Digital twin dynamic control method for coal breaking operation of steep seam drilling

Abstract

The present application relates to a digital twin dynamic control method for coal breaking operation of steeply inclined coal seam drilling, belonging to the technical field of drilling blasting control. The method comprises: obtaining a standardized feature vector after the last drilling blasting of the target steeply inclined coal seam; constructing a digital twin model, inputting the standardized feature vector into the digital twin model, and outputting a predicted drilling state vector; generating a vector deviation between the predicted drilling state and the actual drilling state, and generating a drilling blasting control mode; when the drilling blasting control mode is a parameter reconstruction mode, generating an optimal blasting parameter group; the optimal blasting parameter group is sent to the drilling blasting equipment on site of the target steeply inclined coal seam and executed, and the feedback optimization drilling particle swarm optimization algorithm is obtained. The present application realizes joint prediction based on the digital twin model and the drilling particle swarm optimization algorithm, forms a complete dynamic management and control closed loop, and improves the dynamic control efficiency of the coal breaking operation of the steeply inclined coal seam drilling.

Description

Technical Field

[0001] This invention relates to the field of borehole blasting control technology, and in particular to a digital twin dynamic control method for borehole coal breaking operations in steeply inclined coal seams. Background Technology

[0002] With the continuous improvement of intelligent mining levels in coal mines, drilling operations, fracturing operations, and safety monitoring increasingly rely on digital, automated, and intelligent technologies. Low-level parameter recording and single-machine monitoring are no longer sufficient to meet the operational needs under complex coal seam conditions, especially in steeply dipping coal seams. Due to the combined effects of large dip angles and complex overburden structures, the stress state of the coal body, the evolution of fractures, and the evolution of operational risks all exhibit greater uncertainty. Public research indicates that steeply dipping coal seams are more prone to phenomena such as large-scale roof overhangs, strong dynamic load disturbances, uneven fracture propagation, and coal body fragmentation. Therefore, the same drilling scheme or fracturing parameters often produce significantly different operational results in different areas, making traditional experience-based and static design methods insufficient to meet the requirements of dynamic on-site management.

[0003] Currently, the technologies for controlling borehole blasting mainly fall into three categories: The first category is digital monitoring and remote management technology for drilling rigs, which mainly realizes parameter acquisition, status monitoring, alarm and remote operation and maintenance, but lacks a dedicated digital twin closed loop for the drilling and coal breaking process of steeply inclined coal seams; The second category is automated directional drilling rigs, drilling robots and coiled tubing intelligent drilling technology, which focuses on hole location design, trajectory control and automated equipment execution, but has not yet incorporated the geomechanical constraints, fracture evolution, block size changes and ejection risks under the conditions of steeply inclined coal seams into a unified closed-loop control; The third category is general mine digital twin platform technology, which has the ability to integrate data, 3D modeling and visualization management, but mostly stays at the general platform level and lacks a special dynamic control scheme for drilling and coal breaking operation units in steeply inclined coal seams. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams. The technical solution of this invention is as follows: A digital twin dynamic control method for borehole breaking operations in steeply inclined coal seams includes: S1, collect multiple raw parameter data of preset drilling and blasting parameter types after the last drilling and blasting of the target steeply inclined coal seam, and unify the multiple raw parameter data in time and space to generate a standardized feature vector; S2, construct a digital twin model consisting of multiple twin sub-models, input standardized feature vectors into the digital twin model, and output a predicted borehole state vector from the digital twin model; S3, generate the vector deviation between the predicted borehole state and the actual borehole state based on the predicted borehole state vector and the standardized feature vector, and generate the borehole blasting control mode based on the vector deviation. S4, When the borehole blasting control mode is parameter reconstruction mode, a borehole particle swarm optimization algorithm is generated based on vector deviation and multiple geographical parameters of the target steeply inclined coal seam. The optimal blasting parameter set is generated based on the borehole particle swarm optimization algorithm and the current engineering priority of the target steeply inclined coal seam. S5: The optimal blasting parameter set is sent to the drilling and blasting equipment at the target steeply inclined coal seam site and the drilling and blasting is executed. Based on the feedback of various original parameter data of the preset drilling and blasting parameter type after the drilling and blasting at the target steeply inclined coal seam site, the drilling particle swarm optimization algorithm is optimized, and the optimized drilling particle swarm optimization algorithm is used for the next drilling and blasting.

[0005] Preferably, step S1, when performing spatiotemporal unification of multiple original parameter data and generating standardized feature vectors, includes: S11. For borehole coordinate data and laser scanning point cloud data in various original parameter data, according to the predetermined coordinate system transformation matrix, the borehole coordinate data and laser scanning point cloud data are transformed from the equipment coordinate system to the mine coordinate system to obtain the initial borehole coordinate sequence and initial point cloud data. S12, For geological parameter type data, blasting parameter type data and feedback monitoring parameter type data in various original parameter data, spatial binding is achieved by associating with corresponding spatial locations, and they are mapped to the mine coordinate system to obtain initial geological parameter type data, initial blasting parameter type data and initial feedback monitoring parameter type data; S13, by compensating for the time offset between different acquisition devices through the acquisition device timestamp and synchronization function, the initial borehole coordinate sequence, initial point cloud data, initial geological parameter type data, initial blasting parameter type data and initial feedback monitoring parameter type data are converted into a unified time tag to obtain standard borehole coordinate sequence, standard point cloud data, standard geological parameter type data, standard blasting parameter type data and standard feedback monitoring parameter type data; S14 is a standardized feature vector composed of standard borehole coordinate sequences, standard point cloud data, standard geological parameter type data, standard blasting parameter type data, and standard feedback monitoring parameter type data.

[0006] Preferably, S2 includes: S21, construct a three-dimensional geological structure model, a borehole state vector model, a hybrid prediction model of blasting coal breaking dynamics, an ejection and transport model, and a risk state model, and form a digital twin model from these twin sub-models; S22, the standardized feature vector is input into the digital twin model, and each twin sub-model in the digital twin model outputs its corresponding sub-state. The sub-states output by all twin sub-models constitute the predicted borehole state vector.

[0007] Preferably, S3 includes: S31, Generate the actual borehole state vector based on the standardized feature vector; S32, obtain the Euclidean distance between the predicted borehole state vector and the actual borehole state vector, as the vector deviation between the predicted borehole state and the actual borehole state; S33: Construct multiple deviation intervals based on preset three-level deviation thresholds, determine the deviation interval where the vector deviation is located, and generate a drilling blasting control mode based on the deviation interval where the vector deviation is located.

[0008] Preferably, when the borehole blasting control mode is the parameter reconstruction mode, step S4 includes: S41, Obtain the actual borehole state vector, and construct the initial comprehensive objective function based on the predicted borehole state vector and the actual borehole state vector; S42, adjust the preset weight coefficient group in the initial comprehensive objective function according to the vector deviation to obtain the optimized weight reorganization; S43 generates specific constraints based on various geographical parameters of the target steeply dipping coal seam; S44, replace the preset weight coefficient group in the initial comprehensive objective function with optimization weight reorganization, and generate the borehole particle swarm optimization algorithm in combination with specific constraints; S45, generate the optimal blasting parameter set based on the borehole particle swarm optimization algorithm and the current engineering priority of the target steeply inclined coal seam.

[0009] Preferably, S41 includes: S411, obtain the area of ​​the coal body breaking influence region in the predicted borehole state vector, calculate the difference between the coal body breaking influence region area and the target coal breaking region area to obtain the coal breaking difference value, and calculate the ratio of the coal breaking difference value to the target coal breaking region area as the breaking coverage deviation. S412, obtain the actual block size standard deviation and the actual block size mean in the actual borehole state vector, and calculate the ratio of the actual block size standard deviation to the actual block size mean as the block size non-uniformity. S413, obtain the standard throw range data in the standardized feature vector, calculate the throw area where the standard throw range data exceeds the preset allowable throw area, calculate the difference between the area of ​​the preset allowable throw area and the throw area as the external area, and calculate the ratio of the external area to the preset allowable throw area as the throw over-limit rate. S414. Construct an initial comprehensive objective function based on the preset weight coefficient group, broken coverage deviation, block size unevenness, and over-limit ejection rate.

[0010] Preferably, S42 includes: S421, Based on the relationship between the adjustment coefficient and the vector deviation obtained from pre-training, determine the adjustment coefficient corresponding to each weight coefficient in the preset weight coefficient group in the initial comprehensive objective function; S422, adjust each weight coefficient according to the adjustment coefficient corresponding to each weight coefficient to obtain multiple optimized weights; S423 is an optimization weight reorganization consisting of all optimization weights.

[0011] Preferably, S45 includes: S451, randomly generate multiple initial blasting parameter sets in the parameter feasible region, and configure random particle swarm velocities for each initial blasting parameter set to obtain multiple blasting parameter sets; S452, input each blasting parameter group into the digital twin model to obtain a single-group predicted borehole state vector for each blasting parameter group, and calculate the corresponding fragmentation coverage deviation, block size inhomogeneity and ejection over-limit rate for each blasting parameter group based on the actual borehole state vector and the single-group predicted borehole state vector for each blasting parameter group. S453 uses a Pareto dominance strategy to select multiple non-dominated solutions from multiple blasting parameter sets based on the particle swarm velocity, fragmentation coverage deviation, block size inhomogeneity, and ejection over-limit rate corresponding to each blasting parameter set, and configures all non-dominated solutions in an external archive set. S454: Select a non-dominated solution from the external archive set as the global optimal solution; S455, update the particle swarm velocity of all blasting parameter groups according to the borehole particle swarm optimization algorithm and the global optimal solution; S456, repeat S52-S55 until the maximum number of iterations is reached and then stop iterating; S457, when stopping iteration, select an optimal solution from multiple non-dominated solutions stored in the external archive set as the optimal blasting parameter set based on the current engineering priority of the target steeply inclined coal seam.

[0012] Preferably, S5 includes: S51, the optimal blasting parameter set is sent to the drilling and blasting equipment at the target steeply inclined coal seam and the drilling and blasting is carried out; S52, collect various raw parameter data of the preset borehole blasting parameter type after this borehole blasting in the target steeply inclined coal seam; S53, construct a monitoring parameter data vector based on multiple raw parameter data, and calculate the vector deviation between the monitoring parameter data vector and the predicted borehole state vector as the comprehensive deviation; S54. If the overall deviation is greater than the preset deviation threshold, the model parameters of the digital twin model are corrected according to various original parameter data, and the drilling particle swarm optimization algorithm is re-optimized. The optimized drilling particle swarm optimization algorithm is then used for the next drilling blast.

[0013] Preferably, when S5 sends the optimal blasting parameter set to the drilling blasting equipment at the target steeply inclined coal seam and performs drilling blasting, it further includes: S55, Construct a trajectory deviation correction branch, including: During the process of sending the optimal blasting parameter set to the drilling blasting equipment at the target steeply inclined coal seam site and executing drilling blasting, monitor the drilling trajectory in real time, and generate a correction command when the measured trajectory deviates from the design trajectory and the degree of deviation exceeds the preset deviation threshold. S56, Construct a compensation branch for insufficient coal coverage and unqualified block size in blasting, including: obtaining the block size distribution of coal body from various original parameter data; if the large block rate in the block size distribution exceeds the standard or the coverage area does not meet the target, then generate a compensation blasting command; S57, construct the throwing out of the boundary and vibration exceeding the limit branch, including obtaining the residue throwing range and blasting vibration velocity from various raw parameter data. When the residue throwing range exceeds the allowable area or the blasting vibration velocity is greater than the vibration threshold, a remote limit command is generated. S58 configures the trajectory deviation correction branch, the blasting coal coverage insufficient and block size unqualified compensation branch, and the throwing over the boundary and vibration over the limit branch into the drilling blasting equipment at the target steeply inclined coal seam site. Among them, the trajectory deviation correction branch, the blasting coal coverage insufficient and block size unqualified compensation branch, and the throwing over the boundary and vibration over the limit branch can be triggered independently or in combination. Moreover, the deviation correction command, the compensation blasting command, and the remote limit command are all issued after being verified by the digital twin model.

[0014] All of the above-mentioned optional technical solutions can be combined arbitrarily, and the present invention will not provide a detailed description of the structure after each combination.

[0015] By means of the above solution, the beneficial effects of the present invention are as follows: Compared to traditional borehole blasting control technologies, this invention goes beyond general remote monitoring, data display, or equipment automation. Instead, it specifically targets borehole blasting scenarios in steeply inclined coal seams, establishing a digital twin model that maps multiple state types. Based on the predicted borehole state vector output by the digital twin model, it achieves joint prediction (integrating geomechanical constraints, fracture evolution, block size changes, and ejection risks under steeply inclined coal seam conditions into a closed-loop control). Furthermore, it uses various original parameter data from the field to perform online corrections on the digital twin model and the borehole particle swarm optimization algorithm, forming a complete dynamic control closed loop.

[0016] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, the preferred embodiments of the present invention are described in detail below with reference to the accompanying drawings. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams provided by the present invention. Detailed Implementation

[0018] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0019] The digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams provided in this invention can be executed by any computer, mobile terminal, or server with computing capabilities. For example... Figure 1 As shown in the embodiment of the present invention, the digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams includes: S1 collects various original parameter data of preset drilling and blasting parameter types after the previous drilling and blasting of the target steeply inclined coal seam, and performs spatiotemporal unification of various original parameter data to generate standardized feature vectors.

[0020] Specifically, the preset borehole blasting parameter types include borehole trajectory type, geological parameter type, blasting parameter type, and feedback monitoring parameter type. Among these, the borehole trajectory type data includes borehole coordinate data (acquired in real-time via a measurement-while-drilling system) and laser scanning point cloud data. Geological parameter type data includes coal seam dip angle data, thickness data, uniaxial compressive strength data, and joint and fracture parameter data (geological parameter type data originates from borehole photography and laboratory tests). Blasting parameter type data includes charge amount data, borehole spacing data, row spacing data, minimum resistance line data, plugging length data, and detonation delay data. Feedback monitoring parameter type data includes coal block size distribution data after the previous borehole blast, ejection range data (obtained via UAV or laser scanning), ground vibration velocity data after borehole blasting (detected via a vibration meter), borehole blasting equipment operating status data, accumulation range data, and the actual coal body fragmentation impact area (identified by 3D laser scanning and image recognition technology to determine the actual coal body fragmentation impact area).

[0021] In a specific embodiment, S1, when performing spatiotemporal unification of multiple original parameter data and generating standardized feature vectors, includes: S11. For borehole coordinate data and laser scanning point cloud data in various original parameter data, the borehole coordinate data and laser scanning point cloud data are transformed from the equipment coordinate system to the mine coordinate system according to the predetermined coordinate system transformation matrix to obtain the initial borehole coordinate sequence and initial point cloud data.

[0022] The coordinate system transformation matrix refers to the combination of the rotation matrix and translation vector between the equipment coordinate system and the mine coordinate system.

[0023] S12. For geological parameter type data, blasting parameter type data and feedback monitoring parameter type data in various original parameter data, spatial binding is achieved by associating with corresponding spatial locations, and they are mapped to the mine coordinate system to obtain initial geological parameter type data, initial blasting parameter type data and initial feedback monitoring parameter type data.

[0024] For example, the vibration velocity data in the feedback monitoring parameter type data corresponds to the spatial position of the vibration meter probe. Mapping the probe position of the vibration meter to the mine coordinate system yields the initial vibration velocity data in the mine coordinate system. The same logic applies to other types of geological parameter data, blasting parameter data, and feedback monitoring parameter data.

[0025] S13, by compensating for the time offset between different acquisition devices through the acquisition device timestamp and synchronization function, converts the initial borehole coordinate sequence, initial point cloud data, initial geological parameter type data, initial blasting parameter type data and initial feedback monitoring parameter type data into a unified time tag, and obtains standard borehole coordinate sequence, standard point cloud data, standard geological parameter type data, standard blasting parameter type data and standard feedback monitoring parameter type data.

[0026] Specifically, a synchronization function is a time mapping function used to unify the time base of different acquisition devices. Its purpose is to compensate for time offsets between acquisition devices, ensuring that the starting borehole coordinate sequences, initial point cloud data, initial geological parameter type data, initial blasting parameter type data, and initial feedback monitoring parameter type data from different sources are mapped to the same time label. The following are two synchronization functions: When there is a fixed time offset between acquisition devices, the synchronization function is as follows when calculating the unified timestamp of a certain acquisition device: unified timestamp = original timestamp of the acquisition device + time offset (obtained by synchronization calibration).

[0027] For acquisition devices with clock drift, when calculating the unified time tag for a certain acquisition device, the synchronization function can use linear time mapping: unified time tag = original timestamp of the acquisition device × clock ratio correction factor + time offset parameter, and the time offset parameter is determined by synchronization calibration data.

[0028] S14 is a standardized feature vector composed of standard borehole coordinate sequences, standard point cloud data, standard geological parameter type data, standard blasting parameter type data, and standard feedback monitoring parameter type data.

[0029] Specifically, the standard borehole coordinate sequence, standard point cloud data, standard geological parameter type data, standard blasting parameter type data, and standard feedback monitoring parameter type data are arranged in a set order to form a standardized feature vector.

[0030] S2. Construct a digital twin model consisting of multiple twin sub-models, input standardized feature vectors into the digital twin model, and output a predicted borehole state vector from the digital twin model.

[0031] Specifically, multiple twin models are pre-built in the cloud.

[0032] In one specific embodiment, S2 includes: S21. Construct a three-dimensional geological structure model, a borehole state vector model, a hybrid prediction model of blasting coal breaking dynamics, an ejection and transport model, and a risk state model, and form a digital twin model from these twin sub-models.

[0033] Specifically, the three-dimensional geological structure model represents the spatial morphology of the coal seam, the roof and floor interfaces, and the structural distribution. Its input is standard geological parameter type data, and its output is geological state type data.

[0034] The borehole state vector model takes a standard borehole coordinate sequence and standard point cloud data as inputs, and is used to map the design trajectory and the measured trajectory in real time. Its output is a borehole coordinate sequence.

[0035] The hybrid prediction model for coal breaking dynamics in blasting uses the empirical formulas of the Kuz-Ram model (a mathematical model for predicting rock block size) as the benchmark framework for block size prediction. It employs the GA-LSSVM model (a hybrid prediction model combining genetic algorithm, global optimization, least squares support vector machine, and key hyperparameters) to predict the shape parameters of the Weibull function (a statistical function of the Weibull distribution). This model is then fused with the parameters of the Kuz-Ram model to form the Weibull-GA-LSSVM blast pile morphology prediction model (which can accurately predict the block size distribution of coal after borehole blasting even with limited data). Simultaneously, ANSYS / LS-DYNA finite element analysis is used as a high-precision offline verification tool to simulate crack propagation, vibration, and impact. The inputs to the hybrid prediction model are standard blasting parameter type data and standard feedback monitoring parameter type data. The outputs include parameters for the blasting type such as block size distribution, charge amount, hole spacing, drill pressure, torque, and vibration velocity; parameters for the borehole blasting equipment type such as drill pressure and torque; vibration velocity; and the area of ​​the coal body fragmentation influence zone.

[0036] The output of the ejection and transport model is the ejection distance and the accumulation range, while the input is standard blasting parameter type data and standard feedback monitoring parameter type data.

[0037] The risk status model outputs risk values ​​such as trajectory, block size, ejection, vibration, and equipment, while the inputs are standard borehole coordinate sequences, standard point cloud data, standard geological parameter type data, standard blasting parameter type data, and standard feedback monitoring parameter type data.

[0038] In addition to the specific implementation methods provided in the embodiments of this invention, digital twin models can also be implemented using mechanistic models (such as the Kuz-Ram model), data-driven models (such as the GA-LSSVM model), or hybrid models combining both. Prediction of parameters such as block size distribution, charge quantity, hole spacing, drilling pressure, torque, and vibration velocity can be achieved using a hybrid Kuz-Ram+GA-LSSVM method, or through pure numerical simulation (such as ANSYS / LS-DYNA finite element analysis), deep learning (such as CNN convolutional neural networks, LSTM long short-term memory networks), or empirical rule methods.

[0039] S22, the standardized feature vector is input into the digital twin model, and each twin sub-model in the digital twin model outputs its corresponding sub-state. The sub-states output by all twin sub-models constitute the predicted borehole state vector.

[0040] Specifically, the predicted borehole state vector includes geological state type data such as coal seam dip angle, thickness, and strength output from the 3D geological structure model; borehole coordinate sequence data output from the borehole state vector model; area of ​​coal body fracturing influence zone, block size distribution, charge amount, borehole spacing, drilling pressure, torque, and vibration velocity data output from the blasting coal breaking dynamics hybrid prediction model; throw distance and accumulation range data output from the throw and transport model; and risk values ​​such as trajectory, block size, throw, vibration, and equipment output from the risk state model. The digital twin model continuously updates the predicted borehole state vector based on the standardized feature vectors collected in real time, realizing the state evolution of the digital twin model.

[0041] S3 generates a vector deviation between the predicted borehole state and the actual borehole state based on the predicted borehole state vector and the standardized feature vector, and generates a borehole blasting control mode based on the vector deviation.

[0042] In one specific embodiment, S3 includes: S31, Generate the actual borehole state vector based on the standardized feature vector.

[0043] Specifically, the actual borehole state vector parameter type is consistent with the predicted borehole state vector and satisfies a one-to-one correspondence.

[0044] S32, obtain the Euclidean distance between the predicted borehole state vector and the actual borehole state vector, as the vector deviation between the predicted borehole state and the actual borehole state.

[0045] Specifically, in calculating the predicted borehole state vector With the actual borehole state vector Vector deviation between When, the formula is: ,in, This represents the Euclidean distance. This method can synthesize the differences between multidimensional parameters into a single scalar deviation value.

[0046] S33: Construct multiple deviation intervals based on preset three-level deviation thresholds, determine the deviation interval where the vector deviation is located, and generate a drilling blasting control mode based on the deviation interval where the vector deviation is located.

[0047] Specifically, Table 1 shows the mapping relationship between the vector deviation δt, the preset three-level deviation threshold (δth1<δth2<δth3) and multiple deviation intervals, and the borehole blasting control mode and specific actions.

[0048]

[0049] S4. When the borehole blasting control mode is parameter reconstruction mode, a borehole particle swarm optimization algorithm is generated based on vector deviation and various geographical parameters of the target steeply inclined coal seam. The optimal blasting parameter set is generated based on the borehole particle swarm optimization algorithm and the current engineering priority of the target steeply inclined coal seam.

[0050] Specifically, within the aforementioned bias-driven framework (Table 1), this embodiment of the invention employs a borehole particle swarm optimization algorithm to dynamically optimize blasting parameters and generate the optimal blasting parameter set. Unlike traditional particle swarm optimization algorithms, the innovation of this embodiment lies in the dynamic adjustment of both the preset weight coefficient set and the search strategy (solved by the borehole particle swarm optimization algorithm) based on the vector bias.

[0051] The optimal blasting parameter set includes parameters such as hole spacing, row spacing, charge structure, and detonation sequence. Various geographical parameters include the dip angle of the target steeply dipping coal seam (affecting the throwing direction and distance), joint direction, throwing boundary deviation range, and surrounding rock disturbance range.

[0052] In a specific embodiment, when the borehole blasting control mode is the parameter reconstruction mode, step S4 includes: S41, Obtain the actual borehole state vector, and construct the initial comprehensive objective function based on the predicted borehole state vector and the actual borehole state vector.

[0053] The principle for obtaining the actual borehole state vector is the same as that in S31, and will not be repeated here.

[0054] In one specific embodiment, S41 includes: S411, obtain the area of ​​the coal body breaking influence region in the predicted borehole state vector, calculate the difference between the coal body breaking influence region area and the target coal breaking region area to obtain the coal breaking difference value, and calculate the ratio of the coal breaking difference value to the target coal breaking region area as the breaking coverage deviation.

[0055] Specifically, the target coal breaking area is a pre-set value for the coal breaking area.

[0056] S412, obtain the actual block size standard deviation and the actual block size mean in the actual borehole state vector, and calculate the ratio of the actual block size standard deviation to the actual block size mean as the block size non-uniformity.

[0057] Specifically, the standard deviation and mean of actual block size are data from the coal block size distribution after the last borehole blasting. The smaller the ratio of the standard deviation to the mean of actual block size, the more uniform the actual coal block size.

[0058] S413, obtain the standard throw range data in the standardized feature vector, calculate the throw area where the standard throw range data exceeds the preset allowable throw area, calculate the difference between the area of ​​the preset allowable throw area and the throw area as the external area, and calculate the ratio of the external area to the preset allowable throw area as the throw over-limit rate.

[0059] Specifically, the smaller the ratio of the external area to the preset allowable ejection area, the more controlled the actual borehole blasting is.

[0060] S414. Construct an initial comprehensive objective function based on the preset weight coefficient group, broken coverage deviation, block size unevenness, and over-limit ejection rate.

[0061] Specifically, the initial integrated objective function is to minimize the aforementioned fragmentation cover deviation, block size inhomogeneity, and out-of-bounds ejection rate. The initial integrated objective function is expressed as: Among them, w1, w2 and w3 are all weight coefficients in the preset weight coefficient group, which represent the weight coefficients corresponding to the crushing coverage deviation, block size inhomogeneity and the over-throw rate, respectively; f1, f2 and f3 represent the crushing coverage deviation, block size inhomogeneity and the over-throw rate, respectively.

[0062] S42, adjust the preset weight coefficient group in the initial comprehensive objective function according to the vector deviation to obtain the optimized weight reorganization.

[0063] In one specific embodiment, S42 includes: S421, Based on the relationship between the adjustment coefficient and the vector deviation obtained from pre-training, determine the adjustment coefficient corresponding to each weight coefficient in the preset weight coefficient group in the initial comprehensive objective function.

[0064] Specifically, the relationship between the adjustment coefficient (the specific adjustment value is determined by the relationship between the pre-trained adjustment coefficient and the vector deviation) and the vector deviation can be simply expressed as follows: when the vector deviation is large, increase the weight coefficient corresponding to f3 and decrease the weight coefficient corresponding to f2. When the vector deviation is small, increase the weight coefficient corresponding to f2. The relationship between the adjustment coefficient and the vector deviation ensures that safety is prioritized in abnormal situations.

[0065] S422, adjust each weight coefficient according to the adjustment coefficient corresponding to each weight coefficient to obtain multiple optimized weights.

[0066] Specifically, the optimized weight is obtained by adjusting the weight coefficient wj according to the adjustment coefficient kj corresponding to a certain weight coefficient. When, the formula is: ;in, This represents the vector deviation.

[0067] S423 is an optimization weight reorganization consisting of all optimization weights.

[0068] Specifically, in addition to the aforementioned optimization weight reorganization, the inertia weight in the borehole particle swarm optimization algorithm is adjusted according to the vector deviation: when the vector deviation is large, the inertia weight is increased to enhance the global search capability and avoid getting trapped in local optima. When the vector deviation is small, the inertia weight is decreased to enhance the local fine-tuning capability and accelerate convergence. The adjustment of the inertia weight is used to change the particle search step size and exploration range of the borehole particle swarm optimization algorithm, without changing the optimization objective; it is only used to improve the parameter search efficiency under different vector deviations. The specific relationship between the vector deviation and the adjustment of the inertia weight is determined based on historical empirical values.

[0069] S43 generates specific constraints based on various geographical parameters of the target steeply dipping coal seam.

[0070] Specifically, the target steeply dipping coal seam has several geographical parameters, including the seam dip angle, joint direction, ejection boundary range, and surrounding rock disturbance range. Considering the steep dip angle, well-developed joints, and asymmetric ejection direction of the target steeply dipping coal seam, specific constraints are formed, including the seam dip angle (affecting the ejection direction and range), joint direction (affecting the crack propagation path), ejection boundary range, and surrounding rock disturbance range. These specific constraints serve as feasible region restrictions in the borehole particle swarm optimization algorithm, and together with the initial integrated objective function, constitute a constrained multi-objective optimization model. The formula is expressed as: Where C() represents the constraint function, J represents the dip angle of the coal seam and the joint direction. This indicates that the value has exceeded the specified range. Indicates the range of disturbance to the surrounding rock. denoted by , where U represents the preset safety upper limit and U represents the single-set predicted borehole state vector obtained after the blasting parameter set is input into the digital twin model.

[0071] S44 uses optimization weight reorganization to replace the preset weight coefficient group in the initial comprehensive objective function, and combines it with specific constraints to generate a borehole particle swarm optimization algorithm.

[0072] Specifically, in this embodiment of the invention, the blasting parameter set is not directly constrained for safety, but is indirectly constrained by a single set of predicted borehole state vectors and specific constraints generated by a digital twin model.

[0073] S45, generate the optimal blasting parameter set based on the borehole particle swarm optimization algorithm and the current engineering priority of the target steeply inclined coal seam.

[0074] In one specific embodiment, S45 includes: S451, multiple initial blasting parameter sets are randomly generated within the parameter feasible region, and a random particle swarm velocity is configured for each initial blasting parameter set to obtain multiple blasting parameter sets.

[0075] S452, input each blasting parameter group into the digital twin model to obtain a single-group predicted borehole state vector for each blasting parameter group, and calculate the corresponding fragmentation coverage deviation, block size inhomogeneity and ejection excess rate for each blasting parameter group based on the actual borehole state vector and the single-group predicted borehole state vector for each blasting parameter group.

[0076] Specifically, when calculating the breakage coverage deviation, block size inhomogeneity, and ejection over-limit rate corresponding to a certain blasting parameter set, the principle in S41 can be referred to, which will not be repeated here.

[0077] S453 uses a Pareto dominance strategy to select multiple non-dominated solutions from multiple blasting parameter sets based on the particle swarm velocity, fragmentation coverage deviation, block size inhomogeneity, and ejection over-limit rate corresponding to each blasting parameter set, and configures all non-dominated solutions in an external archive set.

[0078] Specifically, blasting parameter groups with particle swarm velocities less than or equal to a preset particle swarm velocity safety threshold are eliminated. Assuming the two blasting parameter groups that are not eliminated are group a and group b, if the three indicators corresponding to group a—fragmentation coverage deviation, block size inhomogeneity, and ejection over-limit rate—are all better than or equal to group b, and at least one indicator is better than but not equal to group b, then it indicates that group a is Pareto dominant over group b, and group a is a non-dominated solution.

[0079] S454: Select a non-dominated solution from the external archive set as the global optimal solution.

[0080] Specifically, based on the preset weighting coefficients: particle swarm velocity weight, fragmentation and coverage deviation dominance weight, block size dominance weight, and ejection dominance weight, these are weighted and summed with the particle swarm velocity, fragmentation and coverage deviation, block size inhomogeneity, and ejection over-limit rate of each non-dominated solution to obtain the comprehensive score of each non-dominated solution. The non-dominated solution with the smallest comprehensive score is selected as the global optimal solution.

[0081] S455 updates the particle swarm velocity of all blasting parameter groups based on the borehole particle swarm optimization algorithm and the global optimal solution.

[0082] S456, repeat S52-S55 until the maximum number of iterations is reached and then stop iterating.

[0083] Specifically, in addition to stopping iteration when the maximum number of iterations is reached, iteration can also be stopped when the hole drilling particle swarm optimization algorithm reaches Pareto convergence.

[0084] S457, when stopping iteration, select an optimal solution from multiple non-dominated solutions stored in the external archive set as the optimal blasting parameter set based on the current engineering priority of the target steeply inclined coal seam.

[0085] Specifically, the current project priority is determined manually. The current project priority includes either safety priority or efficiency priority. When safety is prioritized, the non-dominated solution with the smallest blasting effect is selected as the optimal solution; when efficiency is prioritized, the non-dominated solution with the largest blasting effect is selected as the optimal solution.

[0086] S5: The optimal blasting parameter set is sent to the drilling and blasting equipment at the target steeply inclined coal seam site and the drilling and blasting is executed. Based on the feedback of various original parameter data of the preset drilling and blasting parameter type after the drilling and blasting at the target steeply inclined coal seam site, the drilling particle swarm optimization algorithm is optimized, and the optimized drilling particle swarm optimization algorithm is used for the next drilling and blasting.

[0087] Specifically, drilling and blasting equipment includes intelligent drilling rigs and charging equipment.

[0088] In one specific embodiment, S5 includes: S51: The optimal blasting parameter set is sent to the drilling and blasting equipment at the target steeply inclined coal seam site and the drilling and blasting is carried out.

[0089] Specifically, the optimal blasting parameter set is embedded in the drilling and blasting command and sent to the drilling and blasting equipment at the target steeply inclined coal seam. The drilling and blasting equipment then performs the drilling and blasting according to the drilling and blasting command.

[0090] S52, collects various raw parameter data of the preset borehole blasting parameter type after the current borehole blasting in the target steeply inclined coal seam.

[0091] S53, construct a monitoring parameter data vector based on multiple raw parameter data, and calculate the vector deviation between the monitoring parameter data vector and the predicted borehole state vector as the comprehensive deviation.

[0092] Specifically, when constructing the monitoring parameter data vector and calculating the vector deviation between the monitoring parameter data vector and the predicted borehole state vector as the comprehensive deviation, the principles of generating the standardized feature vector in step S1 and obtaining the vector deviation between the predicted borehole state and the actual borehole state in step S32 can be referred to respectively, which will not be elaborated here.

[0093] S54. If the overall deviation is greater than the preset deviation threshold, the model parameters of the digital twin model are corrected according to various original parameter data, and the drilling particle swarm optimization algorithm is re-optimized. The optimized drilling particle swarm optimization algorithm is then used for the next drilling blast.

[0094] Specifically, if the overall deviation exceeds the preset deviation threshold, the digital twin model will be updated. The model parameters of the digital twin model will be corrected using various original parameter data obtained from this borehole blasting. For example, the empirical parameters in the Kuz-Ram model will be corrected using the block size distribution data from this borehole blasting, and the weights of the GA-LSSVM model will be updated using the ejection range data from this borehole blasting, etc.

[0095] Based on the above steps, in a specific embodiment, when S5 sends the optimal blasting parameter set to the drilling blasting equipment at the target steeply inclined coal seam site and performs drilling blasting, it further includes: S55, constructing a trajectory deviation correction branch, including: during the process of sending the optimal blasting parameter set to the drilling and blasting equipment at the target steeply inclined coal seam site and executing drilling and blasting, real-time monitoring of the drilling trajectory, and when the measured trajectory deviates from the design trajectory and the degree of deviation exceeds the preset deviation threshold, generating a correction command.

[0096] Specifically, the correction instructions include re-optimizing equipment parameters such as borehole orientation, attitude angle, and borehole inclination angle. A specific example is adjusting the attitude angle of the directional drilling tool to 35°. In this embodiment of the invention, both the measured trajectory and the designed trajectory are sequences composed of multiple points. When determining whether the measured trajectory deviates from the designed trajectory and whether the deviation exceeds a preset deviation threshold, this embodiment of the invention calculates the trajectory deviation d between the designed trajectory sequence and the measured trajectory sequence; the formula for calculating the trajectory deviation d between the designed trajectory sequence and the measured trajectory sequence is: Where n represents the number of borehole coordinate values ​​in the design trajectory sequence and the measured trajectory sequence. This represents the coordinate value of the i-th borehole in the measured trajectory sequence. This represents the coordinate value of the i-th borehole in the design trajectory sequence. When d is greater than a preset deviation threshold, it is determined that the measured trajectory deviates from the design trajectory and the degree of deviation exceeds the preset deviation threshold.

[0097] S56, construct a compensation branch for insufficient coal coverage and unqualified block size in blasting, including: obtaining the block size distribution of coal body from various original parameter data, and generating a compensation blasting command if the large block rate in the block size distribution exceeds the standard or the coverage area does not meet the target.

[0098] Specifically, the compensation blasting command includes automatically increasing the charge per hole (within a safe range) or reducing the hole spacing and row spacing, and recalculating the detonation sequence to achieve compensation blasting.

[0099] S57, construct branches for throwing out of bounds and vibration exceeding limits, including obtaining the residue scattering range and blasting vibration velocity from various raw parameter data. When the residue scattering range exceeds the allowable area or the blasting vibration velocity is greater than the vibration threshold, a remote limiting command is generated.

[0100] Specifically, remote limiting commands include reducing the charge per hole, increasing the plugging length, or adjusting the detonation delay (e.g., changing from a millisecond delay to a half-second delay) to achieve remote limiting control.

[0101] S58 configures the trajectory deviation correction branch, the blasting coal coverage insufficient and block size unqualified compensation branch, and the throwing over the boundary and vibration over the limit branch into the drilling blasting equipment at the target steeply inclined coal seam site. Among them, the trajectory deviation correction branch, the blasting coal coverage insufficient and block size unqualified compensation branch, and the throwing over the boundary and vibration over the limit branch can be triggered independently or in combination. Moreover, the deviation correction command, the compensation blasting command, and the remote limit command are all issued after being verified by the digital twin model.

[0102] Specifically, the correction commands, compensation blasting commands, and remote amplitude limiting commands are all issued after being verified by the digital twin model, which can ensure the safety of on-site implementation.

[0103] This invention is not a general modeling and display platform applicable to the entire mine, but rather a specialized dynamic control scheme for drilling and coal breaking operations in steeply inclined coal seams. It has a clearer technical focus and stronger engineering relevance. All steps in this invention form a complete closed loop, reducing reliance on experience and improving operational accuracy, safety, and stability. Furthermore, this invention features an online feedback and knowledge base reuse mechanism, utilizing the similarity of working conditions to reuse historical experience data, thereby improving the initial configuration efficiency of this invention under similar geological conditions.

[0104] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams, characterized in that, include: S1, collect multiple raw parameter data of preset drilling and blasting parameter types after the last drilling and blasting of the target steeply inclined coal seam, and unify the multiple raw parameter data in time and space to generate a standardized feature vector; S2, construct a digital twin model consisting of multiple twin sub-models, input standardized feature vectors into the digital twin model, and output a predicted borehole state vector from the digital twin model; S3, generate the vector deviation between the predicted borehole state and the actual borehole state based on the predicted borehole state vector and the standardized feature vector, and generate the borehole blasting control mode based on the vector deviation. S4, When the borehole blasting control mode is parameter reconstruction mode, a borehole particle swarm optimization algorithm is generated based on vector deviation and multiple geographical parameters of the target steeply inclined coal seam. The optimal blasting parameter set is generated based on the borehole particle swarm optimization algorithm and the current engineering priority of the target steeply inclined coal seam. S5: The optimal blasting parameter set is sent to the drilling and blasting equipment at the target steeply inclined coal seam site and the drilling and blasting is executed. Based on the feedback of various original parameter data of the preset drilling and blasting parameter type after the drilling and blasting at the target steeply inclined coal seam site, the drilling particle swarm optimization algorithm is optimized, and the optimized drilling particle swarm optimization algorithm is used for the next drilling and blasting.

2. The digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams according to claim 1, characterized in that, The process of S1, which unifies multiple original parameter data in time and space and generates standardized feature vectors, includes: S11. For borehole coordinate data and laser scanning point cloud data in various original parameter data, according to the predetermined coordinate system transformation matrix, the borehole coordinate data and laser scanning point cloud data are transformed from the equipment coordinate system to the mine coordinate system to obtain the initial borehole coordinate sequence and initial point cloud data. S12, For geological parameter type data, blasting parameter type data and feedback monitoring parameter type data in various original parameter data, spatial binding is achieved by associating with corresponding spatial locations, and they are mapped to the mine coordinate system to obtain initial geological parameter type data, initial blasting parameter type data and initial feedback monitoring parameter type data; S13, by compensating for the time offset between different acquisition devices through the acquisition device timestamp and synchronization function, the initial borehole coordinate sequence, initial point cloud data, initial geological parameter type data, initial blasting parameter type data and initial feedback monitoring parameter type data are converted into a unified time tag to obtain standard borehole coordinate sequence, standard point cloud data, standard geological parameter type data, standard blasting parameter type data and standard feedback monitoring parameter type data; S14 is a standardized feature vector composed of standard borehole coordinate sequences, standard point cloud data, standard geological parameter type data, standard blasting parameter type data, and standard feedback monitoring parameter type data.

3. The digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams according to claim 1, characterized in that, S2 includes: S21, construct a three-dimensional geological structure model, a borehole state vector model, a hybrid prediction model of blasting coal breaking dynamics, an ejection and transport model, and a risk state model, and form a digital twin model from these twin sub-models; S22, the standardized feature vector is input into the digital twin model, and each twin sub-model in the digital twin model outputs its corresponding sub-state. The sub-states output by all twin sub-models constitute the predicted borehole state vector.

4. The digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams according to claim 1, characterized in that, S3 includes: S31, Generate the actual borehole state vector based on the standardized feature vector; S32, obtain the Euclidean distance between the predicted borehole state vector and the actual borehole state vector, as the vector deviation between the predicted borehole state and the actual borehole state; S33: Construct multiple deviation intervals based on preset three-level deviation thresholds, determine the deviation interval where the vector deviation is located, and generate a drilling blasting control mode based on the deviation interval where the vector deviation is located.

5. The digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams according to claim 1, characterized in that, When the borehole blasting control mode is parameter reconstruction mode, S4 includes: S41, Obtain the actual borehole state vector, and construct the initial comprehensive objective function based on the predicted borehole state vector and the actual borehole state vector; S42, adjust the preset weight coefficient group in the initial comprehensive objective function according to the vector deviation to obtain the optimized weight reorganization; S43 generates specific constraints based on various geographical parameters of the target steeply dipping coal seam; S44, replace the preset weight coefficient group in the initial comprehensive objective function with optimization weight reorganization, and generate the borehole particle swarm optimization algorithm in combination with specific constraints; S45, Generate the optimal blasting parameter set based on the borehole particle swarm optimization algorithm and the current engineering priority of the target steeply inclined coal seam.

6. The digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams according to claim 5, characterized in that, S41 includes: S411, obtain the area of ​​the coal body breaking influence region in the predicted borehole state vector, calculate the difference between the coal body breaking influence region area and the target coal breaking region area to obtain the coal breaking difference value, and calculate the ratio of the coal breaking difference value to the target coal breaking region area as the breaking coverage deviation. S412, obtain the actual block size standard deviation and the actual block size mean in the actual borehole state vector, and calculate the ratio of the actual block size standard deviation to the actual block size mean as the block size non-uniformity. S413, obtain the standard throw range data in the standardized feature vector, calculate the throw area where the standard throw range data exceeds the preset allowable throw area, calculate the difference between the area of ​​the preset allowable throw area and the throw area as the external area, and calculate the ratio of the external area to the preset allowable throw area as the throw over-limit rate. S414. Construct an initial comprehensive objective function based on the preset weight coefficient group, broken coverage deviation, block size unevenness, and over-exploitation rate.

7. The digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams according to claim 5 or 6, characterized in that, S42 includes: S421, Based on the relationship between the adjustment coefficient and the vector deviation obtained from pre-training, determine the adjustment coefficient corresponding to each weight coefficient in the preset weight coefficient group in the initial comprehensive objective function; S422, adjust each weight coefficient according to the adjustment coefficient corresponding to each weight coefficient to obtain multiple optimized weights; S423 is an optimization weight reorganization consisting of all optimization weights.

8. The digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams according to claim 5 or 6, characterized in that, The S45 includes: S451, randomly generate multiple initial blasting parameter sets in the parameter feasible region, and configure random particle swarm velocities for each initial blasting parameter set to obtain multiple blasting parameter sets; S452, input each blasting parameter group into the digital twin model to obtain a single-group predicted borehole state vector for each blasting parameter group, and calculate the corresponding fragmentation coverage deviation, block size inhomogeneity and ejection over-limit rate for each blasting parameter group based on the actual borehole state vector and the single-group predicted borehole state vector for each blasting parameter group. S453 uses a Pareto dominance strategy to select multiple non-dominated solutions from multiple blasting parameter sets based on the particle swarm velocity, fragmentation coverage deviation, block size inhomogeneity, and ejection over-limit rate corresponding to each blasting parameter set, and configures all non-dominated solutions in an external archive set. S454: Select a non-dominated solution from the external archive set as the global optimal solution; S455, update the particle swarm velocity of all blasting parameter groups according to the borehole particle swarm optimization algorithm and the global optimal solution; S456, repeat S52-S55 until the maximum number of iterations is reached and then stop iterating; S457, when stopping iteration, select an optimal solution from multiple non-dominated solutions stored in the external archive set as the optimal blasting parameter set based on the current engineering priority of the target steeply inclined coal seam.

9. The digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams according to claim 1, characterized in that, S5 includes: S51, the optimal blasting parameter set is sent to the drilling and blasting equipment at the target steeply inclined coal seam and the drilling and blasting is carried out; S52, collect various raw parameter data of the preset borehole blasting parameter type after this borehole blasting in the target steeply inclined coal seam; S53, construct a monitoring parameter data vector based on multiple raw parameter data, and calculate the vector deviation between the monitoring parameter data vector and the predicted borehole state vector as the comprehensive deviation; S54. If the overall deviation is greater than the preset deviation threshold, the model parameters of the digital twin model are corrected according to various original parameter data, and the drilling particle swarm optimization algorithm is re-optimized. The optimized drilling particle swarm optimization algorithm is then used for the next drilling blast.

10. The digital twin dynamic control method for drilling and coal breaking operations in steeply inclined coal seams according to claim 9, characterized in that, When S5 sends the optimal blasting parameter set to the drilling blasting equipment at the target steeply inclined coal seam and executes drilling blasting, it also includes: S55, Construct a trajectory deviation correction branch, including: During the process of sending the optimal blasting parameter set to the drilling blasting equipment at the target steeply inclined coal seam site and executing drilling blasting, monitor the drilling trajectory in real time, and generate a correction command when the measured trajectory deviates from the design trajectory and the degree of deviation exceeds the preset deviation threshold. S56, Construct a compensation branch for insufficient coal coverage and unqualified block size in blasting, including: obtaining the block size distribution of coal body from various original parameter data; if the large block rate in the block size distribution exceeds the standard or the coverage area does not meet the target, then generate a compensation blasting command; S57, construct the throwing out of the boundary and vibration exceeding the limit branch, including obtaining the residue throwing range and blasting vibration velocity from various raw parameter data. When the residue throwing range exceeds the allowable area or the blasting vibration velocity is greater than the vibration threshold, a remote limit command is generated. S58 configures the trajectory deviation correction branch, the blasting coal coverage insufficient and block size unqualified compensation branch, and the throwing over the boundary and vibration over the limit branch into the drilling blasting equipment at the target steeply inclined coal seam site. Among them, the trajectory deviation correction branch, the blasting coal coverage insufficient and block size unqualified compensation branch, and the throwing over the boundary and vibration over the limit branch can be triggered independently or in combination. Moreover, the deviation correction command, the compensation blasting command, and the remote limit command are all issued after being verified by the digital twin model.

Images