A waterlogging level grading coupling calculation method and system in a flood detention basin

Abstract

The application provides a flood storage area waterlogging level grading coupling calculation method and system, comprising: constructing a flood storage area waterlogging level calculation method and an input data set CS; establishing a flood storage area waterlogging level calculation method adaptive matching mechanism based on flood return period grading thresholds, dividing flood grade intervals by setting flood return period grading thresholds, and establishing a return period method mapping function; based on real-time input return period values, dynamically activating the waterlogging level calculation method in the flood storage area waterlogging level calculation method that is adapted to the flood return period; performing waterlogging level analysis operation in the flood storage area waterlogging level calculation method that is adapted to the flood return period, determining a multi-source method waterlogging level set; quantifying the fusion weight coefficients of the multi-source method waterlogging level set adapted to the flood return period; based on the multi-source method waterlogging level set adapted to the flood return period and the fusion weight coefficients, determining the waterlogging level of the flood storage area under different flood return periods.

Description

Technical Field

[0001] This invention relates to the field of flood control design technology and disaster risk management in water conservancy projects, and in particular to a method and system for graded coupled calculation of water level in flood storage and detention areas. Background Technology

[0002] Flood storage and detention areas play a central role in the basin flood control system, and the accurate calculation of their urban flooding levels is a crucial foundation for basin flood control planning, engineering design, and flood management. Currently, the urban flooding level calculation methods recommended in the design flood calculation specifications for water conservancy and hydropower projects mainly include the regulation and storage algorithm and the historical water level correction method. However, these methods have significant limitations: the regulation and storage algorithm performs drainage and regulation calculations time-by-time based on the design flow process, resulting in relatively accurate results, but it relies on high-precision hydrological and drainage data, leading to a sharp drop in accuracy in data-scarce scenarios; the historical water level correction method is limited by the representativeness and reliability of the highest historical water level surveyed, making it difficult to cope with extreme hydrological events; currently, machine learning technology has shown significant application value in the management and research of flood storage and detention areas. While machine learning models have data-driven advantages, they have stringent requirements for the quality and scale of training samples and insufficient generalization ability. More importantly, existing technologies lack method selection criteria for different flood recurrence periods and systematic hierarchical calculation strategies, making it difficult to balance method applicability and calculation accuracy. Meanwhile, the fusion of results from multiple methods lacks a dynamic weighting mechanism, resulting in significant uncertainty and severely limiting its engineering application value. Summary of the Invention

[0003] The purpose of this invention is to address the shortcomings of the prior art by providing a graded coupled calculation method and system for water level in flood storage and detention areas. This method enables adaptive optimization of calculation methods under different flood recurrence periods and weighted fusion of results from multiple methods, thereby improving the accuracy, reliability, and engineering applicability of water level calculation. It provides solid scientific support for refined flood control scheduling in watersheds, safe operation of flood storage and detention areas, and emergency management decision-making.

[0004] To achieve the above objectives, the present invention adopts the following technical solution:

[0005] This invention provides a method for graded coupled calculation of water level in flood storage and detention areas, including:

[0006] S1. Construct a method for calculating water level in flood storage and detention areas and input dataset CS;

[0007] S2. Establish an adaptive matching mechanism for the calculation method of waterlogging level in flood storage and detention areas based on the flood return period classification threshold. By setting the flood return period classification threshold, flood level intervals are divided, and a return period method mapping function is established. Based on the real-time input return period value, the waterlogging level calculation method that is adapted to the flood return period is dynamically activated in the calculation method of waterlogging level in flood storage and detention areas.

[0008] S3. Perform the analytical calculation of water level in the water storage red zone that is compatible with the flood return period in the water level calculation method, and determine the water level set of the multi-source method;

[0009] S4. Fusion weight coefficients of multi-source method for waterlogging level solution set adapted to quantification and flood return period;

[0010] S5. Based on the multi-source method for waterlogging level set and fusion weight coefficient adapted to the flood return period, determine the waterlogging level of the flood storage and detention area under different flood return periods.

[0011] Furthermore, in S1,

[0012] The method for calculating the water level in the flood storage and detention area includes a regulation and storage algorithm, a historical water level correction method, and a machine learning model.

[0013] The input dataset CS includes a rainstorm dataset, a terrain dataset, a drainage dataset, and a model parameter set, specifically:

[0014] ;

[0015] in, PS Data set for heavy rain; DXS For terrain datasets; BZS For drainage dataset; JQS For the model parameter set;

[0016] The rainstorm dataset PS for:

[0017] ;

[0018] in, This refers to daily rainfall observation data; Data on the time-history distribution of design storms for a typical year; This is the initial loss value; Infiltration rate;

[0019] The terrain dataset for:

[0020] ;

[0021] in, This is the water level-volume curve of the flood storage and detention area; This is a water level-area curve; The catchment area of ​​the flood storage and detention area, Z represents the highest historical water accumulation area in the flood storage and detention area. D This is the highest historical water level recorded during the investigation. For elevation data of flood storage and detention areas, For land use and cover data of flood storage and detention areas;

[0022] The drainage dataset for:

[0023] ;

[0024] in, The pumping station is designed with a drainage flow rate. The pumping station is at the designated drainage level. For the scheduling and operation rules of the pumping station;

[0025] The model parameter set for:

[0026] ;

[0027] in, For machine learning model parameters, This represents the number of machine learning parameters.

[0028] Furthermore, in S2,

[0029] The adaptive matching mechanism for calculating water level in flood storage and detention areas based on flood return period classification thresholds includes a flood classification module and a dynamic matching module, specifically:

[0030] The flood level classification module divides flood level intervals by predefined flood return period classification thresholds and establishes a flood level interval set TS;

[0031] , ;

[0032] in, This is the flood recurrence period. The total number of flood levels;

[0033] The dynamic matching module establishes a return period method mapping function based on the water level calculation method and the flood level interval set TS. Based on the real-time input return period value, it dynamically activates the water level calculation method in the flood storage and detention area that is adapted to the flood return period.

[0034] when At that time, the water storage algorithm is activated;

[0035] when At that time, the water storage algorithm and the historical water level correction method are activated;

[0036] when At that time, the water storage algorithm, historical water level correction method and machine learning model are activated;

[0037] when At that time, the machine learning model is activated.

[0038] Furthermore, S3 specifically refers to:

[0039] When executing the aforementioned regulation and storage algorithm, the input dataset CS is invoked. Based on the water balance equation and the water level-volume curve of the flood storage area, the water level for urban flooding is designed by referring to the water level-volume curve from the water storage volume during the red zone period. Specifically:

[0040] , ;

[0041] ;

[0042] in, For the first The flood storage capacity of the flood detention area during a specific period; For the first The flood storage capacity of the flood detention area during a specific period; For the first flood storage and detention area Inflow rate during a given time period; For the first Drainage volume during a given time period; For the first Water level values ​​during the specified time period; To determine the floodwater level calculated by the regulation and storage algorithm, the highest floodwater level value within the time period is taken. , The duration of rainfall that caused flooding in the flood storage and detention area;

[0043] Based on rainfall data and pump station scheduling rules, the inflow rate to the flood storage and detention area is calculated on a time-by-time basis. Drainage volume :

[0044] ;

[0045] ;

[0046] ;

[0047] in, For the first Rainfall data for different time periods; This is the initial loss; For the first Infiltration rate over time; For the calculation period; Let be the water accumulation area of ​​the flood storage and detention area during the i-th time period; For the pumping station Drainage volume during a given time period;

[0048] When executing the historical water level correction method, the input dataset CS is called to calculate the flood water level. Specifically:

[0049] ;

[0050] in, The water level in the flood storage and detention area is calculated using the historical water level correction method. The highest historical water level was determined by on-site investigation; and The measured catchment area and the highest historical water accumulation area of ​​the flood storage and detention area; The difference between the designed frequency of rainfall for the period corresponding to the rainfall duration that causes waterlogging in the flood storage and detention area and the annual maximum precipitation corresponding to the year with the highest historical water level;

[0051] When executing the machine learning model, the input dataset CS is called, and the urban flooding water level is calculated based on the rainstorm data, terrain data, drainage data, and model parameters. The details are as follows:

[0052] .

[0053] Furthermore, in S4, the fusion weight coefficients are obtained using the Delphi method, analytic hierarchy process, CRITIC analysis, entropy weighting method, or principal component analysis.

[0054] Furthermore, S5 specifically includes:

[0055] Consider improving the traditional weighted average method by explicitly combining it with a volatility linear indicator, and adjusting the parameters. By flexibly adjusting the weights, the water level solution set of the multi-source method is fused using the following formula to obtain the water level value of the flood storage and detention area. :

[0056] ;

[0057] ;

[0058] in, To integrate the water level values ​​of flood storage and detention areas using multi-source methods; For the first Weights calculated using each method; For the first The water level value for flooding was calculated using one method; To adjust the parameters; The weighted standard deviation of the solution set for water level in urban flooding reflects the dispersion or volatility of the solution set; is the average value of the water level solution set; m is the number of water levels in the water level solution set;

[0059] The adjustment parameter The rules for selecting values ​​are as follows:

[0060] λ < 0: Penalize fluctuations, pursue the stability of the water level solution set for urban flooding, and reduce variability;

[0061] λ > 0: Encourage volatility, pursue differentiated development of water level solutions for urban flooding, and highlight advantages from an empirical perspective;

[0062] λ = 0: Degenerates into a traditional weighted average.

[0063] Furthermore, a graded coupled calculation system for water level in flood storage and detention areas is implemented using a graded coupled calculation method for water level in flood storage and detention areas, and also includes:

[0064] The basic data input unit is used to construct the water level classification calculation system and parameter set for urban flooding.

[0065] The hierarchical matching unit is used to execute the adaptive matching mechanism of the water level calculation method for flood storage and detention areas based on the flood return period hierarchical threshold. It establishes a return period-method mapping function by dividing the flood level interval; and dynamically activates the water level calculation method adapted to the flood return period based on the real-time input return period value.

[0066] The water level solution set calculation unit performs the water level analysis operation of the flood storage and detention area that is compatible with the flood return period in the water level calculation method system of the flood storage and detention area, and determines the water level solution set of the water level.

[0067] The fusion weight calculation unit is used to calculate the fusion weight coefficients of the multi-source method waterlogging level solution set adapted to the flood return period;

[0068] The waterlogging level coupling unit is used to perform waterlogging level calculations coupled with a multi-source method adapted to the flood return period, and outputs the final waterlogging level value.

[0069] The beneficial effects of this invention are as follows: This invention innovatively constructs a technical framework of data input-hierarchical matching-dissolved set calculation-weight quantization-coupled output, breaking through the limitations of traditional single methods to adapt to all scenarios. It also innovatively proposes a return period-method mapping strategy. Based on the flood return period, it overcomes traditional technical bottlenecks through adaptive method selection and intelligent fusion of multi-source results, achieving a dual improvement in the accuracy and reliability of urban flooding water level calculation. Attached Figure Description

[0070] Figure 1 A flowchart of a graded coupled calculation method for water level in flood storage and detention areas;

[0071] Figure 2 A process flow diagram for a 50-year return period rainstorm and a design net rainfall event.

[0072] Figure 3 A schematic diagram of the final floodwater level calculated for the training dataset of the neural network model;

[0073] Figure 4 A schematic diagram of the floodwater level calculated for the neural network model validation dataset;

[0074] Figure 5 This is a schematic diagram of a coupled calculation system for water level classification in a flood storage and detention area. Detailed Implementation

[0075] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0076] Please see Figure 1 A method for coupled calculation of water level classification in flood storage and detention areas, comprising:

[0077] S1. Construct a method for calculating water level in flood storage and detention areas and input dataset CS;

[0078] S2. Establish an adaptive matching mechanism for the calculation method of waterlogging level in flood storage and detention areas based on the flood return period classification threshold. By setting the flood return period classification threshold, flood level intervals are divided, and a return period method mapping function is established. Based on the real-time input return period value, the waterlogging level calculation method that is adapted to the flood return period is dynamically activated in the calculation method of waterlogging level in flood storage and detention areas.

[0079] S3. Perform the analytical calculation of water level in the water storage red zone that is compatible with the flood return period in the water level calculation method, and determine the water level set of the multi-source method;

[0080] S4. Fusion weight coefficients of multi-source method for waterlogging level solution set adapted to quantification and flood return period;

[0081] S5. Based on the multi-source method for waterlogging level set and fusion weight coefficient adapted to the flood return period, determine the waterlogging level of the flood storage and detention area under different flood return periods.

[0082] In S1,

[0083] The method for calculating the water level in the flood storage and detention area includes a regulation and storage algorithm, a historical water level correction method, and a machine learning model.

[0084] The machine learning model used in this embodiment is an LSTM neural network model.

[0085] The input dataset CS includes a rainstorm dataset, a terrain dataset, a drainage dataset, and a model parameter set, specifically:

[0086] ;

[0087] in, Data set for heavy rain; For terrain datasets; For drainage dataset; For the model parameter set;

[0088] The rainstorm dataset for:

[0089] ;

[0090] in, This refers to daily rainfall observation data; Data on the time-history distribution of design storms for a typical year; This is the initial loss value; Infiltration rate;

[0091] In this embodiment, the rainfall data used are daily rainfall observation data from the first, second, third, and fourth rain gauge stations from 1964 to 2020; following the principle of the least favorable outcome, the maximum 60-day daily rainfall event of 1991 was selected as the design rainfall pattern; the initial loss of the runoff coefficient... The diameter was determined to be 22.5 mm, with a permeability coefficient of... Based on the soil and vegetation conditions of the flood storage and detention area, and using the coefficient adopted by the adjacent flood diversion and storage area, the value was determined to be 1.2 mm / h;

[0092] The terrain dataset for:

[0093] ;

[0094] in, This is the water level-volume curve of the flood storage and detention area; This is a water level-area curve; The catchment area of ​​the flood storage and detention area is 613.98 km². 2 ; This represents the highest historical water accumulation area in the flood storage and detention area, based on a 30m resolution. EMD The measured distance is 409 km. 2 ; To determine the highest historical water level, this embodiment surveyed five water levels within the flood storage and detention area, ranging from 19.7 to 25.24 meters. Take 25.24m; DEM For elevation data of flood storage and detention areas, currently available large-scale raster data is used. DEM The data has a spatial resolution of 30m. LULCFor land use and land cover data in flood storage and detention areas, the 10m resolution global land cover dataset published by ESRI was used.

[0095] The drainage dataset BZS is:

[0096] ;

[0097] in, The pumping station is designed with a drainage flow rate. The pumping station is at the designated drainage level. For the scheduling and operation rules of the pumping station;

[0098] In this embodiment, floodwater in the flood storage and detention area is mainly pumped out to the outer river through the first pumping station during the flood season, with a designed drainage flow rate of 89m³. 3 / s, with a starting and ending point of 22.5m; in combination with the pumping station's drainage capacity and scheduling and operation procedures, the pumping station in this embodiment pumps at 75% of its design flow rate.

[0099] The model parameter set for:

[0100] ;

[0101] in, For machine learning model parameters, This represents the number of machine learning parameters.

[0102] In this embodiment, The parameters of an LSTM neural network model include the time step. Number of neurons in LSTM layer Learning rate Regularization parameters Batch size Number of iterations .

[0103] In S2,

[0104] The adaptive matching mechanism for calculating water level in flood storage and detention areas based on flood return period classification thresholds includes a flood classification module and a dynamic matching module, specifically:

[0105] The flood level classification module divides flood level intervals by predefined flood return period classification thresholds and establishes a set of flood level intervals. ;

[0106] , ;

[0107] in, This is the flood recurrence period. The total number of flood levels;

[0108] In the embodiment, the set of flood level intervals for the flood storage and detention area is as follows: ={10%, 2%, 1%}, .

[0109] The dynamic matching module, based on the water level calculation method and the flood level interval set, Establish a return period method mapping function, and dynamically activate the water level calculation method in the flood storage and detention area that is adapted to the flood return period based on the real-time input return period value.

[0110] when At that time, the water storage algorithm is activated;

[0111] when At that time, the water storage algorithm and the historical water level correction method are activated;

[0112] when At that time, the water storage algorithm, historical water level correction method and machine learning model are activated;

[0113] when At that time, the machine learning model is activated.

[0114] In this embodiment, the water level of the flood storage area is calculated based on the 50-year (2%) flood return period, which will activate the regulation and storage algorithm, the historical water level correction method, and the machine learning model.

[0115] Specifically, S3 is:

[0116] When executing the aforementioned regulation algorithm, the input dataset is called. Based on the water balance equation and the water level-volume curve of the flood storage area, the water level for urban flooding is designed by referring to the water level-volume curve based on the water storage volume during the red zone period. Specifically:

[0117] , ;

[0118] ;

[0119] in, For the first The flood storage capacity of the flood detention area during a specific period; For the first The flood storage capacity of the flood detention area during a specific period; For the first flood storage and detention area Inflow rate during a given time period; For the first Drainage volume during a given time period; For the first Water level values ​​during the specified time period; To determine the floodwater level calculated by the regulation and storage algorithm, the highest floodwater level value within the time period is taken. , The duration of rainfall that caused waterlogging in the flood storage and detention area was recorded as 60 days.

[0120] Based on rainfall data and pump station scheduling rules, the inflow rate to the flood storage and detention area is calculated on a time-by-time basis. Drainage volume :

[0121] ;

[0122] ;

[0123] ;

[0124] in, For the first Rainfall data for different time periods; For initial damage, take 22.5mm; For the first The infiltration rate for that time period is taken as 1.2 mm / h; The calculation period is based on days; Let be the water accumulation area of ​​the flood storage and detention area during the i-th time period; For the pumping station Drainage volume during a given time period;

[0125] The design storm was calculated based on the daily rainfall from four rain gauge stations, using the arithmetic mean as the regional surface rainfall. The sampling principle for the design storm was the historical maximum 7-day, 15-day, and 30-day rainfall. Empirical frequencies were calculated using mathematical formulas, and the alignment type was P-III. The alignment was adapted using an empirical method, yielding the following maximum 7-day design storm values ​​for the flood storage area under a 50-year return period: 586.5 mm, 699.5 mm for the maximum 15-day design storm, and 890.0 mm for the maximum 30-day design storm. The design rainfall pattern was based on the maximum 60-day daily rainfall event from May 18th to July 16th, 1991. Combined with runoff parameters, the design net rainfall event was calculated, and the results are shown in the appendix. Figure 2 Based on the pumping station's drainage capacity and dispatching procedures, analysis and calculations were conducted using 75% of the design flow rate for drainage. The storage capacity of the flood storage and detention area was calculated for each time period. The water level-volume curve of the flood storage and detention area was consulted to obtain the water level for each time period. The highest water level within that time period was taken as the floodwater level calculated by the regulation and storage algorithm. The water level is 23.06m. The water level changes and corresponding values ​​for each time period are shown in Table 1 and Table 2.

[0126] Table 1. Calculation Results of Inland Flood Water Level Using the Regulation and Storage Algorithm

[0127]

[0128] Table 2. Results of Calculation of Floodwater Level Using the Regulation and Storage Algorithm

[0129]

[0130] When executing the historical water level correction method, the input dataset CS is called to calculate the flood water level. Specifically:

[0131] ;

[0132] in, The water level in the flood storage and detention area is calculated using the historical water level correction method. The highest historical water level was determined by on-site investigation; A 汇 and A 积max The measured catchment area and the highest historical water accumulation area of ​​the flood storage and detention area; The difference between the designed frequency of rainfall for the period corresponding to the rainfall duration that causes waterlogging in the flood storage and detention area and the annual maximum precipitation corresponding to the year with the highest historical water level;

[0133] In this embodiment, based on the rainfall characteristics, topographical conditions, and actual number of days of flooding within the flood storage and detention area, the rainfall duration for flooding was determined to be 15 days. According to rainfall analysis of the project area, the years with the largest rainfall were 1954, 1983, 1991, 2016, and 2020. Considering that the frequency of the flooding in 2020 was close to a 50-year return period, and the flooding occurred relatively recently, its survey results are considered reliable. Therefore, 2020 was selected as a typical year, and its highest water level was investigated to be 25.24m. The maximum 15-day design rainfall for a 50-year return period in the flood storage and detention area was 699.5mm, and the maximum 15-day rainfall in 2020 was 668.9mm. Therefore, ΔZ = 30.6mm, and the ratio of the catchment area to the waterlogged area is approximately 1.5. The calculated flooding water level Z, obtained using the historical water level correction method, is... l It is 25.29m.

[0134] When executing the machine learning model, the input dataset CS is called, and the urban flooding water level is calculated based on the rainstorm data, terrain data, drainage data, and model parameters. The details are as follows:

[0135] .

[0136] In this embodiment, considering that there are currently no measured true values ​​of urban flooding water levels in the flood storage and detention area, the urban flooding water levels in the training and validation datasets of the LSTM neural network model are simulated using the HEC-RAS model. The maximum 60-day daily rainfall events of each year from 1964 to 2020 are used as the rainfall process input for the HEC-RAS model to calculate the urban flooding water levels. The urban flooding water levels from 1964 to 2010, the maximum 60-day daily rainfall events of each year, the pumping station drainage flow and scheduling procedures, and 30m... DEM The LSTM neural network model was trained using 10m of land cover data. The Nash efficiency coefficient was used as the evaluation metric to optimize the parameters, and the model parameters were determined as follows: =48, =2 layers (64→32 neurons) =0.001, =0.2, =32, =100, see the graph for training sample calculation results. Figure 3 A trained LSTM neural network model was used, with inputs including urban flooding water levels from 2011 to 2020, the maximum annual 60-day rainfall events, pump station drainage flow rates and scheduling procedures, and 30m... DEM The data, including 10m surface cover data, were used to calculate the water level in the flood storage and detention area. The calculation results for the verification sample are shown below. Figure 4 Using a 50-year return period design net rainfall process as input, the trained LSTM neural network model was used to calculate the 50-year return period flood level in the flood storage and detention area as 25.10m.

[0137] In S4, the fusion weighting coefficients are obtained using the Delphi method, analytic hierarchy process, CRITIC analysis, entropy weighting, or principal component analysis.

[0138] In this embodiment, the Delphi method was used to analyze and determine the dynamic fusion weight coefficients of the three methods of waterlogging level: the water storage algorithm, the historical water level correction method, and the machine learning model, which were 0.4, 0.4, and 0.2, respectively.

[0139] Specifically, S5 is:

[0140] Consider improving the traditional weighted average method by explicitly combining it with a volatility linear indicator, and adjusting the parameters. By flexibly adjusting the weights, the water level solution set of the multi-source method is fused using the following formula to obtain the water level value of the flood storage and detention area. :

[0141] ;

[0142] ;

[0143] in, To integrate the water level values ​​of flood storage and detention areas using multi-source methods; For the first Weights calculated using each method; For the first The water level value for flooding was calculated using one method; To adjust the parameters; The weighted standard deviation of the solution set for water level in urban flooding reflects the dispersion or volatility of the solution set; This represents the average value of the water level set for flooding. The number of water levels in the waterlogging data set;

[0144] The adjustment parameter The rules for selecting values ​​are as follows:

[0145] λ < 0: Penalize fluctuations, pursue the stability of the water level solution set for urban flooding, and reduce variability;

[0146] λ > 0: Encourage volatility, pursue differentiated development of water level solutions for urban flooding, and highlight advantages from an empirical perspective;

[0147] λ = 0: Degenerates into a traditional weighted average.

[0148] Calculations show that =1.071, in order to reduce the variability of the water level solution set for urban flooding, Take -0.1, =24.25m.

[0149] like Figure 5 As shown, a graded coupled calculation system for water level in flood storage and detention areas is implemented using the aforementioned graded coupled calculation method for water level in flood storage and detention areas, and further includes:

[0150] The basic data input unit is used to construct the water level classification calculation system and parameter set for urban flooding.

[0151] The hierarchical matching unit is used to execute the adaptive matching mechanism of the water level calculation method for flood storage and detention areas based on the flood return period hierarchical threshold. It establishes a return period-method mapping function by dividing the flood level interval; and dynamically activates the water level calculation method adapted to the flood return period based on the real-time input return period value.

[0152] The water level solution set calculation unit performs the water level analysis operation of the flood storage and detention area that is compatible with the flood return period in the water level calculation method system of the flood storage and detention area, and determines the water level solution set of the water level.

[0153] The fusion weight calculation unit is used to calculate the fusion weight coefficients of the multi-source method waterlogging level solution set adapted to the flood return period;

[0154] The waterlogging level coupling unit is used to perform waterlogging level calculations coupled with a multi-source method adapted to the flood return period, and outputs the final waterlogging level value.

[0155] The embodiments described above are merely illustrative of implementation methods of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be defined by the appended claims.

Claims

1. A method for calculating the waterlogging level classification coupling in a flood detention basin, characterized in that, The method comprises the following steps: S1, constructing a waterlogging level calculation method for a flood storage and detention area and an input data set CS; S2, establishing an adaptive matching mechanism for the waterlogging level calculation method for the flood storage and detention area based on a flood return period grading threshold, dividing flood grade intervals by setting the flood return period grading threshold, and establishing a return period method mapping function; based on the real-time input return period value, the waterlogging level calculation method in the waterlogging level calculation method for the flood storage and detention area that is adapted to the flood return period is dynamically activated; S3, performing waterlogging level analysis operation in the waterlogging level calculation method for the flood storage and detention area that is adapted to the flood return period, and determining a multi-source method waterlogging level solution set; S4, quantifying the fusion weight coefficient of the multi-source method waterlogging level solution set that is adapted to the flood return period; S5, determining the waterlogging level of the flood storage and detention area under different flood return periods based on the multi-source method waterlogging level solution set and the fusion weight coefficient that are adapted to the flood return period; In S2, The adaptive matching mechanism for the waterlogging level calculation method for the flood storage and detention area based on the flood return period grading threshold comprises a flood grade division module and a dynamic matching module, specifically: The flood grade division module divides the flood grade intervals by predefining the flood return period grading threshold, and establishes a flood grade interval set TS; , k = 1 : K; Wherein, T is the flood return period, and K is the total number of flood grade division; The dynamic matching module establishes a return period method mapping function according to the waterlogging level calculation method and the flood grade interval set TS, and dynamically activates the waterlogging level calculation method in the waterlogging level calculation method for the flood storage and detention area that is adapted to the flood return period based on the real-time input return period value; When T < T1, the storage and regulation algorithm is activated; When T1≤T<T k , activate the regulation and storage algorithm and the historical water level correction method; When T k ≤ T < T K , activate the regulation and storage algorithm, the historical water level correction method and the machine learning model; When T ≥ T K the machine learning model is activated.

2. The waterlogging level grading coupled calculation method in a detention basin according to claim 1, characterized in that: In S1, The waterlogging level calculation method for the flood storage and detention area comprises a storage and regulation algorithm, a historical water level correction method, and a machine learning model; The input data set CS comprises a rainstorm data set, a terrain data set, a drainage data set, and a model parameter set, specifically: ； Wherein, PS is the rainstorm data set; DXS is the terrain data set; BZS is the drainage data set; and JQS is the model parameter set; The rainstorm data set PS is: ； where P p is the daily rainfall observation data; P D is the design storm time distribution data of a typical year; I0is the initial damage value; and f is the infiltration rate. The terrain data set DXS is: ； where V = f(Z) is the water level-storage curve of the flood detention basin; A = f(Z) is the water level-area curve; is the catchment area of the flood detention basin, is the historical maximum water storage area of the flood detention basin, Z D is the historical maximum water storage level of the survey, DEM is the elevation data of the flood detention basin, and LULC is the land use and cover data of the flood detention basin. The drainage data set BZS is: ； wherein Q P is the drainage flow designed for the pumping station, BZ0 is the starting drainage level of the pumping station, is the dispatching operation rule of the pumping station; The model parameter set JQS is: ； wherein, is a machine learning model parameter, is a number of machine learning parameters.

3. The waterlogging level grading coupled calculation method in a detention basin according to claim 2, characterized in that: In S3, When the storage and detention algorithm is executed, the input data set CS is called, and the waterlogging water level Z is designed by the water storage amount of the storage and detention area period and the water level volume curve according to the water balance equation and the water level volume curve of the storage and detention area t , Specifically: ， ； ； V i is the flood storage volume of the flood detention basin in the i th period; V i-1 is the flood storage volume of the flood detention basin in the i-1 th period; W in,i is the inflow of the flood detention basin in the i th period; W out,i is the drainage amount in the i th period; Z ti is the waterlogging water level value in the i th period; Z t is the waterlogging water level calculated by the regulation and storage algorithm, and is the highest waterlogging water level value; i=1:n, and n is the rainfall duration in which the flood detention basin forms waterlogging. According to rainfall data and pump station operation rules, the inflow W of the flood storage and detention area is calculated by time period in,i , drainage capacity W out,i : ； ； ； wherein P i is the rainfall data for the i-th period; I0is the initial damage; i is the infiltration rate for the i-th period; is the calculation period; is the water accumulation area of the i-th period of the flood detention basin; Q i is the i-th period of the pumping station drainage capacity; When performing the historical water level correction method, the input data set CS is called, and the waterlogging water level Z is calculated l , specifically: ； wherein Z l is the waterlogging level in the flood storage and detention area calculated by the historical water level correction method; Z D is the historically highest water level determined by field investigation; A 汇 and A 积max are the calculated catchment area of the flood storage and detention area and the historically highest water area; and ΔZ is the difference between the rainfall intensity of the time period corresponding to the rainfall duration of the formation of waterlogging in the flood storage and detention area and the annual maximum precipitation corresponding to the historically highest water level year. When the machine learning model is executed, the input data set CS is called to calculate the waterlogging water level Z according to the storm data, the terrain data, the drainage data and the model parameters q , as follows: 。 4. The waterlogging level grading coupled calculation method in a detention basin according to claim 3, characterized in that: In S4, the fusion weight coefficient adopts Delphi method, analytic hierarchy process, CRITIC analysis method, entropy weight method, or principal component analysis method.

5. The waterlogging level grading coupled calculation method in a detention basin according to claim 4, characterized in that, In S5, Considering improving the traditional weighted average method, combining with the explicit combination of the wave linear index, adjusting the parameter Flexibly adjusting the weight, using the following formula to fuse the multi-source method waterlogging level solution set, and obtaining the waterlogging level value Z of the flood storage and detention area final : ； ； wherein Z final is the waterlogging level value in the flood storage and detention area by fusing multi-source methods; w j is the weight of the jth calculation method; Z j is the waterlogging level value calculated by the jth method; is the adjustment parameter; is the weighted standard deviation of the waterlogging level solution set, reflecting the dispersion degree or volatility of the solution set; is the average value of the waterlogging level solution set; m is the number of waterlogging levels in the waterlogging level solution set; The adjustment parameter , the value rule is: λ < 0: Punish volatility, pursue the stability of the waterlogging level solution set, and reduce the difference; λ > 0: Encourage volatility, pursue the differential development of the waterlogging level solution set, and highlight the advantage item from the experience angle; λ = 0: Degenerate into traditional weighted average.

6. A waterlogging level grading coupling calculation system in a flood detention basin, characterized in that: A waterlogging level grading coupling calculation method for a flood storage and detention area is adopted, and the method further comprises: A basic data input unit for executing construction of a waterlogging level grading calculation system and a parameter set; The hierarchical matching unit is used for executing an adaptive matching mechanism of a waterlogging level calculation method based on a flood return period hierarchical threshold in a flood storage and detention area. A return period-method mapping function is established through division of a flood level interval. Based on a real-time input return period value, a waterlogging level calculation method adapted to the flood return period is dynamically activated. The waterlogging level solution set calculation unit executes waterlogging level calculation method system analysis operation in the flood storage and detention area adapted to the flood return period, and determines the waterlogging level solution set. The fusion weight calculation unit is used for executing calculation of a fusion weight coefficient of a multi-source method waterlogging level solution set adapted to the flood return period. The waterlogging level coupling unit is used for executing waterlogging level calculation of a multi-source method coupling adapted to the flood return period, and outputs a final waterlogging level value.

Images