A roughness joint calibration method and device, electronic equipment and storage medium

Abstract

This application provides a roughness joint calibration method, apparatus, electronic device, and storage medium. The method includes: constructing a hydrodynamic model; acquiring measured flow sequences for multiple periods at each cross-section; identifying the critical quantile position of each period in the flow quantile threshold curve where it transitions from a flat segment to an upward segment; determining the optimal low flow quantile with the highest synchronicity at the critical quantile position of each cross-section; constructing a low flow constraint requiring that the simulated total flow below the optimal low flow quantile is not lower than the measured total flow; determining the comprehensive weight of each cross-section based on the importance of its management function and the quality of the measured flow data, and constructing an objective function by combining the simulation accuracy of each cross-section under multiple periods; and searching for the optimal combination of roughness parameters that minimizes the objective function under the premise of satisfying the low flow constraint through an optimization algorithm, and outputting it as the calibration result. This method solves the problems of low simulation accuracy and failure to reflect the priority of key cross-sections during low flow periods.

Description

Technical Field

[0001] This application relates to the field of hydrodynamic model parameter calibration technology, and more specifically, to a roughness-rate joint calibration method, apparatus, electronic device, and storage medium. Background Technology

[0002] Roughness is a core parameter characterizing the resistance of the riverbed and boundary conditions to water flow. The accuracy of its value directly affects the simulation effect of hydrodynamic models on hydraulic elements such as water level and flow rate, and is a key foundation for ensuring the scientific nature of water conservancy project scheduling decisions.

[0003] In practical applications, if the roughness coefficient is too large, the hydrodynamic model will overestimate the riverbed resistance, resulting in a smaller flow capacity at the same water level; conversely, a smaller value will lead to a larger flow capacity. Most existing roughness coefficient calibration methods adopt a globally unified optimization strategy, lacking refined control for different hydrological conditions, and there is still room for improvement. Summary of the Invention

[0004] In view of this, embodiments of this application provide a roughness coefficient joint calibration method, apparatus, electronic device, and storage medium to solve the problems of low simulation accuracy during low flow periods and failure to reflect the priority of key sections in traditional roughness coefficient calibration methods.

[0005] In a first aspect, embodiments of this application provide a roughness-rate joint calibration method, the method comprising:

[0006] A hydrodynamic model containing multiple cross sections was constructed, and the measured flow sequences of each cross section under multiple scenarios were obtained.

[0007] For each hydrological event, based on the measured flow sequence of each cross section, the critical quantile position of each cross section in the flow quantile threshold curve is identified as the transition from the flat section to the rising section, and the quantile with the highest synchronicity at the critical quantile position of each cross section is taken as the unified optimal low flow quantile for that event.

[0008] Based on the measured flow threshold corresponding to the optimal low flow quantile, the measured total flow of each section below the measured flow threshold is used as the low flow guarantee benchmark, and a low flow constraint is constructed that requires the simulated total flow to be no less than the low flow guarantee benchmark.

[0009] Based on the importance of the management function of each section and the quality of the measured flow data, the comprehensive weight of each section is determined, and an objective function is constructed based on the comprehensive weight and the simulation accuracy of each section in multiple scenarios.

[0010] The low flow constraint and the objective function are imported into the optimization algorithm. Under the premise of satisfying the low flow constraint, the optimal combination of roughness parameters that minimizes the objective function is searched and output as the calibration result.

[0011] In a feasible implementation, for each hydrological event, based on the measured flow sequence of each cross-section, the critical quantile position of each cross-section in the flow quantile threshold curve transitioning from a flat segment to an upward segment is identified, and the quantile with the highest synchronicity at the critical quantile position of each cross-section is taken as the unified optimal low flow quantile for that event, including:

[0012] Construct a candidate set of low quantile parameters for flow; the candidate set contains multiple candidate quantiles for characterizing the level of ecological baseflow.

[0013] For each hydrological event, the measured flow sequences of each cross section are sorted in ascending order of value;

[0014] Traverse each candidate quantile in the candidate set to determine the target measured threshold corresponding to each section under the candidate quantile;

[0015] The target measured thresholds for each cross section at each candidate quantile are dimensionless, and the first and second derivatives of the dimensionless thresholds with respect to the quantiles are extracted. The first derivative is used to identify the stage of stable flow change, and the second derivative is used to identify the critical quantile position where the flow changes from stable to rising in the stable stage.

[0016] A global scaling function is constructed based on the first and second derivatives extracted from each cross section. The optimal low flow quantile for this event is determined by solving for the candidate quantiles that make the global scaling function reach its maximum value. The global scaling function is used to characterize the synchronicity of flow changes at each cross section, and its function value is positively correlated with the synchronicity.

[0017] In a feasible implementation, the quality of the measured traffic data is determined based on the completeness and stability of the measured traffic sequence. The higher the quality of the measured traffic data, the greater the comprehensive weight. In this case, completeness is characterized by the data missing rate, and stability is characterized by the proportion of outliers. The quality of the measured traffic data is negatively correlated with the missing rate and the proportion of outliers.

[0018] In a feasible implementation, the importance of the management function is divided into several levels according to the cross-section type, including: a high importance level corresponding to major flood control safety cross-sections, a medium importance level corresponding to key ecological nodes or hydraulic control nodes, and a low importance level corresponding to ordinary reference cross-sections; the higher the level of the importance of the management function, the greater the corresponding comprehensive weight.

[0019] In one feasible implementation, the configuration parameters of the optimization algorithm include at least one of the following:

[0020] Roughness parameter range, initial value, number of iterations, search step size, step size decay factor, and constraint violation tolerance.

[0021] In a feasible implementation, the step of searching for the optimal combination of roughness parameters that minimizes the objective function while satisfying the low flow constraint includes:

[0022] Starting from the current roughness parameter value, the probe is performed according to the search step size set in the configuration parameters. If a parameter value that makes the objective function smaller and satisfies the low flow constraint is detected, the probe moves step by step in the direction of the parameter value change.

[0023] In a feasible implementation, based on the comprehensive weights and the simulation accuracy of each section in multiple scenarios, an objective function is constructed, including:

[0024] Based on the comprehensive weight of each monitoring section, the weighted average of the Nash efficiency coefficient of each section under all hydrological events is calculated.

[0025] The objective function is obtained based on the difference between 1 and the weighted average.

[0026] Secondly, embodiments of this application also provide a roughness coefficient calibration apparatus, the apparatus comprising:

[0027] The model module is used to construct a hydrodynamic model containing multiple cross sections and obtain the measured flow sequence of each cross section under multiple scenarios.

[0028] The quantile determination module is used to identify the critical quantile position of each section in the flow quantile threshold curve from the flat section to the rising section for each hydrological event, based on the measured flow sequence of each section, and to take the quantile with the highest synchronicity at the critical quantile position of each section as the unified optimal low flow quantile for that event.

[0029] The first construction module is used to construct a low flow constraint that requires the simulated total flow to be no less than the low flow guarantee benchmark, based on the measured flow threshold corresponding to the optimal low flow quantile and using the measured total flow of each section below the measured flow threshold as the low flow guarantee benchmark.

[0030] The second construction module is used to determine the comprehensive weight of each section based on the importance of the management function of each section and the quality of the measured flow data, and to construct an objective function based on the comprehensive weight and the simulation accuracy of each section in multiple scenarios.

[0031] The calibration module is used to import the low flow constraint and the objective function into the optimization algorithm, search for the optimal combination of roughness parameters that minimizes the objective function under the premise of satisfying the low flow constraint, and output the calibration result.

[0032] In one feasible implementation, the quantile determination module is used, for each hydrological event, to identify, based on the measured flow sequence of each cross-section, the critical quantile position of each cross-section transitioning from a flat segment to an upward segment in the flow quantile threshold curve, and to use the quantile with the highest synchronicity at the critical quantile position of each cross-section as the unified optimal low flow quantile for that event, for the following purposes:

[0033] Construct a candidate set of low quantile parameters for flow; the candidate set contains multiple candidate quantiles for characterizing the level of ecological baseflow.

[0034] For each hydrological event, the measured flow sequences of each cross section are sorted in ascending order of value;

[0035] Traverse each candidate quantile in the candidate set to determine the target measured threshold corresponding to each section under the candidate quantile;

[0036] The target measured thresholds for each cross section at each candidate quantile are dimensionless, and the first and second derivatives of the dimensionless thresholds with respect to the quantiles are extracted. The first derivative is used to identify the stage of stable flow change, and the second derivative is used to identify the critical quantile position where the flow changes from stable to rising in the stable stage.

[0037] A global scaling function is constructed based on the first and second derivatives extracted from each cross section. The optimal low flow quantile for this event is determined by solving for the candidate quantiles that make the global scaling function reach its maximum value. The global scaling function is used to characterize the synchronicity of flow changes at each cross section, and its function value is positively correlated with the synchronicity.

[0038] In a feasible implementation, the quality of the measured traffic data is determined based on the completeness and stability of the measured traffic sequence. The higher the quality of the measured traffic data, the greater the comprehensive weight. In this case, completeness is characterized by the data missing rate, and stability is characterized by the proportion of outliers. The quality of the measured traffic data is negatively correlated with the missing rate and the proportion of outliers.

[0039] In a feasible implementation, the importance of the management function is divided into several levels according to the cross-section type, including: a high importance level corresponding to major flood control safety cross-sections, a medium importance level corresponding to key ecological nodes or hydraulic control nodes, and a low importance level corresponding to ordinary reference cross-sections; the higher the level of the importance of the management function, the greater the corresponding comprehensive weight.

[0040] In one feasible implementation, the configuration parameters of the optimization algorithm include at least one of the following:

[0041] Roughness parameter range, initial value, number of iterations, search step size, step size decay factor, and constraint violation tolerance.

[0042] In a feasible implementation, the calibration module is used to search for the optimal combination of roughness parameters that minimizes the objective function while satisfying the low flow constraint, and is used to:

[0043] Starting from the current roughness parameter value, the probe is performed according to the search step size set in the configuration parameters. If a parameter value that makes the objective function smaller and satisfies the low flow constraint is detected, the probe moves step by step in the direction of the parameter value change.

[0044] In a feasible implementation, the second construction module is used to construct an objective function based on the comprehensive weights and the simulation accuracy of each section in multiple scenarios, for:

[0045] Based on the comprehensive weight of each monitoring section, the weighted average of the Nash efficiency coefficient of each section under all hydrological events is calculated.

[0046] The objective function is obtained based on the difference between 1 and the weighted average.

[0047] Thirdly, embodiments of this application also provide an electronic device, including: a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the roughness joint calibration method as described in any one of the first aspects.

[0048] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the method as described in any one of the first aspects.

[0049] This application provides a roughness coefficient joint calibration method, apparatus, electronic device, and storage medium. By identifying the critical quantile positions of each cross-section in the flow quantile threshold curve where the flow transitions from a flat segment to an upward segment, and taking the quantile with the highest synchronicity at each cross-section's critical quantile position as the optimal low flow quantile for the entire river section, and constructing an objective function by combining cross-section comprehensive weights, and applying a low total flow constraint to the optimization algorithm for solution, the following beneficial effects are achieved:

[0050] To address the issue of insufficient simulation accuracy during low-flow periods, this application uses the measured total flow rate of each cross-section below the optimal low-flow quantile as the low-flow guarantee benchmark. It constructs a low-flow constraint requiring the simulated total flow rate to be no less than this benchmark. This ensures that the optimization process must take into account the overall water volume performance during low-flow periods, avoiding the problem of traditional methods where simulation deviations in low-flow segments are ignored due to the dominance of high-flow data. This improves the simulation reliability of the model during the dry season.

[0051] To address the issue of the inability to prioritize key cross-sections, this application determines a comprehensive weight based on the importance of each cross-section's management function and the quality of measured flow data. Based on this weight, an objective function is constructed, ensuring that key locations such as important urban cross-sections and ecologically sensitive areas receive greater attention during the calibration process, thus achieving differentiated optimization under multi-objective management requirements.

[0052] Compared with existing technologies, this application achieves optimal synchronization of the critical quantile positions of each section in the flow quantile threshold curve when transitioning from the flat segment to the rising segment through a unified optimal low flow quantile identification mechanism. This ensures the simulation accuracy during low flow periods while taking into account the priority of key sections, effectively solving the application limitations of traditional roughness calibration methods in multi-objective management scenarios.

[0053] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0054] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0055] Figure 1 A flowchart of a roughness-rate joint calibration method provided in an embodiment of this application is shown.

[0056] Figure 2 A flowchart of another roughness joint calibration method provided by an embodiment of this application is shown.

[0057] Figure 3 A schematic diagram of the structure of a roughness joint calibration device provided in an embodiment of this application is shown.

[0058] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0060] Roughness is a core parameter characterizing the resistance of the riverbed and boundary conditions to water flow. The accuracy of its value directly affects the simulation effect of hydrodynamic models on hydraulic elements such as water level and flow rate, and is a key foundation for ensuring the scientific nature of water conservancy project scheduling decisions.

[0061] In practical applications, if the roughness coefficient is too large, the model will overestimate the riverbed resistance, resulting in a smaller flow capacity at the same water level; conversely, a smaller value will result in a larger flow capacity. Most existing roughness coefficient calibration methods adopt a globally unified optimization strategy, lacking fine-grained control for different water conditions (especially low flow ranges), and there is still room for improvement.

[0062] Based on this, embodiments of this application provide a roughness and co-calibration method, apparatus, electronic device, and storage medium, which are described below through embodiments.

[0063] To facilitate understanding of this embodiment, a roughness-rate joint calibration method disclosed in this application will first be described in detail. For example... Figure 1 As shown, it includes the following steps:

[0064] Step 101: Construct a hydrodynamic model containing multiple cross sections and obtain the measured flow sequence of each cross section under multiple scenarios.

[0065] First, a one-dimensional hydrodynamic model is constructed, covering the river section under study and including multiple computational cross-sections. These cross-sections correspond to actual monitoring points and are used to simulate the changes in hydraulic elements of the flow at different locations. Simultaneously, measured flow data from these monitoring cross-sections are acquired over multiple hydrological events, each corresponding to an independent flood event or a specific time-series observation sequence. This measured data will serve as the benchmark for subsequent calibration work, used to evaluate the accuracy of the model simulation. Model construction and measured data acquisition are fundamental to the calibration work, providing input for subsequent constraint construction, weight configuration, and parameter optimization.

[0066] Of course, in practical applications, the number of monitoring sections and the selection of hydrological sessions can be flexibly adjusted according to the river network characteristics, data availability, and management objectives of the specific study area. For example, for river sections with long main streams or many tributaries, more representative key sections can be selected for calibration; while for areas with abundant hydrological data, more historical sessions can be selected to improve the robustness of the calibration results.

[0067] Step 102: For each hydrological event, based on the measured flow sequence of each cross section, identify the critical quantile position of each cross section in the flow quantile threshold curve where the transition is from a flat segment to an upward segment, and take the quantile with the highest synchronicity at the critical quantile position of each cross section as the unified optimal low flow quantile for that event.

[0068] The "critical quantile" mentioned here refers to the turning point in the flow distribution where a relatively stable baseflow state begins to rise significantly. In the embodiments of this application, the measured flow series is broken down from the time series and sorted by flow rate from smallest to largest. The low-value, flat area of ​​the flow curve represents the baseflow-dominated state, while the steep increase area represents the flood state. The transition point between the two is of particular significance for identifying the location of ecological baseflow protection—the flow level corresponding to this location can be regarded as the threshold starting point for the river's hydrodynamics to become "active".

[0069] A flow quantile threshold curve is a curve formed by sorting the measured flow series in ascending order of values. The horizontal axis represents the quantile, and the vertical axis represents the corresponding flow value. This curve reflects the overall shape of flow variation with quantiles, where the flat section corresponds to baseflow, the steep rise section corresponds to rising water levels, and the transition section corresponds to the transition region between the two.

[0070] For the same hydrological event, the critical quantile positions at different cross sections may correspond to different quantiles. This step analyzes the synchronicity of the critical quantile positions at each cross section and selects the quantile with the highest synchronicity as the unified optimal low-discharge quantile for that event. The purpose of this is to ensure that all cross sections within the same event use the same quantile as the benchmark position for low-discharge assurance, which facilitates the subsequent construction of unified constraints and reflects the overall hydrological characteristics of the entire river section.

[0071] In practical applications, the identification of critical quantile locations and the analysis of synchronicity can be achieved in various ways. As long as the transition point from base flow to rising water can be accurately located and synchronicity comparisons can be made between multiple cross sections, the purpose of this step can be achieved.

[0072] Step 103: Based on the measured flow threshold corresponding to the optimal low flow quantile, the measured total flow of each section below the measured flow threshold is used as the low flow guarantee benchmark to construct a low flow constraint that requires the simulated total flow to be no less than the low flow guarantee benchmark.

[0073] For each cross-section, all measured flow values ​​below the optimal low flow quantile are summed to obtain a total measured flow. This value is the low flow guarantee benchmark for that cross-section. This benchmark represents the total flow below a certain flow level during a hydrological event, and can be understood as the ecological base flow floor for that cross-section.

[0074] Subsequently, a constraint is constructed for each cross-section, requiring that the sum of the flow values ​​simulated by the hydrodynamic model at that cross-section, which are also below the optimal low flow quantile, cannot be lower than this guarantee benchmark. In other words, the total flow during the low flow period simulated by the model must at least reach the level of the measured total flow.

[0075] The significance of this constraint lies in the fact that it does not require the model to simulate accurately at every single time point, but rather that the model can guarantee water supply during periods of low flow overall. This total quantity constraint is more robust than single-point constraints, and can avoid the problem of constraints being too tight or too loose due to accidental fluctuations at individual time points.

[0076] In practical applications, independent relaxation factors can be set for each cross-section as needed, allowing the simulated total flow rate to be lower than the measured total flow rate within a certain range (the error range set by the relaxation factor) to address potential measurement errors or random factors in the measured data. The value of the relaxation factor can be adjusted based on the data quality of the cross-section or other factors.

[0077] Step 104: Determine the comprehensive weight of each cross-section based on the importance of its management function and the quality of the measured flow data, and construct an objective function based on the comprehensive weight and the simulation accuracy of each cross-section in multiple scenarios.

[0078] The importance of different monitoring sections often varies in actual management. Some sections are located in important urban areas or dangerous sections of dikes, directly related to flood control safety; some sections are key ecological assessment nodes, of great significance to ensuring ecological flow; and some sections may only be ordinary reference sections with relatively low management priority. In order to reflect this difference in the calibration process, it is necessary to assign a weight basis to each section that matches the importance of its management function.

[0079] Meanwhile, the quality of the measured flow data also affects the reliability of the calibration results. If there are many missing or anomalies in the measured data of a certain section, the contribution of the data from that section to the calibration should be reduced accordingly to avoid low-quality data misleading the optimization direction. Therefore, it is necessary to adjust the weight of each section based on the completeness and stability of the measured flow sequence. Completeness can be evaluated by the degree of data missingness, while stability can be evaluated by the amount of anomalies.

[0080] By combining the importance of management functions with the quality of measured traffic data, a comprehensive weight can be obtained for each cross-section. This comprehensive weight reflects both the management status of the cross-section and the reliability of the data itself, enabling subsequent optimization to achieve the effect of "prioritizing important cross-sections and downgrading low-quality data".

[0081] After obtaining the comprehensive weights for each cross-section, a target function reflecting the overall simulation accuracy needs to be constructed. This target function weights and synthesizes the simulation accuracy indicators for each cross-section under each hydrological field. Simulation accuracy indicators can be commonly used evaluation metrics in hydrology, such as the Nash efficiency coefficient; the closer this coefficient is to 1, the better the simulation effect. By weighting and averaging the simulation accuracy indicators for each cross-section and each field according to the comprehensive weights of the corresponding cross-sections, a comprehensive target function can be obtained. The value of this target function can comprehensively evaluate the overall performance of the model across all key cross-sections and multiple fields.

[0082] In practical applications, the classification of the importance of management functions, the evaluation methods for data quality, and the selection of simulation accuracy indicators can all be adjusted according to specific circumstances. For example, cross-sections can be divided into more levels based on the characteristics of the watershed, different statistical indicators can be used to evaluate data quality, or other accuracy evaluation indicators can be selected. As long as the core idea of ​​"assigning weights according to importance and adjusting weights according to data quality" is reflected, the purpose of this step can be achieved.

[0083] Step 105: Import the low flow constraint and the objective function into the optimization algorithm, and search for the optimal combination of roughness parameters that minimizes the objective function while satisfying the low flow constraint, and output it as the calibration result.

[0084] Optimization algorithms are the core tools for solving parametric optimization problems. The optimization algorithm used in this step needs to be able to handle constraints and simultaneously consider the optimization direction of the objective function and the satisfaction of constraints during the parameter search process. The algorithm's input includes configuration parameters such as the search range of the roughness parameter, initial value, number of iterations, and search step size, as well as the low-flow constraints and objective function constructed in the previous steps.

[0085] The optimization process begins with a set of initial roughness parameters. By continuously adjusting the parameter combinations, the algorithm calculates the objective function value and constraint satisfaction after each adjustment, progressively searching in the direction that reduces the objective function without violating the constraints. After each iteration, the algorithm evaluates whether the current parameter combination satisfies the low flow constraints for all cross-sections and all scenarios. Only if the constraints are satisfied will the algorithm accept parameter adjustments that reduce the objective function.

[0086] When the algorithm meets the convergence condition, such as the rate of change of the objective function value being less than a preset threshold multiple times consecutively, and all constraints being strictly satisfied, the iteration stops, and the currently found optimal combination of roughness parameters is output. This parameter combination is the final calibration result, which can be used for subsequent hydrodynamic simulations and engineering applications.

[0087] In this way, the low-flow guarantee requirement is incorporated into the optimization process, ensuring that the final roughness parameters meet the simulation accuracy requirements during the dry season. In practical applications, various optimization algorithms with constraint handling capabilities can be used, such as pattern search algorithms, genetic algorithms, and particle swarm optimization algorithms. Any algorithm capable of parameter optimization under constraints can be used in this step. The specific parameter settings of the algorithm can also be adjusted according to the complexity of the problem and computational resources to achieve better optimization results.

[0088] This application provides a roughness coefficient joint calibration method, apparatus, electronic device, and storage medium. By identifying the critical quantile positions of each cross-section in the flow quantile threshold curve where the flow transitions from a flat segment to an upward segment, and taking the quantile with the highest synchronicity at each cross-section's critical quantile position as the optimal low flow quantile for the entire river section, and constructing an objective function by combining cross-section comprehensive weights, and applying a low total flow constraint to the optimization algorithm for solution, the following beneficial effects are achieved:

[0089] To address the issue of insufficient simulation accuracy during low-flow periods, this application uses the measured total flow rate of each cross-section below the optimal low-flow quantile as the low-flow guarantee benchmark. It constructs a low-flow constraint requiring the simulated total flow rate to be no less than this benchmark. This ensures that the optimization process must take into account the overall water volume performance during low-flow periods, avoiding the problem of traditional methods where simulation deviations in low-flow segments are ignored due to the dominance of high-flow data. This improves the simulation reliability of the model during the dry season.

[0090] To address the issue of the inability to prioritize key cross-sections, this application determines a comprehensive weight based on the importance of each cross-section's management function and the quality of measured flow data. Based on this weight, an objective function is constructed, ensuring that key locations such as important urban cross-sections and ecologically sensitive areas receive greater attention during the calibration process, thus achieving differentiated optimization under multi-objective management requirements.

[0091] Compared with existing technologies, this application achieves optimal synchronization of the critical quantile positions of each section in the flow quantile threshold curve when transitioning from the flat segment to the rising segment through a unified optimal low flow quantile identification mechanism. This ensures the simulation accuracy during low flow periods while taking into account the priority of key sections, effectively solving the application limitations of traditional roughness calibration methods in multi-objective management scenarios.

[0092] In a feasible implementation plan, such as Figure 2As shown, for each hydrological event, based on the measured flow sequence of each cross-section, the critical quantile position of each cross-section in the flow quantile threshold curve transitioning from a flat segment to an upward segment is identified, and the quantile with the highest synchronicity at the critical quantile position of each cross-section is taken as the unified optimal low flow quantile for that event. This includes the following steps:

[0093] Step 201: Construct a candidate set of low quantile parameters for flow; the candidate set contains multiple candidate quantiles used to characterize the level of ecological baseflow.

[0094] Step 202: For each hydrological event, sort the measured flow sequence of each cross section in ascending order of value.

[0095] Step 203: Traverse each candidate quantile in the candidate set and determine the target measured threshold corresponding to each cross section under the candidate quantile.

[0096] Step 204: Dimensionless processing is performed on the target measured thresholds of each cross section at each candidate quantile, and the first and second derivatives of the dimensionless thresholds with respect to the quantiles are extracted; wherein, the first derivative is used to identify the stage of stable flow change, and the second derivative is used to identify the critical quantile position where the flow changes from stable to rising in the stable stage.

[0097] Step 205: Construct a global scaling function based on the first and second derivatives extracted from each cross section. Determine the optimal low flow quantile for this event by solving for the candidate quantiles that make the global scaling function reach its maximum value. The global scaling function is used to characterize the synchronicity of flow changes at each cross section, and its function value is positively correlated with the synchronicity.

[0098] In other words, firstly, a candidate set of low quantile parameters for flow is constructed, which contains multiple candidate quantiles for characterizing the level of ecological baseflow. These candidate quantiles can be set according to actual needs, such as covering the range from extreme low water to general low water, and the step size can be selected according to the accuracy requirements.

[0099] Next, for each hydrological event, the measured flow sequences at each cross-section were sorted in ascending order of value. This was done to break the temporal order and convert the flow data into a flow duration curve, facilitating subsequent analysis of the flow variation with quantiles.

[0100] Then, the candidate quantiles in the candidate set are traversed. For each candidate quantile, the target measured threshold corresponding to that quantile for each cross-section is determined. This threshold is the flow value taken from the sorted sequence at that quantile position.

[0101] After obtaining the target measured thresholds for each cross-section at each candidate quantile, these thresholds are dimensionless to eliminate the magnitude effect caused by differences in absolute flow rates between different cross-sections. Dimensionlessness can be achieved in various ways, such as dividing by the median or average flow rate of each cross-section. Subsequently, the first and second derivatives of the dimensionless thresholds with respect to the quantiles are extracted. The first derivative is used to identify relatively stable phases of flow rate change; the second derivative is used to further pinpoint the point where the flow rate begins to rise from a stable state within the stable phase, i.e., the critical quantile position.

[0102] Finally, a global scaling function is constructed based on the first and second derivatives extracted from each cross-section. This function iterates through all candidate quantiles, calculating the function value for each quantile. The magnitude of this function value reflects the degree of synchronization of flow changes across cross-sections at that quantile; a larger function value indicates more synchronized critical quantile positions across cross-sections. The candidate quantile that maximizes the global scaling function is selected as the unified optimal low flow quantile for that event.

[0103] The formula for the global scaling function is given below:

[0104]

[0105] in, This represents the k-th quantile in the candidate set. express The corresponding global scaling function value, where n is the total number of cross sections in the current session, and s is the s-th monitoring cross section from 1 to n.

[0106] Is the section s in the quantile? Dimensionless threshold (within quantiles) The first derivative of the target measured threshold obtained after dimensionless processing.

[0107] Is the section s in the quantile? Dimensionless threshold (within quantiles) The second derivative of the target measured threshold (obtained after dimensionless processing).

[0108] The slope penalty constant is a preset coefficient used to adjust the degree of penalty imposed on the result by the first derivative.

[0109] exp() is an exponential function with base e, used to convert the first derivative into a penalty factor, where: This indicates that the penalty factor has a negative exponential relationship with the absolute value of the first derivative: the smaller the absolute value of the first derivative, the closer the penalty factor is to 1; the larger the absolute value of the first derivative, the closer the penalty factor is to 0.

[0110] This global scaling function indicates that, for each candidate quantile, the sum of the absolute values ​​of the second derivatives of all cross-sections multiplied by a penalty factor related to the first derivative is the most suitable quantile to serve as the unified optimal low-flow quantile. This optimal low-flow quantile is denoted as... In other words, this global scaling function is looking for a quantile at which all cross-sections are exactly at the critical point of "rising from the steady base current" and have the best synchronization.

[0111] In practical applications, the range and step size of candidate quantiles can be adjusted according to the hydrological characteristics of the watershed. The dimensionless method can also be selected according to the data characteristics. The calculation of the first and second derivatives can be carried out using numerical difference or other approximation methods. As long as the core objective of identifying the critical quantile positions from each cross section and selecting the quantile with the highest synchronicity can be achieved, it can be regarded as an equivalent implementation of this scheme.

[0112] Here, we will illustrate the example with a specific numerical value. Suppose that a flood event in a certain watershed includes 5 monitoring sections, the low flow quantile is set to be between 0.05 and 0.20, and the search step size is 0.01. Then, the constructed candidate quantile set contains 16 candidate values, including 0.05, 0.06, 0.07, ... up to 0.20.

[0113] Taking a specific cross-section as an example, the measured flow sequence for this event has 1000 data points. After sorting the 1000 flow values ​​in ascending order, for a candidate quantile of 0.05, the 50th flow value after sorting is taken as the measured threshold for that cross-section at the 0.05 quantile. Similarly, for the 0.10 quantile, the 100th flow value is taken; for the 0.15 quantile, the 150th flow value is taken, and so on.

[0114] After extracting the measured threshold at each candidate quantile for each cross section, dimensionless processing is performed. For example, if the median of the measured flow sequence at this cross section is 50 cubic meters per second, the measured threshold extracted at the 0.05 quantile is 2.5 cubic meters per second, which becomes 0.05 after dimensionless processing; the measured threshold extracted at the 0.10 quantile is 5 cubic meters per second, which becomes 0.10 after dimensionless processing.

[0115] Subsequently, the first derivative of the dimensionless threshold with respect to the quantiles is calculated. Taking the vicinity of the 0.10 quantile as an example, if the dimensionless value corresponding to the 0.09 quantile is 0.095, the dimensionless value corresponding to the 0.10 quantile is 0.10, and the dimensionless value corresponding to the 0.11 quantile is 0.12, then the first derivative at the 0.10 quantile can be approximated as (0.12-0.095) / (0.11-0.09) = 0.025 / 0.02 = 1.25. The smaller the value of the first derivative, the slower the rate at which the flow rate increases with the quantile in that region, corresponding to a stable phase.

[0116] The second derivative reflects the rate of change of the first derivative. If the first derivative at the 0.09 quantile is 0.8, at the 0.10 quantile it is 1.25, and at the 0.11 quantile it is 2.0, then the second derivative at the 0.10 quantile can be approximated as (2.0-0.8) / (0.11-0.09) = 1.2 / 0.02 = 60. A larger second derivative value indicates a sharp increase in the rate of flow increase near that point, representing the critical quantile position where the flow transitions from a steady base flow to an increasing flow.

[0117] After calculating the first and second derivatives for all cross-sections at each candidate quantile, these derivatives are substituted into the global scaling function. For example, at q=0.12, the sum of the calculated function values ​​of the first and second derivatives for each cross-section is obtained. The sum is 25.6; the sum is 23.8 when q=0.13; the sum is 25.6 when q=0.11. The value is 24.2. By comparison, the function value when q=0.12 is... The maximum value is found, therefore the optimal low flow quantile for this session is determined to be 0.12 (optimal low flow quantile).

[0118] The significance of this 0.12 quantile lies in the fact that, for this flood event, the five cross-sections across the entire river segment achieved optimal spatial synchronicity at the critical quantile position of the transition from baseflow to increased discharge. Subsequent work will use this unified quantile to construct low-discharge constraints for each cross-section.

[0119] Here's an illustrative explanation of the low flow constraint: First, determine the cutoff position for each cross-section based on the optimal low flow quantile. After sorting the measured flow sequence of the cross-section in ascending order, take the flow value located at the optimal low flow quantile as the cutoff point. All flow values ​​before this cutoff point constitute the low flow interval for that cross-section.

[0120] Then, the measured total flow at each cross-section within the low-flow range is calculated—that is, all measured flow values ​​before this dividing point are added together to obtain a sum. This sum is the low-flow guarantee benchmark for that cross-section, representing how much water flowed through those small flows below the optimal low-flow quantile during this flood.

[0121] Similarly, the simulated flow sequence is sorted by value from smallest to largest, and the sum of all simulated flow values ​​before the same boundary point is calculated to obtain the simulated total flow.

[0122] Next, a constraint is constructed for each cross-section: the simulated total flow rate is subtracted from the measured total flow rate, and then a relaxation factor is subtracted, requiring the result to be less than or equal to zero. In other words, the simulated total flow rate cannot be lower than the measured total flow rate minus the relaxation factor.

[0123] This relaxation amount is set independently for each cross-section, initially set to zero. If a cross-section consistently fails to meet the constraints during optimization, it indicates that there may be special cases in the data or model for that cross-section. In this case, the relaxation amount for that cross-section can be appropriately increased to provide some "error tolerance" and allow the optimization process to continue. However, the relaxation amount cannot be increased indefinitely; usually, an upper limit is set, such as not exceeding 10% of the measured total flow. If the constraints still cannot be met even after adjusting to the upper limit, it indicates that there may be a problem with that cross-section, requiring separate handling.

[0124] The advantage of this low-flow constraint on the total volume is that it focuses on the total flow during the entire low-flow period, rather than focusing on the flow value at a single point. Even if the simulated values ​​at individual time points deviate slightly, as long as the overall flow is roughly the same, the constraint is considered satisfied, which better meets the needs of practical applications.

[0125] This application identifies the critical quantile positions of each cross-section using first and second derivatives, and constructs a global scaling function to compare the synchronicity of critical characteristics of all cross-sections at the same quantile. The quantile with the highest synchronicity is taken as the unified optimal low-flow quantile. This ensures that all cross-sections in the same event use the same low-flow guarantee reference position, avoiding the physical inconsistency problem caused by independent point selection for each cross-section in traditional methods, and providing an objective basis for subsequently constructing unified low-flow constraints.

[0126] In a feasible implementation, the importance of the management function is divided into several levels according to the cross-section type, including: a high importance level corresponding to major flood control safety cross-sections, a medium importance level corresponding to key ecological nodes or hydraulic control nodes, and a low importance level corresponding to ordinary reference cross-sections; the higher the level of the importance of the management function, the greater the corresponding comprehensive weight.

[0127] The importance of different monitoring sections often varies in actual management. Some sections are located in important urban areas or dangerous sections of dikes, directly related to flood control safety; some sections are key ecological assessment nodes or tributary confluences, important for ensuring ecological flow; and some sections may only be ordinary reference sections with relatively low management priority. To reflect these differences in the calibration process, a basic weight needs to be assigned to each section, where the weight corresponding to the management function importance of the s-th section is expressed as: (Also known as the base weight).

[0128] Sections involving significant flood control safety, such as urban sections and sections with critical dike sections, are assigned higher weights, corresponding to a high-weight level. Sections of medium importance, such as key ecological assessment nodes or hydraulic control nodes like tributary confluences, are assigned medium weights, corresponding to a medium-weight level. Sections of low importance, such as ordinary tributary sections or reference sections, are assigned relatively lower weights, corresponding to a low-weight level. From high to low weight, the weight values ​​decrease sequentially, thus reflecting the different management priorities of different sections in the subsequent objective function. This hierarchical setting method ensures that key sections receive more attention during the calibration process and provides a clear basis for practical operation.

[0129] In a feasible implementation, the quality of the measured traffic data is determined based on the completeness and stability of the measured traffic sequence. The higher the quality of the measured traffic data, the greater the comprehensive weight. In this case, completeness is characterized by the data missing rate, and stability is characterized by the proportion of outliers. The quality of the measured traffic data is negatively correlated with the missing rate and the proportion of outliers.

[0130] Completeness refers to the degree of continuity of measured flow data over time. If the observation equipment at a certain cross-section malfunctions or data acquisition is interrupted, data gaps will occur. The more data gaps, the more incomplete the measured flow sequence at that cross-section, and the less reliable the statistical characteristics and simulation accuracy assessments derived from it. Therefore, completeness can be quantified by the data missing rate; the lower the missing rate, the more complete the data and the better the data quality.

[0131] Stability refers to the numerical reliability of measured flow data. If the observation equipment at a certain cross-section is interfered with or malfunctions, it may cause abrupt changes in the flow record that significantly deviate from the normal range. These outliers do not represent the actual hydrological process, and if used directly for calibration, they may mislead the optimization direction. Therefore, stability can be quantified by the proportion of outliers; the lower the proportion of outliers, the more stable the data and the better the data quality.

[0132] By combining the aspects of integrity and stability, an evaluation value for the quality of the measured flow data at each cross-section can be obtained. express. The higher the value, the more reliable the measured data of the s-th cross-section, and the higher its weight should be given in the calibration process; conversely, the lower the value, the more reliable the measured data of the s-th cross-section. The lower the weight, the less reliable the measured data of the s-th cross-section is, and it should be assigned a lower weight to avoid the adverse effects of low-quality data on the optimization results. This mechanism reflects the idea of ​​"reducing the weight of low-quality data," which helps to improve the robustness of the calibration results.

[0133] In practical applications, the missing rate can be calculated using conventional statistical methods, and outlier identification can be performed using thresholding, sliding window methods, or other detection algorithms. Any method that can objectively reflect the completeness and stability of the data can be used in this implementation plan.

[0134] In one feasible embodiment, the comprehensive weight of each cross-section can be obtained by multiplying the weight corresponding to the importance of the management function by the quality of the measured traffic data. This can be expressed as:

[0135]

[0136] This comprehensive weighting reflects both the importance of cross-section management and the reliability of the data itself. In practical applications, after configuring the weighting coefficients, normalization verification can be performed to ensure that the sum of the weights of each cross-section matches the number of cross-sections.

[0137] After obtaining the comprehensive weights of each section, it is necessary to construct an objective function that can reflect the overall simulation accuracy.

[0138] For example, based on the comprehensive weights and the simulation accuracy of each section in multiple scenarios, an objective function is constructed, including:

[0139] Based on the comprehensive weight of each monitoring section, the weighted average of the Nash efficiency coefficient of each section under all hydrological events is calculated.

[0140] The objective function is obtained based on the difference between 1 and the weighted average.

[0141] That is, the Nash efficiency coefficient (NSE), commonly used in hydrology, is adopted as the evaluation index for the accuracy of cross-section simulations; the closer the NSE is to 1, the better the simulation effect. To integrate the simulation accuracy of multiple cross-sections and multiple simulations, a weighted average NSE, denoted by F, can be constructed as the core objective function. The mathematical expression of this objective function can be written as:

[0142]

[0143] In this expression, n represents the total number of monitoring sections, and m represents the number of hydrological events. It is the first s The comprehensive weight of each cross section, It is the s-th cross section. The Nash efficiency coefficient for each game. In the denominator... This method serves as a normalization mechanism, ensuring that the objective function value remains within a reasonable range. Through this weighted averaging, the objective function F comprehensively evaluates the model's overall performance across all key sections and multiple simulations; a smaller F-value indicates a better overall simulation effect.

[0144] In one feasible implementation, the configuration parameters of the optimization algorithm include at least one of the following:

[0145] Roughness parameter range, initial value, number of iterations, search step size, step size decay factor, and constraint violation tolerance.

[0146] At this point, the search for the optimal combination of roughness parameters that minimizes the objective function while satisfying the low flow constraint includes:

[0147] Starting from the current roughness parameter value, the probe is performed according to the search step size set in the configuration parameters. If a parameter value that makes the objective function smaller and satisfies the low flow constraint is detected, the probe moves step by step in the direction of the parameter value change.

[0148] The roughness parameter range is used to limit the search space of the parameters, avoiding the algorithm from searching for physically unreasonable roughness values; the initial value refers to the starting point of the roughness parameter used when the algorithm starts searching; the number of iterations controls the number of search rounds of the algorithm; the search step size determines the magnitude of parameter adjustment during each probe; the step size decay factor is used to gradually reduce the step size during the search process, realizing a coarse-to-fine search strategy; the constraint violation tolerance is used to determine whether the parameter combination satisfies the low flow constraint, and the constraint is considered satisfied when the constraint violation amount is less than the tolerance.

[0149] Based on these configuration parameters, the optimization algorithm searches for the optimal combination of roughness parameters through the following steps:

[0150] Starting with the current roughness parameter value, the probe is performed in the parameter space according to the set search step size. The probe is usually performed along each parameter dimension, for example, moving a step size in both the positive and negative directions of each roughness parameter to obtain several new candidate parameter combinations.

[0151] For each new parameter combination detected, the algorithm performs two evaluations: first, it calculates the objective function value under this parameter combination and determines whether it is smaller than the objective function value at the current starting point; second, it checks whether the low-flow constraints of all cross-sections are satisfied under this parameter combination. Only when both conditions are met simultaneously—that is, the new parameters both reduce the objective function value and do not violate any low-flow constraints—will this new parameter be accepted as a better solution.

[0152] Once a better new parameter that meets the conditions is detected, the algorithm enters an accelerated movement phase. Since a better solution has been found along a certain direction, continuing in the same direction may lead to an even better solution. Therefore, the algorithm will start from the new parameter and move one or more steps along the direction of change from the original starting point to the new parameter to accelerate the convergence speed towards the optimal region.

[0153] If no better solution satisfying the conditions can be found in all directions at the current step size, it means that the current region has been searched sufficiently. At this time, the algorithm will reduce the search step size according to the step size decay factor, and perform a more refined local search in the vicinity of the current optimal solution. As the step size gradually decreases, the algorithm's location of the optimal solution becomes more and more accurate.

[0154] Through this iterative process of "probing-moving-reducing step size", the optimization algorithm can gradually approach the optimal combination of roughness parameters that minimizes the objective function while satisfying the low flow constraint.

[0155] Based on the same technical concept, embodiments of this application also provide a roughness and coefficient calibration device, such as... Figure 3 As shown, the device includes:

[0156] Model module 301 is used to construct a hydrodynamic model containing multiple cross sections and obtain the measured flow sequence of each cross section under multiple scenarios.

[0157] The quantile determination module 302 is used to identify, for each hydrological event, the critical quantile position of each cross section in the flow quantile threshold curve where the transition is from a flat segment to an upward segment, based on the measured flow sequence of each cross section, and to take the quantile with the highest synchronicity at the critical quantile position of each cross section as the unified optimal low flow quantile for that event.

[0158] The first construction module 303 is used to construct a low flow constraint that requires the simulated total flow to be no less than the low flow guarantee benchmark, based on the measured flow threshold corresponding to the optimal low flow quantile and using the measured total flow of each section below the measured flow threshold as the low flow guarantee benchmark.

[0159] The second construction module 304 is used to determine the comprehensive weight of each section based on the importance of the management function of each section and the quality of the measured flow data, and to construct an objective function based on the comprehensive weight and the simulation accuracy of each section in multiple scenarios.

[0160] The calibration module 305 is used to import the low flow constraint and the objective function into the optimization algorithm, search for the optimal combination of roughness parameters that minimizes the objective function under the premise of satisfying the low flow constraint, and output it as the calibration result.

[0161] In one feasible implementation, the quantile determination module is used, for each hydrological event, to identify, based on the measured flow sequence of each cross-section, the critical quantile position of each cross-section transitioning from a flat segment to an upward segment in the flow quantile threshold curve, and to use the quantile with the highest synchronicity at the critical quantile position of each cross-section as the unified optimal low flow quantile for that event, for the following purposes:

[0162] Construct a candidate set of low quantile parameters for flow; the candidate set contains multiple candidate quantiles for characterizing the level of ecological baseflow.

[0163] For each hydrological event, the measured flow sequences of each cross section are sorted in ascending order of value.

[0164] Traverse each candidate quantile in the candidate set to determine the target measured threshold corresponding to each cross section under that candidate quantile.

[0165] The target measured thresholds for each cross section at each candidate quantile are dimensionless, and the first and second derivatives of the dimensionless thresholds with respect to the quantiles are extracted. The first derivative is used to identify the stage of stable flow change, and the second derivative is used to identify the critical quantile position where the flow changes from stable to increasing in the stable stage.

[0166] A global scaling function is constructed based on the first and second derivatives extracted from each cross section. The optimal low flow quantile for this event is determined by solving for the candidate quantiles that make the global scaling function reach its maximum value. The global scaling function is used to characterize the synchronicity of flow changes at each cross section, and its function value is positively correlated with the synchronicity.

[0167] In a feasible implementation, the quality of the measured traffic data is determined based on the completeness and stability of the measured traffic sequence. The higher the quality of the measured traffic data, the greater the comprehensive weight. In this case, completeness is characterized by the data missing rate, and stability is characterized by the proportion of outliers. The quality of the measured traffic data is negatively correlated with the missing rate and the proportion of outliers.

[0168] In a feasible implementation, the importance of the management function is divided into several levels according to the cross-section type, including: a high importance level corresponding to major flood control safety cross-sections, a medium importance level corresponding to key ecological nodes or hydraulic control nodes, and a low importance level corresponding to ordinary reference cross-sections; the higher the level of the importance of the management function, the greater the corresponding comprehensive weight.

[0169] In one feasible implementation, the configuration parameters of the optimization algorithm include at least one of the following:

[0170] Roughness parameter range, initial value, number of iterations, search step size, step size decay factor, and constraint violation tolerance.

[0171] In a feasible implementation, the calibration module is used to search for the optimal combination of roughness parameters that minimizes the objective function while satisfying the low flow constraint, and is used to:

[0172] Starting from the current roughness parameter value, the probe is performed according to the search step size set in the configuration parameters. If a parameter value that makes the objective function smaller and satisfies the low flow constraint is detected, the probe moves step by step in the direction of the parameter value change.

[0173] In a feasible implementation, the second construction module is used to construct an objective function based on the comprehensive weights and the simulation accuracy of each section in multiple scenarios, for:

[0174] Based on the comprehensive weight of each monitoring section, the weighted average of the Nash efficiency coefficient of each section under all hydrological events is calculated.

[0175] The objective function is obtained based on the difference between 1 and the weighted average.

[0176] This application provides a roughness coefficient joint calibration method, apparatus, electronic device, and storage medium. By identifying the critical quantile positions of each cross-section in the flow quantile threshold curve where the flow transitions from a flat segment to an upward segment, and taking the quantile with the highest synchronicity at each cross-section's critical quantile position as the optimal low flow quantile for the entire river section, and constructing an objective function by combining cross-section comprehensive weights, and applying a low total flow constraint to the optimization algorithm for solution, the following beneficial effects are achieved:

[0177] To address the issue of insufficient simulation accuracy during low-flow periods, this application uses the measured total flow rate of each cross-section below the optimal low-flow quantile as the low-flow guarantee benchmark. It constructs a low-flow constraint requiring the simulated total flow rate to be no less than this benchmark. This ensures that the optimization process must take into account the overall water volume performance during low-flow periods, avoiding the problem of traditional methods where simulation deviations in low-flow segments are ignored due to the dominance of high-flow data. This improves the simulation reliability of the model during the dry season.

[0178] To address the issue of the inability to prioritize key cross-sections, this application determines a comprehensive weight based on the importance of each cross-section's management function and the quality of measured flow data. Based on this weight, an objective function is constructed, ensuring that key locations such as important urban cross-sections and ecologically sensitive areas receive greater attention during the calibration process, thus achieving differentiated optimization under multi-objective management requirements.

[0179] Compared with existing technologies, this application achieves optimal synchronization of the critical quantile positions of each section in the flow quantile threshold curve when transitioning from the flat segment to the rising segment through a unified optimal low flow quantile identification mechanism. This ensures the simulation accuracy during low flow periods while taking into account the priority of key sections, effectively solving the application limitations of traditional roughness calibration methods in multi-objective management scenarios.

[0180] Figure 4 A schematic diagram of an electronic device provided in this application embodiment includes: a processor 401, a storage medium 402, and a bus 403. The storage medium 402 stores machine-readable instructions executable by the processor 401. When the electronic device runs a roughness joint calibration method as described in the embodiment, the processor 401 and the storage medium 402 communicate through the bus 403, and the processor 401 executes the machine-readable instructions to perform the steps as described in the embodiment.

[0181] In this embodiment, the storage medium 402 may also execute other machine-readable instructions to perform other methods as described in the embodiment. For details on the specific execution steps and principles, please refer to the description of the embodiment, which will not be repeated here.

[0182] This application also provides a computer-readable storage medium storing a computer program that is executed by a processor to perform the steps as described in the embodiments.

[0183] In this embodiment, the computer program, when run by the processor, can also execute other machine-readable instructions to perform other methods as described in the embodiments. For details on the specific execution steps and principles, please refer to the description of the embodiments, which will not be repeated here.

[0184] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the coupling or direct coupling or communication connection shown or discussed may be through some communication interface; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0185] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0186] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0187] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0188] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A roughness joint calibration method, characterized by, The method includes: A hydrodynamic model containing multiple cross sections was constructed, and the measured flow sequences of each cross section under multiple scenarios were obtained. For each hydrological event, based on the measured flow sequence of each cross section, the critical quantile position of each cross section in the flow quantile threshold curve is identified as the transition from the flat section to the rising section, and the quantile with the highest synchronicity at the critical quantile position of each cross section is taken as the unified optimal low flow quantile for that event. Based on the measured flow threshold corresponding to the optimal low flow quantile, the measured total flow of each section below the measured flow threshold is used as the low flow guarantee benchmark, and a low flow constraint is constructed that requires the simulated total flow to be no less than the low flow guarantee benchmark. Based on the importance of the management function of each section and the quality of the measured flow data, the comprehensive weight of each section is determined, and an objective function is constructed based on the comprehensive weight and the simulation accuracy of each section in multiple scenarios. The low flow constraint and the objective function are imported into the optimization algorithm. Under the premise of satisfying the low flow constraint, the optimal combination of roughness parameters that minimizes the objective function is searched and output as the calibration result.

2. The method of claim 1, wherein, For each hydrological event, based on the measured flow sequence of each cross-section, the critical quantile position of each cross-section in the flow quantile threshold curve, where it transitions from a flat segment to an upward segment, is identified. The quantile with the highest synchronicity at each cross-section's critical quantile position is then taken as the unified optimal low flow quantile for that event, including: Construct a candidate set of low quantile parameters for flow; the candidate set contains multiple candidate quantiles for characterizing the level of ecological baseflow. For each hydrological event, the measured flow sequences of each cross section are sorted in ascending order of value; Traverse each candidate quantile in the candidate set to determine the target measured threshold corresponding to each cross section under the candidate quantile. The target measured thresholds for each cross section at each candidate quantile are dimensionless, and the first and second derivatives of the dimensionless thresholds with respect to the quantiles are extracted. The first derivative is used to identify the stage of stable flow change, and the second derivative is used to identify the critical quantile position where the flow changes from stable to rising in the stable stage. A global scaling function is constructed based on the first and second derivatives extracted from each cross section. The optimal low flow quantile for this event is determined by solving for the candidate quantiles that make the global scaling function reach its maximum value. The global scaling function is used to characterize the synchronicity of flow changes at each cross section, and its function value is positively correlated with the synchronicity.

3. The method of claim 1, wherein, The quality of the measured traffic data is determined based on the completeness and stability of the measured traffic sequence. The higher the quality of the measured traffic data, the greater the comprehensive weight. The completeness is characterized by the data missing rate, and the stability is characterized by the proportion of outliers. The quality of the measured traffic data is negatively correlated with the missing rate and the proportion of outliers.

4. The method of claim 1, wherein, The importance of the management function is divided into several levels according to the cross-section type, including: high importance level corresponding to major flood control safety cross-sections, medium importance level corresponding to key ecological nodes or hydraulic control nodes, and low importance level corresponding to ordinary reference cross-sections; the higher the level of the importance of the management function, the greater the corresponding comprehensive weight.

5. The method of claim 1, wherein, The configuration parameters of the optimization algorithm include at least one of the following: Roughness parameter range, initial value, number of iterations, search step size, step size decay factor, and constraint violation tolerance.

6. The method of claim 5, wherein, The step of searching for the optimal combination of roughness parameters that minimizes the objective function while satisfying the low flow constraint includes: Starting from the current roughness parameter value, the probe is performed according to the search step size set in the configuration parameters. If a parameter value that makes the objective function smaller and satisfies the low flow constraint is detected, the probe moves step by step in the direction of the parameter value change.

7. The method according to claim 1, characterized in that, Based on the comprehensive weights and the simulation accuracy of each section in multiple scenarios, an objective function is constructed, including: Based on the comprehensive weight of each monitoring section, the weighted average of the Nash efficiency coefficient of each section under all hydrological events is calculated. The objective function is obtained based on the difference between 1 and the weighted average.

8. A roughness co- calibration device, characterized by The device includes: The model module is used to construct a hydrodynamic model containing multiple cross sections and obtain the measured flow sequence of each cross section under multiple scenarios. The quantile determination module is used to identify the critical quantile position of each section in the flow quantile threshold curve from the flat section to the rising section for each hydrological event, based on the measured flow sequence of each section, and to take the quantile with the highest synchronicity at the critical quantile position of each section as the unified optimal low flow quantile for that event. The first construction module is used to construct a low flow constraint that requires the simulated total flow to be no less than the low flow guarantee benchmark, based on the measured flow threshold corresponding to the optimal low flow quantile and using the measured total flow of each section below the measured flow threshold as the low flow guarantee benchmark. The second construction module is used to determine the comprehensive weight of each section based on the importance of the management function of each section and the quality of the measured flow data, and to construct an objective function based on the comprehensive weight and the simulation accuracy of each section in multiple scenarios. The calibration module is used to import the low flow constraint and the objective function into the optimization algorithm, search for the optimal combination of roughness parameters that minimizes the objective function under the premise of satisfying the low flow constraint, and output the calibration result.

9. An electronic device, comprising: include: The device includes a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is in operation, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the roughness joint calibration method 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 a computer program that, when executed by a processor, performs the steps of the roughness joint calibration method as described in any one of claims 1 to 7.

Images