CN113626500A - Flood stage determination method based on multi-feature indexes - Google Patents

Flood stage determination method based on multi-feature indexes Download PDF

Info

Publication number
CN113626500A
CN113626500A CN202110912605.8A CN202110912605A CN113626500A CN 113626500 A CN113626500 A CN 113626500A CN 202110912605 A CN202110912605 A CN 202110912605A CN 113626500 A CN113626500 A CN 113626500A
Authority
CN
China
Prior art keywords
projection
flood
index
value
function
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202110912605.8A
Other languages
Chinese (zh)
Inventor
张冬冬
徐高洪
戴明龙
李妍清
陈玺
李立平
王含
汪青静
刘冬英
邓鹏鑫
熊丰
黄燕
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Bureau of Hydrology Changjiang Water Resources Commission
Original Assignee
Bureau of Hydrology Changjiang Water Resources Commission
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Bureau of Hydrology Changjiang Water Resources Commission filed Critical Bureau of Hydrology Changjiang Water Resources Commission
Priority to CN202110912605.8A priority Critical patent/CN113626500A/en
Publication of CN113626500A publication Critical patent/CN113626500A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/24—Querying
    • G06F16/245—Query processing
    • G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
    • G06F16/2462—Approximate or statistical queries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/10—Services
    • G06Q50/26—Government or public services
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A10/00—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE at coastal zones; at river basins
    • Y02A10/40—Controlling or monitoring, e.g. of flood or hurricane; Forecasting, e.g. risk assessment or mapping

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Tourism & Hospitality (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Educational Administration (AREA)
  • Strategic Management (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Business, Economics & Management (AREA)
  • Probability & Statistics with Applications (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)
  • Databases & Information Systems (AREA)
  • Fuzzy Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Game Theory and Decision Science (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • General Health & Medical Sciences (AREA)

Abstract

The invention discloses a flood stage determination method based on multi-feature indexes, which comprises the following steps: calculating time interval division, selecting multi-feature indexes, constructing a projection index function, solving an optimal projection direction, calculating projection values of all time intervals and determining flood stages by ordered clustering. According to the method, on the basis of the duration time of the flood season of the clear drainage basin, a projection index function is constructed by adopting a plurality of characteristic indexes capable of reflecting drainage basin precipitation and flood characteristics, a projection value enabling the projection index function to reach the optimum is found through an accelerated genetic algorithm, and the projection value is subjected to cluster analysis through an ordered cluster analysis method, so that the flood stage is determined more objectively, and the defects of single analysis factor and subjective arbitrariness in a conventional analysis method are effectively avoided.

Description

Flood stage determination method based on multi-feature indexes
Technical Field
The invention relates to the technical field of hydrological computing, in particular to a flood stage determination method based on multi-feature indexes.
Background
In recent years, with the improvement of the hydrological meteorological prediction level, the forecast period and the forecast precision of flood forecast are improved to a certain extent, so that the reasonable utilization of flood resources and the comprehensive benefit exertion of hydraulic engineering become possible. When the flood is not generated in the near term, the reservoir appropriately impounds the medium and small flood to reduce the discharge flow, so that the required reservoir capacity of the large flood which is possibly generated in the later period is not greatly influenced, the downstream main flow can be controlled not to exceed the warning water level, and the downstream flood control pressure is effectively reduced. The implementation of the staged scheduling of the reservoir in the flood season is an important way for fully exerting the comprehensive benefits of the reservoir, and the key of the staged scheduling is the reasonable division of the flood season. Therefore, according to the distribution characteristics of the flood in the drainage basin, effective staging of the flood season by adopting a reasonable method is necessary, and scientific and reasonable flood season staging is an important subject for researching stage design flood water and relieving the contradiction between reservoir flood control and prosperity.
Aiming at the problem of flood season flood stage division, the flood season flood stage division is discussed in the Water conservancy and hydropower engineering design flood calculation Specification (SL44-2006), the division of the flood season stage division needs to have a relatively obvious flood cause change rule, each stage flood magnitude should have an obvious difference, and the division is preferably 2-3 stages. Flood staging methods are mainly of two types: weather cause analysis and mathematical statistics. The weather cause analysis is based on basin hydrology and weather analysis, flood seasons are staged according to basin weather characteristics, but the method is complex in combination mode of disaster weather for basins with large water collecting area, so that the method is large in workload and has certain subjectivity in staging. The mathematical statistics method generally adopts annual maximum flood annual distribution statistics, is simple and practical, but requires selection of critical values in a specific analysis process, is subjective and difficult to subdivide flood seasons, and meanwhile, the analysis method only takes a certain characteristic factor of flood as a staging basis, does not organically combine precipitation and other meteorological factors, and has insufficient integrity. In conclusion, the two flood staging methods have advantages and disadvantages, have certain subjectivity, and cannot determine flood staging in a visiting way.
Disclosure of Invention
The invention aims to provide a flood stage determining method based on multiple characteristic indexes, which is different from the prior single-factor analysis, adopts multiple characteristic indexes capable of reflecting the characteristics of drainage basin precipitation and flood to construct a projection index function, realizes parameter optimization through an accelerated genetic algorithm, finds a projection value enabling the projection index function to reach the optimum, and performs cluster analysis on the projection value through an ordered cluster analysis method, thereby more objectively determining the flood stage.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
a flood stage determination method based on multi-feature indexes comprises the following steps:
s1, determining the flood season time of the drainage basin according to the weather characteristics of the drainage basin, selecting the statistical duration, and determining the number of the calculation time periods;
s2, selecting multi-feature indexes representing precipitation and flood characteristics in a flood season, and counting statistical feature values of the indexes in different calculation time periods to form a sample set of the indexes;
s3, according to the sample set of each index obtained in the step S2, carrying out normalization processing on the sample set of each index to obtain a normalized index sample set;
s4, constructing a projection index function according to the normalized index sample set obtained in the step S3;
s5, solving the optimal projection direction of the projection index function by adopting an accelerated genetic algorithm according to the projection index function constructed in the step S4;
s6, calculating the projection value of each calculation time interval according to the optimal projection direction obtained in the step S5;
and S7, performing ordered clustering analysis according to the projection values of the calculation time periods obtained in the step S6, and determining flood periods of the flood season of the drainage basin.
Preferably, in step S1, ten days or months are used as a statistical duration.
Preferably, in step S1, one statistical duration is one calculation period.
Preferably, in step S2, the multi-feature indicators representing the characteristics of the flood season rainfall and the flood water include average ten-day rainfall in many years, a proportion of the current water volume in each ten-day flood season, the number of times of flood in each ten-day, and the maximum number of times of flood peak in each ten-day year.
Preferably, in step 2, one calculation period is one index sample.
Preferably, in step S3, the specific method for performing normalization processing on each index sample set is as follows:
first, the index sample sets obtained in step S2 are each set to { x }*(i, j) | i ═ 1,2, …, n; j ═ 1,2, …, p }, where x is*(i, j) is the jth index value for the ith sample; n and p are respectively the sample capacity and the index number;
then, normalization processing is performed on each index sample set by using the following formula (1), so as to obtain a normalized index sample set { x (i, j) | i ═ 1,2, … n; j ═ 1,2, … p };
x(i,j)＝[x*(i,j)-xmin(j)]/[xmax(j)-xmin(j)] (1)
in formula (1): x (i, j) is the j index normalization value of the ith sample; x is the number of*(i, j) is the j index statistical characteristic value of the ith sample; x is the number ofmin(j) Is the minimum value, x, of the statistical characteristic value of the jth indexmax(j) The maximum value of the j index statistical characteristic value is obtained.
Preferably, in step S4, the specific method for constructing the projection index function is as follows:
first, the normalized index sample set { x (i, j) | i ═ 1,2, … n obtained in step S3; j ═ 1,2, … p } projection to
Figure BDA0003204218310000031
Obtaining a corresponding one-dimensional linear space projection characteristic value set { z (i) ═ 1,2, … n } in a one-dimensional linear space of the projection direction;
wherein z (i) is obtained by the following formula (2):
Figure BDA0003204218310000032
in formula (2): z (i) is x (i, j) is in
Figure BDA0003204218310000033
Projection eigenvalues in the projection direction; a (j) is a unit vector of dimension j, j is 1,2, … p and satisfies
Figure BDA0003204218310000041
x (i, j) is the j index normalization value of the ith sample;
then, classifying according to the one-dimensional scatter diagram of the projection characteristic value set { z (i) }, i ═ 1,2, … n }, and constructing a corresponding projection index function Q (a) according to the classification result;
the constructed projection index function q (a) is specifically shown in the following formulas (3) to (5):
Q(a)＝SzDz (3)
Figure BDA0003204218310000042
Figure BDA0003204218310000043
in formulae (3) to (5): q (a) is a projection index function; szIs the standard deviation of the projection values; dzLocal density, which is a projection value; z (i) is x (i, j) is in
Figure BDA0003204218310000044
Projection eigenvalues in the projection direction; x (i, j) is the j index normalization value of the ith sample; ez is the mean of { z (i) ═ 1,2, … n }; r is the window radius of local density, and the general value R is 0.1Sz；rij| z (i) -z (j) | is the distance between points, rijThe larger the dot density, the smaller the dot density; u (R-R)ij) Is a unit step function when R is more than or equal to RijWhen the unit step function value is 1, when R is<rijThe unit step function value is 0.
Preferably, in step S5, the specific method for solving the optimal projection direction of the projection index function is as follows: solving the optimal solution determination of the projection index function by an accelerated genetic algorithm, namely, by calculating the Q (a) optimal value determination of the following conditions;
Figure BDA0003204218310000045
in formula (6): MaxQ (a) is the maximum value of the projection index function; szIs the standard deviation of the projection values; dzLocal density, which is a projection value; a (j) is a unit vector of dimension j, j is 1,2, … p and satisfies
Figure BDA0003204218310000051
Preferably, in step S6, the specific method for solving the projection value of each calculation period is as follows: the projection value for each calculation period is obtained by substituting the optimum projection direction vector of the projection index function obtained by the accelerated genetic algorithm in step S5 into the projection index function obtained in step S4.
Preferably, in step S7, performing ordered cluster analysis according to the projection value of each calculation time interval obtained in step S6 to determine the flood stage of the flood season of the drainage basin, and the specific method is as follows: the method comprises the steps of firstly carrying out ordered clustering classification on projection values of all calculation time periods, then calculating the sum of squared deviations of all the calculation time periods, then obtaining optimal classification according to an ordered clustering principle, and finally obtaining flood stages of the drainage basin flood period according to the optimal classification result of the ordered clustering of the projection values.
Compared with the prior art, the invention has the advantages that: the flood stage can be determined more objectively, and the defects of single analysis factor and subjective arbitrariness in the conventional analysis method are avoided.
The innovation points of the invention are as follows: the method comprises the steps of firstly constructing a projection index function by adopting a plurality of characteristic indexes capable of reflecting characteristics of drainage basin precipitation and flood, then realizing optimization of parameters through an accelerated genetic algorithm, finding a projection value enabling the projection index function to reach the optimal value, and finally carrying out cluster analysis on the projection value through an ordered cluster analysis method, thereby objectively determining the stage of the flood.
Drawings
FIG. 1 is a flow chart of a flood stage determination method based on multi-feature indexes according to the present invention;
FIG. 2 is a scatter diagram of the projection values of Yichang station in flood season;
fig. 3 is a flood stage achievement diagram obtained by the flood stage determination method of Yichang station based on the multi-characteristic indexes of the invention.
Detailed Description
In order to make the technical means, the creation features, the achievement purposes and the effects of the invention easy to understand, the following description further explains how the invention is implemented by combining the attached drawings and the detailed implementation modes.
Referring to fig. 1, the flood stage determination method based on multi-feature indexes provided by the present invention specifically includes the following steps:
s1, determining the flood season time of the drainage basin according to the weather characteristics of the drainage basin, selecting the statistical duration, and determining the number of the calculation time periods;
s2, selecting indexes representing precipitation and flood characteristics in a flood season, and counting statistical characteristic values of the indexes in different calculation periods to form a sample set of the indexes;
s3, according to the sample set of each index obtained in the step S2, carrying out normalization processing on the sample set of each index to obtain a normalized index sample set;
s4, constructing a projection index function according to the normalized index sample set obtained in the step S3;
s5, solving the optimal projection direction of the projection index function by adopting an accelerated genetic algorithm according to the projection index function constructed in the step S4;
s6, calculating the projection value of each calculation time interval according to the optimal projection direction obtained in the step S5;
and S7, performing ordered clustering analysis according to the projection values of the calculation time periods obtained in the step S6, and determining flood periods of the flood season of the drainage basin.
Specifically, in step S1, ten days or months may be used as a statistical duration.
Specifically, in step S1 of the present invention, a statistical duration is a calculation period.
Specifically, in step S2 of the present invention, the selected multi-feature index that represents precipitation and flood characteristics in flood season specifically includes: the average precipitation in ten days of the year, the proportion of the water volume in each ten days to the flood season, the number of times of flood in each ten days and the maximum number of times of flood peak in each ten days.
Specifically, in step S2 of the present invention, one calculation period is one index sample.
Specifically, in step S3 of the present invention, a specific method for performing normalization processing on each index sample set is as follows:
first, the sample set of each index obtained in step S2 is set to { x }*(i, j) | i ═ 1,2, …, n; j ═ 1,2, …, p }, where x is*(i, j) is the jth index value for the ith sample; n and p are respectively the sample capacity and the index number;
then, the following formula (1) is adopted to perform the sample set normalization processing of each index, so as to obtain a normalized index sample set { x (i, j) | i ═ 1,2, … n; j ═ 1,2, … p };
x(i,j)＝[x*(i,j)-xmin(j)]/[xmax(j)-xmin(j)] (1)
in formula (1): x (i, j) is the j index normalization value of the ith sample; x is the number of*(i, j) is the j index statistical characteristic value of the ith sample; x is the number ofmin(j) Is the minimum value, x, of the statistical characteristic value of the jth indexmax(j) The maximum value of the j index statistical characteristic value is obtained.
The purpose of this step S3 is to eliminate the difference between the index dimensions and unify the variation range of the index dimensions.
Specifically, in step S4 of the present invention, a specific method for constructing the projection index function is as follows:
first, the normalized index sample set { x (i, j) | i ═ 1,2, … n obtained in step S3; j ═ 1,2, … p } projection to
Figure BDA0003204218310000071
Obtaining a corresponding one-dimensional linear space projection characteristic value set { z (i) ═ 1,2, … n } in a one-dimensional linear space of the projection direction;
wherein z (i) can be specifically obtained by the following formula (2):
Figure BDA0003204218310000072
in formula (2): z (i) is x (i, j) is in
Figure BDA0003204218310000073
Projection eigenvalues in the projection direction; a (j) is a unit vector of dimension j, j is 1,2, … p and satisfies
Figure BDA0003204218310000074
x (i, j) is the j index normalization value of the ith sample;
then, classifying according to the one-dimensional scatter diagram of the projection characteristic value set { z (i) }, i ═ 1,2, … n }, and constructing a corresponding projection index function Q (a) according to the classification result;
in the synthesis of the projection values, the spread characteristics of the projection feature value set { z (i) ═ 1,2, … n } are required to be: the local projection points are as dense as possible and are condensed into a plurality of clusters; on the basis of the fact that the projected point groups are as far apart as possible from each other as a whole, the projection index function can be constructed as shown in the following equations (3) to (5):
Q(a)＝SzDz (3)
Figure BDA0003204218310000081
Figure BDA0003204218310000082
in formulae (3) to (5): q (a) is a projection index function; szIs the standard deviation of the projection values; dzLocal density, which is a projection value; z (i) is x (i, j) is in
Figure BDA0003204218310000083
Projection eigenvalues in the projection direction; x (i, j) is the j index normalization value of the ith sample; ez is the mean of { z (i) ═ 1,2, … n }; r is the window radius of local density, and the general value R is 0.1Sz；rij| z (i) -z (j) | is the distance between points, rijThe larger the dot density, the smaller the dot density; u (R-R)ij) Is a unit step function when R is more than or equal to RijWhen the unit step function value is 1, when R is<rijThe unit step function value is 0. Specifically, in step S5 of the present invention, the specific method for solving the optimal projection direction of the projection index function is as follows: and solving the optimal solution determination of the projection index function by introducing an accelerated genetic algorithm.
In practical applications, it is known that when a sample set of index values is given, the projection index function q (a) changes only with the change of the projection direction a; different projection directions reflect different data structure features, and the optimal projection direction is the projection direction which is most likely to expose a certain feature structure of high-dimensional data, so we can estimate the optimal projection direction by solving the projection index function maximization problem, that is, the optimal projection direction is determined by the following formula (6):
Figure BDA0003204218310000091
in formula (6): MaxQ (a) is the maximum value of the projection index function; szIs the standard deviation of the projection values; dzLocal density, which is a projection value; a (j) is a unit vector of dimension j, j is 1,2, … p and satisfies
Figure BDA0003204218310000092
However, the above equation (6) is a complex nonlinear optimization problem using { a (j) }, j ═ 1,2, … p } as an optimization variable, and it is difficult to process by using a conventional optimization method, so that an accelerated genetic algorithm (RAGA) is adopted to solve the optimal solution, that is: and solving a projection index function based on an accelerated genetic algorithm, and calculating the optimal projection direction vector of the projection index function.
Specifically, in step S6 of the present invention, the specific method for solving the projection value of each calculation period is as follows: the projection value for each calculation period is obtained by substituting the optimum projection direction vector of the projection index function obtained by the accelerated genetic algorithm in step S5 into the projection index function obtained in step S4.
Specifically, in step S7 of the present invention, ordered cluster analysis is performed according to the projection values of each calculation time interval obtained in step S6 to determine flood stages of the flood season of the drainage basin, and the specific method is as follows: the method comprises the steps of firstly carrying out ordered clustering classification on projection values of all calculation time periods, then calculating the sum of squared deviations of all the calculation time periods, then obtaining optimal classification according to an ordered clustering principle, and finally obtaining flood stages of the drainage basin flood period according to the optimal classification result of the ordered clustering of the projection values.
For example: for n calculation periods, firstly, projection values of the n calculation periods are classified into the following k classes according to the ordered clustering classification principle, and the specific classification results of the k classes are as follows:
Figure BDA0003204218310000093
then, the dispersion square sum of k classes is calculated respectively by the following formula (7);
Figure BDA0003204218310000094
in formula (7): sjIs the sum of squared deviations of class j;
Figure BDA0003204218310000101
is the jth class mean;
Figure BDA0003204218310000102
is the value of the mth item in the jth class;
then, according to the principle of ordered clustering, obtaining the optimal classification, namely making the variance square sum S of each class under the condition that n and k are fixedjThe smallest is the optimal classification.
And finally, obtaining flood stages of the flood season of the drainage basin according to the optimal classification result.
Taking Yichang station as an example, the following concrete description is how to determine flood season by using the flood season determination method based on multi-characteristic indexes provided by the invention:
step one, determining the flood season time period of a drainage basin according to the weather characteristics of the drainage basin where the Yichang station is located, selecting statistical time length, and determining the number of the calculation time periods;
in this embodiment, since the beginning time of the flood season of the drainage basin at the place of Yichang station in the upstream of the Yangtze river is known to be 5 months to 10 months according to the weather cause of precipitation in the upstream of the Yangtze river, the drainage basin flood season period of Yichang station can be determined to be 5 months to 10 months, the statistical duration is preferably ten days as a calculation period, and therefore, the number of the calculation periods can be determined to be 18 at this time, that is, the calculation periods are respectively: 5, 6, 7, 8, 9, 10 middle, 10;
secondly, selecting indexes representing precipitation and flood characteristics of Yichang station flood season, and counting statistical characteristic values of all indexes in different calculation time periods to form a sample set { x ] of all indexes*(i, j) | i ═ 1,2, …, n; j ═ 1,2, …, p }, where: x is the number of*(i, j) is the jth index value for the ith sample; n and p are respectively the sample capacity and the index number;
in this embodiment, in the field of the place in Yichang at the place upstream of the Yangtze river, according to the characteristics of the upstream flood of the Yangtze river, the average ten-year precipitation in ten days, the proportion of the water volume in ten days to the flood season, the number of times of flood in ten days and the maximum number of times of flood peak in ten days are selected as 4 indexes for representing the characteristics of the flood in the flood season, and the statistical characteristic values of the indexes in the 18 calculation periods determined in the first step are counted, which is shown in table 1 below;
TABLE 1
Figure BDA0003204218310000111
In this embodiment, since there are 18 calculation time periods, 18 sample sets can be formed, that is, n takes a value of 18, and the index takes 4, so p takes a value of 4; thus, the sample set formed by this embodiment is { x }*(i,j)|i＝1,2,…,18；j＝1,2,…,4}；
Thirdly, adopting a normalization formula (1) to perform sampling set { x) of the 4 indexes obtained in the second step*(i, j) | i ═ 1,2, …, 18; j is 1,2, …,4, to get the 4 index sample set;
step four, firstly, the normalized index sample set obtained in the step three is integrated
Figure BDA0003204218310000112
A one-dimensional set of projection values { z (i) } 1,2, … n } for the projection direction, then sorted according to a one-dimensional scatter plot of { z (i) } 1,2, … n } and sorted according to a one-dimensional set of { z (i) } 1,2, … n }, andconstructing a corresponding projection index function according to the scatter diagram classification result;
fifthly, solving the optimal projection direction quantity of the projection index function formed in the fourth step through an accelerated genetic algorithm;
in this embodiment, the solved optimal projection direction quantity a is {0.0283, 0.7081, 0.2826, 0.6465 }.
Sixthly, substituting the optimal projection direction quantity a obtained in the fifth step into the projection index function constructed in the fourth step, and calculating the projection value of each calculation time interval; in the present embodiment, the projection values of 18 calculation periods are calculated, and the table 1 and fig. 2 can be specifically seen.
Seventhly, drawing a projection value one-dimensional scatter diagram corresponding to 18 calculation time periods calculated in the sixth step by taking ten days as time periods, dividing the flood season into front, main and rear 3 time periods according to flood season characteristics, obtaining projection value classification characteristics of 18 calculation time periods according to an ordered clustering method, and obtaining the branch period of the Yichang station flood season according to projection value classification characteristic results of 18 calculation time periods; in this example, the obtained flood season results of Yichang station are as follows: the early season is from 5 to 6 months, the main flood season is from 6 to 9 months, and the late season is from 9 to 10 months.
Finally, the above description is only an embodiment of the present invention, and not intended to limit the scope of the present invention, and all equivalent structures or equivalent processes performed by the present invention or directly or indirectly applied to other related technical fields using the contents of the present specification and the attached drawings are included in the scope of the present invention.

Claims (10)

1. A flood stage determination method based on multi-feature indexes is characterized by comprising the following steps:
s1, determining the flood season time of the drainage basin according to the weather characteristics of the drainage basin, selecting the statistical duration, and determining the number of the calculation time periods;
s2, selecting multi-feature indexes representing precipitation and flood characteristics in a flood season, and counting statistical feature values of the indexes in different calculation time periods to form a sample set of the indexes;
s3, according to each index sample set obtained in the step S2, normalization processing is carried out on each index sample set to obtain a normalized index sample set;
s4, constructing a projection index function according to the normalized index sample set obtained in the step S3;
s5, solving the optimal projection direction of the projection index function by adopting an accelerated genetic algorithm according to the projection index function constructed in the step S4;
s6, calculating the projection value of each calculation time interval according to the optimal projection direction obtained in the step S5;
and S7, performing ordered clustering analysis according to the projection values of the calculation time periods obtained in the step S6, and determining the flood period of the flood season of the basin.
2. The method of claim 1, wherein in step S1, ten days or months are used as a statistical duration.
3. The flood stage determining method according to claim 1, wherein in step S1, a statistical duration is a calculation period.
4. The flood stage determination method based on multi-feature indexes of claim 1, wherein in the step S2, the multi-feature indexes characterizing the flood period precipitation and flood characteristics include average ten-day precipitation for many years, a proportion of the inflow to the flood period in each ten-day, the number of times of flood in each ten-day, and the maximum number of times of flood peak in each ten-day.
5. The flood stage determining method according to claim 1, wherein in the step 2, one calculation period is one index sample.
6. The flood stage determination method based on multi-feature indexes according to claim 1, wherein in the step S3, the specific method for performing normalization processing on each index sample set is as follows:
first, the index sample sets obtained in step S2 are each set to { x }*(i, j) | i ═ 1,2, …, n; j ═ 1,2, …, p }, where x is*(i, j) is the jth index value for the ith sample; n and p are respectively the sample capacity and the index number;
then, normalization processing is performed on each index sample set by using the following formula (1), so as to obtain a normalized index sample set { x (i, j) | i ═ 1,2, … n; j ═ 1,2, … p };
x(i,j)＝[x*(i,j)-xmin(j)]/[xmax(j)-xmin(j)] (1)
in formula (1): x (i, j) is the j index normalization value of the ith sample; x is the number of*(i, j) is the j index statistical characteristic value of the ith sample; x is the number ofmin(j) Is the minimum value, x, of the statistical characteristic value of the jth indexmax(j) The maximum value of the j index statistical characteristic value is obtained.
7. The flood stage determining method according to claim 6, wherein in the step S4, the specific method for constructing the projection index function is as follows:
first, the normalized index sample set { x (i, j) | i ═ 1,2, … n obtained in step S3; j ═ 1,2, … p } projection to
Figure FDA0003204218300000021
Obtaining a corresponding one-dimensional linear space projection characteristic value set { z (i) ═ 1,2, … n } in a one-dimensional linear space of the projection direction;
wherein z (i) can be specifically obtained by the following formula (2):
Figure FDA0003204218300000022
in formula (2): z (i) is x (i, j) is in
Figure FDA0003204218300000023
Projection eigenvalues in the projection direction; a (j) is a unit vector of dimension j, j is 1,2, … p and satisfies
Figure FDA0003204218300000024
x (i, j) is the j index normalization value of the ith sample;
then, classifying according to the one-dimensional scatter diagram of the projection characteristic value set { z (i) }, i ═ 1,2, … n }, and constructing a corresponding projection index function Q (a) according to the classification result;
the constructed projection index function q (a) is specifically shown in the following formulas (3) to (5):
Q(a)＝SzDz (3)
Figure FDA0003204218300000031
Figure FDA0003204218300000032
in formulae (3) to (5): q (a) is a projection index function; szIs the standard deviation of the projection values; dzLocal density, which is a projection value; z (i) is x (i, j) is in
Figure FDA0003204218300000033
Projection eigenvalues in the projection direction; x (i, j) is the j index normalization value of the ith sample; ez is the mean of { z (i) ═ 1,2, … n }; r is the window radius of local density, and the general value R is 0.1Sz；rij| z (i) -z (j) | is the distance between points, rijThe larger the dot density, the smaller the dot density; u (R-R)ij) Is a unit step function when R is more than or equal to RijWhen the unit step function value is 1, when R is<rijThe unit step function value is 0.
8. The flood stage determining method according to claim 7, wherein in the step S5, the specific method for solving the optimal projection direction of the projection index function is as follows: solving the optimal solution determination of the projection index function by an accelerated genetic algorithm, namely, by calculating the Q (a) optimal value determination of the following conditions;
Figure FDA0003204218300000034
in formula (6): MaxQ (a) is the maximum value of the projection index function; szIs the standard deviation of the projection values; dzLocal density, which is a projection value; a (j) is a unit vector of dimension j, j is 1,2, … p and satisfies
Figure FDA0003204218300000041
9. The flood stage determination method based on multi-feature indexes according to claim 1, wherein in the step S6, the specific method for solving the projection value of each calculation period is as follows: the projection value for each calculation period is obtained by substituting the optimum projection direction vector of the projection index function obtained by the accelerated genetic algorithm in step S5 into the projection index function obtained in step S4.
10. The flood stage determination method based on multi-feature indexes according to claim 1, wherein in the step S7, according to the projection values of the calculation time periods obtained in the step S6, ordered clustering analysis is performed to determine the flood stage of the flood season of the drainage basin, and the specific method is as follows: the method comprises the steps of firstly carrying out ordered clustering classification on projection values of all calculation time periods, then calculating the sum of squared deviations of all the calculation time periods, then obtaining optimal classification according to an ordered clustering principle, and finally obtaining flood stages of the drainage basin flood period according to the optimal classification result of the ordered clustering of the projection values.
CN202110912605.8A 2021-08-10 2021-08-10 Flood stage determination method based on multi-feature indexes Pending CN113626500A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110912605.8A CN113626500A (en) 2021-08-10 2021-08-10 Flood stage determination method based on multi-feature indexes

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110912605.8A CN113626500A (en) 2021-08-10 2021-08-10 Flood stage determination method based on multi-feature indexes

Publications (1)

Publication Number Publication Date
CN113626500A true CN113626500A (en) 2021-11-09

Family

ID=78383939

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110912605.8A Pending CN113626500A (en) 2021-08-10 2021-08-10 Flood stage determination method based on multi-feature indexes

Country Status (1)

Country Link
CN (1) CN113626500A (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114065521A (en) * 2021-11-18 2022-02-18 三峡大学 Improved basin flood season staging method, system and storage medium
CN114707879A (en) * 2022-04-18 2022-07-05 江西省水利科学院 Flood season staged dynamic flood limit water level determination method based on pre-discharge rule
CN115841192A (en) * 2023-02-16 2023-03-24 长江水利委员会水文局 Flood season staging method and system based on information entropy theory

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103295079A (en) * 2013-06-09 2013-09-11 国家电网公司 Electric power multi-objective decision support method based on intelligent data mining model
CN106021961A (en) * 2016-06-20 2016-10-12 吉林大学 Urban standard cyclic working condition constructing method based on genetic algorithm optimization
CN108897977A (en) * 2018-08-24 2018-11-27 河海大学 A kind of uncertain attribution method of Runoff Evolution based on the simulation of big Regional Hydrologic
CN112883558A (en) * 2021-01-27 2021-06-01 长江水利委员会水文局 Hydrological model parameter time-varying form construction method
CN113159451A (en) * 2021-05-13 2021-07-23 长江水利委员会水文局 Long-term prediction method for drainage basin drought and flood events based on event knowledge graph construction

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103295079A (en) * 2013-06-09 2013-09-11 国家电网公司 Electric power multi-objective decision support method based on intelligent data mining model
CN106021961A (en) * 2016-06-20 2016-10-12 吉林大学 Urban standard cyclic working condition constructing method based on genetic algorithm optimization
CN108897977A (en) * 2018-08-24 2018-11-27 河海大学 A kind of uncertain attribution method of Runoff Evolution based on the simulation of big Regional Hydrologic
CN112883558A (en) * 2021-01-27 2021-06-01 长江水利委员会水文局 Hydrological model parameter time-varying form construction method
CN113159451A (en) * 2021-05-13 2021-07-23 长江水利委员会水文局 Long-term prediction method for drainage basin drought and flood events based on event knowledge graph construction

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
周研来: "梯级水库群联合优化调度运行方式研究", 中国博士论文全文数据库 工程科技Ⅱ辑, pages 037 - 24 *
王宗志等: "基于模糊时序有效聚类的太湖流域汛期分期研究", 水力发电学报, pages 29 - 34 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114065521A (en) * 2021-11-18 2022-02-18 三峡大学 Improved basin flood season staging method, system and storage medium
CN114065521B (en) * 2021-11-18 2024-06-28 三峡大学 Improved flood season staging method, system and storage medium
CN114707879A (en) * 2022-04-18 2022-07-05 江西省水利科学院 Flood season staged dynamic flood limit water level determination method based on pre-discharge rule
CN114707879B (en) * 2022-04-18 2023-12-08 江西省水利科学院 Flood season stage dynamic flood limit water level determining method based on pre-leakage rule
CN115841192A (en) * 2023-02-16 2023-03-24 长江水利委员会水文局 Flood season staging method and system based on information entropy theory

Similar Documents

Publication Publication Date Title
Liu et al. Random forest solar power forecast based on classification optimization
CN113626500A (en) Flood stage determination method based on multi-feature indexes
CN106650767B (en) Flood forecasting method based on cluster analysis and real-time correction
Xu et al. The spatial and temporal variation features of wind-sun complementarity in China
Rao et al. Regionalization of watersheds by hybrid-cluster analysis
CN110929956B (en) Flood forecasting scheme real-time optimization method based on machine learning
Kumar et al. Regional flood frequency analysis using soft computing techniques
CN113255986B (en) Multi-step daily runoff forecasting method based on meteorological information and deep learning algorithm
CN101673369A (en) Projection pursuit-based method for evaluating flooding risk of drainage pipe network
Li et al. Analysis of the relationship between soil erosion risk and surplus floodwater during flood season
CN112508237B (en) Rain type region division method based on data analysis and real-time rain type prediction method
CN109165693A (en) It is a kind of to sentence knowledge method automatically suitable for dew, frost and the weather phenomenon of icing
CN117977568A (en) Power load prediction method based on nested LSTM and quantile calculation
CN115689051A (en) Method for automatically calibrating SWMM model parameters based on GA algorithm coupling Morris and GLUE
Yin Hazard assessment and regionalization of highway flood disasters in China
Jiang et al. Hydrological characteristic-based methodology for dividing flood seasons: an empirical analysis from China
CN116205136A (en) Large-scale river basin deep learning flood forecasting method based on runoff lag information
CN117332291A (en) Distributed photovoltaic-oriented regional resource monitoring layout method and system
Tayşi et al. Disaggregation of future GCMs to generate IDF curves for the assessment of urban floods
CN111709586B (en) Water level typing design and control method for flood diversion in reservoir basin
CN117058433A (en) Ecological hydrologic partition method based on Gaussian mixture clustering algorithm
CN116863651A (en) Landslide weather early warning method based on rainfall event early warning response dynamic optimization
CN116626780A (en) Weather forecast post-processing method and system for ground weather station set forecast
Huffman et al. An exploratory analysis of wind patterns of Alberta, Canada
Wang et al. Ice-jam flood hazard risk assessment under simulated levee breaches using the random forest algorithm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination