A method for automatic flood identification based on flood volume characteristics and function properties

Abstract

The present invention discloses a flood automatic identification method based on flood volume characteristics and function properties, including obtaining the flood peak flow Q based on the first and second order derivatives of long series runoff data Q(). m ; According to the peak flow Q m The relationship between the set flood peak threshold q1 and the flood peak distance threshold m is used to obtain the flood peak array Q m (); Based on the peak array Q m () Set the threshold n of the maximum duration of flood and determine the calculation period Q n (); Based on the calculation period Q n The second-order derivative of () determines the start and end locations of flood events, and thus the flood time period. Based on the flood time period, the flood peak location is determined. Flood volume characteristics for 3, 5, 7, 15, and 25 hours forward and backward from the flood peak are introduced, and flood events are screened based on the judgment criteria. The advantages are: the introduction of multi-hour flood volume characteristics including peak values, combined with the properties of the function, accurately identifies flood events within a continuous flood process, improving the accuracy and efficiency of flood identification.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrological calculation flood screening, and in particular to a method for automatically identifying flood events based on flood volume characteristics and function properties. Background Art

[0002] Common flood screening methods primarily use Excel and other tools to create rainfall-flood maps. These maps, taking into account rainfall over the same period, use baseflow to identify flood rise and fall points, relying primarily on manual experience to identify flood events. While this method offers some accuracy, it is labor-intensive and labor-intensive. Furthermore, flood screening is subject to subjective judgment, resulting in inconsistent standards for selected flood events, which impacts subsequent hydrological calculations. To address these issues, methods have been developed to classify flood events using the properties of mathematical functions. However, these methods only consider the functional characteristics, not the actual flood volume. This makes them effective only for runoff processes with good data quality and a more pronounced flood profile. However, in actual observed flow events, sudden increases in flow are common, while the flow near the sudden change point is relatively flat. Existing flood identification techniques often misidentify these flow events as flood events, resulting in low flood identification efficiency and high error rates. Summary of the Invention

[0003] The purpose of the present invention is to provide a method for automatically identifying flood events based on flood volume characteristics and function properties, thereby solving the above-mentioned problems existing in the prior art.

[0004] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0005] A method for automatically identifying flood events based on flood volume characteristics and function properties includes the following steps:

[0006] S1. Obtain the peak flow Q based on the first and second derivatives of the long series runoff data Q() m ; According to the peak flow Q m The relationship between the set flood peak threshold q1 and the flood peak distance threshold m is used to obtain the flood peak array Q m ();

[0007] S2, based on the peak array Q m () Set the threshold n of the maximum duration of the flood, and then determine the calculation period Q n (); Based on the calculation period Q n The second-order derivative of () determines the starting and ending positions of the flood, and then determines the flood time period;

[0008] S3. Determine the flood peak location based on the flood time period, introduce flood volume characteristics 3h, 5h, 7h, 15h, and 25h forward and backward from the flood peak location, and implement flood screening based on the judgment conditions.

[0009] Preferably, step S1 specifically includes the following contents:

[0010] S11. Take the long series of runoff data as the research object, denoted as Q();

[0011] S12. Calculate the first-order derivative and the second-order derivative of Q(), denoted as Q′() and Q″() respectively;

[0012] S13. When Q′(i)=0 and Q″(i)<0, the flow at the corresponding position i is recorded as the peak flow Q m ;

[0013] S14, setting the flood peak threshold q1 and the flood peak distance threshold m;

[0014] S15, when Q m >q1 and two adjacent Q m When the distance between them is greater than m, Q m Stored in an array to form the peak array Q m ().

[0015] Preferably, step S2 specifically includes the following contents:

[0016] S21. Set the threshold value n of the maximum duration of the flood, and select the time within the range of the threshold n around the flood peak as the calculation period, recorded as Q n ();

[0017] S22. Calculate Q n The second-order derivative of (), denoted as Q n ″();

[0018] S23, when Q n ″(j)=0 and Q n ″(j-1)<0 and Q″ n When (j+1)>0, the time j is determined to be the starting point of the flood, that is, the location where the flood begins;

[0019] S24, when Q n ″(k)=0 and Q n ″(k-1)>0 and Q″ n When (k+1)<0, the time k is determined to be the ebb and flow point of the flood, that is, the location where the flood ends;

[0020] S25, then the time period from j to k is the flood period.

[0021] Preferably, step S3 specifically includes the following contents:

[0022] S31. Determine the flood peak location of the flood based on the flood time period; record the corresponding flood volume as a3 3 hours forward from the flood peak location, and record the corresponding flood volume as b3 3 hours backward from the flood peak location;

[0023] 5 hours forward from the flood peak, record the corresponding flood volume as a5, and 5 hours backward from the flood peak, record the corresponding flood volume as b5;

[0024] 7 hours forward from the flood peak, the corresponding flood volume is recorded as a7, and 7 hours backward from the flood peak, the corresponding flood volume is recorded as b7;

[0025] 15 hours forward from the flood peak, record the corresponding flood volume as a 15 , 15 hours after the flood peak, record the corresponding flood volume value as b 15 ;

[0026] 25 hours forward from the flood peak, record the corresponding flood volume as a 25 , 25 hours after the flood peak, the corresponding flood volume is recorded as b 25 ;

[0027] S32. If the corresponding flood volume values a and b satisfy the following inequality, the flood event meets the flood volume characteristics of the flood event and is retained; otherwise, the flood event does not meet the flood volume characteristics of the flood event and is eliminated; thereby achieving flood screening, the relevant inequalities are as follows:

[0028]

[0029]

[0030]

[0031]

[0032]

[0033] Preferably, the acquired long series runoff data needs to be preprocessed before step S1, and the specific process is as follows:

[0034] Interpolation processing: For the missing data in the long series of flow data, linear interpolation method is used to process it to ensure the continuity of the long series of flow data;

[0035] Replacement processing: For outliers and mutation points in long series of traffic data, linear interpolation method is used to replace them to improve data rationality.

[0036] The beneficial effects of the present invention are as follows: 1. The method of the present invention realizes the function of automatically dividing flood events. Compared with the method of dividing the rising and falling points of floods by base flow and mainly relying on manual experience to extract flood events, the work efficiency is greatly improved. The screening method is universal and easy to operate, and can be widely used in flood forecasting systems. 2. The method of the present invention has a good recognition effect for runoff data of poor quality, especially for the situation where the flow rate suddenly increases at a certain moment in the measured flow process, but the flow process near the mutation point is relatively smooth, which effectively improves the accuracy of flood screening. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a flow chart of an identification method in an embodiment of the present invention;

[0038] Figure 2 2. Schematic diagram of the flow rate of Yunlong Hydrological Station in 2018 in an embodiment of the present invention;

[0039] Figure 3 This is a schematic diagram of a flood event corresponding to flood peak No. 1 in an embodiment of the present invention;

[0040] Figure 4 This is a schematic diagram of a flood event corresponding to flood peak No. 3 in an embodiment of the present invention;

[0041] Figure 5 This is a schematic diagram of a flood event corresponding to flood peak No. 4 in an embodiment of the present invention;

[0042] Figure 6 This is a schematic diagram of a flood event corresponding to flood peak No. 6 in an embodiment of the present invention;

[0043] Figure 7 This is a schematic diagram of the flood corresponding to flood peak No. 7 in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0045] Example 1

[0046] Such as Figure 1As shown, in this embodiment, by judging the flood volume characteristics of 3h, 5h, 7h, 15h, and 25h containing the peak, and combining them with the function properties, a method for automatically identifying flood events based on flood volume characteristics and function properties is proposed. This method overcomes the problem that existing automatic flood identification methods cannot accurately identify runoff data due to poor quality, and improves the recognition accuracy of automatic flood identification. The method mainly includes four parts:

[0047] 1. Data Preprocessing

[0048] The flow series Q() that needs to be divided into flood events is preprocessed to delete null values, abnormal points, and mutation points. Specifically,

[0049] 1. Interpolation processing: For the missing data in the long series of flow data, linear interpolation method is used to process it to ensure the continuity of the long series of flow data;

[0050] 2. Replacement processing: For abnormal values (such as negative values) and mutation points in long series of traffic data, linear interpolation method is used to replace them to improve data rationality.

[0051] 2. Determine the peak array

[0052] This part corresponds to step S1: Based on the first and second derivatives of the long series runoff data Q(), the peak flow Q is obtained. m ; According to the peak flow Q m The relationship between the set flood peak threshold q1 and the flood peak distance threshold m is used to obtain the flood peak array Q m (). The specific process is:

[0053] 1. Take the long series of runoff data as the research object, denoted as Q();

[0054] 2. Calculate the first and second derivatives of Q(), denoted as Q′() and Q″() respectively;

[0055] 3. When Q′(i)=0 and Q″(i)<0, the flow at the corresponding position i is recorded as the peak flow Q m ;

[0056] 4. Set the flood peak threshold q1 and the flood peak distance threshold m;

[0057] 5. When Q m >q1 and two adjacent Q m When the distance between them is greater than m, Q m Stored in an array to form the peak array Q m ().

[0058] 3. Determine the flood process

[0059] This part corresponds to S2: Based on the peak array Q m () Set the threshold n of the maximum duration of the flood, and then determine the calculation period Q n (); Based on the calculation period Q n The second-order derivative of () determines the starting and ending positions of the flood, and then determines the flood time period. The specific process is:

[0060] 1. Set the threshold value n of the maximum duration of the flood, and select the time within the range of the threshold n around the flood peak as the calculation period, recorded as Q n ();

[0061] 2. Calculate Q n The second-order derivative of (), denoted as Q n ″();

[0062] 3. When Q n ″(j)=0 and Q n ″(j-1)<0 and Q″ n When (j+1)>0, the time j is determined to be the starting point of the flood, that is, the location where the flood begins;

[0063] 4. When Q n ″(k)=0 and Q n ″(k-1)>0 and Q″ n When (k+1)<0, the time k is determined to be the ebb and flow point of the flood, that is, the location where the flood ends;

[0064] 5. From j to k is the flood time period, that is, the flood process.

[0065] 4. Introducing flood volume characteristics to achieve flood screening

[0066] This part corresponds to S3: Based on the flood time period, the flood peak location is determined, the flood volume characteristics of 3 hours, 5 hours, 7 hours, 15 hours, and 25 hours forward and backward from the flood peak location are introduced, and the flood events are screened according to the judgment conditions. The specific process is as follows:

[0067] 1. Determine the flood peak location based on the flood time period; 3 hours forward from the flood peak location, record the corresponding flood volume as a3; 3 hours backward from the flood peak location, record the corresponding flood volume as b3;

[0068] 5 hours forward from the flood peak, record the corresponding flood volume as a5, and 5 hours backward from the flood peak, record the corresponding flood volume as b5;

[0069] 7 hours forward from the flood peak, the corresponding flood volume is recorded as a7, and 7 hours backward from the flood peak, the corresponding flood volume is recorded as b7;

[0070] 15 hours forward from the flood peak, record the corresponding flood volume as a 15 , 15 hours after the flood peak, record the corresponding flood volume value as b 15 ;

[0071] 25 hours forward from the flood peak, record the corresponding flood volume as a 25 , 25 hours after the flood peak, the corresponding flood volume is recorded as b 25 ;

[0072] 2. If the corresponding flood volume values a and b satisfy the following inequality, the flood event meets the flood volume characteristics of the flood event and is retained; otherwise, the flood event does not meet the flood volume characteristics of the flood event and is eliminated; thus, the screening of flood events is achieved. The relevant inequalities are as follows:

[0073]

[0074]

[0075]

[0076]

[0077]

[0078] Example 2

[0079] In this embodiment, the hydrological flow process of the Yunlong hydrological station in the Lancang River Basin from 2018-1-1 0:00 to 2018-12-31 23:00 is selected, such as Figure 2 As shown. From this hydrological flow process, the peak flow greater than 150m 3 / s flood, the specific steps are as follows:

[0080] 1. Data Preprocessing

[0081] Obtain hydrological data and preprocess the data. Collect the flow data of Yunlong Hydrological Station in 2018. First, use linear interpolation to replace the abnormal values (such as negative values) and mutation points in the long series of data. The flow series is represented by Q(), the period length is 1 hour, and the number of elements N = 8760, as shown in the following example: Figure 1 shown.

[0082] 2. Determine the peak array

[0083] By calculating the first and second derivatives of the flow series Q(), the peak flow Q is obtained. m Set the flood peak threshold q1 = 150m 3 / s and the distance threshold of the flood peak m = 120h, extract the flood peak array Qm ().

[0084] First, calculate the first and second order derivatives of Q() to determine the peak flow. When Q′(i)=0 and Q″(i)<0, the flow at the corresponding position i is recorded as the peak flow Q m , set the flood peak threshold to 150m 3 / s, and the flood peak is obtained, as shown in Table 1.

[0085] Table 1 Location and flow of flood peaks

[0086] serial number Flood Peak No. 1 Flood Peak No. 2 Flood Peak No. 3 Flood Peak No. 4 Flood Peak No. 5 Flood Peak No. 6 Flood Peak No. 7 Location of flood peak 4627 4808 4899 5004 5098 5160 5427 Peak flow 164.8 154 171.2 177.7 151 169.6 223.6

[0087] To ensure that each flood peak is in a separate flood process, calculate the flood peak array Q m The distance between each flood peak in ( ) is determined to determine whether it is greater than m. If it is, the flood peak meets the requirements. If it is less than m, the adjacent flood peaks with smaller peak flows are eliminated. In this case, Td is set to 120h. It is found that flood peaks 5 and 6 do not meet the requirements, and flood peak 5 is smaller than flood peak 6, so flood peak 5 is eliminated. This forms a new flood peak sequence, as shown in Table 2.

[0088] Table 2 Location and flow of flood peaks

[0089] serial number Flood Peak No. 1 Flood Peak No. 2 Flood Peak No. 3 Flood Peak No. 4 Flood Peak No. 6 Flood Peak No. 7 Location of flood peak 4627 4808 4899 5004 5160 5427 Peak flow 164.8 154 171.2 177.7 169.6 223.6

[0090] three, Determine the flood process

[0091] Find the location of the flood start and end during a flood process corresponding to the flood peak. First, set the threshold value of the maximum duration of the flood n = 72, and calculate the calculation period Q n The second-order derivative of () is used to filter out the time positions of the flood start and end according to the discrimination conditions. The calculation results are shown in Table 3. The flood process of the flood is shown in Table 4.

[0092] Table 3 Basic information of flood events

[0093] serial number Flood Peak No. 1 Flood Peak No. 2 Flood Peak No. 3 Flood Peak No. 4 Flood Peak No. 6 Flood Peak No. 7 Location of flood peak 4627 4808 4899 5004 5160 5427 Peak flow 164.8 154 171.2 177.7 169.6 223.6 Flood start time 4610 4782 48881 4992 5144 5400 Flood end time 4666 4823 4918 5018 5184 5460

[0094] Table 4 Flow process of flood events

[0095]

[0096]

[0097]

[0098] 4. Introducing flood volume characteristics to achieve flood screening

[0099] Screen flood events based on flood volume characteristics. The floods screened in the third step are divided according to the function properties, without considering the actual flood process. This may lead to the emergence of pseudo-floods that do not conform to the actual flood characteristics. To solve this problem, the flood volume characteristics will be calculated to further screen flood events that conform to the actual flood process and improve the accuracy of automatic flood identification. The details are as follows:

[0100] 1. Calculate the characteristic flood volume of the six floods selected in the third step, including the flood volume at 3h, 5h, 7h, 15h, and 25h of the peak. The flood volume before the peak is recorded as a i The flood volume after the flood peak is recorded as b i , i=3,5,7,15,25;

[0101] 2. Determine whether the relevant formula is satisfied. If so, the flood meets the flood volume characteristics of the flood event and is a flood event. Otherwise, the flood event is eliminated. The relevant formula is as follows:

[0102]

[0103]

[0104]

[0105]

[0106]

[0107] The calculation results are shown in Table 5.

[0108] Table 5 Screening of floods that meet flood characteristics

[0109]

[0110]

[0111] Finally, 5 floods that meet the requirements were screened out, namely the floods corresponding to the No. 1, No. 3, No. 4, No. 6, and No. 7 flood peaks. The corresponding flood processes are as follows: Figures 3 to 7 shown.

[0112] By adopting the above technical solution disclosed in the present invention, the following beneficial effects are obtained:

[0113] The present invention provides a method for automatically identifying flood events based on flood volume characteristics and functional properties. This method automatically classifies flood events. Compared to methods that use baseflow to identify flood rise and fall points and rely primarily on manual experience to extract flood events, this method significantly improves work efficiency. The screening method is versatile and easy to operate, making it widely applicable to flood forecasting systems. The method has excellent recognition results for poor-quality runoff data, especially for situations where the flow rate suddenly increases at a certain moment during measured flow, but the flow rate near the sudden change point is relatively flat, effectively improving the accuracy of flood event screening.

[0114] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for automatically identifying flood events based on flood volume characteristics and function properties, characterized by: The following steps are included: S1. Obtain the peak flow Q based on the first and second derivatives of the long series runoff data Q() m ; According to the peak flow Q m The relationship between the set flood peak threshold q1 and the flood peak distance threshold m is used to obtain the flood peak array Q m (); S2, based on the peak array Q m () Set the threshold n of the maximum duration of the flood, and then determine the calculation period Q n (); Based on the calculation period Q n The second-order derivative of () determines the starting and ending positions of the flood, and then determines the flood time period; S3. Determine the flood peak location based on the flood time period, introduce flood volume characteristics 3 hours, 5 hours, 7 hours, 15 hours, and 25 hours before and after the flood peak location, and implement flood screening based on the judgment conditions; Step S3 specifically includes the following contents: S31. Determine the flood peak location of the flood based on the flood time period; record the corresponding flood volume as a3 3 hours forward from the flood peak location, and record the corresponding flood volume as b3 3 hours backward from the flood peak location; 5 hours forward from the flood peak, record the corresponding flood volume as a5, and 5 hours backward from the flood peak, record the corresponding flood volume as b5; 7 hours forward from the flood peak, the corresponding flood volume is recorded as a7, and 7 hours backward from the flood peak, the corresponding flood volume is recorded as b7; 15 hours forward from the flood peak, record the corresponding flood volume as a 15 , 15 hours after the flood peak, record the corresponding flood volume value as b 15 ; 25 hours forward from the flood peak, record the corresponding flood volume as a 25 , 25 hours after the flood peak, the corresponding flood volume is recorded as b 25 ; S32. If the corresponding flood volume values a and b satisfy the following inequality, the flood event meets the flood volume characteristics of the flood event and is retained; otherwise, the flood event does not meet the flood volume characteristics of the flood event and is eliminated; thereby achieving flood screening, the relevant inequalities are as follows:

2. The method for automatically identifying flood events based on flood volume characteristics and function properties according to claim 1 is characterized by: Step S1 specifically includes the following contents: S11. Take the long series of runoff data as the research object, denoted as Q(); S12. Calculate the first-order derivative and the second-order derivative of Q(), denoted as Q′() and Q″() respectively; S13, when Q ′ When (i) = 0 and Q″(i) < 0, the flow at the corresponding position i is recorded as the peak flow Q m ; S14, setting the flood peak threshold q1 and the flood peak distance threshold m; S15, when Q m >q1 and two adjacent Q m When the distance between them is greater than m, Q m Stored in an array to form the peak array Q m ().

3. The method for automatically identifying flood events based on flood volume characteristics and function properties according to claim 1 is characterized by: Step S2 specifically includes the following contents: S21. Set the threshold value n of the maximum duration of the flood, and select the time within the range of the threshold n around the flood peak as the calculation period, recorded as Q n (); S22. Calculate Q n The second-order derivative of (), denoted as Q n ″(); S23, when Q n ″(j)=0 and Q n ″(j-1)<0 and Q ′ n ′ When (j+1)>0, the time j is determined to be the starting point of the flood, that is, the location where the flood begins; S24, when Q n ″(k)=0 and Q n ″(k-1)>0 and Q ′ n ′ When (k+1)<0, the time k is determined to be the ebb and flow point of the flood, that is, the location where the flood ends; S25, then the time period from j to k is the flood period.

4. The method for automatically identifying flood events based on flood volume characteristics and function properties according to any one of claims 1 to 3, characterized in that: Before step S1, the long series of runoff data obtained needs to be preprocessed. The specific process is as follows: Interpolation processing: For the missing data in the long series of flow data, linear interpolation method is used to process it to ensure the continuity of the long series of flow data; Replacement processing: For outliers and mutation points in long series of traffic data, linear interpolation method is used to replace them to improve data rationality.

Images