CN118095973A - Hydrological abundant encounter probability calculation method based on coupling dimension reduction theory - Google Patents
Hydrological abundant encounter probability calculation method based on coupling dimension reduction theory Download PDFInfo
- Publication number
- CN118095973A CN118095973A CN202410524576.1A CN202410524576A CN118095973A CN 118095973 A CN118095973 A CN 118095973A CN 202410524576 A CN202410524576 A CN 202410524576A CN 118095973 A CN118095973 A CN 118095973A
- Authority
- CN
- China
- Prior art keywords
- function
- probability
- distribution
- climate
- hydrological
- 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
Links
- 230000009467 reduction Effects 0.000 title claims abstract description 24
- 238000004364 calculation method Methods 0.000 title claims abstract description 21
- 230000008878 coupling Effects 0.000 title claims abstract description 16
- 238000010168 coupling process Methods 0.000 title claims abstract description 16
- 238000005859 coupling reaction Methods 0.000 title claims abstract description 16
- 238000009826 distribution Methods 0.000 claims abstract description 69
- 230000000694 effects Effects 0.000 claims abstract description 50
- 230000001186 cumulative effect Effects 0.000 claims abstract description 48
- 238000000034 method Methods 0.000 claims abstract description 40
- 238000005315 distribution function Methods 0.000 claims abstract description 38
- 241000039077 Copula Species 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 66
- 230000006870 function Effects 0.000 claims description 38
- 238000011156 evaluation Methods 0.000 claims description 22
- 238000004422 calculation algorithm Methods 0.000 claims description 12
- 238000005457 optimization Methods 0.000 claims description 10
- 238000011160 research Methods 0.000 claims description 6
- 238000007418 data mining Methods 0.000 claims description 5
- 238000007689 inspection Methods 0.000 claims description 4
- 238000012886 linear function Methods 0.000 claims description 4
- 238000012847 principal component analysis method Methods 0.000 claims description 3
- 238000007637 random forest analysis Methods 0.000 claims description 3
- 238000012935 Averaging Methods 0.000 claims description 2
- 238000013473 artificial intelligence Methods 0.000 claims description 2
- 238000013528 artificial neural network Methods 0.000 claims description 2
- 230000002068 genetic effect Effects 0.000 claims description 2
- 239000002245 particle Substances 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 238000012887 quadratic function Methods 0.000 claims description 2
- 238000012163 sequencing technique Methods 0.000 claims 1
- 238000004458 analytical method Methods 0.000 abstract description 10
- 239000003638 chemical reducing agent Substances 0.000 abstract description 3
- 238000012360 testing method Methods 0.000 description 14
- 230000008859 change Effects 0.000 description 12
- 239000010931 gold Substances 0.000 description 11
- 229910052737 gold Inorganic materials 0.000 description 11
- 230000000052 comparative effect Effects 0.000 description 9
- PCHJSUWPFVWCPO-UHFFFAOYSA-N gold Chemical compound [Au] PCHJSUWPFVWCPO-UHFFFAOYSA-N 0.000 description 9
- 238000001556 precipitation Methods 0.000 description 6
- 238000010276 construction Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000003973 irrigation Methods 0.000 description 3
- 230000002262 irrigation Effects 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 239000000654 additive Substances 0.000 description 2
- 230000000996 additive effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 201000004569 Blindness Diseases 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000001364 causal effect Effects 0.000 description 1
- 238000010835 comparative analysis Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012821 model calculation Methods 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
Landscapes
- Complex Calculations (AREA)
Abstract
The invention discloses a hydrological abundant encounter probability calculation method based on a coupling dimension reduction theory, which comprises the steps of obtaining and preprocessing each index data; obtaining an optimal distribution model; obtaining a relation function of a cumulative distribution function and a cumulative experience frequency; the comprehensive climate index and the comprehensive human activity index are obtained through the high-efficiency dimension reducer; analyzing the cumulative distribution function of each edge distribution after the influence of the changing environment is considered by adopting an optimal distribution model, substituting the cumulative distribution function after the influence of the changing environment into a corresponding relation function, and obtaining a final cumulative distribution function of the edge distribution of the hydrologic climate variable; and calculating the probability of the abundant-blight encounter by adopting a Copula function method. The method and the device fully solve the problems of poor hydrologic frequency precision in a changing environment, insufficient consideration of climate and human activity factors in an interpretation variable, and analysis of the dimension disaster of the interpretation variable by a traditional GAMLSS model, and improve the analysis precision and efficiency of hydrologic abundant encounter in a changing environment background.
Description
Technical Field
The invention belongs to the technical field of hydrologic model calculation, and particularly relates to a hydrologic abundant encounter probability calculation method based on a coupling dimension reduction theory.
Background
The calculation of the hydrological encounter probability among different watersheds or different areas of the same watershed has important significance for the establishment of a flood and drought disaster combined defense scheme among the watersheds and the construction of a water network engineering, for example, the combined flood control scheduling among the watersheds is restricted due to the overlarge probability of the water-rich encounter. For water network engineering, if the probability of encountering water withered in a water supply area and a water receiving area is too large, the running of water diversion across a river basin is not facilitated. The drainage basin hydrologic process consistency is destroyed under the joint influence of global climate change and human activities, and the frequency and intensity of the drainage basin hydrologic abundant meeting are more complex.
Currently, hydrologic abundant encounter probability analysis under hydrologic non-uniformity conditions has formed a multi-dimensional joint distribution simulation method represented by a coupled Generalized Additive Model (GAMLSS) and Copula function, wherein the Generalized Additive Model (GAMLSS) is used to obtain an edge distribution function. However, this method still has the following drawbacks: firstly, the method still estimates the frequency distribution characteristics of the hydrologic variable in a mode of optimizing a certain frequency distribution model (such as P-III distribution), and the accuracy is still to be improved although a certain subjective blindness is avoided in the selection of the distribution model. Secondly, the method often uses a single index or a plurality of indexes (such as time T, rainfall P, air temperature T and the like) as an explanatory variable, and the actions of climate change and human activity cannot be comprehensively considered in the selection of the explanatory variable. However, the climate and human activity indexes are complex and various, if all indexes are considered as explanatory variables, dimension disasters occur, and even the model cannot operate, so that a scientific and efficient dimension reduction method is urgently needed to obtain the explanatory variables reflecting the influence of the climate and human activity. In summary, for hydrologic encounter probability analysis in varying environments, there is a need to increase computational accuracy while maintaining adequate consideration of climate and human activity impact.
Disclosure of Invention
The invention aims to solve the problems that the existing method is low in precision, the effects of climate change and human activity cannot be comprehensively considered, and the variable dimension disaster is analyzed and explained by the traditional GAMLSS model, and provides a calculation method for the hydrological abundant encounter, which can fully reflect the influences of the climate change and the human activity, and can realize the calculation of the hydrological abundant encounter with high precision under the background of a changed environment.
The invention adopts the following technical scheme:
A hydrological abundant encounter probability calculation method based on a coupling dimension reduction theory specifically comprises the following steps:
Step S1: acquiring climate index data, hydrological index data and human activity index data of a research area, and preprocessing data, including quality control, noise removal, interpolation extension and the like;
Step S2: based on hydrological index data, selecting an optimal distribution model of the edge distribution of the research area by adopting a distribution model selector;
step S3: loading a cumulative distribution function CDF (i) and a cumulative experience frequency PE (i) obtained by calculating the optimal distribution model into a feature learner, and obtaining a relation function F (PE (i), CDF (i)) of the cumulative distribution function and the cumulative experience frequency through the feature learner; wherein i represents the number of frequency distribution models;
Step S4: loading the climate index and the human activity index into a high-efficiency dimension reduction device for dimension reduction treatment to obtain a comprehensive climate index KC and a comprehensive human activity index KH after dimension reduction;
Step S5: analyzing a cumulative distribution function CDF_CHO (i) of edge distribution after the influence of a changing environment is considered by adopting a GAMLSS model by taking a comprehensive climate index KC and a comprehensive human activity index KH as final explanatory variables, and substituting the CDF_CHO (i) into a relation function F (PE (i) and a CDF (i)) corresponding to the step S3 to obtain a final cumulative distribution function CDF_CH (i) of the edge distribution of the hydrologic climate variable; and finally, calculating the abundant encounter probability by adopting a Copula function method based on the obtained CDF_CH (i).
Further, the data preprocessing in step S1 includes:
And performing missing value inspection, consistency inspection and white noise treatment on the climate data, the hydrologic data and the human activity index data. And performing interpolation extension on the data in the case of missing data. And if only the individual data is missing, the interpolation is carried out by adopting an adjacent time average value taking method, and when the missing data is more, the interpolation is carried out by adopting a hydrological comparison method and an SDSM statistical downscaling model.
Further, the step S2 includes:
Step S2.1: firstly, adopting a plurality of hydrologic frequency distribution models in a distribution model selector to carry out univariate frequency distribution fitting on a runoff sequence. The selected frequency distribution model comprises 10 two-parameter models and 5 three-parameter models; the two parameter models include GA, GU, IG, LO, LOGNO, NO, RG, WEI, WEI and WEI3; the three parameter model includes GG, BCCG, GIG, PE and P-III. The above 15 models are prior art, and their names are also prior art, and the model abbreviations are not explained here one by one.
All the method and technology sets for realizing the optimization of the distribution model are named as a distribution model selector, namely the distribution model selector comprises ten two-parameter models and five three-parameter models.
Step S2.2: based on the 15 frequency distribution fitting results, performing precision evaluation on each frequency distribution model by adopting Global deviance, AIC, SBC, residual error Mean index Mean, residual error Variance index Variance, residual error Filliben coefficient, residual error bias coefficient Skewness and residual Kurtosis coefficient Kurtosis 8 evaluation indexes. Wherein the smaller the values of Global deviance, AIC and SBC, the closer the Mean and Skewness are to 0, the closer the variance and Filliben are to 1, the closer the kurtosis is to 3, the better the simulation accuracy of the model. If the conclusion of all the evaluation indexes is consistent, the optimal frequency distribution model can be optimized; and if the conclusion of all the evaluation indexes is inconsistent, further performing final model optimization.
Step S2.3: in order to solve the problem that the optimal result cannot be effectively optimized when the multiple indexes are evaluated in the step 2.2 (for example, part of indexes represent that the model 1 is optimal, and the other part of evaluation indexes are considered that the model 2 is optimal), the final optimization of the model result is performed by adopting a TOPSIS comprehensive evaluation method based on entropy weight:
And meet/>
In the method, in the process of the invention,The method comprises the steps that (1) a j-th evaluation index normalization value of an i-th frequency distribution model is obtained, m is the total number of models, n is the number of evaluation index types, and R ij represents the proportion of the j-th evaluation index of the i-th frequency distribution model; s j is the entropy value of the j-th index; /(I)Then the weight occupied by the j index is/isFor the ideal solution of the j-th index,/>Is the negative ideal solution of the j-th index,/>For the distance between each index and the ideal solution,/>For the distance between each index and the negative ideal solution, C is the final model evaluation parameter, the model is more optimal as the value of C is larger, and Q j is the j index value.
Further, the feature learner in step S3 includes a plurality of feature functions for learning a relationship between the cumulative distribution function and the cumulative experience frequency, and selects a final feature function form according to the learning effect.
Further, in step S3, the feature learner includes five feature functions, and if the cumulative distribution function and the cumulative empirical frequency are x and y, respectively, the five function forms are:
Exponential function: y=ae bx
Linear function: y=ax+b
Logarithmic function: y= alnx +b
Polynomial function, for example quadratic function, y=ax 2 +bx+c
Power function: y=ax b
Wherein the parameters a, b, c are determined by artificial intelligence optimization algorithms including, but not limited to, genetic algorithms, particle swarm algorithms, or neural networks.
Further, the step S3 specifically includes:
step S3.1; extracting a cumulative distribution function CDF (i) obtained by calculating an optimal distribution model, and forming a learning pair (PE (i) and CDF (i)) by the cumulative distribution function and the cumulative empirical frequency PE (i);
step S3.2: each group of learning pairs is loaded into a feature learner to learn, and a relation function F (PE (i)) of CDF (i) and PE (i) is obtained through the feature learner.
The invention performs data fitting on the relation between the cumulative distribution function and the cumulative experience frequency through five characteristic functions, and for convenience of description, the invention names a method set for learning the relation rule between the cumulative distribution function and the cumulative experience frequency as a characteristic learner, namely the characteristic learner of the invention comprises five characteristic functions.
The invention relates to a method and a technology set for reducing the dimensions of various climate and human activity indexes to comprehensive indexes of two dimensions, namely a high-efficiency dimension reducer.
Further, the step S4 includes:
Step S4.1: in the high-efficiency dimension reducer, firstly, an Apriori data mining algorithm is adopted to carry out preliminary screening on the climate data and the human activity index data preprocessed in the step S1, and unimportant or insignificant influence factors are removed.
The Apriori data mining algorithm is used to find frequent patterns, associations, correlations, or causal structures that exist between a set of hydrologic elements and a set of climate or human activity index elements, belonging to unsupervised learning. Because the data classification attribute data form is needed in the Apriori data mining algorithm, the K-means clustering algorithm is adopted to classify the original data, and a classification attribute data set is constructed.
Taking runoff data as an example, expressing the collected runoff as a data set R, randomly selecting K runoff values from the data set R as K clustering centers to be divided, and respectively calculating Euclidean distances between any runoff value and any clustering center. And dividing the runoff value into K classes according to the Euclidean distance minimum principle. And averaging all the path values of the K classes obtained by dividing, enabling the path values to be new clustering centers, and repeating the operation until the clustering center values calculated twice in adjacent iteration are not changed, so that a final classification result is obtained.
The method for calculating the euclidean distance is the prior art and will not be described in detail herein.
After the classification attribute data set is constructed, association rules of the climate index, the human activity index and the runoff elements are calculated respectively, and the support degree and the confidence degree of the association rules are analyzed. And eliminating the climate index and the human activity index which do not meet the threshold of the support degree and the confidence degree according to the calculation result of the support degree and the confidence degree of the association rule. Finally, element indexes with obvious influence on hydrology, namely an effective climate influence factor set and an effective human activity influence factor set, can be selected from all the candidate climate indexes and human activity indexes.
The form and calculation method of the support and the confidence belong to the prior art, and are not described in detail here.
Step S4.2: based on the results of the effective climate influence factor set and the effective human activity influence factor set, a random forest model is adopted to analyze the importance of each influence element in each influence factor set, and the importance is ranked. The results of the importance analysis will be used to determine the weight of each influencing element when calculating the integrated climate index KC and the integrated human activity index KH.
Step S4.3: and constructing a comprehensive climate index and a comprehensive human activity index by adopting a principal component analysis method of the coupling information entropy. Taking the comprehensive climate index KC as an example, the main component analysis method of the coupling information entropy comprises the following steps:
① Calculating the correlation coefficient matrix of each climate influence factor
Where cl rs is the correlation coefficient of the r-th and s-th climate influencing factors, r=1, 2, …, k; s=1, 2, …, k; cl rs=clsr.
② Solving the characteristic equation to obtain a characteristic valueAnd are arranged in a descending order to further calculate the corresponding feature vector e cl,r.
③ The main component contribution rate is calculated as follows:
Calculating the accumulated contribution rate of the principal components:
Where p is a feature number satisfying that the cumulative contribution is greater than a threshold value, Is the eigenvalue of item p.
④ Calculating the load of the main components: Where e cl,r,s is the s-th component of vector e cl,r.
⑤ The final weight of each climate factor variable is determined. And according to the linear combination coefficient of each main component and the variance contribution rate of the main component of each climate influence factor variable index, the weight w cl,r of the climate influence factor variable index can be obtained by combining the random forest importance ranking result in the step 42.
⑥ Calculating a climate change index; normalizing each selected representative climate variable to avoid the problem of non-uniform dimension, and then solving the climate change index (normalization) as follows:
wherein KC is the finally obtained standardized climate change index; cc r is the r-th climate-influencing factor variable, and w cl,r is the weight corresponding to the r-th climate-influencing factor variable.
Further, the step S5 includes:
Step S5.1: taking KC and KH corresponding to the step S4 as final explanatory variables, and adopting an optimal distribution model to analyze a cumulative distribution function CDF_CHO (i) of each edge distribution after the influence of a changing environment is considered;
Step S5.2: substituting CDF_CHO (i) into a relation function F (PE (i) and CDF (i)) corresponding to the step S3 to obtain a final cumulative distribution function CDF_CH (i) of the hydrologic climate variable edge distribution; assuming that the relationship function is a linear function, the specific formula is as follows:
CDF_CH(i)=CDF_CHO(i)*a+b
Step S5.3: based on the obtained CDF_CH (i), the joint distribution probability is calculated by adopting a Copula function method.
Step S5.4: calculation of the abundant withering encounter. The corresponding frequencies of the national abundant grade division are P f =37.5% and P k =62.5% (no more than probability), in the method, two hydrologic time sequences are respectively X and Y, the edge distribution functions of the two hydrologic time sequences are u and v respectively, Y pf is the water quantity corresponding to the frequency of P f in the hydrologic time sequence Y, Y pk is the water quantity corresponding to the frequency of P k in the hydrologic time sequence Y, X pf is the water quantity corresponding to the frequency of P f in the hydrologic time sequence X, and X pk is the water quantity corresponding to the frequency of P k in the hydrologic time sequence X. Taking the hydrologic time sequence X as an example, X t≥Xpf is high water, X t≤Xpk is dead water, X pk＜Xt＜Xpf is flat water, wherein X t is the hydrologic quantity of the t year. Then:
The probability of X and Y homoabundant (Feng Feng types) is:
The probability of X and Y being co-planar (flat) is:
the probability of X and Y being co-withered (withered) is:
The probability of X-Feng-Y-plane (Feng Ping type) is:
the probability of X-plane-Y-abundance (Pingfeng) is:
The probability of X to Y withering (withering) is:
The probability of X cumy abundance (cumfeng) is:
The probability of X flat Y dead (flat dead) is:
The probability of X dead Y flat (dead flat) is:
Where u pf、vpf、upk、 vpk is the edge distribution function value corresponding to X pf、Ypf、Xpk、Ypk, and C () is the Copula connection function.
The invention has the beneficial effects that:
1. The coupling distribution model selector and the characteristic learner can remarkably improve the calculation precision of the probability of water resource abundant encounter, and fully solve the problems of poor hydrologic frequency precision and insufficient consideration of climate and human activity factors in interpretation variables in a changing environment.
2. The invention fully considers the influence of climate change and human activity on hydrology, solves the problem of analyzing and explaining variable dimension disaster by the traditional GAMLSS model by a high-efficiency dimension reduction method, and improves the analysis precision and efficiency of hydrological abundant encounter under the background of a changed environment.
Drawings
FIG. 1 is a schematic flow chart of a hydrological abundant encounter probability calculation method of the coupling dimension reduction theory;
FIG. 2 is a graph of preferred results for a model selector model of the present invention;
FIG. 3 is a feature learner result diagram of the present invention, wherein (a) is a result diagram of a large gold station, (b) is a result diagram of a Yajiang station, and (c) is a result diagram of a Lanzhou station;
FIG. 4 is a set of site-specific effective climate-influencing factors and a set of site-specific effective human activity-influencing factors for an embodiment of the present invention, wherein (a) is the set of large-scale climate-influencing factors, (b) is the set of Yajiang-scale climate-influencing factors, (c) is the set of state-scale effective climate-influencing factors, (d) is the set of large-scale human activity-influencing factors, (e) is the set of Yajiang-scale effective human activity-influencing factors, and (f) is the set of state-scale effective human activity-influencing factors;
FIG. 5 is a graph showing calculated results of the probability of a large gold-Lanzhou blight encounter according to an embodiment of the present invention, wherein (a) - (d) in FIG. 5 are the results of comparative test 1 through comparative test 4, respectively, (e) is the result of the present method, and (f) is the actual values between 1956 and 2016;
FIG. 6 is a graph showing calculated results of the probability of a full-fleeing encounter in Lanzhou according to example Yajiang of the present invention, wherein (a) - (d) in FIG. 6 are the results of comparative tests 1-4, respectively, (e) is the result of the present method, and (f) is the actual values between 1956-2016.
Detailed Description
The technical method of the present invention will be further described with reference to the accompanying drawings and specific examples.
Examples: taking the probability of encountering runoff abundant in two groups of hydrologic stations in a northwest water diversion project planning water supply and receiving area as an example, one group is a Dain station (water supply area) -a Lanzhou station (water receiving area), and the other group is a Yajiang station (water supply area) -a Lanzhou station (water receiving area).
As shown in fig. 1, the method for calculating the hydrological abundant encounter probability according to the coupling dimension reduction theory of the embodiment includes the following steps:
Step S1, hydrologic data collect month-by-month flow sequences of the Yangtze river Dajin station, the Yangtze river Yajiang station and the Yangtze river Lanzhou station 1956-2016 for 61 years, and the climatic data collect climatic sequences of corresponding control basin corresponding time provided by a China climatic data service center, wherein the climatic sequences comprise a plurality of climatic indexes such as Precipitation (PRE), average air temperature (TEM-AVE), maximum air temperature (TEM-MAX), minimum air temperature (TEM-MIN), relative Humidity (RHU), evapotranspiration (PET), air pressure data (PRS), sunshine hours (SSD), wind speed (WIN) and the like. Human activity indexes collect Population Density (PD), average person GDP (GDP-AVE), average person water consumption (WC-AVE), construction area (CLA), forest area (FLA), effective Irrigation Area (EIA), reservoir Water Storage Capacity (RWSC), living water supply total quantity (TDWS), production water supply total quantity (TPWS) and the like in the control area range of each site. On this basis, all indexes are divided into low-medium-high value areas, wherein-1 indicates that the index is positioned in the low value area, -2 indicates that the index is positioned in the median area, and-3 indicates that the index is positioned in the high value area.
And performing missing value test, consistency test and white noise treatment on the index data. And performing interpolation extension on the data in the case of missing data. And if only the individual data is missing, the interpolation is carried out by adopting an adjacent time average value taking method, and when the missing data is more, the interpolation is carried out by adopting a hydrological comparison method and an SDSM statistical downscaling model.
In step S2, table 1, table 2, and table 3 show the evaluation results of the fitting accuracy of the distribution model by each single evaluation index in the distribution model selector, wherein the smaller the values of Global deviance, AIC, and SBC, the better the model is, the closer the Mean, skewness value is to the 0 model, the better the Variance, filliben value is, the closer the 1 model is, the better the Kurtosis value is to the 3 model is, but the best model is hardly preferred from the results of tables 1 to 3. Therefore, the invention further adopts a TOPSIS comprehensive evaluation method to optimize the model in the model selector, and the result is shown in figure 2. Fig. 2 shows the preferred results of the frequency distribution models of the large-scale gold station runoff, the Yajiang-scale runoff and the lan-scale gold station runoff, wherein (a) is the calculation result of the comprehensive evaluation index C value, the model performance corresponding to the larger C value is better, (b) the model is assigned based on the order of the C values of the model, the score ranges from 1 to 15, the maximum C value is assigned 15, the second maximum C value is assigned 14, and the minimum C value is assigned 1 by the score, wherein DJ represents the large-scale gold station, YJ represents the Yajiang station and LZ represents the lan-scale gold station. The optimal frequency distribution model of the large gold station runoff is selected to be NO through the distribution model selector, the optimal frequency distribution model of the Yajiang station runoff is selected to be GA, and the optimal frequency distribution model of the Lanzhou station is selected to be GG.
TABLE 1 Dajingzhan single evaluation factor
TABLE 2 Yajiang Single evaluation factors at station
TABLE 3 Lanzhou station single evaluation factor
And S3, after optimizing the optimal distribution models of the Dajin station, the Yajiang station and the Lanzhou station, extracting a cumulative distribution function CDF (i) obtained by calculating each optimal distribution model, and forming a learning pair (PE (i) and a CDF (i)) by the cumulative distribution function CDF (i) and the cumulative empirical frequency PE (i). In the feature learner, five feature functions are provided to perform data fitting on the relationship between the CDF and the PE, and a functional form of the final F (PE (i), CDF (i)) is selected according to the learning effect. The cumulative distribution function CDF and the cumulative empirical frequency PE calculated by the optimal model of each station are shown in table 4, and only a part of the data is shown in table 4 because the data amount is relatively large.
TABLE 4 optimal CDF and PE for each site
Fig. 3 shows the best fit results of the cumulative distribution function CDF and cumulative empirical frequency PE for the large gold station, yajiang station, and the lan station. The learning effect of each site is better, and the fitting precision reaches more than 0.99. The optimal relation function of the Dajin station and the Yajiang station is a polynomial function, and the CDF and PE are respectively x and y, and the form is respectively: y= -0.03748x 2 +1.01683x+0.013 and y= 0.1118x 2 +0.85316x+0.0442. The Lanzhou station optimal relationship function is a unitary linear function, and has the following form: y= 0.9801x-0.0653.
And S4, after the relation function of the CDF and the PE is obtained, loading the climate data and the human activity index data preprocessed in the step S1 into a high-efficiency dimension-reducing device for dimension reduction processing to obtain a comprehensive climate change index KC and a comprehensive human activity index KH.
First, precipitation (PRE), average air temperature (TEM-AVE), maximum air temperature (TEM-MAX), minimum air temperature (TEM-MIN), relative Humidity (RHU), evapotranspiration (PET), barometric pressure data (PRS), solar hours (SSD), and wind speed (WIN) at the location of each hydrologic site are used as candidate climate indicators. Population Density (PD), average-human GDP (GDP-AVE), average-human water consumption (WC-AVE), construction Land Area (CLA), forest Land Area (FLA), effective Irrigation Area (EIA), reservoir Water Storage Capacity (RWSC), total Domestic Water Supply (TDWS) and Total Production Water Supply (TPWS) in the control area range of each hydrologic site are used as alternative human activity indexes.
And then, adopting an Apriori data mining algorithm to screen the collected candidate climate and human activity indexes, removing factors which have insignificant influence on runoff change, so as to obtain an effective climate influence factor set and an effective human activity influence factor set, wherein the result is shown in figure 4. Taking a big gold station as an example, the effective climate influence factors are as follows: precipitation (PRE), average air temperature (TEM-AVE), maximum air temperature (TEM-MAX), relative Humidity (RHU), evapotranspiration (PET), barometric pressure data (PRS) and solar hours (SSD), while minimum air temperature (TEM-MIN) and wind speed (WIN) have been eliminated; the effective human activity influencing factors are as follows: population Density (PD), person average GDP (GDP-AVE), forest Land Area (FLA), effective Irrigation Area (EIA), reservoir Water Storage (RWSC) and Total Domestic Water Supply (TDWS), while Construction Land Area (CLA) and Total Production Water Supply (TPWS) have been eliminated. On this basis, a principal component analysis method was used to construct a climate change index KC and a human activity index KH, and then KC and KH were used as final explanatory variables as shown in tables 5 and 6.
Table 5 comprehensive climate index calculation results for each site (five years before and after)
Table 6 comprehensive human Activity index calculation results for each site (five years before and after)
And S5, analyzing a cumulative distribution function CDF_CHO (i) of each edge distribution after the influence of the changing environment is considered by adopting an optimal distribution model, and substituting the CDF_CHO (i) into a corresponding relation function F (PE (i) and CDF (i)) of each research area to obtain a final cumulative distribution function CDF_CH (i) of the edge distribution of the hydroclimate variable of each research area. And finally, calculating to obtain the withered encounter probability by adopting CVINE-Copula function method based on the obtained CDF_CH (i).
In the comparative analysis, 4 comparative tests were set up in total.
Comparative test 1: the frequency distribution model selects a more mature P-III distribution model, and the analysis of the abundant encounter is directly carried out after the cumulative distribution function of each station is calculated.
Comparative test 2: the model selector is only used for optimizing the frequency distribution model, and then the cumulative distribution function is calculated based on the optimized model for carrying out the analysis of the probability of the abundant encounters.
Comparative test 3: the model selector and the feature learner only use the model selector and the feature learner to learn the optimization and the regular features of the frequency distribution model, and then analyze the abundant encounter based on the accumulated distribution function after learning and correction.
Comparative test 4: the model selector and the feature learner of the invention are used for learning the optimization and the regular features of the frequency distribution model, and taking the influence of the hydrologic inconsistency into consideration, only the time T is used as an explanatory variable, and then the analysis of the abundant encounter is carried out based on the obtained cumulative distribution function.
Fig. 5 is a probability result of a combination of 9 kinds of abundant run-off encounters from a large gold station (water supply area) -a lan station (water receiving area). Figure 6 is a plot of the probability results for a combination of Yajiang station (water supply area) -lan station (water receiving area) runoffs 9 abundant encounters. Table 7 shows the relative error of the combined probability results for 9 kinds of run-off from Dajin station (water supply area) -Lanzhou station (water receiving area) and Table 8 shows the relative error of the combined probability results for 9 kinds of run-off from Yajiang stations (water supply area) -Lanzhou station (water receiving area).
TABLE 7 relative error case of combined probability results for 9 kinds of run-off from Dajin station (Water supply area) -Lanzhou station (Water receiving area)
TABLE 8 Yajiang station (Water supply section) -Lanzhou station (Water receiving section) runoff 9 kinds of abundant encounter combined probability results relative error case
From tables 7 and 8, it can be seen that, in the 9 combinations, the comparison test 1 has no minimum deviation, and the calculation of the abundant encounter result by directly selecting the P-III distribution model is not ideal; among the 9 kinds of abundant encounter combinations of runoffs from the Dajin station (water supply area) -the Lanzhou station (water receiving area), the comparison test 3 combination is superior to the comparison test 4, so that the model selector and the feature learner of the invention are used for learning the optimization and regular features of the frequency distribution model, which is helpful for improving the calculation accuracy, but the time T is not suitable to be used as an explanatory variable; the combination of the 9 kinds of abundant run-off encounters of runoffs from Yajiang stations (water supply area) to Lanzhou stations (water receiving area) is superior to the combination of the comparative test 4, so that the runoffs from Yajiang stations (water supply area) to Lanzhou stations (water receiving area) have the non-uniformity characteristic of time, and the time T is used as an explanatory variable on the basis of the study of the optimization and the regular characteristic of the frequency distribution model by using the model selector and the characteristic learner, thereby being beneficial to improving the calculation precision; the results of the present invention show the greatest advantage in either large-gold (water supply) station-lan (water receiving) station runoff or Yajiang (water supply) station-lan (water receiving) station runoff, with 6 combinations being optimal in the large-gold (water supply) station-lan (water receiving) station runoff 9 wither encounter combinations and 5 combinations being optimal in the Yajiang (water supply) station-lan (water receiving) station runoff 9 wither encounter combinations. In conclusion, the method fully considers the influence of climate change and human activities on hydrology, and can improve the calculation accuracy of the probability of water resource abundant meeting.
The above description is only an example of the present application and is not intended to limit the scope of the present application, and various modifications and variations will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.
Claims (8)
1. A hydrological abundant encounter probability calculation method based on a coupling dimension reduction theory is characterized by comprising the following steps:
Step S1: acquiring climate index data, hydrological index data and human activity index data of a research area, and preprocessing the data;
Step S2: based on hydrological index data, selecting an optimal distribution model of the edge distribution of the research area by adopting a distribution model selector, wherein the distribution model selector comprises a plurality of frequency distribution models;
step S3: the cumulative distribution function and the cumulative experience frequency obtained by calculating the optimal distribution model are loaded into a feature learner, and a relation function of the cumulative distribution function and the cumulative experience frequency is obtained through the feature learner;
Step S4: loading the climate index and the human activity index processed in the step S1 into a high-efficiency dimension reduction device for dimension reduction processing to obtain a comprehensive climate index KC and a comprehensive human activity index KH after dimension reduction;
Step S5: analyzing the cumulative distribution function of each edge distribution after considering the influence of the changing environment by adopting the optimal distribution model of the step S2 by taking the comprehensive climate index KC and the comprehensive human activity index KH of the step S4 as final explanatory variables, and substituting the cumulative distribution function after considering the influence of the changing environment into the corresponding relation function of the step S3 to obtain the final cumulative distribution function of the edge distribution of the hydrologic climate variable; and then calculating the probability of the abundant-blight encounter by adopting a Copula function method.
2. The method for calculating the probability of hydrological abundant encounter according to the coupled dimension reduction theory of claim 1, wherein the feature learner in step S3 includes a plurality of feature functions for learning the relationship between the cumulative distribution function and the cumulative experience frequency, and selects a final feature function form according to the learning effect.
3. The method for calculating the probability of hydrological abundant encounter according to the coupling dimension reduction theory of claim 2, wherein in the step S3, the feature learner includes five feature functions, and if the cumulative distribution function and the cumulative experience frequency are x and y, respectively, the five function forms are:
Exponential function: y=ae bx
Linear function: y=ax+b
Logarithmic function: y= alnx +b
Polynomial function, for example quadratic function, y=ax 2 +bx+c
Power function: y=ax b
Wherein the parameters a, b, c are determined by artificial intelligence optimization algorithms including, but not limited to, genetic algorithms, particle swarm algorithms, or neural networks.
4. The method for calculating the hydrological abundant encounter probability by coupling dimension reduction theory according to claim 1, wherein the step S4 specifically comprises:
step S4.1: adopting an Apriori data mining algorithm to screen the climate index data and the human activity index data preprocessed in the step S1 to obtain an effective climate influence factor set and an effective human activity influence factor set;
step S4.2: analyzing the importance of each influence element in an effective climate influence factor set and an effective human activity influence factor set by adopting a random forest model, and sequencing according to the importance;
Step S4.3: and constructing a comprehensive climate index and a comprehensive human activity index by adopting a principal component analysis method of the coupling information entropy.
5. The method for calculating the probability of hydrologic abundant encounter according to the coupling dimension reduction theory of claim 1, wherein in the step S5, two hydrologic time sequences are set to be X and Y respectively, and the edge distribution functions thereof are u and v respectively, then Y pf is the water amount corresponding to the frequency P f in the hydrologic time sequence Y, Y pk is the water amount corresponding to the frequency P k in the hydrologic time sequence Y, X pf is the water amount corresponding to the frequency P f in the hydrologic time sequence X, and X pk is the water amount corresponding to the frequency P k in the hydrologic time sequence X;
Taking the hydrologic time sequence X as an example, X t≥Xpf is high water, X t≤Xpk is dead water, X pk＜Xt＜Xpf is flat water, wherein X t is the hydrologic quantity of the t year, then: the probability of X and Y homoabundant is: ; the probability of X and Y being leveled is:/> ; The probability of the same withering of X and Y is:/>; The probability of X-Feng-Y-plane is: ; the probability of X flat Y is: ; the probability of X to Y withering is: ; the probability of X withered Y is: ; the probability of X flat Y withered is: ; the probability of X dead Y flat is: Where u pf、vpf、upk、 vpk is the edge distribution function value corresponding to X pf、Ypf、Xpk、Ypk, and C () is the Copula connection function.
6. The method for calculating the hydrological withered encounter probability based on the coupled dimension reduction theory according to claim 1, wherein the data preprocessing in step S1 comprises:
Performing missing value inspection, consistency inspection and white noise treatment on climate index data, hydrological index data and human activity index data; and (3) carrying out interpolation extension on the data aiming at the condition that missing data exists, if only individual data is missing, adopting an adjacent time averaging method to carry out interpolation, and when the missing data is more, adopting a hydrological comparison method and an SDSM statistical downscaling model to carry out interpolation.
7. The method for calculating the hydrological abundant encounter probability based on the coupled dimension reduction theory according to claim 1, wherein the frequency distribution model in step S2 comprises: 10 two-parameter models and 5 three-parameter models; the two parameter models include GA, GU, IG, LO, LOGNO, NO, RG, WEI, WEI and WEI3; the three parameter model includes GG, BCCG, GIG, PE and P-III.
8. The method for calculating hydrological withered encounter probability according to the coupling dimension reduction theory of claim 7, wherein in the step S2, 15 hydrological frequency distribution models are adopted to perform univariate frequency distribution fitting on a runoff sequence, and based on the result of 15 frequency distribution fitting, 8 evaluation indexes including Global deviance, AIC, SBC, residual Mean value Mean, residual Variance index Variance, residual Filliben coefficient, residual bias coefficient Skewness and residual Kurtosis coefficient are adopted to perform precision evaluation on each frequency distribution model; if the conclusion of all the evaluation indexes is consistent, the optimal frequency distribution model is optimized; if the conclusion of all the evaluation indexes is inconsistent, the final optimization of the model result is carried out by adopting a TOPSIS comprehensive evaluation method based on entropy weight.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410524576.1A CN118095973A (en) | 2024-04-29 | 2024-04-29 | Hydrological abundant encounter probability calculation method based on coupling dimension reduction theory |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410524576.1A CN118095973A (en) | 2024-04-29 | 2024-04-29 | Hydrological abundant encounter probability calculation method based on coupling dimension reduction theory |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118095973A true CN118095973A (en) | 2024-05-28 |
Family
ID=91144437
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410524576.1A Pending CN118095973A (en) | 2024-04-29 | 2024-04-29 | Hydrological abundant encounter probability calculation method based on coupling dimension reduction theory |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118095973A (en) |
Citations (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006030013A (en) * | 2004-07-16 | 2006-02-02 | Foundation Of River & Basin Integrated Communications Japan | Nationwide combined radar rainfall information providing system |
| CN105825308A (en) * | 2016-03-24 | 2016-08-03 | 辽宁省水利水电勘测设计研究院 | Reservoir connectivity analysis method based on multi-target analysis |
| CN107423546A (en) * | 2017-04-18 | 2017-12-01 | 武汉大学 | Multivariable hydrological uncertainty processing method based on Copula functions |
| CN110377989A (en) * | 2019-07-08 | 2019-10-25 | 武汉大学 | Two Variational Design flood calculation method of nonuniformity based on hydrothermal reaction coupling balance |
| CN111797129A (en) * | 2020-06-01 | 2020-10-20 | 武汉大学 | Hydrologic drought assessment method under climate change situation |
| CN111898258A (en) * | 2020-07-17 | 2020-11-06 | 武汉大学 | Two-dimensional drought disaster assessment method driven by hydrologic cycle variation |
| CN112819293A (en) * | 2021-01-14 | 2021-05-18 | 中国长江三峡集团有限公司 | Failure early warning analysis method for water reservoir scheduling rule under influence of climate change |
| CN113033014A (en) * | 2021-04-09 | 2021-06-25 | 北京师范大学 | Regional available water supply estimation method considering multi-water-source joint probability distribution |
| CN113052455A (en) * | 2021-03-22 | 2021-06-29 | 广西电网有限责任公司 | Method and device for fine evaluation of watershed runoff hydrological conditions |
| CN113705091A (en) * | 2021-08-23 | 2021-11-26 | 武汉大学 | Non-uniform design flood calculation method considering change environment and scheduling influence |
| CN113887972A (en) * | 2021-10-09 | 2022-01-04 | 水利部牧区水利科学研究所 | Comprehensive drought monitoring and evaluating method based on hydrological process |
| CN115507822A (en) * | 2022-06-09 | 2022-12-23 | 武汉大学 | Flood risk prediction method under hydrographic cyclic variation drive |
| CN115860421A (en) * | 2022-12-26 | 2023-03-28 | 长江水利委员会水文局 | Ecological flow dynamic calculation method and system adaptive to target guarantee rate |
| CN115983132A (en) * | 2023-02-03 | 2023-04-18 | 广西大学 | Response research method of runoff of drainage basin to future climate and land utilization changes |
| CN116452003A (en) * | 2023-05-08 | 2023-07-18 | 郑州大学 | Water right transaction scheme optimization method based on water shortage risk assessment |
| CN116451879A (en) * | 2023-06-16 | 2023-07-18 | 武汉大学 | Drought risk prediction method and system and electronic equipment |
| CN116663719A (en) * | 2023-05-23 | 2023-08-29 | 武汉大学 | Flood risk prediction method, system and electronic equipment |
| CN116756481A (en) * | 2023-07-03 | 2023-09-15 | 中国科学院地理科学与资源研究所 | Watershed future hydrologic drought probability analysis method based on time-varying gain model and Copula |
| CN117035201A (en) * | 2023-10-09 | 2023-11-10 | 南京大学 | Multi-target scheduling rule making method and system for plain river network water engineering cluster |
| CN117057490A (en) * | 2023-10-12 | 2023-11-14 | 武汉大学 | Prediction method and system for wet stress heat wave-flood composite disaster and electronic equipment |
| CN117116382A (en) * | 2023-10-18 | 2023-11-24 | 水利部交通运输部国家能源局南京水利科学研究院 | Water quality space-time prediction method and system for water-bearing lake under influence of diversion engineering |
| CN117648863A (en) * | 2023-12-04 | 2024-03-05 | 西安理工大学 | Regional water supply capacity acquisition method under special drought conditions based on water system communication |
| CN117744501A (en) * | 2024-01-19 | 2024-03-22 | 水利部水利水电规划设计总院 | Water network system regulation node optimal scheduling and decision-making method considering ecological flow |
-
2024
- 2024-04-29 CN CN202410524576.1A patent/CN118095973A/en active Pending
Patent Citations (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006030013A (en) * | 2004-07-16 | 2006-02-02 | Foundation Of River & Basin Integrated Communications Japan | Nationwide combined radar rainfall information providing system |
| CN105825308A (en) * | 2016-03-24 | 2016-08-03 | 辽宁省水利水电勘测设计研究院 | Reservoir connectivity analysis method based on multi-target analysis |
| CN107423546A (en) * | 2017-04-18 | 2017-12-01 | 武汉大学 | Multivariable hydrological uncertainty processing method based on Copula functions |
| CN110377989A (en) * | 2019-07-08 | 2019-10-25 | 武汉大学 | Two Variational Design flood calculation method of nonuniformity based on hydrothermal reaction coupling balance |
| CN111797129A (en) * | 2020-06-01 | 2020-10-20 | 武汉大学 | Hydrologic drought assessment method under climate change situation |
| CN111898258A (en) * | 2020-07-17 | 2020-11-06 | 武汉大学 | Two-dimensional drought disaster assessment method driven by hydrologic cycle variation |
| CN112819293A (en) * | 2021-01-14 | 2021-05-18 | 中国长江三峡集团有限公司 | Failure early warning analysis method for water reservoir scheduling rule under influence of climate change |
| WO2022135265A1 (en) * | 2021-01-14 | 2022-06-30 | 中国长江三峡集团有限公司 | Failure warning and analysis method for reservoir dispatching rules under effects of climate change |
| CN113052455A (en) * | 2021-03-22 | 2021-06-29 | 广西电网有限责任公司 | Method and device for fine evaluation of watershed runoff hydrological conditions |
| CN113033014A (en) * | 2021-04-09 | 2021-06-25 | 北京师范大学 | Regional available water supply estimation method considering multi-water-source joint probability distribution |
| CN113705091A (en) * | 2021-08-23 | 2021-11-26 | 武汉大学 | Non-uniform design flood calculation method considering change environment and scheduling influence |
| CN113887972A (en) * | 2021-10-09 | 2022-01-04 | 水利部牧区水利科学研究所 | Comprehensive drought monitoring and evaluating method based on hydrological process |
| CN115507822A (en) * | 2022-06-09 | 2022-12-23 | 武汉大学 | Flood risk prediction method under hydrographic cyclic variation drive |
| CN115860421A (en) * | 2022-12-26 | 2023-03-28 | 长江水利委员会水文局 | Ecological flow dynamic calculation method and system adaptive to target guarantee rate |
| CN115983132A (en) * | 2023-02-03 | 2023-04-18 | 广西大学 | Response research method of runoff of drainage basin to future climate and land utilization changes |
| CN116452003A (en) * | 2023-05-08 | 2023-07-18 | 郑州大学 | Water right transaction scheme optimization method based on water shortage risk assessment |
| CN116663719A (en) * | 2023-05-23 | 2023-08-29 | 武汉大学 | Flood risk prediction method, system and electronic equipment |
| CN116451879A (en) * | 2023-06-16 | 2023-07-18 | 武汉大学 | Drought risk prediction method and system and electronic equipment |
| CN116756481A (en) * | 2023-07-03 | 2023-09-15 | 中国科学院地理科学与资源研究所 | Watershed future hydrologic drought probability analysis method based on time-varying gain model and Copula |
| CN117035201A (en) * | 2023-10-09 | 2023-11-10 | 南京大学 | Multi-target scheduling rule making method and system for plain river network water engineering cluster |
| CN117057490A (en) * | 2023-10-12 | 2023-11-14 | 武汉大学 | Prediction method and system for wet stress heat wave-flood composite disaster and electronic equipment |
| CN117116382A (en) * | 2023-10-18 | 2023-11-24 | 水利部交通运输部国家能源局南京水利科学研究院 | Water quality space-time prediction method and system for water-bearing lake under influence of diversion engineering |
| CN117648863A (en) * | 2023-12-04 | 2024-03-05 | 西安理工大学 | Regional water supply capacity acquisition method under special drought conditions based on water system communication |
| CN117744501A (en) * | 2024-01-19 | 2024-03-22 | 水利部水利水电规划设计总院 | Water network system regulation node optimal scheduling and decision-making method considering ecological flow |
Non-Patent Citations (1)
| Title |
|---|
| 戴昌军;管光明;孙浩;李翰卿;: "南水北调东线工程优势调水频率分析", 南水北调与水利科技, no. 01, 15 February 2008 (2008-02-15) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111639237B (en) | Electric power communication network risk assessment system based on clustering and association rule mining | |
| CN112749904A (en) | Power distribution network fault risk early warning method and system based on deep learning | |
| CN115829162B (en) | Crop yield prediction method, device, electronic equipment and medium | |
| Khan et al. | Evaluating the performance of several data mining methods for predicting irrigation water requirement | |
| CN111915083B (en) | Wind power prediction method and prediction system based on time layered combination | |
| CN109858180B (en) | Regional crop water consumption space pattern optimization design method | |
| CN116681279B (en) | Forestry pest control risk assessment system based on data analysis | |
| CN117236199B (en) | Method and system for improving water quality and guaranteeing water safety of river and lake in urban water network area | |
| CN117787658B (en) | Water resource scheduling system based on irrigation district river flow abnormal condition | |
| CN114580762A (en) | Hydrological forecast error correction method based on XGboost | |
| CN113052386A (en) | Distributed photovoltaic daily generated energy prediction method and device based on random forest algorithm | |
| CN116562583A (en) | Multidimensional water resource supply and demand prediction method and system | |
| CN114169434A (en) | Load prediction method | |
| CN114414491B (en) | Grass ecology dynamic monitoring and analysis system | |
| CN114723142A (en) | Multi-target land utilization simulation system and method based on non-dominated sorting genetic algorithm and FLUS model | |
| CN113379116A (en) | Cluster and convolutional neural network-based line loss prediction method for transformer area | |
| CN118095973A (en) | Hydrological abundant encounter probability calculation method based on coupling dimension reduction theory | |
| WO2023245399A1 (en) | Rice production potential simulation method based on land system and climate change coupling | |
| CN114091782B (en) | Medium-long term power load prediction method | |
| CN116307038A (en) | Irrigation water consumption prediction method, irrigation water consumption prediction system and computer storage medium | |
| CN115358587A (en) | Regional multi-department collaborative infrastructure planning method and system | |
| Gupta et al. | Markov Chain Model Used in Agricultural Yield Predictions Utilizing on Indian Agriculture | |
| CN115496128A (en) | Urban waterlogging risk forecasting method based on raininess-raininess characteristic parameter combined distribution | |
| Khan et al. | Irrigation water requirement prediction through various data mining techniques applied on a carefully pre-processed dataset | |
| CN107977727A (en) | A kind of method that probability is blocked based on social development and climatic factor prediction cable network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination |