A new prediction method for the degree of closure of a wedge-shaped beam-type grid dam by mud-rock flow

Abstract

The application provides a new type of wedge-shaped beam type grid dam debris flow occlusion degree prediction method, and belongs to the technical field of debris flow disaster prevention and control, and the method comprises the following steps: step one, obtaining basic parameters; step two, designing wedge-shaped beam type grid dam structure parameters; step three, carrying out a water tank model test to obtain measured data of occlusion degree; step four, constructing an occlusion degree prediction model; and step five, applying the model to predict the occlusion degree to calculate the debris flow occlusion degree prediction value. The application takes the wedge-shaped beam type grid dam as the research object, and constructs a debris flow occlusion degree prediction system based on the water tank model test and multi-factor coupling analysis.

Description

Technical Field

[0001] This invention relates to the field of debris flow disaster prevention and control technology, specifically to a novel method for predicting debris flow blockage in wedge-shaped beam grid dams. Background Technology

[0002] Debris flows are a typical sudden geological disaster in mountainous areas, posing a serious threat to towns, transportation routes, water conservancy projects, and more.

[0003] In current debris flow prevention and control, traditional protective structures have significant defects:

[0004] Limitations of solid dams: poor permeability, easy to lose function quickly due to siltation, and require frequent manual dredging; weak impact resistance, easily damaged by long-term debris flow erosion, and may even become a source of secondary disaster materials, making it difficult to meet the long-term disaster prevention needs of mountainous areas.

[0005] Traditional permeable silt-trapping dams have several shortcomings: their designs rely heavily on experience and fail to adequately consider key factors affecting silt flow performance (such as the scale of debris flows, the compatibility of the dam structure with the trajectory of debris flows), leading to strong design subjectivity; their blockage degree and critical blockage discrimination models are incomplete, making them prone to losing their function of intercepting coarse debris flows and discharging fine debris flows due to overflow section blockage; and their structural optimization focuses primarily on stability and scour resistance, failing to improve control capabilities from the perspective of guiding debris flow diversion and energy dissipation, thus limiting their engineering applicability and protective effect.

[0006] With the advancement of major engineering projects in mountainous areas, the requirements for the accuracy and adaptability of debris flow protection structures have significantly increased. There is an urgent need for a technical method that can scientifically predict the closure degree of permeable dams and take into account both structural optimization and functional reliability. Summary of the Invention

[0007] This invention provides a novel method for predicting debris flow blockage degree in wedge-beam grid dams. Taking wedge-beam grid dams as the research object, a debris flow blockage degree prediction system is constructed based on flume model tests and multi-factor coupled analysis.

[0008] A novel method for predicting debris flow blockage degree in a wedge-shaped beam grid dam includes: Step 1: Obtaining basic parameters, which are divided into debris flow characteristic parameters and channel condition parameters. The debris flow characteristic parameters include debris flow unit weight, debris flow solids volume concentration, primary debris flow scale, and minimum particle size of the largest boulder in the debris flow basin. The channel condition parameters include the average channel width of the proposed wedge-shaped beam grid dam cross-section in the debris flow basin and the average longitudinal slope of the channel within the proposed wedge-shaped beam grid dam cross-section. Step 2: Designing the structural parameters of the wedge-shaped beam grid dam, including: determining the wedge angle ratio: this angle ratio is the wedge shape... The ratio of the actual angle of the wedge structure of the beam-type grid dam to 60°; calculation of the dam's reservoir capacity and reservoir capacity ratio: based on the average channel width, the initially set dam height, the wedge structure angle ratio, the minimum particle size of the largest rock, and the average channel longitudinal slope, the reservoir capacity of the wedge-shaped beam-type grid dam is calculated; then, based on the scale of a debris flow, the volume concentration of debris flow solids, the average channel longitudinal slope, the wedge structure angle ratio, the average channel width, the dam height, and the reservoir capacity, the ratio of the total amount of debris flow solids to the reservoir capacity, i.e., the reservoir capacity ratio, is calculated; determination of the dam opening width and relative opening width: the characteristic particle size of the debris flow basin is obtained through field exploration tests. Based on characteristic particle size Define the opening width of the dam body, and then calculate the relationship between the opening width and the characteristic particle size. The ratio of the relative opening width to the closure degree; Step 3: Based on the basic parameters obtained in Step 1 and the structural parameters designed in Step 2, conduct flume model tests to obtain measured data on the closure degree; Step 4: Construct a closure degree prediction model, collect multiple sets of measured closure degree data obtained in Step 3, and use debris flow unit weight, debris flow solid volume concentration, debris flow scale, average channel longitudinal slope gradient, wedge structure angle ratio, relative opening width, and reservoir capacity ratio as control factors. Use multiple regression analysis to analyze the correlation between each control factor and the measured closure degree data, determine the functional form between each control factor and the closure degree, and fit it to obtain the closure degree prediction model; Step 5: Apply the model to predict the closure degree. Input the basic parameters of the debris flow channel to be predicted and the structural parameters of the corresponding wedge beam grid dam into the closure degree prediction model constructed in Step 4, and calculate the predicted value of debris flow closure degree of the wedge beam grid dam under this working condition.

[0009] In this specification, when determining the wedge structure angle ratio in step two, the value of the angle ratio ranges from 0.75 to 1.25.

[0010] In this instruction manual, when calculating the reservoir capacity ratio in step two, the reservoir capacity ratio ranges from 0 to 4.0. The total amount of solid material required for the debris flow is obtained by multiplying the debris flow scale by the volume concentration of solid material in the debris flow.

[0011] In this specification, the characteristic particle size in step two This refers to the particle size value corresponding to a cumulative content of 95% in the overall particle size distribution curve of the debris flow basin. The particle size distribution curve is drawn based on the sieving results of soil samples from the debris flow deposition area and the flow area.

[0012] In this instruction manual, when conducting the flume model test in step three, a test device is set up that includes a hopper, a flume, a wedge-shaped beam grid dam model, and a tailings pool. After the debris flow in the flume has completely stopped moving, the blocked area of ​​the overflow outlet in front of the dam and the total area of ​​the overflow section are recorded immediately, and the ratio of the two is calculated as the measured data of the closure degree.

[0013] In this instruction manual, before using multiple regression analysis in step four, the control factors are first processed to be dimensionless, so that each control factor is on the same order of magnitude, thereby improving the accuracy of the correlation analysis.

[0014] In this specification, when determining the dam opening width in step two, the opening width is based on the characteristic particle size. The setting is 1.62 times.

[0015] In this instruction manual, the number of sets of measured occlusion data collected in step four shall not be less than 70 sets. After fitting the occlusion prediction model, the goodness of fit of the model is verified by calculating the error between the model's predicted values ​​and the measured values ​​to ensure the goodness of fit. ≥0.7.

[0016] In this specification, when designing the structural parameters of the wedge-shaped beam grid dam in step two, a triangular prism connector is also provided at the sharp corner of the wedge structure. The height of the triangular prism connector is consistent with the height of the dam body, which is used to connect the upper and lower rows of wedge structures into a whole and enhance the dam body's resistance to debris flow impact.

[0017] In this manual, when applying the model to predict the closure degree in step five, this method is applicable to debris flow channels in mountainous areas carrying large rocks, and the range of values ​​for the control factor parameters of the working condition to be predicted is consistent with the range of values ​​for the working condition parameters of the flume model test in step three, ensuring the effectiveness of the prediction results.

[0018] The embodiments described in this specification can achieve at least the following beneficial effects:

[0019] Addressing the shortcomings of traditional prediction methods: Overcoming the problem of insufficient consideration of control parameters in traditional empirical methods, this method significantly improves the accuracy and scientific nature of blockage prediction for wedge beam grid dams by systematically integrating multiple key influencing factors, providing quantitative basis for dam design and avoiding the risk of functional failure caused by empirical design.

[0020] Enhance the comprehensive protection capabilities of the dam: Relying on the optimized design of the wedge structure, the dam can divert debris flow and dissipate energy, ensuring the core function of intercepting coarse debris and discharging fine debris, while also having a certain self-cleaning capability, reducing dam failure caused by complete blockage, extending the effective service life, and reducing the cost of later maintenance.

[0021] Enhanced adaptability to different working conditions: It can adapt to different debris flow characteristics (such as different densities and scales) and gully conditions, and can still play a stable role in interception and regulation in complex mountainous environments, providing a reliable disaster protection barrier for downstream towns and major projects, and supporting the long-term operation of the mountainous disaster prevention and mitigation system.

[0022] Promote technical standardization: Establish a complete technical path of experimentation-modeling-prediction to make up for the current situation where the theory lags behind the practice in the field of predicting the blockage degree of through-type silt traps, provide technical reference for the performance research and engineering application of similar through-type dams, and promote the standardized development of debris flow prevention and control technology. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 This is a schematic diagram of a novel wedge-shaped beam grid dam debris flow blockage prediction method involved in some embodiments of the present invention.

[0025] Figure 2 This is a schematic diagram of the structure of a wedge-shaped beam grid dam involved in some embodiments of the present invention.

[0026] Figure 3 This is a schematic diagram comparing the measured and predicted values ​​of the blockage degree when a wedge-shaped beam grid dam intercepts debris flow in some embodiments of the present invention.

[0027] Figure 4 This is a schematic diagram of the particle size distribution curve of the test soil sample involved in some embodiments of the present invention. Detailed Implementation

[0028] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0029] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0030] like Figure 1As shown, this embodiment provides a novel method for predicting the blockage degree of debris flow in a wedge-beam grid dam, including the following steps: Step 1: Obtain basic parameters. The following parameters are collected through field investigation: Debris flow characteristic parameters: The bulk density of debris flow is determined by combining field slurry mixing tests with table lookup. The solid volume concentration of debris flow is calculated based on the bulk density of debris flow, the specific gravity of clear water, and the bulk density of solid matter; According to the "Specification for Investigation of Debris Flow Disaster Prevention Engineering", the scale of a debris flow is calculated by measuring the duration and peak flow of the debris flow; The minimum particle size of the largest rock in the debris flow basin is determined by a watershed particle survey; Channel condition parameters: The average width of the channel of the proposed wedge-beam grid dam is determined by cross-sectional survey, and the average longitudinal slope of the channel within the cross-sectional area is determined by slope measurement; Step 2: Design the structural parameters of the wedge-beam grid dam. This includes: determining the wedge structure angle ratio: this angle ratio is the ratio of the actual angle of the wedge structure of the wedge-shaped grid dam to 60°, determined in conjunction with the expected diversion efficiency requirements of the dam body; calculating the dam body reservoir capacity and reservoir capacity ratio: based on the average channel width, the initially set dam height, the wedge structure angle ratio, the maximum and minimum particle size values ​​of the largest rock, and the longitudinal slope gradient of the channel, the reservoir capacity of the wedge-shaped grid dam is calculated; then, based on the scale of a debris flow, the volume concentration of debris flow solids, the longitudinal slope gradient of the channel, the wedge structure angle ratio, the average channel width, the dam height, and the reservoir capacity, the ratio of the total amount of debris flow solid matter to the reservoir capacity, i.e., the reservoir capacity ratio, is calculated; determining the dam opening width: through field reconnaissance and pit exploration tests, soil samples are collected and sieved in the debris flow deposition and flow areas, and the overall particle size distribution curve of the watershed is plotted to determine the characteristic particle size corresponding to a cumulative content of 95% in the curve. Based on characteristic particle size The dam opening width is set, and then the relative opening width (opening width and characteristic particle size) is obtained. Step 3: Conduct flume model tests to obtain measured data on blockage degree. Construct a test setup including a hopper, flume, wedge-beam grid dam model, and tailings pool. Based on the basic parameters obtained in Step 1 and the structural parameters designed in Step 2, set multiple working conditions (each working condition is a different combination of debris flow characteristic parameters, channel longitudinal slope, and dam structural parameters) to simulate debris flow movement and its interaction with the dam under different working conditions. After the test ends and the debris flow in the flume completely stops moving, record the blocked area of ​​the overflow outlet in front of the wedge-beam grid dam and the total area of ​​the overflow section, and calculate the ratio of the two as the measured data on blockage degree. Step 4: Construct a blockage degree prediction model and collect data from Step 1. The three sets of measured blockage data were obtained. Using debris flow unit weight, debris flow solid volume concentration, debris flow scale, channel longitudinal slope gradient, wedge structure angle ratio, relative opening width, and reservoir capacity ratio as control factors, a multiple regression analysis method was used to analyze the correlation between each control factor and the measured blockage data. The function form was determined and fitted to obtain a blockage prediction model. Step five: Apply the model to predict the blockage. Input the basic parameters of the debris flow channel to be predicted (obtained according to the method in step one) and the structural parameters of the corresponding wedge beam grid dam (designed according to the method in step two) into the blockage prediction model constructed in step four, and calculate the predicted value of debris flow blockage of the wedge beam grid dam under this working condition.

[0031] In some embodiments, when determining the wedge structure angle ratio in step two, the value of the angle ratio ranges from 0.75 to 1.25. The specific value is adjusted by conducting field surveys of the impact intensity of debris flows in the area to be protected, and in conjunction with the expected diversion and energy dissipation effect of the wedge beam grid dam, to ensure that the wedge structure has the preset diversion efficiency.

[0032] In some embodiments, when calculating the reservoir capacity ratio in step two, the reservoir capacity ratio ranges from 0 to 4.0. The total amount of debris flow solid matter required in the calculation process is obtained by multiplying the debris flow scale and the debris flow solid volume concentration to ensure that the reservoir capacity ratio can accurately reflect the matching relationship between the total amount of debris flow solid matter and the reservoir capacity of the dam.

[0033] In some embodiments, when determining the dam opening width in step two, the opening width is based on the characteristic particle size. The 1.62x setting, characteristic particle size By reading the overall particle size distribution curve of the watershed, the particle size value corresponding to the cumulative particle content reaching 95% is determined to ensure that the dam body can achieve its core function of intercepting coarse particles and discharging fine particles.

[0034] In some embodiments, when recording the measured data of occlusion in step three, the measurement and recording of the blocked area of ​​the spillway in front of the dam and the total area of ​​the spillway section are completed within 10 minutes after the debris flow stops moving, so as to eliminate the influence of pore water evaporation and secondary settlement of the sediment on the measurement results and ensure the reliability of the measured data of occlusion.

[0035] In some embodiments, when constructing the occlusion degree prediction model in step four, the number of actual occlusion degree data sets collected shall not be less than 70 sets. Before using multiple regression analysis, the dimensions of each control factor shall be made dimensionless. After the analysis is completed, the model goodness of fit shall be verified by calculating the error between the model prediction value and the actual value, ensuring that the goodness of fit R²≥0.7, and guaranteeing the model prediction accuracy.

[0036] In some embodiments, when designing the structural parameters of the wedge-shaped beam grid dam in step two, a triangular prism connector is also provided at the sharp corner of the wedge structure. The height of the triangular prism is consistent with the height of the dam body, which is used to connect the upper and lower rows of wedge structures into a whole, and at the same time replace some conventional columns to enhance the overall structural stability of the dam body and its resistance to debris flow impact.

[0037] In some embodiments, when obtaining the scale of a debris flow in step one, the calculation method specified in the "Specification for Investigation of Debris Flow Disaster Prevention and Control Engineering" shall be strictly followed, that is, the duration and peak flow of the debris flow shall be obtained through on-site monitoring, and the formula "scale of a debris flow = 0.264 × duration × peak flow" shall be used to calculate the scale, ensuring that the debris flow scale data meets the requirements of industry standards.

[0038] In some embodiments, when constructing the occlusion degree prediction model in step four, a fusion modeling approach is adopted using three algorithms: multiple linear regression, support vector machine regression, and random forest regression. The specific process is as follows: First, the random forest regression model is used to perform feature importance analysis on all control factors, and core control factors that significantly affect occlusion degree are selected. The importance weight of the core control factor is fed back to the multiple linear regression model as the feature weight coefficient when constructing the linear association. Then, the initial occlusion degree prediction value output by the multiple linear regression model is used as an additional input feature and incorporated into the training samples of the support vector machine regression model. This allows the support vector machine regression model to take into account the basic laws of linear association when fitting the nonlinear association between occlusion degree and control factors. At the same time, the nonlinear fitting residual result of the support vector machine regression model is back-transmitted to the random forest regression model to adjust the ensemble weight of the decision tree in the random forest regression model, so that the random forest regression model focuses more on the fitting correction of the nonlinear residual. Through the bidirectional feature and result feedback of the three algorithms, accurate fitting of linear and nonlinear associations is achieved.

[0039] In some embodiments, when constructing the occlusion degree prediction model in step four, model optimization is achieved through hierarchical collaboration and cross-validation of three algorithms: multiple linear regression, support vector machine regression, and random forest regression. Specific operations include: First, dividing the measured occlusion degree data and control factors into training and validation sets; training a multiple linear regression model using the training set data to obtain a basic linear prediction model for occlusion degree; and simultaneously calculating the prediction residuals of this model on the validation set. Second, using the prediction residuals as a new prediction target, training a support vector machine regression model in conjunction with control factors to fit nonlinear correlations that the linear model cannot capture, thus obtaining a residual correction model. Third, using a random forest regression model to analyze the relationship between control factors and measured occlusion degree data. The occlusion degree is fitted across all dimensions to obtain an ensemble learning prediction model. The suitability of the linear base model and the residual correction model is evaluated by validating the random forest regression model using out-of-bag data. In the fourth step, based on the validation results of the random forest regression model, the feature coefficients of the multiple linear regression model and the kernel function parameters of the support vector machine regression model are adjusted. The adjusted linear prediction results are then superimposed with the nonlinear residual correction results to obtain a preliminary fused prediction value. In the fifth step, the preliminary fused prediction value is weighted and fused with the prediction value of the random forest regression model. The weights are dynamically allocated according to the prediction accuracy of the three models on the validation set, ultimately forming an occlusion degree prediction model that takes into account linearity, nonlinearity, and ensemble stability.

[0040] The technical concept of this invention is as follows: The debris flow blockage prediction model proposed in this invention is applicable to specific debris flow channels with a certain amount of large particles, and employs wedge-shaped beam grid dams to intercept and regulate debris flows. Based on the "Design Code for Debris Flow Prevention Engineering," and combined with the site's topographical and geological conditions, the dam site is precisely selected. The dam site should be chosen at the neck of a valley in the middle and lower reaches of the formation zone, where the mouth is narrow and the body is wide, while also controlling upstream tributaries and landslides, ensuring the stability and effectiveness of the dam.

[0041] Large particles refer to particles that can be interlocked to effectively block this type of permeable silt-trapping dam by adjusting the opening size of the wedge-shaped beam grid dam.

[0042] The wedge-shaped beam grid dam model refers to a structural optimization based on a conventional beam grid dam. While maintaining the column structure of the conventional beam grid dam, the crossbeams are designed as wedge-shaped beams with a specific wedge angle. Simultaneously, to leverage the flow diversion capacity of the wedge structure, the wedge beams are given a certain height, which is set to a specific value. To ensure the integrity of the upper and lower rows of wedge-shaped structures and to address stress concentration issues, triangular prisms of the same height as the dam body are installed at the sharp corners of the wedge-shaped structures. These prisms connect the upper and lower rows of wedge-shaped structures into a unified whole while also bearing the impact of debris flows on the sharp corners of the wedge-shaped structures. Because the wedge-shaped structures have a certain extension length, and the triangular prisms can replace the columns of conventional beam-type grid dams, the distance between the columns in the rear row is set to 2.0. The structure of the wedge-shaped beam grid dam is shown in the attached figure. Figure 2 As shown, the wedge-shaped beam grid dam is a permeable barrier structure arranged laterally along the debris flow channel. The overall shape is a combination of continuous periodic wave-shaped beams and vertical columns. The core is composed of three types of components: interception columns 3, connecting columns 2, and wedge-shaped beams 1. Their structure, connection and positional relationship are as follows.

[0043] The dam body forms an overall stable structure through the integrated fixed connection of components: the intercepting column 3 is the vertical load-bearing and main retaining component of the dam body, and its top horizontal end is fixedly connected to one end (beam end) of two symmetrically inclined wedge beams 1 respectively (usually integrally formed by concrete pouring); the wedge beams 1 extending from the adjacent intercepting column 3 converge towards the middle of the dam body, and their ends (i.e., the apex of the V-shaped beam) are fixedly connected to the top end face of the connecting column 2; the bottoms of the intercepting column 3 and the connecting column 2 are embedded in the foundation structure of the ditch foundation to form a vertical support system, so that the dam body and the foundation are connected as a whole.

[0044] The components are arranged in a regular spatial pattern: the intercepting columns 3 are arranged vertically at equal intervals along the length of the dam body and are the main vertical nodes of the structure; the connecting columns 2 are located on the horizontal midline of two adjacent intercepting columns 3 and are distributed in an alternating pattern of "intercepting column 3 → connecting column 2 → intercepting column 3" with the intercepting columns 3; the wedge beams 1 extend inclinedly towards the intercepting columns 3 on both sides with the connecting column 2 as the apex, forming a continuous V-shaped barrier surface, and the gap between adjacent wedge beams 1 is the opening of the dam body (used to pass fine silt), forming a cyclic barrier-diversion spatial layout.

[0045] In this structure, wedge beam 1 is responsible for guiding the debris flow to divert and dissipate energy, and intercepting medium-sized particles. Interception column 3 directly blocks large rocks, while connecting column 2 strengthens the connection stability of adjacent wedge beams 1, ultimately achieving the function of intercepting coarse rocks and discharging fine ones.

[0046] This method includes the following steps.

[0047] Step 1: Select a cross-section suitable for dam construction in the debris flow basin, and determine the debris flow unit weight and solid matter concentration through on-site investigation and testing. The mudslide depth h at the cross-section in front of the dam, and the slope of the gully bed. .

[0048] ;

[0049] —Debris flow unit weight (kN / The specific density (density) can be determined through field surveys, on-site mixing tests, and table lookup (test density 1.6–2.0 t / (The on-site slurry preparation method requires taking no less than 3 sets of debris flow deposit samples at the test point, with each set weighing no less than 5 kg.) —Severity of pure water (kN / ), with a value of 10; —The bulk density of solid materials carried by the mudslide (kN / The experimental value was 26.5.

[0050] Step two: The dam height H of the wedge beam grid dam can be preliminarily determined based on the mud depth h of the debris flow, and the ratio of the debris flow scale to the reservoir capacity of the wedge beam grid dam can be calculated.

[0051] ;

[0052] ;

[0053] Or first determine the initial storage capacity ratio. Determine the dam height of the wedge-shaped beam grid dam.

[0054] ;

[0055] —The average width (m) of the channel of the proposed wedge-shaped grid dam section in the debris flow basin was determined through field investigation. —Proposed height of wedge-shaped grid dam (m); —Angle of the wedge-shaped structure of the wedge-shaped beam grid dam The ratio to 60° (to ensure the wedge structure has a certain shunting efficiency), The range should be 0.75 to 1.25, and it is not affected by the slope of the ditch or other parameters. In actual engineering, the value can be selected according to the comprehensive requirements of the sand interception rate and the degree of blockage. —Scale of a single debris flow ( (This can be determined by referring to the "Specification for Investigation of Debris Flow Disaster Prevention and Control Engineering", based on the duration of the debris flow.) and peak traffic The total volume of slurry in a single debris flow is considered the scale of that debris flow. ; —The minimum particle size (m) of the largest boulder in the debris flow basin was determined through field investigation; —Reservoir capacity of wedge-shaped beam grid dam ( ); —The average gully slope within the cross-section of the proposed wedge-shaped grid dam was determined through field investigation; —The ratio of the total amount of solid debris flow to the reservoir capacity of the wedge-shaped beam grid dam, ranging from 0 to 4.0, can be calculated using Equation 3; —The solid volume concentration of debris flows was determined through field investigation and experiments.

[0056] Step 3: Conduct pit exploration tests through field reconnaissance to obtain particle size distribution curves at various locations (deposition areas, flow areas, etc.) within the debris flow basin. Simultaneously, select a representative overall particle size distribution curve for the basin based on the proportion of large particles. Determine the characteristic particle size of debris flow samples from this basin. (i.e., the particle size value corresponding to a cumulative content of 95% in the full particle size distribution curve) can be used to determine the opening width b of the wedge beam grid dam.

[0057] Step 4: After the test, the debris flow closure degree of the wedge beam grid dam can be calculated using the following formula.

[0058] ;

[0059] In the formula: The allowable flow area of ​​the overflow outlet of the wedge-shaped beam grid dam; The area blocked by the overflow outlet in front of the dam after the debris flow passes over the dam; These are measured values ​​of occlusion. The influence of the high sediment accumulation in front of the dam after the debris flow passes through it is not considered. The calculation is based solely on the actual blockage of the overflow area in front of the dam. The occlusion data is recorded immediately after the debris flow stops moving, and the influence of pore water is not considered.

[0060] Step 5: By combining experimental data with the characteristics of occlusion degree variation with five control factors, and analyzing 70 sets of experimental data, an occlusion degree prediction model is established using Origin software with different function forms, such as multivariate linear function, multivariate power function, multivariate exponential function, and multivariate logarithmic function.

[0061] ;

[0062] In the formula: Here is the predicted occlusion degree, and e is the natural constant. The relative opening width, i.e., the opening width With characteristic particle size The ratio; other symbols are as described above.

[0063] The angle ratio of the wedge-shaped beam grid dam structure A value of 1.0 is recommended for the ratio of the total solid matter from debris flows to the reservoir capacity of the wedge-shaped beam grid dam. The recommended value is 3.0.

[0064] This invention addresses the issue of neglecting debris flow scale in the key control selection factors of existing permeable silt-trapping dams, and proposes a wedge-shaped beam grid dam from a structural optimization perspective. Based on existing research on permeable silt-trapping dams and combined with flume model test data, a novel quantitative calculation method for debris flow closure degree of wedge-shaped beam grid dams is created. It creatively introduces the wedge structure angle ratio and reservoir capacity ratio, providing new evidence for the performance study of permeable silt-trapping dams in intercepting debris flows, and has strong practicality.

[0065] The model is derived from fitting test data of a wedge-beam grid dam flume model. It comprehensively considers the influence of five control factors (debris flow unit weight, debris flow scale, channel longitudinal slope gradient, relative opening width, and wedge structure angle) on the debris flow interception rate. To ensure the applicability of the model, the control factors are dimensionless (debris flow solid volume concentration). Debris flow scale to silt retention dam capacity ratio , channel gradient tangent Relative opening width wedge structure angle ratio ).

[0066] ;

[0067] The overall fitting coefficients of the model are determined by fitting the data from the water tank test.

[0068] Based on the test data of the water tank model, the degree of occlusion There is a functional relationship between it and each control factor.

[0069] occlusion Solid volume concentration with debris flow As increases, it increases, satisfying the functional relationship: ; The fitting index corresponding to the solid volume concentration in debris flows is determined by fitting experimental data.

[0070] occlusion With storage capacity ratio The increase of follows a pattern of first increasing and then decreasing, satisfying the functional relationship: ; The fitting index corresponding to the storage capacity ratio is determined through fitting experimental data.

[0071] occlusion With the gradient of the ditch As increases, decreases, satisfying the functional relationship: ; The fitting index corresponding to the longitudinal slope of the ditch was determined by fitting experimental data.

[0072] occlusion relative opening width of wedge-shaped beam grid dam As increases, decreases, satisfying the functional relationship: ; The fitting index is the relative opening width, determined by fitting experimental data.

[0073] occlusion With the wedge structure angle ratio As increases, it increases, satisfying the functional relationship: ; The fitting index corresponding to the wedge structure angle ratio is determined by fitting experimental data.

[0074] ;

[0075] In the formula: The angle (°) of the wedge structure of the silt trap dam model used in the experiment; , , , , , The symbol is a constant, determined experimentally; the meanings of other symbols are the same as described above.

[0076] Based on the analysis of 70 sets of experimental data, and using Origin software, different function forms were employed, such as multivariate linear functions, multivariate power functions, multivariate exponential functions, and multivariate logarithmic functions, to arrive at the final prediction model for the blockage degree of the wedge-shaped beam grid dam: , The coefficient of determination (goodness of fit) is used to characterize the degree of fit of the model to the experimental data. The value ranges from 0 to 1, and the closer the value is to 1, the better the model fit.

[0077] A Comparison of Measured and Predicted Values ​​of Blockage Degree When Using Wedge-Beam Grid Dams to Intercept Debris Flows: Figure 3 As shown.

[0078] The design parameters of wedge-shaped beam grid dams include the wedge structure angle, the relative opening width of the wedge structure, the height of the wedge structure itself, and the front and rear connecting columns and intercepting columns of the wedge structure.

[0079] To maximize the flow-diverting effect of the wedge structure, the wedge angle is typically set at 60°. The wedge height can be determined based on the minimum axis dimension (min) of the largest particle. , , To maximize the interception and regulation capacity of the wedge-shaped beam grid dam, its relative opening width can be set to 1.62. The wedge-shaped structure will produce a certain stress concentration effect. The entire sand-trapping dam can be connected as a whole by the connecting columns between the wedge-shaped structures and the intercepting columns in the rear row to ensure its stability.

[0080] Design concept of wedge beam grid dam: (1) In accordance with the concept of comprehensive debris flow prevention and control, the through-type sand-blocking dam fully utilizes the immediate effectiveness of geotechnical engineering measures and the continuous strengthening of ecological and environmental protection measures over time; (2) Energy-oriented dissipation: The wedge structure angle guides the debris flow to divert and collide, converting the linear impact kinetic energy into oblique shear energy and eddy current dissipation; (3) Adaptive regulation: The diversion of the wedge structure changes the movement form of particles in the fluid (the direction of movement of the long and short axes) to a certain extent, and the overflow section area is reduced by controlling the structural angle; (4) The peak flow of debris flow is weakened, thereby reducing the impact and damage of debris flow to the downstream, maintaining long-term regulation capability, and mainly using the function of intercepting coarse and discharging fine, and generally playing a synergistic role in combination with solid dam.

[0081] The experimental flume setup mainly consists of a hopper, a flume, a wedge-beam grid dam model, and a tailings pool. The hopper is 0.8m long, 0.8m wide, and 0.8m high, made of stainless steel plate. The flume is 5.0m long, 0.4m wide, and 0.4m high on both sides, with tempered glass sides and a steel plate bottom. The wedge-beam grid dam model is positioned 3.0m from the hopper opening, with four different opening spacings. The tailings pool is 1.0m long, 1.0m wide, and 0.5m high, made of stainless steel plate. The flume's adjustable slope ranges from 0° to 20°. The similarity scale for the flume test is shown in Table 1.

[0082] Table 1. Similarity Scale Table for Water Tank Tests

[0083]

[0084] The particle size distribution curve of the test soil sample is as follows Figure 4 As shown (Note: Based on experimental conditions, this experiment fixed coarse particles (20-50mm accounting for 15%). The maximum particle size for debris flow was determined to be 50mm in this experiment. Large gravel particles were sieved from the original soil sample, and 85% of the total experimental material was retained, with particles smaller than 20mm. The maximum particle size of the material in this experiment was...) =50mm, characteristic particle size =37mm, coefficient of non-uniformity =22.5, curvature coefficient =2.8. Test conditions were designed, and the test parameters are shown in Table 2.

[0085] Table 2. Test Parameter Table

[0086]

[0087] The debris flow was 0.05 in size. Through observation of experimental phenomena and data analysis, it was found that the interception capability was insufficient when the relative opening was too large, thus requiring some compromise. The influence of the wedge structure angle on the degree of closure was explored under the condition of a channel longitudinal slope of 11°. The influence of the channel longitudinal slope on the degree of closure was explored under the condition of a relative opening width of 1.62. Therefore, the final number of experimental groups was 90.

[0088] Experimental Procedure: 1) Model Preparation: Assemble the water tank model, hopper, tailings pool, etc., according to requirements, sieve the materials, and debug the experimental equipment. 2) Material Configuration: Based on the soil sample material required for the experiment and the designed fluid density, prepare the required fluid sample according to a certain ratio of soil sample to water. To ensure that the materials and water are fully mixed, soak the prepared debris flow sample overnight. Before starting the experiment, use a mixer to fully stir the prepared debris flow sample and then pour it into the experimental material tank in a small bucket. 3) Experimental Preparation: According to the designed cross-section location, set up the camera, GoPro camera, and mud level gauge in advance, adjust the shooting direction of the camera, and zero the error of the mud level gauge. Check that the mud level gauge and mud level acquisition instrument are properly connected. One day in advance, set up the wedge-shaped beam grid dam model at the designed location (use glass glue to glue the bottom and sides to prevent the model from being washed away by the debris flow). 4) Start the test: Turn on the cameras mounted on the side and front of the flume, stir the test fluid in the hopper evenly, and then quickly open the hopper valve (release all at once). During the test, drop small balls and monitor the fluid level with a mud level gauge. Record the test process with the camera to facilitate the later acquisition of debris flow movement parameters by combining with other methods. During the test, use two 5000ml measuring buckets to collect debris flow after passing the dam at the flume opening. 5) Data recording: After the fluid in the flume has basically stopped moving, turn off the camera, mud level acquisition instrument, and other equipment. Use a sampling shovel and two 1000ml plastic beakers to collect the silted soil in front of the dam. Record the bulk density of the soil samples in the measuring buckets and beakers. 6) Clean the flume: After the data recording and test sampling are completed, connect the water pipe to start rinsing from the hopper. According to the test group design, change the debris flow fluid properties, debris flow scale, relative opening height, wedge structure angle, and channel gradient, and repeat the above process.

[0089] The experimental data analysis shows that some model test data are shown in Table 3, and some permeable silt trap dam closure degree prediction models are shown in Table 4.

[0090] Table 3. Partial Model Test Data

[0091]

[0092] Table 4. Prediction Model for Blockage Degree of Partially Permeable Silt Retaining Dams

[0093]

[0094] in, The prediction model was proposed by Yuan Dong et al. The prediction model was proposed by Zhou Wenbing et al. The prediction model was proposed by Yang Kaicheng et al. This is the prediction model of the present invention.

[0095] Compared with existing test data of permeable silt-trapping dam models, the evaluation index data of commonly used permeable silt-trapping dam flume model test research are shown in Table 5, and the evaluation index data of wedge beam grid dam flume model are shown in Table 6.

[0096] Table 5. Evaluation Index Data for Commonly Used Permeable Silt Retaining Dam Flume Model Tests

[0097]

[0098] Table 6. Evaluation Index Data for Wedge-Shaped Beam Grid Dam Flume Model

[0099]

[0100] The performance of permeable silt-trapping dams in regulating debris flows relies on a comprehensive evaluation of indicators such as silt trapping rate, blockage degree, and peak flow reduction rate. Through comparative evaluation of these indicators, wedge-beam grid dams demonstrate better control over viscous debris flows and large-scale dilute debris flows. Specific advantages are as follows.

[0101] (1) Wedge-shaped beam grid dams have the functions of high sediment retention rate and controllable blockage degree to a certain extent.

[0102] Adjusting 20kN / In high-density debris flows, the sediment interception rate of traditional beam-type grid dams and lattice dams fluctuates drastically with varying operating conditions (e.g., the lattice dam's interception rate drops sharply from 52% to 17%), reflecting the instability of their structural interception effect; while wedge-shaped beam-type grid dams, under the same density and 0.15 Under large-scale operating conditions, it still maintains a sand interception rate of 91.1% and 78.2%; at 0.30... Under large-scale operating conditions, the sediment interception rate decreases slightly, while the blockage degree shows a gradual decreasing trend, and the peak flow reduction rate also has a considerable value. This also confirms that the blockage degree of the wedge beam grid dam in intercepting small-scale debris flows is not completely blocked, the dam body is not completely ineffective, and still has a certain self-cleaning capacity for subsequent debris flows.

[0103] (2) Better peak flow reduction capability

[0104] In debris flow conditions similar to those of other types of dams (20kN / ), wedge-shaped beam grid dams (20kN / ) -0.10 The peak reduction rate reached 93.8%, far exceeding other permeable silt-trapping dams. Secondly, a comparison showed that under low density conditions, the reduction rate of pile-forest dams and comb-type dams dropped significantly to 40%-50%, while the wedge-beam grid dam maintained a reduction rate of 17kN / m³. Across various scales, the reduction rate remains stable between 40.1% and 70.5%. This stable peak-shaving performance means that wedge-beam grid dams can provide more reliable safety redundancy for downstream protection, and their structure is more adaptable to changes in debris flow properties.

[0105] (3) It has stronger adaptability to working conditions and structural robustness.

[0106] When comparing the performance degradation of different dam types under high and low density conditions, the wedge-beam grid dam shows a more significant advantage. Taking the sediment retention rate as an example, traditional dam types generally experience a precipitous drop in performance when switching from high to low density conditions (e.g., grid dams drop from 52% to 0%). While wedge-beam dams also show a decrease, their performance remains relatively stable under low density and large-scale (17kN / m³) conditions. -0.30 Even under such extreme conditions, it can still maintain a sand interception rate of 74.9% and a peak flow reduction capacity of 50%.

[0107] In summary, wedge-beam grid dams exhibit better sediment retention rates compared to traditional permeable silt-trapping dams. Furthermore, the blockage observed in small-scale interception is not ineffective, as it recovers under the subsequent self-cleaning effect of debris flows. Simultaneously, wedge-beam grid dams demonstrate the best peak flow reduction rate, achieving up to 93.8% reduction for viscous debris flows. While their sediment retention rate for dilute debris flows is lower than that of conventional beam grid dams, the low blockage and high-scale performance data indicate that wedge-beam grid dams still possess good regulation capabilities under large-scale conditions.

[0108] By summarizing the existing empirical models of the interception rate of permeable silt-trapping dams, most of the key parameters selected ignore the influence of debris flow scale. However, the wedge beam grid dam, considering its own structural angle, selects the influence of the change in debris flow scale on its interception rate, further improving the control factors and variation characteristics of the interception and regulation effect of permeable silt-trapping dams, and providing a reference for debris flow disaster prevention and mitigation.

[0109] By summarizing the existing critical closure discrimination models of permeable silt-trapping dams, most of the key parameters selected ignore the influence of debris flow scale. However, wedge beam grid dams, considering their own structural angles, select the influence of debris flow scale changes on critical closure discrimination, further improving the control factors and variation characteristics of the interception and regulation effect of permeable silt-trapping dams, and providing a reference for debris flow disaster prevention and mitigation.

[0110] In one specific embodiment: This embodiment selects a debris flow gully in a mountainous area of ​​western Sichuan as the research object (this gully is a typical debris flow basin carrying large rocks, with an average of 1-2 debris flow disasters per year, and a wedge-shaped beam grid dam is planned to be constructed for disaster prevention). Following the process of acquiring basic parameters, designing dam structural parameters, conducting flume tests, building a prediction model through multi-algorithm fusion, and applying the model for prediction, accurate prediction of debris flow blockage degree is achieved. Step four (building the blockage degree prediction model) integrates three algorithms: Multiple Linear Regression (MLR), Support Vector Machine Regression (SVR), and Random Forest Regression (RFR). Model optimization is achieved through pairwise interactions between algorithms and bidirectional parameter transfer, ultimately outputting the predicted blockage degree value.

[0111] Step 1: Obtain basic parameters

[0112] This step involves conducting field investigations to obtain specific values ​​for debris flow characteristic parameters and channel condition parameters, providing basic data support for subsequent model construction.

[0113] 1. Debris Flow Characteristic Parameters

[0114] Debris flow density (kN / This refers to the weight of a unit volume of debris flow, measured on-site. (Test bulk density range: 1.6–2.0 t / ) Corresponding to 16~20kN / ).

[0115] Clear water severity (kN / ): The value is (constant).

[0116] Solid substance density (kN / The weight per unit volume of solid particles carried by a debris flow, as determined by the experiment, is [value missing]. (constant).

[0117] Debris flow solid volume concentration The volume percentage of solid particles in a debris flow is reflected by the following formula: Substitute numerical values ​​into the calculation: .

[0118] A debris flow of size S ( The total volume of slurry in a single debris flow is calculated based on actual measurements in the "Specifications for Investigation of Debris Flow Disaster Prevention and Control Engineering". .

[0119] Maximum stone minimum particle size value (m): The shortest axis dimension of the largest boulder in the debris flow basin, which was measured to be 0.5m on site.

[0120] 2. Channel condition parameters

[0121] Average channel width B (m): The average horizontal width of the channel across the proposed dam cross-section, measured on-site as 8m.

[0122] Average longitudinal slope of the ditch (°): The average slope of the gully within the proposed dam cross-section, as measured. .

[0123] Step 2: Design the structural parameters of the wedge-shaped beam grid dam.

[0124] Based on the basic parameters in step one, the core structural parameters of the wedge-shaped beam grid dam are designed, and the compatibility index between the dam body and debris flow is determined.

[0125] 1. Wedge structure angle ratio

[0126] Defined as the actual angle of the wedge structure (°) and The ratio is given by the formula: .

[0127] To ensure the efficiency of the diversion, the design Substituting into ( The value range is 0.75 to 1.25.

[0128] 2. Dam reservoir capacity ( )

[0129] The formula for calculating the sediment retention capacity of the dam body is as follows: Where H is the dam height (m), initially set at 5m; , .

[0130] Substitute numerical calculations: numerator: ; Denominator: Storage capacity: .

[0131] 3. Storage capacity ratio

[0132] The ratio of the total solid matter from debris flow to the reservoir capacity of the dam is calculated using the following formula: .

[0133] Substitute numerical calculations: numerator: ; Denominator: Storage capacity ratio: (Value range: 0 to 4.0).

[0134] 4. Characteristic particle size (m)

[0135] This refers to the particle size value corresponding to a cumulative content of 95% in the particle size distribution curve of a debris flow basin, obtained by sieving soil samples through field pit exploration tests. .

[0136] 5. Width of the dam opening b (m)

[0137] Based on the characteristic particle size design, the formula is: Substituting into .

[0138] 6. Relative opening width

[0139] The ratio of the opening width to the characteristic particle size is given by the formula: Substituting, we get .

[0140] Step 3: Conduct a flume model test to obtain measured data on occlusion degree.

[0141] Based on the basic parameters from step one and the structural parameters from step two, a water tank model test device (hopper) was constructed. ,sink Tailings pool The test set up 70 test conditions (covering different debris flow unit weights, scales, gully slopes, etc.) to simulate the interaction between debris flow and dam body.

[0142] After the test, record the area of ​​blockage at the spillway in front of the dam. ( ) and the total area of ​​the overflow section ( ), calculate the measured value of occlusion. The formula is: .

[0143] In this embodiment, the 70 sets of data are divided into 60 training sets (for model training) and 10 test sets (for model validation), one of which is a measured occlusion degree under a typical working condition. (correspond =0.4848、 =1.0、 =0.000171、 =1.6216、 ).

[0144] Step 4: Construct a blockage prediction model

[0145] This step is the core optimization process, which integrates three algorithms: Multiple Linear Regression (MLR), Support Vector Machine Regression (SVR), and Random Forest Regression (RFR). Through feature transfer between MLR and SVR, weight feedback between MLR and RFR, and weighted fusion between SVR and RFR, the algorithms interact with each other to ultimately build a high-precision occlusion prediction model.

[0146] (a) Data preprocessing: dimensionless feature rendering

[0147] Input features Standardization is performed to eliminate dimensional differences; the formula is as follows:

[0148] ;

[0149] For the i-th original feature value (i=1,2,3,4,5, corresponding to...) ); The mean of the i-th feature in the training set is calculated as follows: =0.45、 =1.0、 =0.0002、 =1.60、 =0.18; The standard deviation of the i-th feature in the training set is calculated as follows: =0.05、 =0.1、 =0.00005、 =0.05、 =0.02.

[0150] Taking a typical operating condition as an example, the standardized characteristics are: ; ; ; ; .

[0151] The standardized feature matrix is ​​denoted as .

[0152] (II) Algorithm 1: Multiple Linear Regression Model

[0153] 1. Model building process

[0154] The MLR model is used to establish a linear relationship between features and occlusion, while introducing the feature importance index from RFR feedback to achieve interaction with RFR. The model expression is as follows: ; : Occlusion prediction value of MLR model; The intercept term of the MLR model; : The linear coefficient of the i-th feature in the MLR model; : The standardized value of the i-th input feature; The weight coefficients for the importance of RFR features are determined by the interactive training of MLR and RFR. The comprehensive indicator of feature importance output by RFR is calculated using the following formula: ,in This is the importance value of the i-th feature output by RFR.

[0155] 2. Model Training Process

[0156] 60 training sets and Input the MLR model and solve the coefficients using the least squares method. , The objective function is: ;in, Let be the measured occlusion degree of the j-th training set. is the predicted occlusion value of the MLR model for the j-th training set.

[0157] In the initial stage of training, let _send_ (Without introducing RFR feedback for now), the initial coefficients are obtained by solving: =0.5, =0.2, =0.15, =0.05, =-0.3, =-0.1.

[0158] 3. Model Application Process

[0159] Typical working conditions Substituting into the initial MLR model, we get: ;Calculation yields: .

[0160] (III) Algorithm 2: Support Vector Machine Regression Model

[0161] 1. Model building process

[0162] The SVR model is used to fit the nonlinear relationship between features and occlusion, and the initial predicted values ​​of MLR are... As additional input features, enabling interaction with MLR, the model expression is: ; : The standardized feature matrix; : The initial occlusion prediction value output by the MLR model; : Occlusion prediction value of SVR model; The weight vector of the SVR model (its dimension is consistent with the features after mapping the kernel function). Radial basis function (RBF) mapping maps input features to a high-dimensional space, as shown in the formula: ; : Bandwidth parameter of the RBF kernel function; , The features corresponding to the support vectors and the MLR predicted values; : Bias term of the SVR model.

[0163] 2. Model Training Process

[0164] Fusing features of the training set and Input an SVR model and solve it using the SMO algorithm. and The objective function is:

[0165] ;

[0166] st ;

[0167] : The standardized input feature matrix of the j-th training set; : The initial occlusion prediction value output by the MLR model corresponding to the j-th training set; Penalty coefficient, with a value of 10; : The tolerance of the loss function, with a value of 0.01; Slack variables are used to handle errors in nonlinear fitting.

[0168] Training =0.5, =0.1, the number of support vectors is 25.

[0169] 3. Model Application Process

[0170] Typical working conditions and Substituting 0.50 into the SVR model and mapping the kernel function, we get... =0.8, then: After training, =0.6, substituting gives .

[0171] (iv) Algorithm 3: Random Forest Regression Model

[0172] 1. Model building process

[0173] The RFR model improves prediction accuracy through ensemble learning of multiple decision trees, while also outputting a feature importance index. Feedback is sent to the MLR model to achieve interaction with MLR. The model expression is: ; : Occlusion prediction value of the RFR model; Number of decision trees (trees), with a value of 50; : The predicted closure degree of the t-th decision tree.

[0174] Feature Importance Indicators The calculation method is as follows: the percentage of the total reduction in node impurity of the i-th feature across all decision trees, and the comprehensive index. .

[0175] 2. Model Training Process

[0176] The training set and Input the RFR model, construct 50 decision trees using bootstrap sampling, with the splitting criterion for each decision tree being the mean squared error (MSE), and the objective function being: ;in, This represents the average measured occlusion degree within the node.

[0177] After training is complete, output feature importance values: =0.35 ( ), =0.25 ( ), =0.1 ( ), =0.2 ( ), =0.1 ( ).

[0178] Calculate comprehensive index : .

[0179] 3. Model Application Process

[0180] Typical working conditions Substituting the values ​​from 50 decision trees, we obtain the average of the predicted values ​​from each decision tree. =0.62.

[0181] (v) Algorithm fusion: Achieving pairwise interaction and collaborative prediction

[0182] 1. Optimization of the interaction between MLR and RFR

[0183] The output of RFR =0.2535 is fed back to the MLR model, and retraining is performed to determine the result. If the value is 0.4, then the optimized MLR model is: .

[0184] Substitute the numerical values ​​into the calculation: . , which is the result of the linear term calculation for the initial MLR.

[0185] 2. Weighted fusion of SVR and RFR

[0186] Calculate the fusion weights based on the prediction errors of SVR and RFR. The formula is: ;in, The mean squared error of SVR on the training set ( =0.008), The mean squared error of RFR on the training set ( =0.004).

[0187] Substituting, we get: .

[0188] The final prediction value of the fusion model is: ;in, The correction factor for the optimized MLR value is 0.2, determined by cross-validation of the fusion model.

[0189] Substitute the numerical values ​​into the calculation: .

[0190] In a multi-algorithm interactive closed loop, the loop termination condition refers to the criteria for determining whether the system reaches a preset stable state or performance index after the MLR, SVR, and RFR algorithms have completed iterative interactions through feature transfer, weight feedback, and result fusion, thereby terminating the iteration. Loop termination requires meeting at least one of three conditions: core performance index attainment, parameter change convergence, and iteration count threshold (performance and convergence conditions are given priority). The following detailed explanation of the definition, calculation method, and application scenarios of each condition, in conjunction with this embodiment, is provided below.

[0191] Core performance indicator achievement criteria: This criterion uses the prediction accuracy of the fusion model as the core criterion. When the prediction error or goodness of fit of the fusion model on the validation set reaches a preset threshold, the loop closure is considered complete. In this scheme, the mean absolute percentage error (MAPE) and the coefficient of determination (COP) are selected. As a core performance indicator, its specific definition and threshold are as follows.

[0192] 1. The mean absolute percentage error (MAPE) converges to below the threshold.

[0193] MAPE reflects the relative error between the predicted and measured values ​​of the fusion model, and the formula is: m is the number of samples in the validation set (m=10 in this embodiment); This is the fusion model prediction value for the k-th validation sample; is the measured occlusion value of the k-th verification sample.

[0194] End threshold: Preset (Acceptable error range for debris flow blockage prediction in engineering applications).

[0195] This embodiment verifies that the MAPE calculation of 10 validation sets is as follows: This condition is met.

[0196] 2. Coefficient of determination ( Reaching or exceeding the threshold

[0197] The closer the model is to 1, the better the fit. The formula is: ;in, To verify the mean of the measured values ​​of the set.

[0198] End threshold: Preset The fusion model in this scheme This condition is met.

[0199] Convergence condition for parameter variation: This condition applies to key parameters of interaction between algorithms, such as the feature importance weights of MLR. Fusion weights of SVR and RFR When the rate of change of the parameter in two consecutive iterations is less than a preset threshold, the parameter is considered to have reached a stable state, and the closed loop ends. The specific definitions and thresholds are as follows.

[0200] 1. Rate of change of weight parameters convergence

[0201] Importance weights of RFR features of MLR For example, the formula for the rate of change of parameters is: ; For the t-th iteration value; For the (t-1)th iteration value.

[0202] End threshold: Preset If the parameter change is less than 1%, it is considered stable.

[0203] This embodiment verifies: Initial iteration =0, after the first feedback =0.4, and the feature importance of RFR did not change significantly in the second iteration. Keep it at 0.4. The convergence condition is met.

[0204] 2. Rate of change of predicted values ​​from the fusion model convergence

[0205] The rate of change of the predicted values ​​of the fusion model between two adjacent iterations reflects the stability of the results, and the formula is: ;

[0206] End threshold: Preset The first fusion value in this embodiment =0.72, after the second iteration =0.73, The convergence condition is met.

[0207] Iteration count threshold condition: This condition is a fallback measure to prevent the algorithm from getting stuck in infinite iterations due to abnormal situations. Preset maximum number of iterations. When the actual number of iterations At this point, regardless of whether the performance and parameters converge, the closed loop is forcibly terminated.

[0208] The threshold for this scheme is preset based on the algorithmic interaction characteristics of debris flow blockage prediction. =5 times; In this embodiment, the performance and convergence conditions are met after only 2 iterations, and the threshold is not triggered.

[0209] Priority and decision logic of closed-loop termination conditions

[0210] In practical applications, the determination of the loop termination condition follows the logic of prioritizing performance indicators, followed by convergence conditions, and then using the number of iterations as a fallback: 1. If MAPE and convergence conditions are similar during the iteration process... If both conditions are met, the closed loop is terminated directly, and the prediction result of the current fusion model is output; 2. If the performance indicators are not met, but both the rate of change of parameters and the rate of change of predicted values ​​converge, the model is determined to have entered a stable state, the closed loop is terminated, and the performance indicators are optimized; 3. If neither of the first two conditions is met, but the number of iterations reaches the target, the closed loop is terminated, and the prediction result of the current fusion model is output. If the loop is closed, the optimal model result during the iteration process will be output.

[0211] This termination condition design ensures that the prediction accuracy of the fusion model meets engineering requirements while avoiding the waste of computing resources caused by excessive algorithm iteration. It is the core control criterion for the multi-algorithm interactive closed loop.

[0212] (vi) Core contributions of the algorithm

[0213] 1. MLR algorithm: Establishes a basic linear relationship between features and occlusion, providing an initial prediction benchmark for nonlinear algorithms. At the same time, by receiving feature importance feedback from RFR, it dynamically adjusts the model weights to improve the relevance of linear fitting.

[0214] 2. SVR algorithm: It uses kernel function to realize nonlinear feature mapping, solves the nonlinear correlation between debris flow blockage degree and control factors, and introduces the predicted value of MLR as an additional feature to enrich the input dimension and improve the fitting accuracy.

[0215] 3. RFR algorithm: It reduces the overfitting risk of a single decision tree through ensemble learning, and the output feature importance index provides optimization direction for MLR. At the same time, it is weighted and fused with SVR to combine the advantages of the two nonlinear algorithms and ultimately improve the overall accuracy of the prediction model.

[0216] Step 5: Apply the model to predict the degree of occlusion

[0217] The basic parameters of the debris flow channel to be predicted (obtained according to the method in step one): , , ) and dam structure parameters (designed according to the method in step two: =1.1、 =0.0002、 Substituting (=1.65) into the standardized formula and fusion prediction model, we obtain the standardized features: =0.9、 =1.0、 =0、 =1.0、 =0.5. MLR optimized value: =0.65. SVR predicted value: =0.68. RFR predicted value: =0.70.

[0218] Fusion prediction: .

[0219] The test set validation showed that the prediction error of the fusion model was 4.5%, which is much lower than the error of the single algorithm (MLR is 12%, SVR is 7%, and RFR is 5%), achieving accurate prediction of debris flow blockage.

[0220] The core value of algorithm interaction and fusion: In this embodiment, the three algorithms MLR, SVR, and RFR achieve pairwise interaction through feature transfer, weight feedback, and weighted fusion: 1. Interaction between MLR and SVR: The initial prediction value of MLR serves as an additional input feature for SVR, enabling SVR to consider the linear benchmark when fitting nonlinear relationships and reducing the risk of overfitting; 2. Interaction between MLR and RFR: The feature importance output by RFR is fed back to MLR, dynamically adjusting the weight coefficients of MLR, making MLR pay more attention to features that significantly affect occlusion; 3. Interaction between SVR and RFR: Based on weighted fusion of mean squared error, the nonlinear fitting advantage of SVR and the ensemble learning stability of RFR are combined to ultimately improve the prediction accuracy of the model.

[0221] This multi-algorithm fusion strategy solves the problem that a single algorithm is insufficient in fitting the complex nonlinear relationship of debris flow blockage, and provides a more scientific basis for predicting blockage for the engineering design of wedge-beam grid dams.

[0222] In summary, several specific embodiments of the present invention have been disclosed. Without contradiction, the various embodiments can be freely combined to form new embodiments. That is, embodiments that are alternative solutions can be freely substituted for each other, but cannot be combined with each other; embodiments that are not alternative solutions can be combined with each other. These new embodiments are also part of the substantive content of the present invention.

[0223] The above embodiments describe several specific implementations of the present invention. However, those skilled in the art should understand that various changes or modifications can be made to these implementations without departing from the principles and essence of the present invention, but all such changes and modifications fall within the protection scope of the present invention.

Claims

1. A novel method for predicting debris flow blockage degree in wedge-shaped beam grid dams, characterized in that, include: Step 1: Obtain basic parameters, which are divided into debris flow characteristic parameters and channel condition parameters. The debris flow characteristic parameters include debris flow unit weight, debris flow solid volume concentration, primary debris flow scale, and minimum particle size of the largest boulder in the debris flow basin. The channel condition parameters include the average channel width of the proposed wedge-shaped grid dam section in the debris flow basin and the average longitudinal slope of the channel within the proposed wedge-shaped grid dam section. Step 2: Design the structural parameters of the wedge-shaped beam grid dam, including: Determine the wedge structure angle ratio: This angle ratio is the ratio of the actual angle of the wedge structure of the wedge beam grid dam to 60°; Calculate the dam's reservoir capacity and capacity ratio: Based on the average channel width, the initially set dam height, the wedge structure angle ratio, the minimum particle size of the largest rock, and the average channel longitudinal slope, calculate the reservoir capacity of the wedge beam grid dam; then, based on the scale of a debris flow, the volume concentration of debris flow solids, the average channel longitudinal slope, the wedge structure angle ratio, the average channel width, the dam height, and the reservoir capacity, calculate the ratio of the total amount of debris flow solids to the reservoir capacity, i.e., the capacity ratio. Determining the dam opening width and relative opening width: The characteristic particle size of the debris flow basin was obtained through pit exploration tests conducted during field reconnaissance. Based on characteristic particle size Define the opening width of the dam body, and then calculate the relationship between the opening width and the characteristic particle size. The ratio, i.e., the relative opening width; Step 3: Based on the basic parameters obtained in Step 1 and the structural parameters designed in Step 2, conduct a water tank model test to obtain measured data on the degree of closure; Step 4: Construct a blockage prediction model. Collect multiple sets of measured blockage data obtained in Step 3. Using debris flow unit weight, debris flow solid volume concentration, debris flow scale, average channel longitudinal slope, wedge structure angle ratio, relative opening width, and reservoir capacity ratio as control factors, use multiple regression analysis to analyze the correlation between each control factor and the measured blockage data, determine the functional form between each control factor and the blockage, and fit it to obtain the blockage prediction model. Step 5: Apply the model to predict the blockage degree. Input the basic parameters of the debris flow channel to be predicted and the structural parameters of the corresponding wedge beam grid dam into the blockage degree prediction model constructed in Step 4, and calculate the predicted value of the debris flow blockage degree of the wedge beam grid dam under this working condition.

2. The novel wedge-shaped beam grid dam debris flow blockage prediction method according to claim 1, characterized in that, When determining the wedge structure angle ratio in step two, the value of the angle ratio ranges from 0.75 to 1.

25.

3. The novel wedge-shaped beam grid dam debris flow blockage prediction method according to claim 1, characterized in that, In step two, when calculating the storage capacity ratio, the value of the storage capacity ratio ranges from 0 to 4.

0. The total amount of debris flow solid material required for the calculation is obtained by multiplying the debris flow scale by the volume concentration of debris flow solids.

4. The novel wedge-shaped beam grid dam debris flow blockage prediction method according to claim 1, characterized in that, Characteristic particle size in step two This refers to the particle size value corresponding to a cumulative content of 95% in the overall particle size distribution curve of the debris flow basin. The particle size distribution curve is drawn based on the sieving results of soil samples from the debris flow deposition area and the flow area.

5. The novel wedge-shaped beam grid dam debris flow blockage prediction method according to claim 1, characterized in that, In step three, when conducting the flume model test, a test device is set up that includes a hopper, flume, wedge-shaped beam grid dam model, and tailings pool. After the debris flow in the flume has completely stopped moving, the blocked area of ​​the overflow outlet in front of the dam and the total area of ​​the overflow section are recorded immediately, and the ratio of the two is calculated as the measured data of the closure degree.

6. The novel wedge-shaped beam grid dam debris flow blockage prediction method according to claim 1, characterized in that, Before using multiple regression analysis in step four, the control factors are first processed to be dimensionless, so that each control factor is on the same order of magnitude, thereby improving the accuracy of the correlation analysis.

7. The novel wedge-shaped beam grid dam debris flow blockage prediction method according to claim 1, characterized in that, In step two, when determining the opening width of the dam body, the opening width is based on the characteristic particle size. The setting is 1.62 times.

8. The novel wedge-shaped beam grid dam debris flow blockage prediction method according to claim 1, characterized in that, In step four, at least 70 sets of measured occlusion data are collected. After fitting the occlusion prediction model, the goodness of fit of the model is verified by calculating the error between the model's predicted values ​​and the measured values, ensuring the goodness of fit. ≥0.

7.

9. The novel wedge-shaped beam grid dam debris flow blockage prediction method according to claim 1, characterized in that, In step four, when constructing the occlusion degree prediction model, three algorithms—multiple linear regression, support vector machine regression, and random forest regression—are used for integrated modeling. Specifically, the random forest regression model is used to analyze the feature importance of all control factors, identifying the core control factors that significantly affect occlusion degree. The importance weight of these core control factors is then fed back to the multiple linear regression model as feature weight coefficients when constructing linear associations. The initial occlusion degree prediction value output by the multiple linear regression model is then used as additional input features and incorporated into the training samples of the support vector machine regression model. This allows the support vector machine regression model to consider the fundamental laws of linear associations when fitting the nonlinear association between occlusion degree and control factors. Simultaneously, the nonlinear fitting residuals of the support vector machine regression model are fed back to the random forest regression model, adjusting the ensemble weights of the decision trees in the random forest regression model. This allows the random forest regression model to focus more on correcting the fitting of nonlinear residuals. Through the bidirectional feature and result feedback of the three algorithms, accurate fitting of linear and nonlinear associations is achieved.

10. The novel wedge-shaped beam grid dam debris flow blockage prediction method according to claim 9, characterized in that, In step four, when constructing the closure prediction model, model optimization is achieved through hierarchical collaboration and cross-validation of three algorithms: multiple linear regression, support vector machine regression, and random forest regression. Specific operations include: The first step is to divide the measured occlusion data and control factors into training set and validation set. The training set data is used to train a multiple linear regression model to obtain the basic linear prediction model of occlusion. At the same time, the prediction residual of the model on the validation set is calculated. The second step is to use the predicted residuals as the new prediction target, train the support vector machine regression model in combination with control factors, fit the nonlinear correlations that the linear model cannot capture, and obtain the residual correction model. The third step is to use the random forest regression model to fit the control factors and the measured occlusion in all dimensions to obtain the ensemble learning prediction model. The fit between the linear basic model and the residual correction model is evaluated by using the out-of-bag data validation results of the random forest regression model. The fourth step is to adjust the feature coefficients of the multiple linear regression model and the kernel function parameters of the support vector machine regression model based on the validation results of the random forest regression model. Then, the adjusted linear prediction results are superimposed with the nonlinear residual correction results to obtain the preliminary fused prediction values. The fifth step involves weighted fusion of the preliminary fusion predictions with the predictions from the random forest regression model. The weights are dynamically allocated based on the prediction accuracy of the three models on the validation set, ultimately forming a closure prediction model that takes into account linearity, nonlinearity, and ensemble stability.

Images