Knowledge graph-based blasting charge analysis method and system

Abstract

This invention discloses a method and system for analyzing blasting charges based on knowledge graphs, belonging to the field of knowledge graph technology. The method includes: discretizing the rock mass to be blasted into graph nodes; mapping the spatial distribution characteristics of the rock mass's original microfractures to initial implicit tearing degrees on entity relationship edges to construct an initial geological knowledge graph; mapping the initiation energy to the initial information entropy impulse injected into the nodes; using a Markov entropy flow graph operator to drive the initial information entropy impulse along entity relationship edges for network propagation, and dynamically updating the current implicit tearing degree of the edges based on entropy flow loss to adjust subsequent entropy flow transfer; triggering an explicit topological tearing mechanism when the cumulative damage corresponding to any entity relationship edge exceeds the critical fracture tolerance; after the information entropy dissipation terminates, extracting the residual graph topology and decoding independent rock fragments based on connected components to output the blasting block size analysis results. This method can uniformly characterize fracture structure, blasting propagation, damage evolution, and block size formation process.

Description

Technical Field

[0001] This invention relates to the field of knowledge graph technology, and in particular to a method and system for analyzing explosive charges based on knowledge graphs. Background Technology

[0002] Explosive charge analysis is a crucial technical step in open-pit mining, underground tunnel excavation, water conservancy project excavation, and complex rock mass demolition. The analysis results directly affect the rock mass fragmentation effect, subsequent loading and transportation efficiency, blasting cost control, and construction safety. Existing methods for blasting block size analysis mainly include empirical formula methods, field screening and statistical methods, continuous medium-based numerical simulation methods, and discrete element method (DEM) particle fragmentation simulation methods. While empirical formula methods are simple to apply, they are less adaptable to variations in primary rock fractures, joint development, charge initiation timing, and local stress environments, making it difficult to accurately reflect the actual blasting results under complex geological conditions. Field screening and statistical methods rely on post-blast measurements and cannot provide effective predictions before blasting. Continuous medium numerical simulation, while capable of describing stress wave propagation, faces challenges in modeling complex fracture networks, local fracture separation, and the final fragment topology, resulting in high computational costs. Although the DEM is suitable for describing particle splitting and movement, parameter calibration is complex, and the computational load is enormous, making it difficult to quickly generate block size prediction results applicable to engineering sites.

[0003] Meanwhile, the blasting process is essentially a process in which explosive energy propagates within heterogeneous fractured rock masses, accumulates damage, propagates fractures, and ultimately forms independent fragments. It simultaneously contains spatial structural information, material mechanical properties, local damage evolution relationships, and post-blast topological separation results. Most existing technologies model these factors separately, lacking an analytical method that can unify rock mass structure, fracture distribution, blasting input, propagation feedback, and fracture results into a single computational framework. Especially for rock masses with numerous primary microfractures and complex joint structures, effectively correlating three-dimensional fracture spatial characteristics with blasting propagation paths, adaptive damage evolution, and fragmentation gradation results at a low computational cost remains a prominent technical challenge in current blasting engineering analysis. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a method for analyzing explosive charges based on a knowledge graph, which includes discretizing the rock mass to be blasted into graph nodes, reducing the spatial distribution characteristics of the original micro-fractures in the rock mass to an initial implicit tear degree on the entity relationship edge, and constructing an initial geological knowledge graph; After obtaining the explosive charge, the detonation energy is mapped to the initial information entropy impulse of the injected node; Using a Markov entropy flow graph operator with damage adaptive routing capabilities, the initial information entropy impulse is driven to propagate along the entity relationship edges in the network. The current implicit tear degree of the corresponding edge is dynamically updated by accumulating the information entropy flow through each entity relationship edge. The updated current implicit tear degree is used as an independent variable to adjust the entropy flow transfer of the entity relationship edges. By monitoring the total information entropy accumulated by each rock mass node, when the total information entropy of any edge exceeds the critical fracture tolerance determined by the equivalent mapping based on its static mechanical properties, an explicit topological tearing mechanism is triggered, forcibly disconnecting the entity relationship edges in the initial geological knowledge graph. After the information entropy dissipation in the network terminates, the remaining binary discrete graph topology after the explicit topological tearing mechanism is extracted. Based on the graph theory connected component features, the mutually isolated discrete graph subsets are decoded and mapped into independent rock fragments in physical three-dimensional space, and the engineering-grade blasting block size analysis results are output.

[0006] As a preferred embodiment of the knowledge graph-based explosive charge analysis method of the present invention, the step of dimensionality reduction and mapping of the spatial distribution characteristics of the original microfractures in the rock mass to the initial implicit tear degree on the entity relation edge includes extracting the three-dimensional point cloud data of the area to be analyzed by drilling measurement and reconstructing the three-dimensional fracture surface geometric set by using a triangulation algorithm. In the initial geological knowledge map, adjacent rock mass nodes are connected. and The spatial line segment is the central axis, and the radius is [missing information]. The influence envelope cylinder is used to filter out the set of local crack surfaces that intersect with the influence envelope cylinder; Calculate the sum of the orthographic projection areas of the set of local fracture surfaces in the direction perpendicular to the central axis. ; Using the enveloping cylinder as a short-path propagation path for explosive energy, and taking the proportion of cracks in the cross-section of the path as a criterion, the initial latent tear strength is quantified. in, Represents entity relationship edges The initial latent tear degree.

[0007] As a preferred embodiment of the knowledge graph-based explosive charge analysis method of the present invention, wherein: the initial information entropy impulse includes extracting the first... Explosive charge per blast hole Explosive calorific value per kilogram And the set absolute time for the micro-delay detonation of the detonator. Constructing a discrete inference time step Discrete-time window pulse release function as independent variable : within the time interval The value inside is Its value is 0 in other intervals; compute nodes exist Initial information entropy impulse at time: ; Within the corresponding time interval, continuously add data to the initial geological knowledge map. The node corresponding to each blast hole position is injected with the initial information entropy impulse; in, This refers to the time step for a single simulation. This is the energy-information equivalent constant; This represents the number of time steps obtained by discretizing the effective duration of the detonation effect, used to characterize the range of step lengths in which the detonation energy of the corresponding borehole is continuously injected during the pattern extrapolation process.

[0008] As a preferred embodiment of the knowledge graph-based explosive charge analysis method of the present invention, the Markov entropy flow graph operator is a dynamic message passing model without aftereffects built on a discrete geological knowledge graph, which internally encapsulates a pre-trained Bayesian network. The dynamic cumulative update of the implicit tear degree includes, in the extrapolation time step, extracting the absolute value of the information entropy flow of the current time step node. ; Invoke the Bayesian network to obtain the current hidden tear degree and node stress gradient of all nodes adjacent to the current node. The input matrix includes rock mass type as a feature matrix; the output is the energy loss rate for each adjacent entity relationship edge, and the proportion of energy used to tear the connection relationship for each entity relationship edge. Represents entity relationship edges The probability of energy doing work; Net effective information entropy of subsequent network propagation ; in, This represents the sum of energy losses on all adjacent entity relation edges, where each entity relation edge represents the sum of energy losses on all adjacent entity relation edges. Energy loss is expressed as ; Represents entity relationship edges Energy loss, through entity relationship edges Energy loss rate and Multiplying them together yields the result; An autoregressive cumulative truncation rule is adopted, using the energy loss as the sole driving source of disruptive work, to update entity relation edges. Current hidden tearing degree ; in Represents entity relationship edges The percentage of energy used to tear apart connections.

[0009] As a preferred embodiment of the knowledge graph-based explosive charge analysis method described in this invention, wherein: the updated entity relation edges are used... Given the current latent tearing degree, and simulating the energy release to avoid a high-resistance guiding mechanism, the adjusted non-normalized transfer weights are calculated: ; For nodes All adjacent edges The non-normalized transition weights are subjected to global normalization, and the formal entropy flow transition probability of the next time step is output. ; in, Represents entity relationship edges The basic transition probability, This represents a constant characterizing the sensitivity of a fracture to energy blocking. It is the set of adjacent nodes; The net effective information entropy and the formal entropy flow transition probability are used to... Multiplying, the information entropy flow that actually flows into each adjacent entity relationship edge and propagates downstream is obtained, driving the operator to complete the entropy flow adaptive transfer in a single time step; after completing the calculation of the current time step, the next time step is calculated through the Markov entropy flow graph operator.

[0010] As a preferred embodiment of the knowledge graph-based explosive charge analysis method of the present invention, the method involves: traversing the initial geological knowledge graph, extracting the uniaxial tensile strength, rock elastic modulus, maximum principal stress and minimum principal stress of the spatial location of the target entity relationship edge recorded by the first rock mass node and the second rock mass node connected at both ends of the target entity relationship edge, and extracting the equivalent physical volume of the geometric mapping of the entity relationship edge. Based on the maximum distortion strain energy density theory, the intrinsic tensile strain energy of the first rock mass node and the second rock mass node are calculated respectively, and the minimum value of the two is selected as the interface ultimate tensile strain energy of the solid relationship edge according to the weak surface fracture rule. Extract the maximum and minimum principal stresses, calculate the ratio of the square of the principal stress difference to the rock elastic modulus, and generate the formation deviatoric stress work compensation term. The interface ultimate tensile strain energy is algebraically superimposed with the work compensation term of the formation deviatoric stress, and multiplied by the dimension transformation constant that maps the mechanical Joule dimension to the information theory entropy unit and the equivalent physical volume to calculate the absolute critical fracture tolerance of the entity relation edge.

[0011] As a preferred embodiment of the knowledge graph-based explosive charge analysis method of the present invention, the binarized discrete graph topology includes: extracting the edge relationships after information dissipation cutoff in the network, performing network clustering on the set of rock mass nodes connected by all edges, and dividing them into U mutually isolated discrete graph subsets. Traverse the e-th discrete graph subset For all rock mass nodes within the area, sum the equivalent physical volumes of each node algebraically to obtain the total predicted volume of the e-th independent rock fragment: ; in, represents the e-th discrete map subset; r represents the index of the rock mass node within the map subset; represents the equivalent physical volume of the r-th rock mass node within the subset of the graph; e represents the index of any discrete subset of the U discrete subsets of the graph. Based on the preset standard sieve size range, the total predicted volume of all independently generated rock fragments is grouped by range and the cumulative mass ratio is statistically analyzed, and the blasting block size distribution curve is fitted and output.

[0012] Secondly, the present invention provides a blasting charge analysis system based on a knowledge graph, comprising: a graph construction unit, which discretizes the rock mass to be blasted into graph nodes, reduces the spatial distribution characteristics of the original micro-fractures in the rock mass to an initial implicit tear degree on the entity relationship edge, and constructs an initial geological knowledge graph; after obtaining the blasting charge, it maps the detonation energy to the initial information entropy impulse of the injected node; The computing unit uses a Markov entropy flow graph operator with damage adaptive routing function to drive the initial information entropy impulse to be propagated along the entity relationship edge in the network, and dynamically accumulates and updates the current implicit tear degree of the corresponding edge by the information entropy flow flowing through each entity relationship edge, and uses the updated current implicit tear degree as the independent variable to adjust the entropy flow transfer of the entity relationship edge. The analysis unit monitors the total information entropy accumulated by each rock mass node. When the total information entropy of any edge exceeds the critical fracture tolerance determined by the equivalent mapping based on its static mechanical properties, it triggers an explicit topological tearing mechanism to forcibly disconnect the entity relationship edges in the initial geological knowledge graph. The output unit extracts the binary discrete graph topology that survives the explicit topological tearing mechanism after the information entropy dissipation in the network terminates. Based on the graph theory connected component features, it reverse-decodes and maps the mutually isolated discrete graph subsets into independent rock fragments in physical three-dimensional space, and outputs the engineering-grade blasting block size analysis results.

[0013] Thirdly, the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the knowledge graph-based explosive charge analysis method as described in the first aspect of the present invention.

[0014] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the knowledge graph-based explosive charge analysis method as described in the first aspect of the present invention.

[0015] The beneficial effects of this invention are as follows: By discretizing the rock mass to be blasted into nodes of a geological knowledge graph and mapping the spatial distribution characteristics of primary microfractures to initial implicit tearing degrees on entity relationship edges, a graph-based analysis foundation with unified bearing capacity, mechanical properties, and connection relationships of the rock mass is established. Furthermore, the detonation energy of the charge is mapped to the information entropy impulse on the nodes, and a Markov entropy flow graph operator with damage adaptive routing capability is used to realize the dynamic transmission, loss work, and damage accumulation inference of the blasting action in the fractured rock mass, thereby more realistically reflecting the propagation of energy disturbance and the evolution of local fractures within complex rock masses. Simultaneously, this invention directly transforms local damage results into fractures in the graph connection relationships through an explicit topological tearing mechanism, and decodes independent rock fragments in reverse based on the connected components of the remaining graph. This allows for the direct output of engineering-level blasting block size analysis results without the need for high-consumption discrete element motion solutions, offering advantages such as unified modeling, clear physical meaning, high computational efficiency, and strong adaptability to engineering applications. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. 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.

[0017] Figure 1 This is a flowchart of a knowledge graph-based method for analyzing explosive charges.

[0018] Figure 2 This is a diagram of a computer device used for a knowledge graph-based explosive charge analysis method. Detailed Implementation

[0019] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0020] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0021] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0022] Reference Figure 1 and Figure 2 This is one embodiment of the present invention, which provides a method for analyzing explosive charges based on knowledge graphs, including the following steps: S1: Discretize the rock mass to be blasted into graph nodes, and reduce the spatial distribution characteristics of the original micro-fractures in the rock mass to the initial implicit tear degree on the entity relationship edge, so as to complete the transformation configuration from physical boundary to pure graph theory operator inference boundary and construct the initial geological knowledge graph.

[0023] Specifically, the 3D point cloud data from the drilling measurement while drilling (SWD) of the area to be analyzed is extracted, and a 3D fracture surface geometry set is reconstructed using a triangulation algorithm. SWD is a technology that acquires real-time spatial information of the rock mass surrounding the borehole during drilling. It uses methods such as acoustic logging, imaging logging, or drilling ground-penetrating radar to collect 3D point cloud data within the rock mass, including spatial coordinates, reflection intensity, and lithological information, to reflect the spatial distribution of primary fractures. The triangulation algorithm is a computational geometry method that converts discrete point sets into continuous surface patches. By connecting adjacent points according to topological relationships to form a series of triangular units, a smooth and continuous surface is created. In this technical solution, the discrete 3D point cloud obtained from SWD is input into the triangulation algorithm, which generates a triangular mesh according to the spatial location and local topological relationships of the points, and then combines them to form a complete 3D fracture surface geometry set. In this way, the originally discrete and unstructured fracture point data is digitized and made continuous, enabling accurate representation of the spatial location, extension direction, dip angle, and area information of the fractures, and providing a computational basis for the initial implicit tear degree subsequently mapped to the knowledge graph edges.

[0024] Considering that three-dimensional irregular geological defects cannot be calculated across latitudinal dimensions in mathematical models, an envelope cylindrical space orthographic projection dimensionality reduction algorithm was adopted; in the initial geological knowledge graph, adjacent rock mass nodes are connected. and The spatial line segment is the central axis, and the radius is [missing information]. The influence envelope cylinder is used to screen out the set of local fracture surfaces that intersect with the influence envelope cylinder; by calculating the proportion of the positive projection area of ​​the real three-dimensional physical fracture on the energy straight propagation channel, the complex spatial occlusion effect is accurately and equivalently reduced to a scalar probability attribute on the edge of a one-dimensional map, thus breaking down the underlying data barrier between microscopic geological exploration data and macroscopic map calculation network operations.

[0025] Calculate the sum of the orthographic projection areas of the set of local fracture surfaces in the direction perpendicular to the central axis. The set of local fracture surfaces intersecting the influence envelope cylinder refers to the set of all fracture surfaces that overlap, pass through, cut into, or contact the cylindrical space constructed with the line connecting adjacent rock mass nodes as the central axis and a preset influence radius. The purpose of setting this set of local fracture surfaces is to filter out the fracture surfaces that truly affect the transmission channel corresponding to the current entity relationship edge from the overall fracture structure, thereby reflecting the local structural defects in the neighborhood of that edge and providing a basis for subsequent calculation of the initial implicit tearing degree of that entity relationship edge.

[0026] Using the enveloping cylinder as a short-path propagation path for explosive energy, and taking the proportion of cracks in the cross-section of the path as a criterion, the initial latent tear strength is quantified. in, Represents entity relationship edges The initial latent tear degree.

[0027] It should be noted that the reason why the influence of the envelope cylinder is defined as the short-range propagation path of blasting energy is that the entity relationship edge in the geological knowledge graph only represents the local direct transmission relationship between adjacent rock mass nodes, and does not correspond to the long-distance propagation path in the entire field. For this type of local propagation channel, the weakening of energy transmission by the fracture is mainly manifested in the cutting and occupation of the channel cross section. Therefore, the proportion of fracture occupation in the cross section can be used as a criterion to quantify the initial implicit tearing degree of the entity relationship edge.

[0028] S2: After acquiring the explosive charge, map the detonation energy to the initial information entropy impulse of the injected node.

[0029] Furthermore, the initial information entropy impulse includes the extraction of the first [unclear] in the loading scheme. Explosive charge per blast hole Explosive calorific value per kilogram And the set absolute time for the micro-delay detonation of the detonator. .

[0030] By assuming that the energy release during the explosion process is uniform over a short period of time, a method is constructed to discretely extrapolate the time steps. Discrete-time window pulse release function as independent variable : within the time interval The value inside is The value is 0 in other intervals.

[0031] compute nodes exist Initial information entropy impulse at time: .

[0032] Within the corresponding time interval, continuously add data to the initial geological knowledge map. The nodes corresponding to the positions of the gun holes are injected with the initial information entropy impulse.

[0033] in, This refers to the time step for a single simulation. This is the energy-information equivalent constant; This represents the number of time steps obtained by discretizing the effective duration of the detonation effect, used to characterize the range of step lengths in which the detonation energy of the corresponding borehole is continuously injected during the pattern extrapolation process.

[0034] By introducing the discrete-time window pulse release function as a mathematical technique, the chemical energy of the charge is converted into a pulse data stream that is dynamically injected with each time step of the simulation. The detonation duration of the explosive and the differential delay of the detonator are accurately reproduced at the pure algorithm level, so that the energy injection of different boreholes naturally has a time-dimensional delay difference, thus constructing an extremely accurate physical evolution premise for the spatiotemporal interference superposition effect of stress waves.

[0035] It's important to understand that by dynamically injecting the chemical energy of the explosive charge into the knowledge graph in a computable and traceable manner, the precise simulation of the energy release sequence during the blasting process can be achieved. By extracting the charge amount, heat of explosion, and detonator differential initiation time for each borehole, and combining this with a discrete-time window pulse release function, the total energy is uniformly distributed and injected into the corresponding nodes according to discrete time steps. This preserves the short-time release characteristics of the explosion process while also reflecting the differential initiation time difference between different boreholes. This design can accurately reproduce the spatiotemporal distribution of energy injection at the algorithm level, providing a realistic temporal physical premise for the subsequent propagation of information entropy in the graph, mutual interference, and local stress accumulation. This ensures high accuracy and physical rationality in blasting energy conduction, fracture damage evolution, and final block size prediction.

[0036] S3: Using a Markov entropy flow graph operator with damage adaptive routing function, the initial information entropy impulse is driven to propagate along the entity relationship edges. During the propagation process, the current implicit tear degree of the corresponding edge is dynamically updated by accumulating the information entropy flow flowing through each entity relationship edge. The updated current implicit tear degree is used as the independent variable to adjust the entropy flow transfer of the entity relationship edge, thereby forcing the subsequent information entropy flow to automatically bypass the high resistance zone and undergo adaptive spatial redistribution to the undamaged rock mass, providing a basis for the concentrated evolution of energy for macroscopic fractures.

[0037] Furthermore, the Markov entropy flow graph operator is a non-aftereffect dynamic message-passing model built on a discrete geological knowledge graph. It internally encapsulates a pre-trained Bayesian network and integrates multi-modal energy dissipation mechanisms of plastic absorption, damped oscillations, and crevice depressurization. This is used to iteratively solve the nonlinear propagation and irreversible dissipation process of explosive stress waves in a defective rock mass network within the topological space, time-step by time. It accurately reproduces the nonlinear and irreversible evolution process of explosive stress waves in a defective rock mass network within the topological graph space, making energy flow, damage accumulation, and local dissipation effects quantifiable and traceable at the algorithmic level. This provides a physically sound and highly accurate dynamic basis for subsequent implicit tearing degree updates, path adaptive adjustments, and blasting block size predictions.

[0038] It should be noted that the purpose of using Bayesian networks in this invention is to address the multi-factor coupling and uncertainty issues in the propagation of blasting energy in fractured rock masses. Blasting energy transmission is influenced not only by factors such as the type of rock mass at nodes, mechanical properties, and stress gradients, but also by complex conditional dependencies on the implicit tearing degree of adjacent edges, the entropy of the flowing information, and the distribution of local fractures. Bayesian networks can model the nonlinear and stochastic relationships between these factors through conditional probability, inferring the energy loss rate and tearing energy proportion of edges at each time step. This provides probabilistic adaptive propagation constraints for the Markov entropy flow graph operator, enabling dynamic simulation of local energy dissipation, damage accumulation, and adaptive path adjustment. This ensures that the update of implicit tearing degree and the final block size prediction have physical rationality and high accuracy.

[0039] The dynamic cumulative update of the implicit tear degree includes, in the extrapolation time step, extracting the absolute value of the information entropy flow of the current time step node. ; Invoke the Bayesian network to obtain the current hidden tear degree and node stress gradient of all nodes adjacent to the current node. The input matrix includes rock mass type as a feature matrix; the output is the energy loss rate for each adjacent entity relationship edge, and the proportion of energy used to tear the connection relationship for each entity relationship edge. Represents entity relationship edges The probability of energy being used for work is determined by the input variables, while other energy is lost through oscillations or released through cracks. This is mainly because these input variables have a direct physical relationship with the energy conduction and local damage of edges. The latent tear degree reflects the degree of local damage to the edge, determining the proportion of energy conducted that is converted into damage or dissipation; the node stress gradient represents the direction and intensity of local stress concentration, with high gradient regions being more prone to microcrack propagation and energy dissipation; the information entropy flow represents the total amount of energy currently received by the node, which is the source driving damage and dissipation; the rock mass type reflects the mechanical properties and crack development degree of different materials, which affects the energy transfer and damage characteristics of edges. Bayesian networks utilize the conditional dependencies between these input variables to perform probabilistic inference, thereby calculating the energy loss and work distribution of each adjacent edge at each time step, realizing the dynamic simulation of local energy conduction, dissipation, and damage, and providing a physically reasonable and quantifiable basis for the update of latent tear degree and the adaptive conduction of entropy flow.

[0040] The calculation of nodal stress gradients can be achieved by combining the gravitational stress generated by the rock mass's self-weight, local structural features, and the spatial location of the nodes. Specifically, the method is as follows: First, the gravitational load on the node is calculated based on parameters such as the node's depth and rock mass density, serving as the basic stress. Then, the stress transfer and concentration between neighboring nodes are corrected by considering the elastic modulus, uniaxial tensile strength, and local fracture and joint distribution of the rock mass to which the node belongs, reflecting the influence of structural differences on stress distribution. Finally, the stress difference in each direction is calculated using the three-dimensional spatial coordinates between the node and its neighboring nodes, and divided by the corresponding spatial distance to achieve a normalized rate of change of stress along the spatial direction, thus obtaining the stress gradient vector of the node. This method relies on existing rock mechanics theory, three-dimensional spatial coordinate calculation, and numerical difference or vectorization processing, and can be directly implemented within the existing computational framework.

[0041] Based on the law of conservation of energy, after disregarding the attenuation and dissipation, the net effective information entropy truly used for subsequent network propagation is extracted. .

[0042] in, This represents the sum of energy losses on all adjacent entity relation edges, where each entity relation edge represents the sum of energy losses on all adjacent entity relation edges. Energy loss is expressed as ; Represents entity relationship edges Energy loss, through entity relationship edges Energy loss rate and Multiply them to get the result.

[0043] An autoregressive cumulative truncation rule is adopted, using the energy loss as the sole driving source of disruptive work, to update entity relation edges. Current hidden tearing degree .

[0044] in Represents entity relationship edges The percentage of energy used to tear apart connections.

[0045] To accurately simulate the propagation, dissipation, and destruction of blasting energy in defective rock masses on a discrete geological knowledge graph, a Bayesian network is constructed. This network uses the implicit tearing degree of adjacent edges, node stress gradient, entropy flow through nodes, and rock mass type as inputs. This allows for probabilistic reasoning of the energy loss rate and the proportion of energy used for tearing on each adjacent edge. This enables dynamic calculation of energy transmission and distribution within the graph at each time step, ensuring that energy flow, local damage accumulation, and adaptive path adjustment conform to physical laws. The calculation of node stress gradients incorporates gravitational stress generated by the rock mass's own weight, local structural features, and node spatial location, accurately reflecting the direction and intensity of local stress concentration. This makes the energy loss and tearing probabilities output by the Bayesian network more physically plausible. By applying the principle of energy conservation, the net effective entropy actually used for propagation is extracted, and energy loss is used as the sole driving force for destruction. The implicit tearing degree of edges is updated according to an autoregressive cumulative truncation rule, ensuring that the damage evolution of each edge accurately reflects multimodal effects such as local fracture dissipation, plastic absorption, and damped oscillations. This not only ensures that the nonlinear and irreversible evolution of blasting energy in the topological space is quantifiable and traceable, but also enables the physical-driven update of the latent tearing degree. It provides a highly accurate and physically reasonable dynamic basis for entropy flow adaptive propagation, path selection, and final block size prediction, significantly improving the accuracy and engineering applicability of blasting effect analysis.

[0046] Use the updated entity relationship edge Given the current latent tearing degree, and simulating the energy release to avoid a high-resistance guiding mechanism, the adjusted non-normalized transfer weights are calculated: .

[0047] For nodes All adjacent edges The non-normalized transition weights are subjected to global normalization, and the formal entropy flow transition probability of the next time step is output. .

[0048] in, Represents entity relationship edges The base transition probability (initially the same for each edge). This represents a constant characterizing the sensitivity of a fracture to energy blocking. It is the set of adjacent nodes.

[0049] It's important to note that the inhibitory effect of cracks or damage on energy conduction exhibits a non-linear variation: initially, when the crack is minor, energy can still be transferred smoothly, but as the degree of latent tearing increases, the conduction resistance rises rapidly, eventually potentially approaching complete blockage. The exponential function naturally reflects this physical characteristic of "slow initial decay followed by rapid decline," while ensuring positive weights to maintain reasonable physical meaning in graph propagation. Furthermore, the influence of resistance on energy transfer can be controlled by adjusting the sensitivity coefficient β. Subsequently, global normalization is applied to all adjacent edges of node u to ensure the conservation of the total information entropy output of the node, avoiding unnecessary energy loss. Simultaneously, a competitive relationship is introduced between adjacent edges, causing the information entropy flow to automatically favor edges with lower resistance and less damage, achieving adaptive guidance. The normalized formal transition probability not only reflects the current state of the edge but also embodies the influence of the local environment and damage on energy path selection. It can simulate the nonlinear blocking and adaptive guiding effect of cracks on energy conduction at the algorithm level, so that energy propagates preferentially along low-resistivity channels while ensuring energy conservation, providing a physically reasonable and quantifiable basis for updating the latent tear degree, adjusting the entropy flow path, and predicting the subsequent blasting block size.

[0050] The net effective information entropy and the formal entropy flow transition probability are used to... Multiplying, the information entropy flow that actually flows into each adjacent entity relationship edge and propagates downstream is obtained, driving the operator to complete the entropy flow adaptive transfer in a single time step; after completing the calculation of the current time step, the next time step is calculated through the Markov entropy flow graph operator.

[0051] As a deep reconstruction of the graph operator transmission mechanism, an energy intelligent guidance mechanism based on the "positive feedback avalanche effect" is implemented. In terms of implementation, by constructing the implicit tearing degree and the transfer probability into a positive exponential incentive relationship, the operator is forced to prioritize the allocation of subsequent net effective information entropy to the entity relation edges with higher damage degree. This technique perfectly and equivalently maps the "crack tip stress concentration effect" in real fracture mechanics and the "high-pressure detonation gas wedge splitting effect" in explosion engineering at the bottom layer of the computer discrete operator. That is, the more developed the microcrack channel, the easier it is to attract energy convergence. Thus, under the premise of energy conservation, it simulates the physical multibody dynamic evolution process of the incubation, directional expansion and final connection of the macroscopic main crack with extremely high accuracy.

[0052] S4: By monitoring the total information entropy accumulated by each rock mass node, when the total information entropy of any edge exceeds the critical fracture tolerance determined by the equivalent mapping based on its static mechanical properties, an explicit topological tearing mechanism is triggered to forcibly disconnect the entity relationship edges in the initial geological knowledge graph.

[0053] Furthermore, the entity relationship edges that have fractured are simultaneously reconstructed into virtual free surface sink nodes at the bottom layer of the algorithm. The unidirectional gravitational transfer probability of the original adjacent rock mass nodes pointing to the virtual free surface sink node is constructed to forcibly change the local energy conduction gradient. Thus, in the discrete graph network, the real physical effects of stress waves and detonation gases expanding and unloading onto the new free surface after the physical fracture of the continuous medium generates new cracks are perfectly equivalently simulated.

[0054] Specifically, the critical fracture tolerance includes traversing the initial geological knowledge graph, extracting the uniaxial tensile strength and rock elastic modulus recorded at the first and second rock mass nodes connected at both ends of the target entity relationship edge, as well as the maximum and minimum principal stresses at the spatial location of the entity relationship edge, and extracting the equivalent physical volume of the geometric mapping of the entity relationship edge. In this invention, the various rock mass mechanical parameters and spatial quantities upon which the critical fracture tolerance depends all originate from the physical properties of the rock mass itself and the true mapping of the geological environment. The uniaxial tensile strength and rock elastic modulus reflect the bearing limit and elastic response of the rock mass under external forces, which can be obtained through laboratory core tests, in-situ measurements, or geological data inversion, truly depicting the mechanical nature of the rock mass material. The maximum and minimum principal stresses at the node location reflect the internal stress state of the rock mass under the natural geostress field, which can be obtained through field pressure measurement, strain gauge recording, and regional geomechanical models, and their values ​​determine the failure sensitivity of the local rock mass. The equivalent physical volume of the entity relationship edges maps the topological relationships abstracted from the graph back to the physical space, comprehensively reflecting the actual volume and energy conduction capacity of the rock mass units connected by the edges. By integrating these physical quantities, the critical fracture tolerance not only becomes a parameter for algorithm calculation, but also carries the inherent physical logic of rock mass failure. This makes the fracture judgment of each edge both quantifiable and profoundly reflects the mechanical nature of the rock mass under blasting loads, providing a solid physical foundation for the propagation of information entropy flow, the accumulation of hidden tearing, and the final block size prediction.

[0055] Based on the maximum distortion strain energy density theory, the intrinsic tensile strain energy of the first and second rock mass nodes is calculated separately, and the minimum value of the two is selected as the interface limit tensile strain energy of the solid relationship edge, following the weak surface fracture rule. Firstly, the maximum distortion strain energy density theory posits that local failure of rock mass often occurs in the region of maximum strain energy density; that is, under a given load, fracture or failure will occur when the locally absorbed distortion strain energy of the rock mass reaches its limit. Therefore, by calculating the intrinsic tensile strain energy that a node can withstand, the energy limit of each node under tension can be quantified, thus reflecting the node's own failure potential.

[0056] In this scheme, the intrinsic tensile strain energy is calculated for both ends of the entity relationship edge (the first rock mass node and the second rock mass node), which is actually to consider the bidirectional stress state of the edge connection. Because the two rock masses connected by an edge may have different mechanical properties or different stresses, their respective failure limits are also different. According to the weak-plane fracture rule, the failure tendency always occurs first on the weaker or more stressed side. This reflects the local rock mass mechanical differences, making the edge fracture condition closely related to the mechanical limit of the node itself, and also ensures that the algorithm prioritizes the fracture of the weakest node when the energy distribution is uneven. This achieves the physical quantification of edge fracture, providing a reliable dynamic basis for subsequent implicit tear degree updates, entropy flow propagation, and blasting block size prediction.

[0057] Extract the maximum and minimum principal stresses, calculate the ratio of the square of the principal stress difference to the rock elastic modulus, and generate the formation deviatoric stress work compensation term.

[0058] The interface ultimate tensile strain energy is algebraically superimposed with the work compensation term of the formation deviatoric stress, and multiplied by the dimension transformation constant that maps the mechanical Joule dimension to the information theory entropy unit and the equivalent physical volume to calculate the absolute critical fracture tolerance of the entity relation edge.

[0059] As the underlying algorithm for phase transition judgment criteria, this technique involves cross-disciplinary algebraic superposition mapping of the interfacial ultimate tensile strain energy theory in solid mechanics with the deep geostress work compensation term. This technique strictly anchors the phase transition judgment subject to the "relationship edge" and selects the weakest mechanical property of the interface for energy redline marking according to the "barrel effect" in fracture mechanics. It not only extremely accurately distinguishes the differences in physical fracture difficulty of different rock layer interfaces under different high confining pressure environments, but also provides rigorous computational mechanics judgment support for the explicit fracture at the edge of the map.

[0060] It is worth noting that, in order to rigorously quantify the rock mass failure behavior during blasting and establish clear physical boundaries at the algorithm level, ensuring that the fracture determination of each edge closely corresponds to the mechanical limits of the actual rock mass, parameters such as intrinsic tensile strain energy, work done by formation deviatoric stress, and equivalent volume are introduced for calculation. This allows for a keen distinction between the differences in fracture difficulty at different rock strata interfaces and under high confining pressure environments, thus ensuring that the algorithm responds reasonably to local weak surfaces or mechanical differences. Simultaneously, this design provides a rigorous mechanical basis for judging explicit fractures in the graph, making energy conduction and damage accumulation processes quantifiable and traceable, and providing highly reliable physical support for subsequent information entropy flow adjustment and block size prediction, achieving accurate equivalent simulation of the failure behavior of continuous media using discrete graph networks.

[0061] S5: After the information entropy dissipation in the network terminates, extract the binary discrete graph topology that survives the explicit topological tearing mechanism. Based on the graph theory connected component features, decode and map the mutually isolated discrete graph subsets into independent rock fragments in physical three-dimensional space, and output the engineering-grade blasting block size analysis results.

[0062] Furthermore, the binarized discrete graph topology includes extracting the edge relationships after information dissipation cutoff in the network, performing network clustering on the set of rock mass nodes connected by all edges, and dividing them into U mutually isolated discrete graph subsets.

[0063] Traverse the e-th discrete graph subset For all rock mass nodes within the area, sum the equivalent physical volumes of each node algebraically to obtain the total predicted volume of the e-th independent rock fragment: in, represents the e-th discrete map subset; r represents the index of the rock mass node within the map subset; denoted by , represents the equivalent physical volume of the r-th rock mass node within the subset of the graph; e represents the index of any discrete subset of the U discrete subsets of the graph.

[0064] To transform the discrete graph connectivity results obtained from the algorithm into blasting fragment size information usable in engineering, the predicted volume of each individual rock fragment is grouped according to a preset standard sieve size range. This quantifies the proportional distribution of rocks with different particle sizes. Further statistical analysis of the cumulative mass percentage of each range reflects the proportion and distribution pattern of fragments within the overall rock mass. Fitting these data to generate a blasting fragment size distribution curve not only provides an intuitive basis for engineering judgment but also closely integrates the abstract topological results of the graph calculation with actual construction needs. This achieves a complete mapping from microscopic fracture evolution and localized damage accumulation to macroscopic fragment size distribution, providing a reliable decision-making basis for blasting design, construction optimization, and subsequent loading and processing.

[0065] It abandons the complex post-processing of flying rock collision simulation and post-processing that relies heavily on multibody mechanics ultimate shear force in engineering; it applies the breadth-first search (BFS) algorithm in graph theory to find network connectivity islands that have not been completely cut off by torn edges and to perform equivalent node volume superposition; it reverse-translates the clusters that still maintain topological connectivity after edge breakage into three-dimensional physical rock gradation data of interest in field engineering in an extremely lightweight and accurate manner, completing a perfect physical causal closed loop from "edge breakage" to "block generation".

[0066] This embodiment also provides a knowledge graph-based explosive charge analysis system, including: The graph construction unit discretizes the rock mass to be blasted into graph nodes, and maps the spatial distribution characteristics of the original micro-fractures in the rock mass to the initial implicit tearing degree on the entity relationship edge to construct an initial geological knowledge graph; after obtaining the blasting charge, the detonation energy is mapped to the initial information entropy impulse of the injected node.

[0067] The computing unit uses a Markov entropy flow graph operator with damage adaptive routing function to drive the initial information entropy impulse to be propagated along the entity relationship edges in the network, and dynamically updates the current implicit tear degree of the corresponding edge by accumulating the information entropy flow through each entity relationship edge, and uses the updated current implicit tear degree as the independent variable to adjust the entropy flow transfer of the entity relationship edge.

[0068] The analysis unit monitors the total information entropy accumulated by each rock mass node. When the total information entropy of any edge exceeds the critical fracture tolerance determined by the equivalent mapping based on its static mechanical properties, it triggers an explicit topological tearing mechanism to forcibly disconnect the entity relationship edges in the initial geological knowledge graph.

[0069] The output unit extracts the binary discrete graph topology that survives the explicit topological tearing mechanism after the information entropy dissipation in the network terminates. Based on the graph theory connected component features, it reverse-decodes and maps the mutually isolated discrete graph subsets into independent rock fragments in physical three-dimensional space, and outputs the engineering-grade blasting block size analysis results.

[0070] This embodiment also provides a computer device applicable to the knowledge graph-based explosive charge analysis method, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the knowledge graph-based explosive charge analysis method proposed in the above embodiment.

[0071] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0072] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the knowledge graph-based explosive charge analysis method proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0073] In summary, this invention achieves quantitative simulation of local energy conduction and damage accumulation by: discretizing the blasted rock mass into knowledge graph nodes and mapping the original fractures to the implicit tearing degree of the edges; adaptively guiding the propagation of information entropy flow in the graph using Markov entropy flow operators and Bayesian networks to reflect the dynamic regulation of energy paths by fractures and damage; mapping local damage results to fractures of graph edges through critical fracture tolerance judgment and explicit topological tearing mechanism, and controlling the energy gradient through virtual free surface nodes to achieve equivalent reproduction of continuous medium fracture; finally, calculating the volume of independent rock fragments based on the connected components of the residual graph, and grouping and accumulating statistics in combination with preset sieve size ranges to fit and generate a blasting block size distribution curve, thereby completely simulating the entire process from microscopic fracture evolution to macroscopic fragment distribution at the algorithm level, providing a high-precision and quantifiable physical basis for blasting design and construction.

[0074] It should be noted that the above embodiments are only used to illustrate the technical solutions 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 solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for analyzing explosive charges based on knowledge graphs, characterized in that: This includes discretizing the rock mass to be blasted into map nodes, reducing the spatial distribution characteristics of the original microfractures in the rock mass to an initial implicit tear degree on the entity relationship edge, and constructing an initial geological knowledge map; After obtaining the explosive charge, the detonation energy is mapped to the initial information entropy impulse of the injected node; Using a Markov entropy flow graph operator with damage adaptive routing capabilities, the initial information entropy impulse is driven to propagate along the entity relationship edges in the network. The current implicit tear degree of the corresponding edge is dynamically updated by accumulating the information entropy flow through each entity relationship edge. The updated current implicit tear degree is used as an independent variable to adjust the entropy flow transfer of the entity relationship edges. By monitoring the total information entropy accumulated by each rock mass node, when the total information entropy of any edge exceeds the critical fracture tolerance determined by the equivalent mapping based on its static mechanical properties, an explicit topological tearing mechanism is triggered, forcibly disconnecting the entity relationship edges in the initial geological knowledge graph. After the information entropy dissipation in the network terminates, the remaining binary discrete graph topology after the explicit topological tearing mechanism is extracted. Based on the graph theory connected component features, the mutually isolated discrete graph subsets are decoded and mapped into independent rock fragments in physical three-dimensional space, and the engineering-grade blasting block size analysis results are output.

2. The knowledge graph-based explosive charge analysis method as described in claim 1, characterized in that: The step of dimensionality reduction and mapping of the spatial distribution characteristics of primary microfractures in rock mass to the initial implicit tearing degree on the edge of the entity relationship includes extracting the three-dimensional point cloud data of the area to be analyzed during drilling and reconstructing the geometric set of three-dimensional fracture surfaces using a triangulation algorithm. In the initial geological knowledge map, adjacent rock mass nodes are connected. and The spatial line segment is the central axis, and the radius is [missing information]. The influence envelope cylinder is used to filter out the set of local crack surfaces that intersect with the influence envelope cylinder; Calculate the sum of the orthographic projection areas of the set of local fracture surfaces in the direction perpendicular to the central axis. ; Using the enveloping cylinder as a short-path propagation path for explosive energy, and taking the proportion of cracks in the cross-section of the path as a criterion, the initial latent tear strength is quantified. in, Represents entity relationship edges The initial latent tear degree.

3. The knowledge graph-based explosive charge analysis method as described in claim 2, characterized in that: The initial information entropy impulse includes the extraction of the first [unclear] in the loading scheme. Explosive charge per blast hole Explosive calorific value per kilogram And the set absolute time for the micro-delay detonation of the detonator. Constructing a discrete inference time step Discrete-time window pulse release function as independent variable : within the time interval The value inside is Its value is 0 in other intervals; compute nodes exist Initial information entropy impulse at time: ; Within the corresponding time interval, continuously add data to the initial geological knowledge map. The node corresponding to each blast hole position is injected with the initial information entropy impulse; in, This refers to the time step for a single simulation. This is the energy-information equivalent constant; This represents the number of time steps obtained by discretizing the effective duration of the detonation effect, used to characterize the range of step lengths in which the detonation energy of the corresponding borehole is continuously injected during the pattern extrapolation process.

4. The knowledge graph-based explosive charge analysis method as described in claim 3, characterized in that: The Markov entropy flow graph operator is a dynamic message passing model without aftereffects built on a discrete geological knowledge graph, internally encapsulating a pre-trained Bayesian network. The dynamic cumulative update of the implicit tear degree includes, in the extrapolation time step, extracting the absolute value of the information entropy flow of the current time step node. ; The Bayesian network is invoked to obtain the current hidden tear degree and node stress gradient of all nodes adjacent to the current node. And rock mass type as a feature input matrix; Output the energy loss rate for each adjacent entity relationship edge, and the proportion of energy used by each entity relationship edge to tear the connection; Represents entity relationship edges The probability of energy doing work; Net effective information entropy of subsequent network propagation ; in, This represents the sum of energy losses on all adjacent entity relation edges, where each entity relation edge represents the sum of energy losses on all adjacent entity relation edges. Energy loss is expressed as ; Represents entity relationship edges Energy loss, through entity relationship edges Energy loss rate and Multiplying them together yields the result; An autoregressive cumulative truncation rule is adopted, using the energy loss as the sole driving source of disruptive work, to update entity relation edges. Current hidden tearing degree ; in Represents entity relationship edges The percentage of energy used to tear apart connections.

5. The knowledge graph-based explosive charge analysis method as described in claim 4, characterized in that: Use the updated entity relationship edge Given the current latent tearing degree, and simulating the energy release to avoid a high-resistance guiding mechanism, the adjusted non-normalized transfer weights are calculated: ; For nodes All adjacent edges The non-normalized transition weights are subjected to global normalization, and the formal entropy flow transition probability of the next time step is output. ; in, Represents entity relationship edges The basic transition probability, This represents a constant characterizing the sensitivity of a fracture to energy blocking. It is the set of adjacent nodes; The net effective information entropy and the formal entropy flow transition probability are used to... Multiplying, the information entropy flow that actually flows into each adjacent entity relationship edge and propagates downstream is obtained, driving the operator to complete the entropy flow adaptive transfer in a single time step; after completing the calculation of the current time step, the next time step is calculated through the Markov entropy flow graph operator.

6. The knowledge graph-based explosive charge analysis method as described in claim 5, characterized in that: Traverse the initial geological knowledge graph, extract the uniaxial tensile strength and rock elastic modulus recorded by the first rock mass node and the second rock mass node connected at both ends of the target entity relationship edge, as well as the maximum principal stress and minimum principal stress at the spatial location of the entity relationship edge, and extract the equivalent physical volume of the geometric mapping of the entity relationship edge; Based on the maximum distortion strain energy density theory, the intrinsic tensile strain energy of the first rock mass node and the second rock mass node are calculated respectively, and the minimum value of the two is selected as the interface ultimate tensile strain energy of the solid relationship edge according to the weak surface fracture rule. Extract the maximum and minimum principal stresses, calculate the ratio of the square of the principal stress difference to the rock elastic modulus, and generate the formation deviatoric stress work compensation term. The interface ultimate tensile strain energy is algebraically superimposed with the work compensation term of the formation deviatoric stress, and multiplied by the dimension transformation constant that maps the mechanical Joule dimension to the information theory entropy unit and the equivalent physical volume to calculate the absolute critical fracture tolerance of the entity relation edge.

7. The knowledge graph-based explosive charge analysis method as described in claim 6, characterized in that: The binarized discrete graph topology includes extracting the edge relationships after information dissipation cutoff in the network, performing network clustering on the set of rock mass nodes connected by all edges, and dividing them into U mutually isolated discrete graph subsets. Traverse the e-th discrete graph subset For all rock mass nodes within the area, sum the equivalent physical volumes of each node algebraically to obtain the total predicted volume of the e-th independent rock fragment: ； in, represents the e-th discrete map subset; r represents the index of the rock mass node within the map subset; represents the equivalent physical volume of the r-th rock mass node within the subset of the graph; e represents the index of any discrete subset of the U discrete subsets of the graph. Based on the preset standard sieve size range, the total predicted volume of all independently generated rock fragments is grouped by range and the cumulative mass ratio is statistically analyzed, and the blasting block size distribution curve is fitted and output.

8. A knowledge graph-based explosive charge analysis system, based on the knowledge graph-based explosive charge analysis method according to any one of claims 1 to 7, characterized in that: This includes a graph construction unit that discretizes the rock mass to be blasted into graph nodes, reduces the spatial distribution characteristics of the original microfractures in the rock mass to an initial implicit tear degree on the entity relationship edge, and constructs an initial geological knowledge graph; after obtaining the blasting charge, the detonation energy is mapped to the initial information entropy impulse of the injected node; The computing unit uses a Markov entropy flow graph operator with damage adaptive routing function to drive the initial information entropy impulse to be propagated along the entity relationship edge in the network, and dynamically accumulates and updates the current implicit tear degree of the corresponding edge by the information entropy flow flowing through each entity relationship edge, and uses the updated current implicit tear degree as the independent variable to adjust the entropy flow transfer of the entity relationship edge. The analysis unit monitors the total information entropy accumulated by each rock mass node. When the total information entropy of any edge exceeds the critical fracture tolerance determined by the equivalent mapping based on its static mechanical properties, it triggers an explicit topological tearing mechanism to forcibly disconnect the entity relationship edges in the initial geological knowledge graph. The output unit extracts the binary discrete graph topology that survives the explicit topological tearing mechanism after the information entropy dissipation in the network terminates. Based on the graph theory connected component features, it reverse-decodes and maps the mutually isolated discrete graph subsets into independent rock fragments in physical three-dimensional space, and outputs the engineering-grade blasting block size analysis results.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the knowledge graph-based explosive charge analysis method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the knowledge graph-based explosive charge analysis method according to any one of claims 1 to 7.

Images