A method for evaluating safety of a river embankment flood control project

Abstract

The application provides a kind of river embankment flood control project safety evaluation method, it is related to river embankment safety technical field, specific steps include: using the water level data of same flood period in recent 5 years to construct long short-term memory network model, the water level of future 72 time windows is predicted in rolling; water level characteristic parameter is extracted, and characteristic combination parameter is formed by combining inherent parameters of embankment body; coupling model is constructed, river embankment response under the action of flood is simulated, and stress distribution of water surface is obtained; high stress area is formed by merging the maximum principal stress area, and area ratio is calculated; the area ratio, proximity and compactness of high stress area are calculated, and the overall risk coefficient of river embankment is obtained comprehensively, compared with threshold value to evaluate safety condition.The application can more sensitively capture water level change trend through multi-dimensional data fusion and dynamic analysis mechanism, lay a solid data foundation for river embankment safety evaluation, accurately identify stress concentration area and potential risk point, and improve river embankment safety evaluation from fuzzy qualitative to fine quantitative stage.

Description

Technical Field

[0001] This invention relates to the field of riverbank safety technology, specifically a method for safety evaluation of riverbank flood control projects. Background Technology

[0002] Traditional methods for evaluating the safety of river embankment flood control projects often rely on empirical formulas or simplified models, which have many limitations. Assessments based solely on historical water level data and simple structural parameters cannot accurately reflect the actual stress and safety status of river embankments under extreme weather conditions.

[0003] In existing technologies, such as the method and system for early warning of riverbank slope collapse under the influence of fluctuating water levels disclosed in CN116579266A, the following methods are employed: collecting hydrological and soil data of the study area; improving the groundwater calculation module in the BSTEM model; coupling the improved BSTEM model with the SUTRA model to form a SUBS model, and transferring the groundwater information calculated by the SUTRA model to the BSTEM model; using the SUBS model to predict and analyze the stability of riverbank slopes under the influence of fluctuating water levels, thereby achieving early warning of riverbank slope collapse. This method, by coupling the improved BSTEM model with the SUTRA model to form a SUBS model, can be used for predictive analysis of the stability of various types of riverbank slopes, such as homogeneous and composite slopes, affected by fluctuating water levels, providing a new method for early warning of riverbank slope collapse under the influence of fluctuating water levels.

[0004] However, the following shortcomings still exist: it has not fundamentally broken through the constraints of traditional "separate" analysis. Although the BSTEM model has been improved and coupled with the SUTRA model to form the SUBS model, the model coupling only focuses on the data transfer of the hydrological and soil stability calculation modules. It lacks real-time interactive simulation of hydrodynamic action and structural response, and cannot accurately capture the dynamic coupling effect between water flow scouring, seepage pressure and slope structural deformation when the water level fluctuates, resulting in deviations between the stress distribution calculation results and the actual working conditions.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a safety evaluation method for river embankment flood control projects to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] A method for safety evaluation of river embankment flood control projects, comprising the following steps:

[0009] S1. Obtain hourly water level data for the same flood season in the past 5 years to form a time series dataset. Construct a long short-term memory network model. Train the long short-term memory network model with the water levels of the first m time windows and the water level of the (m+1)th time window in the time series dataset. Obtain the water levels of the current time window and the previous m-1 time windows and input them into the long short-term memory network model. Obtain the water levels of the next 72 time windows through rolling prediction.

[0010] S2. Extract water level characteristic parameters based on the water level in the next 72 time windows, collect the inherent parameters of the embankment, and combine the water level characteristic parameters and the inherent parameters of the embankment to form characteristic combination parameters;

[0011] S3. Construct a coupled model of hydrodynamics and structure, input feature combination parameters to simulate the riverbank response under flood action, and obtain the stress distribution on the water-facing side of the riverbank;

[0012] S4. Extract the maximum principal stress in the stress distribution on the water-facing side of the river embankment, merge them into high-stress areas according to the spatial adjacency rule, obtain the area of ​​each high-stress area on the water-facing side of the river embankment, and calculate the area ratio of the high-stress area.

[0013] S5. Extract the minimum distance between each high-stress zone and the toe and top of the embankment as the proximity distance. Based on the proximity distance, obtain the corresponding structural vulnerability coefficient, calculate the proximity of each high-stress zone, obtain the perimeter of each high-stress zone on the water-facing side of the embankment, and calculate the compactness based on the area and perimeter of each high-stress zone on the water-facing side of the embankment.

[0014] S6. Process the area proportion, proximity and compactness of each high-stress zone to obtain the riverbank risk coefficient of each high-stress zone, and weight them to obtain the overall riverbank risk coefficient. Compare it with the preset threshold to evaluate the safety status of the riverbank.

[0015] Furthermore, the inherent parameters of the levee include levee thickness, height, and slope; the water level characteristic parameters include peak water level, maximum rise rate, and flood duration; and the characteristic combination parameters include levee thickness, height, slope, peak water level, maximum rise rate, and flood duration.

[0016] Furthermore, the specific steps of S1 are as follows:

[0017] 1) Obtain hourly water level data of the same flood season at the confluence of rivers over the past 5 years to form a time series dataset;

[0018] 2) Divide the dataset into training, validation, and test sets in a 7:2:1 ratio;

[0019] 3) Construct a long short-term memory network model, which includes:

[0020] Input layer: The input time series dataset contains water levels for m time windows, where each time window is 1 hour, and m is a natural number ≥ 10;

[0021] Network structure: Build 1-2 layers of long short-term memory network, and add a dropout layer after each layer, with a dropout rate of 0.2 to 0.3;

[0022] Output layer: Set up 1 neuron to predict the water level in the (m+1)th time window;

[0023] 4) Training of the Long Short-Term Memory (LSTM) network model:

[0024] Training parameters include the loss function, optimizer, training epochs, and batch size. Mean squared error is used as the loss function, Adam is used as the optimizer, the number of training epochs is set to 50-100, and the batch size is set to 16-32. The training set is divided into multiple input-output pairs according to the window length m.

[0025] Input: Water levels in the first m time windows of the time series dataset;

[0026] Output: The water level in the (m+1)th time window;

[0027] The predicted value is calculated through "forward propagation" and compared with the actual value to obtain the mean square error;

[0028] Adjusting model parameters through "backpropagation" reduces errors;

[0029] After each training round, the validation set is used to evaluate the effect. If the mean squared error does not decrease for 5-10 consecutive rounds, training is stopped.

[0030] The process of obtaining water levels for the next 72 time windows through rolling forecasts:

[0031] Using the water levels of the current time window and the previous m-1 time windows (a total of m time windows) as input, predict the water level of the next time window. Discard the earliest data in the time window and add the newly predicted water level value to the end of the sequence to form a new input. Continue to predict the water level of the next time window, and repeat the above process until the water level prediction for the next 72 time windows is completed.

[0032] Furthermore, water level characteristic parameters are extracted based on the water levels over the next 72 time windows, using the following formula:

[0033] H peak =max{H1,H2,…,H t ,…,H 72}

[0034] Among them, H peak Peak water level, H tLet be the water level in the t-th future time window, where t is the index of the future time window, t∈[1,72];

[0035]

[0036] ΔH t =H t -H t-1

[0037] Among them, v max For the maximum rate of water level rise, ΔH t H represents the water level increment in the t-th future time window. t-1 Let be the water level in the (t-1)th future time window, and ΔSC be the length of the time window.

[0038] Set a flood level threshold yz, and based on the flood level prediction curve, count the number of windows with water levels ≥ the flood level threshold yz in 72 future time windows, and sum them up to obtain the flood duration T.

[0039] Furthermore, the specific steps of S4 are as follows:

[0040] The upstream face of the river embankment is divided into multiple finite element elements. Based on the stress distribution on the upstream face of the river embankment, the stress components of each finite element element are extracted, including three normal stress components σ. x σ y σ z For each finite element element, the maximum value is selected from its stress components, which is the maximum principal stress σ of the finite element element. max , σ max =nax(σ x ,σ y ,σ z ), σ x ,σ y ,σ z These are the normal stresses in the x-direction, y-direction, and z-direction, respectively.

[0041] σ max >σ th The finite element is labeled as a high-stress element, where σ th The stress threshold;

[0042] High-stress elements are merged using spatial adjacency rules, as follows:

[0043] 1) Randomly select a high-stress element as the starting center;

[0044] 2) Define the spatial adjacency rule as follows: if there is at least one common edge between two adjacent high-stress elements, find high-stress elements that satisfy the spatial adjacency rule with the central element and merge them into the same high-stress region;

[0045] 3) Using the newly merged unit as the new center, repeat step 2 to continuously search for and merge units that satisfy the spatial adjacency rules;

[0046] 4) When no high-stress element conforms to the spatial adjacency rule, stop the iteration and complete the division of a single high-stress region;

[0047] 5) If there are undivided high-stress elements, repeat steps 1-4 until all high-stress elements are divided into a unique high-stress region.

[0048] The formula for calculating the area ratio of high-stress zones is as follows:

[0049]

[0050] Among them, R i Let A be the area percentage of the i-th high-stress region. i Let A be the area of ​​the i-th high-stress zone on the water-facing side of the riverbank. total Let i be the total area of ​​the riverbank's water-facing surface, i∈[1,n], and n be the number of high-stress zones.

[0051] Furthermore, the corresponding structural vulnerability coefficients are obtained based on different proximity distances, using the following formula:

[0052]

[0053] Where, β i Let μ be the structural fragility coefficient of the i-th high-stress zone. foot The anti-slip coefficient of the embankment toe material is μ. top d is the anti-sliding coefficient of the embankment crest material. min,i Let d be the proximity distance of the i-th high-stress zone. foot,i d represents the shortest distance between the i-th high-stress zone and the toe of the dike. top,i Let γ1 be the shortest distance between the i-th high-stress zone and the top of the dike, γ1 be the adjustment coefficient of the dike toe, γ1∈[1.2,1.5], and γ2 be the adjustment coefficient of the dike top, γ2∈[0.8,1.2];

[0054] The proximity of each high-stress zone is calculated using the following formula:

[0055]

[0056] Where, d i The proximity of the i-th high-stress zone;

[0057]

[0058] Among them, JC i For the compactness of the i-th high-stress region, Pi Let A be the perimeter of the i-th high-stress zone on the water-facing side of the riverbank. i Let be the area of ​​the i-th high-stress zone on the water-facing side of the riverbank.

[0059] Furthermore, the area proportion, proximity, and compactness of each high-stress zone are processed to generate the riverbank risk coefficient for each high-stress zone, based on the following formula:

[0060] RI i =ω1R i +ω2d i +ω3JC i

[0061] Among them, RI i Let be the risk coefficient of the river embankment in the i-th high-stress zone;

[0062] In the formula, ω1 is the weighting coefficient of area proportion, ω2 is the weighting coefficient of proximity, and ω3 is the weighting coefficient of compactness. Based on ω1+ω2+ω3=1, let 0<ω3<ω2<ω1<1.

[0063] Furthermore, the overall risk coefficient of the riverbank is obtained by weighting the risk coefficients of each high-stress zone, based on the following formula:

[0064]

[0065] Wherein, RI is the overall risk coefficient of the river embankment;

[0066] The safety status of the river embankment is assessed by comparing the overall risk coefficient with a preset threshold. The specific steps are as follows:

[0067] When 0 <RI<RI th1 The river embankment is classified as low-risk.

[0068] When RI th1 ≤RI≤RI th2 The river embankment is classified as medium-risk.

[0069] When RI>RI th2 The river embankment is classified as high-risk.

[0070] Among them, RI th1 The threshold for classifying low-risk and medium-risk levels, RI th2 The threshold for classifying medium-risk and high-risk levels.

[0071] Compared with the prior art, the beneficial effects of the present invention are:

[0072] This invention obtains meteorological data such as hourly precipitation and precipitation intensity for the next 72 hours from meteorological departments, and simultaneously collects hydrological data such as real-time water level and flow velocity at river confluences. Then, it uses a flood evolution model to make preliminary predictions of water level changes. By using a long short-term memory network to fuse historical water level data from the same period, it iteratively optimizes the initial prediction curve and accurately extracts key indicators such as peak water level and maximum water level rise rate. This multi-dimensional data fusion and dynamic analysis mechanism, compared with existing technologies, can more sensitively capture water level change trends and lay a solid data foundation for riverbank safety assessment.

[0073] By constructing a hydrodynamic-structural coupled finite element model and inputting characteristic combination parameters into the model, the influence of water level changes on the riverbank structure can be realistically simulated. The dynamic coupling effect between water flow scouring, seepage pressure and riverbank structure deformation can be accurately reflected. The stress distribution results obtained are more in line with the actual working conditions, which significantly improves the accuracy of riverbank stress analysis.

[0074] Starting from the finite element level, high-stress zones are identified based on stress distribution, and their area ratio, proximity, compactness, and other multi-dimensional characteristic indicators are calculated in depth. By considering the size of the high-stress zone, its proximity to the embankment toe and top, and its compact shape, the impact on river embankment safety is comprehensively assessed. This allows for the accurate identification of stress concentration areas and potential risk points, elevating river embankment safety assessment from a fuzzy qualitative stage to a refined quantitative stage, and constructing a more scientific and comprehensive risk assessment system. Attached Figure Description

[0075] Figure 1 This is a schematic diagram of the overall method flow of the present invention;

[0076] Figure 2 This is a fitted curve of the structural fragility coefficient and proximity of the present invention;

[0077] Figure 3 This is a fitted curve of proximity distance versus proximity in this invention;

[0078] Figure 4 This is a fitted curve of the area ratio of the present invention and the riverbank risk coefficient;

[0079] Figure 5 This is a fitted curve of proximity and riverbank risk coefficient for the present invention;

[0080] Figure 6 This is a fitted curve of the compactness of the invention and the riverbank risk coefficient. Detailed Implementation

[0081] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0082] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0083] Example 1:

[0084] Please see Figure 1 The present invention provides a technical solution:

[0085] A method for safety evaluation of river embankment flood control projects, comprising the following steps:

[0086] S1. Obtain hourly water level data for the same flood season in the past 5 years to form a time series dataset. Construct a long short-term memory network model. Train the long short-term memory network model with the water levels of the first m time windows and the water level of the (m+1)th time window in the time series dataset. Obtain the water levels of the current time window and the previous m-1 time windows and input them into the long short-term memory network model. Obtain the water levels of the next 72 time windows through rolling prediction.

[0087] Based on the above embodiments, the inherent parameters of the embankment include the embankment thickness, height, and slope; the water level characteristic parameters include peak water level, maximum rise rate, and flood duration; and the characteristic combination parameters include embankment thickness, height, slope, peak water level, maximum rise rate, and flood duration.

[0088] Among them, the thickness of the embankment body refers to the horizontal thickness of the cross section of the embankment body, the height of the embankment body refers to the vertical distance from the toe of the slope to the top of the embankment, and the slope angle refers to the slope angle on the water-facing side and the back water-facing side.

[0089] The methods for collecting data on the horizontal thickness of the dike cross-section, the vertical distance from the toe of the slope to the top of the dike, and the slope angles on the upstream and downstream sides are as follows:

[0090] The horizontal thickness of a cross-section is essentially the "horizontal distance from the water-facing side to the back-facing side." In a measurement coordinate system, the X-axis is usually set perpendicular to the riverbank axis (i.e., the width direction of the cross-section), and the Y-axis is set parallel to the riverbank axis (i.e., the extension direction). The toe of the water-facing slope is the intersection of the water-facing slope and the riverbed, denoted as point U1, and the toe of the back-facing slope is the intersection of the back-facing slope and the ground behind the embankment, denoted as point U2. The coordinates of U1 and U2 in the Y-axis direction are almost the same (because there is no distance along the axis direction within the same cross-section), and the difference in the X-axis coordinates is the horizontal distance between the two points in the "width direction"—which is exactly equal to the horizontal thickness of the embankment. The average of the horizontal thicknesses of multiple cross-sections is calculated as the horizontal thickness of the cross-section.

[0091] Select a leveling benchmark near the dike, denoted as BM, with its elevation being a known value h. BM Erect a leveling rod at point U1 on the water-facing slope toe, read the backsight reading 'a' using the level instrument, and calculate the elevation of point U1. Similarly, measure the elevation of point U2 at the toe of the backwater slope.

[0092] Set up a ruler at the midpoint of the embankment crest and read the elevation h. pd The height of the embankment, h

[0093] Select a regular slope with a vertical height of not less than 2m on the water-facing slope, mark the top of the slope as M and the bottom of the slope as N; use a level to measure the elevation difference Δh between points M and N. MN As the horizontal distance, the horizontal projected distance L between points M and H is measured using a total station. The slope is the ratio of the vertical height to the horizontal distance, and the slope angle is...

[0094] Similarly, the process for obtaining the back slope angle is the same as that for the front slope angle mentioned above.

[0095] Based on the above embodiments, the collected embankment thickness, height, and slope angle all need to be standardized, and all subsequent analyses and calculations are based on the standardized data.

[0096] Based on the above embodiments, the specific steps of S1 are as follows:

[0097] 1) Obtain hourly water level data of the same flood season at the confluence of rivers over the past 5 years to form a time series dataset;

[0098] 2) Divide the dataset into training, validation, and test sets in a 7:2:1 ratio;

[0099] 3) Construct a long short-term memory network model, which includes:

[0100] Input layer: The input time series dataset contains water levels for m time windows, where each time window is 1 hour, and m is a natural number ≥ 10;

[0101] Network structure: Build 1-2 layers of long short-term memory network, and add a dropout layer after each layer, with a dropout rate of 0.2 to 0.3;

[0102] Output layer: Set up 1 neuron to predict the water level in the (m+1)th time window;

[0103] 4) Training of the Long Short-Term Memory (LSTM) network model:

[0104] Training parameters include the loss function, optimizer, training epochs, and batch size. Mean squared error is used as the loss function, Adam is used as the optimizer, the number of training epochs is set to 50-100, and the batch size is set to 16-32. The training set is divided into multiple input-output pairs according to the window length m.

[0105] Input: Water levels in the first m time windows of the time series dataset;

[0106] Output: The water level in the (m+1)th time window;

[0107] The predicted value is calculated through "forward propagation" and compared with the actual value to obtain the mean square error;

[0108] Adjusting model parameters through "backpropagation" reduces errors;

[0109] After each training round, the validation set is used to evaluate the effect. If the mean squared error does not decrease for 5-10 consecutive rounds, training is stopped.

[0110] The process of obtaining water levels for the next 72 time windows through rolling forecasts:

[0111] Using the water levels of the current time window and the previous m-1 time windows (a total of m time windows) as input, predict the water level of the next time window. Discard the earliest data in the time window and add the newly predicted water level value to the end of the sequence to form a new input. Continue to predict the water level of the next time window, and repeat the above process until the water level prediction for the next 72 time windows is completed.

[0112] Based on the above embodiments, water level characteristic parameters are extracted according to the water levels over the next 72 time windows, using the following formula:

[0113] H peak =max{H1,H2,…,H t ,…,H 72}

[0114] Among them, H peak Peak water level, Ht Let be the water level in the t-th future time window, which refers to the average water level within the time window, and t be the index of the future time window, t∈[1,72];

[0115]

[0116] ΔH t =H t -H t-1

[0117] Among them, v max For the maximum rate of water level rise, ΔH t H represents the water level increment in the t-th future time window. t-1 Let be the water level in the (t-1)th future time window, and ΔSC be the length of the time window.

[0118] Set a flood level threshold yz, and based on the flood level prediction curve, count the number of windows with water levels ≥ the flood level threshold yz in 72 future time windows, and sum them up to obtain the flood duration T.

[0119] S3. Construct a coupled model of hydrodynamics and structure, input feature combination parameters to simulate the riverbank response under flood action, and obtain the stress distribution on the water-facing side of the riverbank;

[0120] Based on the above embodiments, the specific steps of S3 are as follows:

[0121] 1) Building the basic model

[0122] Use fluid dynamics software to build a hydrodynamic model, input the geometric boundaries of the river channel and riverbank, including the thickness, height and slope of the riverbank, and define the model dimensions;

[0123] Use structural software to build a river embankment structural model, assign material properties to the embankment body, including strength parameters of concrete / soil, and associate it with the water flow model through a coupling interface;

[0124] 2) Set coupling rules

[0125] The flow model calculates the pressure of the flood on the upstream surface and automatically transfers it to the structural model, which only considers static forces, thus coupling the iterative process.

[0126] 3) Feature combination parameter input

[0127] Embankment geometry parameters: Input the embankment thickness, height, and slope directly into the structural model, and the model will automatically generate the embankment structure of the corresponding dimensions;

[0128] Flood dynamic parameters: peak water level, which is the highest inundation height of the flood; maximum rise rate, which controls how fast the water level rises; flood duration, which sets the time during which the peak water level is maintained.

[0129] 4) Simulation and Result Extraction

[0130] Starting from the initial water level, the water rises at the maximum rate of increase to the peak water level, and the flood continues for the duration specified in the text.

[0131] Automatically record the pressure changes at various points on the water-facing surface over time and generate a pressure distribution map at the peak water level, which serves as the load input for the structural model;

[0132] Import the pressure distribution map at the peak water level output by the flow model into the structural model. The system will automatically apply the pressure at each point to the same position on the upstream side of the riverbank. At the same time, activate the self-weight of the riverbank in the structural model and solve the stress distribution on the upstream side through the mechanical equilibrium equation.

[0133] Export the stress cloud map of the water-facing surface, mark the areas exceeding the allowable stress value of the material as high-stress areas, and directly obtain the maximum stress value and distribution range of the high-stress areas.

[0134] S4. Extract the maximum principal stress in the stress distribution on the water-facing side of the river embankment, merge them into high-stress areas according to the spatial adjacency rule, obtain the area of ​​each high-stress area on the water-facing side of the river embankment, and calculate the area ratio of the high-stress area.

[0135] Based on the above embodiments, the upstream surface of the river embankment is divided into multiple finite element elements. According to the stress distribution on the upstream surface of the river embankment, the stress components of each finite element element are extracted, including three normal stress components σ. x σ y σ z For each finite element element, the maximum value is selected from its stress components, which is the maximum principal stress σ of the finite element element. max , σ max =max(σ x ,σ y ,σ z ), σ x ,σ y ,σ z These are the normal stresses in the x-direction, y-direction, and z-direction, respectively.

[0136] σ max >σ th The finite element is labeled as a high-stress element, where σ th The stress threshold;

[0137] High-stress elements are merged using spatial adjacency rules, as follows:

[0138] 1) Randomly select a high-stress element as the starting center;

[0139] 2) Define the spatial adjacency rule as follows: if there is at least one common edge between two adjacent high-stress elements, find high-stress elements that satisfy the spatial adjacency rule with the central element and merge them into the same high-stress region;

[0140] 3) Using the newly merged unit as the new center, repeat step 2 to continuously search for and merge units that satisfy the spatial adjacency rules;

[0141] 4) When no high-stress element conforms to the spatial adjacency rule, stop the iteration and complete the division of a single high-stress region;

[0142] 5) If there are undivided high-stress elements, repeat steps 1-4 until all high-stress elements are divided into a unique high-stress region.

[0143] The formula for calculating the area ratio of high-stress zones is as follows:

[0144]

[0145] Among them, R i Let A be the area percentage of the i-th high-stress region. i Let A be the area of ​​the i-th high-stress zone on the water-facing side of the riverbank. total Let i be the total area of ​​the riverbank's water-facing surface, i∈[1,n], and n be the number of high-stress zones.

[0146] S5. Extract the minimum distance between each high-stress zone and the toe and top of the embankment as the proximity distance. Based on the proximity distance, obtain the corresponding structural vulnerability coefficient, calculate the proximity of each high-stress zone, obtain the perimeter of each high-stress zone on the water-facing side of the embankment, and calculate the compactness based on the area and perimeter of each high-stress zone on the water-facing side of the embankment.

[0147] Based on the above embodiments, the corresponding structural vulnerability coefficients are obtained according to different proximity distances, using the following formula:

[0148]

[0149] Where, β i Let μ be the structural fragility coefficient of the i-th high-stress zone. foot μ is the anti-slip coefficient of the embankment toe material. top d is the anti-sliding coefficient of the embankment crest material. min,i Let d be the proximity distance of the i-th high-stress zone. foot,i d represents the shortest distance between the i-th high-stress zone and the toe of the dike. top,i This represents the shortest distance between the i-th high-stress zone and the top of the dike.

[0150] Among them, the anti-slip coefficient of the embankment toe material reflects the ability of the material at the embankment toe to resist shear sliding. It is the ratio of the shear strength of the embankment toe material to the normal stress. The larger the value, the more difficult it is for the embankment toe material to slip under vertical load and horizontal thrust, and the stronger the anti-slip stability.

[0151] The shear strength of the soil at the toe of the embankment was directly determined by in-situ direct shear test or vane shear test, and the anti-sliding coefficient of the embankment material was calculated by combining the measured normal stress on site.

[0152] The anti-slip coefficient of the embankment crest material reflects the ability of the embankment crest surface material to resist sliding along the surface. It is usually the coefficient of friction between the embankment crest material and the contacting object (such as rainwater runoff layer, vehicle tires), and is directly related to the anti-slip performance of the embankment crest under local loads (such as vehicles, water accumulation). The larger the value, the less likely the embankment crest is to experience local sliding or collapse.

[0153] For hard materials such as concrete and masonry on the top of the dike, an inclined plane friction test is used (a standard slider is placed on the surface of the top of the dike, and the ratio of frictional force to normal force is measured when the slider slides); for soil dike tops, the shear strength of the surface soil is determined by a ring shear test and converted into the friction coefficient.

[0154] Based on this, it should be noted that:

[0155] When d min,i =d fooy,i At that time, β i =γ1· foot ;

[0156] The toe of the dike is the connection point between the dike and the foundation, bearing all the vertical loads of the dike and the lateral thrust of the water flow. Its anti-sliding stability directly determines whether the dike will slide as a whole. Once the toe becomes unstable, the entire dike will lose its support and collapse, which is a fatal failure. Therefore, by using a high coefficient γ1∈[1.2,1.5], the weight of the relevant parameters of the toe is forcibly increased in the vulnerability assessment, which is perfectly matched with the engineering characteristic that the consequences of the toe failure are irreversible.

[0157] When d min,i =d top,i At that time, β i =γ2·μ top ;

[0158] The main function of the levee crest is to resist flooding and bear temporary loads. Its instability is mostly manifested as local collapse or surface erosion, and will not directly lead to the overall collapse of the levee, which is a localized damage. Therefore, the low coefficient setting of γ2∈[0.8,1.2] objectively reflects the weak impact of levee crest damage on overall stability, which is consistent with the actual characteristic that levee crest damage is repairable.

[0159] When d min,i =dfoot,i =d top,i When the time is right, take the average of the two.

[0160] At this point, the threat posed by the high-stress zone to the toe and crest of the dike is in a critical equilibrium state. The average value calculation neither ignores the fatal risk at the toe nor exaggerates the local impact on the crest, accurately corresponding to the special working condition of balanced threat at both locations.

[0161] Therefore, the above-mentioned functional form is used to express the functional relationship between the proximity of the high-stress zone and the anti-sliding coefficients of the embankment toe material and the embankment crest material.

[0162] In the formula, γ1 is the toe adjustment coefficient, γ1∈[1.2,1.5], and γ2 is the crest adjustment coefficient, γ2∈[0.8,1.2];

[0163] As an "irreparable part," the toe of the dike needs to have a higher safety margin. Statistics show that more than 90% of dike breaches are caused by the instability of the toe. Therefore, an additional 20%-50% weight (i.e., 1.2-1.5 times) should be added to the baseline weight of the adjustment coefficient, and the baseline value of the toe adjustment coefficient should be 1.

[0164] Most breaches at the top of the levee are localized collapses, with a safety factor threshold typically of 1.5 (lower than the toe, due to lower repair difficulty). Actual measurement data shows that even with a levee top anti-sliding coefficient μ... top Even with a 15% decrease, the safety factor remains at 1.6-1.8, indicating a gradual change in risk. Therefore, γ2 does not need to be excessively amplified; a value between 0.8 and 1.2 is sufficient. This avoids exaggerating the impact of localized damage while still reflecting the increased risk to the levee crest under flood inundation.

[0165] Table 1. Changes in proximity with structural vulnerability coefficient and proximity distance

[0166] Serial Number Structural fragility coefficient Proximity Distance Proximity 1 11.70 30.00 0.39 2 12.50 29.75 0.42 3 13.05 29.00 0.45 4 13.68 28.50 0.48 5 14.28 28.00 0.51 6 14.85 27.50 0.54 7 15.39 27.00 0.57 8 15.90 26.50 0.60 9 16.38 26.00 0.63 10 16.83 25.50 0.66 11 17.25 25.00 0.69 12 17.64 24.50 0.72 13 18.00 24.00 0.75 14 18.33 23.50 0.78 15 18.63 23.00 0.81 16 18.90 22.50 0.84 17 19.20 22.00 0.87 18 19.50 21.50 0.90 19 19.80 21.00 0.94 20 20.10 20.50 0.98

[0167] As shown in Table 1, proximity is positively correlated with the structural vulnerability coefficient: as the structural vulnerability coefficient gradually increases from 11.70 to 20.10, the proximity also increases from 0.39 to 0.98, indicating that the higher the structural vulnerability, the stronger the correlation between the proximity risk of the high stress zone and the surrounding area.

[0168] Proximity is negatively correlated with proximity distance: when proximity distance gradually decreases from 30.00 to 20.50, proximity continuously increases from 0.39 to 0.98, indicating that the closer the high-stress area is to the reference object, the higher its spatial proximity correlation.

[0169] according to Figures 2-3As the structural vulnerability coefficient increases, the proximity degree shows a significant upward trend, and the fitted line is a straight line, indicating that the two are significantly positively correlated, that is, the higher the structural vulnerability coefficient, the greater the proximity degree; when the proximity distance increases, the proximity degree continues to decrease, and the fitted line is a straight line, indicating that the two are significantly negatively correlated, that is, the farther the proximity distance, the smaller the proximity degree.

[0170] Based on the above embodiments, the proximity of each high-stress zone is calculated using the following formula:

[0171]

[0172] Where, d i Let represent the proximity of the i-th high-stress zone. The proximity is used to assess the potential threat of the high-stress zone to the toe and top of the embankment by combining two index parameters: structural vulnerability coefficient and proximity distance. The greater the proximity, the higher the degree of danger association between the high-stress zone and the toe and top of the embankment. The stress in the high-stress zone is more likely to be transmitted to the toe and top of the embankment and cause structural damage, thereby increasing the risk level of the embankment.

[0173] Based on this, it should be noted that:

[0174] Structural fragility coefficient β i This reflects the fragility of the embankment structure surrounding the high-stress zone, as indicated by the structural fragility coefficient β. i When the stress level increases, it means that the deformation resistance and crack resistance of the embankment material surrounding the high-stress zone decrease significantly. At this time, even if the stress level in the high-stress zone does not increase significantly, the surrounding fragile structures will find it difficult to resist stress transmission and are prone to damage under stress. This combination of high stress and weak resistance will greatly increase the efficiency of the threat transmission from the high-stress zone to the embankment toe and top, resulting in an increase in proximity.

[0175] Proximity distance d min,i Reflecting the spatial distance relationship between the high-stress zone and the toe and crest of the dike, in terms of proximity distance d min,i When the distance is reduced, meaning the high-stress zone is closer to the toe or crest of the embankment, the stress transmission path within the embankment material is shortened, energy loss is reduced, and stress in the high-stress zone can be transmitted to the toe or crest more directly and efficiently. When the distance of the high-stress zone from the toe is reduced from 5m to 2m, the attenuation of stress transmitted to the toe through contact between soil particles decreases, and the additional stress borne by the toe increases significantly. Since the toe itself is the weakest point of the river embankment, stress concentration over a short distance is more likely to exceed its bearing capacity, leading to increased proximity.

[0176] Therefore, the proximity of high-stress zones is positively correlated with the structural vulnerability coefficient; the proximity of high-stress zones is negatively correlated with proximity distance.

[0177] The threat posed by high-stress zones is not βi and d min,i It is not a simple superposition, but a synergistic effect of vulnerability enhancement and distance reduction: when β i Larger and d min,i At smaller values, the synergy between the two can lead to a product-like amplification of the threat (e.g., β). i Increase by 2 times while d min,i Reduced to half, threat amplified 4 times. (Fractional structure) It can accurately capture this synergistic effect, and its calculation results are highly consistent with the high-risk characteristics of the "high vulnerability + close proximity" combination in actual engineering.

[0178] In summary, the above-mentioned functional form is used to express the functional relationship between the proximity of high-stress zones and the structural vulnerability coefficient and proximity distance.

[0179]

[0180] Among them, JC i For the compactness of the i-th high-stress region, P i Let A be the perimeter of the i-th high-stress zone on the water-facing side of the riverbank. i Let be the area of ​​the i-th high-stress zone on the water-facing side of the riverbank.

[0181] Based on the above embodiments, the area A of the i-th high-stress zone on the water-facing surface of the riverbank... i The spatial extent of the high-stress zone is obtained by accumulating the surface area of ​​the combined high-stress elements; the perimeter P of the i-th high-stress zone on the water-facing surface of the riverbank is also represented. i It is obtained by identifying the outer boundary of the high-stress zone and accumulating the side lengths, and is used to characterize the compactness of the region shape.

[0182] S6. Process the area proportion, proximity and compactness of each high-stress zone to obtain the riverbank risk coefficient of each high-stress zone, and weight them to obtain the overall riverbank risk coefficient. Compare it with the preset threshold to evaluate the safety status of the riverbank.

[0183] Table 2. Variation of Riverbank Risk Coefficient with Area Proportion, Proximity, and Compactness

[0184]

[0185] As shown in Table 2, when the area proportion increases from 0.2 to 0.68, the proximity increases from 0.1 to 0.58, and the compactness increases from 0.1 to 0.58, the riverbank risk coefficient continuously rises from 0.15 to 0.63, with all three maintaining a synchronous monotonically increasing trend. Therefore, the riverbank risk coefficient shows a significant positive correlation with the area proportion, proximity, and compactness; the larger the area, proximity, and compactness of the high-stress zone, the higher the comprehensive risk coefficient of the riverbank.

[0186] according to Figures 4-6 As the area proportion increases, the riverbank risk coefficient shows an upward trend, and the fitted line is a straight line, indicating a significant positive correlation between the two. That is, the higher the area proportion, the greater the riverbank risk coefficient. When the proximity increases, the riverbank risk coefficient continues to rise, and the fitted line is a straight line, indicating a significant positive correlation between the two. That is, the higher the proximity, the greater the riverbank risk coefficient. When the compactness increases, the riverbank risk coefficient also increases, and the fitted line is a straight line, reflecting a significant positive correlation between the two. That is, the higher the compactness, the greater the riverbank risk coefficient.

[0187] Based on the above embodiments, the area proportion, proximity, and compactness of each high-stress zone are processed to generate the riverbank risk coefficient for each high-stress zone, according to the following formula:

[0188] RI i =ω1R i +ω2d i +ω3JC i

[0189] Among them, RI i Let be the risk coefficient of the embankment in the i-th high-stress zone. The risk coefficient is used to assess the risk level of the embankment by combining three indicators: the area ratio of the high-stress zone, its proximity, and its compactness. The larger the risk coefficient, the higher the risk level of the embankment and the lower the safety of the flood control project.

[0190] Based on this, it should be noted that:

[0191] The area ratio R of the high-stress zone i As the stress increases, the high-stress zone expands across the upstream side of the riverbank, meaning more areas bear stresses exceeding the material's allowable limits. For example, if R... i The increase from 10% to 30% indicates that nearly one-third of the water-facing surface is under high stress, the scope of overall structural stress exceeding limits has expanded, and the probability of localized damage (such as cracks and seepage) has increased significantly. This widespread stress exceeding limits will directly exacerbate the overall vulnerability of the river embankment, leading to a decrease in the safety of flood control projects, an increase in the risk level of the river embankment, and consequently an increase in the risk coefficient of the river embankment.

[0192] Proximity d of high stress zone i As the stress level increases, the dangerous correlation between high-stress areas and vulnerable parts such as the levee toe and levee crest intensifies. For example, when d... iThe increase in stress could be due to the high-stress zone being closer to the embankment toe, which is a critical part of the embankment's anti-sliding stability; exceeding the stress limit at the toe can easily lead to overall slippage. Alternatively, the embankment crest material near the high-stress zone may have a lower anti-sliding coefficient, making stress transmission more likely to cause crest collapse. This combined effect of high stress and proximity to weak points significantly increases the risk of local damage spreading to overall instability, leading to reduced safety of flood control projects, increased embankment risk level, and consequently, increased embankment risk coefficient.

[0193] Compactness of high-stress areas JC i When the stress increases, the shape of the high-stress zone becomes more concentrated, and the degree of stress accumulation in the local area intensifies. For example, an increase in stress means that the perimeter of the high-stress zone of the same area is shorter, and the stress distribution is more concentrated in a small area, resulting in a significant increase in the stress peak value in that area. This highly concentrated stress is prone to exceeding the material's bearing capacity limit, triggering local brittle failure, leading to a decrease in the safety of flood control projects, an increase in the risk level of river embankments, and consequently an increase in the risk coefficient of river embankments.

[0194] Therefore, the risk coefficient of the river embankment in the high-stress zone is positively correlated with its area proportion, proximity, and compactness.

[0195] The effects of area proportion, proximity, and compactness on riverbank risk are independent yet cumulative:

[0196] The area ratio reflects the scope effect of risk (the larger the ratio, the more dangerous);

[0197] Proximity reflects the location effect of risk (the closer to the weak point, the more dangerous it is);

[0198] The degree of concentration reflects the intensity of the risk (the more concentrated, the more dangerous).

[0199] All three factors influence risk through different dimensions, and their directions of influence are consistent (all are positively correlated). Their linear superposition can intuitively demonstrate the risk accumulation effect of 1+1>1, which is consistent with the actual understanding that multiple factors synergistically exacerbate risk.

[0200] In summary, the above-mentioned functional form is used to express the functional relationship between the risk coefficient of the river embankment in the high-stress zone and its area proportion, proximity, and compactness.

[0201] In the formula, ω1 is the weighting coefficient of area proportion, ω2 is the weighting coefficient of proximity, and ω3 is the weighting coefficient of compactness.

[0202] Area R i The spatial coverage of high-stress zones has a fundamental impact on risk. Even if the stress concentration in a single area is low (low compactness) or far from weak points (low proximity), a large-scale distribution of high stress can still reduce the overall bearing capacity of the riverbank and increase the probability of concurrent multi-point failures. For example, when R... iWhen the stress level is greater than 50%, more than half of the riverbank's upstream side is under excessive stress, which may trigger a chain reaction of leakage or collapse. The consequences of the damage are global, so its weight ω1 is the highest.

[0203] Proximity d i This reflects the degree of correlation between high-stress areas and the toe and crest of the dike. These areas (such as the toe) are the core of the dike's anti-sliding stability. If high-stress areas are close to them, stress can be directly transmitted to weak points through the materials, accelerating the transformation of local damage into overall instability (such as toe slippage leading to the tilting of the entire dike). However, its impact is still limited to the local-to-global transmission path, and the severity of the consequences is weaker than the spread of large-area risks. Therefore, the weight ω2 is second only to the weight ω1 of the area proportion.

[0204] Compactness JC i This reflects the degree of stress concentration within a small area, primarily affecting the probability of local material failure (e.g., excessively high stress peaks in a small area may lead to localized cracks). However, if the area of ​​this high-stress zone is small (R... i Small) or away from weak points (d i Small-scale damage is unlikely to spread to the entire riverbank, posing a relatively limited threat to overall safety. Therefore, its weight ω3 is the lowest.

[0205] Therefore, based on ω1+ω2+ω3=1, let 0<ω3<ω2<ω1<1.

[0206] As one implementation method, the value range of ω1 is 0.4-0.6, the value range of ω2 is 0.25-0.4, and the value range of ω3 is 0.1-0.25. The specific values ​​are set by the technicians according to the actual situation and are not limited here.

[0207] Based on the above embodiments, the overall risk coefficient of the river embankment is obtained by weighting the risk coefficients of each high-stress zone, according to the following formula:

[0208]

[0209] RI represents the overall risk coefficient of the river embankment.

[0210] Based on the above embodiments, the safety status of the river embankment is assessed by comparing the overall risk coefficient with a preset threshold. The specific steps are as follows:

[0211] When 0 <RI<RI th1 The river embankment is classified as low-risk.

[0212] When RI th1 ≤RI≤RI th2 The river embankment is classified as medium-risk.

[0213] When RI>RIth2 The river embankment is classified as high-risk.

[0214] Among them, RI th1 The threshold for classifying low-risk and medium-risk levels, RI th2 The threshold for classifying medium-risk and high-risk levels.

[0215] Collect historical risk assessment data for similar river embankments, including: the overall risk coefficient of each assessment, the actual safety status of the corresponding period, and the actual safety status including "low risk" corresponding to no damage, "medium risk" corresponding to local cracks, and "high risk" corresponding to embankment collapse.

[0216] The data is cleaned, outliers are removed, and a "RI-Safety Status" sample set is formed;

[0217] Cluster analysis was used to automatically group the RI values ​​in the sample set and identify natural breakpoints in the data: the first breakpoint was the boundary between low and medium risk (corresponding to the mean RI values ​​of no damage and local cracks), which was used as the RI. th1 Initial value; the second breakpoint is the boundary between medium and high risk (corresponding to the mean RI of local cracks and levee breaches), and is used as RI. th2 Initial value.

[0218] The accuracy of the threshold is verified using a validation set. If more than 90% of the low-risk group shows no actual damage and more than 80% of the high-risk group shows serious damage, then the threshold is reasonable.

[0219] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0220] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by software, electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0221] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0222] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for safety evaluation of river embankment flood control projects, characterized in that, The specific steps include: S1. Obtain the hourly water level data of the same flood season in the past 5 years to form a time series data set, and construct a long short-term memory network model to predict the water level of the m+1 time window from the m time window in the time series data set. The water level of the current time window and the previous m-1 time windows is input into the long short-term memory network model, and the water level of the future 72 time windows is obtained by rolling prediction; S2. Extract the water level feature parameters from the water level of the future 72 time windows, collect the inherent parameters of the embankment, and combine the water level feature parameters and the inherent parameters of the embankment to form the feature combination parameters; S3. Construct a coupling model of water flow dynamics and structure, input the feature combination parameters to simulate the response of the river embankment under the action of flood, and obtain the stress distribution on the water surface of the river embankment; S4. Extract the maximum principal stress in the stress distribution on the water surface of the river embankment, merge it according to the spatial adjacency rule to form a high stress area, and obtain the area of each high stress area on the water surface of the river embankment to calculate the area ratio of the high stress area; S5. Extract the minimum distance between each high stress area and the embankment toe and the embankment top as the proximity distance, obtain the corresponding structural vulnerability coefficient according to the proximity distance, calculate the proximity of each high stress area, obtain the perimeter of each high stress area on the water surface of the river embankment, and calculate the compactness according to the area and perimeter of each high stress area on the water surface of the river embankment; S6. Data process the area ratio, proximity and compactness of each high stress area to obtain the river embankment risk coefficient of each high stress area, and obtain the overall risk coefficient of the river embankment by weighting. Compare it with the preset threshold to evaluate the safety condition of the river embankment.

2. The method for evaluating the safety of a river embankment according to claim 1, characterized in that: The inherent parameters of the embankment include the thickness, height and slope of the embankment; the water level feature parameters include the peak water level, maximum rising rate and flood duration; and the feature combination parameters include the thickness, height, slope, peak water level, maximum rising rate and flood duration.

3. The method of claim 2, wherein: The specific steps of S1 are as follows: 1) Obtain the hourly water level data of the same flood season in the past 5 years at the river junction to form a time series data set; 2) Divide the data set into training set, validation set and test set according to the ratio of 7:2:1; 3) Construct a long short-term memory network model, which includes: Input layer: input the water level of m time windows in the time series data set, 1 time window is 1 hour, and m is a natural number greater than or equal to 10; Network structure: build 1-2 layers of long short-term memory network, add a Dropout layer after each layer, and set the dropout rate to 0.2-0.3; Output layer: set 1 neuron to predict the water level of the m+1 time window; 4) Training of long short-term memory network model: Training parameters: including loss function, optimizer, training rounds and batch size, using mean square error as loss function, using Adam as optimizer, setting training rounds to 50-100, and setting batch size to 16-32; cut the training set into multiple input-output pairs according to the window length m: Input: water level of the first m time windows in the time series data set; Output: water level of the m+1 time window; Calculate the predicted value by "forward propagation", and compare it with the true value to get the mean square error; Adjust the model parameters by "back propagation" to reduce the error; Evaluate the effect with the validation set after each round of training, and stop training if the mean square error does not decrease for 5-10 consecutive rounds; Obtain the water level of the future 72 time windows through rolling prediction: Use the water level of the current time window and the previous m-1 time windows as input to predict the water level of the next time window, discard the earliest 1 data in the time window, and add the newly predicted water level value to the end of the sequence to form a new input, continue to predict the water level of the next time window, and repeat the above process until the water level of the future 72 time windows is predicted.

4. The method for evaluating the safety of a river embankment according to claim 3, characterized in that: Extract water level feature parameters according to the water level of the future 72 time windows, and the formula is as follows: H peak = max{H1, H2,..., H t ,...,H 72} wherein H peak is the peak water level, H t is the water level of the tth future time window, t is the index of the future time window, t∈[1,72]; ΔH t = H t - H t-1 where v max is the maximum water level rise rate, AH t is the water level increment of the tth future time window, H t-1 is the water level of the t-1th future time window, and ΔSC is the time window length. Set the water level threshold yz of the flood, based on the water level prediction curve, count the number of windows with water level ≥ water level threshold yz in the 72 future time windows, and accumulate to get the flood duration T.

5. The method of claim 4, wherein: The specific steps of S4 are as follows: The water-facing surface of the river embankment is divided into a plurality of finite element units, stress components of each finite element unit are extracted according to stress distribution of the water-facing surface of the river embankment, including three normal stress components σ x ,σ y ,σ z ,σ max ,σ max ,σ x ,σ y ,σ z ,σ x ,σ y ,σ z are respectively normal stress in x direction, normal stress in y direction, normal stress in z direction; σ max >σ th The finite element is labeled as a high-stress element, where σ th The stress threshold; Merge high stress units by spatial adjacency rules, as follows: 1) Randomly select a high stress unit as the starting center; 2) Define the spatial adjacency rule as: there is at least one common edge between two adjacent high stress units, find the high stress units that meet the spatial adjacency rule with the center unit, and merge them into the same high stress area; 3) Take the newly merged unit as the new center, repeat step 2, and continue to search and merge units that meet the spatial adjacency rule; 4) When there is no high stress unit that meets the spatial adjacency rule, stop iteration and complete the division of a single high stress area; 5) If there are high stress units that have not been divided, repeat steps 1-4 until all high stress units are divided into a unique high stress area, Calculate the area proportion of high stress area, and the formula is as follows: wherein R i is the area ratio of the i-th high stress zone, A i is the area of the i-th high stress zone on the river embankment water-facing surface, A total is the total area of the river embankment water-facing surface, i∈[1,n], and n is the number of high stress zones.

6. The method of claim 5, wherein: According to different proximity distances, the corresponding structural vulnerability coefficients are obtained, and the formula is as follows: wherein β i is the structural vulnerability coefficient of the i-th high stress zone, μ foot is the anti-sliding coefficient of the embankment toe material, μ top is the anti-sliding coefficient of the embankment top material, d min,i is the proximity distance of the i-th high stress zone, d foot,i is the shortest distance between the i-th high stress zone and the embankment toe, d top,i is the shortest distance between the i-th high stress zone and the embankment top, γ1 is the embankment toe adjustment coefficient, γ1 ∈ [1.2, 1.5], γ2 is the embankment top adjustment coefficient, γ2 ∈ [0.8, 1.2]; Calculate the proximity of each high stress area, and the formula is as follows: where d i is the proximity of the ith high stress zone; wherein JC i is the compactness of the i-th high stress zone, P i is the perimeter of the i-th high stress zone on the river embankment water-facing surface, A i is the area of the i-th high stress zone on the river embankment water-facing surface.

7. The method of claim 6, wherein: Process the area proportion, proximity and compactness of each high stress area to generate the levee risk coefficient of each high stress area, and the formula is as follows: RI i = ω1R i + ω2d i + ω3JC i wherein, R1 i is the levee risk factor for the i-th high stress zone; In the formula, ω1 is the weight coefficient of area proportion, ω2 is the weight coefficient of proximity, and ω3 is the weight coefficient of compactness, on the basis of ω1+ω2+ω3=1, let 0<ω3<ω2<ω1<1.

8. The method of claim 7, wherein: Weight the levee risk coefficient of each high stress area to get the overall levee risk coefficient, and the formula is as follows: Where RI is the overall levee risk coefficient; Compare the overall levee risk coefficient with the preset threshold to evaluate the levee safety condition, and the specific steps are as follows: When 0 < RI < Rl th1 , the river embankment is of low risk level; When RI th1 ≤RI≤RI th2 , the river embankment is in a medium risk level; When RI > RI th2 , the river embankment is of high risk level; wherein R1 th1 is a threshold value for dividing the low risk level and the medium risk level, R1 th2 is a threshold value for dividing the medium risk level and the high risk level.

Images