A method for urban waterlogging forecasting based on network forecasting

Abstract

The present invention provides a method for predicting urban waterlogging based on network forecasting. It includes: obtaining the distribution of rain gauges in the target study area and the historical rainfall data and typical rainfall processes of the rain gauges; obtaining the rainstorm waterlogging points and waterlogging depths measured by the rain gauges within the current effective time to obtain the spatial distribution of rainfall, and performing grid division to obtain the subspace distribution of each grid area; establishing an urban waterlogging prediction model based on the historical rainfall data and typical rainfall processes; performing waterlogging prediction based on the urban waterlogging prediction model and in combination with the subspace distribution; judging the difficulty of real-time scenario forecasting for each grid area, and adjusting the corresponding prediction results, thereby forecasting the urban waterlogging situation. The urban waterlogging prediction model is constructed using the rainfall data of the rain gauges, and is divided according to the spatial distribution of the rain gauges. Based on the waterlogging situation of each divided area, the urban waterlogging prediction model is optimized to achieve accurate prediction of urban waterlogging.

Description

Technical Field

[0001] The present invention relates to the field of urban waterlogging forecasting, and in particular to an urban waterlogging forecasting method based on network forecasting. Background Art

[0002] At present, with the acceleration of urbanization and the increase in the surface hardening rate, the city's water storage and retention capacity has weakened. In addition, extreme rainstorms occur frequently during the flood season. Every year during the flood season, "sea-viewing" landscapes appear in various places, which greatly affects urban travel and life and production, and even causes urban flooding disasters, endangering the safety of life and property.

[0003] However, the existing urban flooding calculation models simply predict urban flooding based on historical weather forecasts, which cannot meet the forecast accuracy requirements for urban flooding risk avoidance and disposal capabilities.

[0004] Therefore, the present invention provides a method for predicting urban waterlogging based on network forecasting. Summary of the Invention

[0005] The present invention provides an urban waterlogging forecasting method based on network forecasting, which is used to construct an urban waterlogging prediction model through rainfall data from urban rain gauges, and obtain the spatial distribution of rain gauges through waterlogging points and waterlogging depths, so as to divide the city. Based on the waterlogging situation in each divided area, the urban waterlogging prediction model is optimized respectively, and a forecast is made based on the comprehensive results of all optimization results, thereby achieving a more accurate prediction of urban waterlogging.

[0006] The present invention provides a method for predicting urban waterlogging based on network forecasting, comprising:

[0007] Step 1: Obtain the distribution of rain gauges in the target study area, and obtain historical rainfall data and typical rainfall processes for each rain gauge;

[0008] Step 2: Obtain the rainstorm waterlogging points and waterlogging depths measured by each rain gauge in the target study area within the current effective time, obtain the spatial distribution of rainfall, and divide the target study area into grids to obtain the subspace distribution of each divided grid area;

[0009] Step 3: Establish an urban waterlogging prediction model based on historical rainfall data and typical rainfall processes;

[0010] Step 4: Based on the urban waterlogging prediction model and in combination with the subspace distribution of each divided grid area, predict the waterlogging situation of the corresponding subspace;

[0011] Step 5: Determine the difficulty of real-time scenario forecasting for each divided grid area, adjust the corresponding forecast results, and then forecast the urban flooding situation.

[0012] In one possible implementation, the distribution of rain gauges in the target study area is obtained, and historical rainfall data and typical rainfall processes of each rain gauge are obtained, including:

[0013] Step 11: Obtain the rain gauge station table number and its distribution location in the target study area, and obtain the historical rainfall data of each rain gauge in the target study area. Based on the historical rainfall data, obtain the typical rainfall process of each rain gauge station;

[0014] Step 12: Construct a historical rainfall data table for each rain gauge based on the historical rainfall data and the corresponding typical rainfall process.

[0015] In one possible implementation, the spatial distribution of rainfall is obtained by obtaining the rainstorm water points and water depths measured by each rain gauge in the target study area within the current effective time, including:

[0016] Step 21: Obtain all rainstorm water points and water depths measured by each rain gauge in the target study area within the effective measurement time;

[0017] Step 22: Determine the comprehensive distribution of the rainstorm waterlogging points and the average value of the corresponding waterlogging depths measured by the current rain gauge within the effective measurement time based on the rainstorm waterlogging points and the corresponding waterlogging depths;

[0018] Step 23: Determine the spatial distribution of rainfall in the target study area based on the comprehensive distribution of corresponding rainstorm waterlogging points and the mean waterlogging depth measured by all rain gauges.

[0019] In one possible implementation, the target study area is divided into grids to obtain the subspace distribution of each divided grid area, including:

[0020] Step 01: Based on the spatial distribution of rainfall in the target study area, the target study area is divided into the first grid using Thiessen polygons;

[0021] Step 02: Obtain the rainfall time distribution of the target study area through the historical rainfall data in the historical rainfall data table corresponding to each rain gauge, thereby performing a second grid division on the target study area. The second grid division result is adjusted based on the typical rainfall process to obtain the third grid division result.

[0022] Step 03: Overlap and integrate the first and third meshing results to obtain a comprehensive meshing result.

[0023] Step 04: Based on the comprehensive division results, the subspace distribution of each divided grid area in the target study area is obtained.

[0024] In one possible implementation approach, an urban flooding prediction model is established based on historical rainfall data and typical rainfall processes, including:

[0025] Step 31: Based on the latest rainfall data of the target study area and the corresponding waterlogging points and waterlogging depths, an initial urban waterlogging prediction model is obtained;

[0026] Step 32: Based on the historical rainfall data of the target study area and the corresponding waterlogging points and waterlogging depths, the initial urban waterlogging prediction model is trained to obtain the urban waterlogging prediction model.

[0027] In one possible implementation, based on the urban waterlogging prediction model and in combination with the subspace distribution of each divided grid area, the waterlogging situation of the corresponding subspace is predicted, including:

[0028] Step 41: Obtain the trained urban waterlogging prediction model and perform model correction based on the subspace distribution of each grid area to obtain a corrected urban waterlogging prediction model corresponding to each grid area, thereby obtaining a corrected urban waterlogging prediction set. The corrected urban waterlogging prediction set includes a corrected urban waterlogging prediction model corresponding to each grid area.

[0029] Step 42: Predicting the waterlogging situation of the corresponding divided grid area based on each modified urban waterlogging prediction model to obtain an urban waterlogging prediction set for the target study area;

[0030] The urban waterlogging prediction set includes the spatial number of the subspace corresponding to the divided grid area, the regional location, and the urban waterlogging prediction result of the corresponding subspace.

[0031] In one possible implementation, the difficulty of real-time scenario forecasting for each gridded area is determined, and the corresponding forecast results are adjusted to forecast urban flooding, including:

[0032] Step 51: Obtain several groups of historical forecast rainfall and historical measured rainfall corresponding to different forecast lengths for all rain gauges included in each divided grid area;

[0033] Step 52: Compare each set of historical forecast rainfall and historical measured rainfall of the currently divided grid area with the preset rainfall level;

[0034] If the historical forecast rainfall and the historical measured rainfall in the same group are in the same rainfall level, the corresponding weather scenario and the historical forecast rainfall and historical measured rainfall will be classified into the first forecast category;

[0035] If the historical forecast rainfall and the historical measured rainfall in the same group are not in the same rainfall level, the historical forecast rainfall, the historical measured rainfall and their corresponding two weather scenarios are classified into the second forecast category;

[0036] Step 53: determining a first forecast number based on the same rainfall station in the first forecast category and determining a second forecast number based on the same rainfall station in the second forecast category;

[0037] Step 54: Based on the first forecast number and the second forecast number, obtain a comprehensive rainfall forecast difficulty coefficient for the current grid area;

[0038] ;

[0039] ;in, is the rainfall forecast difficulty coefficient of the g-th rain gauge in the current grid area; m is the number of forecasters for the g-th rain gauge in the current grid area; n is the number of forecast scenarios in the current grid area; The number of qualified forecasts made by the i-th forecaster at the g-th rainfall station in the current grid area under the j-th forecast scenario; is the number of unqualified forecasts made by the i-th forecaster at the g-th rainfall station in the current grid area under the j-th forecast scenario; is the first forecast number corresponding to the g-th rainfall station; is the second forecast number corresponding to the g-th rainfall station; The first weight set for the overall calculation; A second weight set for the average calculation; is the comprehensive rainfall forecast difficulty coefficient in the current grid area; s represents the number of rain gauges in the current grid area; Indicates the maximum forecast error probability in the current divided area grid; represents the average forecast error probability in the current divided area grid;

[0040] Step 55: Based on the comprehensive rainfall forecast difficulty coefficient, obtain the forecast difficulty of the real-time weather scenario of the current divided grid area;

[0041] Step 56: Combine the forecast difficulty of the real-time weather scenario in the current grid area and the forecast error probability of each rain gauge in the current grid area. , adjust the urban waterlogging prediction result of the current grid area, wherein the adjusted urban waterlogging prediction result is the waterlogging prediction result of the current grid area;

[0042] Step 57: Based on the waterlogging prediction result of the current divided grid area, obtain the urban waterlogging prediction result and the corresponding waterlogging index of the current divided grid area, and perform waterlogging forecast.

[0043] In one possible implementation, based on the waterlogging prediction result of the current grid area, the urban waterlogging prediction result and the corresponding waterlogging index of the current grid area are obtained, and waterlogging forecast is performed, including:

[0044] Step 571: Compare the waterlogging prediction result of the current grid area with a preset waterlogging prediction table to obtain a waterlogging index corresponding to the waterlogging prediction result of the current grid area.

[0045] Step 572: Determine whether the waterlogging index is within a first waterlogging preset range;

[0046] If the waterlogging index is within the first waterlogging preset range, it is determined that the current rainfall will not cause regional waterlogging in the current grid area;

[0047] On the contrary, it is judged that the current rainfall will cause waterlogging in the region, and the waterlogging level and waterlogging possibility of the subspace region are determined based on the corresponding level of the waterlogging index;

[0048] Step 573: Based on the waterlogging levels of all divided grid areas of the target study area and the waterlogging possibility of each divided grid area, determine the comprehensive waterlogging level of the target study area and the waterlogging possibility of different target study subspaces, and perform waterlogging forecast.

[0049] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purposes and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

[0050] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0052] Figure 1 This is a flow chart of a method for urban waterlogging forecasting based on network forecasting in an embodiment of the present invention;

[0053] Figure 2 This is a flow chart of obtaining spatial distribution of rainfall in a method for urban waterlogging forecasting based on network forecasting in an embodiment of the present invention;

[0054] Figure 3 This is a flow chart of performing urban waterlogging forecasting based on waterlogging prediction results in an urban waterlogging forecasting method based on network forecasting in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0056] Example 1:

[0057] The embodiment of the present invention provides a method for urban waterlogging forecasting based on network forecasting, such as Figure 1 Shown, including:

[0058] Step 1: Obtain the distribution of rain gauges in the target study area, and obtain historical rainfall data and typical rainfall processes for each rain gauge;

[0059] Step 2: Obtain the rainstorm waterlogging points and waterlogging depths measured by each rain gauge in the target study area within the current effective time, obtain the spatial distribution of rainfall, and divide the target study area into grids to obtain the subspace distribution of each divided grid area;

[0060] Step 3: Establish an urban waterlogging prediction model based on historical rainfall data and typical rainfall processes;

[0061] Step 4: Based on the urban waterlogging prediction model and in combination with the subspace distribution of each divided grid area, predict the waterlogging situation of the corresponding subspace;

[0062] Step 5: Determine the difficulty of real-time scenario forecasting for each divided grid area, adjust the corresponding forecast results, and then forecast the urban flooding situation.

[0063] In this embodiment, the target study area refers to an urban area where urban waterlogging forecasting is required.

[0064] In this embodiment, the distribution of rain gauges refers to the distribution location of each rain gauge, and the distribution location of the rain gauges can be represented by longitude and latitude (geographic coordinates).

[0065] In this embodiment, the historical rainfall data refers to rainfall data of historical rainfall processes collected by rain gauges in the target study area.

[0066] In this embodiment, the typical rainfall process refers to each rain gauge classifying and aggregating rainfall data of the same rainfall type according to the historical rainfall data of the rain gauge, and analyzing the typical rainfall process corresponding to the current rainfall type based on the aggregation results.

[0067] In this embodiment, a rainstorm water accumulation point refers to the coordinates of a location point where water accumulates within the measurement range of the current rain gauge during historical rainfall when the rainfall reaches a certain level.

[0068] In this embodiment, the water depth refers to the water depth at a location where water accumulates within the measurement range of the current rain gauge.

[0069] In this embodiment, the spatial distribution of rainfall refers to determining the spatial distribution of rainfall in the target study area based on the comprehensive distribution of corresponding rainstorm waterlogging points and the average waterlogging depth measured by all rain gauges.

[0070] In this embodiment, the subspace distribution refers to dividing the target research area into multiple grid areas based on the comprehensive division result, and the spatial distribution of each grid area is the corresponding subspace distribution.

[0071] In this embodiment, the urban waterlogging prediction model refers to an urban waterlogging prediction model for the target study area obtained by training an initial urban waterlogging prediction model based on historical rainfall data, waterlogging points, and corresponding waterlogging depths in the target study area.

[0072] In this embodiment, the real-time scenario forecast difficulty refers to the difficulty of determining the real-time rainfall forecast based on the real-time weather scenario transmitted to the comprehensive rainfall forecast difficulty calculation formula.

[0073] In this embodiment, for example, when the prediction result is that the rainfall at location 1 is 10 mm, but due to historical error reports, the historical error report pattern is found and the 10 mm is adjusted, for example, to 8 mm. At this time, a flood forecast can be performed. The reason for the error report may be an inaccurate model or an inaccurate measurement at the rain gauge.

[0074] The beneficial effects of the above technical solution are: building an urban waterlogging prediction model through rainfall data from urban rain gauges, and obtaining the spatial distribution of rain gauges through waterlogging points and water depths, thereby dividing the city, and based on the waterlogging situation in each divided area, optimizing the urban waterlogging prediction model separately, and making forecasts based on the comprehensive results of all optimization results, to achieve more accurate predictions of urban waterlogging.

[0075] Example 2:

[0076] Based on Example 1, the distribution of rain gauges in the target study area is obtained, and the historical rainfall data and typical rainfall processes of each rain gauge are obtained, including:

[0077] Step 11: Obtain the rain gauge station table number and its distribution location in the target study area, and obtain the historical rainfall data of each rain gauge in the target study area. Based on the historical rainfall data, obtain the typical rainfall process of each rain gauge station;

[0078] Step 12: Construct a historical rainfall data table for each rain gauge based on the historical rainfall data and the corresponding typical rainfall process.

[0079] In this embodiment, the target study area refers to an urban area where urban waterlogging forecasting is required.

[0080] In this embodiment, the rain gauge table numbers and their distribution locations refer to the numbers of all rain gauges in the historical rainfall data table of the target study area and the rain gauge distribution locations corresponding to each rain gauge, where the rain gauge distribution locations can be represented by latitude and longitude.

[0081] In this embodiment, the historical rainfall data refers to rainfall data of historical rainfall processes collected by rain gauges in the target study area.

[0082] In this embodiment, the typical rainfall process refers to each rain gauge classifying and aggregating rainfall data of the same rainfall type according to the historical rainfall data of the rain gauge, and analyzing the typical rainfall process corresponding to the current rainfall type based on the aggregation results.

[0083] In this embodiment, the historical rainfall data table refers to a rainfall data table constructed based on historical rainfall data and corresponding typical rainfall processes, wherein each rainfall station corresponds to a rainfall data table.

[0084] The beneficial effects of the above technical solution are: by obtaining rainfall data from urban rain gauges, an accurate urban waterlogging prediction model is constructed, and the urban waterlogging prediction model is optimized through the spatial distribution of rain gauges, and forecasts are made based on the comprehensive results of all optimization results, so as to achieve a more accurate prediction of urban waterlogging.

[0085] Example 3:

[0086] Based on Example 2, the rainstorm water points and water depths measured by each rain gauge in the target study area within the current effective time are obtained to obtain the spatial distribution of rainfall, such as Figure 2 Shown, including:

[0087] Step 21: Obtain all rainstorm water points and water depths measured by each rain gauge in the target study area within the effective measurement time;

[0088] Step 22: Determine the comprehensive distribution of the rainstorm waterlogging points and the average value of the corresponding waterlogging depths measured by the current rain gauge within the effective measurement time based on the rainstorm waterlogging points and the corresponding waterlogging depths;

[0089] Step 23: Determine the spatial distribution of rainfall in the target study area based on the comprehensive distribution of corresponding rainstorm waterlogging points and the mean waterlogging depth measured by all rain gauges.

[0090] In this embodiment, the effective measurement time refers to the time when the current rain gauge performs effective rainfall data measurement.

[0091] In this embodiment, a rainstorm water accumulation point refers to the coordinates of a location point where water accumulates within the measurement range of the current rain gauge during historical rainfall when the rainfall reaches a certain level.

[0092] In this embodiment, the water depth refers to the water depth at a location where water accumulates within the measurement range of the current rain gauge.

[0093] In this embodiment, the comprehensive distribution of rainstorm waterlogging points refers to the spatial distribution of the remaining waterlogging points in the target study area after integrating all rainstorm waterlogging points that have appeared during the effective measurement time at the current rain gauge and eliminating the waterlogging points that no longer accumulate water after being repaired.

[0094] In this embodiment, the mean water depth refers to the average value of all water depths corresponding to each water point during historical rainfall processes.

[0095] In this embodiment, the spatial distribution of rainfall refers to determining the spatial distribution of rainfall in the target study area based on the comprehensive distribution of corresponding rainstorm waterlogging points and the average waterlogging depth measured by all rain gauges.

[0096] The beneficial effects of the above technical solution are: the spatial distribution of rain gauges is obtained through waterlogging points and waterlogging depths, thereby dividing the city, and based on the waterlogging situation in each divided area, the urban waterlogging prediction model is optimized, and then forecasts are made based on the comprehensive results of all optimization results, achieving more accurate predictions of urban waterlogging.

[0097] Example 4:

[0098] Based on Example 1, the target research area is divided into grids to obtain the subspace distribution of each divided grid area, including:

[0099] Step 01: Based on the spatial distribution of rainfall in the target study area, the target study area is divided into the first grid using the Thiessen polygon method;

[0100] Step 02: Obtain the rainfall time distribution of the target study area through the historical rainfall data in the historical rainfall data table corresponding to each rain gauge, thereby performing a second grid division on the target study area. The second grid division result is adjusted based on the typical rainfall process to obtain the third grid division result.

[0101] Step 03: Overlap and integrate the first and third meshing results to obtain a comprehensive meshing result.

[0102] Step 04: Based on the comprehensive division results, the subspace distribution of each divided grid area in the target study area is obtained.

[0103] In this embodiment, the spatial distribution of rainfall refers to determining the spatial distribution of rainfall in the target study area based on the comprehensive distribution of corresponding rainstorm waterlogging points and the average waterlogging depth measured by all rain gauges.

[0104] In this embodiment, the Thiessen polygon refers to dividing the watershed into a number of polygons using the perpendicular bisectors of the lines connecting the rain gauges. The rainfall data can be calculated more accurately after the area division using the Thiessen polygon method.

[0105] In this embodiment, the first grid division refers to performing a first division on the target study area based on the Thiessen polygon method, thereby dividing the target study area into a plurality of grid areas based on the spatial distribution of rainfall.

[0106] In this embodiment, the rainfall time distribution refers to determining the rainfall time distribution based on the rainfall time and rainfall duration corresponding to historical rainfall data.

[0107] In this embodiment, the second grid division is to perform a second division on the target study area based on the rainfall time distribution in the target study area, thereby dividing the target study area into a plurality of grid areas based on the rainfall time.

[0108] In this embodiment, the third grid division is obtained by adjusting the second grid division result based on rainfall conditions in the target study area during a typical rainfall process.

[0109] In this embodiment, result overlapping and result integration refer to overlapping the first grid division result and the third grid division result according to the corresponding longitude and latitude, and sorting the overlapping results, thereby adjusting the division area to obtain a comprehensive grid division result.

[0110] In this embodiment, the comprehensive division result refers to a division result obtained by overlapping and integrating the first grid division result and the third grid division result.

[0111] In this embodiment, the subspace distribution refers to dividing the target research area into multiple grid areas based on the comprehensive division result, and the spatial distribution of each grid area is the corresponding subspace distribution.

[0112] The beneficial effects of the above technical solution are: the spatial distribution of rain gauges is obtained through waterlogging points and waterlogging depths, and the temporal distribution of rainfall is obtained through historical rainfall data. The target study area is divided based on the two distribution results, and the urban waterlogging prediction model is optimized based on the waterlogging situation in each divided area, thereby achieving a more accurate prediction of urban waterlogging.

[0113] Example 5:

[0114] Based on Example 4, an urban waterlogging prediction model is established based on historical rainfall data and typical rainfall processes, including:

[0115] Step 31: Based on the latest rainfall data of the target study area and the corresponding waterlogging points and waterlogging depths, an initial urban waterlogging prediction model is obtained;

[0116] Step 32: Based on the historical rainfall data of the target study area and the corresponding waterlogging points and waterlogging depths, the initial urban waterlogging prediction model is trained to obtain the urban waterlogging prediction model.

[0117] In this embodiment, the rainfall data refers to the rainfall data collected by each rain gauge in the most recent rainfall in the target study area, wherein the rainfall data includes the table number of the measuring rain gauge, the longitude and latitude of the measuring point, the rainfall in the first period, the rainfall in the second period, the rainfall in the third period, etc.

[0118] In this embodiment, the initial urban waterlogging prediction model refers to constructing an initial urban waterlogging prediction model based on the most recent rainfall data and the corresponding waterlogging points and waterlogging depths in the target study area, wherein the input of the initial urban waterlogging prediction model is rainfall data, waterlogging points and waterlogging depths, and the output is the waterlogging prediction result.

[0119] In this embodiment, the urban waterlogging prediction model refers to a more accurate urban waterlogging prediction model for the target study area obtained by training the initial urban waterlogging prediction model based on historical rainfall data, waterlogging points and corresponding waterlogging depths in the target study area.

[0120] The beneficial effects of the above technical solution are: building an urban waterlogging prediction model through historical rainfall data and typical rainfall processes from urban rain gauges, and dividing the city into parts based on the spatial distribution of rain gauges. Based on the waterlogging situation in each divided area, the urban waterlogging prediction model is optimized separately, and forecasts are made based on the comprehensive results of all optimization results, thereby achieving a more accurate prediction of urban waterlogging.

[0121] Example 6:

[0122] Based on Example 5, based on the urban waterlogging prediction model and in combination with the subspace distribution of each divided grid area, the waterlogging situation of the corresponding subspace is predicted, including:

[0123] Step 41: Obtain the trained urban waterlogging prediction model and perform model correction based on the subspace distribution of each grid area to obtain a corrected urban waterlogging prediction model corresponding to each grid area, thereby obtaining a corrected urban waterlogging prediction set. The corrected urban waterlogging prediction set includes a corrected urban waterlogging prediction model corresponding to each grid area.

[0124] Step 42: Predicting the waterlogging situation of the corresponding divided grid area based on each modified urban waterlogging prediction model to obtain an urban waterlogging prediction set for the target study area;

[0125] The urban waterlogging prediction set includes the spatial number of the subspace corresponding to the divided grid area, the regional location, and the urban waterlogging prediction result of the corresponding subspace.

[0126] In this embodiment, the model correction refers to correcting the urban waterlogging prediction model based on the rainfall data of the rain gauges in each subspace of the divided grid area and in combination with the spatial distribution geographical location of the corresponding subspace.

[0127] In this embodiment, the revised urban waterlogging prediction model refers to the revised urban waterlogging prediction model corresponding to the current divided area after the model is revised. Since the geographical locations of different grid areas are different, the revised urban waterlogging prediction models are not necessarily the same.

[0128] In this embodiment, the modified urban waterlogging prediction model set refers to a set consisting of all modified urban waterlogging prediction models.

[0129] In this embodiment, the modified urban waterlogging prediction set includes a modified urban waterlogging prediction model corresponding to each divided grid area.

[0130] In this embodiment, the urban waterlogging prediction set is formed based on the prediction results of each modified urban waterlogging prediction model.

[0131] In this embodiment, the urban waterlogging prediction set includes the space number of the subspace corresponding to the divided grid area, the area location, and the urban waterlogging prediction result of the corresponding subspace.

[0132] The beneficial effects of the above technical solution are: constructing an urban waterlogging prediction model through rainfall data from urban rain gauges for prediction, and obtaining the spatial distribution of rain gauges through waterlogging points and water depths, thereby dividing the city, and based on the waterlogging situation in each divided area, optimizing the urban waterlogging prediction model separately to achieve more accurate prediction of urban waterlogging.

[0133] Example 7:

[0134] Based on Example 6, the difficulty of real-time scenario forecasting for each divided grid area is determined, and the corresponding forecast results are adjusted to forecast urban waterlogging, including:

[0135] Step 51: Obtain several groups of historical forecast rainfall and historical measured rainfall corresponding to different forecast lengths for all rain gauges included in each divided grid area;

[0136] Step 52: Compare each set of historical forecast rainfall and historical measured rainfall of the currently divided grid area with the preset rainfall level;

[0137] If the historical forecast rainfall and the historical measured rainfall in the same group are in the same rainfall level, the corresponding weather scenario and the historical forecast rainfall and historical measured rainfall will be classified into the first forecast category;

[0138] If the historical forecast rainfall and the historical measured rainfall in the same group are not in the same rainfall level, the historical forecast rainfall, the historical measured rainfall and their corresponding two weather scenarios are classified into the second forecast category;

[0139] Step 53: determining a first forecast number based on the same rainfall station in the first forecast category and determining a second forecast number based on the same rainfall station in the second forecast category;

[0140] Step 54: Based on the first forecast number and the second forecast number, obtain a comprehensive rainfall forecast difficulty coefficient for the current grid area;

[0141] ;

[0142] ;in, is the rainfall forecast difficulty coefficient of the g-th rain gauge in the current grid area; m is the number of forecasters for the g-th rain gauge in the current grid area; n is the number of forecast scenarios in the current grid area; The number of qualified forecasts made by the i-th forecaster at the g-th rainfall station in the current grid area under the j-th forecast scenario; is the number of unqualified forecasts made by the i-th forecaster at the g-th rainfall station in the current grid area under the j-th forecast scenario; is the first forecast number corresponding to the g-th rainfall station; is the second forecast number corresponding to the g-th rainfall station; The first weight set for the overall calculation; A second weight set for the average calculation; is the comprehensive rainfall forecast difficulty coefficient in the current grid area; s represents the number of rain gauges in the current grid area; Indicates the maximum forecast error probability in the current divided area grid; represents the average forecast error probability in the current divided area grid;

[0143] Step 55: Based on the comprehensive rainfall forecast difficulty coefficient, obtain the forecast difficulty of the real-time weather scenario of the current divided grid area;

[0144] Step 56: Combine the forecast difficulty of the real-time weather scenario in the current grid area and the forecast error probability of each rain gauge in the current grid area. , adjust the urban waterlogging prediction result of the current grid area, wherein the adjusted urban waterlogging prediction result is the waterlogging prediction result of the current grid area;

[0145] Step 57: Based on the waterlogging prediction result of the current divided grid area, obtain the urban waterlogging prediction result and the corresponding waterlogging index of the current divided grid area, and perform waterlogging forecast.

[0146] In this embodiment, the prediction length period refers to the length of time that the rain gauge predicts rainfall at different times before rainfall. For example, the prediction length period may be 48 hours, 24 hours, or 12 hours before rainfall.

[0147] In this embodiment, the historical forecast rainfall refers to the rainfall predicted by all rainfall stations included in the current grid area when performing rainfall forecasts in the historical forecast process.

[0148] In this embodiment, the historical measured rainfall refers to the rainfall at all rainfall stations included in the target divided grid area during the historical rainfall process.

[0149] In this embodiment, each historical measured rainfall corresponds to one or more historical forecast rainfalls. When each historical measured rainfall corresponds to multiple historical forecast rainfalls, the prediction lengths of the historical forecast rainfalls must be different.

[0150] In this embodiment, the preset rainfall level refers to filtering the corresponding rainfall level in the rainfall database based on the historical forecast rainfall or the historical measured rainfall data. For example, if the historical measured rainfall is 11 mm and the historical forecast rainfall is 12 mm, the corresponding rainfall levels are both level 2, and the rainfall level 2 includes (10 mm, 15 mm).

[0151] In this embodiment, the corresponding weather scenario refers to the weather scenario before, during and after the current rainfall. For example, before the rainfall, it is sunny and the outdoor temperature is 28 degrees Celsius. The rainfall is thunderstorms and heavy rain. After the rainfall, it is cloudy and the outdoor temperature is 20 degrees Celsius.

[0152] In this embodiment, the first forecast classification refers to classifying the corresponding weather scenarios and the historical forecast rainfall and the historical measured rainfall when the historical forecast rainfall and the historical measured rainfall in the same group are at the same rainfall level; the second forecast classification refers to classifying the historical forecast rainfall, the historical measured rainfall and their corresponding two weather scenarios separately when the historical forecast rainfall and the historical measured rainfall in the same group are not at the same rainfall level.

[0153] In this embodiment, the first forecast number refers to the number of corresponding rainfall forecasts made in the same rainfall station based on the first forecast classification; the second forecast number refers to the number of corresponding rainfall forecasts made in the same rainfall station based on the second forecast classification.

[0154] In this embodiment, the comprehensive rainfall forecast difficulty coefficient refers to the rainfall forecast difficulty of the current grid area determined by calculating the average rainfall forecast difficulty coefficients corresponding to all rain gauges in the current grid area.

[0155] In this embodiment, the real-time weather scenario refers to the weather conditions before and during rainfall corresponding to the real-time rainfall in the currently divided grid area.

[0156] In this embodiment, the forecast difficulty refers to the difficulty of determining the real-time rainfall forecast based on the real-time weather scenario transmitted to the comprehensive rainfall forecast difficulty calculation formula.

[0157] In this embodiment, the forecast error probability refers to the probability of a forecast error at the current rain gauge. If the error probability is 80%, and the actual rainfall is always 2 mm different from the forecast rainfall during each error reporting process, then the forecast rainfall needs to be increased or decreased by 2 mm. Adjustments are made based on whether the difference of 2 mm is an increase or decrease, thereby ensuring the accuracy of the forecast.

[0158] In this embodiment, the urban waterlogging prediction result refers to a prediction result obtained by performing a prediction based on the modified urban waterlogging prediction model and combining the rainfall forecast difficulty of each rain gauge.

[0159] In this embodiment, the waterlogging index refers to the waterlogging severity index corresponding to each waterlogging prediction result in the waterlogging prediction table.

[0160] The beneficial effects of the above technical solution are: building an urban waterlogging prediction model through rainfall data from urban rain gauges, dividing the city according to the spatial distribution of rain gauges, and optimizing the urban waterlogging prediction model based on the waterlogging situation in each divided area, and making forecasts based on the comprehensive results of all optimization results, thereby achieving more accurate predictions of urban waterlogging.

[0161] Example 8:

[0162] Based on Example 7, based on the waterlogging prediction results of the current divided grid area, the urban waterlogging prediction results and the corresponding waterlogging index of the current divided grid area are obtained, and waterlogging forecast is performed, such as Figure 3 Shown, including:

[0163] Step 571: Compare the waterlogging prediction result of the current grid area with a preset waterlogging prediction table to obtain a waterlogging index corresponding to the waterlogging prediction result of the current grid area.

[0164] Step 572: Determine whether the waterlogging index is within a first waterlogging preset range;

[0165] If the waterlogging index is within the first waterlogging preset range, it is determined that the current rainfall will not cause regional waterlogging in the current grid area;

[0166] On the contrary, it is judged that the current rainfall will cause waterlogging in the region, and the waterlogging level and waterlogging possibility of the subspace region are determined based on the corresponding level of the waterlogging index;

[0167] Step 573: Based on the waterlogging levels of all divided grid areas of the target study area and the waterlogging possibility of each divided grid area, determine the comprehensive waterlogging level of the target study area and the waterlogging possibility of different target study subspaces, and perform waterlogging forecast.

[0168] In this embodiment, the urban waterlogging prediction result refers to a prediction result obtained by performing a prediction based on the modified urban waterlogging prediction model and combining the rainfall forecast difficulty of each rain gauge.

[0169] In this embodiment, the preset waterlogging prediction table refers to a preset waterlogging prediction table that includes all waterlogging prediction results and a waterlogging index corresponding to each result.

[0170] In this embodiment, the waterlogging index refers to the waterlogging severity index corresponding to each waterlogging prediction result in the waterlogging prediction table.

[0171] In this embodiment, the first preset range of waterlogging refers to a waterlogging index range in which waterlogging will not occur.

[0172] In this embodiment, regional waterlogging means that if the waterlogging index is higher than a first preset waterlogging range, the grid area corresponding to the current waterlogging index will experience regional waterlogging.

[0173] In this embodiment, the waterlogging level refers to the severity level of waterlogging in the grid area corresponding to the current waterlogging index.

[0174] In this embodiment, the possibility of waterlogging refers to the possibility that waterlogging of the current waterlogging level will occur in the grid area corresponding to the current waterlogging index.

[0175] The beneficial effects of the above technical solution are: the spatial distribution of rain gauges is obtained through waterlogging points and waterlogging depths, thereby dividing the city, and based on the waterlogging situation in each divided area, the urban waterlogging prediction model is optimized separately, and waterlogging forecasts are made based on the waterlogging index corresponding to all optimization results, which can achieve more accurate predictions of urban waterlogging.

[0176] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A method for urban waterlogging forecasting based on network forecasting, characterized in that: include: Step 1: Obtain the distribution of rain gauges in the target study area, and obtain historical rainfall data and typical rainfall processes for each rain gauge; Step 2: Obtain the rainstorm waterlogging points and waterlogging depths measured by each rain gauge in the target study area within the current effective time, obtain the spatial distribution of rainfall, and divide the target study area into grids to obtain the subspace distribution of each divided grid area; Step 3: Establish an urban waterlogging prediction model based on historical rainfall data and typical rainfall processes; Step 4: Based on the urban waterlogging prediction model and in combination with the subspace distribution of each divided grid area, predict the waterlogging situation of the corresponding subspace; Step 5: Determine the difficulty of real-time scenario forecasting for each gridded area and adjust the corresponding forecast results to forecast urban waterlogging. Wherein, step 4 includes: Step 41: Obtain the trained urban waterlogging prediction model and perform model correction based on the subspace distribution of each grid area to obtain a corrected urban waterlogging prediction model corresponding to each grid area, thereby obtaining a corrected urban waterlogging prediction set. The corrected urban waterlogging prediction set includes a corrected urban waterlogging prediction model corresponding to each grid area. Step 42: Predicting the waterlogging situation of the corresponding divided grid area based on each modified urban waterlogging prediction model to obtain an urban waterlogging prediction set for the target study area; The urban waterlogging prediction set includes the spatial number of the subspace corresponding to the divided grid area, the regional location and the urban waterlogging prediction result of the corresponding subspace; Among them, the urban waterlogging prediction model is established based on historical rainfall data and typical rainfall processes, including: Step 31: Based on the latest rainfall data of the target study area and the corresponding waterlogging points and waterlogging depths, an initial urban waterlogging prediction model is obtained; Step 32: Based on the historical rainfall data of the target study area and the corresponding waterlogging points and waterlogging depths, the initial urban waterlogging prediction model is trained to obtain the urban waterlogging prediction model.

2. The urban waterlogging forecasting method based on network forecasting according to claim 1, characterized in that: Obtain the distribution of rain gauges in the target study area, and obtain historical rainfall data and typical rainfall processes for each rain gauge, including: Step 11: Obtain the rain gauge station table number and its distribution location in the target study area, and obtain the historical rainfall data of each rain gauge in the target study area. Based on the historical rainfall data, obtain the typical rainfall process of each rain gauge station; Step 12: Construct a historical rainfall data table for each rain gauge based on the historical rainfall data and the corresponding typical rainfall process.

3. The urban waterlogging forecasting method based on network forecasting according to claim 2, characterized in that: Obtain the rainstorm water points and water depths measured by each rain gauge in the target study area within the current effective time, and obtain the spatial distribution of rainfall, including: Step 21: Obtain all rainstorm water points and water depths measured by each rain gauge in the target study area within the effective measurement time; Step 22: Determine the comprehensive distribution of the rainstorm waterlogging points and the average value of the corresponding waterlogging depths measured by the current rain gauge within the effective measurement time based on the rainstorm waterlogging points and the corresponding waterlogging depths; Step 23: Determine the spatial distribution of rainfall in the target study area based on the comprehensive distribution of corresponding rainstorm waterlogging points and the mean waterlogging depth measured by all rain gauges.

4. The urban waterlogging forecasting method based on network forecasting according to claim 1, characterized in that: Divide the target study area into grids and obtain the subspace distribution of each divided grid area, including: Step 01: Based on the spatial distribution of rainfall in the target study area, the target study area is divided into the first grid using Thiessen polygons; Step 02: Obtain the rainfall time distribution of the target study area through the historical rainfall data in the historical rainfall data table corresponding to each rain gauge, thereby performing a second grid division on the target study area. The second grid division result is adjusted based on the typical rainfall process to obtain the third grid division result. Step 03: Overlap and integrate the first and third meshing results to obtain a comprehensive meshing result. Step 04: Based on the comprehensive grid division results, the subspace distribution of each divided grid area in the target study area is obtained.

Images