A city waterlogging risk forecasting method based on joint distribution of rain intensity and rain type characteristic parameters

Abstract

This invention discloses a method for predicting urban flooding risk based on the joint distribution of rainfall intensity and rainfall pattern characteristic parameters, comprising the following steps: (1) selecting a certain number of historical disaster-causing rainfall events, calculating their rainfall pattern parameters, and statistically analyzing the rainfall intensity parameters; (2) classifying the historical disaster-causing rainfall events into various rainfall patterns; (3) training the urban flooding prediction model according to different rainfall patterns; (4) constructing a joint distribution of rainfall intensity and rainfall pattern characteristic parameters using two rainfall pattern parameters and two rainfall intensity parameters; (5) obtaining the predicted value of the rainfall intensity parameter using meteorological forecast events, and obtaining the conditional probability distribution of the rainfall pattern characteristic parameters by combining the joint distribution of rainfall intensity and rainfall pattern characteristic parameters; (6) calculating the conditional probability of various rainfall patterns in future rainfall events; (7) calculating the risk of various rainfall patterns in future rainfall events; and (8) calculating the expected risk of urban flooding in future rainfall events. This invention can improve the accuracy of predicting urban flooding risk caused by medium- and long-term rainfall events.

Description

Technical Field

[0001] This invention relates to the field of urban flooding risk prediction in smart drainage stormwater drainage systems, specifically to a method for predicting urban flooding risk based on the joint distribution of rainfall intensity and rainfall pattern characteristic parameters. Background Technology

[0002] In urban flooding risk prediction, rainfall forecast data is the primary driving variable. For rainfall events in the short term, nowcasting technology can accurately obtain the future rainfall duration distribution, which can then be used to predict urban flooding risk with relatively high accuracy. However, for rainfall events in the longer term, only medium- and long-term weather forecasting technology can be used to predict and analyze future weather patterns, the intensity and extent of large-scale weather processes, and obtain predicted values ​​for rainfall intensity parameters such as total rainfall and duration. Relying solely on rainfall intensity parameters results in low accuracy for urban flooding risk prediction. Furthermore, rainfall pattern parameters, reflecting the unevenness of rainfall, are a crucial factor affecting the accuracy of urban flooding risk prediction. While rainfall pattern and rainfall intensity parameters are not entirely independent, current technologies for predicting urban flooding risk using the joint probability distribution of multiple rainfall characteristic parameters (rainfall intensity and rainfall pattern) are still relatively lacking.

[0003] The patent specification with publication number CN114067019A discloses a rapid pre-production method for urban flooding risk maps coupled with deep learning and numerical simulation. The steps are as follows: 1. Construct an urban flooding numerical simulation model using PCSWMM, with the design rainfall amounts for 2, 5, 10, 50, and 100 years as the input to the model, and the maximum inundation depth and maximum inundation flow velocity as the output to obtain a rainfall dataset; 2. Preprocess the training and testing data to form a training and testing database; 3. Construct a deep convolutional neural network model: first, train the model using the training data as input and output to form a rapid pre-production model for urban flooding risk maps; then, use the test data as input and output to validate the model using the trained model; 4. Based on actual rainfall as input data, use the model constructed in step 3 to predict water accumulation, obtain the maximum inundation depth and maximum inundation flow velocity after rainfall, and calculate and compare the water accumulation prediction results with the actual water accumulation.

[0004] The patent specification with publication number CN111680886A discloses a method for predicting urban flooding risk, including the following steps: Step 1: Dividing the area to be planned into several runoff zones; Step 2: Simulating the surface runoff of each runoff zone under different rainfall conditions in the area to be planned using the SCS-CN model; Step 3: Simulating the surface runoff of each runoff zone under the design storm return period of the stormwater drainage system in the area to be planned using the SCS-CN model, and using it as the drainage capacity of the corresponding runoff zone; Step 4: Calculating the difference between the surface runoff of each runoff zone under different rainfall conditions in the area to be planned and the drainage capacity of each runoff zone under the design storm return period of the stormwater drainage system in the area to be planned; Step 5: Predicting the urban flooding risk of each runoff zone based on the difference, providing a scientific basis for reducing urban flooding risk and updating construction plans.

[0005] Neither of the above two schemes provides a reasonable analysis of the rainfall pattern and the distribution of rainfall pattern parameters, making it impossible to accurately predict the risk of urban flooding caused by medium- and long-term rainfall events. Therefore, by conducting a reasonable analysis of the rainfall pattern and the distribution of rainfall pattern parameters, a more scientific analysis of the risk of urban flooding caused by future rainfall events can be conducted. This is of great significance for guiding urban flood control scheduling, accelerating the flood control response of relevant departments, and facilitating the lives and travel of urban residents. Summary of the Invention

[0006] The purpose of this invention is to provide a method for predicting urban flooding risk based on the joint distribution of rainfall intensity and rainfall pattern characteristic parameters. Using this method, given the predicted values ​​of rainfall intensity parameters for future rainfall events, the distribution of rainfall pattern parameters reflecting rainfall unevenness can be predicted more accurately, thereby improving the accuracy of predicting urban flooding risk caused by medium- and long-term rainfall events.

[0007] A method for predicting urban flooding risk based on the joint distribution of rainfall intensity and rainfall pattern characteristic parameters includes the following steps:

[0008] (1) Select a certain number of historical disaster-causing rainfall events, calculate their rainfall pattern parameters, including the rainfall peak location coefficient r and the Herfindahl-Hirschman index HHI, and statistically analyze the rainfall intensity parameters, including the total rainfall and rainfall duration;

[0009] (2) Classify the historical disaster-causing rainfall events in step (1) into multiple rainfall types according to the calculated HHI and r;

[0010] (3) According to the different rain types in step (2), the corresponding rain-induced waterlogging events are used to train the waterlogging prediction model;

[0011] (4) Use the Vine Copula function to construct the joint distribution of rainfall intensity-rain pattern characteristic parameters from the two rainfall pattern parameters and two rainfall intensity parameters in step (1);

[0012] (5) When predicting the risk of urban flooding in future rainfall events, the predicted value of rainfall intensity parameter is obtained by using meteorological forecast events, and the conditional probability distribution of rainfall pattern characteristic parameter is obtained by combining the joint distribution of rainfall intensity-rain pattern characteristic parameter in step (4).

[0013] (6) Using the conditional probability distribution of the rainfall pattern characteristic parameters and the predicted value of the rainfall intensity parameters obtained in step (5), the conditional probability of various rainfall patterns in future rainfall events is calculated.

[0014] (7) Using the predicted values ​​of the urban flooding prediction models and rainfall intensity parameters corresponding to different rainfall types obtained in step (3), the risks of various rainfall types in future rainfall events are calculated.

[0015] (8) Using the conditional probabilities of various rainfall patterns obtained in step (6) and the risks of various rainfall patterns obtained in step (7), the expected risk of urban flooding in future rainfall events is obtained.

[0016] This scheme analyzes the conditional probability of various rainfall patterns occurring in future rainfall events by combining the distribution of multiple rainfall characteristic parameters, thereby more accurately predicting the risk of urban flooding caused by these events.

[0017] As a preferred method, in step (2), historical disaster-causing rainfall events are divided into nine types based on the calculated HHI and r. Based on HHI, rainfall events are divided into three types: oligopoly, low oligopoly, and competition. Based on r, rainfall events are divided into three types: before, during, and after the rain peak position. These are combined in pairs to form nine types of rainfall events.

[0018] Further optimization shows that the oligopolistic rain peak type corresponds to HHI∈(500,10000], the low oligopolistic rain peak type corresponds to HHI∈(200,500], and the competitive rain peak type corresponds to HHI∈(0,200).

[0019] Further optimization shows that before the rain peak position corresponds to r HHI∈[0,1 / 3], during the rain peak position corresponds to r HHI∈(1 / 3,2 / 3], and after the rain peak position corresponds to r HHI∈(2 / 3,1).

[0020] Preferably, in step (3), the flooding prediction model is trained using the corresponding rainfall-induced flooding events according to the different rainfall types. The expression is as follows:

[0021]

[0022] In the formula, This represents the most unfavorable working condition predicted by the urban flooding prediction model corresponding to the j-th rainfall type during the i-th disastrous rainfall event, i.e., the predicted maximum flood depth at all points globally. This represents the global liquid level before the i-th disastrous rainfall event. This represents the total rainfall of the i-th disastrous rainfall event. This represents the duration of the i-th disastrous rainfall event. This indicates other urban flood-driving factors.

[0023] Further optimization revealed other urban flood drivers, including the operational status of river hydraulic structures and the operational status of rainwater storage facilities.

[0024] Preferably, in step (4), the Vine Copula function is used to construct the joint distribution of rainfall intensity-rainfall pattern characteristic parameters, the expression of which is:

[0025] F(x1,x2,x3,x4)=C(u1,u2,u3,u4)

[0026] In the formula, x1 and x2 represent HHI and r in the rainfall pattern parameters, respectively, x3 and x4 represent the total rainfall and rainfall duration in the rainfall intensity parameters, respectively, and u1, u2, u3, u4 represent their marginal distribution functions.

[0027] Preferably, the expression for the expected risk of urban flooding in future rainfall events in step (8) is:

[0028]

[0029] In the formula, Indicates the response to rainfall event Rf n The worst-case prediction result obtained using the urban flooding prediction model for the i-th rainfall type. Rf represents a rainfall event n The probability of the i-th rain type occurring.

[0030] The beneficial effects of this invention are:

[0031] This invention focuses on the impact of uneven rainfall on urban flooding risk, and improves the accuracy of predicting urban flooding risk caused by medium- and long-term rainfall events. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of a pipe-river drainage system in a certain urban area.

[0033] Figure 2 This is a ground node risk assessment diagram of an urban flooding risk forecasting method based on the joint distribution of rainfall intensity and rainfall pattern characteristic parameters used in the embodiment.

[0034] Figure 3 The image shows the risk effect of rainwater wells in an example of an urban flooding risk forecasting method based on the joint distribution of rainfall intensity and rainfall pattern characteristic parameters.

[0035] Figure 4This is a flowchart of the method of the present invention. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] like Figure 1 As shown, the area of ​​a certain city is approximately 3.22 × 10⁻⁶. 6 m 2 The city's stormwater system has low construction standards, with most pipelines built to standards less than a one-year return period. In addition, some pipe sections are damaged, deformed, or blocked, resulting in low actual water delivery capacity. Furthermore, the stormwater runoff coefficient in this area is as high as 0.76, with a large flow rate, leading to a significant risk of urban flooding in the area.

[0038] This embodiment presents an urban flooding risk forecasting method based on the joint distribution of rainfall intensity and rainfall pattern characteristic parameters, such as... Figure 4 As shown, it includes the following steps:

[0039] (1) From 967 independent rainfall events in the city’s 30-year rainfall data, 162 disaster-causing rainfall events were selected and their rainfall pattern parameters were calculated, including the rainfall peak location coefficient r and the Herfindahl-Hirschman index HHI. Rain intensity parameters, including total rainfall and rainfall duration, were also statistically analyzed.

[0040] (2) Historical disaster-causing rainfall events are divided into 9 types of rainfall patterns based on the calculated HHI and r. Specifically, rainfall events are divided into three types based on HHI: oligopoly type (HHI∈(500,10000]), low oligopoly type (HHI∈(200,500]), and competitive type (HHI∈(0,200]). Rainfall events are divided into three types based on r: before the rain peak position (r HHI∈[0,1 / 3]), middle (r HHI∈(1 / 3,2 / 3]), and after the rain peak position (r HHI∈(2 / 3,1]). The two types are combined into 9 types.

[0041] (3) Based on the different rainfall types, the corresponding rainfall-induced waterlogging events are used to train the waterlogging prediction model. The model expression is as follows:

[0042]

[0043] In the formula, This represents the most unfavorable working condition predicted by the urban flooding prediction model corresponding to the j-th rainfall type during the i-th disastrous rainfall event, i.e., the predicted maximum flood depth at all points globally. This represents the global liquid level before the i-th disastrous rainfall event. This represents the total rainfall of the i-th disastrous rainfall event. This represents the duration of the i-th disastrous rainfall event. This indicates other urban flood-driving factors, including the operational status of river hydraulic structures and rainwater storage facilities.

[0044] (4) The two rainfall pattern parameters and two rainfall intensity parameters in step (1) are used to construct the joint distribution of rainfall intensity-rainfall pattern characteristic parameters using the Vine Copula function. The expression is as follows:

[0045] F(x1,x2,x3,x4)=C(u1,u2,u3,u4)

[0046] In the formula, x1 and x2 represent HHI and r in the rainfall pattern parameters, respectively, x3 and x4 represent the total rainfall and rainfall duration in the rainfall intensity parameters, respectively, and u1, u2, u3, u4 represent their marginal distribution functions.

[0047] (5) Using the Vine Copula function, construct the joint distribution of rainfall intensity-rainfall pattern characteristic parameters from the two rainfall pattern parameters and the two rainfall intensity parameters. The expression is as follows:

[0048] F(x1,x2,x3,x4)=C(u1,u2,u3,u4)

[0049] In the formula, x1 and x2 represent HHI and r in the rainfall pattern parameters, respectively, x3 and x4 represent the total rainfall and rainfall duration in the rainfall intensity parameters, respectively, and u1, u2, u3, u4 represent their marginal distribution functions.

[0050] (6) Using the conditional probability distribution of the rainfall pattern characteristic parameters and the predicted value of the rainfall intensity parameter, the conditional probability of various rainfall patterns in future rainfall events is calculated.

[0051] (7) Using the waterlogging prediction models and rainfall intensity parameters corresponding to different rainfall types, the waterlogging risk of various rainfall types in future rainfall events can be calculated.

[0052] (8) Using the conditional probabilities of various rainfall types and the urban flooding risk of various rainfall types, the expected urban flooding risk of future rainfall events is obtained, expressed as:

[0053]

[0054] In the formula, Indicates the response to rainfall event Rf n The worst-case prediction result obtained using the urban flooding prediction model for the i-th rainfall type. Rf represents a rainfall event nThe probability of the i-th rain type occurring.

[0055] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting urban flooding risk based on the joint distribution of rainfall intensity and rainfall pattern characteristic parameters, characterized in that, Includes the following steps: (1) Select a certain number of historical disaster-causing rainfall events, calculate their rainfall pattern parameters, including the rainfall peak location coefficient r and the Herfindahl-Hirschman index HHI, and statistically analyze the rainfall intensity parameters, including the total rainfall and rainfall duration; (2) The historical disaster-causing rainfall events in step (1) are divided into multiple rainfall types according to the calculated HHI and r. The historical disaster-causing rainfall events are divided into nine types according to the calculated HHI and r. The rainfall events are divided into three types based on HHI: oligopoly, low oligopoly, and competition. The rainfall events are divided into three types based on r: before, middle, and after the rain peak position. The nine types are combined in pairs. The oligopoly rain peak corresponds to HHI∈(500,10000], the low oligopoly rain peak corresponds to HHI∈(200,500], and the competition rain peak corresponds to HHI∈(0,200]. Before the rain peak position, r HHI ∈ [0, 1 / 3]; during the rain peak position, r HHI ∈ (1 / 3, 2 / 3]; after the rain peak position, r HHI ∈ (2 / 3, 1]. (3) According to the different rain types in step (2), the corresponding rain-induced waterlogging events are used to train the waterlogging prediction model; Based on the different rainfall types, the corresponding rainfall-induced flooding events are used to train the flooding prediction model, and its expression is as follows: In the formula, This represents the most unfavorable working condition predicted by the urban flooding prediction model corresponding to the j-th rainfall type in the i-th disastrous rainfall event. This represents the global liquid level before the i-th disastrous rainfall event. This represents the total rainfall of the i-th disastrous rainfall event. This represents the duration of the i-th disastrous rainfall event. Indicates other urban flood driving factors; (4) Using the Vine Copula function, construct the joint distribution of rainfall intensity-rainfall pattern characteristic parameters from the two rainfall pattern parameters and two rainfall intensity parameters in step (1). The expression is as follows: In the formula, H, H1, and r represent the rain pattern parameters respectively. , These represent the total rainfall and rainfall duration in the rainfall intensity parameters, respectively. Let each represent their marginal distribution function; (5) When predicting the risk of waterlogging in future rainfall events, use the weather forecast event to obtain the predicted value of the rainfall intensity parameter, and combine it with the joint distribution of rainfall intensity-rain type characteristic parameters in step (4) to obtain the conditional probability distribution of the rainfall type characteristic parameter; (6) Use the conditional probability distribution of the rainfall type characteristic parameter and the predicted value of the rainfall intensity parameter obtained in step (5) to calculate the conditional probability of various rainfall types in future rainfall events; (7) Use the waterlogging prediction model corresponding to different rainfall types and the predicted value of the rainfall intensity parameter obtained in step (3) to calculate the risk of various rainfall types in future rainfall events. (8) Using the conditional probabilities of various rain types obtained in step (6) and the risks of various rain types obtained in step (7), the expected risk of urban flooding in future rainfall events is obtained.

2. The urban flooding risk forecasting method according to claim 1, characterized in that, Other urban flood drivers include the operational status of river hydraulic structures and stormwater storage facilities.

3. The urban flooding risk forecasting method according to claim 1, characterized in that, The expression for the expected risk of urban flooding in future rainfall events in step (8) is: In the formula, Indicates the response to rainfall events The worst-case prediction result obtained using the urban flooding prediction model for the j-th rainfall type. Indicates rainfall event The probability of the j-th rain type occurring.

Images