Well gun intelligent shooting real-time control management platform

Abstract

This invention discloses a real-time control and management platform for intelligent well firing, relating to the field of well firing control technology. It solves the technical problems of traditional solutions lacking multi-hop communication control between nodes, failing to monitor relay node status in real time, and being unable to dynamically adjust strategies based on real-time data, thus affecting the control effect of the management platform on intelligent well firing. The platform divides the well firing distribution area into multiple zones, with each zone equipped with multi-hop relay nodes, forming a hierarchical transmission path. This solves the problem of limited single-hop communication distance, making it particularly suitable for complex terrains such as mountainous areas and underground environments, expanding coverage through relay hops. Simultaneously, a production mode configuration and full-dimensional status management module enables flexible configuration of production modes, status tracking and visualization of well firing equipment and the entire operation process, operation trajectory playback and event tracing, and quality control management of the firing process, providing comprehensive data support for intelligent firing decision-making and operation process control.

Description

Technical Field

[0001] This invention belongs to the field of well-fired excitation control, specifically a real-time intelligent excitation control and management platform for well-fired excitation. Background Technology

[0002] Well-fire detonation is a key technology in petroleum, geological exploration, and mining. It involves manually burying explosives and detonating them at a designated well depth to generate seismic waves for underground structural detection. Traditional detonation methods rely on manual experience for parameter adjustment, resulting in low efficiency, poor safety, and unstable data quality. Especially in complex terrains (such as mountains and deserts), radio signal obstruction and frequent relocation of relay equipment significantly reduce operational efficiency. Therefore, a more intelligent real-time control and management platform for well-fire detonation is needed.

[0003] Traditional well-fired blasting real-time control and management schemes mostly rely on a single communication mode, such as 4G or Beidou short message, lack multi-hop communication control between nodes, and do not monitor the status of relay nodes in real time. This can easily lead to relay node failures causing delays in blasting commands and missing the best blasting opportunity. At the same time, due to the use of fixed timing blasting, it is impossible to dynamically adjust the strategy based on real-time data. Summary of the Invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a real-time control and management platform for intelligent well-blasting, which solves the technical problems of traditional solutions lacking multi-hop communication control between nodes, failing to monitor the status of relay nodes in real time, and being unable to dynamically adjust strategies based on real-time data, thus affecting the control effect of the management platform on intelligent well-blasting.

[0005] To address the aforementioned problems, a first aspect of the present invention provides a real-time control and management platform for intelligent well-fired excitation, comprising: Communication link construction module: For the distribution of well shots in the region, the distribution area of ​​well shots is divided into areas, and multi-hop relay nodes are set in each area. The well shot data acquisition nodes in the area establish communication links with the corresponding multi-hop relay nodes, and communication links are established between the multi-hop relay nodes. Among them, the communication links are established through 4G+BeiDou dual-mode communication. Communication Management Module: By establishing a link quality assessment function for multi-hop relay transmission, the module analyzes the link quality of multi-hop relay nodes, performs clock skew compensation based on the enhanced time synchronization protocol of PTPv2, and constructs a link health prediction model. Production mode configuration and status full-dimensional management module: It is used to realize flexible configuration of production mode, status tracking and visualization of well drilling equipment and operation process, operation trajectory playback and event tracing, as well as quality control management of the excitation process, providing comprehensive data support and visualization management methods for intelligent excitation decision-making and operation process control.

[0006] Intelligent parameter optimization module: Based on the multispectral imaging module, it collects blasting block size image data to identify block size, establishes a block size distribution statistical model and a well shot basic parameter optimization model, and automatically optimizes the well shot basic parameters. Digital twin construction module: Based on the optimized well shot basic parameters, the dangerous area is analyzed, and a combined digital twin of geology, blasting and production of the dangerous area is established based on the dynamic fluid pressure field, rock mass parameter field and environmental interference field; Intelligent Activation Module: Based on the link health prediction model and the joint digital twin of geology, blasting and production in the dangerous area, an intelligent activation model is constructed to carry out intelligent activation of well blasting.

[0007] Optionally, in one example of the above aspects, the data acquisition node acquires terrain data, rock data, environmental data, and basic parameters of the well shot, including: charge parameters, detonation parameters, and mesh parameters, of the well shot location and the surrounding preset safety area.

[0008] Optionally, in one example of the above aspects, the communication link building module includes: Area division unit: Based on the distribution of well shots and geographical environment, the area is divided into areas so that the number of well shots in the area is within a set threshold. A threshold is set for the total data collection frequency of all well shots in the area. Areas that exceed the threshold are further divided. Node communication configuration unit: Based on the communication protocol and transmission requirements of the well shot data acquisition module, select multi-hop relay node devices that support multiple communication protocols; configure the communication parameters of the well shot data acquisition node, and establish communication between the well shot data acquisition node and the corresponding multi-hop relay node through 4G+BeiDou dual-mode communication.

[0009] Optionally, in one example of the above aspects, the communication management module includes: The communication link evaluation unit analyzes the link quality of multi-hop relay nodes by establishing a link quality evaluation function for multi-hop relay transmission. Based on the link quality assessment coefficient, by setting a threshold for the link quality assessment coefficient, links that are below the threshold are given an early warning and marked. The link conversion unit formulates link conversion rules, including: after receiving an early warning signal, the upload link of a multi-hop relay node selects the maximum value of the communication link quality assessment coefficient between the multi-hop relay nodes and sends the data to the multi-hop relay node corresponding to the maximum value of the quality assessment coefficient for upload; after receiving an early warning signal, the communication link of the data acquisition module is immediately re-established. The link analysis unit, based on the enhanced time synchronization protocol PTPv2, establishes a clock skew compensation model, performs clock skew compensation, and constructs a link health prediction model.

[0010] Alternatively, in one example of the above aspects, the link quality evaluation function is constructed using the following formula: ; Where: LQi is the quality index of the i-th link, RSSIi(t) is the received signal strength (dBm) of the i-th link, RSSImax is the maximum value of the received signal strength in the historical data, BERi(t) is the bit error rate of the i-th link, BERmax is the maximum value of the link bit error rate in the historical data, Δti is the link interruption duration (s), α, β and γ are weighting coefficients, and λ is the time attenuation coefficient.

[0011] Optionally, in one example of the above aspects, based on the enhanced time synchronization protocol PTPv2, a clock skew compensation model is established, clock skew compensation is performed, and a link health prediction model is constructed, including the following steps: An enhanced time synchronization protocol based on PTPv2 is used to establish a clock skew compensation model. Based on the compensated clock deviation, clock deviation compensation is performed; By using historical link signal transmission data, a time window is set, and the link quality index and compensated clock deviation data within the time window are statistically analyzed. For the corresponding operating status of the link in the historical data, the normal operation data and the link failure operation data are labeled. An LSTM model is trained using the labeled data to build a link health prediction model.

[0012] Optionally, in one example of the above aspects, based on the acquisition of blast block size image data by a multispectral imaging module for block size identification, a block size distribution statistical model and a well shot basic parameter optimization model are established, and the well shot basic parameters are automatically optimized, including the following steps: A multispectral imaging module composed of near-infrared laser illumination and a high-speed CMOS camera is used to acquire blast block size image data for block size identification and establish a block size distribution statistical model. By establishing a multi-objective optimization function, an optimization model for the excitation parameters is constructed. Acquire historical data of terrain, rock, and environmental data and basic parameters of well shot locations from data acquisition nodes; calculate the cosine similarity between historical terrain, rock, and environmental data and current terrain, rock, and environmental data; filter historical data with similarity greater than a threshold; and calculate the comprehensive objective function F value of the filtered historical data based on the corresponding blasting block size image data. Calculate the F-values ​​of each set of parameters and the corresponding basic well shot parameters, construct the Kriging surrogate model, use the NSGA-II algorithm for multi-objective optimization, verify the Pareto front solution set, and obtain the optimized basic well shot parameters; set the initiation parameters and mesh parameters according to the optimized basic well shot parameters.

[0013] Optionally, in one example of the above aspects, analyzing the hazardous area based on the optimized well shot basic parameters includes the following steps: Based on the dynamic damage theory of rock mass, a real-time correction formula for blasting safety distance is established; Within the safe zone, a danger zone Ωdanger={(x,y,z)| is defined centered on the well site. ≤Rst, and z≥Hm}, where (x,y,z) are the coordinates in a three-dimensional coordinate system established with the well shot position as the origin, and Hm is the highest position of the blasted rock block.

[0014] Optionally, in one example of the above aspects, a combined digital twin of geology, blasting, and production for a hazardous area is established based on a dynamic fluid pressure field, rock mass parameter field, and environmental disturbance field, including the following steps: Based on the topographic data of the dangerous area where the well shot is located, dynamic fluid pressure field data is extracted, including rock permeability and groundwater level changes; dynamic rock mass parameter field data is extracted from rock data, including rock mass elastic modulus, rock strata interface and fault data; dynamic environmental disturbance field data is extracted from environmental data, including air temperature, air pressure and wind speed. Based on the extracted data, a joint digital twin of the geology, blasting, and production of the hazardous area is established, including: Geological Model: Use Gocad or Surpac to build a 3D geological model. The input is the rock mass elastic modulus, rock layer interface and fault data, and the output is the location of rock mass fracture. Using the locations of rock mass fractures in historical data as labels, the model is trained using labeled data to predict the locations of rock mass fractures. Blasting Model: A blasting dynamics model is established using LS-DYNA or AUTODYN. The inputs are dynamic fluid pressure field data, dynamic rock mass parameter field data, dynamic environmental disturbance field data, and basic well shot parameters. Using measured blasting vibration data from historical data as labels, the model is trained by adding labels to dynamic fluid pressure field data, dynamic rock mass parameter field data, and dynamic environmental disturbance field data, so that the model can predict blasting vibration data. Production Model: The MLP multilayer perceptron is used to establish a mining operation model. The inputs are the blasting volume of the well and the basic parameters of the well. Using the measured blasting efficiency from historical data as labels, an MLP model is trained by adding labels to the well-blasting volume and basic well-blasting parameters, so that the model can predict the blasting efficiency.

[0015] Optionally, in one example of the above aspects, based on the link health prediction model and the joint digital twin of geology, blasting, and production in the hazardous area, an intelligent ignition model is constructed to perform intelligent ignition of well shots, including the following steps: Based on the rock mass fracture location, blasting vibration data and blasting efficiency predicted by the joint digital twin, a 3D convolutional neural network is used to extract spatiotemporal feature vectors from the predicted data. Based on the link health prediction model, all links are predicted to be in normal operation and in failure operation. If a link is in normal operation, an association edge is established between nodes. If a link is in failure operation, the association edge between nodes is disconnected. The GNN graph neural network is used to model the dependencies between nodes on the link topology data, including degree centrality and betweenness centrality. The spatiotemporal feature vectors extracted by the 3D convolutional neural network and the inter-node dependencies of the output of the GNN graph neural network are fused through an attention mechanism to generate a fused feature vector. By sharing the feature extraction networks of the underlying 3D convolutional neural network and the GNN graph neural network, prediction results for different targets are output separately, thus constructing an intelligent activation model: The output target is the allocation of explosive charge to each borehole: By analyzing the fusion feature vector and rock fracture intensity of historical data, and labeling the fusion feature vector and rock fracture intensity of each blast hole based on the explosive charge data of historical data, a regression task network is trained to analyze the explosive charge of each blast hole. The output target is generated according to the firing sequence of the boreholes: By analyzing the fused feature vectors of historical data and link failure operation data, and labeling the fused feature vectors and link failure operation data according to the firing order of the historical data, a Transformer decoder is trained to perform sequence generation tasks and analyze the firing order of the boreholes. The output target is the risk classification of vibration exceeding the standard: By analyzing the fusion feature vectors of historical data and blasting vibration data, the fusion feature vectors and blasting vibration data are labeled according to whether the blast holes have the risk of vibration exceeding the standard. A random forest network is trained to complete the binary classification task and analyze whether the blast holes have the risk of vibration exceeding the standard. Intelligent ignition of well shots is performed based on the results output by the intelligent ignition model.

[0016] Compared with the prior art, the beneficial effects of the present invention are: This invention divides the well-fired equipment distribution area into multiple zones, each with multiple relay nodes to form a hierarchical transmission path. This solves the problem of limited single-hop communication distance, making it particularly suitable for complex terrains such as mountainous and underground areas, expanding coverage through relay hops. Simultaneously, it provides high-speed data transmission, applicable to surface or shallow areas, supporting real-time video monitoring and big data uploading. In remote areas without 4G signals or when 4G networks are congested, it transmits critical control commands via satellite communication, achieving "all-area, all-time" communication coverage and avoiding control failures due to the failure of a single communication mode.

[0017] This invention utilizes a block size distribution statistical model, combined with basic wellbore parameters such as borehole diameter, borehole spacing, and charge quantity, to establish a parameter-block size response relationship, thereby achieving automatic parameter adjustment. This avoids the lag of manual parameter adjustment. By integrating a digital twin with dynamic fluid pressure fields, rock mass parameter fields, and environmental interference fields, it simulates blasting energy distribution. By comparing the actual block size distribution with the simulation results, the parameter model is dynamically corrected. For example, if excessive block size is detected in a certain area, the parameters of surrounding boreholes are automatically adjusted to supplement energy.

[0018] This invention utilizes a combined geological-blasting-production model, integrating geological structure, ignition parameters, and production disturbances into a digital twin to establish a three-dimensional risk field. Traditional methods consider only single factors, while the combined model can identify complex risks, improving the accuracy of early warning. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0020] Figure 1 This is a schematic diagram of the system framework of the present invention; Figure 2 This is a schematic diagram of the communication link construction module framework of the present invention; Figure 3 This is a schematic diagram of the communication management module framework of the present invention. Detailed Implementation

[0021] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. 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.

[0022] Example 1 Please see Figures 1-3 A real-time control and management platform for intelligent well firing is provided, comprising: Communication link construction module: For the distribution of well shots in the region, the distribution area of ​​well shots is divided into areas, and multi-hop relay nodes are set in each area. The well shot data acquisition nodes in the area establish communication links with the corresponding multi-hop relay nodes, and communication links are established between the multi-hop relay nodes. Among them, the communication links are established through 4G+BeiDou dual-mode communication. Communication Management Module: By establishing a link quality assessment function for multi-hop relay transmission, the module analyzes the link quality of multi-hop relay nodes, performs clock skew compensation based on the enhanced time synchronization protocol of PTPv2, and constructs a link health prediction model. Production mode configuration and status full-dimensional management module: It is used to realize flexible configuration of production mode, status tracking and visualization of well drilling equipment and operation process, operation trajectory playback and event tracing, as well as quality control management of the excitation process, providing comprehensive data support and visualization management methods for intelligent excitation decision-making and operation process control.

[0023] Intelligent parameter optimization module: Based on the multispectral imaging module, it collects blasting block size image data to identify block size, establishes a block size distribution statistical model and a well shot basic parameter optimization model, and automatically optimizes the well shot basic parameters. Digital twin construction module: Based on the optimized well shot basic parameters, the dangerous area is analyzed, and a combined digital twin of geology, blasting and production of the dangerous area is established based on the dynamic fluid pressure field, rock mass parameter field and environmental interference field; Intelligent Activation Module: Based on the link health prediction model and the joint digital twin of geology, blasting and production in the dangerous area, an intelligent activation model is constructed to carry out intelligent activation of well blasting.

[0024] Specifically, in this embodiment, the data sent by the data acquisition node to the production mode configuration and status full-dimensional management module through the communication management module includes: Basic data of well-fired equipment: trigger point ID, detonator model, detonator subplate type, electronic detonator manufacturer code, electronic detonator delay parameters, charging preparation time, report timeout time, and vibration alignment parameters; Core production operation data: production mode identifier, time slot sequence number, line number, point number, FO time, TB time, TB deviation value, GPS timestamp, firing waiting time, number of firings, number of design combinations; location and environmental data: wellhead coordinates, real-time position of the gunner, position of the command center, boundary data of the mission area, offset distance; among which, the production mode identifier includes: TD mode, fixed time interval mode, and fixed time slot mode. Status and quality control data: excitation point status identifier, equipment online time, excitation time, working mode, connection status, task progress, equipment power, communication mode, maximum wellhead energy value, abandoned blasting identifier, and abnormal event type; among which, the excitation point status identifier includes: not excitation, excitation completed, multiple excitations, offset excitation, and cannot be excitation, and the abnormal event type includes: excitation abnormality, working status abnormality, quality control abnormality, and system abnormality; Personnel-related data: the gunner's ID number; The production mode configuration and status full-dimensional management module includes: Production mode setting unit: Supports configuration of TD mode, fixed time interval mode and fixed time slot mode. Among them, TD mode adopts priority scheduling, real-time operation and queuing activation mechanism. Fixed time slot mode supports timed independent activation and does not require central control, and parameter setting can be completed easily. Status tracking unit: Real-time tracking of well-fired equipment's online time, firing time, working mode, connection status, task progress, equipment power level, and communication mode; Visualization unit: Visually presents the gunner's position, firing status, and command center position, and simultaneously displays the firing point's unfired, firing completed, multiple firings, offset firing, unfireable status, and mission area range; Trajectory and Event Management Unit: Supports gunner position display, firing playback, travel trajectory playback, playback control and trajectory export operations, while tracking firing anomalies, working status anomalies, quality control anomalies and system anomaly events; Quality control management unit: realizes functions such as offset triggering reminder, FO-TB deviation monitoring, abnormal shot reporting, wellhead time management and energy monitoring; The data collected by the data acquisition node also includes: terrain data, rock data, environmental data, and basic parameters of the well shot location and the surrounding preset safety zone. The basic parameters of the well shot include: detonation parameters and mesh parameters.

[0025] The production mode setting unit supports custom configuration of priority rules through TD mode. The priority rule is to schedule priority according to the pre-set urgency of the task. It uses real-time calculation to process task requests and generate a queuing sequence, and executes the activation operation in sequence according to the queuing sequence. It also supports FTS mode application and switching. The fixed time interval mode supports the setting of time interval parameters and uses a priority scheduling mechanism to ensure that critical tasks are executed first. The fixed time slot mode supports the association configuration of time slot number and corresponding activation point ID. It supports the preset of timed independent activation schemes, which do not require centralized control. The start and end times of the time slot and the association of activation equipment can be configured through the parameter interface. After configuration, it will take effect and be executed automatically. The status tracking unit acquires real-time status data of the well drilling equipment through 4G+BeiDou dual-mode communication. It tracks the equipment's online time, activation and completion timestamps, current working mode identifier, connection signal strength between the equipment and multi-hop relay nodes, task progress, remaining battery power, and communication mode type. It also marks abnormal states and triggers alerts in real time. The task progress includes: number of activations completed, number of activations remaining, and current task sequence number. The remaining battery power includes: real-time percentage and low battery warning threshold. Abnormal states include: offline, low battery, and connection interruption. The visualization unit is built on a GIS geographic information system to create a visual operation interface, visually presenting the gunner's real-time location, firing status, and the specific location of the command center; it synchronously displays the status indicators of each firing point, with offset firing indicating the specific offset distance and the reason for non-firing status indicating the reason; it visually presents the boundary of the mission area and marks the wellhead coordinates and danger zone warning range on the map, supporting map zooming, panning, point selection and querying operations; the real-time location visualization of the gunner includes: latitude and longitude coordinates and dynamically updated movement trajectory; the firing status includes: real-time display of firing preparation, firing in progress, firing completed, and firing anomaly status; the mission area boundary supports custom drawing and import; The trajectory and event management unit provides playback pause, fast forward, rewind, and location control functions for the gunner's historical trajectory display, full playback of the firing process, and travel trajectory playback. It also supports exporting trajectory data in a specified format. Simultaneously, it monitors and tracks various abnormal events in real time, including firing anomalies, operational status anomalies, quality control anomalies, and system anomalies, and automatically records anomaly event details for traceability. Specifically, the historical trajectory display reconstructs the movement path along a timeline; the full firing process playback includes synchronous display of firing time, parameters, and status changes; the travel trajectory playback allows setting playback speed and start / end time intervals; quality control anomalies include FO-TB deviation exceeding limits, off-target firing, and insufficient energy; and anomaly event details include timestamps, associated device IDs, parameter data, and anomaly descriptions. The quality control management unit collects key parameter data during the excitation process in real time and provides graded alerts for misaligned excitations; it calculates and monitors the FO-TB deviation value in real time, compares it with the preset deviation threshold, and automatically marks it as an abnormal shot when it exceeds the threshold and triggers the reporting process; it supports manual and automatic reporting of abnormal shots, and the reported content includes complete information such as excitation point ID, line number, point number, abnormality type, relevant parameter data, and timestamp; it synchronizes and manages wellhead time, calibrated based on the PTPv2 enhanced time synchronization protocol to ensure that the wellhead time error is within the preset range; it monitors the maximum energy value at the wellhead in real time and compares it with the minimum wellhead energy threshold to realize energy compliance verification and low energy warning, while recording energy data for subsequent quality control analysis; among them, the graded alerts for misaligned excitations include: setting warning thresholds according to the misalignment distance to trigger mild, moderate, and severe warnings.

[0026] In this embodiment, the well-fired data distribution area is divided into multiple zones, with each zone equipped with multi-hop relay nodes, forming a hierarchical transmission path of "well-fired data acquisition node → relay node → ... → relay node → control center". This solves the problem of limited single-hop communication distance and is particularly suitable for complex terrains such as mountainous areas and underground locations, expanding the coverage area through relay hops.

[0027] For example, in deep mines, single-hop communication may result in signal loss due to rock strata attenuation, while multi-hop relay can ensure data integrity through relay transmission via underground relay stations.

[0028] A dynamic link selection mechanism is established between relay nodes to automatically switch paths based on real-time link quality. This avoids communication interruptions due to node failure or environmental interference, improving system fault tolerance.

[0029] It provides high-speed data transmission, suitable for ground or shallow areas, and supports real-time video monitoring and big data uploading. In remote areas without 4G signal or when 4G network is congested, it transmits critical control commands via satellite communication. It achieves "all-area, all-time" communication coverage, avoiding control failures due to the failure of a single communication mode.

[0030] By using the PTPv2 enhanced protocol, a master-slave clock architecture is employed, combined with hardware timestamps and transparent clock (TC) or boundary clock (BC) devices, to achieve nanosecond-level synchronization accuracy.

[0031] Well-fired blasting requires strict synchronization, which traditional NTP protocols cannot meet. PTPv2, however, ensures simultaneous detonation of multiple boreholes, avoiding interference from vibration superposition. Clock compensation parameters are dynamically adjusted through a link quality assessment function to eliminate the impact of transmission delay on synchronization. In long-distance or multi-hop transmissions, clock drift introduced by relay hops is automatically corrected to maintain global time consistency.

[0032] By modeling with multi-dimensional metrics, a link quality assessment function is constructed, incorporating parameters such as signal-to-noise ratio, packet loss rate, and latency, to quantify the reliability of each link. This allows for real-time identification of weak links, providing data support for route optimization and fault early warning.

[0033] The forwarding priority of relay nodes is dynamically adjusted based on link quality, prioritizing the use of high-quality links to transmit critical data, optimizing resource allocation, and avoiding data loss or delays caused by low-quality links.

[0034] Based on a block size distribution statistical model and combined with basic well shot parameters such as borehole diameter, borehole spacing, and charge quantity, automatic parameter adjustment is achieved by establishing a parameter-block size response relationship. This avoids the lag of manual parameter adjustment; for example, it can automatically increase the charge quantity in hard rock formations to improve the fracturing effect, or decrease the charge quantity in soft rock to control flyrock.

[0035] The combined digital twin integrates dynamic fluid pressure field, rock mass parameter field, and environmental disturbance field to simulate blasting energy distribution. By comparing the actual block size distribution with the simulation results, the parameter model is dynamically corrected. For example, if it is found that the block size in a certain area is too large, the parameters of the surrounding blast holes are automatically adjusted to supplement the energy.

[0036] By using joint modeling of geology, blasting, and production, a digital twin integrates geological structure, excitation parameters, and production disturbances to establish a three-dimensional risk field. Traditional methods only consider single factors, while the joint model can identify complex risks and improve the accuracy of early warning.

[0037] For example, in an underground mine, the model predicts that a blasting vibration may cause the roof to collapse in a certain roadway, and adjusts the layout of the blast holes in advance and evacuates personnel.

[0038] In one embodiment of the present invention, the data acquisition node acquires terrain data, rock data, environmental data and basic parameters of the well shot location and the surrounding preset safety area. The basic parameters of the well shot include: detonation parameters and mesh parameters.

[0039] The data acquisition node is equipped with a six-dimensional force sensor array, a fiber optic pressure monitoring network, and a distributed acoustic sensing system. It also collects environmental data of the well location based on weather parameters provided by the meteorological department, detects the well data, and sends it to the corresponding multi-hop relay nodes in the area.

[0040] In one embodiment of the present invention, the communication link construction module includes: Collect geographical information on the distribution area of ​​well shots, and provide a basis for subsequent area division and relay node deployment based on the distribution density, working mode and data acquisition frequency of well shots; Area division unit: Based on the distribution of well shots and geographical environment, the area is divided into areas so that the number of well shots in the area is within a set threshold. A threshold is set for the total data collection frequency of all well shots in the area. Areas that exceed the threshold are further divided. Node communication configuration unit: Based on the communication protocol and transmission requirements of the well shot data acquisition module, select multi-hop relay node devices that support multiple communication protocols; configure the communication parameters of the well shot data acquisition node, and establish communication between the well shot data acquisition node and the corresponding multi-hop relay node through 4G+BeiDou dual-mode communication.

[0041] Configure the corresponding receiving parameters on the multi-hop relay node to ensure that the data sent by the well shot data acquisition module can be received correctly, and complete the establishment of the communication link between the well shot data acquisition node and the multi-hop relay node; Configure the relay function parameters of multi-hop relay nodes, including relay routing and forwarding rules, to enable them to achieve multi-hop relay transmission and establish communication links between multi-hop relay nodes; Perform communication tests between adjacent multi-hop relay nodes to check whether the links between multi-hop relays are unobstructed; Based on the test results, the parameters of the relay nodes were adjusted and optimized to ensure the stability and reliability of the multi-hop relay links.

[0042] In one embodiment of the present invention, the communication management module includes: The communication link evaluation unit analyzes the link quality of multi-hop relay nodes by establishing a link quality evaluation function for multi-hop relay transmission. Based on the link quality assessment coefficient, by setting a threshold for the link quality assessment coefficient, links that are below the threshold are given an early warning and marked. The link conversion unit formulates link conversion rules, including: after receiving an early warning signal, the upload link of a multi-hop relay node selects the maximum value of the communication link quality assessment coefficient between the multi-hop relay nodes and sends the data to the multi-hop relay node corresponding to the maximum value of the quality assessment coefficient for upload; after receiving an early warning signal, the communication link of the data acquisition module is immediately re-established. The link analysis unit, based on the enhanced time synchronization protocol PTPv2, establishes a clock skew compensation model, performs clock skew compensation, and constructs a link health prediction model.

[0043] In one embodiment of the present invention, a link quality evaluation function is constructed using the following formula: ; Where: LQi is the quality index of the i-th link, RSSIi(t) is the received signal strength (dBm) of the i-th link, RSSImax is the maximum value of the received signal strength in the historical data, BERi(t) is the bit error rate of the i-th link, BERmax is the maximum value of the bit error rate of the link in the historical data, Δti is the link interruption duration (s), α, β and γ are weighting coefficients, and λ is the time attenuation coefficient (in this embodiment, it is set to 0.1-0.5). In one embodiment of the present invention, based on the enhanced time synchronization protocol PTPv2, a clock skew compensation model is established, clock skew compensation is performed, and a link health prediction model is constructed, including the following steps: An enhanced time synchronization protocol based on PTPv2 is used to establish a clock skew compensation model: Where Δtco is the clock offset after compensation, Δtra is the clock offset before compensation, dr is the link transmission distance, k1 is the communication link compensation parameter, k2 is the temperature compensation parameter, vprop is the signal propagation speed of the communication link, and dT / dt is the temperature change rate. Based on the compensated clock deviation, clock deviation compensation is performed; By using historical link signal transmission data, a time window is set, and the link quality index and compensated clock deviation data within the time window are statistically analyzed. For the corresponding operating status of the link in the historical data, the normal operation data and the link failure operation data are labeled. An LSTM model is trained using the labeled data to build a link health prediction model.

[0044] Example 2 Based on the acquisition of blasted block size image data using a multispectral imaging module, block size identification is performed. A block size distribution statistical model and a well shot basic parameter optimization model are established. Automatic optimization of well shot basic parameters is then performed, including the following steps: A multispectral imaging module, composed of near-infrared laser illumination and a high-speed CMOS camera, acquires blast block size image data for block size identification and establishes a block size distribution statistical model. Where P(d) is the block size distribution function, d is the block size, μd is the logarithmic mean of the block size, and σd is the logarithmic standard deviation of the block size; By establishing a multi-objective optimization function, an optimization model for the activation parameters is constructed: Where F is the comprehensive objective function, w is the corresponding weight coefficient, σtarget is the target value of the logarithmic standard deviation of the block size, P(d>dmax) is the probability that the block size distribution exceeds the standard dmax of the large block, Plimit is the maximum limit value of the probability that the block size distribution exceeds the standard dmax of the large block, vpp is the peak particle size velocity, vlimit is the peak particle size velocity limit value, ηenergy is the energy utilization rate, and ηmin is the minimum energy utilization rate. Acquire historical data of terrain, rock, and environmental data and basic parameters of well shot locations from data acquisition nodes; calculate the cosine similarity between historical terrain, rock, and environmental data and current terrain, rock, and environmental data; filter historical data with similarity greater than a threshold; and calculate the comprehensive objective function F value of the filtered historical data based on the corresponding blasting block size image data. Calculate the F-values ​​of each set of parameters and the corresponding basic well shot parameters, construct the Kriging surrogate model, use the NSGA-II algorithm for multi-objective optimization, verify the Pareto front solution set, and obtain the optimized basic well shot parameters; set the initiation parameters and mesh parameters according to the optimized basic well shot parameters.

[0045] The analysis of the hazardous area based on the optimized well shot parameters includes the following steps: Based on the dynamic damage theory of rock mass, a real-time correction formula for blasting safety distance is established: Where Rst is the blasting safety distance, α is the empirical coefficient, β is the damage accumulation coefficient, D is the rock mass damage variable, τ is the time variable, γ is the pressure correction coefficient, p(t) is the real-time pressure in the wellbore, and pcrit is the critical pressure. Within the safe zone, a danger zone Ωdanger={(x,y,z)| is defined centered on the well site. ≤Rst, and z≥Hm}, where (x,y,z) are the coordinates in a three-dimensional coordinate system established with the well shot position as the origin, and Hm is the highest position of the blasted rock block.

[0046] Example 3 In one embodiment of the present invention, based on the acquisition of blast block size image data by a multispectral imaging module for block size identification, a block size distribution statistical model and a well shot basic parameter optimization model are established, and the well shot basic parameters are automatically optimized, including the following steps: A multispectral imaging module, composed of near-infrared laser illumination and a high-speed CMOS camera, acquires blast block size image data for block size identification and establishes a block size distribution statistical model. Where P(d) is the block size distribution function, d is the block size, μd is the logarithmic mean of the block size, and σd is the logarithmic standard deviation of the block size; By establishing a multi-objective optimization function, an activation parameter optimization model is constructed, and multiple parameters are optimized simultaneously: Where F is the comprehensive objective function, w1, w2 and w3 are the corresponding weight coefficients, σtarget is the target value of the logarithmic standard deviation of the block size, Qmax is the maximum value of the explosive charge, Q is the explosive charge, C is the blasting cost; P(d>dmax) is the probability that the block size distribution exceeds the standard dmax of the large block, Plimit is the maximum limit value of the probability that the block size distribution exceeds the standard dmax of the large block, vpp is the peak particle size velocity (m / s), vlimit is the peak particle size velocity limit value, ηenergy is the energy utilization rate, and ηmin is the minimum energy utilization rate. Acquire historical data of terrain, rock, and environmental data and basic parameters of well shot locations from data acquisition nodes; calculate the cosine similarity between historical terrain, rock, and environmental data and current terrain, rock, and environmental data; filter historical data with similarity greater than a threshold; and calculate the comprehensive objective function F value of the filtered historical data based on the corresponding blasting block size image data. Calculate the F-values ​​of each set of parameters and the corresponding basic well-shot parameters, construct the Kriging surrogate model, use the NSGA-II algorithm for multi-objective optimization, verify the Pareto front solution set, and obtain the optimized basic well-shot parameters; set the charge parameters, initiation parameters, and mesh parameters based on the optimized basic well-shot parameters.

[0047] In one embodiment of the present invention, the dangerous area is analyzed based on the optimized well shot basic parameters, including the following steps: Based on the dynamic damage theory of rock mass, a real-time correction formula for blasting safety distance is established: Where Rst is the blasting safety distance in meters (m), Q is the charge amount in kilograms (kg), M is the equivalent mass density of the rock mass in kg / m³, α is an empirical coefficient (set to 1.2-1.8 in this embodiment), β is the damage accumulation coefficient (set to 0.05-0.15 in this embodiment), D is the rock mass damage variable (range between 0 and 1, set to 0.3 in this embodiment), τ is the time variable, γ is the pressure correction coefficient (set to 0.2-0.5 in this embodiment), p(t) is the real-time pressure in the wellbore in MPa, and pcrit is the critical pressure in MPa. Within the safe zone, a danger zone Ωdanger={(x,y,z)| is defined centered on the well site. ≤Rst, and z≥Hm}, where (x,y,z) are the coordinates in a three-dimensional coordinate system established with the well shot position as the origin, and Hm is the highest position of the blasted rock block.

[0048] In one embodiment of the present invention, a combined digital twin of geology, blasting, and production in a hazardous area is established based on a dynamic fluid pressure field, a rock mass parameter field, and an environmental disturbance field, including the following steps: Based on the topographic data of the dangerous area where the well shot is located, dynamic fluid pressure field data is extracted, including rock permeability and groundwater level changes; dynamic rock mass parameter field data is extracted from rock data, including rock mass elastic modulus, rock strata interface and fault data; dynamic environmental disturbance field data is extracted from environmental data, including air temperature, air pressure and wind speed. Based on the extracted data, a joint digital twin of the geology, blasting, and production of the hazardous area is established, including: Geological Model: Use Gocad or Surpac to build a 3D geological model. The input is the rock mass elastic modulus, rock layer interface and fault data, and the output is the location of rock mass fracture. Using the locations of rock mass fractures in historical data as labels, the model is trained using labeled data to predict the locations of rock mass fractures. Blasting Model: A blasting dynamics model is established using LS-DYNA or AUTODYN. The inputs are dynamic fluid pressure field data, dynamic rock mass parameter field data, dynamic environmental disturbance field data, and basic well shot parameters. Using measured blasting vibration data from historical data as labels, the model is trained by adding labels to dynamic fluid pressure field data, dynamic rock mass parameter field data, and dynamic environmental disturbance field data, so that the model can predict blasting vibration data. Production Model: The MLP multilayer perceptron is used to establish a mining operation model. The inputs are the blasting volume of the well and the basic parameters of the well. Using the measured blasting efficiency from historical data as labels, an MLP model is trained by adding labels to the well-blasting volume and basic well-blasting parameters, so that the model can predict the blasting efficiency.

[0049] In one embodiment of the present invention, based on the link health prediction model and the joint digital twin of geology, blasting and production in the dangerous area, an intelligent activation model is constructed to perform intelligent activation of well shots, including the following steps: Based on the rock mass fracture location, blasting vibration data and blasting efficiency predicted by the joint digital twin, a 3D convolutional neural network is used to extract spatiotemporal feature vectors from the predicted data. Based on the link health prediction model, all links are predicted to be in normal operation and in failure operation. If a link is in normal operation, an association edge is established between nodes. If a link is in failure operation, the association edge between nodes is disconnected. The GNN graph neural network is used to model the dependencies between nodes on the link topology data, including degree centrality and betweenness centrality. The spatiotemporal feature vectors extracted by the 3D convolutional neural network and the inter-node dependencies of the output of the GNN graph neural network are fused through an attention mechanism to generate a fused feature vector. By sharing the feature extraction networks of the underlying 3D convolutional neural network and the GNN graph neural network, prediction results for different targets are output separately, thus constructing an intelligent activation model: The output target is the allocation of explosive charge to each borehole: By analyzing the fusion feature vector and rock fracture intensity of historical data, and labeling the fusion feature vector and rock fracture intensity of each blast hole based on the explosive charge data of historical data, a regression task network is trained to analyze the explosive charge of each blast hole. The output target is generated according to the firing sequence of the boreholes: By analyzing the fused feature vectors of historical data and link failure operation data, and labeling the fused feature vectors and link failure operation data according to the firing order of the historical data, a Transformer decoder is trained to perform sequence generation tasks and analyze the firing order of the boreholes. The output target is the risk classification of vibration exceeding the standard: By analyzing the fusion feature vectors of historical data and blasting vibration data, the fusion feature vectors and blasting vibration data are labeled according to whether the blast holes have the risk of vibration exceeding the standard. A random forest network is trained to complete the binary classification task and analyze whether the blast holes have the risk of vibration exceeding the standard. Intelligent ignition of well shots is performed based on the results output by the intelligent ignition model.

[0050] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A real-time control and management platform for intelligent firing of well shots, characterized in that, include: Communication link construction module: For the distribution of well shots in the region, the distribution area of ​​well shots is divided into areas, and multi-hop relay nodes are set in each area. The well shot data acquisition nodes in the area establish communication links with the corresponding multi-hop relay nodes, and communication links are established between the multi-hop relay nodes. Among them, the communication links are established through 4G+BeiDou dual-mode communication. Communication Management Module: By establishing a link quality assessment function for multi-hop relay transmission, the module analyzes the link quality of multi-hop relay nodes, performs clock skew compensation based on the enhanced time synchronization protocol of PTPv2, and constructs a link health prediction model. Production mode configuration and status full-dimensional management module: used to realize flexible configuration of production mode, status tracking and visualization of well drilling equipment and operation process, operation trajectory playback and event tracing, as well as quality control management of the excitation process, providing comprehensive data support and visualization management means for intelligent excitation decision-making and operation process control; Intelligent parameter optimization module: Based on the multispectral imaging module, it collects blasting block size image data to identify block size, establishes a block size distribution statistical model and a well shot basic parameter optimization model, and automatically optimizes the well shot basic parameters. Digital twin construction module: Based on the optimized well shot basic parameters, the dangerous area is analyzed, and a combined digital twin of geology, blasting and production of the dangerous area is established based on the dynamic fluid pressure field, rock mass parameter field and environmental interference field; Intelligent Activation Module: Based on the link health prediction model and the joint digital twin of geology, blasting and production in the dangerous area, an intelligent activation model is constructed to carry out intelligent activation of well blasting.

2. The intelligent real-time control and management platform for well firing according to claim 1, characterized in that, The data collected by the data acquisition node and sent to the production mode configuration and status full-dimensional management module via the communication management module includes: Basic data of well-fired equipment: trigger point ID, detonator model, detonator subplate type, electronic detonator manufacturer code, electronic detonator delay parameters, charging preparation time, report timeout time, and vibration alignment parameters; Core production operation data: production mode identifier, time slot sequence number, line number, point number, FO time, TB time, TB deviation value, GPS timestamp, firing waiting time, number of firings, number of design combinations; location and environmental data: wellhead coordinates, real-time position of the gunner, position of the command center, boundary data of the mission area, offset distance; among which, the production mode identifier includes: TD mode, fixed time interval mode, and fixed time slot mode. Status and quality control data: excitation point status identifier, equipment online time, excitation time, working mode, connection status, task progress, equipment power, communication mode, maximum wellhead energy value, abandoned blasting identifier, and abnormal event type; among which, the excitation point status identifier includes: not excitation, excitation completed, multiple excitations, offset excitation, and cannot be excitation, and the abnormal event type includes: excitation abnormality, working status abnormality, quality control abnormality, and system abnormality; Personnel-related data: the gunner's ID number; The production mode configuration and status full-dimensional management module includes: Production mode setting unit: Supports configuration of TD mode, fixed time interval mode and fixed time slot mode. Among them, TD mode adopts priority scheduling, real-time operation and queuing activation mechanism. Fixed time slot mode supports timed independent activation and does not require central control, and parameter setting can be completed easily. Status tracking unit: Real-time tracking of well-fired equipment's online time, firing time, working mode, connection status, task progress, equipment power level, and communication mode; Visualization unit: Visually presents the gunner's position, firing status, and command center position, and simultaneously displays the firing point's unfired, firing completed, multiple firings, offset firing, unfireable status, and mission area range; Trajectory and Event Management Unit: Supports gunner position display, firing playback, travel trajectory playback, playback control and trajectory export operations, while tracking firing anomalies, working status anomalies, quality control anomalies and system anomaly events; Quality control management unit: realizes functions such as offset triggering reminder, FO-TB deviation monitoring, abnormal shot reporting, wellhead time management and energy monitoring; The data collected by the data acquisition node also includes: terrain data, rock data, environmental data, and basic parameters of the well shot location and the surrounding preset safety area. The basic parameters of the well shot include: detonation parameters and mesh parameters.

3. The intelligent real-time control and management platform for well firing according to claim 1, characterized in that, The production mode setting unit supports custom configuration of priority rules through TD mode. The priority rule is to schedule priority according to the pre-set urgency of the tasks. It uses real-time computing to process task requests and generate a queuing sequence, and executes the activation operation in sequence according to the queuing sequence. It also supports FTS mode application and switching. The fixed time interval mode supports setting time interval parameters and uses a priority scheduling mechanism to ensure that critical tasks are executed first. The fixed time slot mode supports the association configuration of time slot number and corresponding activation point ID. It supports the preset of timed independent activation schemes, which do not require centralized control. The start and end times of the time slot and the association of activation equipment can be configured through the parameter interface. After configuration, it will take effect and be executed automatically. The status tracking unit acquires real-time status data of the well-blasting equipment based on 4G+BeiDou dual-mode communication. It tracks the equipment's online time, activation and completion timestamps, current working mode identifier, connection signal strength between the equipment and multi-hop relay nodes, task progress, remaining battery power, and communication mode type. It also marks abnormal states and triggers alerts in real time. The task progress includes: number of activations completed, number of activations remaining, and current task sequence number. The remaining battery power includes: real-time percentage and low battery warning threshold. Abnormal states include: offline, low battery, and connection interruption. The visualization unit is based on a GIS geographic information system to construct a visual operation interface, visually presenting the gunner's real-time location, firing status, and the specific location of the command center; it synchronously displays the status indicators of each firing point, with offset firing indicating the specific offset distance and the reason for non-firing status indicating the reason; it visually presents the boundary of the mission area and marks the wellhead coordinates and danger zone warning range on the map, supporting map zooming, panning, point selection and querying operations; the visual presentation of the gunner's real-time location includes: latitude and longitude coordinates and dynamically updated movement trajectory; the firing status includes: real-time display of firing preparation, firing in progress, firing completed, and firing abnormal status; the mission area boundary supports custom drawing and import; The trajectory and event management unit provides functions such as pause, fast forward, rewind, and positioning to specific time points for the playback of the gunner's historical trajectory, the full playback of the firing process, and the playback of the travel trajectory. It also supports the export of trajectory data in a specified format. Simultaneously, it monitors and tracks various abnormal events in real time, including firing anomalies, operational status anomalies, quality control anomalies, and system anomalies, and automatically records details of these anomalies for traceability. Specifically, the historical trajectory playback restores the movement path along a timeline; the full firing process playback includes synchronous display of firing time, parameters, and status changes; the travel trajectory playback allows setting playback speed and start / end time intervals; quality control anomalies include FO-TB deviation exceeding limits, off-target firing, and insufficient energy; and anomaly event details include timestamps, associated device IDs, parameter data, and anomaly descriptions. The quality control management unit collects key parameter data during the excitation process in real time and provides graded alerts for misaligned excitations; it calculates and monitors the FO-TB deviation value in real time, compares it with a preset deviation threshold, and automatically marks it as an abnormal shot when it exceeds the threshold and triggers the reporting process; it supports manual and automatic reporting of abnormal shots, and the reported content includes complete information such as excitation point ID, line number, point number, abnormality type, relevant parameter data, and timestamp; it synchronizes and manages wellhead time, calibrating based on the PTPv2 enhanced time synchronization protocol to ensure that the wellhead time error is within a preset range; it monitors the maximum energy value at the wellhead in real time and compares it with the minimum wellhead energy threshold to achieve energy compliance verification and low energy warning, while recording energy data for subsequent quality control analysis; the graded alerts for misaligned excitations include: setting warning thresholds according to the misalignment distance to trigger mild, moderate, and severe warnings.

4. The intelligent real-time control and management platform for well firing according to claim 1, characterized in that, The communication link construction module includes: Area division unit: Based on the distribution of well shots and geographical environment, the area is divided into areas so that the number of well shots in the area is within a set threshold. A threshold is set for the total data collection frequency of all well shots in the area. Areas that exceed the threshold are further divided. Node communication configuration unit: Based on the communication protocol and transmission requirements of the well shot data acquisition module, select multi-hop relay node devices that support multiple communication protocols; configure the communication parameters of the well shot data acquisition node, and establish communication between the well shot data acquisition node and the corresponding multi-hop relay node through 4G+BeiDou dual-mode communication.

5. The intelligent real-time control and management platform for well firing according to claim 1, characterized in that, The communication management module includes: The communication link evaluation unit analyzes the link quality of multi-hop relay nodes by establishing a link quality evaluation function for multi-hop relay transmission. Based on the link quality assessment coefficient, by setting a threshold for the link quality assessment coefficient, links that are below the threshold are given an early warning and marked. The link conversion unit formulates link conversion rules, including: after receiving an early warning signal, the upload link of a multi-hop relay node selects the maximum value of the communication link quality assessment coefficient between the multi-hop relay nodes and sends the data to the multi-hop relay node corresponding to the maximum value of the quality assessment coefficient for upload; after receiving an early warning signal, the communication link of the data acquisition module is immediately re-established. The link analysis unit, based on the enhanced time synchronization protocol PTPv2, establishes a clock skew compensation model, performs clock skew compensation, and constructs a link health prediction model. The link quality evaluation function is as follows: ； Where: LQi is the quality index of the i-th link, RSSIi(t) is the received signal strength (dBm) of the i-th link, RSSImax is the maximum value of the received signal strength in the historical data, BERi(t) is the bit error rate of the i-th link, BERmax is the maximum value of the link bit error rate in the historical data, Δti is the link interruption duration (s), α, β and γ are weighting coefficients, and λ is the time attenuation coefficient.

6. The intelligent real-time control and management platform for well firing according to claim 1, characterized in that, An enhanced time synchronization protocol based on PTPv2 is used to establish a clock skew compensation model, perform clock skew compensation, and construct a link health prediction model, including the following steps: An enhanced time synchronization protocol based on PTPv2 is used to establish a clock skew compensation model: Where Δtco is the clock offset after compensation, LQi is the quality index of the i-th link, Δtra is the clock offset before compensation, dr is the link transmission distance, k1 is the communication link compensation parameter, k2 is the temperature compensation parameter, vprop is the signal propagation speed of the communication link, and dT / dt is the temperature change rate. Based on the compensated clock deviation, clock deviation compensation is performed; By using historical link signal transmission data, a time window is set, and the link quality index and compensated clock deviation data within the time window are statistically analyzed. For the corresponding operating status of the link in the historical data, the normal operation data and the link failure operation data are labeled. An LSTM model is trained using the labeled data to build a link health prediction model.

7. The intelligent real-time control and management platform for well firing according to claim 1, characterized in that, Based on the acquisition of blasted block size image data using a multispectral imaging module, block size identification is performed. A block size distribution statistical model and a well shot basic parameter optimization model are established. Automatic optimization of well shot basic parameters is then performed, including the following steps: A multispectral imaging module, composed of near-infrared laser illumination and a high-speed CMOS camera, acquires blast block size image data for block size identification and establishes a block size distribution statistical model. Where P(d) is the block size distribution function, d is the block size, μd is the logarithmic mean of the block size, and σd is the logarithmic standard deviation of the block size; By establishing a multi-objective optimization function, an optimization model for the activation parameters is constructed: Where F is the comprehensive objective function w is the corresponding weighting coefficient, σtarget is the target value of the logarithmic standard deviation of the block size, P(d>dmax) is the probability that the block size distribution exceeds the standard dmax of the large block, Plimit is the maximum limit value of the probability that the block size distribution exceeds the standard dmax of the large block, vpp is the peak particle size velocity, vlimit is the peak particle size velocity limit value, ηenergy is the energy utilization rate, and ηmin is the minimum energy utilization rate. Acquire historical data of terrain, rock, and environmental data and basic parameters of well shot locations from data acquisition nodes; calculate the cosine similarity between historical terrain, rock, and environmental data and current terrain, rock, and environmental data; filter historical data with similarity greater than a threshold; and calculate the comprehensive objective function F value of the filtered historical data based on the corresponding blasting block size image data. Calculate the F-values ​​of each set of parameters and the corresponding basic well shot parameters, construct the Kriging surrogate model, use the NSGA-II algorithm for multi-objective optimization, verify the Pareto front solution set, and obtain the optimized basic well shot parameters; set the initiation parameters and mesh parameters according to the optimized basic well shot parameters.

8. The intelligent real-time control and management platform for well firing according to claim 1, characterized in that, The analysis of the hazardous area based on the optimized well shot parameters includes the following steps: Based on the dynamic damage theory of rock mass, a real-time correction formula for blasting safety distance is established: Where Rst is the blasting safety distance, α is the empirical coefficient, β is the damage accumulation coefficient, D is the rock mass damage variable, τ is the time variable, γ is the pressure correction coefficient, p(t) is the real-time pressure in the wellbore, and pcrit is the critical pressure. Within the safe zone, a danger zone Ωdanger={(x,y,z)| is defined centered on the well site. ≤Rst, and z≥Hm}, where (x,y,z) are the coordinates in a three-dimensional coordinate system established with the well shot position as the origin, and Hm is the highest position of the blasted rock block.

9. The intelligent real-time control and management platform for well firing according to claim 1, characterized in that, Based on the dynamic fluid pressure field, rock mass parameter field, and environmental disturbance field, a joint digital twin of geology, blasting, and production in the hazardous area is established, including the following steps: Based on the topographic data of the dangerous area where the well shot is located, dynamic fluid pressure field data is extracted, including rock permeability and groundwater level changes; dynamic rock mass parameter field data is extracted from rock data, including rock mass elastic modulus, rock strata interface and fault data; dynamic environmental disturbance field data is extracted from environmental data, including air temperature, air pressure and wind speed. Based on the extracted data, a joint digital twin of the geology, blasting, and production of the hazardous area is established, including: Geological Model: Use Gocad or Surpac to build a 3D geological model. The input is the rock mass elastic modulus, rock layer interface and fault data, and the output is the location of rock mass fracture. Using the rock mass fracture locations in historical data as labels, the model is trained by adding labels to the rock mass elastic modulus, rock layer interfaces, and fault data, so that the model can predict the rock mass fracture locations. Blasting Model: A blasting dynamics model is established using LS-DYNA or AUTODYN. The inputs are dynamic fluid pressure field data, dynamic rock mass parameter field data, dynamic environmental disturbance field data, and basic well shot parameters. Using measured blasting vibration data from historical data as labels, the model is trained by adding labels to dynamic fluid pressure field data, dynamic rock mass parameter field data, and dynamic environmental disturbance field data, so that the model can predict blasting vibration data. Production Model: The MLP (Multilayer Perceptron) is used to establish a mining operation model. The inputs are the blasting volume and basic parameters of the blasting, and the output is the blasting efficiency. Using the measured blasting efficiency from historical data as labels, the MLP model is trained with the blasting volume and basic parameters of the well and shot after adding labels, so that the model can predict the blasting efficiency.

10. The intelligent real-time control and management platform for well firing according to claim 1, characterized in that, Based on the link health prediction model and the joint digital twin of geology, blasting and production in the hazardous area, an intelligent activation model is constructed to carry out intelligent activation of well blasting, including the following steps: Based on the rock mass fracture location, blasting vibration data and blasting efficiency predicted by the joint digital twin, a 3D convolutional neural network is used to extract spatiotemporal feature vectors from the predicted data. Based on the link health prediction model, all links are predicted to be in normal operation and in failure operation. If a link is in normal operation, an association edge is established between nodes. If a link is in failure operation, the association edge between nodes is disconnected. The GNN graph neural network is used to model the dependencies between nodes on the link topology data, including degree centrality and betweenness centrality. The spatiotemporal feature vectors extracted by the 3D convolutional neural network and the inter-node dependencies of the output of the GNN graph neural network are fused through an attention mechanism to generate a fused feature vector. By sharing the feature extraction networks of the underlying 3D convolutional neural network and the GNN graph neural network, prediction results for different targets are output separately, thus constructing an intelligent activation model: The output target is the allocation of explosive charge to each borehole: By analyzing the fusion feature vector and rock fracture intensity of historical data, and labeling the fusion feature vector and rock fracture intensity of each blast hole based on the explosive charge data of historical data, a regression task network is trained to analyze the explosive charge of each blast hole. The output target is generated according to the firing sequence of the boreholes: By analyzing the fused feature vectors of historical data and link failure operation data, and labeling the fused feature vectors and link failure operation data according to the firing order of the historical data, a Transformer decoder is trained to perform sequence generation tasks and analyze the firing order of the boreholes. The output target is the risk classification of vibration exceeding the standard: By analyzing the fusion feature vectors of historical data and blasting vibration data, the fusion feature vectors and blasting vibration data are labeled according to whether the blast holes have the risk of vibration exceeding the standard. A random forest network is trained to complete the binary classification task and analyze whether the blast holes have the risk of vibration exceeding the standard. Intelligent ignition of well shots is performed based on the results output by the intelligent ignition model.

Images