A river channel base flow mixed segmentation method based on dream optimization algorithm

Abstract

The application discloses a river channel base flow mixed segmentation method based on a dream optimization algorithm, and the method comprises the following steps: firstly, acquiring and processing historical daily runoff data of a river channel; then, calculating an empirical base flow index of the river channel according to the historical daily runoff data of the river channel, constructing a river channel base flow mixed segmentation model, calculating a theoretical base flow index of the river channel, taking the error square sum of the empirical base flow index and the theoretical base flow index as an optimization target, and optimizing parameters of the river channel base flow mixed segmentation model by using the dream optimization algorithm; finally, segmenting the base flow of the river channel by using the river channel base flow mixed segmentation model after the parameter optimization, and outputting a base flow segmentation result; the dream optimization algorithm is used to realize automatic parameter optimization, the defect of deviation accumulation of a single filtering method is avoided, the segmentation efficiency and precision are significantly improved while the physical interpretability is ensured, the interannual difference and time sequence smoothing characteristics are better described, and reliable data support is provided for groundwater recharge evaluation, ecological runoff accounting and basin water resource scheduling.

Description

Technical Field

[0001] This invention belongs to the field of river baseflow segmentation technology, specifically relating to a method for hybrid segmentation of river baseflow based on the dream optimization algorithm. Background Technology

[0002] Baseflow is the long-term, stable, and relatively gentle portion of river runoff, mainly supplied by continuous groundwater and interflow. Scientifically and stably dividing baseflow helps to assess the coupling relationship between groundwater and surface water, formulate ecological discharge indicators, estimate long-term pollutant loads and background concentrations, and provide boundary conditions for reservoir scheduling and water diversion projects.

[0003] Traditional methods for dividing river baseflow include empirical or curve fitting, graphical methods, and digital filtering. Digital filtering is widely used due to its high degree of automation, low implementation cost, and strong repeatability. Eckhart filtering uses the parameter α to characterize the confluence memory (receding slope) and limits the upper limit of baseflow, making it suitable for watersheds with significant storage and release and stable aquifer response. Chapman filtering uses a single parameter k to describe the receding inertia, and has a simple structure and intuitive parameters.

[0004] However, in complex watersheds, a single filter is difficult to maintain good consistency across all years, often resulting in problems such as unstable interannual deviations and large differences between the interannual baseflow index and the target.

[0005] In summary, existing methods for dividing river baseflow have at least the following shortcomings or require improvement: (1) Aligning a certain filter parameter with the annual target by calibrating it alone often sacrifices the daily scale shape, while pursuing daily scale smoothness and peak shape may result in annual scale total error. (2) Traditional Eckhart filtering or Chapman filtering methods usually require manual setting of parameters, which may vary greatly in different watersheds or even different years, and may easily lead to deviations in the segmentation results. Summary of the Invention

[0006] To address the aforementioned problems, this invention provides a hybrid baseflow segmentation method based on the dream optimization algorithm. This method constructs a hybrid segmentation model composed of Eckhart filtering and Chapman filtering, weighting the two baseflow sequences with a weight parameter ω. Then, the sum of the annual squared errors between the empirical baseflow index and the theoretical baseflow index is used as the objective function. The dream optimization algorithm is employed for global optimization to automatically determine the optimal weight ω and parameter α. max ,k and output the result.

[0007] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for mixing and segmenting river baseflow based on a dream optimization algorithm includes the following steps: S1: Acquire and organize historical daily runoff data of the river channel; S2: Calculate the empirical baseflow index of the river based on historical daily runoff data; S3: Construct a river baseflow mixing and segmentation model and calculate the theoretical baseflow index of the river. S4: Taking the minimum sum of squared errors between the empirical baseflow index and the theoretical baseflow index as the optimization objective, the dream optimization algorithm is used to optimize the parameters of the river baseflow mixing and segmentation model. S5: Use the optimized river baseflow mixing and segmentation model to segment the river baseflow and output the baseflow segmentation results.

[0008] In step S1, historical daily runoff data of the river channel is acquired and processed to obtain... i Annual daily flow sequence Q ( i , t ), i Indicates the year sequence number. t It indicates the day number in a year.

[0009] In step S2, the empirical baseflow index of the river is calculated based on historical daily runoff data. The formula for calculating the empirical baseflow index is as follows: (1); In formula (1): BFI target ( i ) indicates the first section of the river channel i The annual experience base current index; Q 90 ( i ) indicates the first i The annual guarantee rate is 90% for daily traffic. Q 50 ( i ) indicates the first i The annual guarantee rate is 50% of the daily traffic.

[0010] In S3, the hybrid segmentation model for the river baseflow is a hybrid segmentation model composed of the Eckhart filter model and the Chapman filter model, with weight parameters... ω The base current sequences of the two models are weighted using the following formula: (2); (3); (4); In equations (2), (3), and (4): α BFI represents the Eckhart attenuation coefficient. max This represents the long-term ratio of the base current to the total flow in the Eckhart filter model.k These are the Chapman filter parameters; ω The weighting is mixed to control the proportion of the two filters. ω ∈(0,1); The first value obtained from the Eckhart filter model is represented by the first value. i Annual base flow; The Chapman filter model calculates the first... i Annual base flow.

[0011] The river baseflow mixing and segmentation model is based on the river channel's first... i Historical daily runoff data for years, used to calculate the river channel's first... i Theoretical baseflow index of the year BFI mix ( i The calculation formula is as follows: (5); In equation (5): Q ( i , t Let be the daily runoff in year i; The first value obtained by the hybrid segmentation filtering model is represented by the first value. i Annual base flow.

[0012] In S4, the objective function for optimizing the parameters of the river baseflow mixing and segmentation model is: (6); In formula (6): J The main objective function is to take the sum of squared annual errors between the empirical baseflow index and the theoretical baseflow index; I This represents the total number of years. This represents the empirical baseflow index of the river in year i.

[0013] In step S4, using the annual mean square error as the objective function, a mixed river baseflow mixing and segmentation method based on the dream optimization algorithm is employed to search for the optimal weights within the given parameter boundaries. ω、 parameter α、 BFI max , k The optimal parameters are obtained.

[0014] The basic idea of ​​the dream optimization algorithm described in this invention is to divide the iterative process into two stages: In the first 90% of iterations, an exploration phase is executed. This involves dividing the population into five subgroups, searching for the optimal solution for each subgroup, and switching between a "forgetting and replenishment strategy" and a "random learning strategy" to randomly perturb and update individuals across multiple dimensions, thereby maintaining diversity and expanding the search range. In the last 10% of iterations, a development phase is entered. All individuals use the current global optimal solution as the core, randomly selecting some dimensions for minor adjustments, while simultaneously performing boundary repair to achieve refined convergence. Throughout the process, the algorithm continuously refreshes the global optimal solution and the optimal objective function value, recording historical optimal values ​​in each generation, and ultimately outputting the global optimal solution, the corresponding optimal objective function value, and a curve showing the optimal value changing with iteration.

[0015] A river baseflow mixing and segmentation system based on a dream optimization algorithm is provided for executing a river baseflow mixing and segmentation method based on a dream optimization algorithm. The system includes: The data acquisition module is used to obtain historical daily runoff data of the river channel; The data processing module is used to process historical daily runoff data of the river. The empirical baseflow index calculation module is used to calculate the empirical baseflow index of a river based on historical daily runoff data. The baseflow segmentation model construction module is used to construct a river baseflow mixing segmentation model and calculate the theoretical baseflow index of the river. The algorithm optimization module is used to optimize the parameters of the river baseflow mixing and segmentation model by using the dream optimization algorithm, with the goal of minimizing the sum of squared errors between the empirical baseflow index and the theoretical baseflow index. The baseflow segmentation execution module is used to perform baseflow segmentation of the river using the optimized river baseflow mixing segmentation model and output the baseflow segmentation results.

[0016] A computer device includes a memory and a processor, which are communicatively connected. The memory stores computer instructions, and the processor executes the computer instructions to perform the aforementioned method for mixing and separating river baseflow based on a dream optimization algorithm.

[0017] A computer-readable storage medium storing computer instructions for causing a computer to execute the aforementioned method for mixing and separating river baseflow based on a dream optimization algorithm.

[0018] The main beneficial effects of this invention are as follows: I. Overcoming the limitations of single-scale filtering while considering multi-scale optimization objectives: Traditional single methods such as Eckhart filtering and Chapman filtering are difficult to balance the accuracy of daily morphology and annual total, often resulting in a trade-off: sacrificing the smoothness of daily runoff when calibrating annual targets, and leading to interannual total error when pursuing daily peak shape. This method innovatively constructs a weighted hybrid model of Eckhart filtering and Chapman filtering, flexibly adjusting the contribution ratio of the two filters through hybrid weights, while simultaneously coupling the Dream Optimization algorithm to globally optimize the core parameters. This design inherits the adaptability of Eckhart filtering to the stable response of aquifers and the intuitiveness of Chapman filtering parameters, while achieving a two-way balance through parameter co-optimization: ensuring a high degree of fit between the annual theoretical baseflow index and the empirical baseflow index, while maintaining the smoothness of the daily baseflow sequence and the rationality of the hydrological process, effectively solving the problem of bias accumulation in single filtering methods.

[0019] II. Automated parameter optimization reduces subjective interference and lowers the application threshold: Traditional baseflow segmentation methods rely on manually setting filter parameters. However, the runoff characteristics vary significantly across different basins and years. Manual parameter calibration is not only inefficient but also prone to introducing systematic errors due to subjectivity, resulting in poor stability of the segmentation results and weak cross-basin adaptability.

[0020] This method aims to minimize the sum of squared annual errors between the empirical baseflow index and the theoretical baseflow index. It achieves automated parameter optimization through the Dream Optimization Algorithm: within a preset range of physically reasonable parameters, the algorithm efficiently searches for the optimal parameter combination through population diversity maintenance in the exploration phase and refined convergence in the development phase. It can output parameter results that conform to hydrological laws without human intervention. This feature greatly reduces the operational threshold of the method, while avoiding the subjectivity of manual parameter selection, making the segmentation results more objective and reliable, and adapting to complex runoff scenarios in different types of watersheds such as mountainous areas, plains, and urbanized areas.

[0021] III. Improve segmentation accuracy and stability to characterize temporal and interannual features: The two-stage iterative mechanism of the Dream Optimization Algorithm endows the parameter optimization with a stronger ability to find the global optimum, enabling it to more accurately pinpoint parameter combinations that suit the runoff characteristics of the target watershed compared to traditional optimization methods. Practical application verification shows that the segmentation results of this method are significantly superior to single-filter methods in core evaluation indicators such as mean absolute error (MAE) and mean square error (MSE); compared to Eckhart filtering and Chapman filtering, the error of this method is significantly reduced. Furthermore, the combination of the hybrid model and intelligent optimization better characterizes the interannual differences and temporal smoothing features of baseflow, avoiding the segmentation bias of single methods in extreme runoff years, making the results more consistent with actual hydrological processes, and providing high-quality data support for long-term hydrological pattern analysis. Attached Figure Description

[0022] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0023] Figure 1 This is a flowchart of the overall process of the present invention; Figure 2 This is the overall flowchart of the dream optimization algorithm of the present invention; Figure 3 The results of segmenting the river baseflow using the method of this invention and the single segmentation method are shown. Detailed Implementation

[0024] Example 1: like Figure 1 As shown, a method for mixing and segmenting river baseflow based on a dream optimization algorithm includes the following steps: S1: Acquire and organize historical daily runoff data of the river channel; S2: Calculate the empirical baseflow index of the river based on historical daily runoff data; S3: Construct a river baseflow mixing and segmentation model and calculate the theoretical baseflow index of the river. S4: Taking the minimum sum of squared errors between the empirical baseflow index and the theoretical baseflow index as the optimization objective, the dream optimization algorithm is used to optimize the parameters of the river baseflow mixing and segmentation model. S5: Use the optimized river baseflow mixing and segmentation model to segment the river baseflow and output the baseflow segmentation results.

[0025] In step S1, by acquiring and organizing historical daily runoff data of the river channel, we can obtain... i Annual daily flow sequence Q ( i , t ), i Indicates the year sequence number. t It indicates the day number in a year.

[0026] In step S2, the empirical baseflow index of the river is calculated based on historical daily runoff data. The formula for calculating the empirical baseflow index is as follows: (1); In formula (1): BFI target ( i ) indicates the first section of the river channel i The annual experience base current index; Q 90 ( i ) indicates the first i The annual guarantee rate is 90% for daily traffic. Q 50 ( i ) indicates the first i The annual guarantee rate is 50% of the daily traffic.

[0027] In S3, the hybrid segmentation model for the river baseflow is a hybrid segmentation model composed of the Eckhart filter model and the Chapman filter model, with weight parameters... ω The base current sequences of the two models are weighted using the following formula: (2); (3); (4); In equations (2), (3), and (4): α BFI represents the Eckhart attenuation coefficient. max This represents the long-term ratio of the base current to the total flow in the Eckhart filter model. k These are the Chapman filter parameters; ω The weighting is mixed to control the proportion of the two filters. ω ∈(0,1); The first value obtained from the Eckhart filter model is represented by the first value. i Annual base flow; The Chapman filter model calculates the first... i Annual base flow.

[0028] The river baseflow mixing and segmentation model is based on the river channel's first... i Historical daily runoff data for years, used to calculate the river channel's first... i Theoretical baseflow index of the year BFI mix ( i The calculation formula is as follows: (5); In equation (5): Q ( i , t Let be the daily runoff in year i; The first value obtained by the hybrid segmentation filtering model is represented by the first value. i Annual base flow.

[0029] In S4, the objective function for optimizing the parameters of the river baseflow mixing and segmentation model is: (6); In formula (6): J The main objective function is to take the sum of squared annual errors between the empirical baseflow index and the theoretical baseflow index; I This represents the total number of years. This represents the empirical baseflow index of the river in year i.

[0030] In step S4, using the annual mean square error as the objective function, a mixed river baseflow mixing and segmentation method based on the dream optimization algorithm is employed to search for the optimal weights within the given parameter boundaries. ω、 parameter α、 BFI max , k The optimal parameters are obtained.

[0031] The specific execution steps of the dream optimization algorithm are as follows: Figure 2 As shown, the basic idea of ​​the dream optimization algorithm described in this invention is to divide the iterative process into two stages: In the first 90% of iterations, an exploration phase is executed. This involves dividing the population into five subgroups, searching for the optimal solution for each subgroup, and switching between a "forgetting and replenishment strategy" and a "random learning strategy" to randomly perturb and update individuals across multiple dimensions, thereby maintaining diversity and expanding the search range. In the last 10% of iterations, a development phase is entered. All individuals use the current global optimal solution as the core, randomly selecting some dimensions for minor adjustments, while simultaneously performing boundary repair to achieve refined convergence. Throughout the process, the algorithm continuously refreshes the global optimal solution and the optimal objective function value, recording historical optimal values ​​in each generation, and ultimately outputting the global optimal solution, the corresponding optimal objective function value, and a curve showing the optimal value changing with iteration.

[0032] A river baseflow mixing and segmentation system based on a dream optimization algorithm is provided for executing the aforementioned river baseflow mixing and segmentation method based on a dream optimization algorithm. The system includes: The data acquisition module is used to obtain historical daily runoff data of the river channel; The data processing module is used to process historical daily runoff data of the river. The empirical baseflow index calculation module is used to calculate the empirical baseflow index of a river based on historical daily runoff data. The baseflow segmentation model construction module is used to construct a river baseflow mixing segmentation model and calculate the theoretical baseflow index of the river. The algorithm optimization module is used to optimize the parameters of the river baseflow mixing and segmentation model by using the dream optimization algorithm, with the goal of minimizing the sum of squared errors between the empirical baseflow index and the theoretical baseflow index. The baseflow segmentation execution module is used to perform baseflow segmentation of the river using the optimized river baseflow mixing segmentation model and output the baseflow segmentation results.

[0033] A computer device includes a memory and a processor, which are communicatively connected. The memory stores computer instructions, and the processor executes the computer instructions to perform the aforementioned method for mixing and separating river baseflow based on a dream optimization algorithm.

[0034] A computer-readable storage medium storing computer instructions for causing a computer to execute the aforementioned method for mixing and separating river baseflow based on a dream optimization algorithm.

[0035] Example 2: In this embodiment, the system uses runoff data from January 1, 2015 to December 31, 2021 (time step of 1 day) to construct a runoff sequence, i.e., I=7.

[0036] The empirical baseflow index of the river was calculated based on historical daily runoff data. The empirical baseflow indices for each year are shown in Table 1, calculated using the formula for calculating the empirical baseflow index.

[0037] Table 1. Empirical Baseflow Index of River Channels in the Basin in Various Years

[0038] With the goal of minimizing the sum of squared annual errors between the empirical baseflow index and the theoretical baseflow index, the dream optimization algorithm was used to optimize the parameters of the river baseflow mixing and segmentation model. Table 2 shows the parameter optimization results, including the parameters of the Eckhart filter and the Chapman filter under the optimal parameters. α BFI max , k All of them fall within a reasonable range, consistent with the "slow change" characteristic of hydrological processes.

[0039] Table 2 Optimization Results of the Mixing and Separation Method for Daily Runoff and Baseflow in the Watershed

[0040] The river baseflow is segmented using a parameter-optimized baseflow mixing and segmentation model, and the baseflow segmentation results are output, such as... Figure 3 As shown, Q represents the original daily runoff data; B_Eck represents the segmentation result of the Eckhart filter method; B_Chap represents the segmentation result of the Chapman filter method; and B_Mix represents the segmentation result of the method of this invention.

[0041] Depend on Figure 3 It can be seen that the method of the present invention is closer to the empirical baseflow index target than the single segmentation method, while maintaining the smoothness and peak shape on a daily scale.

[0042] The evaluation indicators for the calculation results of the single method and the method of the present invention are shown in Table 3. The smaller the value of the evaluation indicator, the smaller the error, indicating that the method is better.

[0043] The evaluation indices MAE and MSE of the method of this invention are both lower than those of the two single filtering methods, indicating that the river baseflow mixing and segmentation method is superior to the single method and the error is significantly reduced.

[0044] Table 3 Evaluation Indicators of Daily Runoff Calculation Results by Single and Mixed Segmentation Methods in Watersheds

[0045] In summary, traditional single filtering methods may have deviations under different runoff conditions, while the river baseflow hybrid segmentation method based on the Dream Optimization Algorithm can give full play to the advantages of both methods. Under the premise of ensuring the physical rationality of the parameters, it significantly reduces the error of baseflow segmentation compared with traditional single methods, demonstrating better accuracy and stability.

Claims

1. A method for mixing and segmenting river baseflow based on a dream optimization algorithm, characterized in that... Includes the following steps: S1: Acquire and organize historical daily runoff data of the river channel; S2: Calculate the empirical baseflow index of the river based on historical daily runoff data; S3: Construct a river baseflow mixing and segmentation model and calculate the theoretical baseflow index of the river. S4: Taking the minimum sum of squared errors between the empirical baseflow index and the theoretical baseflow index as the optimization objective, the dream optimization algorithm is used to optimize the parameters of the river baseflow mixing and segmentation model. S5: Use the optimized river baseflow mixing and segmentation model to segment the river baseflow and output the baseflow segmentation results.

2. The method for river baseflow mixing and segmentation based on dream optimization algorithm according to claim 1, characterized in that: In step S1, historical daily runoff data of the river channel is acquired and processed to obtain... i Annual daily flow sequence Q ( i , t ), i Indicates the year sequence number. t It indicates the day number in a year.

3. The method for river baseflow mixing and segmentation based on dream optimization algorithm according to claim 1, characterized in that: In step S2, the empirical baseflow index of the river is calculated based on historical daily runoff data. The formula for calculating the empirical baseflow index is as follows: （1）； In formula (1): BFI target ( i ) indicates the first section of the river channel i The annual experience base current index; Q 90 ( i ) indicates the first i The annual guarantee rate is 90% for daily traffic. Q 50 ( i ) indicates the first i The annual guarantee rate is 50% for daily traffic.

4. The method for river baseflow mixing and segmentation based on dream optimization algorithm according to claim 1, characterized in that: In S3, the hybrid segmentation model for the river baseflow is a hybrid segmentation model composed of the Eckhart filter model and the Chapman filter model, with weight parameters... ω The base current sequences of the two models are weighted using the following formula: （2）； （3）； （4）； In equations (2), (3), and (4): α BFI represents the Eckhart attenuation coefficient. max This represents the long-term ratio of the base current to the total flow in the Eckhart filter model. k These are the Chapman filter parameters; ω The weighting is mixed to control the proportion of the two filters. ω ∈(0,1); This represents the basic current of the day in year i calculated using the Eckhart filter model; This represents the basic current of the day in year i calculated using the Chapman filter model.

5. The method for river baseflow mixing and segmentation based on dream optimization algorithm according to claim 1, characterized in that: The river baseflow mixing and segmentation model is based on the river channel's first... i Historical daily runoff data for years, used to calculate the river channel's first... i Theoretical baseflow index of the year BFI mix ( i The calculation formula is as follows: （5）； In formula (5): Q ( i , t Let be the daily runoff in year i; The first value obtained by the hybrid segmentation filtering model is represented by the first value. i Annual base flow.

6. The method for river baseflow mixing and segmentation based on dream optimization algorithm according to claim 5, characterized in that: In S4, the objective function for optimizing the parameters of the river baseflow mixing and segmentation model is: （6）； In formula (6): J The main objective function is to take the sum of squared annual errors between the empirical baseflow index and the theoretical baseflow index; I This represents the total number of years. This represents the empirical baseflow index of the river in year i.

7. A river baseflow mixing and segmentation method based on a dream optimization algorithm according to claim 1, characterized in that: In step S4, using the annual mean square error as the objective function, a mixed river baseflow mixing and segmentation method based on the dream optimization algorithm is employed to search for the optimal weights within the given parameter boundaries. ω、 parameter α、 BFI max , k The optimal parameters are obtained.

8. A river baseflow mixing and segmentation system based on a dream optimization algorithm, used to execute the river baseflow mixing and segmentation method based on a dream optimization algorithm as described in any one of claims 1-7, characterized in that, The system includes: The data acquisition module is used to obtain historical daily runoff data of the river channel; The data processing module is used to process historical daily runoff data of the river. The empirical baseflow index calculation module is used to calculate the empirical baseflow index of a river based on historical daily runoff data. The baseflow segmentation model construction module is used to construct a river baseflow mixing segmentation model and calculate the theoretical baseflow index of the river. The algorithm optimization module is used to optimize the parameters of the river baseflow mixing and segmentation model by using the dream optimization algorithm, with the goal of minimizing the sum of squared errors between the empirical baseflow index and the theoretical baseflow index. The baseflow segmentation execution module is used to perform baseflow segmentation of the river using the optimized river baseflow mixing segmentation model and output the baseflow segmentation results.

9. A computer device, characterized in that, It includes a memory and a processor, which are interconnected and communicate with each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the river baseflow mixing and segmentation method based on the dream optimization algorithm as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute any one of claims 1 to 7 of the method for mixing and separating river baseflow based on the dream optimization algorithm.

Images