Optimized layout method of closed conduit-side slope system based on weak intercalated layer control

Abstract

The invention discloses a weak intercalated layer control-based closed conduit-slope system optimization layout method, which comprises the following steps of S1, geological survey and parameter acquisition: accurately positioning a weak intercalated layer through a geological radar and drilling coring, and determining a rock-soil body strength parameter by combining a laboratory rock-soil mechanical test, geological radar data and rock-soil body strength parameters are analyzed through a random forest algorithm, and the interlayer space distribution rule is predicted. By combining geological survey, numerical simulation, intelligent optimization and dynamic monitoring technologies, scientific layout and adjustment of a closed conduit-side slope system are realized, a weak intercalated layer seepage path is accurately cut off, the safety coefficient of the side slope is effectively improved, meanwhile, expansion of a plastic zone is effectively inhibited, settlement of an overlying soil body or large-range deformation of the side slope is reduced, and the stability of the side slope is improved. The seepage stability and the stress slow release capacity of the system are enhanced, the safety of the water diversion project in a complex geological area is improved, and an operable technical path and basis can be provided for layout of similar projects in the future.

Description

Technical Field

[0001] The present invention relates to the technical field of culvert - slope systems, and particularly to an optimized layout method for a culvert - slope system based on the control of weak interlayers. Background Art

[0002] At present, to alleviate the problem of uneven spatio - temporal distribution of water resources, a large number of water diversion projects are carried out in mountainous and hilly areas in the central and western regions, including various forms such as channels, tunnels, and water conveyance culverts. Among them, the water conveyance line often needs to cross high - steep slope areas with large terrain undulations and complex geological structures. Especially in poor geological sections with widely distributed weak interlayers, the slope stability problem has become a key restricting factor for the safe operation of water diversion projects. Weak interlayers often exhibit characteristics such as strong permeability and low strength. Under rainfall conditions, they are prone to becoming the main channels for rainwater infiltration, resulting in the rapid accumulation of pore water pressure, triggering strength weakening and shear slip, and then inducing slope deformation or even overall instability, seriously threatening the safe and stable operation of water conveyance structures. To adapt to complex terrains, some projects adopt shallow - buried closed water conveyance structures (i.e., culverts) as alternative solutions to open channels or tunnels. Such culverts are usually made of reinforced concrete, have good anti - seepage performance, and are buried inside the slope. Although the main function of such culverts is water conveyance, due to their large structural stiffness and moderate burial depth, they inevitably interact with the surrounding strata in the slope - weak interlayer system, showing a coupled influence on the seepage path, stress field distribution, and deformation characteristics.

[0003] However, in existing engineering designs, culverts are mostly regarded as single water conveyance structures, ignoring their regulatory role in the mechanical response of slopes, and lacking systematic analysis and optimization methods, resulting in a lack of scientific basis during the layout process and the problem of unreasonable layout positions. On the one hand, the layout of culverts may trigger new stress concentration areas due to disturbing the original seepage path, and even induce the settlement of overlying soil. On the other hand, if the seepage path in the weak interlayer cannot be reasonably cut off or the development of its plastic zone cannot be controlled, the positive effect of the culvert on slope stability may not be effectively exerted. Moreover, most existing studies focus on the disturbance of the slope body during tunnel construction or the influence of drainage - type channel systems on slope stability, lacking in - depth research on the fluid - solid coupling mechanism and stability regulation effect between culverts and weak - interlayer slope systems during the operation period. Considering the above - mentioned situations, this application proposes an optimized layout method for a culvert - slope system based on the control of weak interlayers. Summary of the Invention

[0004] Based on the technical problems existing in the background art, the present invention proposes an optimized layout method for a culvert - slope system based on the control of weak interlayers.

[0005] An optimized layout method for a culvert - slope system based on the control of weak interlayers proposed by the present invention includes the following steps:

[0006] S1: Geological Survey and Parameter Acquisition: Use ground penetrating radar and core drilling to accurately locate the weak interlayer. Combine laboratory geotechnical mechanics tests to determine the strength parameters of the rock and soil mass. Use the random forest algorithm to analyze the ground penetrating radar data and the strength parameters of the rock and soil mass to predict the spatial distribution law of the interlayer;

[0007] S2: Numerical Model Construction: Based on the monitoring data in S1, establish a three-dimensional fluid-solid coupling model of the slope-culvert. Set the seepage-stress coupling boundary conditions, and use the particle swarm optimization algorithm (PSO) to invert the true parameters of the rock and soil mass, and dynamically calibrate the accuracy of the three-dimensional fluid-solid coupling model of the slope-culvert;

[0008] S3: Rainfall Seepage Simulation: Through the three-dimensional fluid-solid coupling model of the slope-culvert established in S2, simulate the process of pore water pressure accumulation and plastic zone expansion under different rainfall intensities, and combine the LSTM time series model to predict the long-term evolution trend of the seepage field;

[0009] S4: Generation of Intelligent Culvert Layout Scheme: According to the analysis results of the seepage field and stress field in S3, give priority to laying the culvert in the key turning area of the weak interlayer or above the seepage convergence path, and the burial depth is 1.5-2 times the thickness of the interlayer. Use the genetic algorithm (GA) to automatically optimize the layout coordinates with the maximization of FoS as the goal, and screen out the culvert layout area that can most weaken the weakening effect of the weak interlayer and inhibit the evolution of the slip zone;

[0010] S5: Seepage Regulation and Multi-field Verification: Simulate the local seepage field around the culvert through CFD, optimize the layout of drainage holes to form an efficient "interference zone". At the same time, combine digital twin technology to compare the numerical simulation results with the real-time monitoring data, and dynamically correct the seepage control strategy;

[0011] S6: Anti-disturbance Test under Extreme Conditions: Simulate the slope response under the compound condition of rainstorm-earthquake through the three-dimensional fluid-solid coupling model of the slope-culvert, and introduce the robust control theory to optimize the parameter tolerance of the scheme to ensure the system stability under extreme conditions;

[0012] S7: Intelligent Construction and Real-time Monitoring: During the construction period, use microseismic monitoring and fiber optic sensing (BOTDR) to control the excavation disturbance, deploy an Internet of Things (IoT) sensor network (GNSS, piezometer) to collect displacement and pore water pressure data in real time, and automatically track the development of slope surface cracks through UAV aerial survey and AI image recognition;

[0013] S8: Dynamic Feedback and Correction: Based on the Bayesian update algorithm, fuse the monitoring data in S7 and dynamically correct the parameters of the three-dimensional fluid-solid coupling model of the slope-culvert;

[0014] S9: Effect Evaluation: Compare the simulation prediction and the measured data, evaluate the improvement effect of slope stability, and construct a knowledge graph (Neo4j) to archive geological parameters, optimization schemes and monitoring data.

[0015] Preferably, in S1, the random forest algorithm is used to analyze the geological radar data and the geotechnical strength parameters, and the specific logical steps for predicting the spatial distribution law of the interlayer are as follows:

[0016] S101: Collect geological radar data, geotechnical parameters and auxiliary data. The geological radar data includes waveform amplitude and two-way travel time raw data. The geotechnical parameters include cohesion (c) and internal friction angle (φ) strength parameters. The auxiliary data includes topographic elevation, groundwater level and existing geological profiles, and extract characteristic data from the collected data. The extracted characteristic data can significantly reflect the physical and mechanical properties and spatial distribution law of the weak interlayer;

[0017] S102: Divide the data set collected in S101 into a training set (70%), a validation set (15%) and a test set (15%), and generate multiple decision trees through bootstrap sampling. Each decision tree only uses some random features, and input the training set, validation set and test set into the decision tree in turn to train, validate and test the decision tree;

[0018] S103: Input the geological radar data and geotechnical parameters of the unexplored area into the trained decision tree. The decision tree generates a probability map of the existence of the interlayer and outputs it. If the probability > 80%, it is determined as an interlayer.

[0019] Preferably, the specific logical steps of S2 are as follows:

[0020] S201: Establish a three-dimensional grid of slope - culvert based on the exploration data in S1, and densify the weak interlayer area;

[0021] S202: Conduct fluid - solid coupling solution and PSO parameter inversion. The formula used for fluid - solid coupling solution is as follows:

[0022] Seepage field: where k is the permeability coefficient, h is the water head, S s is the storage rate, and t is the time;

[0023] Stress field: where σ ' is the effective stress tensor, p is the pore water pressure, i is the unit tensor, ρ is the density of the geotechnical body, and g is the gravitational acceleration vector;

[0024] The process of PSO parameter inversion is as follows:

[0025] S2021: Initialize the particle swarm and randomly generate N groups of parameter combinations;

[0026] S2022: Substitute each group of parameters into the three - dimensional grid of slope - culvert established in S201, calculate f(θ), and evaluate the fitness;

[0027] S2023: Record the historical optimal solution (P best ) of each particle and the global optimal solution (g best );

[0028] S2024: Adjust the particle positions;

[0029] S2025: Reach the maximum number of iterations, termination condition;

[0030] The formulas used in its entire process are as follows:

[0031] Objective function:

[0032] where θ = [E, v, c, φ, k], is the parameter vector to be inverted, and w1 and w2 are the weight coefficients of pore pressure and displacement;

[0033]

[0034] where w is the inertia weight; c1, c2 are learning factors; r1, r2 ∼ U(0, 1);

[0035] S203: Calculate the partial derivatives of the objective function with respect to each parameter Sort to determine the dominant parameters;

[0036] S204: Substitute the inverted parameters into the slope - culvert three - dimensional grid, compare the displacement / pore pressure curves of simulation and monitoring, displacement error ≤ 10%, pore pressure error ≤ 15%;

[0037] S205: When the deviation between the newly added monitoring data and the simulation results exceeds the threshold, dynamically update the model parameters.

[0038] Preferably, the specific logical steps of S3 are as follows:

[0039] S301: Input rainfall conditions into the slope - culvert three - dimensional fluid - solid coupling model. The parameters input for the rainfall conditions are three intensities: light rain (5 mm / h), moderate rain (20 mm / h), and heavy rain (50 mm / h), with a duration of 6 - 72 hours;

[0040] S302: Coupled - solve the seepage field and Biot consolidation, output the pore water pressure p(x, t) and plastic strain εp(x, t), and record the pore pressure time - history data {pt} and plastic zone area Ap(t) at key points (such as the middle of the weak interlayer),

[0041] The formulas used are as follows:

[0042] Seepage field control equation:

[0043] where k is the permeability coefficient, h is the water head, S s is the storage rate, t is the time, z is the elevation head, ρw is the density of water, g is the acceleration due to gravity, and p is the pore water pressure;

[0044] The formula for judging the plastic zone is as follows: where τ is the shear stress and σ ' is the effective normal stress;

[0045] S303: Predict the long-term seepage field through the LSTM model. Input the historical rainfall sequence {R t-24 , …… R t} and the pore pressure monitoring data {P t-24 , …… P t} into the LSTM model in the prior art. The LSTM model outputs the pore pressure prediction for the next 72 hours

[0046] Preferably, in step S4, the genetic algorithm (GA) is used to automatically optimize the layout coordinates with the maximization of FoS as the goal, and the canal layout area that can most weaken the weakening effect of the soft interlayer and inhibit the evolution of the sliding zone is screened out. The specific logical steps are as follows:

[0047] S401: Calculate the maximized safety factor (F0S), where X = (x, y, z) is the center coordinate of the canal, τ max = c + σ ' tanφ is the maximum shear strength, and τ actual is the actual shear stress;

[0048] S402: Set the constraint conditions that the canal shall not exceed the slope range, x ∈ [x min , x max , y ∈ [y min , y max , z ∈ [z min , z max ;

[0049] S403: Initialize the population. Each individual is the canal coordinate X = (x, y, z), using real number coding, and randomly generate N groups of coordinates to ensure that the geometric constraints are satisfied;

[0050] S404: Call the three-dimensional fluid-solid coupling model of slope-canal, and for each individual Xi, run the model to calculate FoS and the plastic zone area Ap;

[0051] S405: If the engineering constraints (insufficient burial depth) are violated, then significantly reduce the fitness, and its calculation formula is: Fitness = FOS - λ·Penalty, where λ is the penalty coefficient;

[0052] S406: Minimize the plastic strain (εp) corresponding to the high FoS region calculated in S401 to directly weaken the weakening effect of the soft interlayer, thereby screening out the optimal layout area for the culvert.

[0053] Preferably, the specific logical steps of S5 are as follows:

[0054] S501: Extract the three-dimensional geometry (BIM model), the permeability coefficient k of the soft interlayer, and the boundary flow velocity of the culvert and its surrounding soft interlayer, and input them into the CFD model;

[0055] S502: The CFD model uses parametric scanning and the response surface method (RSM) to find a drainage hole layout plan that can reduce the head gradient in the soft interlayer area by more than 30%;

[0056] S503: The CFD model outputs the optimal drainage hole configuration and the expected seepage field distribution;

[0057] S504: Real-time collect the pore water pressure p(t) and the flow velocity u(t), and the CFD model receives the monitoring data and calculates the theoretical seepage field under the current state;

[0058] S505: Compare the real-time monitored data with the simulated data and trigger the following actions: If the deviation > 10%, calibrate k and the effective area Ae of the drainage hole with PSO, and automatically adjust the opening of the drainage hole through the electric valve to maintain p < p crit , realizing the adaptive control of the pore water pressure.

[0059] Preferably, the specific logical steps of S6 are as follows:

[0060] S601: Input the rainstorm and earthquake parameters into the three-dimensional fluid-structure interaction model of the slope-culvert for simulation. The rainstorm is a rainfall intensity of 50 mm / h for 72 hours, and the slope surface is set as a flow boundary Input the El Centro wave (PGA = 0.3g) for vibration, and set the bottom as a free field boundary;

[0061] S602: Use the coupled Biot equation for dynamic expansion and perform plastic judgment. The formula is:

[0062]

[0063] Plastic judgment: where τ is the shear stress, σ ' is the effective normal stress;

[0064] S603: Monitor the horizontal displacement δx at the top of the slope, and evaluate the liquefaction risk according to the pore pressure ratio, r u > 0.8 is dangerous;

[0065] S604: Optimize the parameters of the scheme according to robust control theory: displacement δx < 50 mm, pore pressure ratio r u < 0.6, parameter fluctuation range k ∈ [k0 ± 30%], c ∈ [c0 ± 20%], and modify the parameters of the three-dimensional fluid-solid coupling model of the slope-culvert.

[0066] Preferably, in the above S7, for microseismic monitoring, use microseismic probes, arrange them around the potential slip surface of the slope, the number is ≥ 8, the coverage extends 1.5 times the slope height outside the excavation surface, for distributed optical fiber sensing, arrange distributed optical fibers along the axis of the culvert and the weak interlayer, the spatial resolution ≤ 0.5 m, and the strain measurement accuracy is ±5 με;

[0067] When controlling the excavation disturbance, it is necessary to carry out risk level classification as follows:

[0068] Index Safe (Level I) Warning (Level II) Dangerous (Level III) Microseismic event rate <5 times per hour 5 - 10 times per hour > 10 times per hour Maximum fiber strain <200 με 200 - 500 με > 500 με

[0069] According to the risk level response measures, during level I, carry out normal excavation, the monitoring frequency is 1 time / 30 minutes, during level II, reduce the excavation footage, during level III, immediately suspend the excavation, implement temporary support, and inject chemical grout to reinforce the weak interlayer.

[0070] Preferably, in the above S8, the Bayesian update algorithm is used to dynamically correct the key parameters θ = [k, c, φ, E] of the three-dimensional fluid-solid coupling model of the slope-culvert, and the formula it uses is:

[0071] Among them, P(θ) is the prior distribution of the parameter, p(θ|D) is the likelihood function, which is used for the matching degree between the monitoring data and the simulation results, Among them, σ u σ p is the standard deviation of the monitoring error.

[0072] Compared with the existing technologies, the beneficial effects of the present invention are:

[0073] 1. By combining geological exploration and numerical simulation, reasonably arrange the culvert, adjust the seepage path and hydraulic environment, significantly slow down the accumulation speed of pore water pressure, effectively reduce the risk of seepage-induced instability in the weak interlayer area, and reduce the possibility of slope sliding and deformation;

[0074] 2. Through the collaborative optimization of the three-dimensional fluid-solid coupling model and the intelligent algorithm for the layout position of the culvert, make it in the key turning area of the weak interlayer or above the seepage convergence path, effectively cut off the original seepage path, control the flow velocity and scale of the groundwater flow channel, form an "interference zone" with the function of water retention or flow diversion, effectively inhibit the expansion of the plastic zone, and at the same time enhance the seepage stability and stress relief ability of the system;

[0075] 3. By combining digital twin technology with real-time monitoring data, the model parameters are dynamically corrected and the drainage volume is adaptively regulated, so that under high-intensity rainfall conditions, the optimally arranged culverts show stronger anti-disturbance ability, effectively slowing down the development speed of the plastic zone, reducing the settlement of the overlying soil or large-scale deformation of the slope, enhancing the redundancy of the system, and improving the safety of the water diversion project in complex geological areas;

[0076] The present invention realizes the scientific layout and adjustment of the culvert-slope system by combining geological exploration, numerical simulation, intelligent optimization and dynamic monitoring technologies, accurately cuts off the seepage path of the weak interlayer, effectively improves the safety factor of the slope, and at the same time effectively inhibits the expansion of the plastic zone, reduces the settlement of the overlying soil or large-scale deformation of the slope, enhances the seepage stability and stress relief ability of the system, improves the safety of the water diversion project in complex geological areas, and can provide an operable technical path and basis for the layout of future similar projects. Brief Description of the Drawings

[0077] Figure 1 It is a flowchart of an optimized layout method for a culvert-slope system based on weak interlayer control proposed by the present invention. Detailed Embodiments

[0078] The present invention will be further explained below with reference to specific embodiments.

[0079] Embodiment

[0080] Refer to Figure 1 , this embodiment proposes an optimized layout method for a culvert-slope system based on weak interlayer control, including the following steps:

[0081] S1: Geological survey and parameter acquisition: accurately locate the weak interlayer through ground penetrating radar and core drilling, determine the strength parameters of the rock and soil mass by combining laboratory geotechnical mechanics tests, and use the random forest algorithm to analyze the ground penetrating radar data and the strength parameters of the rock and soil mass to predict the spatial distribution law of the interlayer;

[0082] The specific logical steps of using the random forest algorithm to analyze the ground penetrating radar data and the strength parameters of the rock and soil mass to predict the spatial distribution law of the interlayer are as follows:

[0083] S101: Collect ground penetrating radar data, rock and soil mass parameters and auxiliary data. The ground penetrating radar data includes waveform amplitude and two-way travel time raw data. The rock and soil mass parameters include cohesion (c), internal friction angle (φ) strength parameters. The auxiliary data includes terrain elevation, groundwater level and existing geological profiles, and extract characteristic data from the collected data. The extracted characteristic data can obviously reflect the physical and mechanical properties and spatial distribution law of the weak interlayer;

[0084] S102: Divide the dataset collected in S101 into a training set (70%), a validation set (15%), and a test set (15%). Generate multiple decision trees through bootstrap sampling. Each decision tree only uses some random features, and input the training set, validation set, and test set into the decision tree in sequence to train, validate, and test the decision tree;

[0085] S103: Input the ground penetrating radar data and geotechnical parameters of the unexplored area into the trained decision tree. The decision tree generates a probability map of the existence of the interlayer and outputs it. If the probability > 80%, it is determined as an interlayer;

[0086] S2: Numerical model construction: Based on the monitoring data in S1, establish a three-dimensional fluid-solid coupling model of the slope-culvert, set the seepage-stress coupling boundary conditions, and use the particle swarm optimization algorithm (PSO) to invert the true parameters of the geotechnical body and dynamically calibrate the accuracy of the three-dimensional fluid-solid coupling model of the slope-culvert;

[0087] The specific logical steps are as follows:

[0088] S201: Based on the exploration data in S1, establish a three-dimensional grid of the slope-culvert and densify the weak interlayer area;

[0089] S202: Conduct fluid-solid coupling solution and PSO parameter inversion. The formula used for the fluid-solid coupling solution is as follows:

[0090] Seepage field: where k is the permeability coefficient, h is the water head, S s is the storage rate, and t is the time;

[0091] Stress field: where σ ' is the effective stress tensor, p is the pore water pressure, i is the unit tensor, ρ is the density of the geotechnical body, and g is the gravitational acceleration vector;

[0092] The process of PSO parameter inversion is as follows:

[0093] S2021: Initialize the particle swarm and randomly generate N groups of parameter combinations;

[0094] S2022: Substitute each group of parameters into the three-dimensional grid of the slope-culvert established in S201, calculate f(θ), and evaluate the fitness;

[0095] S2023: Record the historical optimal solution (P best ) of each particle and the global optimal solution (g best );

[0096] S2024: Adjust the particle positions;

[0097] S2025: Reach the maximum number of iterations, which is the termination condition;

[0098] The formulas used in the whole process are as follows:

[0099] Objective function:

[0100] where θ = [E, v, c, φ, k] is the parameter vector to be inverted, and w1 and w2 are the weight coefficients of pore pressure and displacement;

[0101]

[0102] where w is the inertia weight; c1, c2 are learning factors; r1, r2 ~ U(0, 1);

[0103] S203: Calculate the partial derivatives of the objective function with respect to each parameter Sort to determine the dominant parameters;

[0104] S204: Substitute the inverted parameters into the slope-culvert three-dimensional grid, and compare the displacement / pore pressure curves of simulation and monitoring. The displacement error ≤ 10%, and the pore pressure error ≤ 15%;

[0105] S205: When the deviation between the newly added monitoring data and the simulation results exceeds the threshold, the model parameters are dynamically updated;

[0106] S3: Rainfall seepage simulation: Simulate the process of pore water pressure accumulation and plastic zone expansion under different rainfall intensities through the slope-culvert three-dimensional fluid-solid coupling model established in S2, and predict the long-term seepage field evolution trend in combination with the LSTM time series model;

[0107] The specific logical steps are as follows:

[0108] S301: Input the rainfall conditions into the slope-culvert three-dimensional fluid-solid coupling model. The parameters input for the rainfall conditions are three intensities: light rain (5 mm / h), moderate rain (20 mm / h), and heavy rain (50 mm / h), with a duration of 6 - 72 hours;

[0109] S302: Coupled solution of the seepage field and Biot consolidation, output the pore water pressure p(x, t) and plastic strain εp(x, t), and record the pore pressure time history data {pt} and plastic zone area Ap(t) at key points (such as the middle of the weak interlayer),

[0110] The formulas used are as follows:

[0111] Seepage field control equation:

[0112] where k is the permeability coefficient, h is the water head, S s is the storage rate, t is the time, z is the elevation head, ρw is the density of water, g is the acceleration due to gravity, and p is the pore water pressure;

[0113] The plastic zone judgment formula is as follows: Where τ is the shear stress, and σ ' is the effective normal stress;

[0114] S303: Predict the long-term seepage field through the LSTM model. Input the historical rainfall sequence {R t-24 , …… R t} and the pore pressure monitoring data {P t-24 , …… P t} into the LSTM model in the prior art. The LSTM model outputs the pore pressure prediction for the next 72 hours

[0115] S4: Generate the intelligent layout plan for the culvert: According to the analysis results of the seepage field and stress field in S3, preferentially lay the culvert in the key turning area of the weak interlayer or above the seepage convergence path, with a buried depth of 1.5 - 2 times the thickness of the interlayer, and use the genetic algorithm (GA) to automatically optimize the layout coordinates with the maximization of FoS as the goal, and screen out the culvert layout area that can most weaken the weakening effect of the weak interlayer and inhibit the evolution of the sliding zone;

[0116] Use the genetic algorithm (GA) to automatically optimize the layout coordinates with the maximization of FoS as the goal, and screen out the culvert layout area that can most weaken the weakening effect of the weak interlayer and inhibit the evolution of the sliding zone. The specific logical steps are as follows:

[0117] S401: Calculate the maximum safety factor (F0S), where X = (x, y, z) is the center coordinate of the culvert, and τ max = c + σ ' tanφ is the maximum shear strength, and τ actual is the actual shear stress;

[0118] S402: Set the constraint conditions. The culvert shall not exceed the slope range, x ∈ [x min , x max , y ∈ [y min , y max , z ∈ [z min , z max ;

[0119] S403: Initialize the population. Each individual is the culvert coordinate X = (x, y, z), using real number coding, and randomly generate N groups of coordinates to ensure that the geometric constraints are met;

[0120] S404: Call the three-dimensional fluid-solid coupling model of the slope - culvert, and for each individual Xi, run the model to calculate FoS and the plastic zone area Ap;

[0121] S405: If the engineering constraints (insufficient burial depth) are violated, the fitness is significantly reduced, and its calculation formula is: Fitness = FOS - λ·Penalty, where λ is the penalty coefficient;

[0122] S406: Minimize the plastic strain (εp) corresponding to the high FoS region calculated in S401, directly weaken the weakening effect of the weak interlayer, and thus screen out the optimal layout area of the culvert;

[0123] S5: Seepage regulation and multi-field verification: Through CFD simulation of the local seepage field around the culvert, optimize the layout of drainage holes to form an efficient "interference zone", and at the same time, combined with digital twin technology, compare the numerical simulation results with the real-time monitoring data to dynamically correct the seepage control strategy;

[0124] The specific logical steps are as follows:

[0125] S501: Extract the three-dimensional geometry (BIM model), the permeability coefficient k of the weak interlayer, and the boundary flow velocity of the culvert and its surrounding areas, and input them into the CFD model;

[0126] S502: The CFD model uses parametric scanning and the response surface method (RSM) to find a drainage hole layout plan that can reduce the hydraulic head gradient in the weak interlayer area by more than 30%;

[0127] S503: The CFD model outputs the optimal drainage hole configuration and the expected seepage field distribution;

[0128] S504: Real-time collect the pore water pressure p(t) and the flow velocity u(t), and the CFD model receives the monitoring data and calculates the theoretical seepage field under the current state;

[0129] S505: Compare the real-time monitored data with the simulated data and trigger the following actions: If the deviation > 10%, calibrate the permeability coefficient k of the weak interlayer and the effective area Ae of the drainage hole with PSO, and automatically adjust the opening of the drainage hole through an electric valve to maintain p < p crit , to achieve adaptive control of the pore water pressure;

[0130] S6: Extreme condition anti-disturbance test: Simulate the slope response under the compound condition of rainstorm - earthquake through the three-dimensional fluid - solid coupling model of slope - culvert, and introduce robust control theory to optimize the parameter tolerance of the scheme to ensure the system stability under extreme conditions;

[0131] The specific logical steps are as follows:

[0132] S601: Input the rainstorm and earthquake parameters into the three-dimensional fluid - solid coupling model of slope - culvert for simulation. The rainstorm is a rainfall intensity of 50 mm / h, lasting for 72 hours, and the slope surface is set as a flow boundary Vibration input of El Centro wave (PGA = 0.3g), with the bottom set as the free-field boundary;

[0133] S602: Use the coupled Biot equation for dynamic expansion and perform plasticity judgment. The formula is:

[0134]

[0135] Plasticity judgment: where τ is the shear stress and σ ' is the effective normal stress;

[0136] S603: Monitor the horizontal displacement δx at the top of the slope and evaluate the liquefaction risk according to the pore pressure ratio. r > 0.8 is dangerous; u >0.8 is dangerous;

[0137] S604: Optimize the scheme parameters according to the robust control theory: displacement δx < 50mm, pore pressure ratio r u <0.6, the parameter fluctuation range k ∈ [k0 ± 30%], c ∈ [c0 ± 20%], and modify the parameters of the slope-culvert three-dimensional fluid-solid coupling model;

[0138] S7: Intelligent construction and real-time monitoring: During the construction period, use microseismic monitoring and fiber optic sensing (BOTDR) to control the excavation disturbance, deploy an Internet of Things (IoT) sensor network (GNSS, piezometers) to collect displacement and pore water pressure data in real time, and automatically track the development of slope surface cracks through UAV aerial survey and AI image recognition;

[0139] For microseismic monitoring, it is a microseismic probe, arranged around the potential slip surface of the slope, with the number ≥ 8, and the coverage extends 1.5 times the slope height outside the excavation surface. Distributed optical fibers are arranged along the axis of the culvert and the weak interlayer for fiber optic sensing, with a spatial resolution ≤ 0.5m and a strain measurement accuracy of ±5με;

[0140] When controlling the excavation disturbance, it is necessary to conduct risk level classification as follows:

[0141] Index Safe (Level I) Warning (Level II) Dangerous (Level III) Microseismic event rate <5 times per hour 5 - 10 times per hour > 10 times per hour Maximum fiber strain <200 με 200 - 500 με > 500 με

[0142] According to the risk level response measures, during level I, normal excavation is carried out with a monitoring frequency of 1 time per 30 minutes. During level II, the excavation footage is reduced. During level III, excavation is immediately suspended, temporary support is implemented, and chemical grout is injected to reinforce the weak interlayer;

[0143] S8: Dynamic feedback and correction: Based on the Bayesian update algorithm, fuse the monitoring data of S7 and dynamically correct the parameters of the slope-culvert three-dimensional fluid-solid coupling model;

[0144] Among them, the Bayesian update algorithm is used to dynamically correct the key parameters θ = [k, c, φ, E] of the three-dimensional fluid-solid coupling model of the slope-culvert. The formula it uses is as follows:

[0145] Among them, P(θ) is the prior distribution of the parameters, and p(θ|D) is the likelihood function, which is used to measure the matching degree between the monitoring data and the simulation results. Among them, σ u , σ p is the standard deviation of the monitoring error;

[0146] S9: Effect evaluation: Compare the simulation prediction with the measured data, evaluate the improvement effect of the slope stability, and construct a knowledge graph (Neo4j) to file the geological parameters, optimization schemes, and monitoring data;

[0147] Through the combination of geological exploration, numerical simulation, intelligent optimization, and dynamic monitoring technologies, this embodiment realizes the scientific layout and adjustment of the culvert-slope system, accurately cuts off the seepage path of the weak interlayer, effectively improves the safety factor of the slope, simultaneously effectively inhibits the expansion of the plastic zone, reduces the settlement of the overlying soil or large-scale deformation of the slope, enhances the seepage stability and stress relief ability of the system, improves the safety of the water diversion project in complex geological areas, and can provide an operable technical path and basis for the layout of future similar projects.

[0148] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, makes equivalent substitutions or changes, and should be covered by the protection scope of the present invention.

Claims

1. An optimization layout method for a culvert-slope system based on the control of weak interlayers, characterized in that It includes the following steps: S1: Geological survey and parameter acquisition: Use ground-penetrating radar and core drilling to accurately locate the soft interlayer. Combine laboratory geotechnical mechanics tests to determine the strength parameters of the rock and soil mass. Analyze the ground-penetrating radar data and the strength parameters of the rock and soil mass using the random forest algorithm to predict the spatial distribution law of the interlayer; S2: Numerical model construction: Based on the monitoring data in S1, establish a three-dimensional fluid-solid coupling model of the slope-culvert. Set the seepage-stress coupling boundary conditions, and use the particle swarm optimization algorithm (PSO) to invert the true parameters of the rock and soil mass, and dynamically calibrate the accuracy of the three-dimensional fluid-solid coupling model of the slope-culvert; S3: Rainfall seepage simulation: Use the three-dimensional fluid-solid coupling model of the slope-culvert established in S2 to simulate the process of pore water pressure accumulation and plastic zone expansion under different rainfall intensities, and combine the LSTM time series model to predict the long-term evolution trend of the seepage field; S4: Generation of intelligent layout plan for the culvert: According to the analysis results of the seepage field and stress field in S3, preferentially lay the culvert in the key turning area of the soft interlayer or above the seepage convergence path, and the buried depth is 1.5-2 times the thickness of the interlayer. Use the genetic algorithm (GA) to automatically optimize the layout coordinates with the maximization of FoS as the goal, and screen out the culvert layout area that can most weaken the weakening effect of the soft interlayer and inhibit the evolution of the slip zone; S5: Seepage regulation and multi-field verification: Use CFD to simulate the local seepage field around the culvert, optimize the arrangement of drainage holes to form an efficient "interference zone", and at the same time combine digital twin technology to compare the numerical simulation results with the real-time monitoring data, and dynamically correct the seepage control strategy; S6: Anti-disturbance test under extreme working conditions: Use the three-dimensional fluid-solid coupling model of the slope-culvert to simulate the slope response under the combined working conditions of rainstorm-earthquake, and introduce robust control theory to optimize the parameter tolerance of the scheme to ensure the system stability under extreme conditions; S7: Intelligent construction and real-time monitoring: During the construction period, use microseismic monitoring and fiber optic sensing to control the excavation disturbance, arrange an Internet of Things (IoT) sensor network to collect displacement and pore water pressure data in real time, and automatically track the development of slope surface cracks through UAV aerial survey and AI image recognition; S8: Dynamic feedback and correction: Based on the Bayesian update algorithm, fuse the monitoring data in S7, and dynamically correct the parameters of the three-dimensional fluid-solid coupling model of the slope-culvert; S9: Effect evaluation: Compare the simulation prediction and the measured data, evaluate the improvement effect of slope stability, and construct a knowledge graph (Neo4j) to archive geological parameters, optimization schemes and monitoring data.

2. The optimized layout method of the culvert-slope system based on the control of weak interlayers according to claim 1, characterized in that, In the above S1, the specific logical steps of using the random forest algorithm to analyze the ground-penetrating radar data and the strength parameters of the rock and soil mass to predict the spatial distribution law of the interlayer are as follows: S101: Collect ground-penetrating radar data, rock and soil parameters and auxiliary data. The ground-penetrating radar data includes waveform amplitude and two-way travel time raw data. The rock and soil parameters include cohesion (c), internal friction angle (φ) strength parameters. The auxiliary data includes terrain elevation, groundwater level and existing geological profiles, and extract feature data from the collected data. The extracted feature data can significantly reflect the physical and mechanical properties and spatial distribution law of the soft interlayer; S102: Divide the data set collected in S101 into a training set (70%), a validation set (15%), and a test set (15%), and generate multiple decision trees through bootstrap sampling. Each decision tree only uses some random features, and the training set, validation set, and test set are sequentially input into the decision tree for training, validation, and testing of the decision tree. S103: Input the ground penetrating radar data and geotechnical parameters of the unexplored area into the trained decision tree. The decision tree generates a map of the probability of the existence of the interlayer and outputs it. If the probability > 80%, it is determined as an interlayer.

3. The optimized layout method of a culvert-slope system based on the control of weak interlayers according to claim 1, characterized in that, The specific logical steps of S2 are as follows: S201: Based on the exploration data in S1, establish a three-dimensional grid of the slope-culvert, and densify the area of the soft interlayer. S202: Perform fluid-solid coupling solution and PSO parameter inversion. The formula used for the fluid-solid coupling solution is as follows: Seepage field: where k is the permeability coefficient, h is the hydraulic head, S s is the storage rate, and t is the time; Stress field: where σ ' is the effective stress tensor, p is the pore water pressure, i is the unit tensor, ρ is the density of the rock and soil mass, and g is the gravitational acceleration vector; The process of POS parameter inversion is as follows: S2021: Initialize the particle swarm and randomly generate N groups of parameter combinations. S2022: Substitute each group of parameters into the three-dimensional grid of the slope-culvert established in S201, calculate f(θ), and evaluate the fitness. S2023: Record the historical optimal solution (P best ) of each particle and the global optimal solution (g best ); S2024: Adjust the particle position. S2025: Reach the maximum number of iterations, which is the termination condition. The formula used in its entire process is as follows: Objective function: where θ = [E, v, c, φ, k], which is the parameter vector to be inverted, and w1 and w2 are the weight coefficients of pore pressure and displacement. where w is the inertia weight; c1, c2 are the learning factors; r1, r2 ~ U(0,1). S203: Calculate the partial derivatives of the objective function with respect to each parameter Sort to determine the dominant parameters; S204: Substitute the inverted parameters into the three-dimensional grid of the slope-culvert, and compare the displacement / pore pressure curves of the simulation and the monitoring. The displacement error ≤ 10%, and the pore pressure error ≤ 15%. S205: When the deviation between the newly added monitoring data and the simulation result exceeds the threshold, the model parameters are dynamically updated.

4. A method for optimizing the layout of a culvert-slope system based on the control of soft interlayers, characterized in that, The specific logical steps of S3 are as follows: S301: Input rainfall conditions into the three-dimensional fluid-solid coupling model of the slope-culvert. The input parameters for the rainfall conditions are three intensities: light rain (5 mm / h), moderate rain (20 mm / h), and heavy rain (50 mm / h), with a duration of 6 - 72 hours. S302: Couple and solve the seepage field and Biot consolidation, output the pore water pressure p(x,t) and the plastic strain εp(x,t), and record the pore pressure time history data {pt} and the plastic zone area Ap(t) of the key points. The formula used is as follows: Seepage field control equation: where k is the permeability coefficient, h is the hydraulic head, S s is the storage rate, t is the time, z is the elevation head, ρw is the density of water, g is the acceleration due to gravity, and p is the pore water pressure; The formula for judging the plastic zone is as follows: where τ is the shear stress and σ ' is the effective normal stress; S303: Predict the long-term seepage field through the LSTM model, and input the historical rainfall sequence {R t-24 ,……R t} and the pore pressure monitoring data {P t-24 ,……P t} into the LSTM model in the prior art, and the LSTM model outputs the pore pressure prediction for the next 72 hours 5. A method for optimizing the layout of a culvert-slope system based on the control of weak interlayers, characterized in that, In S4, the genetic algorithm (GA) is used to automatically optimize the layout coordinates with the maximization of FoS as the goal, and screen out the culvert layout area that can most weaken the weakening effect of the soft interlayer and inhibit the evolution of the slip zone. The specific logical steps are as follows: S401: Calculate the maximum safety factor (F0S), where X = (x, y, z) is the central coordinate of the culvert, and τ max = c + σ ' tanφ is the maximum shear strength, and τ actual is the actual shear stress; S402: Set constraint conditions that the culvert cannot exceed the slope range, x ∈ [x min , x max , y ∈ [y min , y max , z ∈ [z min , z max ; S403: Initialize the population. Each individual is the culvert coordinate X = (x, y, z), using real number coding, and randomly generate N groups of coordinates to ensure compliance with geometric constraints. S404: Call the three-dimensional fluid-solid coupling model of the slope-culvert, and for each individual Xi, run the model to calculate FoS and the plastic zone area Ap. S405: If the engineering constraints are violated, the fitness is greatly reduced. The calculation formula is: Fitness = FOS - λ·Penalty, where λ is the penalty coefficient. S406: Minimize the plastic strain (εp) corresponding to the high FoS region calculated in S401 to directly weaken the weakening effect of the soft interlayer, thereby screening out the optimal layout area of the culvert.

6. The optimized layout method of a culvert-slope system based on weak interlayer control according to claim 1, characterized in that The specific logical steps of S5 are as follows: S501: Extract the three-dimensional geometry (BIM model), permeability coefficient k, and boundary flow velocity of the culvert and the surrounding soft interlayer, and input them into the CFD model; S502: The CFD model uses parametric scanning and the response surface method (RSM) to find a drainage hole layout plan that can reduce the hydraulic head gradient in the soft interlayer area by 30%+; S503: The CFD model outputs the optimal drainage hole configuration and the expected seepage field distribution; S504: Real-time collect the pore water pressure p(t) and flow velocity u(t), and the CFD model receives the monitoring data to calculate the theoretical seepage field under the current state; S505: Compare the real-time data and the simulated data of the monitoring, and trigger the following actions: If the deviation > 10%, calibrate k and the effective area Ae of the drainage hole with PSO, and automatically adjust the opening of the drainage hole through the electric valve to maintain p < p crit , and achieve the adaptive control of the pore water pressure.

7. A method for optimizing the layout of a culvert-slope system based on the control of weak interlayers, characterized in that, The specific logical steps of S6 are as follows: S601: Input the rainstorm and earthquake parameters into the 3D fluid-solid coupling model of slope-culvert for simulation. The rainstorm has a rainfall intensity of 50 mm / h and lasts for 72 hours, and the slope surface is set as a flow boundary. The El Centro wave (PGA = 0.3g) is input for vibration, and the bottom is set as a free-field boundary; S602: Use the coupled Biot equation for dynamic expansion and perform plastic judgment. The formula is: Plasticity judgment: where τ is the shear stress and σ ' is the effective normal stress; S603: Monitor the horizontal displacement δx at the slope crest and evaluate the liquefaction risk based on the pore pressure ratio r > 0.8 is dangerous; u > 0.8 is dangerous; S604: Optimize the parameters of the scheme according to the robust control theory: displacement δx < 50 mm, pore pressure ratio r u <0.6, with the parameter fluctuation range k ∈ [k0 ± 30%] and c ∈ [c0 ± 20%], modify the parameters of the three-dimensional fluid-solid coupling model of the slope-culvert.

8. A method for optimizing the layout of a culvert-slope system based on the control of weak interlayers, characterized in that, In S7, microseismic monitoring is carried out using microseismic probes. They are arranged around the potential slip surface of the slope, with the number ≥8, and the coverage extends 1.5 times the slope height outside the excavation surface. Distributed optical fibers are arranged along the axis of the culvert and the soft interlayer, with a spatial resolution ≤0.5m and a strain measurement accuracy of ±5με; When controlling the excavation disturbance, it is necessary to carry out risk level classification as shown in the following table: According to the risk level response measures, normal excavation is carried out at level I, with a monitoring frequency of 1 time per 30 minutes, the excavation footage is reduced at level II, and excavation is immediately suspended at level III, temporary support is implemented, and chemical grout is injected to reinforce the soft interlayer.

9. The optimized layout method of the culvert-slope system based on the control of weak interlayers according to claim 1, characterized in that, In the S8, the Bayesian update algorithm is used to dynamically correct the key parameters θ = [k, c, φ, E] of the three-dimensional fluid-structure interaction model of the slope-culvert. The formula it uses is as follows: where P(θ) is the prior distribution of the parameter, and p(θ|D) is the likelihood function, which is used to monitor the matching degree between the measured data and the simulation results. where σ u , σ p is the standard deviation of the monitoring error.

Images