An engineering geology regional disaster early warning method and system based on digital twinning

Abstract

The application relates to the technical field of digital twinning and disaster early warning, in particular to an engineering geological regional disaster early warning method and system based on digital twinning; the method comprises the following steps: constructing a structure evolution type digital twinning body, dividing structure units with homogeneous mechanical response, establishing a directional stress propagation matrix and a coupled dynamic evolution equation, simulating the mechanical state and evolution process of regional geological bodies; constructing a disturbance trigger chain propagation matrix and calculating the trigger chain strength and evolution rate of each unit, identifying a potential chain instability path and realizing hierarchical early warning in combination with a dynamic threshold model; constructing a structure safety redundancy dynamic attenuation model, identifying an evolution acceleration area and establishing a structure instability evolution index for multi-physical quantity fusion evaluation; actively applying a virtual micro disturbance, determining an irreversible evolution stage by analyzing the nonlinear amplification and residual offset of the redundancy response, and forming a closed-loop dynamic early warning mechanism. The application realizes high-reliability disaster early warning from trend prediction to active verification.

Description

Technical Field

[0001] This invention relates to the field of digital twin and disaster early warning technology, specifically to a method and system for early warning of engineering geological areas based on digital twins. Background Technology

[0002] Early warning of geological disasters in engineering areas is an important technical field to ensure the safety of infrastructure construction and operation. As my country's transportation, water conservancy and urbanization construction continues to extend to areas with complex geological conditions, the hidden, sudden and chain-like characteristics of geological disasters have put forward higher requirements for early warning technology. At present, engineering geological monitoring and early warning technology has developed rapidly and has been widely used in various geotechnical engineering practices. Automated monitoring methods based on sensor networks can acquire rich stress, displacement and hydrological data. Combined with engineering geological survey results and historical disaster data, an early warning workflow based on data collection and threshold judgment has been formed.

[0003] Chinese invention patent application CN121009686A discloses a geological disaster simulation method and system based on a digital twin simulation platform. By employing a dynamic weighted fusion algorithm combined with information entropy and reliability scoring, the accuracy and robustness of data fusion are improved. A digital twin model is constructed based on multi-source data inversion and 3D geological modeling technology, updating geological parameters in real time to reflect changes in the geological body's state. A multi-level, tiered early warning model is introduced, integrating physical models and deep learning predictions to enhance the timeliness and foresight of early warnings. The accuracy of the tiered disaster simulation model is dynamically selected, and a risk heat map is generated, improving the spatial precision and decision support capabilities for geological disaster risk identification.

[0004] Current engineering practice relies mainly on field monitoring and indoor tests to understand the mechanical behavior of geological bodies. However, due to limitations in the density of monitoring points and testing conditions, it is difficult to fully capture the continuous evolution trajectory of the internal state parameters of geological structures. The rise of digital twin technology has provided new possibilities for realizing the full-process simulation and dynamic mapping of the mechanical behavior of geological bodies. By constructing a virtual model that is linked with the physical entity in real time, it is expected to break through the spatiotemporal limitations of traditional analysis methods and open up a more scientific technical path for disaster early warning in engineering geological areas. Summary of the Invention

[0005] The purpose of this invention is to address the problems existing in the background technology by proposing a method and system for early warning of engineering geological regional disasters based on digital twins.

[0006] The technical solution of this invention: a method for early warning of engineering geological regional disasters based on digital twins, comprising the following specific implementation steps:

[0007] S1. Construct a structural evolution-type regional geological digital twin. By dividing the structural units according to mechanical consistency, establishing the directional stress propagation matrix between units, constructing coupled dynamic evolution equations and combining adaptive error feedback, the calculable evolution and real-time synchronous update of the regional geological body can be realized.

[0008] S2. Based on the constructed structural evolution-type regional geological digital twin, multi-source natural and anthropogenic disturbances are uniformly quantified as inputs. By coupling the relationship between structural integrity and asymmetric stress propagation, a disaster trigger chain propagation matrix is ​​constructed. The trigger chain strength and its time evolution rate of each structural unit are calculated, potential chain instability paths are identified, and a dynamic threshold mechanism based on regional statistical characteristics is adopted to achieve hierarchical early warning output.

[0009] S3. Based on the obtained trigger chain propagation strength and evolution trend, a dynamic decay model of structural safety redundancy is constructed. By establishing a continuous redundancy function, time derivative analysis and spatial gradient identification mechanism, the regional structural safety reserve is transformed from static evaluation to a differentiable evolutionary variable. Trigger chain strength coupling correction is introduced to achieve accurate identification of the accelerated dissipation region of structural redundancy. Finally, a structural instability evolution index is constructed to output graded risk results.

[0010] S4. Based on the identified accelerated structural evolution zone, by actively applying controlled micro-perturbations within the digital twin, the nonlinear amplification and residual offset of the structural redundancy response are analyzed to identify irreversible evolution stages and form a closed-loop dynamic early warning mechanism, thereby realizing disaster early warning from trend judgment to active verification.

[0011] Preferably, the process of constructing a structural evolution-type regional geological digital twin in step S1 includes:

[0012] Based on the stratigraphic structure, geophysical parameters and stress gradient, the region is divided into mechanically responsive homogeneous structural units. Each unit is defined with an extended state vector containing stress, strain, shear stress, pore water pressure and damage index.

[0013] An asymmetric stress transfer topology matrix is ​​constructed based on structural units. Combining unit contact area, mechanical distance, attenuation coefficient and directional correction factor, the mechanical coupling propagation between adjacent units is simulated.

[0014] A coupled dynamic evolution equation for the regional structure is constructed, integrating the nonlinear elastoplastic response within the unit, the influence of neighborhood coupling, and the input of external disturbances, to realize the time-varying evolution calculation of the structural state of the entire region;

[0015] An adaptive error feedback mechanism is introduced to compare the measured monitoring data with the model prediction results and make corrections through a dynamic gain matrix, thereby achieving synchronous updates between the digital twin and the actual engineering status.

[0016] Preferably, the partitioning condition for homogeneous structural units with mechanical response is:

[0017] Under given historical perturbation conditions, the rate of change of stress response gradient meets the preset allowable threshold for stress gradient and the rate of change of rock physical parameters meets the preset allowable threshold for parameter change.

[0018] The damage index in the extended state vector is obtained by integrating the strain history. When the damage index reaches the material's ultimate damage value, it indicates that the element is close to failure.

[0019] The stress transfer coefficients in the asymmetric stress transfer topology matrix are determined based on the structural contact weighting coefficient, the effective mechanical distance between the centroids of the elements, the directional correction factor, and the stress propagation attenuation coefficient.

[0020] The directional correction factor is determined by the cosine of the angle between the principal stress direction and the line connecting the elements.

[0021] Preferably, the process of constructing the catastrophe triggering chain propagation matrix in step S2 by coupling the relationship between structural integrity and asymmetric stress propagation includes:

[0022] Multi-source disturbances such as rainfall, earthquakes, and construction are classified and modeled. Different physical quantities are normalized by historical extreme values ​​and uniformly quantified by weighting coefficients to construct unit-level disturbance input vectors.

[0023] Based on the established stress transfer matrix between structural units, a structural integrity modulation function, distance attenuation, and direction correction factor are introduced to construct a disturbance trigger chain propagation matrix, so that the disturbance propagation intensity is simultaneously affected by the unit degradation state and mechanical direction.

[0024] By calculating the trigger chain strength index and its time change rate of each structural unit using the propagation matrix, a directed graph of the structural network is constructed to identify regions of accelerated growth of trigger chains and potential catastrophic paths.

[0025] A dynamic threshold model is constructed based on the mean and standard deviation of the trigger chain strength across the entire region. The warning levels are divided by combining the chain length and evolution rate. When the unit strength exceeds the adaptive threshold, the graded warning results are output.

[0026] Preferably, the structural integrity modulation function is constructed based on the structural integrity index of the unit. The closer the integrity index is to zero, the more severe the damage. The modulation function determines the degree to which the integrity change amplifies the propagation intensity through the damage sensitivity adjustment coefficient.

[0027] The trigger chain strength index is calculated by accumulating the propagation contribution from other units through the propagation matrix, and the identification of the accelerated growth region of the trigger chain is based on time series analysis of the strength evolution rate;

[0028] In the dynamic threshold model, the adaptive threshold is the sum of the mean strength of the trigger chain across the entire region and the adjustment coefficient multiplied by the standard deviation.

[0029] Preferably, the process of constructing the structural safety redundancy dynamic attenuation model in step S3 includes:

[0030] Based on the real-time stress field and pore water pressure field of digital twin structural units, a dynamic effective shear strength model is constructed, and the equivalent shear stress of the unit is calculated. A continuous structure safety redundancy function is defined, and the safety state of discrete units is mapped to a spatial continuous redundancy field.

[0031] The time derivative of the structural safety redundancy is calculated, and the redundancy decay rate and decay acceleration are obtained respectively to identify whether there is an accelerated dissipation trend of redundancy. At the same time, the obtained trigger chain strength is introduced to construct an exponential correction redundancy function so that the safety reserve reflects the influence of internal stress degradation and external disturbance propagation.

[0032] A spatial gradient function is constructed based on a continuous redundant field to identify redundant abrupt change regions and potential slip concentration zones. An evolution concentration index is constructed through the regional average decay rate to determine whether a unit belongs to an accelerated degradation zone.

[0033] A structural instability evolution index is constructed, which includes corrected redundancy, decay rate, decay acceleration and trigger chain strength. The weight parameters are calibrated using historical sample data to achieve multi-physical variable fusion evaluation, and risk classification is output based on the index change trend and critical threshold.

[0034] Preferably, the physical meaning of the structural safety redundancy function is that a value greater than zero indicates the existence of a safety reserve, a value approaching zero indicates approaching the critical point, and a value less than zero indicates entering the theoretical instability region;

[0035] The redundant decay acceleration is calculated in discrete time form. When the redundant decay rate is less than the set decay rate threshold and the decay acceleration is less than the set decay acceleration threshold, it is determined to be accelerated decay.

[0036] The evolution concentration index is the ratio of the unit decay rate to the average decay rate of the region. Combined with three composite conditions—redundancy below a preset redundancy threshold, decay rate below a decay rate threshold, and decay acceleration below a decay acceleration threshold—the unit is determined to be in the structural evolution acceleration zone.

[0037] Preferably, the process of actively applying controlled micro-perturbations within the digital twin in step S4 includes:

[0038] Based on the output evolution acceleration zone element, construct a micro-perturbation state vector containing stress, pore pressure and strain, and use a hierarchical incremental loading strategy to perform virtual perturbation on the target element;

[0039] The change in redundancy of the unit is calculated under virtual disturbance, the disturbance response amplification factor and the sensitivity nonlinearity index are defined, and the nonlinear response trend is used to determine whether the unit has entered the high sensitivity stage.

[0040] After the disturbance is removed, the residual redundancy offset is calculated, and the irreversible evolution index is constructed by combining the nonlinear sensitivity index to determine whether the unit has entered the irreversible evolution stage. The regional critical evolution zone is identified by continuous high-value units.

[0041] A closed-loop dynamic early warning mechanism is constructed based on the regional critical zone. The early warning level is output through dual judgment, namely the evolution acceleration zone and the irreversible evolution stage. The dynamic adjustment and closed-loop feedback are realized by using the irreversible exponential change rate.

[0042] Preferably, the disturbance response amplification factor is the ratio of the redundancy correction before and after the disturbance is applied. If the ratio exceeds the set amplification threshold, it indicates that the structure exhibits a nonlinear amplification response.

[0043] The sensitive nonlinearity index is calculated based on the redundancy changes under different disturbance levels, and combined with the set sensitive threshold judgment unit, it enters the high-sensitivity nonlinearity stage;

[0044] The residual redundancy offset is the difference between the redundancy restored after the disturbance is removed and the redundancy before the disturbance. If the offset exceeds the set residual offset threshold, it indicates that there is cumulative damage to the structure.

[0045] The irreversible evolution index is constructed by fusing perturbation-sensitive weights and residual offset weights. If the set critical threshold is exceeded, the unit is determined to enter an irreversible evolution state.

[0046] When several consecutive units enter an irreversible evolutionary state, the region formed by these units is identified as the critical evolution zone.

[0047] The technical solution of this invention: A digital twin-based engineering geological regional disaster early warning system, used to execute the aforementioned digital twin-based engineering geological regional disaster early warning method, comprising:

[0048] The digital twin modeling and state update module is used to construct a structural evolution digital twin of an engineering geological area, perform comprehensive state modeling of regional geological units, integrate multi-source geological data, realize the dynamic evolution of the digital twin through a continuous time update mechanism, and generate state vectors of each unit in real time.

[0049] The perturbation chain analysis and dynamic evolution module is used to analyze the propagation path and impact range of potential perturbations within the digital twin, generate a trigger chain strength field, quantify the transmission impact of perturbations on each structural unit, process time series data to identify the starting point, acceleration zone and potential expansion region of perturbation propagation, and form a dynamic evolution state diagram.

[0050] The structural safety redundancy decay and evolution acceleration identification module is used to calculate the safety redundancy of each structural unit and its time evolution trend based on the digital twin. It identifies potential instability regions by redundancy decay rate, acceleration and spatial gradient, forms a continuous safety redundancy field and automatically identifies the evolution acceleration region, and generates a comprehensive instability evolution index.

[0051] The active perturbation injection and critical phase verification module is used to apply controlled micro-perturbations within the digital twin, simulate the nonlinear response of the unit under the perturbation, calculate the residual offset after the perturbation is removed, form an irreversible evolution index, identify highly sensitive units and regional critical zones to achieve closed-loop verification of the dual judgment mechanism, dynamically update the early warning level based on the evolution index, and output visualized early warning information.

[0052] Compared with the prior art, the above-mentioned technical solution of the present invention has the following beneficial technical effects:

[0053] This invention designs a method and system for early warning of regional disasters in engineering geology based on digital twins. By constructing a structurally evolving digital twin, the engineering geological body is transformed from a static model into a dynamic and computable system with mechanical evolution capabilities, realizing real-time synchronous updates and long-term stable simulation of regional geological conditions, providing a high-precision data foundation for disaster early warning. By establishing an asymmetric stress propagation matrix and coupled dynamic evolution equations, the directionality and attenuation characteristics of stress transmission within the geological body are realistically reflected, significantly improving the accuracy of simulating mechanical behavior under complex geological conditions. By constructing a disturbance trigger chain propagation matrix and introducing a structural integrity modulation function, unified quantification and chain propagation analysis of multi-source natural and anthropogenic disturbances are achieved, enabling the identification of potential instability paths and areas of accelerated growth of trigger chains, thus extending the early warning time from post-disaster. The response is brought forward to the pre-disaster trend identification stage; by establishing a structural safety redundancy dynamic decay model, the safety reserve is transformed from a static indicator into a differentiable evolutionary variable. Combined with redundancy decay rate, acceleration, and spatial gradient identification, the precise location of accelerated dissipation regions and potential slip surfaces is achieved; furthermore, by actively applying virtual micro-perturbations and analyzing the nonlinear amplification and residual offset of the redundant response, it is possible to effectively identify whether the structure has entered the irreversible evolution stage, forming a closed-loop dynamic early warning mechanism, which solves the limitation of traditional methods that can only passively monitor and cannot actively verify; this invention organically combines multi-physics coupling analysis, chain propagation modeling, and active verification technology to achieve continuous structured analysis from trend prediction to criticality determination, providing operable, verifiable, and dynamically updated intelligent decision support for engineering geological disaster prevention and control. Attached Figure Description

[0054] Figure 1 This is a flowchart of a method for early warning of engineering geological regional disasters based on digital twins proposed in this invention;

[0055] Figure 2 This is a system architecture diagram of an engineering geological regional disaster early warning system based on digital twin proposed in this invention. Detailed Implementation

[0056] Example 1, as Figure 1 As shown, the present invention proposes a method for early warning of engineering geological regional disasters based on digital twins, which includes the following specific implementation steps:

[0057] S1. Construct a structural evolution-based regional geological digital twin. By dividing the structural units according to mechanical consistency, establishing the directional stress propagation matrix between units, constructing coupled dynamic evolution equations, and combining adaptive error feedback, the computational evolution and real-time synchronous update of the regional geological body are realized. This provides basic data and structural models for disaster evolution analysis and early warning, ensuring the long-term stability, operability, and close resemblance to actual engineering conditions of the twin. The specific implementation process is as follows:

[0058] S11. Based on the stratigraphic structure, geophysical parameters, and stress gradient, the region is divided into structural units with homogeneous mechanical responses. Each unit defines an extended state vector, including stress, strain, shear stress, pore water pressure, and damage index, to achieve a calculable characterization of the mechanical state and material degradation within the unit. Specifically:

[0059] The principle of "homogeneous region of mechanical response" is adopted, that is, under given historical perturbation conditions, the rate of change of stress response gradient satisfies: ;

[0060] Simultaneously satisfying the rate of change of rock physical parameters: ;

[0061] When the above conditions are met within a certain region, it is divided into the same structural unit. ;

[0062] in, Represents the spatial stress gradient; This represents the allowable threshold for stress gradient, determined by historical monitoring statistics. Represents the spatial gradient of the elastic modulus; This indicates the allowable threshold for parameter changes;

[0063] Define an extended state vector for each structural unit:

[0064] ;

[0065] ;

[0066] when When this time, it indicates that the unit is close to destruction;

[0067] in, Indicates equivalent normal stress; Represents the equivalent shear stress; Indicates the equivalent total strain; Indicates pore water pressure; Indicates the structural integrity index; This represents the cumulative damage variable, obtained by integrating the strain history. Indicates the material's ultimate damage value;

[0068] S12. Based on the structural units, construct an asymmetric stress transfer topology matrix. Combine the unit contact area, mechanical distance, attenuation coefficient, and directionality correction factor to simulate the mechanical coupling propagation between adjacent units. This enables the digital twin to realistically reflect the directionality and attenuation characteristics of stress transfer within the region. Specifically:

[0069] Considering the directionality and non-uniformity of stress propagation in engineering geology, an asymmetric propagation matrix is ​​established:

[0070] ; ;

[0071] ;

[0072] The propagation intensity is maximum when the propagation direction is consistent with the principal stress direction; it is zero when the direction is opposite.

[0073] in, This represents the stress transfer matrix between structural elements; Indicates from unit Passed to The stress transfer coefficient; This represents the structural contact weighting coefficient, calculated based on the contact area of ​​the rock strata and the properties of the interface. This represents the effective mechanical distance between the centroids of the unit cells; Indicates the directional correction factor; This represents the stress propagation attenuation coefficient, which is determined by the rock mass's elastic modulus and damping characteristics. This represents the angle between the principal stress direction and the line connecting the elements, obtained through measuring the stress direction at the point of measurement or through finite element calculation.

[0074] S13. Construct a regional structural coupled dynamic evolution equation, integrating the nonlinear elastoplastic response within the unit, the influence of neighborhood coupling, and the input of external disturbances, to realize the time-varying evolution calculation of the structural state of the entire region, providing a dynamic basis for disaster path analysis, specifically:

[0075] Based on the above, a regional dynamics model is constructed:

[0076] ;

[0077] ;

[0078] in, The nonlinear evolution function of the element itself (elastoplastic + damage evolution) is defined based on the material constitutive model and historical damage characteristics; The function representing the influence of neighboring elements on this element is obtained through a coupled stress propagation model. This indicates external disturbance inputs, such as monitoring data or simulated inputs related to rainfall, construction loads, vibration, or groundwater fluctuations. This indicates the change in externally applied normal stress; This indicates the change in externally applied shear stress; This indicates changes in pore water pressure caused by rainfall or seepage.

[0079] S14. An adaptive error feedback mechanism is introduced, comparing the measured monitoring data with the model prediction results and correcting them through a dynamic gain matrix. This achieves synchronous updates between the digital twin and the actual engineering state, ensuring the long-term stable operation of the twin and its continuous close resemblance to the real state. Specifically:

[0080] To prevent model drift, a correction mechanism is constructed:

[0081] ;

[0082] Set the gain matrix update rules: ;

[0083] ;

[0084] in, This represents the predicted state vector of the i-th structural unit in the digital twin at the k-th time step. This represents the actual monitoring data state vector of the i-th structural unit at the k-th time step; This represents the dynamic gain matrix, which is the adaptive correction gain matrix of the i-th structural unit at the k-th time step. The initial value can be set as the identity matrix or an empirical value, and then iteratively updated through error feedback. This represents the state vector of the i-th structural unit after correction at the (k+1)-th time step. This represents the prediction error vector of the i-th unit at the k-th time step; This represents the gain update adjustment coefficient, which controls the sensitivity of the error to gain adjustment. It can be obtained through historical monitoring error statistics or model simulation calibration. This represents the gain matrix of the i-th structural unit after error feedback update, which is used for the next twin state correction, i.e. twin state correction at the k+1 time step. Its initial value can be set as the identity matrix or an empirical value.

[0085] S2. Based on the structural evolution-type regional geological digital twin constructed in step S1, multi-source natural and anthropogenic disturbances are uniformly quantified as inputs. By coupling structural integrity and asymmetric stress propagation relationships, a catastrophe trigger chain propagation matrix is ​​constructed. The trigger chain strength and its temporal evolution rate of each structural unit are calculated, potential chain-like instability paths are identified, and a dynamic threshold mechanism based on regional statistical characteristics is used to achieve hierarchical early warning output. This realizes a continuous structured analysis process from disturbance input to chain evolution and then to risk assessment. The specific implementation process is as follows:

[0086] S21. Classify and model multi-source disturbances such as rainfall, earthquakes, and construction. Normalize different physical quantities using historical extreme values ​​and combine them with weighting coefficients for unified quantification, constructing a unit-level disturbance input vector. Simultaneously, smooth out high-frequency noise through a time window, ensuring that the disturbance data can stably participate in subsequent propagation calculations within the digital twin network. Specifically:

[0087] Based on engineering geological monitoring and historical disaster data, disturbances are classified into three categories:

[0088] Natural disturbances: rainfall, groundwater level fluctuations, and seismic input;

[0089] Human disturbance: construction loads, blasting, excavation;

[0090] Coupled disturbances: The superposition of natural and anthropogenic disturbances affects the regional structure;

[0091] Each type of disturbance is defined as a time series input:

[0092] ;

[0093] To ensure that different types of perturbations can be computed simultaneously in the twin, a normalization process is defined:

[0094] ;

[0095] in, Representing structural units The perturbation vector; This represents the quantization value of the k-th type of disturbance; m represents the number of disturbance categories; This represents the normalized quantized value of the k-th type of perturbation; This represents the historical maximum disturbance value, ensuring amplitude standardization. The perturbation weights are determined based on historical statistical data and expert evaluation.

[0096] S22. Based on the stress transfer matrix between structural elements established in step S1, a structural integrity modulation function, distance attenuation, and direction correction factor are introduced to construct a disturbance triggering chain propagation matrix. This allows the disturbance propagation intensity to be simultaneously affected by the element degradation state and mechanical direction, achieving chain propagation modeling with physical constraint characteristics. Specifically:

[0097] The asymmetric stress propagation matrix and the stress transfer matrix between structural elements in step S1 Based on this, construct the perturbation trigger chain propagation matrix: ;

[0098] ;

[0099] ; ;

[0100] in, The trigger chain propagation matrix represents time t; Indicates from structural unit spread to Trigger chain strength; This represents the normalized perturbation input value, which is a unit. A standardized expression of the current disturbance intensity; The structural integrity index typically ranges from 0 to 1, with values ​​closer to 0 indicating more severe damage. Indicates the integrity modulation function; This represents the damage sensitivity adjustment coefficient, which determines the degree to which changes in integrity amplify the propagation intensity; Indicates the propagation correction function;

[0101] S23. Calculate the trigger chain strength index and its time rate of change for each structural unit using the propagation matrix, construct a directed graph of the structural network, identify regions of accelerated growth of trigger chains and potential catastrophic paths, and distinguish between local anomalies and chain-cascaded anomalies, specifically:

[0102] Through the propagation matrix Calculate the trigger chain strength for each unit: ;

[0103] Simultaneously calculate the intensity evolution rate: ;

[0104] For the entire region unit Time series analysis was performed to identify regions of accelerated growth in the trigger chain (highly sensitive areas). Through trigger chain connectivity analysis, potential catastrophic paths were extracted.

[0105] Constructing a directed graph ,node Let E be a cell, and let E be an edge. transfer;

[0106] Use depth-first search or shortest path algorithms to identify chain-like paths from disturbance sources to critical units;

[0107] in, Representation unit Current cumulative trigger chain strength; Indicates a pair from other units The propagation contribution; n represents the total number of result units; This indicates the rate of change of the trigger chain strength over time, used to determine whether it has entered the accelerated growth phase;

[0108] S24. A dynamic threshold model is constructed based on the mean and standard deviation of the trigger chain strength across the entire region. Early warning levels are determined by combining chain length and evolution rate. When the unit strength exceeds the adaptive threshold, a graded early warning result is output, realizing a risk assessment mechanism that dynamically adjusts with changes in the overall structural state. Specifically:

[0109] Define dynamic threshold: ;

[0110] For any unit ,like If so, it is determined that the state has entered a high-risk critical state;

[0111] in, This indicates the trigger chain strength after smoothing through a time window, used to eliminate short-period disturbance noise; Indicates the critical threshold of the unit; This represents the standard deviation of the trigger chain strength across the entire region; This represents the adjustment coefficient, which controls the sensitivity of the early warning system.

[0112] S3. Based on the trigger chain propagation strength and evolution trend obtained in step S2, a dynamic attenuation model of structural safety redundancy is constructed. By establishing a continuous redundancy function, time derivative analysis, and spatial gradient identification mechanism, the regional structural safety reserve is transformed from a static evaluation into a differentiable evolutionary variable. Trigger chain strength coupling correction is introduced to achieve accurate identification of the accelerated dissipation region of structural redundancy. Finally, a structural instability evolution index is constructed to output graded risk results, realizing the logical transition from disturbance propagation analysis to instability formation identification. The specific implementation process is as follows:

[0113] S31. Based on the real-time stress field and pore water pressure field of digital twin structural units, a dynamic effective shear strength model is constructed, and the equivalent shear stress of the unit is calculated. Then, a continuous structure safety redundancy function is defined, mapping the safety state of discrete units to a spatially continuous redundant field, specifically:

[0114] For each structural unit An improved effective stress shear model is adopted:

[0115] ;

[0116] Define the real-time equivalent shear stress: ;

[0117] And define the structural safety redundancy function: ;

[0118] Its physical meaning is:

[0119] No risk of stress;

[0120] There are safety reserves;

[0121] It approaches the critical point;

[0122] It enters the theoretical instability zone;

[0123] Mapping element redundancy values ​​to spatially continuous functions And weighted interpolation is used:

[0124] ;

[0125] in, Let represent the dynamic ultimate shear strength of the i-th element at time t, calculated using the improved Mohr-Coulomb effective stress model; The cohesion of the i-th unit is derived from indoor triaxial tests, in-situ shear tests, or historical survey databases. This represents the total normal stress of the i-th element at time t; The pore water pressure in the i-th cell at time t is derived from the hydrological coupling calculation model. The internal friction angle of the i-th element is derived from the geotechnical parameter database; This represents the actual equivalent shear stress of the i-th element at time t; Represents the shear stress component of the i-th element; Indicates the safety redundancy of the i-th unit; Represents the continuous space safety redundancy field function; This represents the spatial interpolation weight function. ;

[0126] S32. Calculate the time derivative of the structural safety redundancy, obtain the redundancy decay rate and decay acceleration respectively, identify whether there is an accelerated dissipation trend of redundancy, and simultaneously introduce the trigger chain strength obtained in step S2 to construct an exponentially modified redundancy function, so that the safety reserve simultaneously reflects the influence of internal stress degradation and external disturbance propagation, thereby forming a dynamically coupled redundancy decay dynamic model, specifically:

[0127] Calculate the redundancy decay rate: ;

[0128] It should be noted that, in discrete time: ;

[0129] when This indicates that safety reserves are decreasing;

[0130] Calculate the redundant decay acceleration: ;

[0131] Its discrete form is: ;

[0132] like: and This indicates that not only does redundancy decrease, but it is also "accelerating its decay";

[0133] Consider the trigger chain strength in step S2 Constructing corrected redundancy: When disturbance propagation intensifies, the equivalent redundancy automatically decreases;

[0134] in, Indicates the safety redundancy decay rate; Indicates safety redundancy decay acceleration; This indicates the time step, i.e., the sampling period of the monitoring system; This indicates the equivalent safety redundancy after trigger chain correction; This represents the trigger chain coupling amplification factor, calibrated based on historical disaster samples; Let (x, y, z) represent the k-th discrete-time sampling point; (x, y, z) represent the three-dimensional spatial coordinates.

[0135] S33. Based on a continuous redundant field, a spatial gradient function is constructed to identify redundant abrupt change regions and potential slip concentration zones. An evolution concentration index is then constructed using the regional average decay rate to determine whether a unit belongs to an accelerated degradation zone. Under the combined conditions of low redundancy, negative decay rate, and negative acceleration, it is labeled as a structural evolution acceleration zone, achieving spatial-scale risk cluster identification. Specifically:

[0136] Obtaining redundant spatial gradients If a region has redundancy below the set first threshold and gradient greater than the set second threshold, it indicates that the structural safety reserve has abruptly changed in space, which may form a potential slip surface or failure zone.

[0137] Define the evolutionary concentration index ;

[0138] like If the redundancy decays faster than the regional average, it indicates that the unit belongs to the "evolutionary concentration zone".

[0139] Combining the three conditions: , and If all conditions are met simultaneously, it is determined to be a "structural evolution acceleration zone";

[0140] in, This represents the gradient of the safety redundancy field space, i.e., the intensity of change in the redundancy space; This represents the evolutionary concentration index, which is the ratio of the unit decay rate to the regional average level. This represents the average absolute value of the decay rate of all cells within the region. Indicates the redundancy threshold; Indicates the decay rate threshold; Indicates the decay acceleration threshold;

[0141] S34. Construct a structural instability evolution index that includes corrected redundancy, decay rate, decay acceleration, and trigger chain strength. Weight parameters are calibrated using historical sample data to achieve multi-physical variable fusion assessment. Risk classification is output based on the index's changing trend and critical threshold, providing quantitative criteria for subsequent active disturbance verification and early warning issuance. Specifically:

[0142] To form an output-ready engineering criterion, a structural instability evolution index is constructed:

[0143] ;

[0144] in, Indicates the structural instability evolution index; , , and These represent the weight coefficients of each indicator, which are determined by regression calibration based on historical disaster data, satisfying the following conditions: ;

[0145] It should be noted that the structural instability evolution index is combined with the redundancy decay characteristics to classify units and regions. In low-risk areas, the index is lower than the regional average, indicating sufficient redundancy and slow decay. In medium-risk areas, the index is higher than the average but has not entered the accelerated evolution stage, suggesting that potential changes need to be monitored. In high-risk areas, the redundancy decay is significant and the trigger chain strength is increasing, indicating that the structural evolution is accelerating and may lead to local instability. In critical areas, the index continues to rise and the redundancy is close to zero, showing a clear trend of macro-instability. Preventive or intervention measures should be taken immediately to achieve dynamic, continuous, and operable engineering early warning.

[0146] S4. Based on the structural evolution acceleration zone identified in step S3, controlled micro-perturbations are actively applied within the digital twin to analyze the nonlinear amplification and residual offset of the structural redundancy response, identify the irreversible evolution stage, and form a closed-loop dynamic early warning mechanism to achieve highly reliable disaster early warning from trend judgment to active verification. The specific implementation process is as follows:

[0147] S41. Based on the evolution acceleration zone element output in step S3, construct a micro-perturbation state vector, including stress, pore pressure, and strain, and employ a graded incremental loading strategy to virtually perturb the target element in order to probe the structural response capability and critical load-bearing characteristics of potentially sensitive areas, specifically:

[0148] Based on the "accelerated structural evolution zone" and structural instability evolution index output in step S3 High-value units: Micro-perturbation injection is performed only in this region to ensure computational efficiency and focus on potentially dangerous cells;

[0149] For each unit Construct a perturbation vector: ;

[0150] Employing multi-level incremental perturbations: ;

[0151] in, This represents the set of units in the accelerated structural evolution region; The threshold for the evolution index can be determined by historical disaster data or expert experience. Represents the state vector of the micro-perturbation; This indicates minute stress disturbances; This indicates water pressure disturbance in minute pores; Indicates minute strain disturbances; Indicates the disturbance level or classification coefficient, with values ​​as follows: (Can be added according to project requirements); The magnitude of the fundamental disturbance vector is represented by historical stress variation statistics or simulation using a digital twin physical model.

[0152] S42. Calculate the change in unit redundancy under virtual disturbance, define the disturbance response amplification factor and the sensitivity nonlinearity index, and determine whether the unit has entered a highly sensitive stage by the nonlinear response trend. This provides a dynamic basis for determining irreversible evolution and enables early identification of potential instability points. Specifically:

[0153] Disturbance After application, calculate the corrected redundancy:

[0154] ;

[0155] It should be noted that the function This function maps the state vector of micro-perturbations applied in a digital twin to the change in element redundancy correction. Its core function is to reflect the amplification effect of perturbations under structural nonlinear conditions. This function comprehensively considers material nonlinearity, inter-element coupling relationship, and historical stress and pore pressure state. By weighting and nonlinearly mapping micro-stress, pore pressure, and strain perturbations, it calculates the correction value of element redundancy after perturbation, enabling the system to quantify the sensitivity of the structure to micro-perturbations, thereby providing a reliable basis for subsequent nonlinear amplification factor and critical sensitivity stage determination.

[0156] Define the disturbance response amplification factor:

[0157] ;

[0158] like This indicates that the structure exhibits a nonlinear amplified response;

[0159] Calculate the sensitivity nonlinearity index:

[0160] ;

[0161] like (The set sensitivity threshold) indicates that the unit has entered the "highly sensitive nonlinear stage", which corresponds to the potential instability critical point;

[0162] in, Indicates the number of seconds after the disturbance is applied. The level of redundancy correction reflects the state of unit redundancy after changes in disturbance, taking into account nonlinear effects; Indicates the disturbance response amplification factor; Indicates a disturbance-sensitive nonlinear index;

[0163] S43. After disturbance removal, calculate the residual redundancy offset, construct an irreversible evolution index in conjunction with the nonlinear sensitivity index, determine whether the unit has entered the irreversible evolution stage, and identify regional-level critical evolution zones through continuous high-value units, providing a basis for the operational early warning core area of ​​the project. Specifically:

[0164] After applying and removing the perturbation, calculate the residual redundancy change:

[0165] ;

[0166] like This indicates that the structure has not fully recovered and there is cumulative damage.

[0167] Constructing an irreversible evolution index: ;

[0168] like If this happens, the unit enters an irreversible evolutionary state;

[0169] If multiple consecutive units satisfy (The set critical threshold) forms a "critical evolution zone", which is the high-risk core area within the digital twin and serves as the target for early warning.

[0170] in, This indicates the redundancy restored by the unit after the disturbance is removed, and is used to determine whether there is residual damage to the structure after the disturbance. This represents the residual redundant offset after the disturbance is removed; Indicates the irreversible evolution index; This represents the perturbation-sensitive weight, reflecting the contribution of nonlinear amplification to irreversibility; This represents the residual offset weight, reflecting the contribution of accumulated damage to irreversibility;

[0171] S44. A closed-loop dynamic early warning mechanism is constructed based on regional critical zones. Early warning levels are output through dual determination (evolutionary acceleration zone and irreversible evolution stage), and dynamic adjustment and closed-loop feedback are achieved using the irreversible exponential rate of change. This enables the early warning system to be self-verifiable and continuously updated. Specifically:

[0172] The warning is triggered only when the following conditions are met: step S3 determines that the evolution is in an accelerated phase; step S4 determines that the evolution is in an irreversible stage.

[0173] Based on the rate of change of the irreversible evolution index, the early warning level (low / medium / high / critical) is dynamically adjusted according to the growth rate.

[0174] Example 2, as Figure 2 As shown, the present invention proposes a digital twin-based disaster early warning system for engineering geological areas, which is used to execute a digital twin-based disaster early warning method for engineering geological areas proposed in Embodiment 1. It includes: a digital twin modeling and state update module, a disturbance chain analysis and dynamic evolution module, a structural safety redundancy attenuation and evolution acceleration identification module, and an active disturbance injection and critical phase verification module.

[0175] The digital twin modeling and state update module is used to construct a structural evolution digital twin of an engineering geological area and perform comprehensive state modeling of regional geological units. The module integrates multi-source geological data, including topography, soil and rock properties, hydrological information and historical disaster records. Through a continuous time update mechanism, it realizes the dynamic evolution of the digital twin and can generate state vectors of each unit in real time, reflecting key parameters such as stress, strain and pore water pressure, providing a high-precision data foundation for subsequent early warning analysis.

[0176] The perturbation chain analysis and dynamic evolution module is responsible for analyzing the propagation path and impact range of potential perturbations within the digital twin, generating a trigger chain intensity field, quantifying the transmission impact of perturbations on each structural unit, processing time series data, identifying the starting point, acceleration zone, and potential expansion region of perturbation propagation, and forming a dynamic evolution state diagram. This module can simulate the coupling effect of multi-source perturbations, realize the prediction of local risk accumulation and regional response, provide input for structural safety redundancy analysis, and ensure that perturbation analysis is closely coupled with the real-time state of the digital twin.

[0177] The structural safety redundancy decay and evolution acceleration identification module calculates the safety redundancy of each structural unit and its temporal evolution trend based on the digital twin. It identifies potential instability regions by redundancy decay rate, acceleration, and spatial gradient. It can form a continuous safety redundancy field and automatically identify evolution acceleration zones. At the same time, it generates a comprehensive instability evolution index and sorts and classifies high-risk units within the region. It embeds the disturbance coupling effect into the redundancy model to achieve synchronous analysis of redundancy consumption and disturbance propagation.

[0178] The active perturbation injection and critical phase verification module simulates the nonlinear response of the unit under perturbation by applying controlled micro-perturbations within the digital twin, and calculates the residual offset after the perturbation is removed to form an irreversible evolution index. It can identify highly sensitive units and regional critical zones, and realize closed-loop verification of the dual judgment mechanism. It dynamically updates the early warning level according to the evolution index and outputs visualized early warning information and risk maps, providing accurate basis for engineering management and emergency decision-making.

[0179] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for early warning of regional disasters in engineering geology based on digital twins, characterized in that, The specific implementation steps include the following: S1. Construct a structural evolution-type regional geological digital twin. By dividing the structural units according to mechanical consistency, establishing the directional stress propagation matrix between units, constructing coupled dynamic evolution equations and combining adaptive error feedback, the calculable evolution and real-time synchronous update of the regional geological body can be realized. S2. Based on the constructed structural evolution-type regional geological digital twin, multi-source natural and anthropogenic disturbances are uniformly quantified as inputs. By coupling the relationship between structural integrity and asymmetric stress propagation, a disaster trigger chain propagation matrix is ​​constructed. The trigger chain strength and its time evolution rate of each structural unit are calculated, potential chain instability paths are identified, and a dynamic threshold mechanism based on regional statistical characteristics is adopted to achieve hierarchical early warning output. S3. Based on the obtained trigger chain propagation strength and evolution trend, a dynamic decay model of structural safety redundancy is constructed. By establishing a continuous redundancy function, time derivative analysis and spatial gradient identification mechanism, the regional structural safety reserve is transformed from static evaluation to a differentiable evolutionary variable. Trigger chain strength coupling correction is introduced to achieve accurate identification of the accelerated dissipation region of structural redundancy. Finally, a structural instability evolution index is constructed to output graded risk results. S4. Based on the identified accelerated structural evolution zone, by actively applying controlled micro-perturbations within the digital twin, the nonlinear amplification and residual offset of the structural redundancy response are analyzed to identify irreversible evolution stages and form a closed-loop dynamic early warning mechanism, thereby realizing disaster early warning from trend judgment to active verification.

2. The method for early warning of engineering geological regional disasters based on digital twins according to claim 1, characterized in that, The process of constructing a structural evolution-based regional geological digital twin in step S1 includes: Based on the stratigraphic structure, geophysical parameters and stress gradient, the region is divided into mechanically responsive homogeneous structural units. Each unit is defined with an extended state vector containing stress, strain, shear stress, pore water pressure and damage index. An asymmetric stress transfer topology matrix is ​​constructed based on structural units. Combining unit contact area, mechanical distance, attenuation coefficient and directional correction factor, the mechanical coupling propagation between adjacent units is simulated. A coupled dynamic evolution equation for the regional structure is constructed, integrating the nonlinear elastoplastic response within the unit, the influence of neighborhood coupling, and the input of external disturbances, to realize the time-varying evolution calculation of the structural state of the entire region; An adaptive error feedback mechanism is introduced to compare the measured monitoring data with the model prediction results and make corrections through a dynamic gain matrix, thereby achieving synchronous updates between the digital twin and the actual engineering status.

3. The method for early warning of engineering geological regional disasters based on digital twins according to claim 2, characterized in that, The partitioning condition for homogeneous structural units with mechanical response is: Under given historical perturbation conditions, the rate of change of stress response gradient meets the preset allowable threshold for stress gradient and the rate of change of rock physical parameters meets the preset allowable threshold for parameter change. The damage index in the extended state vector is obtained by integrating the strain history. When the damage index reaches the material's ultimate damage value, it indicates that the element is close to failure. The stress transfer coefficients in the asymmetric stress transfer topology matrix are determined based on the structural contact weighting coefficient, the effective mechanical distance between the centroids of the elements, the directional correction factor, and the stress propagation attenuation coefficient. The directional correction factor is determined by the cosine of the angle between the principal stress direction and the line connecting the elements.

4. The method for early warning of engineering geological regional disasters based on digital twins according to claim 3, characterized in that, Step S2 involves constructing the catastrophe triggering chain propagation matrix by coupling the structural integrity and asymmetric stress propagation relationship. Multi-source disturbances such as rainfall, earthquakes, and construction are classified and modeled. Different physical quantities are normalized by historical extreme values ​​and uniformly quantified by weighting coefficients to construct unit-level disturbance input vectors. Based on the established stress transfer matrix between structural units, a structural integrity modulation function, distance attenuation, and direction correction factor are introduced to construct a disturbance trigger chain propagation matrix, so that the disturbance propagation intensity is simultaneously affected by the unit degradation state and mechanical direction. By calculating the trigger chain strength index and its time change rate of each structural unit using the propagation matrix, a directed graph of the structural network is constructed to identify regions of accelerated growth of trigger chains and potential catastrophic paths. A dynamic threshold model is constructed based on the mean and standard deviation of the trigger chain strength across the entire region. The warning levels are divided by combining the chain length and evolution rate. When the unit strength exceeds the adaptive threshold, the graded warning results are output.

5. A method for early warning of engineering geological regional disasters based on digital twins according to claim 4, characterized in that, The structural integrity modulation function is constructed based on the structural integrity index of the unit. The closer the integrity index is to zero, the more severe the damage. The modulation function determines the degree to which the integrity change amplifies the propagation intensity through the damage sensitivity adjustment coefficient. The trigger chain strength index is calculated by accumulating the propagation contribution from other units through the propagation matrix, and the identification of the accelerated growth region of the trigger chain is based on time series analysis of the strength evolution rate; In the dynamic threshold model, the adaptive threshold is the sum of the mean strength of the trigger chain across the entire region and the adjustment coefficient multiplied by the standard deviation.

6. A method for early warning of engineering geological regional disasters based on digital twins according to claim 5, characterized in that, The process of constructing the structural safety redundancy dynamic attenuation model in step S3 includes: Based on the real-time stress field and pore water pressure field of digital twin structural units, a dynamic effective shear strength model is constructed, and the equivalent shear stress of the unit is calculated. A continuous structure safety redundancy function is defined, and the safety state of discrete units is mapped to a spatial continuous redundancy field. The time derivative of the structural safety redundancy is calculated, and the redundancy decay rate and decay acceleration are obtained respectively to identify whether there is an accelerated dissipation trend of redundancy. At the same time, the obtained trigger chain strength is introduced to construct an exponential correction redundancy function so that the safety reserve reflects the influence of internal stress degradation and external disturbance propagation. A spatial gradient function is constructed based on a continuous redundant field to identify redundant abrupt change regions and potential slip concentration zones. An evolution concentration index is constructed through the regional average decay rate to determine whether a unit belongs to an accelerated degradation zone. A structural instability evolution index is constructed, which includes corrected redundancy, decay rate, decay acceleration and trigger chain strength. The weight parameters are calibrated using historical sample data to achieve multi-physical variable fusion evaluation, and risk classification is output based on the index change trend and critical threshold.

7. A method for early warning of engineering geological regional disasters based on digital twins according to claim 6, characterized in that, The physical meaning of the structural safety redundancy function is that a value greater than zero indicates the existence of a safety reserve, a value approaching zero indicates approaching the critical point, and a value less than zero indicates entering the theoretical instability region. The redundant decay acceleration is calculated in discrete time form. When the redundant decay rate is less than the set decay rate threshold and the decay acceleration is less than the set decay acceleration threshold, it is determined to be accelerated decay. The evolution concentration index is the ratio of the unit decay rate to the average decay rate of the region. Combined with three composite conditions—redundancy below a preset redundancy threshold, decay rate below a decay rate threshold, and decay acceleration below a decay acceleration threshold—the unit is determined to be in the structural evolution acceleration zone.

8. A method for early warning of engineering geological regional disasters based on digital twins according to claim 7, characterized in that, Step S4 involves actively applying controlled micro-perturbations within the digital twin, including: Based on the output evolution acceleration zone element, construct a micro-perturbation state vector containing stress, pore pressure and strain, and use a hierarchical incremental loading strategy to perform virtual perturbation on the target element; The change in redundancy of the unit is calculated under virtual disturbance, the disturbance response amplification factor and the sensitivity nonlinearity index are defined, and the nonlinear response trend is used to determine whether the unit has entered the high sensitivity stage. After the disturbance is removed, the residual redundancy offset is calculated, and the irreversible evolution index is constructed by combining the nonlinear sensitivity index to determine whether the unit has entered the irreversible evolution stage. The regional critical evolution zone is identified by continuous high-value units. A closed-loop dynamic early warning mechanism is constructed based on the regional critical zone. The early warning level is output through dual judgment, namely the evolution acceleration zone and the irreversible evolution stage. The dynamic adjustment and closed-loop feedback are realized by using the irreversible exponential change rate.

9. A method for early warning of engineering geological regional disasters based on digital twins according to claim 8, characterized in that, The disturbance response amplification factor is the ratio of the redundancy correction before and after the disturbance is applied. If this ratio exceeds the set amplification threshold, it indicates that the structure exhibits a nonlinear amplification response. The sensitive nonlinearity index is calculated based on the redundancy changes under different disturbance levels, and combined with the set sensitive threshold judgment unit, it enters the high-sensitivity nonlinearity stage; The residual redundancy offset is the difference between the redundancy restored after the disturbance is removed and the redundancy before the disturbance. If the offset exceeds the set residual offset threshold, it indicates that there is cumulative damage to the structure. The irreversible evolution index is constructed by fusing perturbation-sensitive weights and residual offset weights. If the set critical threshold is exceeded, the unit is determined to enter an irreversible evolution state. When several consecutive units enter an irreversible evolutionary state, the region formed by these units is identified as the critical evolution zone.

10. A digital twin-based engineering geological regional disaster early warning system, used to execute the digital twin-based engineering geological regional disaster early warning method according to any one of claims 1 to 9, characterized in that, include: The digital twin modeling and state update module is used to construct a structural evolution digital twin of an engineering geological area, perform comprehensive state modeling of regional geological units, integrate multi-source geological data, realize the dynamic evolution of the digital twin through a continuous time update mechanism, and generate state vectors of each unit in real time. The perturbation chain analysis and dynamic evolution module is used to analyze the propagation path and impact range of potential perturbations within the digital twin, generate a trigger chain strength field, quantify the transmission impact of perturbations on each structural unit, process time series data to identify the starting point, acceleration zone and potential expansion region of perturbation propagation, and form a dynamic evolution state diagram. The structural safety redundancy decay and evolution acceleration identification module is used to calculate the safety redundancy of each structural unit and its time evolution trend based on the digital twin. It identifies potential instability regions by redundancy decay rate, acceleration and spatial gradient, forms a continuous safety redundancy field and automatically identifies the evolution acceleration region, and generates a comprehensive instability evolution index. The active perturbation injection and critical phase verification module is used to apply controlled micro-perturbations within the digital twin, simulate the nonlinear response of the unit under the perturbation, calculate the residual offset after the perturbation is removed, form an irreversible evolution index, identify highly sensitive units and regional critical zones to achieve closed-loop verification of the dual judgment mechanism, dynamically update the early warning level based on the evolution index, and output visualized early warning information.

Images