A Smart Control System and Method for Downhole Operation Safety

By using a comprehensive downhole sensing unit to monitor downhole liquid levels, gas content, and well cover structural status in real time, combined with identity verification and time period verification, the shortcomings of unique identity binding and real-time detection in the downhole operation safety management system are solved, realizing multi-constraint proactive access and safety early warning for downhole operations.

CN122656834BActive Publication Date: 2026-11-17XIAMEN SMART XIANGAN DATA CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202611160898.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-08-03
Publication Date
2026-11-17
Estimated Expiration
2046-08-03

AI Technical Summary

Technical Problem

Existing underground operation safety management systems rely on manual approval, lack unique identity binding and mandatory detection, and cannot monitor the underground environment and manhole cover structure in real time, making it difficult to prevent safety hazards.

Method used

The system employs a comprehensive downhole sensing unit to monitor downhole liquid levels, gas content, and manhole cover structural status in real time. Through identity verification, time period verification, and multiple detection mechanisms, it ensures that identity, time period, environmental, and structural safety conditions are met before unlocking, thus achieving proactive access with multiple constraints.

Benefits of technology

It achieves multi-constraint proactive access for downhole operations, avoids unauthorized downhole entry and risky downhole entry in dangerous environments, fills the blind spot in structural health status detection before manhole cover opening, reduces the false judgment rate, adapts to complex working conditions and provides advanced early warning.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122656834B_ABST
    Figure CN122656834B_ABST
Patent Text Reader

Abstract

This invention discloses an intelligent control system and method for downhole operation safety. The method includes: responding to an opening request received by the wellhead locking device, acquiring the requester's identity credentials, current time domain information, and comprehensive downhole safety parameters; the comprehensive downhole safety parameters include at least the liquid level, hazardous gas concentration, and the integrity value of the wellhead cap structure; determining whether the current request simultaneously meets five conditions: identity has been approved, the time is within the permitted period, the liquid level is below the safety limit, the hazardous gas concentration is below the safety limit, and the integrity value of the wellhead cap structure is valid; controlling the locking device to release the lock only when all five conditions are met, otherwise maintaining the lock. This invention sets identity verification, time period control, and mandatory detection of the pre-entry environment and the state of the wellhead cap structure as preconditions for unlocking, realizing proactive access control for confined space operations and eliminating unauthorized entry into the well and risky entry into the well in dangerous environments from the source.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of safety technology for confined space operations, specifically to an intelligent control system and method for safety management in downhole operations. Background Technology

[0002] Underground operations (such as those in urban utility tunnels, coal mine roadways, and power tunnels—work in confined spaces) are an indispensable part of daily operations in municipal maintenance, energy transmission, and communication support. As the only physical passage between the surface and the underground space, the opening and closing of manhole covers directly affects the safety of workers. For a long time, safety management in underground operations has relied primarily on mechanical locks combined with manual approval procedures; however, this traditional model is no longer adequate for current safety management needs.

[0003] The core flaws of this model are: firstly, the lack of unique identification of personnel means that unauthorized personnel can use the universal key, making it difficult to trace unauthorized entry into the well; secondly, the safety confirmation process before entering the well relies on manual execution and lacks a mandatory detection mechanism linked to the lock. When the liquid level in the well rises suddenly due to heavy rain, or when toxic gases accumulate in a sealed environment for a long time, or when the well cover is structurally damaged by vehicles, workers may open the well cover without knowing the situation, which could easily lead to serious accidents such as poisoning, suffocation, drowning, or collapse.

[0004] In recent years, although there have been solutions for monitoring the condition of manhole covers, these solutions mainly focus on alarms after the manhole cover is illegally opened or moved. This is a "passive response" mode and fails to link the active detection of the structural health of the manhole cover, the real-time monitoring of underground environmental parameters, and the unlocking control of the manhole cover locking device. Therefore, it is impossible to complete a comprehensive safety access determination before unlocking. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent control system and method for downhole operation safety, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A method for intelligent control of downhole operation safety includes: Qualification verification steps: In response to the opening request received by the wellhead interlocking device, obtain the identity credentials of the request initiator, the current time domain information, and the downhole multi-dimensional status parameters; the downhole multi-dimensional status parameters include at least the downhole liquid level, the downhole harmful gas content, and the integrity characterization value of the well cover / seal structure. Comprehensive license determination steps: Determine whether the current request for activation simultaneously meets the following license conditions: (i) The identity credentials belong to a pre-approved list of operational personnel; (ii) The current time domain information falls within the pre-defined allowed operation time window; (iii) The downhole liquid level is lower than the preset liquid level safety boundary value, and the downhole harmful gas content is lower than the preset gas safety boundary value, and the integrity characterization value of the cap structure is valid; Execution response steps: If all the above conditions are met, a release command is issued to the wellhead locking device; if any condition is not met, the locking state is maintained and a corresponding blocking prompt message is output. The maintenance of the locking state only applies to wellhead locking devices that are currently in the locking state and does not affect wellhead locking devices that are already in the releasing state from continuing to maintain the releasing state.

[0007] Furthermore, the integrity characterization value of the sealing structure is obtained in the following manner: Spatial distribution parameters of the geomagnetic field are collected at multiple locations around the wellhead. The instantaneous geomagnetic field components in each direction at each location are compared with the pre-stored reference geomagnetic field distribution to obtain the magnetic field distortion measurement at each location. When the magnetic field distortion measurement at at least two locations exceeds the predetermined distortion boundary value at the same time, it is determined that the wellhead cover has undergone unauthorized displacement, and the integrity characterization value of the cover structure is marked as failure. Furthermore, the spatial attitude angle parameter of the wellhead cover, the ambient light transmittance parameter of the underside of the wellhead cover, and the vibration response parameters of multiple different parts of the wellhead cover are picked up; the statistical feature values ​​of each parameter are extracted with a preset time window, and the statistical feature values ​​are combined into a feature vector. Based on the feature vector, pattern classification is performed. When the classification result is a significant pose deviation or physical property damage, the integrity characterization value of the cover structure is marked as failure. Furthermore, a response value sequence is constructed using the energy fluctuation of each vibration response parameter within a continuous time window. The statistical dispersion and statistical central value of the response value sequence are calculated and compared with the pre-established benchmark dispersion and benchmark central value. When the preset physical property damage judgment criterion or foundation settlement judgment criterion is met, the integrity characterization value of the cover structure is marked as failure.

[0008] Furthermore, the method of picking up the spatial distribution parameters of the geomagnetic field at multiple locations around the wellhead specifically involves: picking up the geomagnetic field components at three locations uniformly arranged at 120° angles to each other on the same horizontal circumference of the inner wall of the wellhead; comparing the instantaneous geomagnetic field components in each direction at each location with the pre-stored reference geomagnetic field components, and calculating the normalized deviation in each direction and obtaining the comprehensive distortion metric; when the comprehensive distortion metric at at least two of the three locations simultaneously exceeds the predetermined distortion boundary value, it is determined that the wellhead cap has undergone unauthorized displacement.

[0009] Furthermore, the process of acquiring the spatial attitude angle parameter of the wellhead cover, the ambient light transmittance parameter on the underside of the wellhead cover, and the vibration response parameters of multiple different parts of the wellhead cover specifically involves: continuously acquiring the above parameters at a predetermined sampling frequency; constructing an analysis window with a fixed number of continuous sampling points; calculating the arithmetic mean of the spatial attitude angle parameter, the median of the ambient light transmittance parameter, and the peak value and root mean square value of each vibration response parameter within each analysis window; combining the above statistical feature values ​​into a feature vector; inputting the feature vector into a trained statistical classification model; the output categories of the statistical classification model include normal state, slight pose deviation state, significant deviation state, and physical property damage state; when the output is a significant deviation state or a physical property damage state, the integrity characterization value of the cover structure is marked as failure.

[0010] Furthermore, the construction of the response value sequence based on the energy fluctuation of each vibration response parameter within a continuous time window specifically includes the following steps: The triaxial acceleration response signals at multiple radial locations of the wellhead cap are collected, and an analysis time window is formed by a predetermined number of continuous sampling points. The vibration energy value at each location within each time window is calculated. Take a predetermined number of consecutive analysis time windows, calculate the vibration energy difference between adjacent time windows, and then calculate the total vibration energy fluctuation of each part within the consecutive time windows; The total vibration energy fluctuations at all locations are calculated to obtain the comprehensive response value of the current analysis unit. The comprehensive response values ​​of a predetermined number of analysis units are continuously picked up to form a response value sequence, and the statistical center value and statistical dispersion of the response value sequence are calculated. The statistical central value and statistical dispersion are compared with the pre-established benchmark central value and benchmark dispersion, respectively: when the statistical central value is greater than the benchmark central value and the ratio of the statistical dispersion to the benchmark dispersion reaches a first predetermined boundary value, it is determined to be physical property damage; when the ratio of the statistical central value to the benchmark central value reaches a second predetermined boundary value and the absolute value of the deviation of the ratio of the statistical dispersion to the benchmark dispersion from 1 is less than a third predetermined boundary value, it is determined to be foundation settlement; when any of the above determinations are met, the integrity characterization value of the cap structure is marked as failure.

[0011] Furthermore, the downhole multi-dimensional state parameters also include oxygen content and combustible gas content; the permissible condition (iii) also includes oxygen content being within a predetermined safe range and combustible gas content being lower than a preset combustible gas safety boundary value. When any of the parameters, such as the downhole fluid level, harmful gas content, oxygen content, or combustible gas content, reaches a predetermined percentage of its corresponding safety boundary value, an early warning signal is sent to the remote monitoring platform; when any parameter reaches or exceeds its corresponding safety boundary value, or when the integrity characterization value of the cap structure fails, a prohibition release command is issued to the wellhead locking device and the warning device is activated to issue a warning.

[0012] Furthermore, the liquid level safety boundary value and the gas safety boundary value are set by the remote monitoring platform according to the category of the wellhead cover and sent to the edge adjudication unit for storage; the category of the wellhead cover includes at least drainage cover, power cover and gas cover, and the liquid level safety boundary value and gas safety boundary value corresponding to different categories are configured independently.

[0013] Furthermore, the method also includes a post-release operation process recording and tracking step: The process of workers entering the well is recorded by the image acquisition unit. The video image frames are used as input, and the feature extraction network identifies the wearing status of the workers' head protective gear. The temporal correlation tracking model continuously records the trajectory of the workers. When the operation duration exceeds the allowed operation window, a timeout warning is sent to the remote monitoring platform; if the wellhead locking device is currently in the locked state, it is prohibited from performing a new release action; if the wellhead locking device is currently in the released state, it remains in the released state until the personnel downhole have safely returned. Only after the remote monitoring platform confirms that all personnel have evacuated can the locking state be re-entered.

[0014] Furthermore, the downhole liquid level is obtained by submersible liquid level sensing or ultrasonic liquid level detection, and the harmful gas content is obtained by electrochemical sensing or infrared absorption sensing. The edge adjudication unit that performs the permission determination step transmits data to the remote monitoring platform through a narrowband remote communication link. When the signal strength of the narrowband remote communication link is lower than the preset strength threshold value, it automatically switches to the cellular communication backup link.

[0015] To achieve the above objectives, the present invention provides the following technical solution: A system for implementing intelligent control methods for downhole operation safety includes: The access verification unit, in response to the opening request received by the wellhead locking device, obtains the identity credentials of the request initiator, the current time domain information, and downhole multi-dimensional status parameters; the downhole multi-dimensional status parameters include at least the downhole liquid level height, the downhole harmful gas content, and the integrity characterization value of the cap structure. Comprehensive license determination steps: Determine whether the current request for activation simultaneously meets the following license conditions: (i) The identity credentials belong to a pre-approved list of operational personnel; (ii) The current time domain information falls within the pre-defined allowed operation time window; (iii) The downhole liquid level is lower than the preset liquid level safety boundary value, and the downhole harmful gas content is lower than the preset gas safety boundary value, and the integrity characterization value of the cap structure is valid; If the above conditions are met simultaneously, the unlock control unit issues a release command to the wellhead locking device; if any condition is not met, the locking state is maintained and a corresponding blocking prompt message is output. The maintenance of the locking state only applies to wellhead locking devices that are currently in the locking state and does not affect wellhead locking devices that are already in the releasing state from continuing to maintain the releasing state.

[0016] Compared with the prior art, the beneficial effects of the present invention are: This invention implements multi-constraint proactive access control, achieving source management: Existing technologies often only conduct environmental monitoring after the well cover (manhole cover) is opened or passively trigger alarms after an accident. This invention sets identity verification, time period verification, downhole environmental parameter detection, and well cover structural status detection as preconditions for unlocking: Unlocking is not possible if the identity credential is not on the approved list, if the current time domain information is not within the permitted time period, if the downhole liquid level or gas level is not up to standard, or if the well cover has physical damage or settlement. If any condition is not met, the well cover cannot be opened, physically preventing unauthorized entry into the well, entry during unauthorized time periods, and risky entry into the well under hazardous environments. This represents a fundamental shift from "post-event response" to "pre-event prevention."

[0017] This invention proactively identifies the structural condition of manhole covers before unlocking, filling a technological gap: existing manhole cover monitoring solutions only trigger alarms after the cover is illegally opened or moved, a reactive "responding after an anomaly occurs" mode that cannot identify physical damage or potential settlement hazards before the cover is opened. This invention proactively acquires the integrity characteristics of the manhole cover structure before unlocking through a three-pronged mechanism: First, it collects geomagnetic field components at multiple locations around the manhole wall and compares them with a benchmark value, determining whether unauthorized displacement has occurred based on the degree of magnetic field distortion; second, it collects the cover's pitch angle, roll angle, ambient light transmittance, and vibration response signals, extracting statistical features to classify and identify significant displacement or damage; third, it constructs a response value sequence based on the energy change of the vibration response signal within a continuous time window, determining physical damage or foundation settlement by comparing the statistical dispersion with the central value. These three mechanisms are independent yet complementary, elevating the detection of the manhole cover structure condition from "post-event alarm" to "proactive screening before unlocking," filling the technological gap in existing technologies where the structural health status of the manhole cover cannot be known before opening.

[0018] This invention features a multi-independent detection and fusion judgment mechanism, significantly reducing the false alarm rate. The determination of the seal's structural integrity does not rely on a single detection method, but rather utilizes three independent detection channels operating in parallel based on different physical principles: magnetic field distortion detection (based on the disturbance of the Earth's magnetic field by the movement of ferromagnetic materials), attitude-transmittance-vibration joint classification and recognition (based on pattern classification using multi-dimensional statistical features), and vibration response statistical discrimination (based on quantitative analysis of statistical parameters related to energy changes). These three mechanisms reflect the structural state of the seal from different dimensions. When any channel determines a structural anomaly, the seal's structural integrity characterization value is marked as faulty and unlocking is prohibited; only when all three are determined to be intact can the next stage proceed. This "triple independent detection, one-vote veto" fusion judgment mechanism effectively avoids false alarms caused by environmental interference (such as the passage of surrounding ferromagnetic objects or occasional impact vibrations) or missed alarms due to detection blind spots caused by single detection methods.

[0019] This invention features a dynamically configurable safety boundary value classification system to adapt to complex operating conditions: Different well types (drainage, power, and gas) have fundamentally different safety requirements for environmental parameters: the liquid level boundary value for drainage wells can be appropriately relaxed, but the gas boundary value should be more stringent; gas wells are extremely sensitive to methane concentration; and power wells have strict requirements for both liquid level and gas. For the same well type, its safety boundary values ​​should also be dynamically adjusted under different seasons (rainy and dry seasons) and operating conditions. This invention uses a remote monitoring platform to set differentiated liquid level and gas safety limits according to the cap type, and supports dynamic correction based on historical data before issuing them to the edge adjudication unit. This ensures that the access judgment rules match the actual operating conditions, avoiding "one-size-fits-all" boundary values ​​that lead to "locking where it should not be locked" or "locking where it should not be locked incorrectly."

[0020] This invention features edge computing and remote collaboration, ensuring reliable response: Downhole multi-source data is filtered and preprocessed in real time by the edge adjudication unit and fused and judged locally. Even if the communication link is interrupted, the edge adjudication unit can still perform locking decisions based on locally stored boundary values ​​and whitelists, ensuring that basic security functions do not fail; the remote monitoring platform performs final authorization and global boundary value management, combining the advantages of rapid local response and unified central control.

[0021] This invention provides proactive early warning and preventative measures: when the downhole liquid level or gas concentration approaches the safety limit, an early warning signal is sent, allowing maintenance personnel to arrange for drainage or ventilation in advance. Environmental parameters are adjusted to a safe range before allowing personnel to enter the well, avoiding delays caused by temporary measures. Furthermore, this invention employs a tiered response strategy when the operation duration exceeds the permitted operating time window: after an overtime occurs, the system sends an overtime notification to the remote monitoring platform and blocks new opening requests. Simultaneously, it automatically identifies the current wellhead locking device status—if it is locked, new release actions are prohibited; if it is released, the released status is maintained, ensuring the return passage for personnel already downhole is unaffected. Locking is only restored after all personnel have safely evacuated. This tiered response logic ensures compliant control of operation time while fully protecting the escape rights of downhole personnel, avoiding the safety hazard of personnel being trapped underground due to improper management. Attached Figure Description

[0022] Figure 1 This is a schematic diagram of the structure of the present invention.

[0023] Figure 2 This is a schematic diagram of the system structure of the present invention.

[0024] In the diagram: 1-Wellhead, 2-Sealing cap, 3-Downhole integrated sensing unit, 4-Wellhead locking device. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Please see Figure 1 The present invention provides a technical solution: I. System Overall Architecture This embodiment provides an intelligent control system for downhole operation safety, the core components of which are as follows: The system includes a closable cover 2 at the wellhead 1 and a wellhead locking device 4 for opening and locking. A downhole integrated sensing unit 3 is installed inside the cover 2.

[0027] Wellhead locking device 4 (an electromechanical lock installed between the well cover and the well base, with a built-in safety chip and real-time clock), downhole integrated sensing unit 3 (including liquid level acquisition component, gas concentration acquisition component, magnetic field acquisition component, attitude acquisition component, light transmission intensity acquisition component and vibration response acquisition component), edge adjudication unit (industrial-grade embedded computer), remote monitoring platform, communication module, alarm device and controller.

[0028] The system workflow is as follows: After the operator initiates an opening request through a near-field mobile terminal or a remote platform, the system first acquires downhole safety parameters through the various acquisition components of the downhole integrated sensing unit; the edge adjudication unit performs filtering preprocessing and local fusion discrimination on the multi-source data, and uploads the processed feature data to the remote monitoring platform; the platform comprehensively determines the identity, time period, environmental parameters, and cap structure status; an unlocking command is issued only when all conditions are met; if any condition is not met, the lock remains closed.

[0029] II. Formulaic Implementation of Pre-Unlock Inspection of the Sealing Structure Status 2.1 Multi-location geomagnetic field component acquisition and benchmark establishment On the same horizontal circumference of the wellhead inner wall, three locations evenly distributed at 120° angles to each other were used to collect geomagnetic field components (based on the principle of anisotropic magnetoresistance, range ±8 Gauss, resolution better than 0.01 μT). Each collection point started working before unlocking, continuously sensing whether there was any unauthorized displacement or abnormal positional shift of the cap.

[0030] After the system is installed and debugged, under stable conditions with the cover properly closed and no external ferromagnetic disturbance, continuous data collection is performed. Q Group sample (sampling period 1 second, total 1 hour) Q =3600). No. p One collection location ( p =1,2,3) in the first q Axial ( q =1,2,3, corresponding to the three orthogonal directions of the spatial rectangular coordinate system respectively) t The sample value is B p,q ( t ).

[0031] Calculate the baseline mean value for each axis at each acquisition location:

[0032] in: For the first p The location is in q The standard deviation of the axial magnetic field, in μT, is used to characterize the normal fluctuation amplitude of the magnetic field in each axis under steady-state conditions.

[0033] 2.2 Real-time Deviation Normalization and Comprehensive Distortion Measurement Calculation During the real-time monitoring phase, magnetic field components are collected each time. B p,q ( t After that, calculate its deviation from the benchmark value and normalize it:

[0034] in: d p,q ( t ) is the first p The location is in q Axial, t The normalized deviation at any given time (dimensionless) represents the degree by which the current magnetic field deviates from the reference relative to the normal fluctuation range. B p,q ( t ) for t The real-time magnetic field component readings at each moment. The physical meaning of the above normalization operation is: to convert the original magnetic field deviation (unit μT) into a dimensionless quantity scaled by a reference standard deviation, making the deviations of each axis comparable. For example, when the real-time magnetic field value of a certain axis deviates from the reference value by exactly 1 reference standard deviation, the normalized deviation is 1.0; when it deviates by 2 reference standard deviations, the normalized deviation is 2.0.

[0035] No. p The comprehensive distortion metric for each acquisition location is defined as the square root of the sum of the squares of the three axial normalized deviations (i.e., the Euclidean norm), used to comprehensively characterize the degree of magnetic field distortion at that location:

[0036] Where: Δ p ( t ) is the first p The location is in t The overall distortion measure at any given time (dimensionless) indicates that the larger the value, the more significant the magnetic field distortion at that location.

[0037] Based on electromagnetic field simulation and on-site calibration experiments, when the cover is horizontally displaced by 12cm, the axial magnetic field components at each acquisition location change significantly. The overall distortion metric Δ... p ( t The measured value is approximately 3.2. Set the judgment boundary value. k =3.0 (Reserving a safety margin to ensure that the judgment is triggered before the displacement reaches 12cm), the first p Each acquisition location outputs a binary indicator:

[0038] in: I p ( t ) is the first p The location is in t The binary determination indicator at any given time is set to 1 to indicate that a significant displacement of the cover has been determined at that position, and to 0 to indicate that no significant displacement has been determined. k =3.0 is the boundary value for judging the comprehensive distortion metric.

[0039] Employing a 3-out-of-2 voting logic eliminates occasional single-point misjudgments:

[0040] Among them: Λ( t ) for t The sum of the judgment values ​​at the three acquisition positions at time t, with a value range of 0, 1, 2, and 3. If Λ( t If the value is greater than or equal to 2, it is determined that the cap has undergone unauthorized displacement exceeding 12cm, and the cap structure integrity indicator is invalidated. S cover If the setting is invalid, the locking device remains locked. The reason for using the three-out-of-two voting logic is that a single acquisition position may cause occasional misjudgments due to local magnetic field anomalies, while when any two of the three acquisition positions are simultaneously determined to exceed the boundary value, the probability of the cover actually shifting increases significantly.

[0041] As a supplement to the detection of magnetic field distortion, this embodiment further verifies the structural status of the manhole cover by combining attitude-transmission-vibration signal acquisition and classification.

[0042] 2.3 Attitude-Light Transmission-Vibration Combined Signal Acquisition The pitch angle was acquired using a six-axis inertial acquisition method at the center of the lower surface of the cover. and roll angle ψ (Unit: °). Under normal operating conditions, the cap is in a horizontal position. When the cover tilts due to foundation settlement or damage, and ψ The absolute value increases.

[0043] Ambient light transmittance was collected at the edge of the lower surface of the cap. (Unit: Lux), photosensitive side facing up. In the normally closed state, the cap is opaque. When the cap is damaged by cracks, holes, or other defects, ambient light can penetrate into the well. Significantly increased.

[0044] Vibration response signals were collected at six locations: the center of the cap and five locations at a radius of 1 / 2 from the center (one every 72°). u j( j =1,2,…,6 (unit: V or m / s²). These locations continuously sense the vibration response characteristics of the cover under external stimuli such as vehicle running over it and people stepping on it.

[0045] Edge adjudication unit at sampling frequency f s Each signal is acquired at 200 Hz, and an analysis window is formed every 0.5 seconds (i.e., 100 sampling points) to calculate the statistical characteristics once.

[0046] 2.4 Feature Vector Construction Calculate the statistics for 100 sampling points within each analysis window.

[0047] Attitude characteristics: Take the arithmetic mean of pitch and roll angles:

[0048] in: To analyze the pitch angle within the time window ϕ The arithmetic mean of (unit: °). To analyze the roll angle within the time window ψ The arithmetic mean of (unit: °). k The sampling point number is used. The reason for using the arithmetic mean instead of the instantaneous value is that a single sampling point may have noise fluctuations, while the mean of 100 points can stably represent the average posture of the cover within 0.5 seconds.

[0049] Light transmission characteristics: Take the median illuminance:

[0050] in: To analyze the illuminance within the window The median (in Lux). The median was used instead of the mean to mitigate occasional, transient strong light interference.

[0051] Vibration characteristics: For the vibration response signal at each acquisition location, the peak value and root mean square value are calculated respectively.

[0052] in: For the first j The peak value (unit: V or m / s²) at each location within the analysis window reflects the maximum amplitude of the vibration.

[0053]

[0054] in: For the first j The root mean square value (unit: V or m / s²) at each location within the analysis window reflects the average energy of the vibration.

[0055] Finally, an 11-dimensional feature vector is constructed:

[0056] in: It is an 11-dimensional feature vector, whose components correspond to the mean pitch angle, mean roll angle, median illuminance, vibration peak values ​​at six positions, and root mean square vibration values ​​at the first two positions.

[0057] 2.5 Statistical Classification and Identification A statistical classification model using radial basis function kernels is employed. The kernel function is defined as:

[0058] in: The radial basis function kernel is used to measure two eigenvectors. and The similarity between them; The square of the Euclidean distance between the two eigenvectors; The kernel width parameter. The value of affects the smoothness of the classification boundary, as verified by cross-validation. The classification effect is optimal at that time.

[0059] During the training phase, 2,500 labeled samples were collected for each of the four states: (0) Normal state - the cap is intact and in the correct position; (1) Slightly tilted state - the pitch or roll angle is between 5° and 15°; (2) Significantly offset state - the pitch or roll angle is greater than 15°, or the cap has obvious horizontal displacement; (3) Physically damaged state - the cap has visible cracks or holes.

[0060] Regularization parameters were determined through cross-validation. C =8. C The tolerance of the control model for classification errors of training samples: C The larger the value, the more rigorous the fit to the training samples, but the generalization ability may decrease. C If the value is too small, it may result in underfitting. C =8 strikes a balance between training accuracy and generalization ability.

[0061] After training, the decision function is:

[0062] in: F (x) is the classification decision function, which outputs the category label; The number of support vectors; For the first mThe Lagrange multipliers of the support vectors are obtained by solving a convex quadratic programming problem. For the first m The category label of each sample (in a one-to-many strategy, it indicates whether the sample belongs to the target category); b This is a bias term.

[0063] In the online identification phase, the feature vector x, constructed in real time, is input into the decision function, and the output label is mapped as follows: 0 - normal state, 1 - slightly tilted state, 2 - significantly offset state, and 3 - damaged state. When the output is 2 (significantly offset state) or 3 (damaged state), the integrity of the cap structure is indicated. S cover If the setting is disabled, the locking device remains locked.

[0064] 2.6 Statistical Judgment of Vibration Response This method serves as a supplementary validation to statistical classification and identification, using statistical parameters of vibration response to quantitatively assess the state of the cap structure. The physical principle of this method is as follows: when the cap structure is intact, the vibration energy distribution between each sampling location is consistent; when the cap suffers physical damage, the transmission of vibration energy in the crack area is hindered, and the dispersion of energy distribution between different locations increases; when the cap experiences foundation settlement, the entire cap sinks, causing an overall increase in vibration energy at each location, but the dispersion does not change significantly.

[0065] (1) Vibration energy calculation Triaxial acceleration was collected at six vibration response acquisition locations. a j,r ( t (Unit: m / s²), where j =1,2,…,6 are the data collection location numbers. r ∈{ x , y , z} indicates the axis. Based on the sampling frequency. f s =250Hz sampling, with each second constituting an analysis window ( N =250 sampling points).

[0066] Definition of the first j The position in the n Vibrational energy within a window:

[0067] in: E j ( n ) is the first j The position in the n Vibrational energy within a window, in units of (m / s²)²;N =250 is the number of sampling points within each analysis window; k The sampling point number within the window; a j,x ( k ), a j,y ( k ), a j,z ( k ) are respectively the first j The position in the k The triaxial acceleration values ​​at each sampling point. The purpose of averaging the sum of squares is to eliminate the dimensional effects caused by different window lengths.

[0068] (2) Energy difference in continuous window Take continuous N w =8 windows, with the starting window number being... n 0, calculate the energy difference between adjacent windows:

[0069] Where: Δ E j ( m ) is the first j The position in the m Energy difference between adjacent window pairs, in units of (m / s²)²; N w =8 represents the number of consecutive windows; m This is the difference index. The above operation yielded... N w−1 =7 differences, reflecting the changing trend and fluctuation amplitude of vibration energy at each acquisition location within 8 consecutive windows.

[0070] (3) Total change and comprehensive response value No. j Total vibration energy change at each sampling location within these 8 windows:

[0071] in: V j For the first j The total change in vibrational energy at each location, expressed in m / s², represents the overall fluctuation of vibrational energy at that location over 8 seconds. The changes at the six locations form a vector V = [ V 1, V 2,…, V 6] T .

[0072] The overall response value of this analysis unit:

[0073] in: C The overall response value for a single analysis unit, in units of (m / s²)²; Let V be the 2-norm of vector V. This value reflects both the amplitude of vibration changes at each position and the consistency of changes between positions.

[0074] (4) Multiple statistical analyses and benchmark establishment Continuous data acquisition K =18 analysis units, resulting in the response value sequence Γ1,Γ2,…,Γ18.

[0075] Calculate the statistical center (sample mean) of the sequence:

[0076] in: The statistical center value (sample mean) of the response value sequence, in units of (m / s²)², reflects the average level of the cover vibration response during this period; K =18 represents the total number of analysis units.

[0077] Calculate the statistical dispersion (sample standard deviation) of the sequence:

[0078] in: The statistical dispersion (sample standard deviation) of the response value sequence, in units of (m / s²)², reflects the degree of fluctuation of the response values ​​between each analysis unit, i.e., the stability of the vibration response.

[0079] After the system is installed and debugged, it is run continuously for 7 days under normal conditions with the cap intact, and the average value of the 7 days is taken as the baseline center value. and benchmark dispersion .

[0080] (5) Judgment criteria The criteria for judging physical damage are:

[0081] in: l 1 = 1.25 is the first boundary value for determining physical property damage; The reference center value; s Γ0 The baseline dispersion is determined based on the following: continuous monitoring of normal capped cells for 3 months, and the dispersion ratio. s Γ / s Γ0The 95th percentile is 1.08; for monitoring of caps with known cracks, the minimum ratio is 1.28, and 1.25 is chosen to balance sensitivity and resistance to false alarms.

[0082] The criteria for determining foundation settlement are:

[0083] in: l 2=1.25 is the second boundary value for determining basic settlement; l =3=0.12 is the third boundary value for basic settlement determination, representing the upper limit of the acceptable deviation of the dispersion ratio relative to 1. Mean ratio boundary value l The basis for determining 2=1.25 is: continuous monitoring of normal capped cells for 3 months. The 95th percentile is 1.11; for capped monitoring with known settlement of more than 2 cm, this ratio is the smallest at 1.22. An absolute deviation of the dispersion ratio from 1 of less than 0.12 means that the variation in standard deviation does not exceed 12%, which is within the normal fluctuation range.

[0084] When any of the above conditions are met, the integrity of the cap structure is indicated. S cover If the setting is disabled, the locking device remains locked. Besides the manhole cover's structural condition, downhole environmental parameters (liquid level, harmful gas concentration, etc.) are also key factors determining whether the lock can be opened. The following provides a detailed explanation of the perception and determination of downhole environmental parameters.

[0085] III. Formulaic Implementation of Downhole Environmental Parameter Sensing and Judgment 3.1 Environmental Parameter Collection Downhole liquid level height is obtained using an immersion-type liquid level acquisition method (range 0~10m, accuracy ±0.5%FS). h (Unit: m), sampling frequency 10 Hz. Hydrogen sulfide concentration was obtained using electrochemical acquisition (range 0~100 ppm). C H2S (Unit: ppm), sampling frequency 1Hz. Methane concentration was obtained using infrared absorption sampling (range 0~100% LEL). C CH4 (Unit: %LEL), sampling frequency 1Hz. Oxygen concentration was obtained using electrochemical acquisition (range 0~30% VOL). C O2 (Unit: %VOL), sampling frequency 1Hz.

[0086] 3.2 Environmental Parameter Filtering Preprocessing The raw data from each acquisition channel may contain transient spikes in interference, which are eliminated using a recursive averaging filtering algorithm:

[0087] in: for t Parameter values ​​after filtering at any time (units depend on the specific parameters); x ( t ) represents the original sampled value of the current time domain information. N f The length of the filter window is 10 for liquid level and 5 for gas. T s Sampling period (unit: s); i This represents the forward sampling sequence number. The physical meaning of recursive averaging filtering is: taking the current time domain information and the past... N f The arithmetic mean of the data from -1 sampling times is used as the filtered output of the current time domain information, which has a good effect on suppressing high-frequency random noise.

[0088] After filtering, reliable environmental parameters for the current time domain information are obtained: .

[0089] 3.3 Security Boundary Value Configuration The remote monitoring platform has a built-in security boundary value database, which sets boundary values ​​for different capping categories:

[0090] in: H max For liquid level safety limits (unit: m); For the safety limit of hydrogen sulfide (unit: ppm); The values ​​represent the safety limits for methane (unit: %LEL). Each boundary value is determined based on relevant industry safety standards and on-site measured data, and can be remotely updated and distributed via a remote platform.

[0091] 3.4 Comprehensive assessment of environmental safety Liquid level safety sign:

[0092] in: E liq This is a liquid level safety indicator; a value of 1 indicates that the liquid level meets the standard, and a value of 0 indicates that the liquid level exceeds the standard. for t The liquid level height after filtering at any given time.

[0093] Comprehensive gas safety sign:

[0094] in: E gasThis is a comprehensive gas safety indicator; a value of 1 indicates that all gas parameters meet the standards, while a value of 0 indicates that at least one gas parameter exceeds the standard. They are respectively t The concentrations of hydrogen sulfide, methane, and oxygen after filtering at different times; This is the lower limit of oxygen concentration. This represents the upper limit of oxygen concentration.

[0095] Comprehensive environmental safety signs:

[0096] in: E env This is a comprehensive environmental safety indicator, with 1 indicating that the environment is completely safe and 0 indicating that there is an environmental risk.

[0097] 3.5 Early Warning The edge adjudication unit monitors the changing trends of various environmental parameters in real time, and will detect when any parameter reaches its corresponding safety limit. or At that time, an early warning signal is sent to the remote monitoring platform, among which or =80%.

[0098] The warning trigger conditions are:

[0099] in: For the first k The warning indicator for this environmental parameter is set to 1, indicating that a warning has been triggered. x k ( t ) is the first k The current filtered value of the environmental parameter; X max For its corresponding safety limit (take the following values ​​for oxygen concentration respectively) and ).

[0100] When any parameter reaches or exceeds the safety limit ( xk ( t )≥ X max When the oxygen concentration exceeds the specified range, the edge adjudication unit immediately sends a prohibition unlocking command to the wellhead locking device and drives the alarm device to issue an alarm.

[0101] IV. Comprehensive Access Determination and Unlocking Control 4.1 Identity and Time Period Verification Authentication Flags P id The wellhead locking device compares the requester's identity credentials with the locally stored authorized list. If they match, it sets the value to 1; otherwise, it sets the value to 0.

[0102] Time period verification mark P time : Current time domain information t License time window with local storage [ t start , t end Compare, if t start ≤ t ≤ t end If the result is 1, then take 1; otherwise, take 0.

[0103] 4.2 Comprehensive Admission Judgment Edge adjudication units or remote monitoring platforms perform comprehensive access control determination:

[0104] Where: Unlock is the comprehensive access control result, with 1 indicating that unlocking is allowed and 0 indicating that unlocking is prohibited; P id For authentication purposes; P time This is a time-period verification indicator; E env It serves as a comprehensive environmental safety sign. S cover The integrity indicator of the sealing structure (derived from the triple detection in Part Two, when magnetic field distortion detection is not triggered, classification and identification output is not in a significantly offset state or a physical property damage state, and vibration response statistical judgment does not trigger damage or settlement judgment) S cover To be valid (take 1); when any of the above is triggered, S cover (Invalidated (set to 0)).

[0105] When Unlock=1 (i.e., all four flags are true simultaneously), the controller sends an unlocking command to the wellhead locking device, the locking device is released, and the cover can be opened.

[0106] When Unlock=0 (any flag is false), the locking device maintains the locked state and outputs the corresponding prohibition reason prompt.

[0107] V. Complete Workflow Example Scenario setting: A drainage cover (No. WH-003) is permitted to be used from 9:00 to 12:00 on June 21, 2026. The approving personnel are Zhang and Li.

[0108] Safety limits for liquid level in drainage caps H max =0.50m, Safety Limits .

[0109] At 9:10, Zhang arrived at the scene and initiated the request to open the door. Identity verification: Zhang is on the whitelist. P id =1.

[0110] Time period verification: 9:10 falls within the range of [9:00, 12:00]. P time =1.

[0111] Environmental assessment: Liquid level , It is 2 ppm (<10 ppm). The percentage was 20.8% (within the range of 19.5% to 23.5%), and all individual items met the standards. E env =1.

[0112] Structural determination: The combined distortion metrics at the three magnetic field acquisition locations are 0.8, 0.9, and 0.7, respectively, all less than the boundary value of 3.0, Λ=0<2, indicating no triggering; classification and identification output label 0 (normal state); vibration response statistics. None of them triggered a damage or settlement determination. S cover =1.

[0113] All four indicators are 1, Unlock=1, the lock is released, and Zhang is allowed to go down the well.

[0114] Time 10:00 (Simulated abnormal scenario - liquid level exceeding the limit): Suppose that due to heavy rain, the downhole fluid level rises to 0.62m. Mr. Li attempts to unlock the well, and the system detects this. The locking device remains locked and displays a message: "Downhole fluid level exceeds standard, do not go down the well."

[0115] At 11:45: The operation has reached 90% of the permitted duration, and the platform sends a "timeout is imminent" notification to Zhang's APP.

[0116] At 12:00: The system detected that the operation time had reached the upper limit of the permitted time window, automatically generated an overtime record, and reported it to the remote monitoring platform. From this moment on, the system blocked any new opening requests; at this time, the wellhead locking device remained in the released state, and the return passage for personnel Zhang underground remained unobstructed. At 12:03, after Zhang safely exited the well, the remote monitoring platform confirmed that all personnel had evacuated before the wellhead locking device could be locked, ending the current operation cycle, and all process data was automatically archived.

[0117] like Figure 2As shown, a system for implementing intelligent control methods for downhole operation safety includes: The access verification unit, in response to the opening request received by the wellhead locking device, obtains the identity credentials of the request initiator, the current time domain information, and downhole multi-dimensional status parameters; the downhole multi-dimensional status parameters include at least the downhole liquid level height, the downhole harmful gas content, and the integrity characterization value of the cap structure. Comprehensive license determination steps: Determine whether the current request for activation simultaneously meets the following license conditions: (i) The identity credentials belong to a pre-approved list of operational personnel; (ii) The current time domain information falls within the pre-defined allowed operation time window; (iii) The downhole liquid level is lower than the preset liquid level safety boundary value, and the downhole harmful gas content is lower than the preset gas safety boundary value, and the integrity characterization value of the cap structure is valid; If the above conditions are met simultaneously, the unlock control unit sends a release command to the wellhead locking device; if any condition is not met, the locked state is maintained and a corresponding blocking prompt message is output.

[0118] This invention achieves at least the following: First, the proactive access logic of "four constraints and no unlocking if standards are not met" eliminates the risk of going down the mine at the source. In existing technologies, whether it is traditional mechanical locks or electronic locks that have emerged in recent years, their core function is "unlocking after verifying identity." Whether the underground environment is safe and whether the sealing structure is intact depends entirely on the subjective judgment of the workers before they go down into the well or on temporary spot checks with handheld instruments. There is a lack of a mandatory detection mechanism linked to the locks. If the workers are negligent or ignore safety hazards in order to meet the deadline, accidents such as poisoning, suffocation, and drowning may occur.

[0119] This invention sets five conditions—identity verification, time period verification, liquid level detection, gas concentration detection, and cap structure status detection—as prerequisites for unlocking: all five conditions must be met before the lock can be released. Specifically, the lock will not unlock if the identity credential is not on the authorized list, if the current time information is not within the permission window, if the downhole liquid level or gas level is below standard, or if the cap has physical damage or settlement.

[0120] This proactive access logic of "no unlocking if not meeting the standards" achieves a qualitative change from the soft constraint of "informing about risks and reminding people to pay attention" to the hard constraint of "physical inability to unlock if the conditions are not met," thus simultaneously eliminating risky underground activities in dangerous environments from both the institutional design and physical enforcement levels.

[0121] 2. A "triple independent inspection + one-vote veto" mechanism for determining the integrity of the sealing structure: proactively identifying structural defects before unlocking, filling technical blind spots. Existing cap monitoring solutions, regardless of the detection method used, all trigger an alarm based on the logic of "an alarm being issued after the cap is opened or moved," which is a reactive, "responding only after an anomaly occurs" model. This means that when a cap has developed cracks due to long-term pressure but has not yet collapsed, this invention obtains the cap's structural integrity characterization value in parallel through three independent detection channels based on different physical principles before unlocking, and uses this characterization value as a rejection factor for unlocking, achieving three breakthroughs in principle: (1) Reversal of detection timing: The post-event response mode of "detection only after the cover is opened or moved" is reversed to the pre-event prevention mode of "active screening before unlocking", filling the time gap in the prior art where the structural state cannot be known before the cover is opened.

[0122] (2) Redundancy in detection dimensions: The three detection channels are based on different physical principles: magnetic field distortion detection (using the disturbance of the Earth's magnetic field by the movement of ferromagnetic materials), attitude-transmittance-vibration joint classification and recognition (using pattern classification based on multidimensional statistical features), and vibration response statistical discrimination (using statistical parameters of energy change for quantitative analysis). The three are independent of each other and map the structural state of the cap from different physical dimensions. This forms a "triple redundancy, mutual verification" judgment system, so that an occasional misjudgment by any single channel will not lead to an overall misjudgment.

[0123] (3) Enhanced judgment mechanism: A "one-vote veto" strategy is adopted: when any channel judges the structure to be abnormal, the structural integrity characterization value is marked as failure, and the cover cannot be physically opened. The three detection channels are independent and the judgment conditions are not coupled to each other, which is equivalent to setting up "three parallel locks" for the safety of the cover structure. As long as one lock judges to be abnormal, the entire system is prohibited from unlocking, which greatly reduces the probability of security vulnerabilities.

[0124] Third, the synergistic effect of the "environment + structure" dual physical quantity pre-sensing achieves a "1+1>2" level of safe access. This invention uses both environmental parameter detection and sealing structure detection as prerequisites for unlocking, achieving collaborative sensing and joint determination of two different physical safety parameters. The effect of this "environment + structure" joint operation permission mechanism far exceeds the sum of the effects of implementing either parameter individually. Even if the structural inspection passes, the lock will remain locked if the liquid level or gas level is below the standard. When the sealing structure is abnormal, the lock will remain locked even if all environmental parameters meet the standards. The safety indicators of both are combined using a logical AND operation to form a comprehensive environmental safety indicator. Any abnormality in either one constitutes an independent veto condition for unlocking.

[0125] This dual physical quantity-based veto access mechanism elevates the safety guarantee of downhole operations from "single-point protection" to "multi-point joint defense," achieving a qualitative leap in safety.

[0126] IV. Adaptive configuration of dynamic and differentiated safety boundary values: This solves the technical bias of "one-size-fits-all" boundary values ​​that are divorced from actual working conditions. In existing downhole operation safety management systems, the safety boundary values ​​for environmental parameters mostly adopt nationally unified standards or factory-fixed values. This static configuration method has a commonly overlooked but extremely dangerous flaw: it is normal for the liquid level in drainage wells to rise briefly during the rainy season, but fixed boundary values ​​may prohibit unlocking even in non-dangerous conditions, leading to maintenance delays; while gas wells are extremely sensitive to methane concentration, if the fixed boundary values ​​are set with reference to drainage well standards, the boundary values ​​may be too high to effectively prohibit entry into the well, resulting in catastrophic consequences.

[0127] This invention enables a remote monitoring platform to independently configure differentiated safety limits for liquid level, hydrogen sulfide, methane, and oxygen based on the cover category (drainage, electricity, and gas). It also supports dynamic correction based on historical data before distributing these limits to the edge adjudication unit. This "classified configuration, dynamic correction" boundary value management mechanism ensures that the access judgment rules always match the actual operating conditions, avoiding both security failures due to "unlocked locks" and maintenance delays due to "incorrectly locked locks," thus achieving a dynamic balance between security and availability.

[0128] V. Reliable Response Architecture for Edge Computing and Remote Collaboration: Admission Determination Can Still Be Performed Independently Even Offline Existing IoT lock control systems generally rely on cloud platforms for centralized judgment. Once the communication link is interrupted, the lock loses its access judgment capability or degenerates into a simple local card verification mode, losing the comprehensive judgment capability of environment and structure.

[0129] This invention employs a two-layer decision-making architecture where an edge adjudication unit and a remote monitoring platform work in tandem. Downhole multi-source sensing data is first filtered and preprocessed by the edge adjudication unit, followed by local fusion and discrimination. Even if the narrowband remote communication link is interrupted, the edge adjudication unit can still independently execute a complete four-fold access determination based on locally stored security boundary values, authorized lists, and permitted time windows, maintaining the locking decision. The remote monitoring platform then performs final authorization and global boundary value management. This two-layer architecture ensures that basic security functions remain intact during communication interruptions, while simultaneously achieving complementary advantages between rapid local response and centralized unified control.

[0130] VI. The organic combination of advanced early warning and gradual risk intervention When the downhole environmental parameters approach a predetermined percentage of the safety limit, an early warning signal is sent, allowing maintenance personnel to arrange for drainage or ventilation in advance, adjusting the environmental parameters to a safe range before unlocking the well. This three-tiered, progressive intervention strategy—"early warning when approaching the boundary value, locking when reaching the boundary value, and rejection when structural abnormalities occur"—expands safety management from "binary judgment at the critical point" to "trend perception during the approach process," reserving sufficient window of opportunity for management personnel to handle situations and avoiding operational delays due to temporary measures.

[0131] In this invention, any parts not described herein are prior art, or may be the same as prior art, or may be known art, or may be implemented using prior art, and will not be described in detail here.

[0132] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent control and management of safety in downhole operations, characterized in that, include: Qualification verification steps: In response to the opening request received by the wellhead interlocking device, obtain the identity credentials of the request initiator, the current time domain information, and downhole multi-dimensional status parameters; the downhole multi-dimensional status parameters include at least the downhole liquid level, the downhole harmful gas content, and the integrity characterization value of the cap structure. Comprehensive license determination steps: Determine whether the current request for activation simultaneously meets the following license conditions: (i) The identity credentials belong to a pre-approved list of operational personnel; (ii) The current time domain information falls within the pre-defined allowed operation time window; (iii) The downhole liquid level is lower than the preset liquid level safety boundary value, and the downhole harmful gas content is lower than the preset gas safety boundary value, and the integrity characterization value of the cap structure is valid; Execution response steps: If the above conditions are met simultaneously, a release command is issued to the wellhead locking device; if any condition is not met, the locking state is maintained and the corresponding blocking prompt information is output. The maintenance of the locking state only applies to wellhead locking devices that are currently in the locking state and does not affect wellhead locking devices that are already in the releasing state from continuing to maintain the releasing state. The integrity characterization value of the cap structure is obtained through the following method: Spatial distribution parameters of the geomagnetic field are collected at multiple locations around the wellhead. The instantaneous geomagnetic field components in each direction at each location are compared with the pre-stored reference geomagnetic field distribution to obtain the magnetic field distortion measurement at each location. When the magnetic field distortion measurement at at least two locations exceeds the predetermined distortion boundary value at the same time, it is determined that the wellhead cover has undergone unauthorized displacement, and the integrity characterization value of the cover structure is marked as failure. Furthermore, the spatial attitude angle parameter of the wellhead cover, the ambient light transmittance parameter of the underside of the wellhead cover, and the vibration response parameters of multiple different parts of the wellhead cover are picked up; the statistical feature values ​​of each parameter are extracted with a preset time window, and the statistical feature values ​​are combined into a feature vector. Based on the feature vector, pattern classification is performed. When the classification result is a significant pose deviation or physical property damage, the integrity characterization value of the cover structure is marked as failure. Furthermore, a response value sequence is constructed using the energy fluctuation of each vibration response parameter within a continuous time window. The statistical dispersion and statistical central value of the response value sequence are calculated and compared with the pre-established benchmark dispersion and benchmark central value. When the preset physical property damage judgment criterion or foundation settlement judgment criterion is met, the integrity characterization value of the cover structure is marked as failure.

2. The intelligent control method for downhole operation safety as described in claim 1, characterized in that, The process of collecting spatial distribution parameters of the geomagnetic field at multiple locations around the wellhead specifically involves: collecting geomagnetic field components at three locations uniformly arranged at 120° angles to each other on the same horizontal circumference of the inner wall of the wellhead; comparing the instantaneous geomagnetic field components in each direction at each location with the pre-stored reference geomagnetic field components, and calculating the normalized deviation in each direction and obtaining the comprehensive distortion metric; when the comprehensive distortion metric at at least two of the three locations simultaneously exceeds the predetermined distortion boundary value, it is determined that the wellhead cap has undergone unauthorized displacement.

3. The intelligent control method for downhole operation safety as described in claim 1, characterized in that, The process of acquiring spatial attitude angle parameters of the wellhead cap, ambient light transmittance parameters on the underside of the wellhead cap, and vibration response parameters from multiple different locations on the wellhead cap involves: continuously acquiring these parameters at a predetermined sampling frequency; constructing an analysis window using a fixed number of continuous sampling points; calculating the arithmetic mean of the spatial attitude angle parameters, the median of the ambient light transmittance parameters, and the peak and root mean square values ​​of each vibration response parameter within each analysis window; combining these statistical feature values ​​into a feature vector; inputting the feature vector into a trained statistical classification model; the output categories of the statistical classification model include normal state, slight pose deviation state, significant deviation state, and physical property damage state; when the output is a significant deviation state or a physical property damage state, the integrity characterization value of the cap structure is marked as failure.

4. The intelligent control method for downhole operation safety as described in claim 1, characterized in that, The construction of the response value sequence based on the energy fluctuation of each vibration response parameter within a continuous time window specifically includes the following steps: The triaxial acceleration response signals at multiple radial locations of the wellhead cap are collected, and an analysis time window is formed by a predetermined number of continuous sampling points. The vibration energy value at each location within each time window is calculated. Take a predetermined number of consecutive analysis time windows, calculate the vibration energy difference between adjacent time windows, and then calculate the total vibration energy fluctuation of each part within the consecutive time windows; The total vibration energy fluctuations at all locations are calculated to obtain the comprehensive response value of the current analysis unit. The comprehensive response values ​​of a predetermined number of analysis units are continuously picked up to form a response value sequence, and the statistical center value and statistical dispersion of the response value sequence are calculated. The statistical central value and statistical dispersion are compared with the pre-established benchmark central value and benchmark dispersion, respectively: when the statistical central value is greater than the benchmark central value and the ratio of the statistical dispersion to the benchmark dispersion reaches a first predetermined boundary value, it is determined to be physical property damage; when the ratio of the statistical central value to the benchmark central value reaches a second predetermined boundary value and the absolute value of the deviation of the ratio of the statistical dispersion to the benchmark dispersion from 1 is less than a third predetermined boundary value, it is determined to be foundation settlement; when any of the above determinations are met, the integrity characterization value of the cap structure is marked as failure.

5. The intelligent control method for downhole operation safety as described in claim 1, characterized in that, The downhole multidimensional state parameters also include oxygen content and combustible gas content; the permissible condition (iii) also includes oxygen content being within a predetermined safe range and combustible gas content being lower than a preset combustible gas safety boundary value. When any of the parameters, such as the downhole fluid level, harmful gas content, oxygen content, or combustible gas content, reaches a predetermined percentage of its corresponding safety boundary value, an early warning signal is sent to the remote monitoring platform; when any parameter reaches or exceeds its corresponding safety boundary value, or when the integrity characterization value of the cap structure fails, a prohibition release command is issued to the wellhead locking device and the warning device is activated to issue a warning.

6. The intelligent control method for downhole operation safety according to claim 1, characterized in that, The liquid level safety boundary value and gas safety boundary value are set by the remote monitoring platform according to the category of the wellhead cover and sent to the edge adjudication unit for storage; the category of the wellhead cover includes at least drainage cover, power cover and gas cover, and the liquid level safety boundary value and gas safety boundary value corresponding to different categories are configured independently.

7. The intelligent control method for downhole operation safety according to claim 1, characterized in that, The method also includes post-release operation process recording and tracking steps: The process of workers entering the well is recorded by the image acquisition unit. The video image frames are used as input, and the feature extraction network identifies the wearing status of the workers' head protective gear. The temporal correlation tracking model continuously records the trajectory of the workers. When the operation duration exceeds the allowed operation window, a timeout warning is sent to the remote monitoring platform, and a locking prohibition command is issued to the wellhead locking device. The wellhead locking prohibition command only prevents new opening requests after the command is issued, and does not affect the current state of the wellhead locking device that is already in the released state, ensuring that downhole personnel can return to the surface at any time.

8. The intelligent control method for downhole operation safety according to claim 1, characterized in that, The downhole liquid level is obtained by submersible liquid level sensing or ultrasonic liquid level detection, and the harmful gas content is obtained by electrochemical sensing or infrared absorption sensing. The edge adjudication unit that performs the permission determination step transmits data to the remote monitoring platform through a narrowband remote communication link. When the signal strength of the narrowband remote communication link is lower than the preset strength boundary value, it automatically switches to the cellular communication backup link.

9. A system for executing the intelligent control method for downhole operation safety according to any one of claims 1 to 8, characterized in that, include: The access verification unit, in response to the opening request received by the wellhead locking device, obtains the identity credentials of the request initiator, the current time domain information, and downhole multi-dimensional status parameters; the downhole multi-dimensional status parameters include at least the downhole liquid level height, the downhole harmful gas content, and the integrity characterization value of the cap structure. Comprehensive license determination steps: Determine whether the current request for activation simultaneously meets the following license conditions: (i) The identity credentials belong to a pre-approved list of operational personnel; (ii) The current time domain information falls within the pre-defined allowed operation time window; (iii) The downhole liquid level is lower than the preset liquid level safety boundary value, and the downhole harmful gas content is lower than the preset gas safety boundary value, and the integrity characterization value of the cap structure is valid; If the above conditions are met simultaneously, the unlock control unit sends a release command to the wellhead locking device; if any condition is not met, the locking state is maintained and the corresponding blocking prompt information is output. The maintenance of the locking state only applies to the wellhead locking device that is currently in the locking state and does not affect the wellhead locking device that is already in the releasing state from continuing to maintain the releasing state. The integrity characterization value of the cap structure is obtained through the following method: Spatial distribution parameters of the geomagnetic field are collected at multiple locations around the wellhead. The instantaneous geomagnetic field components in each direction at each location are compared with the pre-stored reference geomagnetic field distribution to obtain the magnetic field distortion measurement at each location. When the magnetic field distortion measurement at at least two locations exceeds the predetermined distortion boundary value at the same time, it is determined that the wellhead cover has undergone unauthorized displacement, and the integrity characterization value of the cover structure is marked as failure. Furthermore, the spatial attitude angle parameter of the wellhead cover, the ambient light transmittance parameter of the underside of the wellhead cover, and the vibration response parameters of multiple different parts of the wellhead cover are picked up; the statistical feature values ​​of each parameter are extracted with a preset time window, and the statistical feature values ​​are combined into a feature vector. Based on the feature vector, pattern classification is performed. When the classification result is a significant pose deviation or physical property damage, the integrity characterization value of the cover structure is marked as failure. Furthermore, a response value sequence is constructed using the energy fluctuation of each vibration response parameter within a continuous time window. The statistical dispersion and statistical central value of the response value sequence are calculated and compared with the pre-established benchmark dispersion and benchmark central value. When the preset physical property damage judgment criterion or foundation settlement judgment criterion is met, the integrity characterization value of the cover structure is marked as failure.

Citation Information

Patent Citations

  • Well lid linkage system

    CN108560607A

  • Intelligent well lid monitoring system based on Internet of Things architecture

    CN111986360A