Vertical line sediment transport rate calculation method and system based on support vector machine algorithm

Abstract

The invention discloses a vertical line sediment transport rate calculation method and system based on a support vector machine algorithm, and the method comprises the steps: collecting historical section water and sediment distribution data, and dividing a test set and a verification set based on data representativeness; calculating distribution parameters according to historical section water and sediment distribution data, wherein the distribution parameters comprise linear coefficient item parameters, wake flow function item parameters, bottom concentration item parameters and index distribution item parameters of all vertical lines; based on the test set and the verification set, establishing a support vector machine (SVR) algorithm prediction model, and constructing a prediction relationship between the flow velocity, the sand content and the water depth data of a single point and the distribution parameters; using a support vector machine (SVR) algorithm prediction model to reconstruct section water and sediment distribution data, using the reconstructed section water and sediment distribution data to calculate the vertical line sediment transport rate, and verifying the accuracy of the calculated vertical line sediment transport rate; and calculating the vertical line sediment transport rate on line. According to the method, accurate reconstruction of the vertical line sediment transport rate based on single-point data is realized, and technical support is provided for efficient and real-time hydrological sediment transport monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrological research and river channel management, and specifically to a method and system for calculating the vertical sediment transport rate based on the support vector machine algorithm. Background Art

[0002] The accurate measurement of the vertical sediment transport rate is crucial for analyzing river erosion and deposition changes, predicting riverbed evolution, and evaluating soil erosion. Traditionally, the six-point method, as a precise method for measuring the vertical sediment load, has been widely used. This method samples and analyzes at six specific positions on the vertical line. Although relatively accurate data can be obtained, it requires manual sampling at each point, which is a cumbersome and time-consuming process, and consumes a large amount of human and material resources, making it difficult to meet the requirements of real-time monitoring and rapid analysis.

[0003] With the development of online measurement technology, the measurement of the vertical sediment transport rate is gradually moving towards automation and real-time. However, the core problem has become prominent. Online measurement needs to deduce the vertical sediment transport rate based on the flow velocity and sediment concentration at a single point of online measurement. Existing methods generally rely on direct linear relationship fitting. However, due to the complex non-linear relationship between the vertical flow velocity distribution, sediment concentration distribution and the flow velocity and sediment concentration at a single point, the linear fitting accuracy is poor. This also results in the inability to meet the actual application requirements when using online flow velocity and sediment concentration measurement equipment for hydrological analysis. Therefore, how to accurately deduce the vertical sediment transport rate from the flow velocity and sediment concentration at a single point has become a key problem in improving the effectiveness of online measurement equipment. Summary of the Invention

[0004] To solve the problem in the prior art that the accuracy of deducing the vertical sediment transport rate from the flow velocity and sediment concentration at a single point is poor and it cannot be actually used, the present invention proposes a method and system for calculating the vertical sediment transport rate based on the support vector machine algorithm, which can accurately calculate the vertical sediment transport rate online.

[0005] To achieve the above object, the present invention is implemented through the following technical solutions:

[0006] The method for calculating the vertical sediment transport rate based on the support vector machine algorithm of the present invention includes the following operations:

[0007] Collect historical cross-section water and sediment distribution data, and divide the test set and the validation set based on data representativeness;

[0008] Calculate the distribution parameters according to the historical cross-section water and sediment distribution data, including the linear coefficient term, wake function term, bottom concentration term, and exponential distribution term of each vertical line;

[0009] Based on the test set and the validation set, establish a support vector machine SVR algorithm prediction model, and construct a prediction relationship between the flow velocity, sediment concentration, and water depth data at a single point and the distribution parameters;

[0010] Using the support vector machine SVR algorithm prediction model to reconstruct the water and sediment distribution data of the cross-section, calculating the sediment transport rate per vertical line using the reconstructed water and sediment distribution data of the cross-section, and comparing it with the measured sediment transport rate per vertical line to verify the accuracy of the calculated sediment transport rate per vertical line;

[0011] Online detecting the velocity, sediment concentration and water depth data of a single point, and calculating the sediment transport rate per vertical line based on the support vector machine SVR algorithm prediction model.

[0012] A further improvement of the present invention lies in: collecting the historical water and sediment distribution data of the cross-section, and dividing the test set and the verification set based on data representativeness, including: collecting the historical water and sediment distribution data of the cross-section, dividing the historical water and sediment distribution data of the cross-section into the water and sediment distribution data of the flood season and the water and sediment distribution data of the dry season, uniformly sampling the water and sediment distribution data of the flood season and the water and sediment distribution data of the dry season according to the water depth of each vertical line of the cross-section to obtain typical data, selecting 70% of the typical data as the test set, and the remaining 30% of the typical data as the verification set, where: the historical water and sediment distribution data of the cross-section includes the vertical line velocity distribution, sediment concentration distribution, and water depth data of the cross-section, where the water depth data is the water depth of each vertical line of the cross-section, the sediment concentration distribution is the six-point sediment concentration of each vertical line, the vertical line velocity distribution of the cross-section is the six-point velocity, and the six-point sediment concentration and the six-point velocity are the sediment concentration and velocity at the surface layer, relative water depths of 0.2H, 0.4H, 0.6H, 0.8H, and the bottom layer, where H is the water depth of the vertical line.

[0013] A further improvement of the present invention lies in: calculating the distribution parameters according to the historical water and sediment distribution data of the cross-section, specifically including:

[0014] According to the vertical line velocity distribution of the cross-section, calculating the linear coefficient term and the wake function term by linear fitting, and the expression is:

[0015] ;

[0016] ;

[0017] ;

[0018] where, is the friction velocity, is the wake function term, is the linear coefficient term, is the water depth, is the vertical height, that is, the distance from the bed surface, is the constant term, is the velocity;

[0019] According to the sediment concentration distribution and water depth data, calculating the bottom concentration term and the exponential distribution term, and the expression is:

[0020] ;

[0021] ;

[0022] ;

[0023] The relationship existing in the bottom concentration when sediment is in vertical equilibrium distribution is expressed as:

[0024] ;

[0025] Among them, is the concentration of suspended sediment at a set position, that is, the sediment concentration, is the bottom concentration term, is the negative value of the ratio of sediment settling velocity to sediment diffusion coefficient, that is, the exponential distribution term, is the settling velocity of sediment particles, is the reference point, is the coefficient of flow turbulence diffusion or energy dissipation, is the height near the water surface or the bed surface, is the natural constant.

[0026] A further improvement of the present invention lies in: based on the test set and the validation set, establishing a support vector machine SVR algorithm prediction model, and constructing the prediction relationship between the flow velocity, sediment concentration and water depth data of a single point and the distribution parameters, specifically including: According to the divided test set, determine the independent variables, including the flow velocity, sediment concentration and water depth at 0.6H of each vertical line in the test set; Determine the dependent variables, including the linear coefficient term, wake function term, bottom concentration term and exponential distribution term calculated at 0.6H of each vertical line in the test set; Establish a support vector machine SVR algorithm prediction model, substitute the independent variables and the dependent variables to train the support vector machine SVR algorithm prediction model, and obtain the model parameters; Use the validation set to perform the verification calculation of the support vector machine SVR algorithm prediction model; Based on the verification calculation results, perform the performance evaluation of the support vector machine SVR algorithm prediction model.

[0027] A further improvement of the present invention lies in: using the support vector machine SVR algorithm prediction model to reconstruct the cross-section water and sediment distribution data, calculating the vertical sediment transport rate using the reconstructed cross-section water and sediment distribution data, and comparing it with the measured vertical sediment transport rate to verify the accuracy of the calculated vertical sediment transport rate, specifically including:

[0028] According to the measured cross-section water and sediment distribution data, calculate the vertical average flow velocity and the vertical average sediment concentration by the following formula, and calculate the measured vertical sediment transport rate according to the vertical average flow velocity and the vertical average sediment concentration;

[0029] ;

[0030] ;

[0031] The measured sediment transport rate per vertical line is as follows:

[0032] ;

[0033] In the formula: is the average velocity per vertical line, is the velocity at position on the vertical line, is the relative water depth, with values of 0, 0.2H, 0.4H, 0.6H, 0.8H, 1H, where H is the water depth of the vertical line, , , , , , are the velocities at relative water depths of 0, 0.2H, 0.4H, 0.6H, 0.8H, 1H respectively, is the average sediment concentration per vertical line, is the sediment concentration at position on the vertical line, is the layering weight coefficient. For the surface layer and the bottom layer, takes 1, and for the remaining intermediate layers, takes 2, is the point position;

[0034] According to the support vector machine SVR algorithm prediction model, using the velocity, sediment concentration, and water depth data at 0.6H, calculate the distribution parameters;

[0035] Based on the calculated distribution parameters, reconstruct the vertical velocity distribution and sediment concentration distribution of the cross-section, determine the six-point sediment concentration and six-point velocity, calculate and determine the average sediment concentration per vertical line and the average velocity per vertical line through vertical integration, and obtain the reconstructed sediment transport rate per vertical line from the average sediment concentration per vertical line and the average velocity per vertical line;

[0036] Calculate the root mean square error and Nash-Sutcliffe efficiency coefficient of the difference between the reconstructed sediment transport rate per vertical line and the measured sediment transport rate per vertical line to evaluate the accuracy of the support vector machine SVR algorithm prediction model.

[0037] The vertical sediment transport rate calculation system based on the support vector machine algorithm of the present invention includes:

[0038] A collection module for collecting historical cross-section water and sediment distribution data and dividing the test set and validation set based on data representativeness;

[0039] A parameter calculation module for calculating distribution parameters according to historical cross-section water and sediment distribution data, including the linear coefficient term, wake function term, bottom concentration term, and exponential distribution term of each vertical line;

[0040] A model construction module, which is used to establish a prediction model of the support vector machine SVR algorithm based on a test set and a validation set, and construct a prediction relationship between the flow velocity, sediment concentration, and water depth data of a single point and the distribution parameters.

[0041] A model verification module, which is used to reconstruct the cross-section water and sediment distribution data by using the support vector machine SVR algorithm prediction model, calculate the vertical sediment transport rate by using the reconstructed cross-section water and sediment distribution data, and compare it with the measured vertical sediment transport rate to verify the accuracy of the calculated vertical sediment transport rate.

[0042] An online calculation module, which is used to online detect the flow velocity, sediment concentration, and water depth data of a single point, and calculate the vertical sediment transport rate based on the support vector machine SVR algorithm prediction model.

[0043] A further improvement of the present invention lies in that the operations performed by the acquisition module include: collecting historical cross-section water and sediment distribution data, dividing the historical cross-section water and sediment distribution data into two groups of sequences, namely flood-season cross-section water and sediment distribution data and dry-season cross-section water and sediment distribution data, uniformly sampling the flood-season cross-section water and sediment distribution data and the dry-season cross-section water and sediment distribution data according to the water depth of each vertical line of the cross-section to obtain typical data, selecting 70% of the typical data as the test set, and the remaining 30% of the typical data as the validation set, where: the historical cross-section water and sediment distribution data includes the cross-section vertical line flow velocity distribution, sediment concentration distribution, and water depth data, where the water depth data is the water depth of each vertical line of the cross-section, the sediment concentration distribution is the six-point sediment concentration of each vertical line, the cross-section vertical line flow velocity distribution is the six-point flow velocity, and the six-point sediment concentration and the six-point flow velocity are the sediment concentration and flow velocity at the surface layer, relative water depths of 0.2H, 0.4H, 0.6H, 0.8H, and the bottom layer, where H is the water depth of the vertical line.

[0044] A further improvement of the present invention lies in that the operations performed by the parameter calculation module specifically include:

[0045] According to the cross-section vertical line flow velocity distribution, calculate the linear coefficient term and the wake function term by using linear fitting, and the expression is:

[0046] ;

[0047] ;

[0048] ;

[0049] Among them, is the friction velocity, is the wake function term, is the linear coefficient term, is the water depth, is the vertical height, that is, the distance from the bed surface, is the constant term, is the flow velocity;

[0050] According to the sediment concentration distribution and water depth data, calculate the bottom concentration term and the exponential distribution term. The expressions are as follows:

[0051] ;

[0052] ;

[0053] ;

[0054] The relationship existing in the bottom concentration when the sediment is in vertical equilibrium distribution is expressed as:

[0055] ;

[0056] Among them, is the suspended sediment concentration at the set position, that is, the sediment concentration, is the parameter of the bottom concentration term, is the negative value of the ratio of the sediment settling velocity to the sediment diffusion coefficient, that is, the exponential distribution term, is the settling velocity of the sediment particles, is the reference point, is the coefficient of water flow turbulence diffusion or energy dissipation, is the height near the water surface or the bed surface, is the natural constant.

[0057] A further improvement of the present invention lies in that the specific operations performed by the model construction module include: According to the divided test set, determine the independent variables, including the flow velocity, sediment concentration, and water depth at 0.6H of each vertical line in the test set; Determine the dependent variables, including the calculated linear coefficient term, wake function term, bottom concentration term, and exponential distribution term at 0.6H of each vertical line in the test set; Establish a support vector machine SVR algorithm prediction model, substitute the independent variables and the dependent variables to train the support vector machine SVR algorithm prediction model, and obtain the model parameters; Use the validation set to perform the verification calculation of the support vector machine SVR algorithm prediction model; Based on the verification calculation results, perform the performance evaluation of the support vector machine SVR algorithm prediction model.

[0058] A further improvement of the present invention lies in that the specific operations performed by the model verification module include:

[0059] According to the measured cross-sectional water and sediment distribution data, calculate the vertical average flow velocity and the vertical average sediment concentration by the following formula, and calculate the measured vertical sediment transport rate according to the vertical average flow velocity and the vertical average sediment concentration;

[0060] ;

[0061] ;

[0062] The measured sediment transport rate per vertical line is:

[0063] ;

[0064] In the formula: is the average velocity of the vertical line, is the velocity at the position on the vertical line, is the relative water depth, with values of 0, 0.2H, 0.4H, 0.6H, 0.8H, 1H, where H is the water depth of the vertical line, , , , , , are the velocities at the relative water depths of 0, 0.2H, 0.4H, 0.6H, 0.8H, 1H respectively, is the average sediment concentration of the vertical line, is the sediment concentration at the position on the vertical line, is the layer weighting coefficient. For the surface layer and the bottom layer, takes 1, and for the remaining intermediate layers, takes 2, is the point position;

[0065] According to the support vector machine SVR algorithm prediction model, using the velocity, sediment concentration, and water depth data at 0.6H, calculate the distribution parameters;

[0066] Based on the calculated distribution parameters, reconstruct the vertical velocity distribution and sediment concentration distribution of the cross-section, determine the six-point sediment concentration and six-point velocity, calculate the average sediment concentration and average velocity of the vertical line through vertical integration, and obtain the reconstructed sediment transport rate per vertical line from the average sediment concentration and average velocity of the vertical line;

[0067] Calculate the root mean square error and Nash-Sutcliffe efficiency coefficient of the difference between the reconstructed sediment transport rate per vertical line and the measured sediment transport rate per vertical line to evaluate the accuracy of the support vector machine SVR algorithm prediction model.

[0068] The beneficial effects of the present invention are as follows: By establishing a support vector machine SVR algorithm prediction model, the present invention fits and optimizes the relevant distribution parameters, determines the optimal measurement points based on the distribution parameter fitting, and reconstructs the cross-section water and sediment distribution data, realizing the accurate calculation of the sediment transport rate per vertical line. The present invention solves the key problem of the association between vertical point selection and sediment transport rate in on-line measurement, realizes the accurate reconstruction of the sediment transport rate per vertical line based on single-point data, and provides technical support for efficient and real-time hydrological sediment monitoring. Description of the Drawings

[0069] Figure 1 It is a schematic flow chart of the method in the embodiment of the present invention; Figure 2 It is a schematic diagram for comparing the actual value and the predicted value of the linear coefficient term in the embodiment of the present invention; Figure 3 It is a comparison chart of the reconstructed vertical sediment transport rate and the measured vertical sediment transport rate in the embodiment of the present invention. Detailed implementation manners

[0070] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0071] As Figure 1 shown, the vertical sediment transport rate calculation method based on the support vector machine algorithm in this embodiment includes the following steps:

[0072] Step 1: Collect historical cross-section water and sediment distribution data, and divide the test set and the validation set based on data representativeness. The historical cross-section water and sediment distribution data includes cross-section vertical velocity distribution, sediment concentration distribution, and water depth data. Among them, the water depth data is the water depth of each vertical line in the cross-section, the sediment concentration distribution is the six-point sediment concentration of each vertical line, and the cross-section vertical velocity distribution is the six-point velocity. The six-point sediment concentration and the six-point velocity are the sediment concentration and velocity at the surface layer, relative water depths of 0.2H, 0.4H, 0.6H, 0.8H, and the bottom layer, where H is the vertical water depth.

[0073] Divide the historical cross-section water and sediment distribution data into two groups of sequences: flood season cross-section water and sediment distribution data and dry season cross-section water and sediment distribution data. Uniform sampling is performed according to the water depth of each vertical line in the cross-section to obtain typical data. 70% of the typical data is selected as the test set, and the remaining 30% of the typical data is used as the validation set. Tables 1 and 2 are the cross-section water and sediment distribution data tables for 2018 and 2019.

[0074] Table 1: Cross-section water and sediment distribution data table for 2018

[0075] Table 2: Cross-section water and sediment distribution data table for 2019

[0076] The samples selected in this embodiment are shown in Tables 3 and 4:

[0077] Table 3: Sampling data in the dry season

[0078] Table 4: Sampling data in the high-water season

[0079] Step 2: Calculate the distribution parameters according to the historical cross-section water and sediment distribution data, including the linear coefficient term, wake function term, bottom concentration term, and exponential distribution term for each vertical line. The specific steps are as follows:

[0080] Step 2.1: Determine the linear coefficient term and the wake function term. Include the calculation expression for the vertical velocity distribution of the cross-section in logarithmic form:

[0081] ;

[0082] Among them, is the friction velocity, is the flow velocity, is the equivalent roughness height of the bed surface, with a value of 0.01 times the water depth, i.e., 0.01h, is the water depth, is the vertical height, i.e., the distance from the bed surface, is the constant term.

[0083] After conversion:

[0084] ;

[0085] ;

[0086] ;

[0087] Among them, is the wake function term, is the linear coefficient term.

[0088] Step 2.2: Determine the bottom concentration term and the exponential distribution term.

[0089] ;

[0090] ;

[0091] ;

[0092] The relationship of the bottom concentration when the sediment is in vertical equilibrium distribution is expressed as:

[0093] ;

[0094] Among them, is the concentration of suspended sediment at the set position, i.e., the sediment concentration, is the bottom concentration term, is the negative value of the ratio of the sediment settling velocity to the sediment diffusion coefficient, i.e., the exponential distribution term, is the settling velocity of sediment particles, is the reference point, is the coefficient of water flow turbulence diffusion or energy dissipation, is the height near the water surface or the bed surface, is the natural constant.

[0095] Step 2.3: Calculate the linear coefficient term, wake function term, bottom concentration term, and exponential distribution term for each vertical line according to the historical cross-section water and sediment distribution data.

[0096] Obtain the corresponding linear coefficient term and wake function term based on the vertical line velocity distribution of the cross-section:

[0097] Substitute the velocity data of the six points of the surface layer, 0.2H, 0.4H, 0.6H, 0.8H, and the bottom layer of the vertical line into respectively, and use linear fitting to obtain the linear coefficient term and the wake function term of each vertical line of the cross-section.

[0098] Substitute the sediment concentrations of the six points of the surface layer, 0.2H, 0.4H, 0.6H, 0.8H, and the bottom layer of the vertical line into respectively, and use linear fitting to obtain the bottom concentration term and the exponential distribution term of each vertical line of the cross-section.

[0099] Step 3: Based on the test set and the validation set, establish a support vector machine SVR algorithm prediction model to construct the prediction relationship between the velocity, sediment concentration, and water depth data of a single point and the distribution parameters. The specific steps are as follows:

[0100] Step 3.1: Select the representative point on the vertical line. The representative point on the vertical line can be taken at the middle position of the vertical line water depth. In this embodiment, the point at 0.6H water depth is selected, and the velocity, sediment concentration, and water depth values at 0.6H are used as independent variables and substituted into the support vector machine SVR algorithm for fitting calculation.

[0101] Step 3.2: Use the support vector machine SVR algorithm for small sample machine learning to establish a prediction relationship between the velocity and sediment concentration at 0.6H and the linear coefficient term parameter, wake function term parameter, bottom concentration term parameter, and exponential distribution term. In actual processing, mainly adjust the three parameters of the regularization parameter (BoxConstraint), kernel function scaling factor (KernelScale), and maximum tolerance error (Epsilon).

[0102] Given the training data set , is the independent variable, is the dependent variable. In the present invention, includes water depth, sediment concentration, and flow velocity, includes the linear coefficient term, wake function term, bottom concentration term, and exponential distribution term of each vertical line.

[0103] The goal of the support vector machine SVR algorithm prediction model is to find the function , where is the sample, is the weight vector, is the bias term, satisfying:

[0104] · The deviation between the predicted value and the true value does not exceed , is the maximum tolerance error.

[0105] · Through the slack variables and to process samples that exceed .

[0106] Objective function and constraints:

[0107] ;

[0108] ;

[0109] where: is the weight vector, controlling the direction of the regression hyperplane, is the regularization parameter, balancing the weights of the model complexity and training error, , are the slack variables, allowing samples to exceed the interval band ( is the maximum tolerance error).

[0110] Introduce Lagrange multipliers , (corresponding to the constraint conditions) and , (corresponding to the non-negativity of the slack variables) to construct the Lagrangian function: ;

[0111] where: is the sample, that is, the independent variable, is the dependent variable, is the sample ordinal number, is the bias term, is the total number of samples.

[0112] Substitute the above conditions into the Lagrangian function to obtain the dual problem:

[0113] ;

[0114] Wherein: , , , are Lagrange multipliers, is a sample, is the sample ordinal number.

[0115] Constraint conditions:

[0116] ;

[0117] Wherein: is a kernel function for handling non - linear regression. Only when , for the corresponding sample is a support vector.

[0118] The final regression function is:

[0119] ;

[0120] Where the bias term is calculated through support vectors:

[0121] 1. Select samples or that satisfy. .

[0122] 2. Use to solve for .

[0123] The support vector machine SVR algorithm prediction model can flexibly handle linear and non - linear regression problems while ensuring the generalization ability of the model.

[0124] According to the divided test set, the independent variables are the flow velocity, sediment concentration, and water depth at 0.6H of each vertical line in the test set, and the dependent variables are the linear coefficient term parameter, wake function term parameter, bottom concentration term parameter, and exponential distribution term calculated at 0.6H of each vertical line. A support vector machine SVR prediction model is established. Substitute the independent variables and dependent variables to train the support vector machine SVR algorithm prediction model to obtain the regularization parameter , the kernel function scaling factor K, and the maximum tolerance error of the three parameters. Use the validation set to perform the verification calculation of the support vector machine SVR algorithm prediction model. Based on the verification calculation results, calculate the performance indicators of the validation set and the test set results, including the correlation index R 2 and the mean square error index MSE.

[0125] The calculation expression of R 2 is:

[0126] ;

[0127] Where: is the sum of squared residuals, representing the deviation between the model predicted value and the true value; is the total sum of squares, representing the total variability of the true values relative to their mean; is the mean of the true values.

[0128] The MSE formula is:

[0129] ;

[0130] Where: represents the true value of the -th sample, represents the predicted value of the -th sample, and N represents the total number of samples.

[0131] If the correlation is less than the set value or the mean squared error is greater than the set value, then adjust the regularization parameter , the kernel function scaling factor K, and the maximum tolerance error and recalculate after taking their values until the correlation index R 2 and the mean squared error index MSE meet the requirements.

[0132] As Figure 2 shown, taking the linear coefficient term as an example, the optimal parameters are obtained: , , .

[0133] Step 4, use the support vector machine SVR algorithm to predict the reconstructed cross-section water and sediment distribution data, calculate the vertical sediment transport rate using the reconstructed cross-section water and sediment distribution data, and compare it with the measured vertical sediment transport rate to verify the accuracy of the calculated vertical sediment transport rate.

[0134] Step 4.1, calculate the vertical average velocity and the vertical average sediment concentration based on the measured cross-section water and sediment distribution data, and calculate the measured vertical sediment transport rate based on the vertical average velocity and the vertical average sediment concentration;

[0135] ;

[0136] ;

[0137] The measured vertical sediment transport rate is:

[0138] ;

[0139] In the formula: is the vertical average velocity, is the flow velocity at a position on the vertical line , is the relative water depth, taking values of 0, 0.2H, 0.4H, 0.6H, 0.8H, 1H, where 0 represents the surface layer, 1H represents the bottom layer, and H is the vertical line water depth , , , , , are the flow velocities at relative water depths of 0, 0.2H, 0.4H, 0.6H, 0.8H, 1H respectively is the average sediment concentration of the vertical line is the position on the vertical line where the sediment concentration is is the layering weight coefficient. For the surface layer and the bottom layer take 1. For the remaining intermediate layers take 2 is the point position, taking values from 1 to 6, corresponding to the relative water depths: 0, 0.2H, 0.4H, 0.6H, 0.8H, 1H

[0140] Step 4.2, according to the support vector machine SVR algorithm prediction model, using the flow velocity, sediment concentration and water depth data at 0.6H, calculate the linear coefficient term , the wake function term , the bottom concentration term and the exponential distribution term .

[0141] According to Step 2.1 and Step 2.2, reconstruct the vertical line flow velocity distribution and sediment concentration distribution, determine the flow velocity and sediment concentration at each point of the vertical line water depth, calculate and determine the average sediment concentration and average flow velocity of the vertical line, and obtain the reconstructed vertical line sediment transport rate from the average sediment concentration and average flow velocity of the vertical line

[0142] Reconstruction results of the flow velocity and sediment concentration distributions along the water depth direction

[0143] ;

[0144] ;

[0145] The reconstructed average flow velocity of the vertical line and the average sediment concentration of the vertical line are

[0146] ;

[0147] ;

[0148] The reconstructed vertical line sediment transport rate is

[0149] 。

[0150] Step 4.3. Verify the accuracy of the support vector machine SVR algorithm prediction model according to the reconstructed vertical sediment transport rate. This includes: calculating the root mean square error (RRMSE) and Nash-Sutcliffe efficiency coefficient (NSE) of the difference between the reconstructed vertical sediment transport rate and the measured vertical sediment transport rate, and evaluating the accuracy of the support vector machine SVR algorithm prediction model.

[0151] The calculation expression of the root mean square error (RMSE) is:

[0152] ;

[0153] where N is the total number of samples, represents the true value of the th sample, that is, the measured vertical sediment transport rate, represents the predicted value of the th sample, that is, the reconstructed vertical sediment transport rate.

[0154] ;

[0155] where is the mean of the true values, that is, the mean of the measured vertical sediment transport rates.

[0156] The calculation expression of the Nash-Sutcliffe efficiency coefficient (NSE) is:

[0157] ;

[0158] where is the mean of the true values, that is, the mean of the measured vertical sediment transport rates.

[0159] The calculation results of the sediment transport rate (reconstructed vertical sediment transport rate) and the verification sediment transport rate (measured vertical sediment transport rate) are as follows in the table: Table 5: Calculation results of reconstructed vertical and verification sediment transport rates

[0160] The calculation results of MSE and NSE are shown in Figure 3, in the figure, the abscissa of each data point represents the calculated sediment transport rate, and the ordinate represents the verified sediment transport rate. When the data points basically fall on the 45° line, it means that the calculated sediment transport rate is equal to the verified sediment transport rate. When the data points deviate and fall below the 45° line, it means that the calculated sediment transport rate is larger than the verified sediment transport rate. When the data points deviate and fall above the 45° line, it means that the calculated sediment transport rate is smaller than the verified sediment transport rate. When RRMSE < 0.3 and NSE > 0.75, it is evaluated that the prediction model accuracy of the support vector machine SVR algorithm is relatively high. The calculation results of this embodiment are: MSE = 0.7955, RRMSE = 0.23, NSE = 0.954.

[0161] Step 5, online detect the flow velocity, sediment concentration and water depth data of a single point, and calculate the vertical sediment transport rate based on the prediction model of the support vector machine SVR algorithm.

[0162] Those skilled in the art of this technology can understand that, unless otherwise defined, all terms (including technical terms and scientific terms) used here have the same meaning as the general understanding of those of ordinary skill in the art to which the present invention belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless defined as here.

[0163] The specific embodiments described above have further elaborated on the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only the specific embodiments of the present invention and is not used to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for calculating the vertical sediment transport rate based on the support vector machine algorithm, characterized in that: Including the following operations: Collect historical cross-section water and sediment distribution data, and divide the test set and validation set based on data representativeness; Calculate distribution parameters according to the historical cross-section water and sediment distribution data, including the linear coefficient term, wake function term, bottom concentration term, and exponential distribution term for each vertical line; Based on the test set and validation set, establish a support vector machine SVR algorithm prediction model, and construct the prediction relationship between the single-point flow velocity, sediment concentration, and water depth data and the distribution parameters; Use the support vector machine SVR algorithm prediction model to reconstruct the cross-section water and sediment distribution data, calculate the vertical sediment transport rate using the reconstructed cross-section water and sediment distribution data, and compare it with the measured vertical sediment transport rate to verify the accuracy of the calculated vertical sediment transport rate; Online detect the single-point flow velocity, sediment concentration, and water depth data, and calculate the vertical sediment transport rate based on the support vector machine SVR algorithm prediction model.

2. The method for calculating the vertical sediment transport rate based on the support vector machine algorithm according to claim 1, characterized in that: The collection of historical cross-section water and sediment distribution data and the division of the test set and validation set based on data representativeness include: collecting historical cross-section water and sediment distribution data, dividing the historical cross-section water and sediment distribution data into flood-season cross-section water and sediment distribution data and dry-season cross-section water and sediment distribution data, uniformly sampling the flood-season cross-section water and sediment distribution data and dry-season cross-section water and sediment distribution data according to the water depth of each vertical line of the cross-section to obtain typical data, selecting 70% of the typical data as the test set, and the remaining 30% of the typical data as the validation set, where: the historical cross-section water and sediment distribution data include the cross-section vertical line flow velocity distribution, sediment concentration distribution, and water depth data, where the water depth data is the water depth of each vertical line of the cross-section, the sediment concentration distribution is the six-point sediment concentration of each vertical line, the cross-section vertical line flow velocity distribution is the six-point flow velocity, and the six-point sediment concentration and six-point flow velocity are the sediment concentration and flow velocity at the surface layer, relative water depths of 0.2H, 0.4H, 0.6H, 0.8H, and the bottom layer, where H is the water depth of the vertical line.

3. The method for calculating the vertical sediment transport rate based on the support vector machine algorithm according to claim 2, characterized in that: The calculation of distribution parameters according to the historical cross-section water and sediment distribution data specifically includes: According to the cross-section vertical line flow velocity distribution, use linear fitting to calculate the linear coefficient term and wake function term, and the expression is: ； ； ； Among them, is the friction velocity, is the wake function term, is the linear coefficient term, is the water depth, is the vertical height, i.e., the distance from the bed surface, is the constant term, is the flow velocity; According to the sediment concentration distribution and water depth data, calculate the bottom concentration term and exponential distribution term, and the expression is: ； ； ； The relationship existing in the bottom concentration when the sediment is in vertical equilibrium distribution is expressed as: ； wherein, is the suspended sediment concentration at a set position, i.e., the sediment concentration, is the bottom concentration term, is the negative value of the ratio of the sediment settling velocity to the sediment diffusion coefficient, i.e., the exponential distribution term, is the settling velocity of sediment particles, is the reference point, is the coefficient of water flow turbulent diffusion or energy dissipation, is the height near the water surface or the bed surface, is the natural constant.

4. The method for calculating the vertical sediment transport rate based on the support vector machine algorithm according to claim 2, wherein: Based on the test set and validation set, establish a support vector machine SVR algorithm prediction model, and construct the prediction relationship between the single-point flow velocity, sediment concentration, and water depth data and the distribution parameters. Specifically include: According to the divided test set, determine the independent variables, including the flow velocity, sediment concentration, and water depth at 0.6H of each vertical line in the test set; Determine the dependent variables, including the linear coefficient term, wake function term, bottom concentration term, and exponential distribution term calculated at 0.6H of each vertical line in the test set; Establish a support vector machine SVR algorithm prediction model, substitute the independent variables and dependent variables to train the support vector machine SVR algorithm prediction model, and obtain the model parameters; Use the validation set to perform the verification calculation of the support vector machine SVR algorithm prediction model; Based on the verification calculation results, perform the performance evaluation of the support vector machine SVR algorithm prediction model.

5. The method for calculating the vertical sediment transport rate based on the support vector machine algorithm according to claim 3, characterized in that: Use the support vector machine SVR algorithm prediction model to reconstruct the water and sediment distribution data of the cross-section. Calculate the sediment transport rate per vertical line using the reconstructed water and sediment distribution data of the cross-section, and compare it with the measured sediment transport rate per vertical line to verify the accuracy of the calculated sediment transport rate per vertical line. Specifically, it includes: According to the measured water and sediment distribution data of the cross-section, calculate the average velocity per vertical line and the average sediment concentration per vertical line using the following formula. Calculate the measured sediment transport rate per vertical line based on the average velocity per vertical line and the average sediment concentration per vertical line; ； ； The measured sediment transport rate per vertical line is: ； In the formula: is the average velocity along the vertical line, is the velocity at the position along the vertical line, is the relative water depth, with values of 0, 0.2H, 0.4H, 0.6H, 0.8H, 1H, where H is the water depth of the vertical line, , , , , , are the velocities at relative water depths of 0, 0.2H, 0.4H, 0.6H, 0.8H, 1H respectively, is the average sediment concentration along the vertical line, is the sediment concentration at the position along the vertical line, is the stratification weight coefficient. For the surface layer and the bottom layer, takes 1, and for the remaining intermediate layers, takes 2, is the point position; According to the support vector machine SVR algorithm prediction model, calculate the distribution parameters using the velocity, sediment concentration, and water depth data at 0.6H; Based on the calculated distribution parameters, reconstruct the vertical velocity distribution and sediment concentration distribution of the cross-section, determine the six-point sediment concentration and six-point velocity. Calculate the average sediment concentration per vertical line and the average velocity per vertical line through vertical integration. Obtain the reconstructed sediment transport rate per vertical line from the average sediment concentration per vertical line and the average velocity per vertical line; Calculate the root mean square error and Nash-Sutcliffe efficiency coefficient of the difference between the reconstructed sediment transport rate per vertical line and the measured sediment transport rate per vertical line to evaluate the accuracy of the support vector machine SVR algorithm prediction model.

6. A vertical sediment transport rate calculation system based on the support vector machine algorithm, characterized in that: It includes: A collection module for collecting historical water and sediment distribution data of the cross-section and dividing the test set and validation set based on data representativeness; A parameter calculation module for calculating distribution parameters according to historical water and sediment distribution data of the cross-section, including the linear coefficient term, wake function term, bottom concentration term, and exponential distribution term of each vertical line; A model construction module for establishing a support vector machine SVR algorithm prediction model based on the test set and validation set, and constructing the prediction relationship between the velocity, sediment concentration, and water depth data of a single point and the distribution parameters; A model verification module for using the support vector machine SVR algorithm prediction model to reconstruct the water and sediment distribution data of the cross-section, calculating the sediment transport rate per vertical line using the reconstructed water and sediment distribution data of the cross-section, and comparing it with the measured sediment transport rate per vertical line to verify the accuracy of the calculated sediment transport rate per vertical line; An online calculation module for online detecting the velocity, sediment concentration, and water depth data of a single point, and calculating the sediment transport rate per vertical line based on the support vector machine SVR algorithm prediction model.

7. The vertical sediment transport rate calculation system based on the support vector machine algorithm according to claim 6, characterized in that: The operations performed by the collection module include: collecting historical water and sediment distribution data of the cross-section, dividing the historical water and sediment distribution data into two groups of sequences: flood season water and sediment distribution data and dry season water and sediment distribution data. Uniformly sample the flood season water and sediment distribution data and dry season water and sediment distribution data according to the water depth of each vertical line of the cross-section to obtain typical data. Select 70% of the typical data as the test set, and the remaining 30% of the typical data as the validation set. Among them: the historical water and sediment distribution data of the cross-section includes the vertical velocity distribution of the cross-section, sediment concentration distribution, and water depth data. Among them, the water depth data is the water depth of each vertical line of the cross-section, the sediment concentration distribution is the six-point sediment concentration of each vertical line, the vertical velocity distribution of the cross-section is the six-point velocity, and the six-point sediment concentration and six-point velocity are the sediment concentration and velocity at the surface layer, relative water depths of 0.2H, 0.4H, 0.6H, 0.8H, and the bottom layer, where H is the water depth of the vertical line.

8. The vertical sediment transport rate calculation system based on the support vector machine algorithm according to claim 7, characterized in that: The operations performed by the parameter calculation module specifically include: According to the vertical velocity distribution of the cross-section, calculate the linear coefficient term and wake function term using linear fitting. The expression is: ； ； ； Among them, is the friction velocity, is the wake function term, is the linear coefficient term, is the water depth, is the vertical height, that is, the distance from the bed surface, is the constant term, is the flow velocity; Calculate the bottom concentration term and the exponential distribution term according to the sediment concentration distribution and water depth data. The expression is as follows: ； ； ； The relationship of the bottom concentration when the sediment is in vertical equilibrium distribution is expressed as: ； wherein, is the suspended sediment concentration at the set position, i.e., the sediment concentration, is the bottom concentration term, is the negative value of the ratio of the sediment settling velocity to the sediment diffusion coefficient, i.e., the exponential distribution term, is the settling velocity of sediment particles, is the reference point, is the coefficient of water flow turbulent diffusion or energy dissipation, is the height near the water surface or the bed surface, is the natural constant.

9. The vertical sediment transport rate calculation system based on the support vector machine algorithm according to claim 7, characterized in that: The specific operations performed by the model construction module include: According to the divided test set, determine the independent variables, including the flow velocity, sediment concentration, and water depth at 0.6H of each vertical line in the test set; Determine the dependent variables, including the linear coefficient term, wake function term, bottom concentration term, and exponential distribution term calculated at 0.6H of each vertical line in the test set; Establish a support vector machine SVR algorithm prediction model, substitute the independent variables and dependent variables to train the support vector machine SVR algorithm prediction model, and obtain the model parameters; Use the validation set to perform the verification calculation of the support vector machine SVR algorithm prediction model; Based on the verification calculation results, evaluate the performance of the support vector machine SVR algorithm prediction model.

10. The vertical sediment transport rate calculation system based on the support vector machine algorithm according to claim 8, characterized in that: The specific operations performed by the model verification module include: According to the measured cross-section water and sediment distribution data, calculate the vertical average flow velocity and vertical average sediment concentration by the following formula, and calculate the measured vertical sediment transport rate according to the vertical average flow velocity and vertical average sediment concentration; ； ； The measured vertical sediment transport rate is: ； Wherein: is the average velocity of the vertical line; is the velocity at the position on the vertical line; is the relative water depth, taking values of 0, 0.2H, 0.4H, 0.6H, 0.8H, 1H, where H is the water depth of the vertical line; , , , , , are the velocities at relative water depths of 0, 0.2H, 0.4H, 0.6H, 0.8H, 1H respectively; is the average sediment concentration of the vertical line, and is the sediment concentration at the position on the vertical line; is the stratification weight coefficient. For the surface layer and the bottom layer, takes 1, and for the remaining intermediate layers, takes 2; is the point position; According to the support vector machine SVR algorithm prediction model, calculate the distribution parameters with the flow velocity, sediment concentration, and water depth data at 0.6H; Based on the calculated distribution parameters, reconstruct the vertical velocity distribution and sediment concentration distribution of the cross-section, determine the six-point sediment concentration and six-point flow velocity, calculate the vertical average sediment concentration and vertical average flow velocity through vertical integration, and obtain the reconstructed vertical sediment transport rate from the vertical average sediment concentration and vertical average flow velocity; Calculate the root mean square error and Nash-Sutcliffe efficiency coefficient of the difference between the reconstructed vertical sediment transport rate and the measured vertical sediment transport rate, and evaluate the accuracy of the support vector machine SVR algorithm prediction model.

Images