Injection-production dynamic real-time tracking and regulation method

Abstract

This invention relates to the field of injection and production monitoring technology, and particularly to a method for real-time tracking and control of injection and production dynamics. The method includes: controlling the injection well to perform limited-amplitude micro-disturbance injection according to a coded sequence in each preset tracking cycle, generating an injection reference signal; simultaneously acquiring the response signal of the production well, constructing and analyzing the injection and production disturbance response curve based on both, generating a connectivity consistency value and a response difference value characterizing the connectivity state between wells, and constructing an injection and production dynamic fingerprint map based on their temporal changes; dividing the fingerprint map into a stable sub-map and a fluctuating sub-map, identifying abnormal categories such as enhanced crossflow or limited injection / supply when fluctuations persist; automatically determining the target control combination and generating a graded control strategy accordingly; updating the fingerprint map after executing the strategy, with the disappearance of fluctuations and the return of the characterization value to a stable range as the control completion indicator, otherwise iterative optimization. This invention achieves active monitoring, intelligent diagnosis, and closed-loop control of injection and production dynamics, improving the accuracy and timeliness of oilfield management.

Description

Technical Field

[0001] This invention relates to the field of injection and production monitoring technology, and in particular to a method for dynamic real-time tracking and control of injection and production. Background Technology

[0002] Traditional injection-production dynamic monitoring methods mainly rely on periodic well tests, production data statistics, and manual experience analysis. These methods typically suffer from problems such as strong monitoring lag, coarse interpretation of results, and difficulty in identifying early anomalies. For example, conventional injection-production data analysis is often based on daily or monthly average data, failing to capture the dynamic changes of underground fluids on hourly or even minute-scale scales; the judgment of common problems such as crossflow and injection limitation heavily relies on the personal experience of engineers, lacking objective and quantitative real-time diagnostic methods; and control measures are mostly ex-post adjustments, with slow response and inability to quickly and accurately evaluate the control effects. With the development of smart oilfield technology, real-time data acquisition and monitoring systems have been widely used, providing a data foundation for dynamic analysis. However, some existing real-time analysis methods focus on trend monitoring of single parameters or simple correlation analysis, failing to integrate bidirectional, multi-parameter disturbance response information of injection and production at the system level.

[0003] Therefore, it is difficult to clearly characterize the subtle changes and physical nature of the dynamic connectivity between wells. Thus, there is an urgent need in this field for an integrated method that can achieve real-time and precise tracking of injection and production dynamics, intelligent early identification of anomalies, and automatic execution of closed-loop control and effect verification, in order to overcome the shortcomings of existing technologies in terms of real-time performance, accuracy, and intelligence, thereby optimizing the injection and production structure and improving the efficiency of oilfield development. Summary of the Invention

[0004] To address this, the present invention provides a dynamic real-time tracking and control method for injection and production, which overcomes the problem in the prior art that does not perform synchronous response analysis on injection wells and production wells, resulting in the inability to accurately capture subtle changes in dynamic connectivity between wells, causing a serious lag in the identification of abnormal states and a lack of targeted control measures, ultimately affecting the efficiency of injection and production development.

[0005] To achieve the above objectives, the present invention provides a method for dynamic real-time tracking and control of injection and production, comprising:

[0006] Step S1: Obtain the injection flow rate and injection pressure of the injection well, and obtain the bottom hole pressure, production rate and water cut of the production well;

[0007] Step S2: Within each preset tracking cycle, the injection well is controlled to perform limited micro-disturbance injection, so that the injection flow rate and injection pressure change according to a preset coding sequence, and the disturbance signal corresponding to the coding sequence is recorded as the injection reference signal;

[0008] Step S3: Obtain the production well response signal corresponding to the injection reference signal time. The response signal includes a bottom hole pressure change sequence and a water cut change sequence. Generate an injection-production disturbance response curve based on the injection reference signal and the response signal.

[0009] Step S4: Analyze the injection-production disturbance response curve, determine the connectivity consistency characterization value and response difference characterization value of the combination of injection well and production well based on the analysis results, and generate a dynamic fingerprint map of injection and production based on the changes of the connectivity consistency characterization value and response difference characterization value in the time dimension.

[0010] Step S5: Divide the injection and sampling dynamic fingerprint into a stable sub-graph and a fluctuating sub-graph, and identify the injection and sampling abnormal feature category when the duration of the fluctuating sub-graph is longer than a preset duration. The injection and sampling abnormal feature category includes the crossflow enhancement category and the injection supply limitation category.

[0011] Step S6: Determine the target control well combination based on the injection and production anomaly characteristic category, and generate the corresponding graded control strategy. The graded control strategy includes reducing the injection intensity of the corresponding injection well, and redistributing the injection intensity between different layers and / or applying pulsed injection to eliminate the restricted state.

[0012] Step S7: After executing the hierarchical control strategy, update the injection-collection dynamic fingerprint. When the duration of the fluctuation subgraph falls back to within the preset duration and the connectivity consistency representation value and the response difference representation value return to the stable benchmark range, the control is determined to be completed; otherwise, iteratively update the hierarchical control strategy.

[0013] As a preferred technical solution for the dynamic real-time tracking and control method of injection and production, the control of the injection well to perform limited micro-disturbance injection within each preset tracking cycle includes:

[0014] Within each preset tracking period, the injection flow rate and injection pressure of the injection well are modulated according to a preset coding sequence to generate a serialized perturbation with identifiable characteristics.

[0015] The amplitude change of the micro-perturbation injection does not exceed a preset amplitude threshold;

[0016] The preset encoding sequence is a pseudo-random binary sequence or a sequence consisting of alternating positive and negative steps, and a complete encoding sequence is executed within one preset tracking period.

[0017] As a preferred technical solution for the injection-production dynamic real-time tracking and control method, generating an injection-production disturbance response curve based on the injection reference signal and the response signal includes:

[0018] Real-time acquisition of bottom hole pressure and water cut data of the produced well, and time synchronization and alignment with the injection reference signal;

[0019] From the time-synchronized bottom hole pressure data, extract the pressure change sequence within the same time period as the injection reference signal, and use it as the first response signal;

[0020] From the time-synchronized and aligned moisture content data, extract the moisture content change sequence within the same time period as the injected reference signal, and use it as the second response signal;

[0021] The pressure response curve generated based on the first response signal, and the moisture content response curve generated based on the second response signal;

[0022] The injection-production disturbance response curve includes the pressure response curve and the water cut response curve.

[0023] As a preferred technical solution for the injection-production dynamic real-time tracking and control method, the connectivity consistency characterization value and response difference characterization value of the injection well and the production well combination are determined based on the analysis results, including:

[0024] Correlation analysis was performed on the pressure response curve and the water content response curve respectively to obtain the maximum cross-correlation coefficient between the injection reference signal and the corresponding response signal. The maximum cross-correlation coefficient corresponding to the pressure response curve was used as the connectivity consistency characterization value.

[0025] Time delay estimation is performed on the pressure response curve and the moisture content response curve respectively to obtain the corresponding pressure time delay and moisture content time delay;

[0026] The difference between the pressure time delay and the moisture content time delay is calculated and used as the response difference characterization value.

[0027] As a preferred technical solution for the injection-progression dynamic real-time tracking and control method, a continuous tracking period is used as a time series. The connectivity consistency characterization value and the response difference characterization value corresponding to each tracking period are plotted on the same chart to generate a trajectory diagram representing the evolution of the two over time, which is recorded as the injection-progression dynamic fingerprint.

[0028] As a preferred technical solution for the injection-production dynamic real-time tracking and control method, the injection-production dynamic fingerprint is divided into a stable subgraph and a fluctuating subgraph, including:

[0029] The state is divided according to the fluctuation amplitude of the connectivity consistency characterization value and the response difference characterization value in the injection and sampling dynamic fingerprint map;

[0030] When the change magnitude of the connectivity consistency characterization value and the response difference characterization value is less than the corresponding stability threshold in multiple consecutive tracking periods, the graph segment corresponding to the consecutive time period is divided into the stable subgraph.

[0031] When the change magnitude of at least one of the connectivity consistency characterization value and the response difference characterization value is greater than or equal to the corresponding stability threshold, the graph segment corresponding to that time period is divided into the fluctuation subgraph.

[0032] As a preferred technical solution for the injection-production dynamic real-time tracking and control method, when the duration of the fluctuation subgraph exceeds a preset duration, the injection-production abnormality feature category is identified, including:

[0033] When the number of consecutive tracking cycles corresponding to the fluctuation subgraph exceeds a preset duration threshold, the changing trends of the connectivity consistency representation value and the response difference representation value corresponding to the fluctuation subgraph are extracted.

[0034] If the connectivity consistency characterization value shows a monotonically increasing trend and the response difference characterization value shows a monotonically decreasing trend, then it is identified as a crosstalk enhancement category;

[0035] If the connectivity consistency characterization value shows a monotonically decreasing trend and the response difference characterization value shows a monotonically increasing trend, then it is identified as a supply-limited category.

[0036] As a preferred technical solution for the injection-production dynamic real-time tracking and control method, the target control well combination is determined according to the injection-production anomaly characteristic category, and a corresponding hierarchical control strategy is generated, including:

[0037] Based on the identified abnormal feature categories, the corresponding injection well-production well combinations are selected as the target control combinations;

[0038] If the target control combination is of the crossflow enhancement category, then a graded control strategy is generated to reduce the injection intensity of the corresponding injection well and / or redistribute the injection intensity between different layers.

[0039] If the target control combination is a supply-limited category, the generated graded control strategy includes increasing the injection intensity of the corresponding injection well and / or applying a graded control strategy of pulsed injection.

[0040] As a preferred technical solution for the injection-production dynamic real-time tracking and control method, determining whether the update is complete based on the updated injection-production dynamic fingerprint after executing the hierarchical control strategy includes:

[0041] After executing the hierarchical control strategy, the injection-progression dynamic fingerprint is updated based on the newly acquired data in several tracking cycles.

[0042] In the updated injection-collection dynamic fingerprint graph, it is determined whether the duration of the fluctuation subgraph is less than or equal to the preset duration, and whether the connectivity consistency characterization value and the response difference characterization value are both within the range of the stable benchmark.

[0043] If all of the above conditions are met, then the regulation targeting the current abnormal characteristics is considered complete.

[0044] Compared with existing technologies, the advantages of this invention are as follows: by applying identifiable, limited-amplitude micro-perturbations to the injection well in each tracking cycle, the formation response is actively stimulated, and pressure and water cut change signals of the production well are collected simultaneously; then, by using correlation analysis and time delay estimation techniques, the dynamic correlation between injection and production is quantified into connectivity consistency characterization values ​​and response difference characterization values, and an injection-production dynamic fingerprint map that can intuitively reflect the health status of the system is constructed based on its time series; by automatically identifying abnormal fluctuation patterns in the fingerprint map, the system can diagnose the type and severity of problems in real time, and automatically generate and execute graded control strategies; finally, by continuously updating the fingerprint map and verifying whether fluctuations are eliminated and whether parameters return to a stable baseline, the early identification capability and precise treatment efficiency of anomalies such as crossflow and injection limitation are significantly improved, thereby optimizing the injection-production structure, improving displacement effect and oilfield development benefits. Attached Figure Description

[0045] Fig. 1 This is a flowchart of the injection-progression dynamic real-time tracking and control method according to an embodiment of the present invention;

[0046] Fig. 2 This is a flowchart illustrating the generation of injection-import disturbance response curves based on the injection reference signal and the response signal, as described in an embodiment of the present invention.

[0047] Fig. 3 This is a logical diagram of dividing the injection and sampling dynamic fingerprint graph into a stable subgraph and a fluctuating subgraph in an embodiment of the present invention. Detailed Implementation

[0048] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0049] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0050] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0051] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0052] Please see Figs. 1-3 As shown, the present invention provides a method for dynamic real-time tracking and control of injection and extraction, comprising:

[0053] Step S1: Obtain the injection flow rate and injection pressure of the injection well, and obtain the bottom hole pressure, production rate and water cut of the production well;

[0054] In step S2, within each preset tracking period, the injection flow rate and injection pressure of the injection well are modulated by micro-perturbation injection not exceeding a preset amplitude threshold according to the preset encoding sequence, so as to generate a serialized perturbation with identifiable characteristics. A complete encoding sequence is executed within one preset tracking period.

[0055] In implementation, a preset amplitude threshold is a key constraint for implementing micro-disturbances, used to define the scope of the micro-disturbances. The preset amplitude threshold is usually not a fixed value, but rather an upper limit of variation determined comprehensively based on the specific oilfield's injection well conditions, formation properties, and current injection-production stability. In one implementation, this threshold can be set to a stable state. During a relatively stable phase of oilfield injection and production, small disturbances with amplitudes of 0.5%, 1%, 2%, and 3% can be applied, while simultaneously monitoring fluctuations in bottom hole pressure and production volume of key production wells.

[0056] By analyzing these response data, disturbance amplitudes that can be effectively captured by high-precision sensors without causing significant fluctuations in the production process or exceeding the formation's elastic response range are selected. This amplitude is established as the preset amplitude threshold, ensuring that the disturbance signal is sufficiently small to avoid impacting normal production and the formation, while also being effectively separated from background noise. The principle is to operate within the system's linear or weakly nonlinear response range, thereby enhancing the reliability and accuracy of subsequent stimulus-response analysis. This ensures the generation of active test signals with excellent signal-to-noise ratios for precise analysis without interfering with normal oilfield production.

[0057] Regarding the preset coding sequence, a pseudo-random binary sequence or a sequence composed of alternating positive and negative steps is used in implementation. When using a random binary sequence, a switching command that appears random but actually has precise mathematical regularity is simulated, such as: increase-decrease-decrease-increase… The purpose is to utilize the excellent autocorrelation characteristics of this sequence so that, even in the presence of noise interference, the correlation and time delay information with the injection signal can still be extracted with high precision from the response signal of the produced well. When using a sequence composed of alternating positive and negative steps, the injection parameters are made to undergo periodic rises and falls like steps, first increasing several times and then decreasing several times. The purpose is to generate a periodic disturbance pattern that is distinctive in the time domain and easy to intuitively identify and track, facilitating rapid trend comparison and initial anomaly judgment. Regardless of the sequence used, it is required that a complete coding sequence must be executed within a preset tracking period.

[0058] By using a preset sequence, it can be ensured that each tracking cycle is an independent and complete stimulus-response analysis unit, so that the injected reference signals generated in different cycles have a consistent baseline structure and duration. This allows dynamic analysis to be based on periodic and standardized testing, avoiding analysis bias caused by incomplete signals or cycle misalignment.

[0059] In this invention, by setting reasonable perturbation amplitude boundaries and using structured coding sequences to complete excitation within a fixed period, a stable, reliable and repeatable active detection signal source is provided for the entire method. This is the physical basis for subsequent accurate correlation analysis, dynamic fingerprint generation and intelligent control.

[0060] Step S3, specifically, involves real-time acquisition of bottom hole pressure and water cut data from the produced well, and time synchronization alignment with the injection reference signal; extracting the pressure change sequence within the same time period as the injection reference signal from the time-synchronized bottom hole pressure data, as the first response signal; extracting the water cut change sequence within the same time period as the injection reference signal from the time-synchronized water cut data, as the second response signal; and generating a pressure response curve based on the first response signal and a water cut response curve based on the second response signal.

[0061] It is understandable that the micro-disturbance signal applied by the injection well propagates through the formation to the production well with a physical time lag, and the data acquisition system may have internal clock differences. Therefore, time synchronization alignment is required before analysis, and the injection reference signal and the original data streams of bottom hole pressure and water cut from the production well are time-stamp aligned and corrected. In practice, this can be achieved by inserting synchronization markers into the signal or by using the absolute timestamp of the data acquisition system. This can eliminate the impact of system time difference and transmission delay on causal relationship analysis, ensuring that the extracted response sequence corresponds precisely to the injection disturbance on the time axis, thereby guaranteeing the physical authenticity and analytical reliability of the constructed response curve.

[0062] It should be understood that the extraction of pressure change and water content change sequences does not directly use the original observation values. Instead, it involves extracting segments from the aligned data that completely overlap with the time window of the injected reference signal within the current tracking period, and calculating their relative changes relative to the start of the period or the average value of the previous period. This forms a sequence that can directly reflect the changing trend caused by the disturbance, thereby separating the background trend and long-term fluctuations, highlighting the dynamic response directly caused by this active micro-disturbance, and improving the sensitivity and accuracy of subsequent correlation analysis.

[0063] When generating injection-import disturbance response curves, the injection reference signal sequence processed as described above is used as the independent variable, and the extracted first and second response signals are used as the dependent variables respectively. Their correspondence is plotted in the time-parameter variation coordinate system to form two independent response curves.

[0064] Understandably, defining both as injection-production disturbance response curves provides a multi-dimensional and complementary perspective on the response. The pressure response curve primarily characterizes the dynamic relationship between formation energy transfer and pressure conduction, and is sensitive to changes in connectivity and reservoir properties; the water cut response curve more directly reflects the sweep and propulsion of the injected fluid, and is sensitive to dominant water channels or crossflow phenomena. By setting up a parallel hyperbola, the assessment of injection-production dynamics becomes more comprehensive and three-dimensional, enabling cross-validation of analytical conclusions and enhancing the robustness of anomaly identification and feature judgment.

[0065] Step S4, specifically, performs correlation analysis on the pressure response curve and the water content response curve respectively, obtains the maximum cross-correlation coefficient between the injected reference signal and the corresponding response signal respectively, and uses the maximum cross-correlation coefficient corresponding to the pressure response curve as the connectivity consistency characterization value;

[0066] Time delay estimation is performed on the pressure response curve and the moisture content response curve respectively to obtain the corresponding pressure time delay and moisture content time delay; the difference between the pressure time delay and the moisture content time delay is calculated as the response difference characterization value.

[0067] In implementation, a cross-correlation analysis method is used. The injected reference signal sequence and the first response signal are slid point by point along the time axis, and the Pearson correlation coefficient is calculated, thus obtaining a cross-correlation function sequence in which the correlation coefficient changes with the time offset. Peak detection is performed on this sequence, and the correlation coefficient corresponding to the largest peak is extracted as the maximum cross-correlation coefficient characterizing the connectivity of the pressure response. The same operation is performed on the water content response signal to obtain another maximum cross-correlation coefficient characterizing the connectivity of the water content response.

[0068] In injection-production systems, pressure transmission is the most fundamental driving force and immediate characteristic of fluid flow. The smoothness of pressure wave transmission, i.e., the consistency of pressure response, directly determines the connectivity of the main channels in the formation. While changes in water cut are important, they are a subsequent derivative result, reflecting more the sweep efficiency of the aqueous fluid. Therefore, when assessing the fundamental state of inter-well connectivity, the correlation coefficient of the pressure response has a more basic and core indicative significance; only the maximum cross-correlation coefficient corresponding to the pressure response curve is selected as the characteristic value of connectivity consistency.

[0069] Understandably, cross-correlation analysis can effectively extract the mathematical characteristics of signal similarity and time delay from a noisy background. The magnitude of the maximum cross-correlation coefficient directly quantifies the fidelity of the injected disturbance signal reproduced at the production end. The closer the correlation coefficient is to 1, the higher the consistency of the two signal waveforms, that is, the smoother the dynamic connection path established between the injection well and the production well through the formation, and the more synchronized the response.

[0070] Separate time delay estimations are performed concurrently to obtain the time delay. During the cross-correlation analysis described above, the time offset corresponding to the maximum cross-correlation coefficient represents the time delay of the response signal relative to the injected reference signal. Specifically, the time offset extracted from the cross-correlation function of the pressure response is the pressure time delay, characterizing the time required for the pressure wave to propagate from the injection well to the production well; the time offset extracted from the cross-correlation function of the water cut response is the water cut time delay, representing the time required for the water phase fluid front or water cut change signal to advance from the injection well to the production well. This accurately captures the difference in propagation speed between pressure and water cut in the formation.

[0071] Understandably, pressure waves propagate rapidly via the elasticity of the rock, while fluid propagation is controlled by pore structure, permeability, and two-phase flow resistance, resulting in a much slower speed. Therefore, these two time delays themselves carry rich information about reservoir and fluid properties. The response difference characterization value eliminates the influence of absolute distance factors such as well spacing, focusing directly on the speed difference between the two mechanisms of pressure transmission and fluid propagation. In homogeneous formations, this difference should remain relatively stable. When the underground flow environment changes, such as the emergence of a water-dominant channel, the water flow velocity will abnormally increase, leading to a significant shortening of the water cut time delay, thus reducing the time delay difference. Conversely, when injection is obstructed or reservoir conditions deteriorate, fluid propagation becomes more difficult, and the water cut time delay may abnormally increase, leading to a larger time delay difference. Therefore, this characterization value has a unique indicative role in identifying specific flow anomaly patterns and provides a core parameter that complements the connectivity consistency characterization value, specifically designed to capture changes in flow differences.

[0072] Specifically, using continuous tracking periods as a time series, the connectivity consistency representation value and the response difference representation value corresponding to each tracking period are plotted on the same chart to generate a trajectory diagram representing the evolution of the two over time, which is recorded as the injection and sampling dynamic fingerprint diagram.

[0073] In implementation, the horizontal axis represents natural time or continuous tracking cycle number, and the vertical axis represents the connectivity consistency characterization value and response difference characterization value, respectively. After each tracking cycle, the two calculated characterization values ​​are taken as a data point and marked at the corresponding time position in the chart. The data points of consecutive cycles are connected by lines to form two curves that evolve over time, which enables long-term trend management of injection and collection dynamics and historical status tracing.

[0074] By stitching together the analysis results of each short period in chronological order, slow trends, periodic patterns, or abrupt events that cannot be observed within a single period can be revealed. This visualization method transforms massive, abstract data analysis results into a highly condensed comprehensive graph. The overall shape, trend, and local anomalies of the waveform together constitute a unique dynamic identifier for the well group, providing the most intuitive and reliable graphical basis for subsequent stable state classification, automatic anomaly identification, and evaluation of control effects.

[0075] Step S5, specifically, classify the states according to the fluctuation amplitude of the connectivity consistency characterization value and the response difference characterization value in the injection and acquisition dynamic fingerprint map. When the change amplitude of the connectivity consistency characterization value and the response difference characterization value in multiple consecutive tracking periods is less than the corresponding stability threshold, the map segment corresponding to the consecutive time period is divided into the stable sub-map. When the change amplitude of at least one of the connectivity consistency characterization value and the response difference characterization value is greater than or equal to the corresponding stability threshold, the map segment corresponding to the time period is divided into the fluctuation sub-map.

[0076] In practice, stability thresholds are typically not theoretically calculated values, but rather obtained through statistical calibration based on historical data from a specific oilfield injection-production unit during a known stable production phase. In one implementation, a historical period verified by both field experience and production data as stable is selected, and the connectivity consistency characterization value and response difference characterization value sequences calculated for all tracking cycles within that period are extracted. The variation amplitudes of these two sequences are calculated, preferably by adding 2 to 3 times the standard deviation to the mean. These two amplitude limits are then set as the stability thresholds for the corresponding characterization values. It can be understood that using the natural fluctuation range of characterization values ​​from a historical stable phase as a standard for judging whether future cycles will remain stable allows for the establishment of a dynamic benchmark that matches the specific geological conditions and production status of the oilfield. This ensures that the state classification adapts to the individual characteristics of different oilfields while avoiding subjective arbitrariness.

[0077] Classifying states based on fluctuation amplitude is essentially applying the principles of statistical process control to perform online pattern recognition on time series data. Specifically, the system uses a sliding time window to calculate the change amplitude of two characteristic values ​​over a consecutive number of tracking periods in real time. Generally, 5 to 10 consecutive periods are selected, and this real-time amplitude is compared with the corresponding stability threshold. Setting the number of consecutive periods aims to prevent misjudgments of the state due to accidental noise or minor fluctuations in a single period, ensuring that the identified state changes have a trend. The system determines that the time period is in a stable state only when the change amplitudes of both characteristic values ​​within the consecutive window are consistently below their respective thresholds, and marks the corresponding segment of the fingerprint as a stable subgraph.

[0078] Accordingly, the implementation defines that if the change in either of the two characteristic values ​​reaches or exceeds the corresponding stability threshold within a certain time period, the system immediately classifies that time period as a fluctuating state, and the corresponding segment is marked as a fluctuation subgraph. It is understandable that a significant deviation of any key indicator may indicate that the system is beginning to deviate from a stable equilibrium state, requiring attention. The stable subgraph corresponds to the system's healthy baseline period, during which only continuous monitoring is needed without intervention; the fluctuation subgraph corresponds to the system's diagnostic window period. Once its duration exceeds expectations, it triggers subsequent in-depth anomaly identification processes, thereby achieving intelligent, quantitative, and segmented management of the long-term operating status of the injection and production system. This transforms continuous monitoring data streams into structured stable-fluctuating event sequences, greatly improving the efficiency of state perception and the timeliness of early warning under large amounts of data.

[0079] Specifically, when the duration of the fluctuation subgraph exceeds a preset duration, the injection and sampling anomaly feature category is identified. When the number of consecutive tracking cycles corresponding to the fluctuation subgraph exceeds a preset duration threshold, the changing trends of the connectivity consistency characterization value and the response difference characterization value corresponding to the fluctuation subgraph are extracted.

[0080] If the connectivity consistency characterization value shows a monotonically increasing trend and the response difference characterization value shows a monotonically decreasing trend, it is identified as a crossflow enhancement category; if the connectivity consistency characterization value shows a monotonically decreasing trend and the response difference characterization value shows a monotonically increasing trend, it is identified as a supply-limited category.

[0081] In implementation, the preset duration threshold is often calibrated based on the self-regulation and fluctuation duration characteristics exhibited by the oilfield injection and production system during historical stable operation. In one implementation, by statistically analyzing the duration of occasional, short-lived fluctuation segments during a large number of historical stable subgraph periods, and counting them in terms of the number of continuous tracking cycles (e.g., these short-lived fluctuations mostly last for several cycles), 95% of the typical duration of such short-lived fluctuations is set as the preset duration threshold. It is understood that short-term fluctuations that the system can quickly recover from through self-regulation in a healthy state are separated from persistent anomalies requiring external intervention. The purpose of setting this threshold is to avoid overreacting to short-lived, random disturbances or normal adjustment processes, ensuring that the classification and diagnosis process is only initiated when the abnormal state lasts for a sufficiently long time, exceeding the system's self-healing capacity. This significantly improves the stability and reliability of the anomaly identification process and reduces false alarms.

[0082] Extracting the trend corresponding to the fluctuation subplot essentially involves performing trend fitting analysis on two sequences: the connectivity consistency representation value and the response difference representation value, within the continuous fluctuation period. Specifically, linear regression can be used to determine whether the overall statistical trend of each representation value sequence within the fluctuation time window is monotonically increasing, monotonically decreasing, or without a significant trend. This allows for the extraction of directional core features representing the evolution of the system state from the seemingly chaotic fluctuation signal. It is understandable that different underground physical processes, such as crossflow formation and injection constraint, drive these two key parameters to produce specific combinations of directional evolution patterns. Trend analysis can effectively remove random noise and capture these patterns.

[0083] Based on the fundamental laws of seepage mechanics, when the dominant flow channel (crossflow) is enhanced, fluid flow resistance decreases, and both pressure transmission and fluid propulsion accelerate, with the velocity difference between the two narrowing. Mathematically, this manifests as continuously improving connectivity and a continuously decreasing time delay difference. Conversely, when injection is obstructed or reservoir conditions deteriorate, flow resistance increases, the response weakens and slows, and the difference between the two becomes more pronounced. This manifests as continuously deteriorating connectivity and a continuously increasing time delay difference. This invention uses a combination of two parameters for discrimination, forming cross-validation. This makes the identification of crossflow or confinement no longer dependent on the accidental changes of a single indicator, thereby greatly improving the accuracy and anti-interference ability of diagnostic conclusions and laying a solid and reliable decision-making foundation for implementing differentiated and targeted control.

[0084] Step S6, specifically, involves selecting the corresponding injection-production well combination as the target control combination based on the identified abnormal feature categories. For target combinations with enhanced crossflow, the injection intensity is reduced and / or redistributed. Instructions are sent to the downhole intelligent regulator or surface flow control device of the target injection well to reduce the total injection volume by a preset margin, based on the current injection intensity. For example, the injection volume is reduced by 5% to 10% according to the preset first-level margin. Simultaneously, if the well has stratified injection capability, a layer reassignment instruction is generated, for example, reducing the injection ratio of high-permeability layers and correspondingly increasing the injection ratio of low-permeability layers. Crossflow is essentially the ineffective circulation of injected water in high-permeability channels. By reducing the total intensity, the driving energy is weakened, and by redistributing the flow between layers, the fluid is redirected. This blocks the dominant channels from both the quantity and direction dimensions, achieving active suppression and flow field guidance of crossflow phenomena.

[0085] For target combinations of injection-limited categories, injection intensity is increased and / or pulsed injection is applied to enhance the injection intensity of the target injection well, for example, by increasing it by 5% to 15% according to a preset second-level amplitude. Simultaneously, a periodic pulsed injection scheme can be generated, instructing the injection equipment to alternate between periods of high-pressure injection and periods of low-pressure or no injection, utilizing the pressure excitation effect to overcome near-wellbore blockage or activate low-permeability areas. Limitations often stem from insufficient injection pressure or increased flow resistance. By increasing energy supply and applying periodic pressure fluctuations, flow bottlenecks are overcome and the effective injection range is expanded. The larger the response difference characterization value, the more difficult fluid propulsion is; therefore, a higher pulse pressure upper limit, a longer single high-pressure injection duration, or an increased number of pulse cycles are set. The specific mapping relationship can be calibrated through finite-number simulations or field pilot tests for the target reservoir.

[0086] In practice, the adjustment ranges in each level of control strategy are not fixed values. Their setting is an engineering calibration process deeply coupled with the specific oilfield injection-production unit. Under the premise of ensuring wellbore integrity and formation safety, the goal is to find a control level that can have a clear and monitorable impact on the current reservoir dynamics. Implementation typically relies on a comprehensive analysis of two types of data: static geological and engineering data of the oilfield, such as reservoir permeability range, fracture pressure, and tubing pressure limits, which constitute the absolute boundary of the control; and dynamic control test data of the injection-production unit during historical stable periods. Through a limited number of step-by-step injection intensity adjustment tests, the minimum and effective adjustment amounts when key parameters of the production well produce clear and repeatable responses are observed and recorded, thereby determining the effective starting point and reasonable range of the control.

[0087] For example, by conducting injection intensity tests on the target well group with gradual, stepwise variations, an effective adjustment window can be determined that elicits a reliable formation response without inducing production fluctuations. The purpose of this process is to ensure that automatically generated control commands are both safe and effective, thereby guaranteeing that the intensity of each control command is within the safe range allowed by the geological and engineering conditions of the well group.

[0088] Step S7, specifically, after executing the hierarchical control strategy for several tracking cycles, the injection-sampling dynamic fingerprint is updated based on the newly acquired data; it is determined whether the duration of the fluctuation subgraph in the updated injection-sampling dynamic fingerprint is less than or equal to the preset duration, and whether the connectivity consistency representation value and the response difference representation value are both within the stable benchmark range; if the above conditions are met simultaneously, it is determined that the control for the current abnormal feature is completed; otherwise, the hierarchical control strategy is iteratively updated.

[0089] In practice, the preset duration is usually determined by referring to the recovery cycle of operating condition fluctuations in oilfield production management and combining it with historical data analysis of this method. The statistical analysis of the typical number of tracking cycles experienced from the execution of control to the disappearance of the fluctuation subgraph in previous successful control cases is used as the lower limit of the statistics as this threshold. Generally, the value is 3 to 5 cycles. The preset duration is used to determine whether the current system has responded within a reasonable time.

[0090] The stable benchmark range is based on data from a large number of historical stable subgraph phases. Statistical analysis is performed on connectivity consistency representation values ​​and response difference representation values, calculating their mean and standard deviation. The interval formed by adding or subtracting a certain number of standard deviations from the mean is taken as the benchmark range. Using the historical health state fluctuation range of the system itself as the regression target ensures the individualization and objectivity of the judgment criteria, while also meeting the time and state requirements.

[0091] After the control strategy is implemented, several tracking cycles continue to run, during which data is continuously collected and new injection-production dynamic fingerprint maps are generated. Subsequently, the algorithm automatically detects the existence and duration of fluctuation subgraphs in the new fingerprint map and reads the two most recently calculated characterization values. The fluctuation duration, connectivity value, and difference value are compared and judged with the two preset scales mentioned above. It is understandable that by introducing a necessary effect observation period, misjudgments due to the instantaneous response of the system or control delays are avoided. The response and adjustment of the formation require a certain amount of physical time; only after observation for several consecutive cycles can it be confirmed whether the system has established a new, stable dynamic equilibrium.

[0092] The iterative update of the hierarchical control strategy treats each control attempt as the start of a new observation and diagnosis cycle, strictly adhering to the complete methodology of this implementation for repetitive calculations and judgments. Specifically, if the updated injection-progression dynamic fingerprint map indicates that control is incomplete, the host computer does not simply repeat the control instructions from the previous round. Instead, it uses the newly generated dynamic fingerprint map, which has not yet stabilized, as input for the current system state and again executes the complete analysis process from state division to anomaly identification. Based on the connectivity consistency characterization value and response difference characterization value calculated from the latest data, and the trend characteristics they reflect, the system re-evaluates the type and degree of the abnormal state. Based on the results of this re-evaluation, it initiates a new round of target control combination screening and hierarchical control strategy generation logic. This process will continue cyclically until the system continuously monitors newly generated fingerprint maps that meet all the criteria for control completion, thus achieving intelligent adaptive iteration of dynamically adjusting intervention measures based on real-time system feedback.

[0093] By applying identifiable, limited-amplitude micro-perturbations to the injection wells during each tracking cycle, the system actively stimulates formation responses and simultaneously acquires pressure and water cut change signals from the production wells. Then, using correlation analysis and time delay estimation techniques, the dynamic correlation between injection and production is quantified into connectivity consistency and response difference characterization values. Based on these time series, a dynamic injection-production fingerprint map is constructed that intuitively reflects the system's health status. By automatically identifying abnormal fluctuation patterns in the fingerprint map, the system can diagnose problem types and severity in real time and automatically generate and execute graded control strategies. Finally, by continuously updating the fingerprint map and verifying whether fluctuations have been eliminated and parameters have returned to stable baselines, the system significantly improves its early identification capability and precise treatment efficiency for anomalies such as crossflow and injection limitation, thereby optimizing the injection-production structure, improving displacement effects, and enhancing oilfield development benefits.

[0094] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0095] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for dynamic real-time tracking and control of injection and extraction, characterized in that, include: Obtain the injection flow rate and injection pressure of the injection well, and obtain the bottom hole pressure, production rate and water cut of the production well; Within each preset tracking cycle, the injection well is controlled to perform limited micro-disturbance injection, so that the injection flow rate and injection pressure change according to a preset coding sequence, and the disturbance signal corresponding to the coding sequence is recorded as the injection reference signal; Obtain the production well response signal corresponding to the injection reference signal time. The response signal includes a bottom hole pressure change sequence and a water cut change sequence. Generate an injection-production disturbance response curve based on the injection reference signal and the response signal. The injection-production disturbance response curve is analyzed, and the connectivity consistency characterization value and response difference characterization value of the combination of injection well and production well are determined based on the analysis results. Based on the changes of the connectivity consistency characterization value and response difference characterization value in the time dimension, a dynamic fingerprint map of injection and production is generated. The injection and sampling dynamic fingerprint is divided into a stable subgraph and a fluctuating subgraph. When the duration of the fluctuating subgraph is longer than a preset duration, the injection and sampling abnormal feature categories are identified. The injection and sampling abnormal feature categories include crossflow enhancement category and injection supply restriction category. The target control well combination is determined based on the injection and production anomaly characteristics category, and a corresponding graded control strategy is generated. The graded control strategy includes reducing the injection intensity of the corresponding injection well, and redistributing the injection intensity between different layers and / or applying pulsed injection to eliminate the restricted state. After executing the hierarchical control strategy, the injection-progression dynamic fingerprint is updated. When the duration of the fluctuation subgraph falls back to within the preset duration and the connectivity consistency representation value and the response difference representation value return to the stable benchmark range, the control is determined to be completed; otherwise, the hierarchical control strategy is iteratively updated.

2. The injection-production dynamic real-time tracking and control method according to claim 1, characterized in that, The control injection well to perform limited micro-disturbance injection within each preset tracking cycle includes: Within each preset tracking period, the injection flow rate and injection pressure of the injection well are modulated according to a preset coding sequence to generate a serialized perturbation with identifiable characteristics. The amplitude change of the micro-perturbation injection does not exceed a preset amplitude threshold; The preset encoding sequence is a pseudo-random binary sequence or a sequence consisting of alternating positive and negative steps, and a complete encoding sequence is executed within one preset tracking period.

3. The injection-production dynamic real-time tracking and control method according to claim 2, characterized in that, Generate an injection-progression perturbation response curve based on the injection reference signal and the response signal, including: Real-time acquisition of bottom hole pressure and water cut data of the produced well, and time synchronization and alignment with the injection reference signal; From the time-synchronized bottom hole pressure data, extract the pressure change sequence within the same time period as the injection reference signal, and use it as the first response signal; From the time-synchronized and aligned moisture content data, extract the moisture content change sequence within the same time period as the injected reference signal, and use it as the second response signal; The pressure response curve generated based on the first response signal, and the moisture content response curve generated based on the second response signal; The injection-production disturbance response curve includes the pressure response curve and the water cut response curve.

4. The injection-production dynamic real-time tracking and control method according to claim 3, characterized in that, Based on the analysis results, the connectivity consistency characterization value and response difference characterization value of the injection well and production well combination are determined, including: Correlation analysis was performed on the pressure response curve and the water content response curve respectively to obtain the maximum cross-correlation coefficient between the injection reference signal and the corresponding response signal. The maximum cross-correlation coefficient corresponding to the pressure response curve was used as the connectivity consistency characterization value. Time delay estimation is performed on the pressure response curve and the moisture content response curve respectively to obtain the corresponding pressure time delay and moisture content time delay; The difference between the pressure time delay and the moisture content time delay is calculated and used as the response difference characterization value.

5. The injection-production dynamic real-time tracking and control method according to claim 4, characterized in that, Using continuous tracking periods as a time series, the connectivity consistency representation value and the response difference representation value corresponding to each tracking period are plotted on the same chart to generate a trajectory diagram representing the evolution of the two over time, which is recorded as the injection and sampling dynamic fingerprint.

6. The injection-production dynamic real-time tracking and control method according to claim 5, characterized in that, The injection-collection dynamic fingerprint is divided into a stable subgraph and a fluctuating subgraph, including: The state is divided according to the fluctuation amplitude of the connectivity consistency characterization value and the response difference characterization value in the injection and sampling dynamic fingerprint map; When the change magnitude of the connectivity consistency characterization value and the response difference characterization value is less than the corresponding stability threshold in multiple consecutive tracking periods, the graph segment corresponding to the consecutive time period is divided into the stable subgraph. When the change magnitude of at least one of the connectivity consistency characterization value and the response difference characterization value is greater than or equal to the corresponding stability threshold, the graph segment corresponding to that time period is divided into the fluctuation subgraph.

7. The injection-production dynamic real-time tracking and control method according to claim 6, characterized in that, When the duration of the fluctuation subgraph exceeds a preset duration, the categories of injection and sampling anomalies are identified, including: When the number of consecutive tracking cycles corresponding to the fluctuation subgraph exceeds a preset duration threshold, the changing trends of the connectivity consistency representation value and the response difference representation value corresponding to the fluctuation subgraph are extracted. If the connectivity consistency characterization value shows a monotonically increasing trend and the response difference characterization value shows a monotonically decreasing trend, then it is identified as a crosstalk enhancement category; If the connectivity consistency characterization value shows a monotonically decreasing trend and the response difference characterization value shows a monotonically increasing trend, then it is identified as a supply-limited category.

8. The injection-production dynamic real-time tracking and control method according to claim 7, characterized in that, The target control well combination is determined based on the injection-production anomaly characteristic category, and a corresponding graded control strategy is generated, including: Based on the identified abnormal feature categories, the corresponding injection well-production well combinations are selected as the target control combinations; If the target control combination is of the crossflow enhancement category, then a graded control strategy is generated to reduce the injection intensity of the corresponding injection well and / or redistribute the injection intensity between different layers. If the target control combination is a supply-limited category, the generated graded control strategy includes increasing the injection intensity of the corresponding injection well and / or applying a graded control strategy of pulsed injection.

9. The injection-production dynamic real-time tracking and control method according to claim 8, characterized in that, Whether the update is complete is determined based on the updated injection-progression dynamic fingerprint after the execution of the tiered control strategy, including: After several tracking cycles following the execution of the tiered control strategy, the injection-progression dynamic fingerprint is updated based on the newly acquired data; In the updated injection-collection dynamic fingerprint graph, it is determined whether the duration of the fluctuation subgraph is less than or equal to the preset duration, and whether the connectivity consistency characterization value and the response difference characterization value are both within the range of the stable benchmark. If all of the above conditions are met, then the regulation targeting the current abnormal characteristics is considered complete.

Images