A method and system for fine simulation of runoff generation considering spatial heterogeneity

Abstract

The present invention provides a method and system for fine simulation of runoff generation considering spatial heterogeneity, and relates to the technical field of hydrological simulation. The method includes: dividing the urban underlying surface and natural underlying surface of the target area, generating a land use type map; establishing an urban hydrological model based on the urban underlying surface, dividing different sub-catchments according to the urban river / pipeline network information, and calculating the surface runoff generation of different sub-catchments respectively; establishing an urban river / pipeline network hydrodynamic model based on the urban river / pipeline network information and the surface runoff generation calculation results of different sub-catchments; dividing several sub-basins according to the digital elevation DEM data of the area to which the natural underlying surface belongs, and calculating the runoff generation of each sub-basin respectively; establishing a river confluence model based on the runoff generation calculation results of each sub-basin, and using the river confluence model to find the confluence point; coupling the urban river / pipeline network hydrodynamic model and the river confluence model based on the hydrodynamic relationship and confluence logic relationship of the confluence point.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrological simulation, and in particular to a method and system for fine simulation of runoff generation taking into account spatial heterogeneity. Background Art

[0002] Affected by human activities, urban underlying surfaces are mainly composed of impermeable artificial materials, and the runoff is formed much faster than the natural underlying surface composed of vegetation and bare soil. In addition, the urban underlying surface structure is complex, including a variety of artificial facilities such as drainage systems, roads, and buildings, and the convergence mechanism is significantly different from the natural underlying surface. At the same time, the population and facility density in cities are high, and the economic losses of urban waterlogging are often relatively large. Therefore, a special urban hydrological model needs to be established for cities.

[0003] The more common urban hydrological models currently only model small-scale urban areas, and are unable to describe the spatial heterogeneity of large-scale urban natural and urban underlying surfaces. In the modeling process, the flow in natural rivers is usually generalized as boundary conditions, and the process of waiting for cross-sectional flow data often cannot meet the requirements of forecast timeliness. Summary of the invention

[0004] Purpose of the invention: Based on the basin hydrological model and the urban hydrological model, a method for fine simulation of runoff generation that takes into account spatial heterogeneity is proposed, and a system for implementing the above method is further proposed. By dividing the underlying surface into two different types, urban underlying surface and natural underlying surface, different models are established for different underlying surfaces, and coupled calculations are performed through confluence points. At the same time, the optimization algorithm is used to calibrate the parameters of the coupling model, which improves the efficiency of model parameter calibration and model fitting accuracy, thereby effectively solving the above problems existing in the prior art.

[0005] First, a method for fine simulation of runoff generation considering spatial heterogeneity is proposed, which includes the following steps:

[0006] S1. Based on the digital elevation DEM data, remote sensing image data, and urban river / pipeline network information of the target area, the underlying surface type of the target area is divided into urban underlying surface and natural underlying surface; and a land use type map of the target area is generated;

[0007] S2, establishing an urban hydrological model based on the urban underlying surface, dividing different sub-catchments according to the urban river / pipeline network information, and calculating the surface runoff of different sub-catchments respectively;

[0008] S3. Establishing a hydrodynamic model of the urban river / pipeline network based on the urban river / pipeline network information and the surface runoff calculation results of different sub-catchments;

[0009] S4, divide several sub-basins according to the digital elevation DEM data of the area to which the natural underlying surface belongs, and calculate the runoff generation for each sub-basin respectively;

[0010] S5. Based on the flow calculation results of each sub-basin, a river confluence model is established; the confluence point between the urban river / pipeline network and the natural river is found using the river confluence model;

[0011] S6. Based on the hydrodynamic relationship of the confluence point and / or the confluence logic relationship, coupling the urban river / pipeline network hydrodynamic model and the river confluence model;

[0012] S7. Use the coupled model to simulate runoff generation in the target area.

[0013] In a further embodiment of the first aspect, in step S1, a K-Means clustering algorithm or a supervised classification algorithm is used to divide the underlying surface into two categories: urban underlying surface and natural underlying surface.

[0014] In a further embodiment of the first aspect, the underlying surface is divided into two categories: urban underlying surface and natural underlying surface using a K-Means clustering algorithm, specifically including:

[0015] S1-a-1, randomly divide the pixels in the remote sensing image data into k clusters, and calculate the initial center of each cluster;

[0016] S1-a-2, calculate the distance between each pixel and each initial center, and select the initial center closest to each pixel as the new center to form a new cluster;

[0017] S1-a-3, calculate the average value of the vector features of each new cluster, and use the average value of the vector features as the new center of the cluster;

[0018] S1-a-4, repeat steps S1-a-2 to S1-a-3, continue to form new clusters and calculate new centers, iterate until the center of the cluster no longer changes, and form the underlying surface type classification result;

[0019] S1-a-5, visually inspect whether there are errors in the underlying surface type classification result formed in step S1-a-4, and manually correct them if there are errors;

[0020] S1-a-6. Filter the underlying surface type classification result to eliminate small isolated pixel points in the underlying surface type classification result.

[0021] In a further embodiment of the first aspect, the underlying surface is divided into two categories: urban underlying surface and natural underlying surface using a supervised classification algorithm, specifically including:

[0022] S1-b-1, manually select a number of pixels whose feature significance meets the predetermined requirements and classify them as urban underlying surface or natural underlying surface, and create the classification results as a training sample set;

[0023] S1-b-2, constructing an underlying surface classification model, using a maximum likelihood method and / or a support vector machine algorithm and / or a random forest algorithm, and using the training sample set to train the underlying surface classification model;

[0024] S1-b-3, using the trained underlying surface classification model to classify the entire remote sensing image data, automatically classify each pixel into an urban underlying surface or a natural underlying surface, and form an underlying surface type classification result;

[0025] S1-b-4. Visually check whether there are any errors in the underlying surface type classification results. If there are any errors, correct them manually;

[0026] S1-b-5. Filter the underlying surface type classification result to eliminate small isolated pixel points in the underlying surface type classification result.

[0027] In a further embodiment of the first aspect, the sub-catchment area in step S2 includes a permeable area, an impermeable area with depressions, and an impermeable area without depressions; surface runoff generation is calculated for the permeable area, the impermeable area with depressions, and the impermeable area without depressions respectively;

[0028] For permeable areas, when the rainfall meets the surface infiltration conditions, water begins to accumulate on the ground, and when it exceeds the storage capacity of the depression, surface runoff is formed. At this time, the flow rate is The calculation formula is as follows:

[0029]

[0030] In the formula, i represents rainfall intensity; f represents the surface infiltration rate; represents the sampling interval of rainfall intensity;

[0031] For impermeable areas with depressions, runoff is formed after the rainfall meets the maximum depression storage capacity on the ground. The calculation formula is as follows:

[0032]

[0033] In the formula, P represents rainfall; D represents depression storage capacity;

[0034] For impermeable areas without depressions, rainfall is converted into runoff in addition to ground evaporation. When rainfall is greater than evaporation, runoff is formed. The calculation formula is as follows:

[0035]

[0036] Where P represents rainfall and E represents evaporation.

[0037] In a further embodiment of the first aspect, the urban river / pipeline network hydrodynamic model is represented by a nonlinear partial differential equation:

[0038]

[0039] For each time step, the finite difference method is used to solve the problem, so the net inflow and net outflow are averaged within each time step. and Respectively represent the initial value and final value of the water depth in a period, and the nonlinear partial differential equation is updated as follows:

[0040]

[0041] Where d is the water depth; A is the sub-basin area; V = A × d, V is the surface water volume; is the net rain; W is the sub-basin overland flow width; n is the Manning roughness coefficient; is the surface water storage depth; S is the sub-basin width; WCON is the flow routing parameter, .

[0042] In a further embodiment of the first aspect, in step S4, runoff generation calculation is performed on each sub-basin, including calculating initial soil moisture distribution and calculating rainfall-runoff relationship;

[0043] The initial soil moisture distribution is composed of the average initial soil moisture content in the basin. And its corresponding ordinate A is expressed as:

[0044]

[0045]

[0046] Where WMM represents the maximum water storage capacity at a single point in the basin; B represents the index of the water storage capacity-area distribution curve;

[0047] The mathematical form of the water storage capacity-area allocation curve is as follows:

[0048]

[0049] In the formula, f represents the runoff area; F represents the total basin area; It indicates the water storage capacity at a single point in the basin;

[0050] The average initial soil moisture content in the watershed is calculated After the corresponding ordinate A, the flow rate R is calculated according to the rainfall PE:

[0051] When A+PE≤WMM, the flow rate is ;

[0052] When A+PE>WMM, the flow rate is ;

[0053] Where WM represents the average tensile water capacity of the basin.

[0054] In a further embodiment of the first aspect, the river confluence model in step S5 includes surface runoff confluence, subsoil flow confluence, underground runoff confluence, total inflow per unit area of ​​river network, river network confluence per unit area, and river confluence below unit area;

[0055] Surface runoff confluence The expression is as follows:

[0056]

[0057] Where, CS represents the surface runoff decay coefficient; represents the surface runoff at time t; U represents the unit conversion coefficient, , represents the sampling interval of rainfall intensity, and F represents the total basin area;

[0058] Soil flow confluence The expression is as follows:

[0059]

[0060] Where, CI represents the soil flow recession coefficient; represents the subsoil flow runoff at time t;

[0061] Subsurface runoff confluence The expression is as follows:

[0062]

[0063] Where, CG represents the underground runoff recession coefficient; represents the underground runoff at time t;

[0064] Total inflow per unit area Surface runoff confluence , mid-stream confluence Merge with underground runoff sum:

[0065]

[0066] River network confluence per unit area The expression is as follows:

[0067]

[0068] In the formula, CR represents the water storage coefficient of the river network; L represents the lag time; Represents the outlet flow rate per unit area at time t-1.

[0069] In a further embodiment of the first aspect, in step S6, coupling the urban river / pipeline network hydrodynamic model and the river confluence model specifically includes:

[0070] When the water level at the pipe outlet is higher than the river level, the pipe drainage flows into the river, and the flow rate of the river downstream of the outlet is equal to the sum of the flow rate discharged into the river by the pipe network and the original flow rate of the river;

[0071] When the water level of the river is higher than the water level of the drain outlet, the drainage of the pipeline is blocked and the river water flows back into the pipeline. At this time, the water flow is pushed back into the pipe network. The flow of the river downstream of the drain outlet is equal to the original flow of the river minus the flow pushed back into the pipe network.

[0072] According to a second aspect of the present invention, a fine flow simulation system is proposed, which includes a division unit, a generation unit, a first calculation unit, a first model building unit, a second calculation unit, a second model building unit, a coupling unit, and a simulation result output unit.

[0073] The division unit divides the underlying surface type of the target area based on the digital elevation DEM data, remote sensing image data, and urban river / pipeline network information of the target area, and the underlying surface type is divided into urban underlying surface and natural underlying surface;

[0074] The generating unit generates a land use type map of the target area based on the underlying surface type of the target area;

[0075] The first calculation unit establishes an urban hydrological model based on the urban underlying surface, divides different sub-catchments according to the urban river / pipeline network information, and calculates the surface runoff of different sub-catchments respectively;

[0076] The first model building unit establishes a hydrodynamic model of the urban river / pipeline network based on the urban river / pipeline network information and the surface runoff calculation results of different sub-catchments;

[0077] The second calculation unit divides several sub-basins according to the digital elevation DEM data of the area to which the natural underlying surface belongs, and calculates the runoff generation for each sub-basin respectively;

[0078] The second model building unit establishes a river confluence model based on the flow calculation results of each sub-basin; the confluence point of the urban river / pipeline network and the natural river is found using the river confluence model;

[0079] The coupling unit couples the urban river / pipeline network hydrodynamic model and the river confluence model based on the hydrodynamic relationship of the confluence point and / or the confluence logic relationship;

[0080] The simulation result output unit uses the coupled model to simulate the flow generation in the target area.

[0081] The fine flow simulation system can be used to execute the fine flow simulation method disclosed in the first aspect and its further embodiments.

[0082] Beneficial effect: The present invention proposes a method for fine simulation of runoff generation taking into account spatial heterogeneity. By dividing two different types of underlying surfaces and modeling them separately for coupling, the spatial heterogeneity of the complex environment where the natural underlying surface and the urban underlying surface coexist is taken into account. This can effectively solve the problem that traditional urban hydrological models are unable to describe the spatial heterogeneity of the complex environment where the natural underlying surface and the urban underlying surface coexist in large cities. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] Figure 1 It is a flow chart of the fine simulation method of flow generation in an embodiment of the present invention.

[0084] Figure 2 This is a schematic diagram of a situation in which the water level at the pipeline drain outlet is higher than the river water level in an embodiment of the present invention.

[0085] Figure 3 This is a schematic diagram of a situation in which the water level of the river channel is higher than the water level of the drainage outlet in an embodiment of the present invention. DETAILED DESCRIPTION

[0086] In the following description, a large number of specific details are provided to provide a more thorough understanding of the present invention. However, it is apparent to those skilled in the art that the present invention can be implemented without one or more of these details. In other examples, in order to avoid confusion with the present invention, some technical features known in the art are not described.

[0087] In view of the fact that the watershed hydrological model assumes that the underlying surface within the unit is homogeneous and is less affected by human activities and it is difficult to express the relationship between rainfall and urban waterlogging; and the urban hydrological model cannot describe the spatial heterogeneity of the complex environment where the natural underlying surface and the urban underlying surface of a large city intersect and coexist, this embodiment proposes a fine simulation method for runoff generation that takes into account spatial heterogeneity based on the watershed hydrological model and the urban hydrological model. By dividing the underlying surface into two different types, urban underlying surface and natural underlying surface, different models are established for different underlying surfaces, and coupled calculations are performed through confluence points. At the same time, the optimization algorithm is used to calibrate the parameters of the coupling model, which improves the efficiency of model parameter calibration and the accuracy of model fitting. The following embodiment specifically illustrates this fine simulation method for runoff generation that takes into account spatial heterogeneity.

[0088] Embodiment 1:

[0089] This embodiment proposes a method for fine simulation of runoff generation considering spatial heterogeneity. The flow chart is shown in Figure 1 As shown, the steps are as follows:

[0090] Step 1: Collect digital elevation DEM data, remote sensing image data, and urban river / pipeline network information of the target area, and divide the underlying surface type of the target area, which is divided into urban underlying surface and natural underlying surface; generate a land use type map of the target area.

[0091] Step 2: Establish an urban hydrological model based on the urban underlying surface, divide different sub-catchments according to the pipe network information of the urban underlying surface, and classify the sub-catchments into permeable areas, impermeable areas with depressions, and impermeable areas without depressions for surface runoff calculation.

[0092] Step 3: Based on the urban river / pipeline network information and the surface runoff calculation results of different sub-catchments, a hydrodynamic model of the urban river / pipeline network is established.

[0093] Step 4: Divide the natural underlying surface into several sub-basins based on the digital elevation DEM data of the area, and calculate the runoff and divide the water source for each sub-basin.

[0094] Step 5: Based on the flow calculation results of each sub-basin, the Muskingum method or the one- or two-dimensional hydrodynamic model of the river is used to establish a river confluence model to perform confluence calculations. The confluence model is used to find the confluence point between the urban river / pipeline network and the natural river.

[0095] Step 6: Couple the urban river / pipeline network hydrodynamic model with the river confluence model based on the hydrodynamic relationship of the confluence point or the confluence logic relationship. Figure 2 As shown in the figure, the pipe drainage flows into the river, and the flow rate of the river downstream of the outlet is equal to the sum of the flow rate discharged into the river by the pipe network and the original flow rate of the river. Figure 3 As shown in the figure), the drainage of the pipeline is blocked, and the river water flows back into the pipeline. At this time, the water flow is pushed back into the pipe network. The flow of the river downstream of the drainage outlet is equal to the original flow of the river minus the flow pushed back into the pipe network.

[0096] Step 7: Automatically calibrate the coupled model based on the automatic optimization algorithm, and use the coupled model to simulate runoff generation in the target area.

[0097] Embodiment 2:

[0098] Based on Example 1, this example discloses two feasible solutions for dividing the underlying surface type of the target area in step 1. It should be noted that the two solutions disclosed in this example are only two preferred and feasible solutions, and do not mean to deliberately limit the protection scope of this application.

[0099] Based on the land use type map of the study area, the K-Means clustering algorithm or supervised learning method can be used to divide the underlying surface into two categories: urban underlying surface and natural underlying surface. The urban underlying surface refers to the ground surface that is greatly affected by human activities, has a mature drainage system, and is mostly covered by hardened ground and buildings. The natural underlying surface refers to the natural surface such as loose soil covered by vegetation or open wasteland, water area, etc. that is basically not interfered with by human activities.

[0100] The classification of the underlying surface type in the target area can be achieved by using the idea of ​​K-Means clustering algorithm, and its main implementation methods are as follows:

[0101] a. All remote sensing image pixels are randomly divided into k clusters, and the initial center of each cluster is calculated;

[0102] b. Calculate the distance between each pixel and each initial center. Each pixel selects the nearest initial center as the new center to form a new cluster.

[0103] c. Calculate the average value of the vector features of each new cluster generated in step b, and use this value as the new center of the cluster;

[0104] d. Repeat steps b and c to continue forming new clusters and calculating new centers, iterating until the center of the cluster no longer changes, thus forming the underlying surface type classification result.

[0105] e. Visual inspection: Check whether there are any errors in the underlying surface type classification results formed in step d. If there are any errors, correct them manually;

[0106] f. Filtering: Use filtering methods to eliminate small isolated pixels in the above-mentioned underlying surface type classification results, so that the underlying surface type is continuous and complete, and a land use type map is generated.

[0107] The classification of the underlying surface type in the target area can also be achieved by using the idea of ​​supervised classification algorithm, and its main implementation methods are as follows:

[0108] a. Create training samples. Manually select pixels with obvious features and classify them as urban underlying surfaces or natural underlying surfaces to create training sample sets. When creating training samples, areas with obvious features and as single pixel spectral features as possible should be selected as training samples.

[0109] b. training the model by using a maximum likelihood method, a support vector machine (SVM) algorithm, or a random forest algorithm, etc., and performing training on the training sample set created in step a;

[0110] c. Classification: classify the entire remote sensing image according to the model trained in step b, and classify each pixel into an urban underlying surface category or a natural underlying surface category to form an underlying surface type classification result;

[0111] d. Visual inspection: check whether there are any errors in the underlying surface type classification results formed in step c. If there are any errors, correct them manually;

[0112] e. Filtering: Use filtering methods to eliminate small isolated pixels in the above-mentioned underlying surface type classification results, make the underlying surface type continuous and complete, and generate a land use type map.

[0113] Embodiment 3:

[0114] On the basis of the previous embodiments, this embodiment discloses some calculation details, including the calculation methods for permeable areas, impermeable areas with depressions, and impermeable areas without depressions, the calculation method of urban river / pipeline network hydrodynamic model, sub-basin flow calculation and water source division calculation. It should be noted that the two schemes disclosed in this embodiment are only two preferred and feasible schemes, and do not mean to deliberately limit the scope of protection of this application.

[0115] The calculation methods for permeable areas, impermeable areas with depressions, and impermeable areas without depressions are as follows:

[0116] For permeable areas, when the rainfall meets the surface infiltration conditions, water begins to accumulate on the ground. When it exceeds the storage capacity of the depression, surface runoff is formed. The flow calculation formula is:

[0117]

[0118] Where R1 is the flow rate of the permeable area (mm);

[0119] i is the rainfall intensity (mm / hr);

[0120] f is the surface infiltration rate (mm / hr).

[0121] For the flow generation in impermeable areas with depressions, runoff can be formed after the rainfall meets the maximum depression storage capacity on the ground. The flow generation calculation formula is:

[0122]

[0123] Where R2 is the flow rate produced in the impervious area with depression storage capacity (mm);

[0124] P is rainfall (mm);

[0125] D is the depression volume (mm).

[0126] For impermeable areas without depressions, rainfall is basically converted into runoff except for ground evaporation. When rainfall is greater than evaporation, runoff can be formed. The runoff calculation formula is:

[0127]

[0128] Where, R3 is the flow rate of the non-depression permeable water storage area (mm);

[0129] P is rainfall (mm);

[0130] E is the evaporation amount (mm).

[0131] The calculation method of the urban river / pipeline network hydrodynamic model is as follows:

[0132] Under the same conditions, runoff is formed in the order of impervious areas without depressions, impervious areas with depressions, and permeable areas. Each subcatchment area uses a nonlinear reservoir model to calculate runoff according to the three surface types divided above. The nonlinear reservoir obtains the confluence result by simultaneously solving the continuity equation and the Manning formula.

[0133] The continuity equation is shown below:

[0134]

[0135] Where d is the water depth;

[0136] A is the sub-basin area;

[0137] V=A×d, V is the amount of surface water;

[0138] for pure rain;

[0139] Q is the outgoing flow.

[0140] The Manning formula is shown below:

[0141]

[0142] Where W is the sub-basin overland flow width;

[0143] n is the Manning roughness coefficient;

[0144] It is the depth of surface water storage;

[0145] S is the width of the subbasin.

[0146] Combining the above two equations, we get a nonlinear partial differential equation:

[0147]

[0148] The meanings of the physical quantities in the formula are the same as above.

[0149] In the above formula, the sub-basin overland flow width W, slope S and roughness n are combined into a parameter WCON, which is called the flow routing parameter:

[0150]

[0151] For each time step, the finite difference method is used to solve the problem. Therefore, the net inflow and net outflow must be averaged within each time step. The subscripts 1 and 2 represent the initial and final values ​​of the water depth in a period (Δt), respectively. The equation becomes

[0152]

[0153] The water depth d can be obtained by solving it using the New-Raphson iteration method.

[0154] The methods for calculating sub-basin runoff and dividing water sources are as follows:

[0155] The water storage capacity-area distribution curve is used to consider the uneven distribution of soil water shortage. The water storage capacity-area distribution curve can be used to determine the total runoff volume when the storage is fully stored and the rainfall is evenly distributed.

[0156] Practice shows that for closed basins, the basin water storage capacity-area distribution curve is preferably parabolic, and its form is:

[0157]

[0158] in:

[0159] f is the runoff area, km 2 ;

[0160] F is the total basin area, km 2 ;

[0161] is the water storage capacity at a single point in the basin, mm;

[0162] WMM is the maximum water storage capacity at a single point in the basin, mm;

[0163] B is the exponent of the water storage capacity-area distribution curve.

[0164] The calculation of rainfall runoff is divided into two steps: initial soil moisture distribution and calculation and rainfall runoff relationship calculation. For the distribution and calculation of initial soil moisture, there are two formulas, which can be simplified from the above formula:

[0165]

[0166]

[0167] in:

[0168] is the average initial soil moisture content in the watershed, mm;

[0169] A is the ordinate corresponding to the average initial soil moisture content in the basin, mm.

[0170] After calculating the initial soil wetness W0 of the basin and its corresponding ordinate A, the flow rate can be calculated based on the rainfall PE.

[0171] When A+PE≤WMM:

[0172]

[0173] When A+PE>WMM:

[0174]

[0175] Compared with the two-source water division, the three-source water division removes FC and uses a free water reservoir structure to solve the water source division problem. The free water reservoir structure takes into account the vertical regulation and storage function of the aeration zone. The total runoff R calculated by the full-storage runoff model first enters the free water reservoir for regulation and then divides the water source. The free water reservoir is equipped with two outlets on the runoff area, one is a side outlet, forming the soil flow RS; the other is a downward outlet, forming underground runoff RG. According to the concept of full-storage runoff, runoff is only possible on the runoff area FR. The runoff area is variable, so the bottom width FR of the free water reservoir is also variable.

[0176] The three-source water division structure uses a basin free water storage capacity-area curve similar to the basin storage capacity-area allocation curve to consider the problem of uneven distribution of free water storage capacity in the basin. The line shape of the basin free water storage capacity-area allocation curve is:

[0177]

[0178] in:

[0179] is the free water storage capacity at a single point in the basin, mm;

[0180] MS is the maximum free water storage capacity at a single point in the basin, mm;

[0181] EX is the exponent of the basin free water storage capacity-area distribution curve.

[0182] Similar to the basin storage capacity-area distribution curve, the basin free water storage capacity-area curve can also be calculated by integration of the initial value ordinate, surface runoff, subsoil flow and groundwater runoff.

[0183] The calculation method of the maximum free water storage capacity MS at a single point in a basin is similar to that of WMM:

[0184]

[0185] After assuming S0, the corresponding ordinate value AU can be obtained as:

[0186]

[0187] The runoff area FR is:

[0188]

[0189] The method for calculating surface runoff is similar to that for calculating flow production.

[0190] When PE+AU<MS, surface runoff RS is:

[0191]

[0192] When PE+AU≥MS, surface runoff RS is:

[0193]

[0194] The free water storage capacity for this period is:

[0195]

[0196] The corresponding subsurface flow and subsurface runoff are:

[0197]

[0198]

[0199] Among them, FR0 and FR are the ratios of runoff areas in the previous period and the current period.

[0200] The automatic optimization algorithm in step 7 can automatically calibrate the model parameters by using SCE-UA algorithm, simulated annealing algorithm, genetic algorithm, particle swarm algorithm, grey wolf algorithm, whale optimization algorithm, etc. This is not the innovation of the present invention, and there is sufficient information for those skilled in the art, so it will not be described in detail.

[0201] Embodiment 4:

[0202] Based on Example 3, this example discloses a feasible method for calculating river confluence. It should be noted that the two solutions disclosed in this example are only two preferred and feasible solutions, and do not mean to deliberately limit the protection scope of this application.

[0203] The confluence calculation of the three water sources is divided into six parts: surface runoff confluence, subsoil runoff confluence, groundwater runoff confluence, total inflow per unit area of ​​river network, confluence per unit area of ​​river network and confluence of rivers below unit area. Each part is explained one by one.

[0204] For surface runoff confluence, the linear reservoir calculation formula is:

[0205]

[0206] in:

[0207] QS is surface runoff, m3 / s;

[0208] CS is the surface runoff decay coefficient;

[0209] RS is the surface runoff, mm.

[0210] For the mid-flow confluence, the linear reservoir calculation formula is:

[0211]

[0212] in:

[0213] QI is soil flow, m3 / s;

[0214] CI is the soil flow recession coefficient;

[0215] RI is the soil flow runoff, mm.

[0216] For underground runoff confluence, the calculation formula for the linear reservoir is:

[0217]

[0218] in:

[0219] QG is underground runoff, m3 / s;

[0220] CG is the subsurface runoff recession coefficient;

[0221] RG is underground runoff, mm.

[0222] The total inflow of a river network per unit area is the sum of surface runoff, subsurface flow and underground runoff:

[0223]

[0224] The unit area river network confluence adopts the hysteresis algorithm. The calculation formula is:

[0225]

[0226] in:

[0227] Q is the outlet flow rate per unit area, m3 / s;

[0228] CR is the water storage recession coefficient of the river network;

[0229] L is the lag time, h.

[0230] Muskingum segmented continuous algorithm is used for river confluence below unit area. The parameters are channel storage coefficient KE(h) and flow weight factor XE, and the parameters of each unit river section take the same value. In order to ensure the two linear conditions of Muskingum method, KE≈Δt is taken for each unit river section. The formula is as follows:

[0231]

[0232] For the hydrodynamic confluence method, it is consistent with the method used in step S4.

[0233] Embodiment 5:

[0234] This embodiment discloses a fine simulation system for flow generation, which consists of eight parts: a division unit, a generation unit, a first calculation unit, a first model building unit, a second calculation unit, a second model building unit, a coupling unit, and a simulation result output unit.

[0235] The division unit divides the underlying surface type of the target area based on the digital elevation DEM data, remote sensing image data, and urban river / pipeline network information of the target area, and the underlying surface type is divided into urban underlying surface and natural underlying surface. The generation unit generates a land use type map of the target area based on the underlying surface type of the target area. The first calculation unit establishes an urban hydrological model based on the urban underlying surface, divides different sub-catchments according to the urban river / pipeline network information, and calculates the surface runoff of different sub-catchments respectively. The first model construction unit establishes an urban river / pipeline network hydrodynamic model based on the urban river / pipeline network information and the surface runoff calculation results of different sub-catchments. The second calculation unit divides several sub-basins according to the digital elevation DEM data of the area to which the natural underlying surface belongs, and calculates the runoff of each sub-basin respectively. The second model construction unit establishes a river confluence model based on the runoff calculation results of each sub-basin; and uses the river confluence model to find the confluence point of the urban river / pipeline network and the natural river. The coupling unit couples the urban river / pipeline network hydrodynamic model and the river confluence model based on the hydrodynamic relationship and / or confluence logic relationship of the confluence point. The simulation result output unit uses the coupled model to simulate the flow generation in the target area. The fine flow generation simulation system can automatically execute the fine flow generation simulation method considering spatial heterogeneity disclosed in the above-mentioned embodiments 1 to 4, and this process is not described in detail.

[0236] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the present invention itself. Various changes may be made to it in form and detail without departing from the spirit and scope of the present invention as defined in the appended claims.

Claims

1. A method for fine simulation of runoff generation considering spatial heterogeneity, characterized in that: The steps include: S1. Based on the digital elevation DEM data, remote sensing image data, and urban river / pipeline network information of the target area, the underlying surface type of the target area is divided into urban underlying surface and natural underlying surface; Generate a land use type map of the target area; S2, establishing an urban hydrological model based on the urban underlying surface, dividing different sub-catchments according to the urban river / pipeline network information, and calculating the surface runoff of different sub-catchments respectively; S3. Establishing a hydrodynamic model of the urban river / pipeline network based on the urban river / pipeline network information and the surface runoff calculation results of different sub-catchments; S4, divide several sub-basins according to the digital elevation DEM data of the area to which the natural underlying surface belongs, and calculate the runoff generation for each sub-basin respectively; S5. Based on the flow calculation results of each sub-basin, a river confluence model is established; the confluence point between the urban river / pipeline network and the natural river is found using the river confluence model; The river confluence model includes surface runoff confluence, subsoil flow confluence, underground runoff confluence, total inflow of river network per unit area, river network confluence per unit area, and river confluence below unit area; Surface runoff confluence The expression is as follows: ； Where, CS represents the surface runoff decay coefficient; represents the surface runoff at time t; U represents the unit conversion coefficient, , represents the sampling interval of rainfall intensity, and F represents the total basin area; Soil flow confluence The expression is as follows: ； Where, CI represents the soil flow recession coefficient; represents the subsoil flow runoff at time t; Subsurface runoff confluence The expression is as follows: ； Where, CG represents the underground runoff recession coefficient; represents the underground runoff at time t; Total inflow per unit area Surface runoff confluence , mid-stream confluence Merge with underground runoff sum: ； River network confluence per unit area The expression is as follows: ； In the formula, CR represents the water storage coefficient of the river network; L represents the lag time; represents the outlet flow rate per unit area at time t-1; S6. Based on the hydrodynamic relationship of the confluence point and / or the confluence logic relationship, coupling the urban river / pipeline network hydrodynamic model and the river confluence model, specifically including: When the water level at the pipe outlet is higher than the river level, the pipe drainage flows into the river, and the flow rate of the river downstream of the outlet is equal to the sum of the flow rate discharged into the river by the pipe network and the original flow rate of the river; When the water level of the river is higher than the water level of the drainage outlet, the drainage of the pipeline is blocked, and the river water flows back into the pipeline. At this time, the water flow is pushed back into the pipe network. The flow of the river downstream of the drainage outlet is equal to the original flow of the river minus the flow pushed back into the pipe network. S7. Use the coupled model to simulate runoff generation in the target area.

2. The method for fine simulation of runoff generation considering spatial heterogeneity according to claim 1 is characterized in that: In step S1, the underlying surface is divided into two categories: urban underlying surface and natural underlying surface using a K-Means clustering algorithm or a supervised classification algorithm.

3. The method for fine simulation of runoff generation considering spatial heterogeneity according to claim 2 is characterized in that: The K-Means clustering algorithm is used to divide the underlying surface into two categories: urban underlying surface and natural underlying surface, including: S1-a-1, randomly divide the pixels in the remote sensing image data into k clusters, and calculate the initial center of each cluster; S1-a-2, calculate the distance between each pixel and each initial center, and select the initial center closest to each pixel as the new center to form a new cluster; S1-a-3, calculating the average vector feature value of each new cluster, and taking the average vector feature value as the center of the new cluster; S1-a-4, repeat steps S1-a-2 to S1-a-3, continue to form new clusters and calculate new centers, iterate until the center of the cluster no longer changes, and form the underlying surface type classification result; S1-a-5, visually inspect whether there are errors in the underlying surface type classification result formed in step S1-a-4, and manually correct them if there are errors; S1-a-6. Filter the underlying surface type classification result to eliminate small isolated pixel points in the underlying surface type classification result.

4. The method for fine simulation of runoff generation considering spatial heterogeneity according to claim 2 is characterized in that: The supervised classification algorithm is used to divide the underlying surface into two categories: urban underlying surface and natural underlying surface, including: S1-b-1, manually select a number of pixels whose feature significance meets the predetermined requirements and classify them as urban underlying surface or natural underlying surface, and create the classification results as a training sample set; S1-b-2, constructing an underlying surface classification model, using a maximum likelihood method and / or a support vector machine algorithm and / or a random forest algorithm, and using the training sample set to train the underlying surface classification model; S1-b-3, using the trained underlying surface classification model to classify the entire remote sensing image data, automatically classify each pixel into an urban underlying surface or a natural underlying surface, and form an underlying surface type classification result; S1-b-4. Visually check whether there are any errors in the underlying surface type classification results. If there are any errors, correct them manually; S1-b-5. Filter the underlying surface type classification result to eliminate small isolated pixel points in the underlying surface type classification result.

5. The method for fine simulation of runoff generation considering spatial heterogeneity according to claim 1 is characterized in that: The sub-catchment area in step S2 includes a permeable area, an impermeable area with depressions, and an impermeable area without depressions; surface runoff generation is calculated for the permeable area, the impermeable area with depressions, and the impermeable area without depressions respectively; For permeable areas, when the rainfall meets the surface infiltration conditions, water begins to accumulate on the ground, and when it exceeds the storage capacity of the depression, surface runoff is formed. At this time, the flow rate is The calculation formula is as follows: ； In the formula, i represents rainfall intensity; f represents the surface infiltration rate; represents the sampling interval of rainfall intensity; For impermeable areas with depressions, runoff is formed after the rainfall meets the maximum depression storage capacity on the ground. The calculation formula is as follows: ； In the formula, P represents rainfall; D represents depression storage capacity; For impermeable areas without depressions, rainfall is converted into runoff in addition to ground evaporation. When rainfall is greater than evaporation, runoff is formed. The calculation formula is as follows: ； Where P represents rainfall and E represents evaporation.

6. The method for fine simulation of runoff generation considering spatial heterogeneity according to claim 5 is characterized in that: The urban river / pipeline network hydrodynamic model is represented by a nonlinear partial differential equation: ； For each time step, the finite difference method is used to solve the problem, so the net inflow and net outflow are averaged within each time step. and Respectively represent the initial value and final value of the water depth in a period, and the nonlinear partial differential equation is updated as follows: ； Where d is the water depth; A is the sub-basin area; is the net rain; W is the sub-basin overland flow width; n is the Manning roughness coefficient; is the surface water storage depth; S is the sub-basin width; WCON is the flow routing parameter, .

7. The method for fine simulation of runoff generation considering spatial heterogeneity according to claim 1 is characterized in that: In step S4, the runoff generation of each sub-basin is calculated, including calculating the initial soil moisture distribution and calculating the rainfall-runoff relationship; The initial soil moisture distribution is composed of the average initial soil moisture content in the basin. And its corresponding ordinate A is expressed as: ； ； Where WMM represents the maximum water storage capacity at a single point in the basin; B represents the index of the water storage capacity-area distribution curve; The mathematical form of the water storage capacity-area allocation curve is as follows: ； In the formula, f represents the runoff area; F represents the total basin area; It indicates the water storage capacity at a single point in the basin; The average initial soil moisture content in the watershed is calculated After the corresponding ordinate A, the flow rate R is calculated according to the rainfall PE: When A+PE≤WMM, the flow rate is ; When A+PE>WMM, the flow rate is ; Where WM represents the average tensile water capacity of the basin.

8. A fine simulation system for runoff generation, which can be used to implement the fine simulation method for runoff generation considering spatial heterogeneity as described in any one of claims 1 to 7; characterized in that: The fine flow simulation system includes: A division unit; the division unit divides the underlying surface type of the target area based on the digital elevation DEM data, remote sensing image data, and urban river / pipeline network information of the target area, and the underlying surface type is divided into an urban underlying surface and a natural underlying surface; Generating unit; the generating unit generates a land use type map of the target area based on the underlying surface type of the target area; A first calculation unit; the first calculation unit establishes an urban hydrological model based on the urban underlying surface, divides different sub-catchments according to the urban river / pipeline network information, and calculates the surface runoff of different sub-catchments respectively; A first model building unit; the first model building unit establishes a hydrodynamic model of the urban river / pipeline network based on the urban river / pipeline network information and the surface runoff calculation results of different sub-catchments; The second calculation unit: the second calculation unit divides a plurality of sub-basins according to the digital elevation DEM data of the area to which the natural underlying surface belongs, and performs flow generation calculation on each sub-basin respectively; The second model building unit: The second model building unit establishes a river confluence model based on the flow calculation results of each sub-basin; and uses the river confluence model to find the confluence point between the urban river / pipeline network and the natural river; A coupling unit; the coupling unit couples the urban river / pipeline network hydrodynamic model and the river confluence model based on the hydrodynamic relationship of the confluence point and / or the confluence logic relationship; A simulation result output unit; the simulation result output unit uses the coupled model to simulate the flow generation in the target area.

Images