaps, and friction modifiers), partially or completely shutting down (closing) probe well 102, delaying or modifying the completion schedule for probe well 102, and performing new or “repeated fracturing” hydraulic fracturing operations on probe well 102. It should be understood that this is a non-exhaustive list of defensive intervention protocols. It should also be understood that two or more of these defensive intervention protocols may be performed simultaneously or sequentially, and defensive intervention protocols may be applied to multiple probe wells 102 as part of a comprehensive plan covering multiple potentially impactful probe and activating wells 102, 104.

[0026] Prior to hydraulic fracturing, an operator using the FDI intervention system 100 at workstation 122 can connect analysis module 114 to control station 120 and a selected number of pressure sensors 108 in ignition well 104 and probe well 102. Once hydraulic fracturing has commenced, analysis module 114 can poll control station 120 and pressure sensors 108 continuously or periodically. In some embodiments, analysis module 114 polls the pressure sensors at intervals between once per second and once every fifteen minutes. In an exemplary embodiment, analysis module 114 pulls the pressure sensor 108 every thirty seconds. Raw data from control station 120 and pressure sensors 108 is provided to analysis module 114 for processing. Analysis module 114 is typically configured to detect anomalies in pressure measurements obtained from the pressure sensors in probe well 102. In some embodiments, analysis module 114 applies simple rule-based analysis, where recommended actions are determined based on inputs received from control station 120 and pressure sensors 108. In other implementations, the analysis module 114 invokes machine learning, simulation physics engines, or statistical functions to detect FDI events based on pressure anomalies and autonomously determine the causal relationship between the FDI events and one or more characteristics of the hydraulic fracturing operation and the well.

[0027] Therefore, refer to Figure 2 The analysis module 114 of the FDI intervention system 100 is typically configured to perform optimized well control operations 200 by receiving the following: (i) inputs of field data from box 202 (e.g., pressure sensor 108, automation control 110); (ii) information from a historical database at box 204, which correlates the economic impact of past production enhancements and interventions in relevant hydrocarbon-producing geological structures; and (iii) information regarding planned hydraulic fracturing operations to be performed on activating well 104 at box 206, and potential defensive intervention protocols that can be deployed on exploratory well 102. The analysis module 114 is optimally configured to apply machine learning and neural networks to the various inputs at box 208 to generate one or more recommendations at box 210. The recommended well control protocols can be implemented manually or automatically to optimize hydrocarbon production from exploratory well 102 and activating well 104. Once the selected well control protocol has been put into operation, the results of the operation are studied at box 212 and used to update the input of analysis module 114 for further iteration of FDI intervention system 100.

[0028] Turn Figure 3The diagram illustrates a process flow chart for the predictive analytics model development process 300. The process begins at step 302, where historical data related to the asset (e.g., pressure readings from exploratory well 102 and excitation well 104) are collected. At step 304, features and parameters of the model are developed based on multiple factors related to hydrocarbon production from the well, including, for example, production targets, completion strategies, well spacing, well construction, drilling technology and progress, well depletion and stress, and reservoir-specific properties (e.g., porosity, depth, etc.).

[0029] At step 306, based on these features, parameters, and historical data, the model development process 300 finds correlations between the features, historical data, and evidence of actual FDI events occurring in the historical data. Confirmatory data to determine the likelihood of FDI events can be obtained using tracer fluid mechanisms, optical fibers, pressure response analysis, and production response analysis. Based on these correlations, process 300 ranks the features and parameters at step 308.

[0030] At step 310, the process uses machine learning algorithms to build a predictive model, which may include support vector machines (SVM), random forests, and artificial neural networks. At step 310, the predictive model is iteratively built based on multiple inputs, including completion strategies, normalized completion parameters, well characteristics, reservoir quality, distance, and depletion history. The predictive model is configured to output multiple probabilities, including the risk of an FDI event, the cost and availability of potential defensive intervention protocols to mitigate the harm caused by the FDI event, the risk of interruption of production in exploration well 102 if no defensive intervention protocol is implemented, and the risk of interruption and delay of production caused by the implementation of one or more defensive intervention protocols. Importantly, the predictive model can be configured to produce composite predictions that include the probability of a particular event occurring and the relative costs and benefits associated with those events and potential interventions. In this way, the computer-implemented model can be configured to output a predictive array or spectrum that includes both probability and cost / benefit factors. For example, analysis module 114 may determine that, in order to mitigate the damage caused by a highly unlikely FDI event (but which would result in significant disruption if the FDI event were to occur), a defensive intervention protocol that demonstrates a significant risk of causing minor disruption to production from exploration well 102 should be deployed.

[0031] It is important to note that, in certain circumstances, the analysis module 114 can determine that a particular FDI event will be beneficial to the exploration well 102. For example, if the analysis module 114 determines that an FDI event will increase or otherwise enhance hydrocarbon production from the exploration well 102, the analysis module 114 can generate recommendations that include the potential benefits realized by the occurrence of the predicted FDI event (e.g., a “negative” value within the cost determination structure). The state or operation of the exploration well 102 can be automatically adjusted in response to the recommendations from the analysis module 114 to optimize the benefits received from the predicted FDI event.

[0032] At step 312, a selected set of recommendations (e.g., whether to implement recommended defensive intervention protocols) are implemented on at least some of the probe well 102 and the activation well 104. Once implemented, the results of the hydraulic fracturing operation on the activation well 104 and the impact (if any) on the probe well 102 are measured. This information may include changes in downhole pressure in the probe well 102 indicating an FDI event, the cost of production losses from the probe well 102, the complexities of the hydraulic fracturing operation on the activation well 104, and the cost of implementing defensive intervention protocols on the probe well 102. Then, at step 310, this information can be stored, processed, analyzed, and used as input for the next iteration of the predictive model.

[0033] Next turn Figure 4 The diagram illustrates a process flow chart of a method 400 for automatically controlling a probe well 102 using an FDI intervention system 100. Method 400 begins at step 402, where a “candidate” probe well 102 is selected for analysis using the FDI intervention system 100. The candidate well is selected before the next stage of a completion operation (e.g., hydraulic fracturing) on ​​the excitation well 104. Once the candidate probe well 102 is selected, method 400 is divided into two sequences, which can be executed in parallel or sequentially. In one sequence, at step 404, the FDI intervention system 100 determines the probability of an FDI event occurring at the candidate probe well 102 during the upcoming completion stage on the excitation well 104. At step 406, if an FDI event occurs and production from the candidate probe well 102 is interrupted, the FDI intervention system 100 provides a prediction of the costs resulting from the production loss. In this way, if candidate exploration well 102 remains online without defensive intervention during the next phase completed on activating well 104, the FDI intervention system 100 generates a “risk-weighted production loss” that can be caused by an FDI event.

[0034] In another sequence, at step 408, if candidate probe well 102 is shut down or if a defensive intervention protocol is applied, the FDI intervention system 100 estimates the production delay. At step 410, the FDI intervention system 100 estimates the economic impact of the production delay caused by shutting down candidate probe well 102 or applying a defensive intervention that temporarily interrupts or reduces production from probe well 102. The costs calculated at step 410 may include the costs of materials and labor used to implement the defensive intervention protocol.

[0035] At step 412, the FDI intervention system 100 analyzes the risk-weighted costs of intervening in and not intervening in candidate exploratory well 102. If the anticipated loss from shutting down or intervening in production from candidate exploratory well 102 exceeds the risk-weighted loss from unmitigated FDI events affecting candidate exploratory well 102, the FDI intervention system 100 recommends keeping candidate exploratory well 102 online at step 414 during the upcoming completion phase on activating well 104. However, if the FDI intervention system 100 determines that the risk-weighted loss from FDI events exceeds the costs of shutting down or applying defensive intervention protocols to candidate exploratory well 102, the FDI intervention system 100 recommends applying defensive protocols to candidate exploratory well 102 at step 416.

[0036] In some implementations, steps 402-416 are automated, and the recommendations in steps 414 and 416 are executed without human intervention by sending appropriate command signals to the automation control 110 and the well intervention mechanism 112. In other implementations, the FDI intervention system 100 is configured to generate written reports, visual displays, or other human-oriented outputs without automatically implementing the recommendations from step 412. The operator can then manually apply the recommendations based on a set of choices made by the analysis module 114.

[0037] In the presence of multiple exploration wells 102, method 400 proceeds to step 418, where the FDI intervention system 100 determines whether all candidate exploration wells 102 have been evaluated using method 400. Once all candidate exploration wells 102 have been evaluated using method 400, the method proceeds to step 420 and performs the next processing stage of the completed operation on the activated well 104. In some embodiments, the FDI intervention system 100 is configured to automatically initiate the next stage of the processing operation on the activated well 104 by sending appropriate command signals to the hydraulic fracturing equipment 118 and the control station 120.

[0038] Turn Figure 5The diagram illustrates a process flow chart of the defensive intervention protocol applied in step 416 of method 400. At step 502, the FDI intervention system 100 determines whether candidate probe well 102 should be temporarily shut down at step 504, or whether defensive intervention should be applied to the candidate probe well at step 506. If the FDI intervention system 100 recommends shutting down candidate probe well 102 at step 504, the FDI intervention system 100 sends appropriate command signals to the automated controls for candidate probe well 102 to shut down the well (e.g., via an automated throttle or control valve).

[0039] If the FDI intervention system 100 recommends applying a defensive intervention, it provides the recommended intervention based on predictive analytics derived from machine learning. Once the recommended defensive intervention is identified, method 500 moves to step 508 and applies the intervention. In an exemplary embodiment, the defensive intervention is automatically applied by the FDI intervention system 100 via signals sent to automation controls 110 and well intervention mechanism 112. As mentioned above, the application of the selected defensive intervention can also be manually applied by an operator in response to a recommendation report generated by the FDI intervention system 100. In some embodiments, the FDI intervention system 100 is configured to present multiple defensive intervention options for human operators to consider.

[0040] Once the selected defensive intervention has been applied, method 500 proceeds to step 510, where the FDI intervention system 100 determines whether the completion phase on excitation well 104 has been completed. Method 500 loops back to step 508 until the completion phase ends. Once the completion phase on excitation well 104 is completed, method 500 moves to step 512 to determine whether the implemented defensive intervention should be removed or withdrawn. In some cases, the FDI intervention system 100 may determine that leaving the defensive intervention in the appropriate location on candidate probe well 102 is more effective when anticipating subsequent completion phase activity on excitation well 104.

[0041] If the FDI intervention system 100 determines that the defensive intervention should remain in place, method 500 proceeds to step 514. If the FDI intervention system 100 determines that the defensive intervention should be withdrawn, method 500 proceeds to step 516 and puts candidate probe well 102 back into production by opening the well or removing the defensive intervention. Method 500 then proceeds to step 514, where information recorded in probe well 102 and activating well 104 is used to update the predictive model used by the FDI intervention system 100. At step 518, method 500 is reset for the next completion phase on activating well 104.

[0042] Therefore, in these exemplary embodiments, the FDI intervention system 100 determines the likelihood of an FDI event occurring between the excitation well 104 and one or more probe wells 102, assesses or develops one or more defensive intervention protocols designed to protect the potentially affected probe wells 102, compares the relative economic impact of deploying various defensive intervention protocols, and then controls the operation of the excitation well 104 and probe wells 102 according to the selected well control protocol, based on the determination of which option presents the lowest risk-weighted cost (adverse economic impact) for the probe well 102. While the FDI intervention system 100 is well-suited for use in conjunction with FDI events triggered by hydraulic fracturing, it can also find utility in monitoring and optimizing injection procedures implemented during enhanced oil recovery (EOR) operations.

[0043] It should be understood that although many features and advantages of various embodiments of the invention, as well as details of the structure and function of various embodiments of the invention, have been set forth in the foregoing description, this disclosure is merely illustrative and detailed changes may be made to the fullest extent indicated by the broad general meaning of the terms expressed in the appended claims, particularly in terms of the structure and arrangement of parts within the principles of the invention.

Claims

1. A method of controlling the operation of a boundary well located near a stimulation well, the stimulation well undergoing a hydraulic fracturing operation, the hydraulic fracturing operation capable of producing a fracturing-driven interference (FDI) event to the boundary well, wherein the method is directed to optimizing the economic recovery of hydrocarbons from the stimulation well and the boundary well, the method characterized by the steps of: providing an FDI intervention system, the FDI intervention system including a computer-implemented predictive model for determining the risk of the FDI event occurring during the hydraulic fracturing operation; computing an FDI event cost weighted by the risk of the FDI event affecting production from the boundary well; computing a defense intervention implementation cost of applying a defense intervention to the boundary well to mitigate the harm from an FDI event; computing a cost comparison based on the comparison of the defense intervention implementation cost to the risk-weighted FDI event cost; and automatically controlling the operation of the boundary well with the FDI intervention system based on the cost comparison. applying the defense intervention to the boundary well if the computed cost comparison determines that the defense intervention implementation cost is less than the risk-weighted FDI event cost.

2. The method of claim 1, wherein the step of automatically controlling operation of the edge well comprises:

3. The method of claim 2, wherein applying the defense intervention to the boundary well includes shutting in the boundary well.

4. The method of claim 2, wherein applying the defense intervention to the boundary well includes injecting a pressurized fluid into the boundary well to increase the pressure within the boundary well.

5. The method of claim 4, wherein applying the defense intervention to the boundary well includes performing a re-fracturing operation on the boundary well. not applying the defense intervention to the boundary well if the computed cost comparison determines that the defense intervention implementation cost is greater than the risk-weighted FDI event cost.

6. The method of claim 1, wherein the step of automatically controlling operation of the edge well comprises:

7. The method of claim 1, wherein the step of computing a defense intervention implementation cost includes evaluating a delayed production cost from temporarily shutting in the boundary well.

8. The method of claim 7, wherein the step of computing a defense intervention implementation cost further includes evaluating a material and labor cost of implementing a defense intervention protocol.

9. The method of claim 1, wherein the step of providing an FDI intervention system, the FDI intervention system including a computer-implemented predictive model for determining the risk of the FDI event occurring during the hydraulic fracturing operation, further includes developing the computer-implemented predictive model using machine learning.

10. The method of claim 9, wherein the step of developing the computer-implemented predictive model using machine learning includes associating the risk of an FDI event with feature engineered inputs.

11. The method of claim 10, wherein the step of developing the computer-implemented predictive model using machine learning further includes using an artificial neural network, a support vector machine, or a random forest determination. ​ 12. The method of claim 9, wherein the step of developing the computer-implemented predictive model using machine learning comprises correlating risk of FDI events with anomalies detected within the spurring well or the flanking well.

13. The method of claim 9, wherein the step of developing the computer- implemented predictive model using machine learning comprises correlating risk of FDI events based on a completion strategy of the spurring well.

14. The method of claim 9, wherein the step of developing the computer- implemented predictive model using machine learning comprises correlating risk of FDI events based on a set of wellbore characteristics of the spurring well.

Images