A fracturing operation real-time monitoring and intelligent regulation and control system, method and storage medium

Abstract

The application discloses a fracturing operation real-time monitoring and intelligent regulation and control system and method and a storage medium, and belongs to the field of oil and gas field hydraulic fracturing operation monitoring and regulation and control. In view of the inconsistency of time and stage benchmarks of multi-source data, a "state machine + sliding window multi-parameter voting" is adopted to realize stage alignment based on a pump injection scheme table, and mismatch misjudgment is eliminated. In view of stage boundary jitter, transition is introduced to eliminate jitter and reduce alarm fatigue. In view of the lack of stage semantics in rule research and judgment, "stage gating filtering + priority grading" is adopted to realize differentiated research and judgment of the same phenomenon in different stages, and cross-stage false reports are reduced. In view of the disconnection between monitoring and treatment, treatment suggestions are fed back to the operation regulation and control end to form a closed loop.

Description

Technical Field

[0001] This invention relates to the field of monitoring and control technology for hydraulic fracturing operations in oil and gas fields, specifically to a real-time monitoring and intelligent control system, method, and storage medium for fracturing operations, and specifically to the stage identification and alignment, hierarchical rule analysis, and closed-loop processing of multi-source real-time data during fracturing operations. Background Technology

[0002] Hydraulic fracturing is a core technology for enhancing oil and gas reservoir production. During operation, it is essential to monitor multiple parameters in real time, including wellhead pressure, pump pressure, displacement, and sand concentration, and dynamically adjust these parameters accordingly to ensure both operational safety (preventing bottomhole pressure over-limits, sand shedding, and sand blockage) and operational quality (ensuring net pressure and sand concentration). The fracturing process typically involves stages such as pre-flush fluid, sand carrying, displacement, and shut-in pressure reduction. Each stage corresponds to different design displacement, design sand concentration, and design fluid volume in the pumping plan table. Therefore, accurately aligning real-time monitoring data with the stage references in the pumping plan table is a prerequisite for effective monitoring and control.

[0003] To address the aforementioned needs, several solutions have been proposed in the existing technology, but all have limitations: Chinese patent document CN113818857A uses stage liquid volume as a single threshold to drive stage advancement and forced intervention, without incorporating multiple parameters such as discharge rate and sand concentration into the matching, and lacks a de-shaking mechanism at stage boundaries, making it prone to stage misjudgment due to parameter fluctuations; Chinese patent document CN121765554A overlays historical construction curves to generate a standard curve and upper and lower envelope surfaces and aligns them according to the time axis for early warning, but does not recognize the inherent stage division of the pumping scheme table, nor does it distinguish the semantic differences of different stages for the same phenomenon (such as pressure rise); Yuan Bin et al. (Yuan Bin, Zhao Mingze, Meng Siwei, Zhang Wei, Zheng He. Intelligent identification and early warning method for multiple complex events in horizontal well fracturing [J]. Petroleum Exploration and Development, 2023, 50(6):1298-1306.) The deep learning model based on Unet++ and Att-BiLSTM identifies stage events and predicts pressure. It is a black box model, relies on a large number of labeled samples, and is not bound to the pump injection scheme design stage. It cannot provide interpretable stage gating semantics and disposal suggestions. Chinese patent document CN115410347A only targets the single abnormal early warning of sand blockage and does not construct a complete monitoring and control link of "collection - stage alignment - gating judgment - closed-loop disposal".

[0004] In summary, the existing solutions have the following systemic defects: (1) The phase advancement is driven by time or a single liquid volume, which can easily lead to misjudgment of the phase in cases of time mismatch caused by pauses, network outages, and data re-entry; (2) There are transition zones for discharge and sand concentration at the phase boundaries, and the single threshold frequently jumps, causing alarm fatigue; (3) The rule judgment lacks phase gating and does not take the phase as a pre-gating condition, resulting in a large number of cross-phase false alarms; (4) Monitoring and disposal are disconnected, stopping at alarm display and failing to form a closed loop for feedback construction; (5) The data acquisition deployment is heavy and prone to crashing, and the data cannot be automatically replenished after interruption. The high-frequency data at the second level is repeatedly written, the retrieval is slow, and the file expands. There is an urgent need for a real-time monitoring and control system for fracturing construction that aligns multi-source real-time data under a unified time and phase benchmark, performs gated hierarchical judgment based on this, forms a closed-loop disposal, and takes into account both lightweight acquisition and reliable storage. Summary of the Invention

[0005] The purpose of this invention is to provide a real-time monitoring and intelligent control system, method, and storage medium for fracturing operations. This invention addresses the inconsistency between time and stage benchmarks in multi-source data by using a pump injection scheme table as a benchmark and employing a "state machine + sliding window multi-parameter voting" approach to achieve stage alignment and eliminate mismatches and misjudgments. For stage boundary jitter, it introduces transition de-jittering to reduce alarm fatigue. To address the lack of stage semantics in rule analysis, it uses "stage gating filtering + priority grading" to achieve differentiated analysis of the same phenomenon across stages, reducing cross-stage false alarms. Finally, to address the disconnect between monitoring and response, it feeds back response suggestions to the construction control end to form a closed loop.

[0006] The technical problem to be solved by the present invention is achieved by the following technical solution: a real-time monitoring and intelligent control system for fracturing construction, including a data acquisition module, a stage alignment module, a rule analysis module, an early warning and disposal module, and a data storage module; The data acquisition module is used to periodically... Acquire multi-source construction data and associate and merge them with construction wells and timestamps to form a real-time data stream. The multi-source construction data includes at least wellhead pressure, pump pressure, discharge rate, sand concentration and corresponding construction timestamps. The data storage module is used to partition data by "hash number + construction date", perform batch transaction writing and self-healing deduplication when building a unique index, and provide a traceable data access interface for each module. The stage alignment module is used to identify stages based on the pumping scheme table using a "state machine + sliding window multi-parameter voting" approach, introduce transition debouncing, and output stage identifiers, cumulative fluid volume, and pump stop / well shut-in status under a unified time reference. The pumping scheme table includes the design discharge rate, design sand concentration, and design liquid volume for each stage; The rule analysis module is used to perform stage-based gating and priority-based analysis based on a preset rule base, and output the phenomenon, current value, handling suggestions, and priority attributes. Each rule in the preset rule base has a stage gating attribute, priority attribute, phenomenon description, handling suggestion and rule type; The early warning and response module is used to generate graded alarms based on the urgency of the response recommendations and to feed back the response recommendations to the data acquisition module, the stage alignment module, and the construction control terminal.

[0007] Preferably, the present invention also includes a manual intervention module, which is used to manually align the stage indicators, cumulative fluid volume, and pump stop / well shut-off status under a unified time reference when the stage alignment module cannot accurately align them. This is used for forced switching of real-time mode and reuse of intervention records in post-replay mode.

[0008] This invention also discloses a method for real-time monitoring and intelligent control of fracturing operations, which employs the aforementioned real-time monitoring and intelligent control system for fracturing operations and includes the following steps: S1. Multi-source data acquisition and merging The system acquires multi-source construction data from the construction data platform in real time using a periodic Tpoll, and then associates and merges the data according to the construction well and timestamp to form a real-time data stream with a unified time sequence. The multi-source construction data includes at least wellhead pressure, pump pressure, discharge rate, sand concentration and corresponding construction timestamp. S2, Phase Alignment The current pumping stage is maintained by a state machine. The measured data points of step S1 are mapped to the corresponding design stage in the pumping scheme table by means of "sliding window multi-parameter voting". Jump de-bouncing is introduced, a stage identifier under a unified time reference is established, and the stage identifier, stage cumulative liquid volume, pump stop status and well shut-in status under a unified time reference are output. The pumping scheme table includes the design discharge rate, design sand concentration, and design liquid volume for each stage; S3, Rule Analysis Input the stage identifier, cumulative liquid volume, pump stop status and well shut-in status of step S2, load the preset rule base, perform stage-gated filtering on the rules according to the stage identifier of step S2, and then compare the measured values ​​of the condition fields with the thresholds one by one to obtain the judgment results. Output the phenomenon description, current measured value, treatment suggestions and priority. Each rule in the preset rule base has a stage gating attribute, priority attribute, phenomenon description, handling suggestion and rule type; S4, Closed-loop processing Based on the output of step S3, a graded alarm is generated according to the urgency of the handling recommendations, and the handling recommendations are fed back to the data acquisition module, the stage alignment module, and the construction control terminal to form a closed-loop monitoring and control link.

[0009] Preferably, in step S1 of this invention, the data acquisition employs the following lightweight mechanism: (1) Pure HTTP authentication and data collection The system completes the identity authentication of the intranet industrial platform using an HTTP request and directly calls the platform's real-time operating status interface to poll for data, without starting any browser process throughout the process. (2) Concurrency of real-time streaming and historical data backup Upon startup, the background concurrently retrieves all historical second-level data from the platform, without blocking real-time polling; When the data acquisition pauses and resumes beyond the threshold or the gap reopens, automatic re-filling is triggered. (3) Discontinuity gap index retreat and filling If the historical recovery fails, do not give up. Retry according to the exponential backoff interval until successful. The backoff interval is calculated according to the following formula and is capped at 60 seconds. (10) In the formula, For the first The waiting interval before the next retry is s; Let s be the initial interval; (4) Whitelist filtering of operating wells Real-time operating results are filtered according to the platform's "currently operating wells" whitelist, collecting only wells actually under construction to ensure consistency with the objects displayed on the platform page. If the whitelist is not loaded, filtering is disabled. This invention completes data collection at a pure HTTP layer and uses concurrent breakpoints to fill gaps. Partitioning and deduplication storage support high-frequency writes at the second level, ensuring data integrity and traceability. This invention allows the collection end to complete platform authentication and polling without starting a browser process; breakpoint gaps can be automatically filled; and high-frequency data at the second level is written using partitioning and unique indexes for self-healing deduplication, avoiding duplicate writes, slower retrieval, and file bloat.

[0010] Preferably, in step S2 of this invention, the following sub-steps are performed for each measured data point: (1) Multi-parameter voting via sliding window The maintenance capacity for discharge volume and sand concentration is respectively: A sliding window queue is used to enqueue points one by one and remove points at the head of the queue that exceed the window. The window mean is taken as the input for the stage decision to suppress single-point jitter. (1) In the formula, For the first in the window Measured displacement of each data point, m³ / min; The mean displacement measured within the window is m³ / min; For the first in the window Measured sand concentration at each data point, kg / m³; The average sand concentration measured within the window is expressed in kg / m³. The number of sliding windows is [number]. (2) Pump shutdown detection and well shut-in identification Pumping is stopped when the average discharge rate of the window is lower than the pump stop discharge rate threshold. (2) In the formula, The pump stoppage discharge threshold is set in m³ / min. If the average discharge rate of the window is lower than this threshold, the pump is stopped. During the pump stoppage, the liquid volume is not accumulated and the stage is not switched, but the current stage information continues to be output. When the pump stoppage status continues to reach the preset duration threshold, the well shut-in flag is set, and pressure drop monitoring rules are executed during the well shut-in pressure drop stage. (3) Cumulative actual liquid volume When the pump is running, the actual liquid volume of the current stage is accumulated, but accumulation is not performed if the time interval between adjacent data points is greater than the data interruption threshold. (3) In the formula, The current cumulative actual liquid volume, in m³; The time interval between adjacent data points is s; Where t is the current data point collection time, t prev That is, t-1 is the time of the previous data point collection; Let s be the data interruption threshold. V acc (t-1) This represents the cumulative actual liquid volume corresponding to the previous data point. (4) Stage transition determination When a next stage exists, calculate the sand concentration matching indicator, discharge volume matching indicator, and liquid volume compliance indicator separately: (4) (5) (6) In the formula, M C M is a marker for sand concentration matching. Q For displacement matching; M V This is the liquid volume matching flag, and can be either 0 or 1. The sand concentration for the next stage is specified in kg / m³. The sand concentration tolerance is expressed in kg / m³. The displacement for the next stage is designed in m³ / min; The displacement tolerance percentage is %; The designed liquid volume for the current stage is in m³. The volume ratio is dimensionless. The transition decision is based on matching sand concentration and discharge rate, or meeting liquid volume requirements: (7) In the formula, This is the jump trigger flag, either 0 or 1; Setting it to 0 indicates that the transition condition is not met. A value of 1 indicates that the condition is true; (5) Transition de-jitter To suppress false transitions caused by stage boundary parameter jitter, transition conditions must be met continuously for a preset number of data points before a transition is submitted: (8) In the formula, The number of data points that continuously satisfy the transition condition; To debouncing the points, the number is reset to zero if the condition is interrupted; when... Submit the jump at the appropriate time; (6) Timeout anomaly detection If the actual duration of the current stage exceeds 1.5 times the designed duration and the cumulative liquid volume is less than 50% of the designed liquid volume, it is judged as a timeout anomaly, and an inquiry is sent to the operator in real-time mode: (9) In the formula, The actual duration of the stage is in seconds. The design duration is in seconds (s). (s), The design displacement for the current stage is m³ / min; if the above formula is satisfied, it is judged as a timeout abnormality, and in real-time mode, the operator is asked whether to force entry into the next stage. (7) Unification of manual intervention and playback In real-time mode, when the algorithm cannot determine the stage or determines that there is a timeout, it sends a query to the operator and receives a forced stage switching instruction. After the forced switch, the current stage's cumulative liquid volume is reset and the manual intervention record is recorded. In post-event playback mode, pre-recorded human intervention records are automatically applied, and when all sand-carrying stages are completed, the current sand concentration is below the threshold and it is not the last stage, the remaining stages are automatically identified as replacement stages. The real-time and playback modes share the same core alignment algorithm.

[0011] Preferably, in step S2 of this invention, the number of transition debouncing points is determined according to the following priority: Debounce points Determined according to the following rules: Set a default value; when the difference between the design sand concentration and the design discharge rate in two adjacent stages does not exceed... And the displacement difference does not exceed hour, The default value is set to a preset multiple; when the liquid volume reaches the standard mark. hour, Take 1; When the liquid volume compliance flag is 1, the preset minimum number of de-vibration points is used to skip de-vibration and directly jump to the next stage; when the design sand concentration difference between two adjacent stages does not exceed the preset sand concentration tolerance and the design discharge rate difference does not exceed the preset discharge rate tolerance percentage, a preset multiple of the preset default de-vibration point number is used; otherwise, the preset default de-vibration point number is used. That is, when the liquid volume compliance flag MV=1, the number of de-vibration points... Choose 1, skip debouncing and submit the transition directly; when MV=0 and the design sand concentration difference between two adjacent stages does not exceed the sand concentration tolerance ΔC, and the design discharge rate difference does not exceed the discharge rate tolerance percentage. hour, Take a preset multiple of the default debouncing points; otherwise... Use the preset default debouncing points.

[0012] The pump stop detection is determined when the average discharge rate of the window is lower than the preset pump stop discharge rate threshold. During the pump stop period, no liquid volume is accumulated and no stage is switched. When the pump stop continues for a preset duration threshold, a well shut-off flag is set.

[0013] Preferably, in step S3 of this invention, the rule evaluation is performed according to the following mechanism: (1) Stage gating (authoritative gating): The "applicable stage" attribute of the rule is used as the authoritative gating field. Before the analysis, the rule is first filtered by stage gating according to the current stage identifier. The phase flag is not retained in the rule sub-conditions to avoid duplicate or contradictory gating. The applicable stage values ​​include all stages, pre-fluid, sand carrying, displacement, and shut-in pressure drop. (2) Single condition and compound condition: The rule condition is a single condition or a compound condition. A compound condition is when multiple sub-conditions are satisfied at the same time. The measured value of each condition field is calculated and compared with the threshold. (3) Priority classification: The priority is divided into multiple levels, such as P0, P1...PN, and is executed in order of priority from high to low. When any P0 level rule is hit, only the P0 level judgment result is presented to suppress low priority noise, but other concurrent P0 level results are not discarded. (4) Decoupling of phenomenon and treatment: When a rule is hit, the phenomenon description, the current measured value of the corresponding condition field when it is triggered, the treatment suggestion and priority are recorded at the same time, so that they can be displayed in the panel and report simultaneously; (5) Deduplication: When the same rule is hit multiple times in the data stream, only the first hit result is retained to avoid duplicate alarms.

[0014] Preferably, in step S3 of this invention, the rule base includes rules for bottom hole pressure exceeding limits, bottom hole pressure being too low, bottom hole pressure dropping sharply, abnormal rise in bottom hole pressure during the sand-carrying stage, insufficient net pressure, wellhead pressure exceeding limits, abnormal rise in wellhead pressure during the sand-carrying stage, pressure rise during the high sand ratio stage, low discharge rate during the sand-carrying stage, sand blockage determination, pressure rise during the displacement stage, pressure buildup, excessively high net pressure, sand concentration exceeding the design upper limit, sand concentration lower than the design, discharge rate exceeding the design upper limit, and shut-in pressure drop rate monitoring rules.

[0015] Preferably, in step S4 of this invention, the highest-level rule is displayed upon a single hit, while other-level rules are displayed only after a preset number of consecutive hits. The interval between two adjacent alarms for the same rule is not less than a preset duration, and the same alarm is not displayed again until it is cleared. When any highest-level rule is hit, only the highest-level judgment result is presented. When multiple highest-level rules are hit at the same time, the handling suggestion is to take the highest-level instruction according to the priority of urgency. In step S4, the proposed solutions are categorized by urgency into emergency pump shutdown, parameter adjustment, replacement of the clear water sluice blockage, alarm notification, and monitoring notification, and are distributed to the construction control terminal, the stage alignment module, and the data acquisition module. This invention ensures that when the same rule is repeatedly hit in the data stream, only the first hit result is retained, and alarm presentation is controlled based on the number of consecutive hits and the minimum presentation interval, suppressing alarm flashing caused by momentary jitter.

[0016] The present invention also discloses a storage medium storing a program thereon, which, when executed by a processor, implements the steps of the above-described method for real-time monitoring and intelligent control of fracturing operations.

[0017] Compared with the prior art, the beneficial effects of the present invention are: Effect 1 (Unified stage benchmarks reduce misjudgments): This invention significantly reduces false jumps and cross-stage false alarms near stage boundaries. Specifically, it involves offline playback of historical construction data from the same well, counting the number of false jumps and cross-stage false alarms within a ±30s window at the stage boundary, and comparing the results with a comparative scheme that uses stage fluid volume as a single threshold to drive stage advancement without jump debounce. The statistical values ​​of this invention are lower than those of the comparative scheme. The playback data is taken from historical construction records on the construction data platform.

[0018] Related features: Stage alignment outputs stage identifiers under a unified time base using "state machine + sliding window multi-parameter voting" and applies continuous point debouncing, doubling the debouncing for adjacent stages with the same parameters and forcibly eliminating debouncing when the liquid volume meets the standard; rule analysis uses this stage identifier as the authoritative gating.

[0019] Effect 2 (Stage-based gating and differentiated analysis to reduce false alarms across stages): This invention enables the same pressure rise phenomenon to obtain different semantic judgment results at different stages, significantly reducing alarm fatigue. Specifically, it involves offline playback of the same historical construction data and counting the number of alarms triggered by the same wellhead pressure rise phenomenon in the sand-carrying stage and the displacement stage. After stage-gated filtering, the rules of this invention only generate alarms within the gated stage, and the number of false alarms across stages is lower than that of the comparison scheme without stage-gated filtering.

[0020] Related features: Rules have stage gating attributes and are graded by priority. When P0 is hit, low-priority noise is suppressed. The same rule is deduplicated and alarm confirmation count is added.

[0021] Effect 3 (Closed-loop processing, improved timeliness): This invention transforms the judgment conclusions into actionable instructions that can be fed back into construction, shortening the link between anomaly identification and action. Specifically, after an alarm is triggered, the action suggestions are directly distributed by the early warning and action module to the construction control module, the stage alignment module, and the data acquisition module according to the action type pre-marked by the rules, without the need for manual relay; the highest-level rule is presented on the first hit.

[0022] Related features: The early warning and response module feeds back response suggestions to the data collection, alignment and construction control end, and connects real-time mode and post-event playback mode with manual intervention recording. Attached Figure Description

[0023] Figure 1 A block diagram showing the modular components of a real-time monitoring and intelligent control system for fracturing operations; Figure 2 A flowchart illustrating the process of aligning stages; Figure 3 A flowchart illustrating the rule-based assessment process; Figure 4 A flowchart illustrating the steps of a real-time monitoring and intelligent control method for fracturing operations. Detailed Implementation

[0024] The present invention will be further described below with reference to the accompanying drawings and embodiments, but the scope of the invention is not limited thereto.

[0025] like Figure 1 As shown, a real-time monitoring and intelligent control system for fracturing operations includes a data acquisition module, a stage alignment module, a rule analysis module, an early warning and handling module, and a data storage module. This solution uses an online monitoring and control system for the fracturing process as its carrier, and can operate within the intranet environment of the fracturing site. During the operation, it provides real-time stage identification, anomaly alarms, and handling suggestions, forming a closed loop with the manual intervention process.

[0026] The data acquisition module is used to periodically... Multi-source construction data is acquired and aggregated into a real-time data stream by association with construction wells and timestamps. The multi-source construction data includes at least wellhead pressure, pump pressure, discharge rate, sand concentration, and corresponding construction timestamps. In this embodiment, the data acquisition module pulls real-time snapshots from the construction data platform at a cycle of approximately 1 second and aggregates them into a real-time data stream by association with construction wells and timestamps. For historical gaps not covered by the platform, breakpoint restoration is used to fill in the gaps in the optional implementation.

[0027] The data storage module is used to partition data by "well number + construction date", perform batch transaction writing, and self-heal deduplication when building a unique index, providing a traceable data access interface for each module. The data storage module provides a unified data access interface for each module, and its storage objects are the original multi-source real-time data (wellhead pressure, pump pressure, discharge rate, sand concentration, construction timestamp, etc.) and its metadata written by the data acquisition module. In addition, the intermediate judgment results and manual intervention records generated by other modules can also be persisted through this module for traceability and playback. Therefore, the storage objects are not limited to the data of the data acquisition module.

[0028] The data storage module employs a partitioning and deduplication mechanism: data is partitioned and stored according to hash number and construction date. A unique index is created for "hash number + construction date + collection time" during table creation. Writes utilize batch transactions, supporting high-frequency writes at the second level and fast retrieval, and providing traceable data access interfaces for each module. "Batch transactions" refers to encapsulating multiple database write operations within the same transaction and committing them together. This ensures that the batch of writes either all succeed or all fail and roll back, thus maintaining data consistency under high-frequency writes at the second level and significantly reducing disk I / O overhead, avoiding file bloat and slower retrieval caused by committing row by row. Specifically, data storage uses SQLite's pre-write log mode, partitioned by hash number and construction date, and a unique index is created for "hash number + construction date + collection time." If index creation fails, duplicate rows are deleted first (keeping the earliest row) before rebuilding, followed by batch transaction writes. The unique index's self-healing deduplication, combined with batch transaction writes, supports high-frequency writes at the second level and fast retrieval.

[0029] The stage alignment module is used to identify stages based on the pumping scheme table using a "state machine + sliding window multi-parameter voting" method, introduces transition debouncing, and outputs stage identifiers, cumulative fluid volume, and pump stop / well shut-in status under a unified time reference.

[0030] The pumping scheme table includes the design discharge rate, design sand concentration, and design liquid volume for each stage.

[0031] In this embodiment, the alignment configuration takes a sliding window Nw of 5 points, a sand concentration tolerance ΔC of 30 kg / m³, and a discharge tolerance δ. Q 20%, liquid volume completion ratio r c=0.95. For each measured data point: maintain the sliding window of displacement and sand concentration and take the average value; if the average displacement of the window is lower than 0.05m³ / min, the pump is stopped and the well is shut off if the pump stops for 30s; calculate the matching flag and transition decision according to the formula (4)-(7) below; the number of de-shaking points is 3 by default, 6 is taken for adjacent stages where the design sand concentration and displacement are close, and 1 is taken for direct transition when the liquid volume meets the standard; if the time interval between adjacent points is greater than 300s, the liquid volume is not accumulated; if the actual duration exceeds 1.5 times the design duration and the accumulated liquid volume is less than 50% of the design liquid volume, ask the operator in real-time mode whether to force entry into the next stage.

[0032] "Sliding window 5 points" refers to retaining the 5 most recent data points within the window and calculating their average. For example, when the continuously collected displacements are 2.0, 2.1, 1.9, 2.0, and 2.0 m³ / min, the average value of the window is approximately 2.0 m³ / min. Using this average value as the stage judgment input can eliminate single-point fluctuations.

[0033] The state machine transitions in the following sequence: "Unidentified → Pre-flush fluid → Sand carrying (may contain multiple stages) → Displacement → Pump stop / well shut-off". When the fluid volume reaches the target, it can skip the sand concentration / displacement matching and wait for forced transition.

[0034] In layman's terms, the state machine manages the construction process in the order of "unidentified → pre-fluid → sand carrying (which may contain multiple stages) → displacement → pump shutdown / well shut-in": At the start of construction, it first identifies and enters the pre-fluid stage, then progresses to sand carrying and displacement, and enters the well shut-in pressure drop stage after construction is completed; when pump shutdown is detected, the stage counting is paused (no accumulation of fluid volume, no switching of stages), and after the discharge volume recovers and remains stable for a period exceeding the preset time threshold, it automatically jumps back to the last stage before pump shutdown and continues execution; the well shut-in pressure drop stage is the final state of construction and does not jump back.

[0035] The rule analysis module is used to perform stage-based gating and priority-based analysis based on a preset rule base, and output the phenomenon, current value, handling suggestions, and priority attributes.

[0036] Each rule in the preset rule base has a stage gating attribute, priority attribute, phenomenon description, handling suggestion and rule type.

[0037] In this embodiment, the rule analysis module loads a default rule library (such as bottom hole pressure exceeding limit P0 emergency pump stoppage, bottom hole pressure sudden drop P0, bottom hole pressure rise during sand carrying stage P1 sand blockage precursor, shut-in pressure drop monitoring P2, etc.). Each rule has stage-gated attributes (all stages / pre-flush / sand carrying / displacement / shut-in pressure drop) and priority (P0–P3).

[0038] When making a judgment, first filter according to the current stage identifier, and then execute according to priority order; after a hit, record the phenomenon description, the current measured value (such as the actual value of the bottom hole pressure), the handling suggestions and priority; if the same rule is hit multiple times, only the first one is retained; if there is a P0 hit, only the P0 result is presented.

[0039] The early warning and response module is used to generate graded alarms based on the urgency of the response recommendations and to feed back the response recommendations to the data acquisition module, the stage alignment module, and the construction control terminal.

[0040] In this embodiment, the early warning and response module distinguishes between emergency pump stoppage, parameter adjustment, water sluice block replacement, alarm, and monitoring based on the analysis results, and generates construction control instructions. Before an alarm is displayed, the same rule must be hit a preset number of times (e.g., 3 times), and the interval between two consecutive alarms must be no less than 2 seconds. The type of response suggestion is pre-marked based on the nature of the hit rule: rules involving personal and equipment safety that require immediate pumping stoppage are marked as "emergency pump stoppage"; rules requiring adjustment of pumping parameters (e.g., flow rate, sand ratio) are marked as "parameter adjustment"; rules requiring the injection of clean water to replace mortar in the pipe section are marked as "water sluice block replacement"; rules only requiring operator attention without immediate action are marked as "alarm prompt" or "monitoring prompt". The early warning and response module generates and distributes corresponding instructions to the construction control terminal or the corresponding module for execution according to the response type marked for each rule.

[0041] The fracturing construction real-time monitoring and intelligent control system also includes a manual intervention module. The manual intervention module is used to manually align the stages when the stage alignment module cannot align them accurately, so as to output stage identifiers, cumulative fluid volume, and pump stop / well shut-in status under a unified time reference. This is used for forced switching of real-time mode and reuse of intervention records in post-event playback mode.

[0042] like Figures 2-4 As shown, a method for real-time monitoring and intelligent control of fracturing operations includes the following steps: S1. Multi-source data acquisition and merging The data acquisition module periodically (Example 1s, synchronized with platform data at the second level) Multi-source construction data is obtained in real time from the construction data platform and associated and merged according to the construction well and timestamp to form a real-time data stream with a unified time sequence. The multi-source construction data includes at least wellhead pressure, pump pressure, discharge rate, sand concentration and corresponding construction timestamp.

[0043] The data acquisition module also maintains the data storage channel, writing real-time data into the storage module.

[0044] In one implementation, data acquisition can employ the following lightweight mechanism: (1) Pure HTTP authentication and data collection The system completes the identity authentication of the intranet industrial platform with HTTP requests (login credentials are transmitted with symmetric encryption, multi-step single sign-on, token calculation and carrying), and directly calls the platform's real-time working status interface to poll data. No browser process is started throughout the process, and the deployment size is significantly smaller than that of the headless browser solution. (2) Concurrency of real-time streaming and historical data backup Upon startup, the background concurrently retrieves the platform's historical full-volume, second-level data (this historical interface only supports POST requests, and the request body must be constructed according to the platform's conventions and include session credentials), without blocking real-time polling; When data acquisition is paused and resumed for more than a threshold or when a gap is reopened, data replenishment is automatically triggered. That is, when the data acquisition pause time exceeds the threshold (e.g., 10 seconds), the missing construction pumping data will be automatically replenished when data acquisition resumes. Similarly, when the data acquisition interruption time exceeds the threshold, the missing data will also be automatically replenished.

[0045] "Replenishment" refers to the process of retrieving and supplementing historical data that was not collected during periods of suspension, network outage, or data loss, in order to ensure the continuity of construction data. Specifically, when the system starts, it concurrently retrieves all historical data from the platform at the second level in the background; when the data collection pause and resume exceeds a preset threshold or a data gap is detected, replenishment is automatically triggered, and the replenishment range is determined by the start and end timestamps and hash numbers of the gap.

[0046] (3) Discontinuity gap index retreat and filling If the history recovery fails, do not give up. The retry backoff interval of the history recovery increases exponentially and is capped. Retry according to the exponential backoff interval until success. The backoff interval is calculated according to the following formula and is capped at 60 seconds. (10) In the formula, For the first The waiting interval before the next retry is s; The initial interval is s; in this embodiment, d0 is 5s. All values ​​in the embodiment (such as acquisition cycle, pump stop threshold, number of anti-shake points, retreat limit, etc.) are initial engineering values ​​and are configurable. They can be adjusted according to the actual data patterns of the target well and the construction platform during deployment, and are not fixed values.

[0047] "Exponential backoff" refers to a compensation retry strategy where, after each failed retry, the next waiting interval increases exponentially (e.g., a power of 2). In this embodiment, the backoff interval is calculated according to formula (10) and capped at a preset upper limit (60s) to avoid increasing the platform burden or causing frequent request alarms due to high-frequency retries. Real-time streaming and historical backoff occur concurrently, and exponential backoff continues until success, ensuring data integrity.

[0048] (4) Whitelist filtering of operating wells Real-time operating results are filtered according to the platform's "currently operating wells" whitelist, collecting only wells that are actually under construction to ensure consistency with the objects displayed on the platform page. If the whitelist is not loaded, the filtering is downgraded to no filtering.

[0049] The "currently operating well" whitelist refers to the set of well numbers maintained by on-site personnel in the construction operation platform that are currently undergoing fracturing operations. It is obtained by calling the "currently operating well" query interface provided by the construction operation platform when the system starts, caching the returned results as a whitelist. This whitelist is used to filter real-time operational results, collecting only wells actually under construction, ensuring consistency between the collected data and the platform's display. "Degradation" means that if the whitelist has not yet been successfully loaded, the system will not perform this filtering (filtering means collecting all returned results). This "degradation" process avoids complete interruption of data collection due to whitelist acquisition failure, ensuring that the system can still operate normally and data collection is uninterrupted even when the configuration is not ready.

[0050] Step S2, Stage Alignment like Figure 2 As shown, the current pumping stage is maintained by a state machine. The measured data points of step S1 are mapped to the corresponding design stage in the pumping scheme table by means of "sliding window multi-parameter voting". The stage identifier, stage cumulative liquid volume, pump stop status and well shut-in status under a unified time reference are output, and a stage identifier under a unified time reference is established. The pumping scheme table includes the design discharge rate, design sand concentration, and design liquid volume for each stage; The stage alignment module holds the pumping scheme table (design displacement, design sand concentration, and design fluid volume for each stage), maintains the current pumping stage using a state machine, and performs the following sub-steps for each measured data point. Here, "stage alignment" refers to mapping the measured data points to the corresponding design stages (pre-fluid / sand carrying / displacement / shut-in pressure drop) in the pumping scheme table, establishing stage identifiers under a unified time reference, which is distinct from the "merging" semantic in data acquisition.

[0051] The purpose of using a state machine to maintain the current pumping stage is to preserve the historical state of the construction process and to ensure that each pumping stage proceeds in an orderly manner according to predetermined conditions. It provides a unified state description for the stage attribution of each measured data point, and the current state of the state machine is the current stage identifier.

[0052] To complete the task of "mapping measured data points to the corresponding design stages in the pump injection scheme table and establishing stage identifiers under a unified time benchmark", the following sub-steps need to be executed in sequence: sliding window multi-parameter voting, pump stop / well shut-in identification, cumulative actual fluid volume, stage transition judgment and transition de-shaking. That is, the mapping is completed by the collaboration of the aforementioned sub-steps, rather than by a single direct assignment.

[0053] (1) Multi-parameter voting via sliding window The maintenance capacity for discharge volume and sand concentration is respectively: A sliding window queue is used to enqueue points one by one and remove points at the head of the queue that exceed the window. The window mean is taken as the input for the stage decision to suppress single-point jitter. In this embodiment, the sliding window size N w Take 5, sand concentration tolerance ΔC is taken as 30 kg / m³, and discharge tolerance δ is taken as 30 kg / m³. Q Take 20%; where ΔC can be 20–50 kg / m³, δ Q It can be 10–30%, N w The sliding window size can be configured within the range of 3–10 points according to the noise level of the construction data platform. The selection of the sliding window size should balance the smoothing effect and the response speed: if the window is too large, it will introduce a large judgment delay and weaken the monitoring sensitivity to signal abrupt changes (such as sand blockage precursors); if the window is too small, the smoothing will be insufficient and it will be easily affected by single-point noise interference. Generally, 3–10 points is appropriate.

[0054] The selection and implementation of the sliding window are illustrated below with an example: The window size Nw is the number of data points retained within the window. Each new measured data point is added to the tail of the window queue, while older points exceeding the window length are removed from the head of the queue, ensuring the window always retains the most recent data points. For example, with a window size of 5, the window consists of the 5 most recent consecutive data points, and their arithmetic mean is used as the input for determining the stage of displacement and sand concentration at that moment. The purpose of window mean filtering is to eliminate the interference of single-point noise such as platform normalization error and instantaneous spikes on stage determination and rule analysis, i.e., to smooth the real-time data to suppress single-point jitter. The value of Nw is independent of the total number of data points: regardless of the cumulative number of data points collected in the well, the window always retains only the most recent Nw data points for mean calculation; the value of Nw should balance the smoothing effect and response speed, generally ranging from 3 to 10 points.

[0055] (1) In the formula, For the first in the window Measured displacement of each data point, m³ / min; The mean displacement measured within the window is m³ / min; For the first in the window Measured sand concentration at each data point, kg / m³; The average sand concentration measured within the window is expressed in kg / m³. Size of the sliding window (number of data points), number of points, alignment configuration, configurable.

[0056] (2) Pump shutdown detection and well shut-in identification Pumping is stopped when the average discharge rate of the window is lower than the pump stop discharge rate threshold. (2) In the formula, The pump stoppage discharge threshold, in m³ / min, is an alignment logic constant. A pump stoppage is determined when the average discharge rate within a window falls below this threshold. A pump stoppage is also determined when the average discharge rate within a window falls below 0.05 m³ / min. During the pump stoppage, no liquid volume is accumulated, and no stage is switched, but the current stage information continues to be output to prevent interference with stage counting during the pump stoppage / liquid replenishment process. When the pump stoppage status continues for a preset duration threshold, a well shut-in flag is set, allowing the rule analysis module to execute pressure drop monitoring rules during the well shut-in pressure drop stage. The pump stoppage discharge threshold and the well shut-in duration threshold are initial engineering values ​​and can be adjusted according to the target well's discharge capacity and on-site conditions; they are not fixed values. The values ​​given in this example are only illustrative.

[0057] Table 1. Rule Base (Example)

[0058] In the table, the rule name is the rule type, and the trigger condition is the phenomenon description. Pressure drop monitoring rules are one category in the rule base listed in the table, specifically the last row. The phase constraint in the rule trigger condition is always given in the "Gated Stage" column, and the trigger condition itself does not repeatedly include the phase flag.

[0059] Pumps will be stopped if the discharge rate is below 0.05 m³ / min. These thresholds are initial values ​​for the project and can be adjusted according to the target well discharge rate and on-site conditions. They are not fixed values.

[0060] The methods for obtaining the corresponding parameters in the table are all existing technologies. For example, the bottom hole pressure and net pressure are calculated by the data acquisition module: using the measured wellhead pressure, pump pressure, discharge rate, and sand concentration as input, the model is solved segment by segment along the wellbore from the wellhead to the bottom hole according to the wellbore pressure-temperature distribution model. That is, the hydrostatic pressure increment is superimposed segment by segment along the wellbore from the wellhead to the bottom hole and the frictional pressure drop is deducted. After iterative coupling of the temperature field and pressure field until convergence, the pressure value of the bottom hole node is taken as the bottom hole pressure; the net pressure is the difference between the bottom hole pressure and the formation closure pressure. Specifically: using the measured wellhead pressure, pump pressure, discharge rate, sand concentration, and injection temperature as input, the pressure and temperature distribution are solved segment by segment along the wellbore from the wellhead to the bottom hole. For the i-th segment of the wellbore, its pressure increment consists of the hydrostatic pressure increment and the frictional pressure drop. The wellbore pressure distribution equation is as follows: p(i+1)=p(i)+ρ(i)·g·cosθ(i)·Δz λ(i)·ρ(i)·v(i)²·Δz / (2·D(i)) In the formula, p(i) and p(i+1) are the pressures at the beginning and end of the i-th well section, respectively, in Pa; ρ(i) is the density of the fracturing fluid or proppant-carrying fluid in the i-th well section at the average temperature and average pressure of that section, in kg / m³; g is the acceleration due to gravity, taken as 9.81 m / s²; θ(i) is the inclination angle of the i-th well section, in °; Δz is the segment length, in m; λ(i) is the friction coefficient along the i-th well section, determined by the generalized Reynolds number of that section and the rheological parameters (consistency coefficient and flow index) of the fracturing fluid; v(i) is the flow velocity in the i-th well section, in m / s, v(i) = Q / A(i), where Q is the measured discharge rate, in m³ / s; A(i) is the flow cross-sectional area of ​​that section, in m²; D(i) is the hydraulic diameter of the flow channel of that section, in m; the flow cross-sectional area and hydraulic diameter are calculated based on the inner diameter of the casing or tubing according to the fracturing fluid injection method.

[0061] The temperature increment of the same i-th segment of the wellbore consists of heat generated by friction, work done by gravity, and radial heat loss. The wellbore temperature distribution equation is as follows: T(i+1)=T(i)+Δp_f(i) / (ρ(i)·cp(i))+g·Δz / cp(i) ΔT_loss(i) In the formula, T(i) and T(i+1) are the temperatures at the beginning and end of the i-th wellbore, respectively, in K; cp(i) is the average isobaric specific heat capacity of the fracturing fluid or sand-carrying fluid in this section, in J / (kg·K); Δp_f(i) is the frictional pressure drop along the section, in Pa; ΔT_loss(i) is the radial heat loss of this section to the formation, in K, which is proportional to the difference between the fluid temperature and the formation temperature in this section. The proportionality coefficient is determined by the comprehensive heat transfer coefficient, hydraulic channel size, mass flow rate and specific heat capacity of this section. The comprehensive heat transfer coefficient is determined by the flow channel size, the thermal conductivity of the tubing and casing and cement sheath, the formation thermal conductivity and the formation thermal diffusivity.

[0062] The temperature and pressure distribution in the wellbore are coupled: the density, friction coefficient, and flow velocity in the pressure equation all change with the average temperature of the section, and the friction heat generation in the temperature equation also changes with the pressure of the section. Therefore, the pressure and temperature equations are iteratively calculated for each section until convergence. The convergence criterion is that the pressure change in the section is less than 0.01 MPa and the temperature change is less than 0.1 K. The calculation is advanced segment by segment from the wellhead to the bottom of the well, and the pressure value at the end of the last segment is taken as the bottom hole pressure.

[0063] The formation closure pressure is determined by interpreting the pump shutdown pressure drop from a small-scale fracturing test, and its value is the difference between the instantaneous pump shutdown pressure and the formation pressure; the net pressure is the difference between the bottom hole pressure and the formation closure pressure. The bottom hole pressure and net pressure are updated point by point with the collected construction data and are used as the judgment quantities for the corresponding rules in step S3.

[0064] (3) Cumulative actual liquid volume When the pump is running, the actual liquid volume of the current stage is accumulated. If the time interval between adjacent data points is greater than the data interruption threshold, the data volume is not accumulated (this is considered data interruption protection). (3) In the formula, The current cumulative actual liquid volume, in m³, is the integral of displacement and time. The time interval between adjacent data points, s, is calculated from the timestamp; , t is the current data point collection time. prev That is, t-1 is the time of the previous data point collection. s is the data interruption threshold and an alignment logic constant. V acc (t-1) This represents the cumulative actual liquid volume corresponding to the previous data point.

[0065] In the formula The time interval between adjacent data points is the difference between the current data point acquisition time and the previous data point acquisition time; where t is the current data point acquisition time and tprev is the previous data point acquisition time. The data interruption threshold is the maximum allowable time interval between adjacent points. When the time interval between adjacent data points exceeds this threshold, it is considered a data interruption and the liquid volume is not accumulated. That is, if the time interval between adjacent data points is greater than 300s, the liquid volume is not accumulated to avoid the false increase of liquid volume during the network interruption and re-encoding period; the example uses 300s, and the threshold can be configured.

[0066] (4) Stage transition determination When a next stage exists, calculate the sand concentration matching indicator, discharge volume matching indicator, and liquid volume compliance indicator separately: (4) (5) (6) In the formula, M C M is a marker for sand concentration matching. Q For displacement matching; M V This is the liquid volume matching flag, and can be either 0 or 1. For the next stage, design sand concentration (kg / m³) and pumping scheme table; Sand concentration tolerance, kg / m³, aligned configuration, configurable; For the next stage, design displacement (m³ / min) and pumping scheme table; This represents the displacement tolerance percentage, %; it is configurable and aligned with the configuration. The current stage design liquid volume (m³) and pumping scheme table are provided. For liquid volume completion ratio, dimensionless, aligned configuration, configurable; The method for determining whether there is a "next stage": If the current stage is the last pumping stage of the post-processing scheme (or is in the shut-in pressure reduction stage), then there is no next stage and no transition judgment is performed; otherwise, the design values ​​of the next stage (design sand concentration, design discharge rate, design fluid volume) of the current stage are used for matching.

[0067] The sand concentration matching flag is set to 1 to indicate that the difference between the average sand concentration in the window and the design sand concentration in the next stage does not exceed the sand concentration tolerance, and 0 indicates that it exceeds the tolerance; the discharge matching flag is set to 1 to indicate that the percentage difference between the average discharge rate in the window and the design discharge rate in the next stage does not exceed the discharge rate tolerance percentage, and 0 indicates that it exceeds the tolerance; the liquid volume compliance flag is set to 1 to indicate that the cumulative actual liquid volume in the current stage reaches the product of the design liquid volume and the liquid volume completion ratio, and 0 indicates that it has not been reached. The above flags are set to 0 or 1 to represent the two mutually exclusive Boolean results of "established / not established" for each matching condition, which are used for transition decision to combine according to formula (7).

[0068] The transition decision is based on matching sand concentration and discharge rate, or meeting liquid volume requirements: (7) In the formula, This is the jump trigger flag, either 0 or 1; Setting it to 0 indicates that the transition condition is not met. A value of 1 indicates that the condition is true; A "transition" refers to the state transition action of a state machine switching from the current pumping stage to the next pumping stage. In simple terms, it means whether a pumping stage has changed. If the stage changes, it is a "transition"; if it does not change, no "transition" has occurred. The triggering condition is whether the transition condition is met: when the transition condition is met, the stage identifier changes from the current stage to its successor stage and the stage count is transferred; otherwise, the current stage is maintained.

[0069] A jump trigger flag of 1 indicates that the jump condition is met and a stage switch should occur, while a flag of 0 indicates that the condition is not met and the current stage is maintained. The jump decision is executed according to formula (7): when the sand concentration and discharge rate are matched, or when the liquid volume reaches the completion ratio, the jump trigger flag is set to 1. When the liquid volume reaches the standard, the de-vibration is skipped and the jump is performed directly to prevent abnormal parameter retention and avoid construction stagnation. The number of de-vibration points for adjacent stages where the design sand concentration and discharge rate are close is doubled to prevent false jumps in the stepped sand addition scenario.

[0070] (5) Transition de-jitter To suppress false transitions caused by stage boundary parameter jitter, transition conditions must be met continuously for a preset number of data points before a transition is submitted: (8) In the formula, The number of data points that continuously satisfy the transition condition is given by the number of consecutive data points, which is determined by continuous statistics. To debouncing the number of points, a constant and a judgment logic, the value is cleared to zero upon interruption of the condition; when... Time-based transition submission. Debounce points. Determined according to the following rules: default ; When the design sand concentration and design discharge rate of two adjacent stages do not exceed the sand concentration difference, And the displacement difference does not exceed When the design sand concentration and design discharge rate are close to each other in two adjacent stages, The default value is doubled (Example 6) to suppress erroneous transitions between adjacent stages with similar parameters in the stepped sand addition scenario.

[0071] When the liquid volume reaches the standard mark hour, Set to 1 (skip debouncing and jump directly) to prevent the current stage from being stuck for a long time when the parameters are abnormal.

[0072] The anti-shaking points and their judgment rules are all initial engineering values ​​and are configurable. They can be adjusted according to the actual working conditions of the target well during deployment and are not mandatory standard values. In the example, the default value is 3, the adjacent stages with the same parameter are 2 times, which is 6, and the liquid volume reaches the standard, which is only an example.

[0073] In this embodiment, when the liquid volume compliance flag MV=1, the value is 1, skipping the de-jittering and directly jumping over; otherwise, when the design sand concentration and design discharge rate of two adjacent stages are close to each other (the sand concentration difference does not exceed the tolerance ΔC and the discharge rate difference does not exceed the tolerance percentage δQ), the value is twice the default value; in other cases, the default value of 3 is used. Liquid volume compliance means that the actual pumped liquid volume has been fully pumped according to the designed liquid volume value for that stage in the pumping procedure.

[0074] In the formula N D To reduce jitter points, The number of data points that continuously satisfy the transition condition (Example N) D Take 3 (configurable); when Not less than N D When a transition is committed, the condition is cleared upon interruption, thereby suppressing erroneous transitions caused by fluctuations in stage boundary parameters.

[0075] When the liquid volume meets the standard, the number of de-vibration points is set to 1, meaning that the data point is directly submitted for transition and skips the continuous judgment. The purpose is that if the construction parameters are abnormal, causing the sand concentration and discharge rate to fail to enter the design range of the next stage for a long time, the cumulative liquid volume will continue to increase. At this time, using "reaching the design liquid volume completion ratio" as the basis for forced advancement can prevent the construction from being stuck in the current stage for a long time and avoid stagnation.

[0076] The number of de-jittering points refers to the "number of consecutive satisfactions" required for stage transition determination. It is used to eliminate false transitions caused by instantaneous fluctuations in boundary parameters. The number of de-jittering points and its multiple for "adjacent stages with the same parameter" (e.g., default 2 times, expandable to 3 times) can be configured differently according to stage type (e.g., stepped sand-adding stage, continuous sand-carrying stage) to further match the severity of boundary parameter fluctuations in different stages and suppress false transitions in scenarios such as stepped sand-adding. Specifically: ① For continuous sand-carrying stages with stable boundary parameters, the number of de-jittering points can be set to 1 times for "adjacent stages with the same parameter" and the number of de-jittering points can be relatively small (e.g., transition is achieved after 3 consecutive satisfactions); ② For stepped sand-adding stages with progressively changing boundary fluctuations and severe boundary fluctuations, the number of de-jittering points can be increased to 2-3 times and the number of de-jittering points can be increased (e.g., transition is achieved only after 6-9 consecutive satisfactions) to suppress false transitions in stepped sand-adding scenarios; ③ When the liquid volume reaches the standard and forced advancement is initiated, the number of de-jittering points is set to 1, and the transition is directly submitted as a fallback to prevent construction from being stuck in the current stage for a long time. The above multiples and points can be adjusted according to stage type in the configuration.

[0077] (6) Timeout anomaly detection If the actual duration of the current stage exceeds 1.5 times the designed duration and the cumulative liquid volume is less than 50% of the designed liquid volume, it is judged as a timeout anomaly, and an inquiry is sent to the operator in real-time mode: (9) In the formula, The actual duration of the stage is in seconds. The design duration, in seconds, is calculated from the time and scheme table; (s), The design displacement for the current stage is m³ / min; if the above formula is satisfied, it is judged as a timeout anomaly, and in real-time mode, the operator is asked whether to force entry into the next stage.

[0078] (7) Unification of manual intervention and playback In real-time mode, when the algorithm cannot determine the stage or determines that there is a timeout, it sends a query to the operator and receives a forced stage switching instruction. After the forced switch, the current stage's cumulative liquid volume is reset and the manual intervention record is recorded. In post-event playback mode, pre-recorded human intervention records are automatically applied, and when all sand-carrying stages are completed, the current sand concentration is below the threshold and it is not the last stage, the remaining stages are automatically identified as replacement stages. Both real-time and playback modes share the same core alignment algorithm. The phase alignment module outputs phase identifiers, cumulative liquid volume for each phase, pump stop status, and well shut-in status under a unified time reference, ensuring that subsequent rule analysis is based on a consistent time and phase reference.

[0079] "The same alignment algorithm is used for both real-time and playback modes" means that the alignment algorithm can be used for both real-time analysis at the fracturing site (real-time mode) and post-fracturing data playback (post-playback mode). Both real-time and playback modes perform the same alignment processes on the measured data, including sliding window voting, pump stop / well shut-in identification, fluid volume accumulation, transition judgment, and debouncing. The only difference is the external input for stage decisions—in real-time mode, a forced switching command is issued manually in real time, while in playback mode, pre-recorded manual intervention records are automatically applied. This setup ensures that real-time monitoring and post-playback provide completely consistent stage divisions for the same construction process, avoiding discrepancies between playback conclusions and actual field conditions due to different judgment logics.

[0080] The final output of the stage alignment module is the stage identifier, stage cumulative fluid volume, pump stop status, and well shut-in status under a unified time benchmark. These outputs serve as inputs for rule analysis (step S3), ensuring that subsequent rule analysis is based on a consistent time and stage benchmark.

[0081] S3, Rule Analysis like Figure 3 As shown, input the stage identifier, cumulative liquid volume of the stage, pump stop status and well shut-in status of step S2, load the rule base, perform stage-gated filtering on the rules according to the stage identifier of step S2, and then compare the measured values ​​of the condition fields with the thresholds one by one to obtain the judgment results, and output the phenomenon description, current measured value, treatment suggestions and priority. Each rule in the rule base has a stage gating attribute, priority attribute, phenomenon description, treatment suggestion and rule type; in addition, a single rule base can be loaded in groups by block, layer or process type, so that the stage gating attribute of the rule switches with the process type, so as to adapt to the scenario where the same system serves multiple blocks and multiple fracturing processes.

[0082] The rule base shown above is a general base and is not grouped and loaded according to blocks, layers, or process types. However, in practice, some rules in the rule base can be grouped and loaded according to process types, and the stage gating attribute of the rules can be switched according to the process type, as shown in the table below. In the table, "Temporary Plugging Section" is an extended stage of the basic stage enumeration (full stage, pre-fluid, sand carrying, displacement, shut-in pressure drop): when the process type is temporary plugging-to-fracturing, the temporary plugging section is included in the stage gating enumeration as an independent extended stage, and the corresponding pressure relief / stabilization rules in the rule base use "Temporary Plugging Section" as their gating stage; the gating stage of other rules is still taken from the basic stage enumeration.

[0083] Table 2. Schematic diagram of rule base grouping and loading by process type and stage gating switching.

[0084] The "Phenomenon Description" is a textual description of the construction phenomenon that occurs when the rule is triggered (e.g., "abnormal rise in bottom hole pressure during the sand-carrying stage"). The "Trigger Condition" in the rule table is a comparison between the measured value and the threshold, and the "Continuation Condition" is the duration or number of points required to maintain the hit. The phenomenon description is used to show the operator "what happened," the trigger / continuation condition is the basis for determining that the phenomenon occurred, i.e., "why it happened," and the handling suggestions are used to guide "how to handle the phenomenon."

[0085] The input to step S3 is the stage identifier, cumulative liquid volume of the stage, pump stop / well shut-in status, and unified benchmark real-time data output from step S2. The output of S3 is the judgment result of "phenomenon description + current measured value + treatment suggestion + priority". This result is the basis for step S4 "closed-loop treatment and manual intervention". That is, after step S3 performs stage-gated filtering on the rules based on the stage identifier of step S2, it compares the measured values ​​of the condition fields one by one with the threshold to obtain the judgment result.

[0086] The rule analysis module loads the rule base. Each rule includes a stage gating attribute, priority attribute, phenomenon description, handling suggestion, and rule type. Rule analysis is performed according to the following mechanism: (1) Stage Gating (Authoritative Gating): The "Applicable Stage" attribute of the rule is used as the authoritative gating field. Before analysis, the rule is first filtered by stage gating according to the current stage identifier. The phase flag is not retained in the rule sub-conditions (the phase flag has been normalized to the "Applicable Stage" attribute) to avoid duplicate or contradictory gating. The applicable stage values ​​include all stages, pre-fluid, sand carrying, displacement, and shut-in pressure drop; the phase flag is uniformly normalized to the "Applicable Stage" authoritative field to eliminate duplicate / contradictory gating and prevent false alarms.

[0087] In this invention, "phase" refers to the stage of fracturing operation (i.e., pre-flush fluid, proppant carrying, displacement, shut-in pressure drop, etc.). "Phase marker" is the condition / marker used to determine "which phase we are currently in." It can be understood as a "stage identifier."

[0088] (2) Single condition and compound condition: The rule condition can be a single condition or a compound condition (a compound condition is when multiple sub-conditions are satisfied at the same time). The measured value of each condition field is calculated and compared with the threshold. This method determines whether there are any anomalies in the current fracturing operation based on "rule conditions" set in the rule base. Anomalies refer to whether a physical quantity (such as bottomhole pressure, wellhead pressure, displacement, or sand concentration) is in an abnormal state relative to its design value or threshold, thus deciding whether to trigger corresponding alarms and handling suggestions. For example, a single condition, Pwh>50MPa, means that the wellhead pressure exceeds the design value or threshold. Once this condition is met, subsequent alarms and handling suggestions will be triggered. During operation, the relationship between the measured values ​​of candidate physical quantities and their preset thresholds / design values ​​is determined, and the measured values ​​of each condition field are compared with the thresholds to determine whether a match has occurred.

[0089] (3) Priority classification: In this embodiment, the priority is divided into four levels: P0, P1, P2 and P3, and they are executed in order from high to low priority. When any P0 level rule is hit, only the P0 level judgment result is presented to suppress low priority noise, but other concurrent P0 level results are not discarded. That is, when any P0 level rule is hit, only the P0 level judgment result is presented. When multiple P0 level rules are hit at the same time, the handling suggestion is to take the highest priority instruction according to the priority of emergency pump stop, replacement of clear water sluice block, parameter adjustment and alarm prompt. In this embodiment, priorities are divided into four levels: P0, P1, P2, and P3, based on the urgency of the situation. P0 is the highest level (e.g., emergency pump stop rules), and P3 is the lowest. The number of priorities is not mandatory; it's a configurable engineering feature and doesn't require four levels. P0 is the highest-level rule. Triggering a P0 rule indicates a very urgent anomaly requiring pump stoppage. Therefore, this urgent anomaly should be presented to the user, while other lower-level anomalies should be temporarily stopped (discarded) to avoid affecting the display of P0 results. However, if other P0 results exist, they should not be discarded. If multiple P0 rules are triggered simultaneously, the recommended handling order is "emergency pump stoppage > replacement of clear water slug > parameter adjustment > alarm notification," prioritizing the highest urgency level instruction. Simultaneously, the complete descriptions of all P0 rules and their current measured values ​​should be presented to avoid conflicting instructions for the operator. The appearance of "P0" in the text indicates that the rule belongs to the highest urgency level. Only P0 rules are displayed when any P0 rule is triggered, suppressing lower-priority noise and highlighting the emergency. The alarm will only appear after a preset number of consecutive hits with an interval of no less than a preset duration, thus suppressing alarm flashing.

[0090] In addition, the priority grading can be expanded to more levels, or "P0 suppression" can be changed to be suppressed according to the type of handling suggestion (such as emergency pump stop rules) instead of priority level, in order to adapt to different preferences of different process sites for the way alarms are presented in order of priority.

[0091] (4) Decoupling of phenomenon and treatment: When a rule is hit, the phenomenon description, the current measured value of the corresponding condition field when it is triggered, the treatment suggestion and priority are recorded at the same time, so that they can be displayed in the panel and report simultaneously; (5) Deduplication: When the same rule is hit multiple times in the data stream, only the first hit result is retained to avoid duplicate alarms.

[0092] The default rule base covers rules for bottom hole pressure exceeding limits, bottom hole pressure too low, bottom hole pressure sudden drop, abnormal rise in bottom hole pressure during the sand-carrying stage (a precursor to sand blockage), insufficient net pressure, wellhead pressure exceeding limits, abnormal rise in wellhead pressure during the sand-carrying stage, pressure rise during the high sand ratio stage, low discharge rate during the sand-carrying stage, sand blockage determination (sharp pressure rise), pressure rise during the displacement stage (sand bridge), pressure buildup (high pressure and low discharge rate), excessively high net pressure, sand concentration exceeding the design limit, sand concentration below the design limit, discharge rate exceeding the design limit, and monitoring of shut-in pressure drop rate.

[0093] As shown in Table 3, the complete triggering conditions (threshold, duration / acknowledgment duration, and gating stage) for three of the P0 core rules are given below. The values ​​of the triggering conditions are all initial engineering values ​​and are configurable. They will be calibrated based on the wellbore properties and pumping design scheme of the target well. Table 3. P0 Rule Base Illustration

[0094] The above rules, like the others, employ the same stage gating, priority levels P0–P3, single / compound conditions, phenomenon and treatment decoupling, and deduplication mechanisms.

[0095] Each rule in the default rule base has a clear judgment criterion, namely "trigger condition (comparison of measured value and threshold)," and its judgment criterion is the same as the three P0 example rules mentioned above. After a rule is hit, a judgment result is generated consisting of "phenomenon description + current measured value + handling suggestion + priority," which is the "rule judgment result." Taking "bottom hole pressure exceeding limit" as an example: when the measured effective bottom hole pressure is greater than 120MPa and lasts for no less than 2 seconds within a continuous window, the P0 rule is hit, and the handling suggestion of "emergency pump stop, output shutdown command" is output and presented immediately.

[0096] The "rules mentioned above" are just three example rules at the P0 level. The remaining rules are other rules besides the three P0 level rules mentioned above. The priority level is divided into four levels, P0-P3, as mentioned above.

[0097] S4, Closed-loop processing like Figure 4 As shown, the early warning and response module generates tiered alarms based on the urgency of the response recommendations according to the output of the rule analysis module, and feeds the response recommendations back to the data acquisition module, the stage alignment module, and the construction control terminal, forming a closed-loop monitoring and control link.

[0098] The handling recommendations are categorized by urgency into emergency pump shutdown, parameter adjustment, replacement of clear water sluice blockage, alarm prompts, and monitoring prompts. Corresponding construction control instructions are generated based on the type of handling recommendation. Alarm presentation uses a confirmation counting mechanism; the same rule must be hit a preset number of times consecutively before an alarm is triggered, and the interval between two adjacent alarms for the same rule is not less than a preset duration to suppress alarm flashing caused by momentary fluctuations.

[0099] In this embodiment, the presentation of graded alarms adopts the "graded confirmation count" standard: P0 level rules are presented immediately upon a single hit, so as not to delay emergency situations due to confirmation delays; P1-P3 level rules need to be hit consecutively a preset number of times (3 times in the example, configurable) before being presented; the interval between two adjacent alarms of the same rule is not less than a preset duration (2 seconds in the example, configurable); the same alarm will not be presented again during the period before it is cleared (the alarm deduplication is the responsibility of the early warning and handling module).

[0100] The purpose of feeding back the handling suggestions to the stage alignment module and the data acquisition module is mainly to form a closed loop: the feedback to the stage alignment module is used to trigger handling involving stage changes (such as forced switching or replacing the clear water sluice gate), and the feedback to the data acquisition module is used to trigger an increase in the acquisition frequency in a high-risk continuous state (such as increasing it from 1s to 0.5s), thereby connecting monitoring, analysis and control.

[0101] In this embodiment, the real-time mode records the intervention reason label, which is used to generate a structured intervention review report during post-event playback / debriefing for construction summary and process optimization.

[0102] This embodiment provides a storage medium on which a program is stored. When the program is executed by a processor, it implements the steps of the real-time monitoring and intelligent control method for fracturing operations in this embodiment.

[0103] This invention addresses the problem of inconsistent time / stage benchmarks in multi-source data by using stage alignment as a unified time benchmark and employing a "state machine + sliding window multi-parameter voting" approach to eliminate misjudgments caused by multi-source data mismatch.

[0104] This invention addresses the problem of false alarms caused by stage boundary jitter by employing "transition de-jittering + double de-jittering of adjacent stages with the same parameters + liquid volume-forced jitter-free operation" to suppress false boundary transitions and reduce alarm fatigue.

[0105] This invention addresses the problem of lacking stage semantics in rule-based judgment. It adopts "stage-gated filtering + P0–P3 priority grading and P0 suppression" to achieve differentiated judgment of the same phenomenon in different stages and reduce false alarms across stages.

[0106] This invention addresses the problem of disconnect between monitoring and response by employing a closed-loop approach: "response suggestions are fed back to each module and the construction control end, with manual intervention connecting real-time and playback," thereby improving response timeliness.

[0107] This invention addresses the issues of heavy data collection and deployment, breakpoint gaps, and storage bloat by employing a method of "pure HTTP data collection + concurrent breakpoint recovery; partitioning + unique index self-healing deduplication," resulting in lightweight and stable operation with complete and traceable data.

Claims

1. A real-time monitoring and intelligent control system for fracturing operations, characterized in that, It includes a data acquisition module, a stage alignment module, a rule analysis module, an early warning and response module, and a data storage module; The data acquisition module is used to periodically... Acquire multi-source construction data and associate and merge them with construction wells and timestamps to form a real-time data stream. The multi-source construction data includes at least wellhead pressure, pump pressure, discharge rate, sand concentration and corresponding construction timestamps. The data storage module is used to partition data by "hash number + construction date", perform batch transaction writing and self-healing deduplication when building a unique index, and provide a traceable data access interface for each module. The stage alignment module is used to identify stages based on the pumping scheme table using "state machine + sliding window multi-parameter voting", introduces transition debouncing, and outputs stage identifier, cumulative liquid volume, and pump stop / well shut-in status under a unified time reference. The pumping scheme table includes the design discharge rate, design sand concentration, and design liquid volume for each stage; The rule analysis module is used to perform stage-based gating and priority-based analysis based on a preset rule base, and output the phenomenon, current value, handling suggestions, and priority attributes. Each rule in the preset rule base has a stage gating attribute, priority attribute, phenomenon description, handling suggestion and rule type; The early warning and response module is used to generate graded alarms based on the urgency of the response recommendations and to feed back the response recommendations to the data acquisition module, the stage alignment module, and the construction control terminal.

2. The real-time monitoring and intelligent control system for fracturing operations according to claim 1, characterized in that, It also includes a manual intervention module, which is used to manually align the stages when the stage alignment module cannot align them accurately, so as to output stage identifiers, cumulative fluid volume, and pump stop / well shut-in status under a unified time reference, and to reuse intervention records for forced switching in real-time mode and post-replay mode.

3. A method for real-time monitoring and intelligent control of fracturing operations, employing the real-time monitoring and intelligent control system for fracturing operations as described in claim 1 or claim 2, characterized in that, Includes the following steps: S1. Multi-source data acquisition and merging The system acquires multi-source construction data from the construction data platform in real time using a periodic Tpoll, and then associates and merges the data according to the construction well and timestamp to form a real-time data stream with a unified time sequence. The multi-source construction data includes at least wellhead pressure, pump pressure, discharge rate, sand concentration and corresponding construction timestamp. S2, Phase Alignment The current pumping stage is maintained by a state machine. The measured data points of step S1 are mapped to the corresponding design stage in the pumping scheme table by means of "sliding window multi-parameter voting". Transition debouncing is introduced, a stage identifier under a unified time reference is established, and the stage identifier, stage cumulative liquid volume, pump stop status and well shut-in status under a unified time reference are output. The pumping scheme table includes the design discharge rate, design sand concentration, and design liquid volume for each stage; S3, Rule Analysis Input the stage identifier, cumulative liquid volume, pump stop status and well shut-in status of step S2, load the preset rule base, perform stage-gated filtering on the rules according to the stage identifier of step S2, and then compare the measured values ​​of the condition fields with the thresholds one by one to obtain the judgment results. Output the phenomenon description, current measured value, treatment suggestions and priority. Each rule in the preset rule base has a stage gating attribute, priority attribute, phenomenon description, handling suggestion and rule type; S4, Closed-loop processing Based on the output of step S3, a graded alarm is generated according to the urgency of the handling recommendations, and the handling recommendations are fed back to the data acquisition module, the stage alignment module, and the construction control terminal to form a closed-loop monitoring and control link.

4. The method for real-time monitoring and intelligent control of fracturing operations according to claim 3, characterized in that, In step S1, the data acquisition employs the following lightweight mechanism: (1) Pure HTTP authentication and data collection The system completes the identity authentication of the intranet industrial platform using an HTTP request and directly calls the platform's real-time operating status interface to poll for data, without starting any browser process throughout the process. (2) Concurrency of real-time streaming and historical data backup Upon startup, the background concurrently retrieves all historical second-level data from the platform, without blocking real-time polling; When the data acquisition pauses and resumes beyond the threshold or the gap reopens, automatic re-filling is triggered. (3) Gap index retreat and filling If the historical recovery fails, do not give up. Retry according to the exponential backoff interval until successful. The backoff interval is calculated according to the following formula and is capped at 60 seconds. （10） In the formula, For the first The waiting interval before the next retry is s; Let s be the initial interval; (4) Whitelist filtering of operating wells Real-time operating results are filtered according to the platform's "currently operating wells" whitelist, collecting only wells that are actually under construction to ensure consistency with the objects displayed on the platform page. If the whitelist is not loaded, the filtering is downgraded to no filtering.

5. The method for real-time monitoring and intelligent control of fracturing operations according to claim 3, characterized in that, In step S2, the following sub-steps are performed for each measured data point: (1) Multi-parameter voting via sliding window The maintenance capacity for discharge volume and sand concentration is as follows: A sliding window queue is used to enqueue points one by one and remove points at the head of the queue that exceed the window. The window mean is taken as the input for the stage decision to suppress single-point jitter. （1） In the formula, For the first in the window Measured displacement of each data point, m³ / min; The mean displacement measured within the window is m³ / min; For the first in the window Measured sand concentration at each data point, kg / m³; The average sand concentration measured within the window is expressed in kg / m³. The number of sliding windows is [number]. (2) Pump shutdown detection and well shut-in identification Pumping is stopped when the average discharge rate of the window is lower than the pump stop discharge rate threshold. （2） In the formula, The pump stoppage discharge threshold is set in m³ / min. If the average discharge rate of the window is lower than this threshold, the pump is stopped. During the pump stoppage, the liquid volume is not accumulated and the stage is not switched, but the current stage information continues to be output. When the pump stoppage status continues to reach the preset duration threshold, the well shut-in flag is set, and pressure drop monitoring rules are executed during the well shut-in pressure drop stage. (3) Cumulative actual liquid volume When the pump is running, the actual liquid volume of the current stage is accumulated, but accumulation is not performed if the time interval between adjacent data points is greater than the data interruption threshold. （3） In the formula, The current cumulative actual liquid volume, in m³; The time interval between adjacent data points is s; t is the current data point collection time. prev That is, t-1 is the time of the previous data point collection. Let s be the data interruption threshold. V acc (t-1) This represents the cumulative actual liquid volume corresponding to the previous data point; (4) Stage transition determination When a next stage exists, calculate the sand concentration matching indicator, discharge volume matching indicator, and liquid volume compliance indicator separately: （4） （5） （6） In the formula, M C M is a marker for sand concentration matching. Q For displacement matching; M V This is the liquid volume matching flag, and can be either 0 or 1. The sand concentration for the next stage is specified in kg / m³. The sand concentration tolerance is expressed in kg / m³. The displacement for the next stage is designed in m³ / min; The displacement tolerance percentage is %; The designed liquid volume for the current stage is in m³. The volume ratio is dimensionless. The transition decision is based on matching sand concentration and discharge rate, or meeting liquid volume requirements: （7） In the formula, This is the jump trigger flag, either 0 or 1; Setting it to 0 indicates that the transition condition is not met. A value of 1 indicates that the condition is true; (5) Transition de-jitter To suppress false transitions caused by stage boundary parameter jitter, transition conditions must be met continuously for a preset number of data points before a transition is submitted: （8） In the formula, The number of data points that continuously satisfy the transition condition; To debouncing the points, the count is zeroed out if the condition is interrupted; when... Submit the jump at the appropriate time; (6) Timeout anomaly detection If the actual duration of the current stage exceeds 1.5 times the designed duration and the cumulative liquid volume is less than 50% of the designed liquid volume, it is judged as a timeout anomaly, and an inquiry is sent to the operator in real-time mode: （9） In the formula, The actual duration of the stage is in seconds. The design duration is in seconds (s). (s), The design displacement for the current stage is m³ / min; if the above formula is satisfied, it is judged as a timeout abnormality, and in real-time mode, the operator is asked whether to force entry into the next stage. (7) Unification of manual intervention and playback In real-time mode, when the algorithm cannot determine the stage or determines that there is a timeout, it sends a query to the operator and receives a forced stage switching instruction. After the forced switch, the current stage's cumulative liquid volume is reset and the manual intervention record is recorded. In post-event playback mode, pre-recorded human intervention records are automatically applied, and when all sand-carrying stages are completed, the current sand concentration is below the threshold and it is not the last stage, the remaining stages are automatically identified as replacement stages. The real-time and playback modes share the same core alignment algorithm.

6. The method for real-time monitoring and intelligent control of fracturing operations according to claim 5, characterized in that, In step S2, the number of transition debouncing points is determined according to the following priority: Debounce points Determined according to the following rules: Set a default value; when the difference between the design sand concentration and the design discharge rate in two adjacent stages does not exceed... And the displacement difference does not exceed hour, The default value is set to a preset multiple; when the liquid volume reaches the standard mark. hour, Take 1; When the liquid volume meets the standard, the preset minimum number of de-vibration points is used to skip de-vibration and directly jump to the next step; when the difference in design sand concentration between two adjacent stages does not exceed the preset sand concentration tolerance and the difference in design discharge volume does not exceed the preset discharge volume tolerance percentage, the preset multiple of the preset default number of de-vibration points is used; in other cases, the preset default number of de-vibration points is used.

7. The method for real-time monitoring and intelligent control of fracturing operations according to claim 3, characterized in that, In step S3, rule analysis is performed according to the following mechanism: (1) Stage gating: The "applicable stage" attribute of the rule is used as the authoritative gating field. Before the analysis, the rule is first filtered by stage gating according to the current stage identifier. The phase flag is not retained in the rule sub-conditions to avoid duplicate or contradictory gating. The applicable stage values ​​include all stages, pre-fluid, sand carrying, displacement, and shut-in pressure drop. (2) Single condition and compound condition: The rule condition is a single condition or a compound condition. A compound condition is when multiple sub-conditions are satisfied at the same time. The measured value of each condition field is calculated and compared with the threshold. (3) Priority classification: The priority is divided into multiple levels, such as P0, P1...PN, and is executed in order of priority from high to low. When any P0 level rule is hit, only the P0 level judgment result is presented to suppress low priority noise, but other concurrent P0 level results are not discarded. (4) Decoupling of phenomenon and treatment: When a rule is hit, the phenomenon description, the current measured value of the corresponding condition field when it is triggered, the treatment suggestion and priority are recorded at the same time, so that they can be displayed in the panel and report simultaneously; (5) Deduplication: When the same rule is hit multiple times in the data stream, only the first hit result is retained to avoid duplicate alarms.

8. The method for real-time monitoring and intelligent control of fracturing operations according to claim 7, characterized in that, In step S3, the rule base includes rules for bottom hole pressure exceeding limits, bottom hole pressure too low, bottom hole pressure sudden drop, abnormal rise in bottom hole pressure during the sand-carrying stage, insufficient net pressure, wellhead pressure exceeding limits, abnormal rise in wellhead pressure during the sand-carrying stage, pressure rise during the high sand ratio stage, low discharge rate during the sand-carrying stage, sand blockage judgment, pressure rise during the displacement stage, pressure buildup, excessively high net pressure, sand concentration exceeding the design limit, sand concentration lower than the design limit, discharge rate exceeding the design limit, and shut-in pressure drop rate monitoring rules.

9. The method for real-time monitoring and intelligent control of fracturing operations according to claim 3, characterized in that: In step S4, the highest level rule is displayed upon a single hit, while other level rules are displayed only after a preset number of consecutive hits. The interval between two adjacent alarms for the same rule is not less than a preset duration, and the same alarm will not be displayed again until it is cleared. When any highest-level rule is hit, only the highest-level judgment result is presented. When multiple highest-level rules are hit at the same time, the handling suggestion is to take the highest-level instruction according to the priority of urgency. In step S4, the proposed solutions are categorized by urgency into emergency pump shutdown, parameter adjustment, replacement of clear water sluice blockage, alarm prompts, and monitoring prompts, and are distributed to the construction control terminal, the stage alignment module, and the data acquisition module.

10. A storage medium, characterized in that: It stores a program that, when executed by a processor, implements the steps of the real-time monitoring and intelligent control method for fracturing operations as described in any one of claims 3-9.

Images