A Method for Evaluating the Stability of Wandering Riverbeds Based on Multi-Source Data Fusion

Abstract

This invention discloses a method for evaluating the stability of wandering riverbeds based on multi-source data fusion. It includes the following steps: Step 1: Data collection; Step 2: Calculation of flow and sediment concentration at various monitoring sections of the river segment during the flood season; Step 3: Calculation of the average flood season scour intensity parameter for each sub-segment in year k; Step 4: Determination of the river width and depth at the river segment scale; Step 5: Calculation of the river segment scale facies coefficient parameter for each sub-segment in year k; Step 6: Determination of the multi-year channel oscillation discrimination index (CMI) value for each sub-segment; Step 7: Classification of riverbed stability based on the CMI parameter; Step 8: Evaluation of the riverbed stability of each sub-segment in the study river segment using satellite remote sensing imagery, resulting in a riverbed stability evaluation map of the study river segment. This invention has the advantage of quantitatively analyzing the impact of changes in natural conditions and strong human activities on the stability of wandering riverbeds.

Description

Technical Field

[0001] This invention relates to the field of water conservancy and hydropower engineering technology, specifically to a method for evaluating the stability of wandering riverbeds based on multi-source data fusion. Background Technology

[0002] Wandering rivers have unique geomorphological characteristics and complex water and sediment transport and riverbed evolution, such as unstable main channel, drastic changes in river morphology, and large scouring and deposition. Wandering rivers are widely found around the world, such as the Red River in North America, the Segundo River in South America, the Tana River in Europe, and the Brahmaputra River in South Asia. Typical wandering rivers in my country include the Danjiangkou-Zhongxiang section of the Han River, the upper reaches of the Tarim River, the Xiaobei section of the middle reaches of the Yellow River, and the lower reaches of the Mengjin River. There are two main conditions for the formation of wandering rivers: (1) strong water flow; (2) easy scouring and deposition of the riverbed. Due to these two conditions, the riverbed becomes wide and shallow, and the sandbars move and shrink significantly, resulting in scattered water flow and unstable main channel. At present, research on riverbed stability at home and abroad involves disciplines such as hydrodynamics, sediment transport mechanics, and river geomorphology. The main research methods include remote sensing image data analysis, measured data analysis, mathematical models, and physical models. Existing research typically focuses on qualitative analyses of simple water and sediment transport patterns, sediment transport mechanisms, and river stability characteristics. Therefore, quantitative evaluation indicators for riverbed stability have not yet been established, and systematic and effective methods for evaluating riverbed stability are lacking. Furthermore, existing research generally lacks visualization capabilities, failing to intuitively display information related to riverbed stability.

[0003] The dramatic horizontal undulation of wandering river channels negatively impacts their lateral stability, posing a serious threat to flood control safety on both banks. Riverbed stability is a crucial indicator in the management and protection of wandering rivers. There is an urgent need to develop a multi-source data fusion-based method for evaluating the stability of wandering riverbeds. This method would quantitatively analyze the impact of changes in natural conditions and strong human activities on riverbed stability and visually represent riverbed stability information. With the future development of the Yangtze River Economic Belt and the ecological protection and high-quality development of the Yellow River Basin, maintaining the stability of wandering riverbeds is of paramount importance. The role of riverbed stability assessment in wandering river channel management projects and disaster prevention and mitigation cannot be ignored.

[0004] Therefore, there is an urgent need to develop a multi-source data fusion method for evaluating the stability of wandering riverbeds that can quantitatively analyze the impact of changes in natural conditions and strong human activities on the stability of the riverbed, and intuitively display the riverbed stability information. Summary of the Invention

[0005] The purpose of this invention is to provide a method for evaluating the stability of wandering riverbeds based on multi-source data fusion. This method can quantitatively analyze the impact of changes in natural conditions and strong human activities on the stability of wandering riverbeds and intuitively display riverbed stability information. This invention utilizes hydrological data during the flood season and topographic data from the corresponding annual river channel survey sections to propose the Channel Motion Discriminant Index (CMI) as a riverbed stability parameter. It evaluates the riverbed stability of various regions in the studied river section and visualizes the riverbed stability information by combining satellite remote sensing data. This provides a scientific basis and technology for river management and flood control and disaster reduction of wandering rivers.

[0006] To achieve the above objectives, the technical solution of the present invention is: a method for evaluating the stability of wandering riverbeds based on multi-source data fusion, characterized by comprising the following steps:

[0007] Step 1: Collect information;

[0008] Collect historical post-flood topographic data (starting distance and elevation of measuring points at the unified measurement section, and distance between adjacent sections) and historical daily average hydrological data (flow, sediment concentration and water level) of water level stations and hydrological stations along the main / tributary streams. Interpret satellite remote sensing image data and statistically study the years in which the river channel plane swing occurred in the river section.

[0009] Step 2: Calculate the flow and sediment concentration at each monitoring section of the river section during the flood season;

[0010] Based on historical post-flood topographic data and historical daily average hydrological data of each monitoring section of the river section, a one-dimensional water and sediment mathematical model was used to calculate the flow and sediment concentration process of the monitoring sections along the river during the flood season. The specific method is as follows:

[0011] Step 2.1: Establish a one-dimensional mathematical model of water and sediment;

[0012] Step 2.2: Given the upstream and downstream boundary conditions and riverbed boundary conditions of the one-dimensional water and sediment mathematical model;

[0013] Step 2.3: Calculate the flow and sediment concentration at each monitoring section during the flood season;

[0014] Step 3: Calculate the average water flow scour intensity parameters of each sub-section of the studied river segment during the flood season in year k;

[0015] The average water flow scour intensity parameters of each sub-section in the study river section during the flood season in year k are calculated and used as the water and sediment conditions in the CMI (Channel Motion Discriminant Index).

[0016] Step 4: Determine the river section dimensions, flat river width, and flat water depth for each sub-river section;

[0017] The width and depth of the riverbed at each cross-section were determined, and based on the cross-sectional width and depth, the width of the riverbed at the segmental scale for each sub-section was further determined. and the depth of the flat beach

[0018] Step 5: Calculate the river facies coefficient parameters at the river scale for each sub-river segment in the k-th year of the study river segment;

[0019] Based on the river section-scale flat-shoal width and flat-shoal depth of each sub-river section determined in step 4, the river section-scale facies coefficient parameters of each sub-river section in the study section in year k are calculated. The riverbed boundary condition, used as a discriminant index for riverbed horizontal swing (CMI), is calculated using the following formula:

[0020]

[0021] In the formula: k is the year; The width of the flat river (m) is the river section dimension of the sub-river segment; The flat-shoal water depth (m) of the river section;

[0022] Step 6: Determine the value of the multi-year channel horizontal swing discrimination index (CMI) for each sub-river section;

[0023] Step 7: Use logistic regression to classify the riverbed stability parameter CMI of each sub-section in the study river section;

[0024] Step 7.1: Set the threshold for the logistic regression model. A probability greater than the model threshold is defined as 1, and a probability less than the model threshold is defined as 0.

[0025] Step 7.2: Perform binary classification on the model regression results to obtain the confusion matrix of the logistic regression model binary classification results under different model thresholds;

[0026] Step 7.3: Use accuracy (ACC) to test the precision of the logistic regression model's binary classification results under different model thresholds, and select the model threshold that maximizes ACC.

[0027] Step 7.4: Based on the output probability of river branching, classify the stability level of the estuary flow path into "stable", "basically stable", "slightly unstable", "unstable" and "extremely unstable".

[0028] Step 8: Evaluate the riverbed stability of each sub-section of the study river by combining satellite remote sensing images, and obtain the riverbed stability evaluation map of the study river section;

[0029] Step 8.1: Calculate the channel plane swing discrimination index (CMI) for each sub-section of the studied river section in a certain year, thereby obtaining the estuary flow path stability level of each sub-section.

[0030] Step 8.2: Different colors are used to represent different levels on the stability evaluation map of the studied river section to classify the levels.

[0031] In the above technical solution, in step 2.1, a one-dimensional hydrodynamic model is established, and the main governing equations include:

[0032] Continuity equation for water flow:

[0033] Water flow equation:

[0034]

[0035] Suspended sediment transport equation:

[0036] Riverbed scouring and deposition equation:

[0037] Where: Q is the flow rate (m) 3 / s); Z is the water level (m); A0 and A are the cross-sectional area of ​​the riverbed erosion and deposition and the cross-sectional area of ​​the water flow, respectively (m²). 2 B is the width of the river (m); α f S is the water flow correction factor; S is the cross-sectional average sediment concentration (kg / m³). 3 );ρ m Density of turbid water (kg / m³) 3 ), ρ m =ρ f +(Δρ / ρ s )S, where Δρ=ρ s -ρ f , ρ s ρ f These represent the density of clear water and the density of sediment, respectively; ρ1, q1, and u1 are the lateral inflow and outflow densities per unit river length (kg / m³). 3 ) and inflow / outflow flow rate (m 3 / s), the component of the lateral inflow and outflow velocity in the mainstream direction (m / s); J f For hydraulic gradient; J l The local resistance caused by cross-sectional expansion; h c The submersion depth of the centroid of the cross-section is given by g (m); g, x, and t are the acceleration due to gravity (m / s²). 2 ), distance along the path (m) and time (s); U is the average flow velocity at the cross section (m / s); N is the number of suspended sediment groups; S lκ Sediment concentration (kg / m³) for lateral inflow and outflow groups 3 ); ρ' is the dry density of the bed sand (kg / m³) 3 );S *κ S κ The grouping carrying capacity (kg / m³) of suspended sediment in the k-th particle size group3 ) and sand content (kg / m 3 );ω κ α κ These are the settling velocity (m / s) and recovery saturation coefficient of the suspended sediment in the k-th particle size group.

[0038] In the above technical solution, in step 2.2, given the upstream and downstream boundary conditions of the one-dimensional water and sediment mathematical model and the riverbed boundary conditions, the following is achieved: assuming the riverbed is a fixed bed and using post-flood topographic data as the riverbed boundary conditions; inputting the flood season flow process of the inlet section as the upstream boundary conditions; setting the flood season water level process of the outlet section as the downstream boundary conditions.

[0039] In the above technical solution, in step 2.3, the process of calculating the flow and sediment concentration of each monitoring section during the flood season is implemented as follows: the main channel and the beach are divided on the calculation section, and then different code feature value tables are used at the nodes of the main channel and the beach; where the main channel = 0 and the beach = 1; the roughness of the main channel is determined by the relationship between flow and roughness, while the roughness of the high beach and the low beach is taken as a fixed value; the Preissmann scheme is used for discrete calculation to obtain the flow elements of each section; the explicit upwind scheme is used for discrete calculation to obtain the flow and sediment concentration of each section during the flood season.

[0040] In the above technical solution, in step 3, the average water flow scour intensity parameters of each sub-river segment in the study river segment during the flood season in year k are calculated. The specific method is as follows:

[0041] Step 3.1: Divide the study river section into n sub-segments with small length differences. The length of each sub-segment is 10%-15% of the length of the study river section, and the starting section of the sub-segment must be a general measurement section.

[0042] Step 3.2: Based on the flow rate and sediment concentration of each monitoring section determined in Step 2, calculate the average flood season scour intensity of each sub-river segment in year k.

[0043]

[0044] In the formula: k is the year; The average flood season flow (m³) at the initial cross-section of the sub-river section in year k. 3 / s); The average suspended sediment concentration (kg / m³) during the flood season of the kth year at the initial unified measurement section of the sub-river section. 3 ).

[0045] In the above technical solution, step 4 involves determining the river section scale and flat river width for each sub-river segment. and the depth of the flat beach The specific method is as follows:

[0046] Step 4.1: Determine the flat beach elevation of each survey section;

[0047] The specific principles for determining the flat beach elevations of each survey section are as follows:

[0048] When the shoal lips on both sides of the main channel of a certain cross section are obvious, the elevation of the lower side of the shoal lips on both sides of the main channel of the cross section shall be taken as the flat beach elevation.

[0049] When the beach lips on both sides of the main channel of a certain section are not obvious, the beach elevation of the adjacent sections before and after should be comprehensively determined to ensure that the beach lip elevation does not change significantly along the way.

[0050] Step 4.2: Determine the cross-sectional dimensions of each survey section, including the width and depth of the riverbed.

[0051] Step 4.3: Calculate the river section scale, flat river width, and flat water depth of the sub-river section.

[0052] In the above technical solution, in step 4.2, the cross-sectional dimensions of each survey section, the width of the riverbed and the water depth of the riverbed, are determined as follows:

[0053] The width of the flat river at the cross section is the distance between the two banks of the flat river at that cross section. The area enclosed by the water level of the flat river at the cross section and the main channel is the area of ​​the flat river at the cross section. The water depth of the flat river at the cross section = the area of ​​the flat river at the cross section / the width of the flat river at the cross section.

[0054] In the above technical solution, in step 4.3, the river section-scale flat-shoal width and flat-shoal water depth of the sub-river section are calculated as follows: A method combining geometric mean based on logarithmic transformation and weighted average of cross-sectional spacing is used to calculate the river section-scale flat-shoal width and flat-shoal water depth. The calculation formula is as follows:

[0055]

[0056] In the formula: These are the characteristic parameters of the river channel at the river section scale, including the river width and water depth at the river section scale. x represents the width (m) and depth (m) of the flat river at the i-th cross-section; i The distance from the i-th cross-section to the starting cross-section is (m); N represents the number of cross-sections; and L is the length of the calculated river segment (m).

[0057] In the above technical solution, in step 6, the value of the multi-year channel horizontal oscillation discrimination index (CMI) for each sub-river segment is determined as follows:

[0058] Based on the data sets of the average flood season scour intensity parameters and river facies coefficient parameters of each sub-section of the study river section obtained in steps 3 and 5, the discriminant equation for river channel planar swing is established as follows:

[0059]

[0060] In the formula: This indicates the intensity of water flow and the sediment transport capacity of the river channel. This characterizes the lateral stability of the riverbed and the mobility of the riverbanks;

[0061] Data points above the discriminant equation indicate that channel plane undulation has occurred, while data points below the discriminant equation indicate that channel plane undulation has not occurred. A parameter λ is introduced to evaluate the reliability of the equation, as shown in the following equation:

[0062]

[0063] In the formula: λ is the boundary parameter; N c N represents the number of points within the correctly defined boundary region of the coordinate system. w The number of points located in the erroneous boundary region of the coordinate system; N T Total points;

[0064] By adjusting the undetermined parameters a and b in the river channel planar swing discrimination equation, the number N points distributed in the correct boundary region is made possible. c At most, the number of points N in the error demarcation region can be controlled simultaneously. w The increase of λ; when λ is greater than 0.5, the equation for the river channel outlet boundary can be considered accurate and reasonable; under the premise of λ>0.5, the calibrated river channel plane swing discrimination equation is transformed to obtain the river channel plane swing discrimination index CMI for each sub-river segment:

[0065]

[0066] CMI = a (10)

[0067] When CMI≥a, it indicates that there is a high probability that the river channel will swing.

[0068] When CMI < a, it indicates that the channel plane will not swing in this sub-section.

[0069] In the above technical solution, in step 7, the logistic regression method is used to classify the riverbed stability parameter CMI of each sub-segment of the studied river section, as follows: The final optimized logistic regression model is obtained using the Bayesian optimization algorithm, and the model thresholds are set to 0.25, 0.5, and 0.75. The calculated probability greater than the model threshold is defined as 1, and the probability less than the model threshold is defined as 0. The model regression results are binary classified to obtain the confusion matrix of the logistic regression model binary classification results under different model thresholds. The accuracy ACC is used to test the accuracy of the logistic regression model binary classification results under different model thresholds (ACC is the ratio of the number of correctly predicted samples to the total number of samples). The model threshold that maximizes ACC is selected, and then the riverbed swing probability output under the model threshold is segmented to classify the estuary flow path stability level.

[0070] The multi-source data includes measured data and remote sensing data.

[0071] The present invention has the following advantages and beneficial effects:

[0072] (1) The method is systematic and effective, and closely related to engineering practice;

[0073] This invention establishes an empirical relationship between the average flood season scour intensity parameter and the river section-scale facies coefficient parameter based on the formation conditions of wandering rivers. Then, using this empirical relationship, it derives the Channel Motion Discriminant Index (CMI) for the studied river section, serving as a riverbed stability parameter. The magnitude of the CMI is then used to determine whether there is a high probability of channel motion occurring in different areas of the studied river section. The riverbed stability parameter proposed in this invention has a strong theoretical and statistical basis, fully reflecting the influence mechanism of flow intensity and riverbed boundary conditions on the formation process of wandering rivers. The evaluation results of this invention are at the river section scale (providing intuitive quantitative analysis), more comprehensively reflecting the riverbed stability of a river section. This overcomes the shortcomings of existing technologies that only use cross-sectional stability coefficients for qualitative analysis, failing to provide intuitive quantitative analysis.

[0074] (2) It can provide a highly operable reference for the study of riverbed stability of wandering rivers;

[0075] This invention derives the Channel Motion Discriminant Index (CMI) by establishing empirical relationships; it also visualizes the riverbed stability information of different regions in the study section by combining satellite remote sensing image data; this method yields unique data processing results and has strong operability and engineering application value. Attached Figure Description

[0076] Figure 1 This is a diagram showing the results of the shoal and channel division at the Shanglijin and Xihekou sections of a certain river downstream of a certain river in an embodiment of the present invention.

[0077] Figure 2 The calculation results of the flood season flow and sediment concentration of the Cha 3 section in this embodiment of the invention for 2007 and 2015 are shown.

[0078] Figure 3 This is a diagram showing the determination of the flat beach elevation of the Lijin and Xihekou sections in 1990 in an embodiment of the present invention.

[0079] Figure 4 This is a graph showing the relationship between river phase coefficient and water flow scour intensity in the Lijin-Cha3 river section from 1976 to 2018 in an embodiment of the present invention.

[0080] Figure 5 The images show the riverbed stability evaluation figures for the Lijin-Cha3 river section in 2007 and 2015 in this embodiment of the invention.

[0081] Figure 6 This is a flowchart of the method of the present invention.

[0082] exist Figure 2 In this context, Q represents the flow rate (m). 3 / s); S is the average sand content of the cross section (kg / m³). 3 ).

[0083] exist Figure 3 In the middle, B bf and H bf These refer to the width and depth of the Pingtan River, respectively.

[0084] Figure 5 The numbers 1, 2, ..., 9 in the text represent the sub-segment numbers into which the river segment is divided. Figure 5 If it is divided into 9 sub-river sections, then there are sub-river sections 1, 2, ..., 9. Detailed Implementation

[0085] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings. However, these descriptions do not constitute a limitation of the present invention and are merely illustrative. The advantages of the present invention will become clearer and easier to understand through this description.

[0086] Example

[0087] The present invention will now be described in detail using an example of its application to the stability evaluation of a wandering riverbed. This invention will also provide guidance for the application of the invention to the stability evaluation of other wandering riverbeds.

[0088] Currently, there is no systematic research on the evaluation of riverbed stability in a certain river; most studies are qualitative.

[0089] like Figure 1 As shown, this embodiment uses the method of the present invention to evaluate the stability of a wandering riverbed, including the following steps:

[0090] Step 1: Collect historical post-flood topographic data (distance between measuring points and elevation, and distance between adjacent sections) for each unified measurement section of the study river segment, as well as historical daily average hydrological data (discharge, sediment concentration, and water level) for water level and hydrological stations along the main / tributary streams. Interpret satellite remote sensing imagery and statistically analyze the years in which the river channel experienced horizontal undulation. Taking the Lijin-Cha3 section of a certain river's lower reaches as the study segment, with the Lijin station at the inlet and the Cha3 section at the outlet, historical post-flood topographic data for 41 unified measurement sections of this river segment from 1976 to 2018 were collected. Simultaneously, historical daily average hydrological data for the Lijin and Cha3 sections were also collected.

[0091] Step 2, based on historical post-flood topographic data and historical daily average hydrological data of each monitoring section of the research river section, uses a one-dimensional water and sediment mathematical model to calculate the flow and sediment concentration processes of the monitoring sections along the river during the flood season. This step further includes:

[0092] Step 2.1: Establish a one-dimensional mathematical model of water and sediment; the main governing equations include:

[0093] Continuity equation for water flow:

[0094] Water flow equation:

[0095]

[0096] Suspended sediment transport equation:

[0097] Riverbed scouring and deposition equation:

[0098] In the formula: Q is the flow rate (m³ / s). 3 / s); Z is the water level (m); A0 and A are the cross-sectional area of ​​the riverbed erosion and deposition and the cross-sectional area of ​​the water flow, respectively (m²). 2 B is the width of the river (m); α f S is the water flow correction factor; S is the cross-sectional average sediment concentration (kg / m³). 3 );ρ m Density of turbid water (kg / m³) 3 ), ρ m =ρ f +(Δρ / ρ s )S, where Δρ=ρ s -ρ f , ρ s ρ f These represent the density of clear water and the density of sediment, respectively; ρ1, q1, and u1 are the lateral inflow and outflow densities per unit river length (kg / m³). 3 ) and inflow / outflow flow rate (m 3 / s), the component of the lateral inflow and outflow velocity in the mainstream direction (m / s); J fFor hydraulic gradient; J l The local resistance caused by cross-sectional expansion; h c The submersion depth of the centroid of the cross-section is given by g (m); g, x, and t are the acceleration due to gravity (m / s²). 2 ), distance along the path (m) and time (s); U is the average flow velocity at the cross section (m / s); N is the number of suspended sediment groups; S lκ Sediment concentration (kg / m³) for lateral inflow and outflow groups 3 ); ρ' is the dry density of the bed sand (kg / m³) 3 );S *κ S κ The grouping carrying capacity (kg / m³) of suspended sediment in the k-th particle size group 3 ) and sand content (kg / m 3 );ω κ α κ These are the settling velocity (m / s) and recovery saturation coefficient of the suspended sediment in the k-th particle size group.

[0099] Step 2.2: Given the upstream and downstream boundary conditions and riverbed boundary conditions of the one-dimensional hydro-sediment mathematical model. Input the flood season flow process of the Lijin section as the upstream boundary condition, and input the flood season water level process of the Cha3 section as the downstream boundary condition. Assume that the riverbed of the Lijin-Cha3 section is a fixed bed, and use the post-flood topographic data of various statistical sections of this section in the same year as the riverbed boundary conditions.

[0100] Step 2.3: Calculate the flood season flow and sediment concentration processes at each statistical cross-section from 1976 to 2018. First, divide the cross-section into the main channel and the floodplain. Then, at the nodes of the main channel and the floodplain, use different code characteristic value tables (main channel = 0, floodplain = 1) (e.g., ...). Figure 1 (As shown); secondly, the roughness of the main channel is determined using the relationship between flow rate and roughness, while the roughness of the high and low shoals is taken as a fixed value; then, the Preissmann scheme is used for discrete calculation to obtain the flow elements of each cross-section; finally, the explicit upwind scheme is used for discrete calculation to obtain the flow rate and sediment concentration process of each cross-section during the flood season (e.g., Figure 2 (As shown).

[0101] Step 3: Calculate the average flood season scour intensity parameters of each sub-section in the study river segment during the k-th year as the water and sediment conditions in the Channel Motion Discriminant Index (CMI). This step further includes:

[0102] Step 3.1: Divide the Lijin-Cha3 river section into 9 sub-segments with relatively small length differences. The length of each sub-segment ranges from 9.93 to 11.31 km. The starting section of each sub-segment is the unified measurement section.

[0103] Step 3.2: Based on the flow rate and sediment concentration of each monitoring section determined in Step 2, calculate the average flood season scour intensity of each sub-river segment in year k.

[0104]

[0105] In the formula: k is the year; The average flood season flow (m³) at the initial cross-section of the sub-river section in year k. 3 / s); The average suspended sediment concentration (kg / m³) during the flood season of the kth year at the initial unified measurement section of the sub-river section. 3 The average scouring intensity of the flood season for each sub-river section from 1976 to 2018 was calculated sequentially using this method.

[0106] Step 4: Determine the flat river width and flat river depth at each survey section, and further determine the flat river width at the section level for each sub-river segment based on the cross-sectional scale of the flat river width and flat river depth. and the depth of the flat beach This step further includes:

[0107] Step 4.1: Determine the flat beach elevation of each survey section. For example... Figure 3 As shown, the flat beach elevation at the Lijin section in 1990 was 13.82m, and the flat beach elevation at the Xihekou section was 8.83m. The specific principles for determining the flat beach elevation of each surveyed section are as follows:

[0108] When the shoal lips on both sides of the main channel of a certain cross section are obvious, the elevation of the lower side of the shoal lips on both sides of the main channel of the cross section shall be taken as the flat beach elevation.

[0109] When the beach lips on both sides of the main channel of a certain cross section are not obvious, the beach elevation is determined by referring to the elevation of the adjacent cross sections before and after, so as to minimize the change in the beach lip elevation along the way.

[0110] Step 4.2: Determine the cross-sectional dimensions of each survey section, including the width and depth of the riverbed. For example... Figure 3 As shown, in 1990, the width and depth of the Pingtan River at the Lijin section were 511.8m and 3.02m, respectively, while the width and depth of the Pingtan River at the Xihekou section were 350.9m and 2.88m, respectively.

[0111] Step 4.3: Using a combination of geometric mean based on logarithmic transformation and weighted average of cross-sectional spacing, the river width and water depth at the flat beach scale of the nine sub-river segments are calculated.

[0112]

[0113] In the formula: These are the characteristic parameters of the river channel at the river section scale, including the river width and water depth at the river section scale. Let be the width (m) and depth (m) of the flat river at the i-th cross-section; Let x be the characteristic parameters of the flat riverbed at the i-th and i+1-th cross sections; i The distance (m) from the i-th cross-section to the starting cross-section; N represents the number of cross-sections; L is the length of the calculated river segment (m). In this embodiment, N = 41.

[0114] Step 5: Based on the river section-scale flatland width and flatland depth of each sub-section determined in Step 4, calculate the river section-scale facies coefficient parameters of each sub-section in the study section from 1976 to 2018. Riverbed boundary conditions are used as a discriminant in the Channel Motion Index (CMI).

[0115]

[0116] In the formula: k is the year. In this embodiment, k = 1976, 1977, ..., 2018; The width of the flat river (m) is the river section dimension of the sub-river segment; The flat-shoal water depth (m) is the river section-scale depth of the sub-river.

[0117] Step 6: Determine the multi-year channel horizontal oscillation discrimination index (CMI) value for each sub-river segment. The specific implementation method in this embodiment is as follows:

[0118] Based on the data sets of the average flood season scour intensity parameters and river facies coefficient parameters of each sub-river segment obtained in steps 3 and 5, the relationship between river facies coefficient and scour intensity of the Lijin-Cha 3 river segment from 1976 to 2018 was plotted, thereby establishing the river channel planar swing discrimination equation:

[0119]

[0120] In the formula: This indicates the intensity of water flow and the sediment transport capacity of the river channel. This characterizes the lateral stability of the riverbed and the mobility of the riverbanks. Data points above the discriminant equation indicate that channel planar undulation has occurred, while those below indicate that channel planar undulation has not occurred. A parameter λ is introduced to evaluate the reliability of the equation:

[0121]

[0122] In the formula: λ is the boundary parameter; N c N represents the number of points within the correctly defined boundary region of the coordinate system. w The number of points located in the erroneous boundary region of the coordinate system; N T Let N be the total number of points. By adjusting the undetermined parameters a and b in the river channel planar swing discrimination equation, the number of points N distributed in the correct boundary region is made possible.c At most, the number of points N in the error demarcation region can be controlled simultaneously. w The increase of λ. When λ is greater than 0.5, the equation for the river channel distributary boundary can be considered accurate and reasonable. Under the premise that λ>0.5, the calibrated river channel plane swing discrimination equation is transformed to obtain the river channel plane swing discrimination index CMI for each sub-river segment:

[0123]

[0124] CMI = a(10)

[0125] When CMI ≥ a, there is a high probability that the river channel will wobble. For example... Figure 4 As shown ( Figure 4 The most crucial step in obtaining the Channel Plane Swing Discriminant Index (CMI) is that the data points representing "During Plane Swing" and "Without Plane Swing" show a clear boundary, thus yielding a boundary equation. Based on this boundary equation, the Channel Plane Swing Discriminant Index (CMI) is further obtained. By plotting the relationship between the river phase coefficient and the water flow scour intensity, it can be seen that the data points show a clear boundary. Calibration yields a = 38.16, b = -0.39, and λ = 0.74. Therefore, the CMI of the sub-river segment in this embodiment is ≥ 38.16, indicating that there is a high probability that channel plane sway will occur in this embodiment.

[0126] Step 7: The riverbed stability parameter CMI of each sub-segment in the study river section is classified using logistic regression. The specific implementation method in this embodiment is as follows:

[0127] The CMI values ​​of the sub-river segments were input into the logistic regression model. The accuracy of binary classification calculations was compared using three model thresholds: 0.25, 0.5, and 0.75. Setting the model thresholds to 0.25, 0.5, and 0.75 yielded overall accuracy (ACC) values ​​of 0.86, 0.97, and 0.89, respectively. Therefore, the model threshold should be set to 0.5. Based on the calculation results, the riverbed stability level standards were set from low to high as follows: when the output probability P... A When the value is less than 0.21, it is classified as "stable"; when the output probability is 0.21 ≤ P... A When the probability is less than 0.42, it is classified as "basically stable"; when the output probability is 0.42 ≤ P... A When the probability is less than 0.61, it is classified as "sub-unstable"; when the output probability is 0.61 ≤ P... A When the probability is less than 0.83, it is classified as "unstable"; when the output probability is 0.83 ≤ P... A When the value is less than 1, it is classified as "extremely unstable".

[0128] Step 8: The riverbed stability of each sub-segment within the study river section is evaluated using satellite remote sensing imagery, resulting in a riverbed stability evaluation map of the study river section. This allows for timely implementation of engineering or non-engineering measures based on the riverbed stability information displayed on the map, reducing losses caused by lateral channel shifting. The evaluation results are validated using data from 2007 and 2015. In 2007, sub-segment 9 of the Lijin-Cha3 river section experienced channel shifting, while no channel shifting occurred in the Lijin-Cha3 river section in 2015. The evaluation information is as follows: Figure 5 As shown ( Figure 5 The riverbed stability zoning was visualized. The results showed that: (1) The riverbed stability evaluation map was consistent with the actual situation. In 2015, the overall riverbed stability of the Lijin-Cha3 section was higher than that in 2007. In 2007, sub-sections No. 4, No. 8 and No. 9 all reached the "extremely unstable" level; (2) In 2015, the riverbed stability level of sub-sections No. 7 and No. 9 was "unstable", indicating that it had not reached but was close to the critical level of the riverbed level swing. Therefore, these two sub-sections should be the key areas for future flood control projects; (3) The flow path stability level of sub-section No. 8 decreased from "extremely unstable" in 2007 to "basically stable" in 2015. The reason may be that in 2011, an artificial straightening project was carried out near the upstream of Qingjia 8, which greatly improved the stability of sub-section No. 8.

[0129] In summary, this embodiment employs the method of the present invention, utilizing hydrological data from the flood season and topographic data from the corresponding annual river channel survey sections, to propose the Channel Motion Discriminant Index (CMI) as a riverbed stability parameter. This index is used to evaluate the riverbed stability of various regions within the studied river section, and riverbed stability information is visualized using satellite remote sensing data. This provides a scientific basis and technology for river management and flood control and disaster reduction in wandering rivers. The method of the present invention has a strong theoretical and statistical foundation, fully reflecting the influence mechanism of flow intensity and riverbed boundary conditions on the formation process of wandering rivers. The data processing results are unique, demonstrating strong operability and engineering application value.

[0130] The above-described embodiment, using the Lijin-Cha3 section of a certain river's lower reaches as the research section, is merely an illustrative example of the technical solution of this invention. The method for evaluating the stability of wandering riverbeds based on multi-source data fusion involved in this invention is not limited to the content described in the above embodiments, but is defined by the scope of the claims. Any modifications, additions, or equivalent substitutions made by those skilled in the art based on these embodiments are within the scope of protection claimed by the claims of this invention.

[0131] All other unspecified parts belong to the prior art.

Claims

1. A method for evaluating the stability of wandering riverbeds based on multi-source data fusion, characterized by: Includes the following steps, Step 1: Collect information; Collect and study historical post-flood topographic data of each monitoring section of the river section, as well as historical daily average hydrological data of water level stations and hydrological stations along the main / tributary streams. Interpret satellite remote sensing image data and statistically study the years in which the river channel plane swing occurred in the river section. Step 2: Calculate the flow and sediment concentration at each monitoring section of the river section during the flood season; Based on historical post-flood topographic data and historical daily average hydrological data of each monitoring section of the river section, a one-dimensional water and sediment mathematical model was used to calculate the flow and sediment concentration process of the monitoring sections along the river during the flood season. The specific method is as follows: Step 2.1: Establish a one-dimensional mathematical model of water and sediment; Step 2.2: Given the upstream and downstream boundary conditions and riverbed boundary conditions of the one-dimensional water and sediment mathematical model; Step 2.3: Calculate the flow and sediment concentration at each monitoring section during the flood season; Step 3: Calculate the average water flow scour intensity parameters of each sub-section of the studied river segment during the flood season in year k; The average water flow scour intensity parameters of each sub-section in the study river section during the flood season in year k are calculated and used as the water and sediment conditions in the CMI (Channel Motion Discriminant Index). Step 4: Determine the river section dimensions, flat river width, and flat water depth for each sub-river section; The width and depth of the riverbed at each cross-section were determined, and based on the cross-sectional width and depth, the width of the riverbed at the segmental scale for each sub-section was further determined. and the depth of the flat beach Step 5: Calculate the river facies coefficient parameters at the river scale for each sub-river segment in the k-th year of the study river segment; Based on the river section-scale flat-shoal width and flat-shoal depth of each sub-river section determined in step 4, the river section-scale facies coefficient parameters of each sub-river section in the study section in year k are calculated. The riverbed boundary condition, used as a discriminant index for riverbed horizontal swing (CMI), is calculated using the following formula: In the formula: k is the year; The width of the flat river section is the river width of the sub-river segment, in meters; The flat-shoal water depth of the river section is measured in meters. Step 6: Determine the value of the multi-year channel horizontal swing discrimination index (CMI) for each sub-river section; Step 7: Use logistic regression to classify the riverbed stability parameter CMI of each sub-section in the study river section; Step 7.1: Set the threshold for the logistic regression model. A probability greater than the model threshold is defined as 1, and a probability less than the model threshold is defined as 0. Step 7.2: Perform binary classification on the model regression results to obtain the confusion matrix of the logistic regression model binary classification results under different model thresholds; Step 7.3: Use the accuracy (ACC) test to examine the precision of the logistic regression model's binary classification results under different model thresholds, and select the model threshold that maximizes the ACC. Step 7.4: Based on the output probability of river branching, classify the stability level of the estuary flow path into "stable", "basically stable", "slightly unstable", "unstable" and "extremely unstable"; Step 8: Evaluate the riverbed stability of each sub-section of the study river by combining satellite remote sensing images, and obtain the riverbed stability evaluation map of the study river section; Step 8.1: Calculate the channel plane swing discrimination index (CMI) for each sub-section of the studied river section in a certain year, thereby obtaining the estuary flow path stability level of each sub-section. Step 8.2: Different colors are used to represent different levels on the stability evaluation map of the studied river section to classify the levels.

2. The method for evaluating the stability of wandering riverbeds based on multi-source data fusion according to claim 1, characterized in that: In step 2.1, a one-dimensional hydrodynamic model is established, and the governing equations include: Continuity equation for water flow: Water flow equation: Suspended sediment transport equation: Riverbed scouring and deposition equation: Where: Q is the flow rate, m 3 / s; Z is the water level, m; A0 and A are the cross-sectional areas of riverbed erosion and deposition and the cross-sectional area of ​​water flow, respectively, m. 2 B represents the width of the river, in meters; α f S is the water flow correction factor; S is the cross-sectional average sediment concentration, kg / m³. 3 ;ρ m The density of turbid water is kg / m³. 3 , ρ m =ρ f +(Δρ / ρ s )S, where Δρ=ρ s -ρ f , ρ s ρ f Here, ρ1 represents the density of clear water and the density of sediment, respectively; ρ1, q1, and u1 represent the lateral inflow / outflow density per unit river length, the lateral inflow / outflow flow rate per unit river length, and the component of the lateral inflow / outflow velocity in the mainstream direction, respectively, with units of kg / m2. 3 m 3 / s, m / s; J f For hydraulic gradient; J l The local resistance caused by cross-sectional expansion; h c The submerged depth of the centroid of the cross-section is given in meters (m); g, x, and t represent the acceleration due to gravity, the distance traveled, and time, respectively, with units of m / s. 2 m, s; U is the average cross-sectional velocity, m / s; N is the number of suspended sediment groups; S lκ Sediment concentration for lateral inflow and outflow groups, kg / m³ 3 ρ' is the dry density of the bed sand, kg / m³ 3 S *κ S κ The sediment carrying capacity and sediment concentration of the suspended sediment in the k-th particle size group are given in kg / m³. 3 kg / m 3 ;ω κ α κ These represent the settling velocity and recovery saturation coefficient of suspended sediment in the k-th particle size group.

3. The method for evaluating the stability of wandering riverbeds based on multi-source data fusion according to claim 2, characterized in that: In step 2.2, given the upstream and downstream boundary conditions and riverbed boundary conditions of the one-dimensional water and sediment mathematical model, the following is implemented: assuming the riverbed is a fixed bed and using post-flood topographic data as the riverbed boundary conditions; inputting the flood season flow process at the inlet section as the upstream boundary condition; setting the flood season water level process at the outlet section as the downstream boundary condition.

4. The method for evaluating the stability of wandering riverbeds based on multi-source data fusion according to claim 3, characterized in that: In step 2.3, the flow and sediment concentration processes of each monitoring section during the flood season are calculated as follows: the main channel and the floodplain are divided on the calculation section, and then different code feature values ​​are used to represent the nodes of the main channel and the floodplain; where the main channel = 0 and the floodplain = 1; the roughness of the main channel is determined by the relationship between flow and roughness, while the roughness of the high and low floodplains is taken as a fixed value; the Preissmann scheme is used for discrete calculation to obtain the flow elements of each section; the explicit upwind scheme is used for discrete calculation to obtain the flow and sediment concentration processes of each section during the flood season.

5. The method for evaluating the stability of wandering riverbeds based on multi-source data fusion according to claim 4, characterized in that: In step 3, the average flood scour intensity parameters of each sub-section of the studied river section in year k are calculated. The specific method is as follows: Step 3.1: Divide the study river section into n sub-segments with small differences in length. The length of each sub-segment is 10%-15% of the length of the study river section, and the starting section of the sub-segment must be a general measurement section. Step 3.2: Based on the flow and sediment concentration data of each monitoring section determined in Step 2, calculate the average flood season scour intensity of each sub-river segment in year k. In the formula: k is the year; Let m be the average flood season flow at the initial cross-section of the sub-river section in year k. 3 / s; The average suspended sediment concentration during the flood season in year k at the initial measurement section of the sub-river section is given in kg / m³. 3 .

6. The method for evaluating the stability of a wandering riverbed based on multi-source data fusion according to claim 5, characterized in that: In step 4, the river section dimensions and flat river width of each sub-river segment are determined. and the depth of the flat beach The specific method is as follows: Step 4.1: Determine the flat beach elevation of each survey section; The elevation of the flat beach at each survey section is determined according to the following principles: When the shoal lips on both sides of the main channel of a certain cross section are obvious, the elevation of the lower side of the shoal lips on both sides of the main channel of the cross section shall be taken as the flat beach elevation. When the beach lips on both sides of the main channel of a certain section are not obvious, the beach elevation is determined by referring to the flat beach elevation of the adjacent sections before and after, so that the beach lip elevation does not change significantly along the way. Step 4.2: Determine the cross-sectional dimensions of each survey section, including the width and depth of the riverbed. Step 4.3: Calculate the river section scale, flat river width, and flat water depth of the sub-river section.

7. The method for evaluating the stability of a wandering riverbed based on multi-source data fusion according to claim 5, characterized in that: In step 4.2, the cross-sectional dimensions of each survey section, including the width and depth of the riverbed, are determined as follows: The width of the flat river at the cross section is the distance between the two banks of the flat river at that cross section. The area enclosed by the water level of the flat river at the cross section and the main channel is the area of ​​the flat river at the cross section. The water depth of the flat river at the cross section = the area of ​​the flat river at the cross section / the width of the flat river at the cross section.

8. The method for evaluating the stability of a wandering riverbed based on multi-source data fusion according to claim 7, characterized in that: In step 4.3, the river section-scale flat-shoal width and flat-shoal depth of the sub-river segment are calculated as follows: A method combining geometric mean based on logarithmic transformation and weighted average of cross-sectional spacing is used to calculate the river section-scale flat-shoal width and flat-shoal depth. The calculation formula is as follows: In the formula: These are the characteristic parameters of the riverbed and channel at the river section scale, including the river section-scale width of the riverbed and channel. water depth x represents the width and depth of the flat river at the i-th cross-section, in meters (m) and meters (m), respectively. i The distance from the i-th cross-section to the starting cross-section is m; N represents the number of cross-sections; and L is the length of the calculated river segment, m.

9. The method for evaluating the stability of a wandering riverbed based on multi-source data fusion according to claim 8, characterized in that: In step 6, the multi-year channel horizontal oscillation discrimination index (CMI) value for each sub-river segment is determined, as follows: Based on the data sets of the average flood season scour intensity parameters and river facies coefficient parameters of each sub-section of the study river section obtained in steps 3 and 5, the discriminant equation for river channel planar swing is established as follows: In the formula: This indicates the intensity of water flow and the sediment transport capacity of the river channel. This characterizes the lateral stability of the riverbed and the mobility of the riverbanks; Data points above the discriminant equation indicate that channel plane undulation has occurred, while data points below the discriminant equation indicate that channel plane undulation has not occurred. A parameter λ is introduced to evaluate the reliability of the equation, as shown in the following equation: In the formula: λ is the boundary parameter; N c N represents the number of points within the correctly defined boundary region of the coordinate system. w N represents the number of points located in the erroneous boundary region of the coordinate system. T Total points; By adjusting the undetermined parameters a and b in the river channel planar swing discrimination equation, the number N points distributed in the correct boundary region is made possible. c At most, the number of points N in the error demarcation region can be controlled simultaneously. w The increase of λ; when λ is greater than 0.5, the equation for the river channel outlet boundary is considered accurate and reasonable; under the premise that λ>0.5, the calibrated river channel plane swing discrimination equation is transformed to obtain the river channel plane swing discrimination index CMI for each sub-river segment: When CMI≥a, it indicates that the channel plane will swing in this sub-section; When CMI < a, it indicates that the channel plane will not swing in this sub-section.

Images