Reservoir bank slope disaster chain real-time prevention and control method

Abstract

The invention belongs to the technical field of geotechnical engineering and artificial intelligence crossing, and particularly discloses a reservoir bank slope disaster chain real-time prevention and control method, which comprises the steps of determining a plurality of disaster nodes of a reservoir bank slope, and constructing a graph model of the reservoir bank slope disaster chain according to the plurality of disaster nodes; determining a correlation degree and a mechanical transfer coefficient between two disaster nodes corresponding to each edge in the graph model, and determining a coupling weight of each edge according to the correlation degree and the mechanical transfer coefficient; presetting initial disaster nodes and termination disaster nodes of a plurality of disaster chains in the graph model, and determining the shortest path between each pair of initial disaster node and termination disaster node and the path betweenness centrality of the shortest path according to the coupling weight; and determining a key score of each shortest path according to the path betweenness centrality, the loss expected value and the dynamic propagation risk probability. The method can interfere with the triggering and propagation links of a disaster chain, weaken the disaster propagation energy, improve the global toughness of the reservoir bank slope, and realize real-time prevention and control.

Description

Technical Field

[0001] The present invention belongs to the cross - technical field of geotechnical engineering and artificial intelligence, and discloses a real - time prevention and control method for the disaster chain of reservoir bank slopes, specifically a real - time prevention and control method for the disaster chain of reservoir bank slopes based on dynamic coupling network and multi - level blocking, which is used for the real - time prevention and control and treatment of the disaster chain in reservoir bank slope projects. Background Technique

[0002] Due to the long - term influence of various factors such as long - term immersion of reservoir water, water - level fluctuation, and geological structure, reservoir bank slopes are extremely prone to geological disasters such as landslides. Once a landslide occurs, it may trigger a surge wave, and the surge wave hitting the dam may lead to dam break, forming a complex disaster chain. The disaster chain of reservoir bank slopes (such as "rainfall infiltration → seepage - stress coupling → local slip → overall instability → secondary surge disaster") has characteristics such as multi - factor dynamic coupling, cascade propagation non - linearity, and blocking time - effect sensitivity.

[0003] Domestic and foreign research often uses static network theory to construct disaster chain models, abstracting disaster events such as landslides, surge waves, and dam breaks as nodes, and historical relevance as edge weights. However, its edge weights are fixed based on historical data and cannot dynamically integrate real - time monitoring data (such as rainfall intensity, displacement rate) and geotechnical mechanical parameters (such as elastic modulus, permeability coefficient). The constructed network topology does not reflect physical mechanisms such as stress transfer in rock and soil masses and hydraulic gradient changes, resulting in the mechanical driving process of the propagation path being simplified to a probability relationship. In addition, traditional research on reservoir bank slope disasters mostly focuses on the mechanism analysis of single disaster types (such as landslides or dam breaks) or single mechanical models, while ignoring the dynamic propagation law of the cascade effect of multiple disaster types (such as the chain reaction of landslide → surge wave → dam break). The traditional single - disaster - type research paradigm cannot comprehensively reveal the cascade effect and propagation law among various disaster types in the disaster chain, resulting in prevention and control strategies focusing on strengthening a single disaster node (such as grouting of the landslide body), and it is difficult to meet the requirements for effective prevention and control of reservoir bank slope disasters.

[0004] In summary, the existing prevention and control methods for the disaster chain of reservoir bank slopes have the following problems:

[0005] (1) Existing disaster chain models (such as event trees, Bayesian networks) are mostly based on historical data to construct static topological structures, with fixed edge weights, and cannot real - time integrate dynamic changes such as hydrological fluctuations (such as water - level sudden - drop rate>0.5 m / d) and deterioration of geotechnical parameters (such as cohesion attenuation).

[0006] (2) There is insufficient dynamic coupling between real - time monitoring data (such as Beidou displacement meters, piezometers) and network models. In the process of calculating key paths, traditional path - search algorithms (such as Dijkstra algorithm) need to globally update the network, with high time complexity and unable to meet the requirements of real - time prevention and control.

[0007] (3) Existing blocking methods (such as anti-slide pile reinforcement, grouting to stop leaks) are mostly single-point prevention and control, ignoring the cascade propagation characteristics of the disaster chain, with insufficient suppression of the propagation path and no dynamic adjustment of the blocking intensity for high-weight edges (such as the seepage path of soft interlayers). Summary of the Invention

[0008] The purpose of the present invention is to provide a real-time prevention and control method for the disaster chain of reservoir bank slopes to solve the technical problem that existing disaster prevention and control methods for reservoir bank slopes are difficult to combine real-time data for real-time prevention and control of the disaster chain.

[0009] The present invention provides a real-time prevention and control method for the disaster chain of reservoir bank slopes, including:

[0010] Step 1: Determine multiple disaster nodes of the reservoir bank slope, and construct a graph model of the disaster chain of the reservoir bank slope according to the multiple disaster nodes;

[0011] Step 2: Determine the correlation degree and mechanical transmission coefficient between every two disaster nodes corresponding to each edge in the graph model, and determine the coupling weight of each edge according to the correlation degree and mechanical transmission coefficient;

[0012] Step 3: Preset the starting disaster node and the ending disaster node of multiple disaster chains in the graph model, and determine the shortest path and the betweenness centrality of the shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight;

[0013] Step 4: Determine the criticality score of each shortest path according to the betweenness centrality, loss expectation value, and dynamic propagation risk probability.

[0014] Preferably, after step 4, it further includes:

[0015] Determine the blocking threshold according to the criticality score of each shortest path, and determine the critical path as the shortest path with a criticality score greater than the blocking threshold;

[0016] Perform risk blocking on the critical path.

[0017] Preferably, determining the blocking threshold according to the criticality score of each shortest path specifically includes:

[0018] Obtain the maximum criticality score and determine the average value of all criticality scores;

[0019] Determine the blocking threshold according to the maximum criticality score and the average value of all criticality scores.

[0020] Preferably, after performing risk blocking on the critical path, it further includes:

[0021] Determine the pre-blocking toughness index of the critical path according to the criticality score of the critical path before risk blocking;

[0022] Determine the post-blocking resilience index of the critical path according to the criticality score of the critical path after risk blocking;

[0023] Determine the effectiveness of risk blocking according to the pre-blocking resilience index and the post-blocking resilience index.

[0024] Preferably, determine the mechanical transfer coefficient between two disaster nodes corresponding to each edge in the graph model, specifically:

[0025] Determine the elastic modulus at two disaster nodes corresponding to each edge in the graph model;

[0026] Determine the tangent modulus of the stress-strain curve of the rock mass on the reservoir bank slope, and determine the mechanical transfer coefficient between two disaster nodes corresponding to each edge in the graph model according to the elastic modulus and the tangent modulus.

[0027] Preferably, determine the coupling weight of each edge according to the correlation degree and the mechanical transfer coefficient, specifically:

[0028] Determine the bearing capacity of the rock mass according to the critical load of the rock mass on the reservoir bank slope and the maximum load borne at the current moment;

[0029] Determine the coupling weight of each edge according to the product of the correlation degree, the mechanical transfer coefficient and the bearing capacity of the rock mass.

[0030] Preferably, determine the shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight, specifically:

[0031] Determine the initial shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight;

[0032] When the change amount of the coupling weight in the initial shortest path is greater than a preset threshold, determine the disaster nodes corresponding to the adjacent edges of the coupling weight;

[0033] Update the initial shortest path according to the disaster nodes corresponding to the adjacent edges of the coupling weight to obtain the shortest path between each pair of starting disaster nodes and ending disaster nodes.

[0034] Preferably, determine the initial shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight, specifically:

[0035] Determine the reciprocal of the bearing capacity of the rock mass according to the critical load of the rock mass on the reservoir bank slope and the maximum load borne at the current moment;

[0036] Determine the activation probability of the disaster nodes corresponding to the coupling weight;

[0037] Determine the initial shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight, the reciprocal of the rock mass bearing capacity, and the activation probability.

[0038] Preferably, determine the criticality score of each shortest path according to the betweenness centrality of the path, the expected loss value, and the dynamic propagation risk probability, specifically:

[0039] Determine the inhibition parameter of each shortest path according to the length of each shortest path and the average value of the lengths of all shortest paths;

[0040] Determine the criticality score of each shortest path according to the inhibition parameter of each shortest path, the betweenness centrality of the path, the expected loss value, and the dynamic propagation risk probability.

[0041] Preferably, the method for determining the dynamic propagation risk probability is as follows:

[0042] Determine the path association strength according to the product of multiple coupling weights on each shortest path;

[0043] Determine the probability activation superposition amount according to the activation probabilities of multiple disaster nodes on each shortest path;

[0044] Determine the dynamic propagation risk probability of each shortest path according to the path association strength and the probability activation superposition amount.

[0045] The real-time prevention and control method for the disaster chain of the reservoir bank slope of the present invention has the following beneficial effects compared with the prior art:

[0046] (1) The present invention combines the mechanical transfer coefficient and the correlation degree, and real-time corrects the coupling weight of each edge, realizing the real-time update of the coupling weight with the dynamic monitoring data, and solving the problem that the edge weight of the existing static network cannot reflect the dynamic evolution.

[0047] (2) The criticality score of the shortest path of the present invention integrates the betweenness centrality of the path, the expected loss value, and the dynamic propagation risk probability, and solves the decision deviation problem of a single index.

[0048] (3) After determining the criticality score of each shortest path, the present invention determines the critical path according to the criticality score and performs risk blocking on the critical path. And when performing risk blocking, a three-level blocking rule is proposed. Through the three major mechanisms of hydraulic regulation (reducing seepage pressure), mechanical reinforcement (increasing the critical load), and energy dissipation (constructing redundant paths), the triggering and propagation links of the disaster chain are directly intervened, the energy of disaster propagation is weakened, and the global resilience of the system is improved to achieve real-time prevention and control. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a schematic flowchart of the real-time prevention and control method for the disaster chain of the reservoir bank slope in the embodiment of the present invention.

[0050] Figure 2 This is a comparison chart of the key scores and the change rate of the toughness index after blocking in Embodiments 1 to 3 of the present invention. Detailed implementation manners

[0051] In the following description, specific details such as specific system structures and technologies are presented for the purpose of illustration rather than limitation, so as to thoroughly understand the embodiments of the present invention. However, those skilled in the art should clearly understand that the present invention can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from obstructing the description of the present invention.

[0052] An embodiment of the present invention provides a real-time prevention and control method for the disaster chain of the reservoir bank slope, as Figure 1 shown, including:

[0053] Step 1: Determine multiple disaster nodes of the reservoir bank slope, and construct a graph model of the disaster chain of the reservoir bank slope according to the multiple disaster nodes.

[0054] In the embodiment of the present invention, the disaster nodes are determined and the disaster node set is constructed according to the trigger - propagation - amplification process of the disaster chain of the reservoir bank slope , where represents the total number of disaster nodes, realizing full-chain coverage from the disaster-causing factor ( ) → carrier medium ( ) → disaster entity ( ) → prevention and control feedback ( ), as shown in Table 1.

[0055] Table 1 Disaster node division

[0056] Construct a graph model of the disaster chain of the reservoir bank slope according to the above disaster nodes, specifically:

[0057] If the occurrence of disaster node causes the occurrence of disaster node , then a directed edge is constructed between the two disaster nodes, and the set of edges is denoted as , and the constructed graph model (i.e., the complex network model) is represented as .

[0058] In the embodiment of the present invention, the data of the 12 disaster nodes in Table 1 are dimensionlessized to obtain dimensionless data , as shown in formula (1).

[0059] (1)

[0060] In the formula, is the disaster node at the dimensionless data at the moment, is the disaster node at the data at the moment, is the average value of the historical data of the disaster node , is the disaster node the standard deviation of the historical data, is the time decay factor at the moment and , where is the decay coefficient, taking 0.05 - 0.2 / day, is the current moment, is the moment when the data is actually collected.

[0061] In the embodiments of the present invention, the subsequent calculations all use the dimensionless data .

[0062] Step 2: Determine the correlation degree and mechanical transfer coefficient between two disaster nodes corresponding to each edge in the graph model, and determine the coupling weight of each edge according to the correlation degree and mechanical transfer coefficient, including:

[0063] Step 2.1: Determine the correlation degree between two disaster nodes corresponding to each edge in the graph model, specifically:

[0064] Step 2.1.1: Determine the difference value of the monitoring data between two disaster nodes corresponding to each edge in the graph model .

[0065] Where is the difference value between the disaster node and the disaster node at the moment, , are respectively the monitoring data of the disaster node and the disaster node at the moment.

[0066] Step 2.1.2: Determine the correlation degree between two disaster nodes corresponding to each edge in the graph model according to the time decay factor and the difference value .

[0067] In the embodiments of the present invention, a correlation degree based on machine learning is constructed for quantifying the disaster node and the disaster node The dynamic correlation strength between The larger the value, the more similar the behaviors of the two disaster nodes are during the dynamic evolution of the disaster.

[0068] (2)

[0069] In the formula, represents the time window length, that is, the total number of moments involved in the calculation; is the disaster node and the disaster node at the moment, the difference value; and are respectively at the moment, the monitoring data of the disaster node and the disaster node is the disaster node and the disaster node the global minimum difference value; is the disaster node and the disaster node the global maximum difference value; is the resolution coefficient, usually taking to balance sensitivity and stability; is the dynamic weight at the moment, reflecting the importance of different moments to the correlation degree, .

[0070] Step 2.2. Determine the mechanical transfer coefficient corresponding to each edge in the graph model between two disaster nodes, specifically:

[0071] Step 2.2.1. Determine the elastic modulus at each of the two disaster nodes corresponding to each edge in the graph model.

[0072] Exemplarily, determine the elastic modulus at the disaster node in the graph model and the elastic modulus at the disaster node .

[0073] Step 2.2.2. Determine the tangent modulus of the stress-strain curve of the rock mass on the reservoir bank slope, and determine the mechanical transfer coefficient corresponding to each edge in the graph model between two disaster nodes according to the elastic modulus and the tangent modulus .

[0074] In the embodiment of the present invention, the geotechnical mechanical transfer coefficient between the disaster node and the disaster node is introduced to quantify the disaster node The mechanical correlation strength with the disaster node is as shown in formula (3). The mechanical transfer efficiency between disaster nodes is corrected by the mechanical properties (elastic modulus, stress-strain relationship) of the material on the integral path. Embedding the mechanical response of the geological structure into the graph model makes the cascade process of "landslide → surge → dam break" not only depend on the event correlation degree but also be controlled by the actual mechanical behavior of the rock and soil mass:

[0075] (3)

[0076] In the formula, is the effective transfer path length between the disaster node and the disaster node ; are the elastic moduli at the disaster node and the disaster node respectively; is the tangent modulus of the stress-strain curve of the rock mass on the reservoir bank slope, is the stress, is the strain; is the infinitesimal length on the integral path.

[0077] Step 2.3: Determine the coupling weight of each edge according to the correlation degree and the mechanical transfer coefficient, specifically:

[0078] Step 2.3.1: Determine the bearing capacity of the rock mass according to the quotient of the critical load of the rock mass on the reservoir bank slope and the maximum load borne at the current moment.

[0079] Step 2.3.2: Determine the coupling weight of each edge according to the product of the correlation degree , the mechanical transfer coefficient and the bearing capacity of the rock mass. This coupling weight is a dynamic coupling weight, as shown in formula (4).

[0080] (4)

[0081] In the formula, is the coupling weight between the disaster node and the disaster node . When the value of is relatively high, it indicates a strong correlation between the disaster node and the disaster node , which may be the key path for the propagation of the disaster chain; When the value is relatively low, it indicates a weak correlation between the disaster node and the disaster node , which can be used as a secondary path for disaster propagation.

[0082] Step 3: Preset the starting disaster nodes and ending disaster nodes of multiple disaster chains in the graph model, and determine the shortest path and the betweenness centrality of the shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight, where the shortest path is the determined disaster chain.

[0083] Step 3.1: Preset the starting disaster nodes and ending disaster nodes of multiple disaster chains in the graph model, and determine the shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight. Specifically:

[0084] Step 3.1.1: Preset the starting disaster nodes and ending disaster nodes of multiple disaster chains in the graph model. Exemplarily, the starting disaster node of one of the preset disaster chains is and the ending disaster node is .

[0085] Step 3.1.2: Determine the shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight. Specifically:

[0086] A. Determine the initial shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight.

[0087] Exemplarily, the method for determining the initial shortest path between the starting disaster node and the ending disaster node is specifically as follows: Determine the reciprocal of the rock mass bearing capacity according to the critical load of the reservoir bank slope rock mass and the maximum load borne at the current moment; then determine the activation probability of the disaster node corresponding to the coupling weight; finally, determine the initial shortest path between the starting disaster node and the ending disaster node according to the coupling weight, the reciprocal of the rock mass bearing capacity, and the activation probability, as shown in formula (5).

[0088] (5)

[0089] In the formula, is the initial shortest path between the starting disaster node and the ending disaster node at time ; is the coupling weight between the starting disaster node and the ending disaster node at time ; is the starting disaster node at The activation probability at a moment, which ranges from 0 to 1; is the reciprocal of the bearing capacity of the rock mass determined according to the critical load of the rock mass slope on the reservoir bank and the maximum load borne at the moment, reflecting the node stability; , represents the adjustment coefficient, which is defaulted according to engineering experience , to balance the propagation risk and the mechanical failure risk.

[0090] In the method for determining the initial shortest path according to the embodiment of the present invention, the activation probability, the adjustment coefficient, and the ratio of the maximum load to the critical load are introduced, and the corresponding initial shortest path is comprehensively calculated, and the obtained result is more accurate.

[0091] B. When the change amount of the coupling weight in the initial shortest path is greater than the preset threshold, determine the disaster node corresponding to the adjacent edge of the coupling weight.

[0092] Since the shortest path is dynamically changing, the existing path search algorithm Dijkstra based on artificial intelligence needs to globally recalculate all paths. The embodiment of the present invention corrects the Dijkstra algorithm. In the graph model, only the disaster nodes with changed coupling weights are locally updated to improve the calculation efficiency.

[0093] The corrected Dijkstra algorithm according to the embodiment of the present invention realizes incremental calculation through the following improvement: when the change amount of the coupling weight of the directed edge composed of the disaster node to the disaster node in the initial shortest path is greater than the preset threshold, and .

[0094] C. Update the initial shortest path according to the disaster node corresponding to the adjacent edge of the coupling weight to obtain the shortest path between each pair of starting disaster nodes and ending disaster nodes.

[0095] Let be a disaster node in the affected disaster node set, traverse all the disaster nodes in the affected disaster node set to find a shorter path, and update the initial shortest path to obtain the finally determined shortest path between each pair of starting disaster nodes and ending disaster nodes.

[0096] Exemplarily, update the initial shortest path according to formula (6).

[0097] (6)

[0098] Wherein, is the shortest path between the disaster node at time and the disaster node ; is the initial shortest path between the disaster node at time and the disaster node ; is the shortest path between the disaster node at time and the disaster node ; is the shortest path between the disaster node at time and the disaster node ; and are determined according to formula (5).

[0099] The embodiment of the present invention is based on the sliding time window (i.e., changes) and increment, and only updates the paths with the coupling weight change amount greater than the preset threshold, reduces the calculation complexity, improves the update efficiency of the paths in the graph model, and significantly improves the real-time performance and accuracy of the disaster chain path recognition.

[0100] Step 3.2, determining the betweenness centrality of the shortest path.

[0101] Assume that the shortest path is , the embodiment of the present invention introduces a time decay factor , calculates the dynamic betweenness centrality of the disaster node and the betweenness centrality of the path. Since the historical contribution of the disaster nodes in the path gradually weakens over time, the embodiment of the present invention pays more attention to the recent activities of the disaster chain.

[0102] (7)

[0103] Wherein, is the betweenness centrality of the shortest path ; is the total number of disaster nodes in the shortest path ; is the disaster node in the shortest path , and ; is the dynamic betweenness centrality of the disaster node in the shortest path ; is The time decay factor at a moment and , where is the attenuation coefficient, taking 0.05 - 0.2 / day, is the current moment, is the moment when the data is actually collected; is the number of the shortest paths passing through the disaster node ; is the number of the shortest paths between the disaster node and the disaster node .

[0104] The shortest paths with high betweenness centrality are the "main roads" for disaster conduction. Blocking such shortest paths can slow down the disaster diffusion to the greatest extent.

[0105] Step 4. Determine the criticality score of each path according to the betweenness centrality, the expected loss value, and the dynamic propagation risk probability.

[0106] Before performing Step 4, it is necessary to determine the expected loss value and the dynamic propagation risk probability first.

[0107] In the embodiment of the present invention, an expected loss function is constructed to determine the expected loss value of the shortest path , which is used to quantify the upper limit of the direct economic loss and the indirect instability diffusion risk of the shortest path.

[0108] (8)

[0109] In the formula, represents the direct economic loss caused by the failure of the disaster node , which is assigned by expert evaluation; is the direct economic loss of the most vulnerable disaster node in the shortest path, reflecting the single-point failure risk in the worst case; is the penalty coefficient for load instability (default ); is the current load of the disaster node (such as the stress of slope rock mass, seepage pressure); is the critical load (instability threshold) of the disaster node .

[0110] The expected loss value of the embodiment of the present invention takes into account both the direct economic loss of the most serious single point and the systematic instability risk, and adjusts the prevention and control priority through . For the paths with high values obtained, it is necessary to reinforce the end disaster nodes and reduce the load pressure.

[0111] The shortest path Dynamic propagation risk probability The risk is jointly determined by the path association strength (structural attribute) and the probability activation superposition effect (dynamic state). The dynamic propagation risk probability is determined as follows: Determine the path association strength according to the product of multiple coupling weights on the shortest path; Determine the probability activation superposition amount according to the activation probabilities of multiple disaster nodes on the shortest path; Determine the dynamic propagation risk probability of the shortest path according to the path association strength and the probability activation superposition amount , as shown in formula (9).

[0112] (9)

[0113] In the formula is the coupling weight of the edge between the disaster node and the disaster node at time is the path association strength, indicating that the overall propagation risk of the shortest path is the product of the coupling weights of each edge. If any one coupling weight decreases, the propagation ability of the entire shortest path will be significantly weakened; is the probability activation superposition amount, represents the natural exponential function, is the activation probability of the disaster node at time, indicating the activation state of the disaster node at time (such as the landslide trigger probability). The shortest path with a high needs to be monitored preferentially because it may trigger a chain of disasters due to the instability of a single node.

[0114] Step 4.1: Determine the suppression parameter of each shortest path according to the length of each shortest path and the mean value of the lengths of all shortest paths.

[0115] Step 4.2: Determine the criticality score of each shortest path according to the suppression parameter, betweenness centrality, loss expectation value, and dynamic propagation risk probability of each shortest path. Taking one of the shortest paths as an example, its criticality score is shown in formula (10).

[0116] (10)

[0117] In the formula, is the criticality score of the shortest path ; is the index weight, default , which can be adjusted dynamically; is the shortest path Length; is the average length of all the shortest paths; is an inhibition parameter used to inhibit overly long shortest paths, is for controlling the inhibition intensity, with a default value of .

[0118] The criticality score of the embodiment of the present invention comprehensively evaluates the criticality of the shortest paths from three dimensions: propagation ability (dynamic propagation risk probability), topological importance (path betweenness centrality), and consequence severity (expected loss value), avoiding single - indicator deviation, and fusing mechanical state, topology, and temporal data in real - time to adapt to disaster evolution.

[0119] After step 4, the embodiment of the present invention further includes:

[0120] Step 5: Determine the blocking threshold according to the criticality score of each shortest path, and determine the shortest paths with criticality scores greater than the blocking threshold as critical paths.

[0121] Among them, determining the blocking threshold according to the criticality score of each shortest path is specifically as follows: Obtain the maximum criticality score and determine the mean value of the criticality scores of all shortest paths; then determine the blocking threshold according to the maximum criticality score and the mean value of all criticality scores, as shown in formula (11).

[0122] (11)

[0123] In the formula, is the blocking threshold, and its essence is an adaptive dynamic blocking threshold; is the maximum criticality score; is the total number of all shortest paths; is the mean value of the criticality scores of all shortest paths; when , it is determined that is a critical path and triggers blocking, ensuring that blocking measures are taken only for paths that are significantly higher than the current average risk and have extremely high criticality scores themselves, optimizing resource allocation. The designed prevents both insufficient response to a single extreme risk and frequent triggering of blocking due to local fluctuations, and can also be dynamically adjusted according to the real - time state of the graph model.

[0124] Step 6: Perform risk blocking on the critical paths.

[0125] The risk blocking strategy of the embodiment of the present invention is shown in Table 2.

[0126] Table 2 Risk blocking strategy for the disaster chain of the reservoir bank slope

[0127] After the key path is blocked for risk in the embodiments of the present invention, the following steps are further included:

[0128] Step 7: Determine the pre-blocking resilience index of the key path according to the criticality score of the key path before risk blocking; determine the post-blocking resilience index of the key path according to the criticality score of the key path after risk blocking.

[0129] In the embodiments of the present invention, the pre-blocking resilience index and the post-blocking resilience index are determined according to the following formula (12).

[0130] (12)

[0131] In the formula, is the resilience index; is the shortest path of the criticality score; is the maximum criticality score; is the total number of all shortest paths. The higher the value, the stronger the resilience of the graph model.

[0132] Step 8: Determine the effectiveness of risk blocking according to the pre-blocking resilience index and the post-blocking resilience index, as shown in formula (13).

[0133] (13)

[0134] In the formula, is the pre-blocking resilience index, is the post-blocking resilience index. When the blocking measure is effective, the improvement of system resilience can significantly reduce disaster losses; when the blocking effect is insufficient, the strategy needs to be adjusted or the blocking intensity needs to be enhanced.

[0135] The present invention constructs a multi-level adaptive risk blocking strategy, and through a closed-loop mechanism of multi-level blocking → resilience optimization → parameter self-learning, realizes the intelligent prevention and control of the disaster chain of the reservoir bank slope and the improvement of system resilience, showing significant advantages in terms of disaster suppression efficiency, engineering cost optimization and interdisciplinary integration, and providing an innovative solution for engineering safety in complex geological environments. The specific effects are as follows:

[0136] (1) The present invention fuses real-time dynamic monitoring data such as hydrological fluctuations (such as the water level drop rate > 0.5 m / d) and deterioration of geotechnical parameters (such as cohesion attenuation), and realizes the dynamic adjustment of the coupling weight with the update of the data. The present invention introduces a mechanical transfer coefficient to quantify the influence of stress diffusion in the fault zone and permeability change of the weak interlayer on the propagation of the disaster chain, and the physical interpretability of the blocking strategy is significantly enhanced.

[0137] (2) The present invention introduces a criticality scoring method that combines betweenness centrality of the fusion path, expected loss value, and dynamic propagation risk probability to evaluate the disaster situation of the shortest path, solving the problem of decision-making bias with a single indicator.

[0138] (3) The present invention uses the collaborative optimization of a multi-level risk blocking strategy to solve the problem that traditional blocking strategies are mostly single-point reinforcement (such as anti-slide piles) with poor blocking effects. The present invention proposes a three-level adaptive risk blocking strategy, which directly intervenes in the triggering and propagation links of the disaster chain through three major mechanisms: hydraulic regulation (reducing seepage pressure), mechanical reinforcement (increasing critical load), and energy dissipation (constructing redundant paths), weakening the energy of disaster propagation and enhancing the global resilience of the system to achieve real-time prevention and control.

[0139] The effectiveness of the method of the present invention will be verified with more specific embodiments below.

[0140] Embodiment 1

[0141] Taking the landslide - surge - dam breach disaster chain of the slope on the bank of a reservoir in a certain domestic area as an example, the calculation results of the disaster nodes and coupling weights of the key path of this disaster chain are as follows, as shown in Table 3.

[0142] Table 3 Definition of disaster nodes and coupling weights in Embodiment 1

[0143] Calculate the criticality score according to the above results , as shown in Table 4.

[0144] Table 4 Calculation results of criticality score and blocking threshold in Embodiment 1

[0145] As can be seen from Table 4, and , triggering the first-level blocking condition, and the measure taken is grouting to reduce the seepage pressure, which will be reduced from 0.83 to 0.3. The blocking effect after implementing the blocking strategy is shown in Table 5, and the resilience index has increased by 94.3%, proving that the blocking measure is effective and the disaster loss can be significantly reduced after blocking.

[0146] Table 5 Verification of blocking effect in Embodiment 1

[0147] Embodiment 2

[0148] Taking the rainfall - seepage - landslide disaster chain of the slope on the bank of a reservoir in a certain domestic area as an example, the calculation results of the disaster nodes and coupling weights of the key path of this disaster chain are shown in Table 6.

[0149] Table 6 Definition of disaster nodes and coupling weights in Example 2

[0150] Calculate the criticality score based on the above results , as shown in Table 7

[0151] Table 7 Calculation results of criticality score and blocking threshold in Example 2

[0152] As can be seen from Table 7 and , triggering the secondary blocking condition, and the measure taken is to reinforce the disaster node , increase the critical load by 20%. The blocking effect after implementing the blocking strategy is shown in Table 8, and the toughness index has increased by 140%, proving that the blocking measure is effective

[0153] Table 8 Verification of blocking effect in Example 2

[0154] Example 3

[0155] Taking the earthquake - fault activation - dam break disaster chain of a hydropower project reservoir bank slope in China as an example, the calculation results of the disaster nodes and coupling weights on the critical path are shown in Table 9

[0156] Table 9 Definition of disaster nodes and coupling weights in Example 3

[0157] Calculate the criticality score based on the above results , as shown in Table 10

[0158] Table 10 Calculation results of criticality score and blocking threshold in Example 3

[0159] As can be seen from Table 10 and , triggering the tertiary blocking condition, and the measure taken is to add seismic piers nodes , used to monitor the mechanical response of the piers in real - time. Thus, the new path means that the seismic energy bypasses the fault fracture zone ( ), and is directly transmitted to the dam body through the piers, reducing the load of the original path. The blocking effect after implementing the blocking strategy is shown in Table 11, and the toughness index has increased by 130%, proving that the blocking measure is effective and can significantly reduce the disaster losses after blocking

[0160] Table 11 Verification of blocking effect in Example 3

[0161] The comparison charts of the key scores and the change rates of the toughness index after blocking in the above-mentioned Examples 1 to 3 are as Figure 2 shown.

[0162] The improvement of toughness after blocking in the above three examples all exceeded 90%, meeting the goal. By means of the dynamic blocking threshold and the multi-level blocking strategy, the key score was significantly reduced (average decrease -56.8%). The blocking measures directly intervened in the triggering and propagation links of the disaster chain through three major mechanisms: hydraulic regulation (reducing seepage pressure), mechanical reinforcement (increasing the critical load), and energy dissipation (constructing redundant paths), weakened the disaster propagation energy, and enhanced the global toughness of the system. The scientificity and practicability of the multi-level blocking strategy in complex engineering scenarios were verified through multi-scenario disaster chain data.

[0163] The above are only several embodiments of the present invention, and do not impose any form of limitation on the present invention. Although the present invention is disclosed as above with preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art, without departing from the scope of the technical solution of the present invention, making some changes or modifications using the technical content disclosed above is equivalent to equivalent implementation cases and all fall within the scope of the technical solution.

Claims

1. A real-time prevention and control method for the disaster chain of the reservoir bank slope, characterized in that, Including: Step 1: Determine multiple disaster nodes of the reservoir bank slope, and construct a graph model of the reservoir bank slope disaster chain according to the multiple disaster nodes; Step 2: Determine the correlation degree and mechanical transfer coefficient between two disaster nodes corresponding to each edge in the graph model, and determine the coupling weight of each edge according to the correlation degree and mechanical transfer coefficient; Step 3: Preset the starting disaster nodes and ending disaster nodes of multiple disaster chains in the graph model, and determine the shortest path and the betweenness centrality of the shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight; Step 4: Determine the criticality score of each shortest path according to the betweenness centrality, loss expectation value and dynamic propagation risk probability.

2. The real-time prevention and control method for the disaster chain of the reservoir bank slope according to claim 1, characterized in that After step 4, it further includes: Determine the blocking threshold according to the criticality score of each shortest path, and determine the shortest path with the criticality score greater than the blocking threshold as the critical path; Perform risk blocking on the critical path.

3. The real-time prevention and control method for the disaster chain of the reservoir bank slope according to claim 2, characterized in that, Determine the blocking threshold according to the criticality score of each shortest path, specifically: Obtain the maximum criticality score and determine the average value of all criticality scores; Determine the blocking threshold according to the maximum criticality score and the average value of all criticality scores.

4. The real-time prevention and control method for the disaster chain of the reservoir bank slope according to claim 2, wherein After performing risk blocking on the critical path, it further includes: Determine the pre-blocking resilience index of the critical path according to the criticality score of the critical path before risk blocking; Determine the post-blocking resilience index of the critical path according to the criticality score of the critical path after risk blocking; Determine the effectiveness of risk blocking according to the pre-blocking resilience index and the post-blocking resilience index.

5. The real-time prevention and control method for the disaster chain of the reservoir bank slope according to claim 1, characterized in that Determine the mechanical transfer coefficient between two disaster nodes corresponding to each edge in the graph model, specifically: Determine the elastic modulus at two disaster nodes corresponding to each edge in the graph model; Determine the tangent modulus of the stress-strain curve of the reservoir bank slope rock mass, and determine the mechanical transfer coefficient between two disaster nodes corresponding to each edge in the graph model according to the elastic modulus and the tangent modulus.

6. The real-time prevention and control method for the disaster chain of the reservoir bank slope according to claim 1, characterized in that Determine the coupling weight of each edge according to the correlation degree and mechanical transfer coefficient, specifically: Determine the rock mass bearing capacity according to the critical load of the reservoir bank slope rock mass and the maximum load borne at the current moment; Determine the coupling weight of each edge according to the product of the correlation degree, mechanical transfer coefficient and the rock mass bearing capacity.

7. The real-time prevention and control method for the disaster chain of the reservoir bank slope according to claim 1, characterized in that, Determine the shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight, specifically: Determine the initial shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight; When the change amount of the coupling weight in the initial shortest path is greater than the preset threshold, determine the disaster nodes corresponding to the adjacent edges of the coupling weight; Update the initial shortest path according to the disaster nodes corresponding to the adjacent edges of the coupling weight to obtain the shortest path between each pair of starting disaster nodes and ending disaster nodes.

8. The real-time prevention and control method for the disaster chain of the reservoir bank slope according to claim 7, characterized in that, Determine the initial shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight, specifically: Determine the reciprocal of the rock mass bearing capacity according to the critical load of the reservoir bank slope rock mass and the maximum load borne at the current moment; Determine the activation probability of the disaster nodes corresponding to the coupling weight; Determine the initial shortest path between each pair of starting disaster nodes and ending disaster nodes according to the coupling weight, the reciprocal of the bearing capacity of the rock mass, and the activation probability.

9. The real-time prevention and control method for the disaster chain of the reservoir bank slope according to claim 1, characterized in that Determine the criticality score of each shortest path according to the betweenness centrality, the expected loss value, and the dynamic propagation risk probability, specifically: Determine the suppression parameter of each shortest path according to the length of each shortest path and the average value of the lengths of all shortest paths; Determine the criticality score of each shortest path according to the suppression parameter, the betweenness centrality, the expected loss value, and the dynamic propagation risk probability of each shortest path.

10. The real-time prevention and control method for the disaster chain of the reservoir bank slope according to claim 9, characterized in that, The method for determining the dynamic propagation risk probability is: Determine the path correlation strength according to the product of multiple coupling weights on the shortest path; Determine the probability activation superposition amount according to the activation probabilities of multiple disaster nodes on the shortest path; Determine the dynamic propagation risk probability of the shortest path according to the path correlation strength and the probability activation superposition amount.

Images