CN109919362B - Medium-and-long-term runoff forecasting method considering hydraulic engineering scheduling influence - Google Patents
Medium-and-long-term runoff forecasting method considering hydraulic engineering scheduling influence Download PDFInfo
- Publication number
- CN109919362B CN109919362B CN201910114775.4A CN201910114775A CN109919362B CN 109919362 B CN109919362 B CN 109919362B CN 201910114775 A CN201910114775 A CN 201910114775A CN 109919362 B CN109919362 B CN 109919362B
- Authority
- CN
- China
- Prior art keywords
- forecasting
- hydraulic engineering
- runoff
- value
- particle
- 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.)
- Expired - Fee Related
Links
- 238000013277 forecasting method Methods 0.000 title claims abstract description 11
- 238000000034 method Methods 0.000 claims abstract description 34
- 238000005457 optimization Methods 0.000 claims abstract description 16
- 230000007774 longterm Effects 0.000 claims abstract description 14
- 239000002245 particle Substances 0.000 claims description 59
- 238000012549 training Methods 0.000 claims description 27
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 21
- 238000012360 testing method Methods 0.000 claims description 19
- 238000011156 evaluation Methods 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000010219 correlation analysis Methods 0.000 claims description 5
- 239000000126 substance Substances 0.000 claims description 5
- 238000011144 upstream manufacturing Methods 0.000 claims description 4
- 150000001875 compounds Chemical class 0.000 claims description 3
- 238000007689 inspection Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 230000002265 prevention Effects 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 3
- 238000012216 screening Methods 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a medium-and-long-term runoff forecasting method considering the dispatching influence of hydraulic engineering, and relates to the technical field of hydrological forecasting. The method considers the influence of medium and large-scale hydraulic engineering scheduling on medium and long-term runoff forecasting, improves the forecasting precision and the practicability of the medium and long-term runoff forecasting compared with the conventional common method, has the advantages of simple compiling, less parameter setting and strong global optimization capability, and can effectively avoid the problems of low forecasting precision and the like caused by large calculated amount, long time consumption, large required sample amount, easy falling of a forecasting result into local optimization. Can be used as an effective method for forecasting long-term runoff in a specific drainage basin.
Description
Technical Field
The invention relates to the technical field of hydrologic forecasting, in particular to a medium-long term runoff forecasting method considering the dispatching influence of hydraulic engineering.
Background
The long-term runoff forecasting process in the drainage basin has certain space-time uncertainty due to the comprehensive influence of various factors such as climate, weather, underlying surface, human activities and the like.
Traditional medium-and-long-term runoff forecasting methods such as a physical cause method, a mathematical statistic method and a regression analysis method and modern emerging forecasting methods such as a fuzzy analysis method, a grey system method, a neural network and the like usually use meteorological factors as alternative factors, select a suitable forecasting factor set for a specific basin from the meteorological factors, and construct a relation between the forecasting factor set and runoff of the basin.
However, in some specific watersheds, the influence of human activities on the runoff of the watersheds is large, and particularly the medium-long term runoff of the watersheds is directly influenced by the dispatching of medium-large hydraulic engineering. The existing method mostly takes meteorological factors as forecasting factors, mostly considers weather and how the climatic factors influence medium and long term runoff from a natural level, and has less consideration on how human activities influence the medium and long term runoff.
Disclosure of Invention
The invention aims to provide a medium-and-long-term runoff forecasting method considering the dispatching influence of hydraulic engineering, so that the problems in the prior art are solved.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
a medium-long term runoff forecasting method considering hydraulic engineering scheduling influence comprises the following steps:
s1, screening weather forecast factors;
s2, selecting medium and large hydraulic engineering which forecasts main flow and branch flow at the upstream of the section and has certain influence on runoff, and calculating hydraulic engineering influence factors according to the following steps:
s201, calculating the full storage rate of each hydraulic engineering according to the following formula:
p＝wl/WL
wherein the content of the first and second substances,
wl represents the monthly mean of the actual water levels of the selected months of the selected reservoir;
WL represents the characteristic water level of the reservoir project, for the water project with flood prevention task in flood season, the WL value of the flood season is the flood limit water level of the water project, and the WL value of the non-flood season is the normal water storage level; for reservoir engineering without flood control tasks in flood seasons, WL values of the flood seasons and WL values of the flood seasons are normal water storage levels;
p represents the full storage rate of the hydraulic engineering, and the value range is [0-1 ];
s202, calculating the full storage rate of all hydraulic engineering of the basin according to the following formula:
wherein the content of the first and second substances,
piis the full rate, RC, of hydraulic engineering iiRepresenting the total storage capacity of the hydraulic engineering i, wherein TRC is the total storage capacity of all the hydraulic engineering, and P is the full storage rate of all the hydraulic engineering in a drainage basin and is a hydraulic engineering influence factor;
s3, normalizing the weather forecast factor and the hydraulic engineering influence factor to obtain a normalized value of the forecast factor;
s4, selecting runoff data of a basin to be forecasted in S years as a historical sample, taking the runoff data of the previous N years in the historical sample and the normalization value of the forecasting factor of the selected corresponding time as a training set, and taking the runoff data of the next M years and the normalization value of the forecasting factor of the selected corresponding time as a test set, wherein S is N + M, N is more than M, and S, N, M is positive integers;
s5, determining the optimal values of the error punishment parameter C, the kernel parameter sigma and the insensitive loss coefficient epsilon in the SVR model by adopting a particle swarm optimization algorithm, inputting the optimal values into the SVR model, training the SVR model by utilizing a training set, checking by utilizing a test set, obtaining a forecasting model if the optimal values meet the requirements, or reselecting part of forecasting factors, and then training the model by utilizing a new training test set until the model reaches the standard;
and S6, forecasting the runoff to be forecasted by using the forecasting model.
Preferably, S1 includes the steps of:
s101, performing correlation analysis on the historical multinomial circulation index data and the historical runoff data of the basin to be forecasted according to a formula to obtain correlation coefficients between annual runoff y and forecasting factors x:
in the formula (I), the compound is shown in the specification,
ρia correlation coefficient representing the ith factor and the annual runoff quantity y,
represents the average value of the factor over a number of years,
the mean value of the run-off of the sample is shown,
l is the runoff sample year;
s102, selecting a circulation index with a large correlation coefficient and a certain physical significance as a factor to serve as a weather forecasting factor.
Preferably, in S3, the normalization processing is specifically performed according to the following formula:
wherein, ytAVG, STD represent the value of a predictor at a time within a selected time period, the average value of the values of the predictors selected within the selected time period, and the standard deviation of the values of the predictors selected within the selected time period, respectively.
Preferably, in S4, the determining the optimal values of the error penalty parameter C, the kernel parameter σ, and the insensitive loss coefficient ∈ in the SVR model by using the particle swarm optimization algorithm includes the following steps:
s401, determining the value ranges of the three parameters of C, sigma and epsilon;
s402, initializing a particle swarm, namely setting the scale, the iteration times, the position and the speed of the particle swarm;
s403, determining a fitness evaluation function, and calculating a fitness value of each particle by using the fitness function;
s404, comparing the fitness of the current position of each particle with the fitness of the historical best position of each particle, and determining the current best position
S405, comparing the fitness of the current optimal position of each particle with the fitness of the current optimal position of the whole group to determine the current optimal position;
s406, updating the particle speed and the particle position;
s407, judging whether an end condition is met, if not, turning to S403; if the global optimal solution is satisfied, outputting a global optimal solution, wherein the global optimal solution at the moment is the optimal solution of three parameters, namely C, sigma and epsilon; the end condition is that the number of iterations reaches an upper limit.
Preferably, S403 specifically selects the certainty coefficient as a fitness evaluation function, and calculates the fitness value of each particle using the following formula as an evaluation function:
where DC is the deterministic coefficient, yc(i) To predict value, yo(i) In order to be the actual value of the measurement,
is the average value of measured values of years, and n is the number of years of the training set sample.
Preferably, S406 specifically is:
the particle velocity is updated according to the following equation:
vi+1＝wvi+c1r1(localbest-xi)
+c2r2(globalbest-xi)
the particle position is updated according to the following formula:
xi+1＝xi+vi+1
wherein i represents the number of iterations, xiBits representing particles at the ith iterationV. position ofiRepresenting the velocity of the particle at the i-th iteration, r1，r2Two random numbers, c, expressed between (0, 1)1c2The expression is a speed-increasing factor, the value of which is generally 2, w is a dynamic weight factor, and the value range of the factor is [0.4, 0.9 ]]。
Preferably, the dynamic weight factor w is dynamically updated according to the following formula:
wherein, winiAnd wendThe initial value and the final value of the dynamic weight factor are respectively 0.9 and 0.4, G is the iteration number of the PSO algorithm, and G is the current iteration number.
Preferably, in S4, the testing is performed by using the test set, and if the requirements are satisfied, the forecasting model is obtained, where the requirements to be satisfied are: and evaluating the model according to the pitch-average consistency rate, wherein the pitch-average consistency rate of the forecast model is required to be more than 60%. Wherein, the calculation formula of the pitch coincidence rate is as follows:
R＝Nsame/N×100％
wherein R is the rate of consistency from the distance, N is the total years participating in the verification, and N issameThe years that the signs of the measured runoff and the predicted runoff pitch are the same are shown.
The invention has the beneficial effects that: the method for forecasting the medium-and-long-term runoff considering the influence of the hydraulic engineering scheduling considers the influence of medium-and-large-scale hydraulic engineering scheduling on the medium-and-long-term runoff forecasting, improves the forecasting precision and the practicability of the medium-and-long-term runoff forecasting compared with the conventional common method, has the advantages of simple compiling, less parameter setting and strong global optimization capability, and can effectively avoid the problems of large calculation amount, long time consumption, large required sample amount, low forecasting precision caused by the fact that a forecasting result is easy to fall into local optimization and the like. Can be used as an effective method for forecasting long-term runoff in a specific drainage basin.
Drawings
FIG. 1 is a flow chart of a medium-and-long-term runoff forecasting method considering the scheduling influence of hydraulic engineering, provided by the invention;
FIG. 2 is a schematic flow chart of finding the optimal parameter values of the SVR model by the particle swarm optimization algorithm.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the invention, are intended for purposes of illustration only and are not intended to limit the scope of the invention.
As shown in fig. 1, an embodiment of the present invention provides a method for forecasting medium and long term runoff considering scheduling influence of a hydraulic engineering, including the following steps:
s1, screening weather forecast factors;
s2, selecting medium and large hydraulic engineering which forecasts main flow and branch flow at the upstream of the section and has certain influence on runoff, and calculating hydraulic engineering influence factors according to the following steps:
s201, calculating the full storage rate of each hydraulic engineering according to the following formula:
p＝wl/WL
wherein the content of the first and second substances,
wl represents the monthly mean of the actual water levels of the selected months of the selected reservoir;
WL represents the characteristic water level of the reservoir project, for the water project with flood prevention task in flood season, the WL value of the flood season is the flood limit water level of the water project, and the WL value of the non-flood season is the normal water storage level; for reservoir engineering without flood control tasks in flood seasons, WL values of the flood seasons and WL values of the flood seasons are normal water storage levels;
p represents the full storage rate of the hydraulic engineering, and the value range is [0-1 ];
s202, calculating the full storage rate of all hydraulic engineering of the basin according to the following formula:
wherein the content of the first and second substances,
pito the water conservancyFill rate, RC, of project iiRepresenting the total storage capacity of the hydraulic engineering i, wherein TRC is the total storage capacity of all the hydraulic engineering, and P is the full storage rate of all the hydraulic engineering in a drainage basin and is a hydraulic engineering influence factor;
s3, normalizing the weather forecast factor and the hydraulic engineering influence factor to obtain a normalized value of the forecast factor;
s4, selecting runoff data of a basin to be forecasted in S years as a historical sample, taking the runoff data of the previous N years in the historical sample and the normalization value of the forecasting factor of the selected corresponding time as a training set, and taking the runoff data of the next M years and the normalization value of the forecasting factor of the selected corresponding time as a test set, wherein S is N + M, N is more than M, and S, N, M is positive integers;
s5, determining the optimal values of the error punishment parameter C, the kernel parameter sigma and the insensitive loss coefficient epsilon in the SVR model by adopting a particle swarm optimization algorithm, inputting the optimal values into the SVR model, training the SVR model by utilizing a training set, checking by utilizing a test set, obtaining a forecasting model if the optimal values meet the requirements, or reselecting part of forecasting factors, and then training the model by utilizing a new training test set until the model reaches the standard;
and S6, forecasting the runoff to be forecasted by using the forecasting model.
In the method, firstly, weather forecasting factors are selected from 130 circulation factors, then hydraulic engineering influence factors are calculated, all the obtained forecasting factors are used as the input of an SVR model, and the input data is subjected to standardization processing by using a Z-Score method; selecting historical runoff data of years as historical samples, and dividing the historical samples into two types, wherein one type is a training set. One is a test set, and the training set is used for training an SVR model; when the SVR model is trained, firstly, the particle swarm optimization algorithm is adopted to determine the values of three parameters of C, sigma and epsilon in the SVR model, and the method comprises the following steps: setting the particle swarm scale, randomly setting an initial position for each particle, then calculating the individual optimal position of each particle and the global optimal fitness of the particle swarm, updating the speed and the position of each particle, continuously iterating until the expected prediction precision is reached or the iteration time upper limit is reached, and obtaining the coordinate value of the particle with the highest position, namely the optimal solution of the SVR model parameters C, sigma and epsilon; and substituting the obtained C, sigma and epsilon parameters into the SVR model, and training the model by using a training set. And after the model training is finished, determining the prediction precision of the model by using the test set, and analyzing errors. And finally, predicting the future flow value of the basin to be predicted.
In the invention, a hydraulic engineering influence factor is provided by combining a scheduling mode of a hydraulic engineering, and the influence of medium and large-scale hydraulic engineering scheduling on medium and long-term runoff forecasting is considered. Compared with the conventional method, the forecasting precision and the practicability of the medium-and-long-term forecasting method are improved, the provided method is simple to compile, few in parameter setting and strong in global optimization capability, and the problems that the forecasting precision is not high due to the fact that the calculated amount is large, the time consumption is long, the required sample amount is large, and the forecasting result is prone to falling into local optimization can be effectively avoided.
In the present invention, S1 may include the following steps:
s101, performing correlation analysis on the historical multinomial circulation index data and the historical runoff data of the basin to be forecasted according to a formula to obtain correlation coefficients between annual runoff y and forecasting factors x:
in the formula (I), the compound is shown in the specification,
ρia correlation coefficient representing the ith factor and the annual runoff quantity y,
represents the average value of the factor over a number of years,
the mean value of the run-off of the sample is shown,
l is the runoff sample year;
s102, selecting a circulation index with a large correlation coefficient and a certain physical significance as a factor to serve as a weather forecasting factor.
In S3, the normalization process is specifically performed according to the following formula:
wherein, ytAVG, STD represent the value of a predictor at a time within a selected time period, the average value of the values of the predictors selected within the selected time period, and the standard deviation of the values of the predictors selected within the selected time period, respectively.
As shown in fig. 2, in S4, the determining the optimal values of the error penalty parameter C, the kernel parameter σ, and the insensitive loss coefficient ∈ in the SVR model by using the particle swarm optimization algorithm includes the following steps:
s401, determining the value ranges of the three parameters of C, sigma and epsilon;
s402, initializing a particle swarm, namely setting the scale, the iteration times, the position and the speed of the particle swarm;
s403, determining a fitness evaluation function, and calculating a fitness value of each particle by using the fitness function;
s404, comparing the fitness of the current position of each particle with the fitness of the historical best position of each particle, and determining the current best position
S405, comparing the fitness of the current optimal position of each particle with the fitness of the current optimal position of the whole group to determine the current optimal position;
s406, updating the particle speed and the particle position;
s407, judging whether an end condition is met, if not, turning to S403; if the global optimal solution is satisfied, outputting a global optimal solution, wherein the global optimal solution at the moment is the optimal solution of three parameters, namely C, sigma and epsilon; the end condition is that the number of iterations reaches an upper limit.
S403 specifically, selecting the certainty coefficient as a fitness evaluation function, and calculating the fitness value of each particle using the following formula as an evaluation function:
where DC is the deterministic coefficient, yc(i) To predict value, yo(i) In order to be the actual value of the measurement,
is the average value of measured values of years, and n is the number of years of the training set sample.
S406 specifically includes:
the particle velocity is updated according to the following equation:
vi+1＝wvi+c1r1(localbest-xi)
+c2r2(globalbest-xi)
the particle position is updated according to the following formula:
xi+1＝xi+vi+1
wherein i represents the number of iterations, xiIndicating the position of the particle at the i-th iteration, viRepresenting the velocity of the particle at the i-th iteration, r1，r2Two random numbers, c, expressed between (0, 1)1c2The expression is a speed-increasing factor, the value of which is generally 2, w is a dynamic weight factor, and the value range of the factor is [0.4, 0.9 ]]。
Wherein, the dynamic weighting factor w can be dynamically updated according to the following formula:
wherein, winiAnd wendThe initial value and the final value of the dynamic weight factor are respectively 0.9 and 0.4, G is the iteration number of the PSO algorithm, and G is the current iteration number.
In S4, the testing set is used for performing the inspection, and if the requirements are met, a forecasting model is obtained, where the requirements to be met are: and evaluating the model according to the pitch-average consistency rate, wherein the pitch-average consistency rate of the forecast model is required to be more than 60%. The calculation formula of the pitch-average consistency rate is as follows:
R＝Nsame/N×100％
wherein R is the rate of consistency from the distance, N is the total years participating in the verification, and N issameThe years that the signs of the measured runoff and the predicted runoff pitch are the same are shown.
DETAILED DESCRIPTION OF EMBODIMENT (S) OF INVENTION
In the embodiment, the runoff of the 7-month entering reservoir of the Danjiang estuary reservoir of the Hanjiang river basin is forecasted, and the method provided by the invention is implemented as follows:
step 1, acquiring hundreds of weather system index sets (namely 88 atmospheric circulation indexes, 26 sea temperature indexes and 16 other indexes, namely 130 factors in each month) in the past year from an official website of a national weather center as historical multi-item circulation index data, and after carrying out correlation analysis on the historical multi-item circulation index data and the historical runoff data of a basin to be forecasted, sequencing the obtained correlation coefficients from large to small according to absolute values.
The month forecast selects the meteorological factors of the first half year for correlation analysis, 130 factors in each month account for 6 months, so that 780 weather factors participating in correlation calculation are displayed, because the weather factors are not all displayed at the place with excessive number, the meteorological factors with the maximum absolute values of the first 20 items are displayed,as shown in the following table:
then, the preliminarily selected forecasting factors are subjected to physical cause analysis, the meteorological indexes with large correlation coefficient and certain physical influence on the forecasting basin are selected as the meteorological forecasting factors, and meanwhile, the forecasting factors selected by a runoff forecasting model of 7 months are supplemented,as shown in the following table:
step 2, selecting upstream engineering influence factors of the Ankang reservoir, calculating 1 seat, and displaying the monthly average water level of the Ankang reservoir in 6 months per year in 2001 + 2016 and the full storage rate obtained by calculation according to the following table:
because only one reservoir is selected to calculate the engineering influence factor, the full storage rate of the healthy reservoir is the engineering influence factor of the watershed, and the method is as follows:
step 3, normalizing the values of the weather forecasting factors and the hydraulic engineering influence factors to obtain normalized values of the forecasting factors, wherein the normalized values are shown in the following table:
and 4, selecting runoff data of the drainage basin to be forecasted in 16 years (2001-2016 years) as a historical sample, taking the runoff data of the previous 11 years (2001-2011 years) and the normalized value of the forecasting factor of the selected corresponding time in the historical sample as a training set, and taking the runoff data of the next 5 years (2011-2016) and the normalized value of the forecasting factor of the selected corresponding time as a test set.
Step 5, in the implementation process, the value ranges of the three parameters C, sigma and epsilon adopt related data in the existing method, wherein the maximum value of the particle position is set to be (100, 200, 100), and the minimum value is set to be (0.01, 0.01, 0.01); the maximum value of the particle velocity modulation range is set to (10, 1, 10), and the minimum value is set to (-10, -1, -10). The particle swarm size is set to 300, the number of particle swarm iterations is set to 1000, and the position and the speed of the particle take random values as the initial position and the initial speed of the particle in the value range set in the step 5-1). The upper limit of the number of iterations is 1000. Through calculation, the optimal values of the three parameters of C, sigma and epsilon are as follows: 0.01, 37.6758, 100
Step 6, inputting the optimal values of C, sigma and epsilon obtained by the PSO algorithm into an SVR model for training, using a test set for inspection to meet set requirements, obtaining a forecasting model, and using the forecasting model to forecast runoff of the year to be forecasted, wherein the forecasting result is as follows:
therefore, the method provided by the invention is adopted to represent the model accuracy by the average relative error and the distance consistent rate on the training set and the test set, and the following table shows that:
if the direct influence of the scheduling of medium and large-scale hydraulic engineering on medium and long-term runoff of warehousing runoff of the Danjiang river mouth reservoir is not considered, the forecasting is carried out by adopting the existing method only considering weather forecasting factors, and the forecasting precision is shown in the following table:
therefore, under the condition that the influence of the scheduling of medium and large-scale hydraulic engineering on the medium and long-term runoff of the watershed is considered, the average relative error of the runoff forecast on the test set obtained by forecasting the runoff of 7-month input into the reservoir of the Danjiang watershed of the Hanjiang watershed is improved by 10%, the precision is greatly improved, the model forecasting precision can be improved to better describe the condition of future runoff, and the runoff forecast has higher reference value for daily scheduling of the Danjiang watershed reservoir.
By adopting the technical scheme disclosed by the invention, the following beneficial effects are obtained: the method for forecasting the medium-and-long-term runoff considering the influence of the hydraulic engineering scheduling considers the influence of medium-and-large-scale hydraulic engineering scheduling on the medium-and-long-term runoff forecasting, improves the forecasting precision and the practicability of the medium-and-long-term runoff forecasting compared with the conventional common method, has the advantages of simple compiling, less parameter setting and strong global optimization capability, and can effectively avoid the problems of large calculation amount, long time consumption, large required sample amount, low forecasting precision caused by the fact that a forecasting result is easy to fall into local optimization and the like. Can be used as an effective method for forecasting long-term runoff in a specific drainage basin.
The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and improvements can be made without departing from the principle of the present invention, and such modifications and improvements should also be considered within the scope of the present invention.
Claims (7)
1. A medium-long term runoff forecasting method considering hydraulic engineering scheduling influence is characterized by comprising the following steps:
s1, screening weather forecast factors;
s2, selecting medium and large hydraulic engineering which forecasts main flow and branch flow at the upstream of the section and has certain influence on runoff, and calculating hydraulic engineering influence factors according to the following steps:
s201, calculating the full storage rate of each hydraulic engineering according to the following formula:
p＝wl/WL
wherein the content of the first and second substances,
wl represents the monthly mean of the actual water levels of the selected months of the selected hydraulic project;
WL represents the characteristic water level of the hydraulic engineering, for the hydraulic engineering with flood prevention task in flood season, the WL value in the flood season is the flood limit water level of the hydraulic engineering, and the WL value in the non-flood season is the normal water storage level; for the hydraulic engineering without flood control tasks in the flood season, WL values of the flood season and the non-flood season are both normal water storage levels;
p represents the full storage rate of the hydraulic engineering, and the value range is [0-1 ];
s202, calculating the full storage rate of all hydraulic engineering of the basin according to the following formula:
wherein n represents the number of hydraulic works, piIs the full rate, RC, of hydraulic engineering iiRepresenting the total storage capacity of the hydraulic engineering i, wherein TRC is the total storage capacity of all the hydraulic engineering, and P is the full storage rate of all the hydraulic engineering in a drainage basin and is a hydraulic engineering influence factor;
s3, normalizing the weather forecast factor and the hydraulic engineering influence factor to obtain a normalized value of the forecast factor;
s4, selecting runoff data of a basin to be forecasted in S years as a historical sample, taking the runoff data of the previous N years in the historical sample and the normalization value of the forecasting factor of the selected corresponding time as a training set, and taking the runoff data of the next M years and the normalization value of the forecasting factor of the selected corresponding time as a test set, wherein S is N + M, N is more than M, and S, N, M is positive integers;
s5, determining the optimal values of the error punishment parameter C, the kernel parameter sigma and the insensitive loss coefficient epsilon in the SVR model by adopting a particle swarm optimization algorithm, inputting the optimal values into the SVR model, training the SVR model by utilizing a training set, checking by utilizing a test set, obtaining a forecasting model if the optimal values meet the requirements, or reselecting part of forecasting factors, and then training the model by utilizing a new training test set until the model reaches the standard;
s6, forecasting the runoff to be forecasted by using the forecasting model;
in step S4, the test set is used for inspection, and if the requirements are met, a forecasting model is obtained, where the requirements to be met are: evaluating the model by the distance and average consistency rate, wherein the distance and average consistency rate of the forecasting model is required to be more than 60%, and the calculation formula of the distance and average consistency rate is as follows:
R＝Nsame/N×100％
wherein R is the rate of consistency from the distance, N is the total years participating in the verification, and N issameThe years that the signs of the measured runoff and the predicted runoff pitch are the same are shown.
2. The method for forecasting the medium and long term runoff considering the dispatching influence of the hydraulic engineering as claimed in claim 1, wherein the step S1 comprises the following steps:
s101, performing correlation analysis on the historical multinomial circulation index data and the historical runoff data of the basin to be forecasted according to a formula to obtain correlation coefficients between annual runoff y and forecasting factors x:
in the formula (I), the compound is shown in the specification,
ρia correlation coefficient representing the ith factor and the annual runoff quantity y,
represents the average value of the factor over a number of years,
the mean value of the run-off of the sample is shown,
l is the runoff sample year;
s102, selecting a circulation index with a large correlation coefficient and a certain physical significance as a weather forecasting factor.
3. The method for forecasting the medium-and-long-term runoff considering the dispatching influence of the hydraulic engineering according to claim 1, wherein in the step S3, the normalization processing is specifically performed according to the following formula:
wherein, ytAVG, STD represent the value of a predictor at a time within a selected time period, the average value of the values of the predictors selected within the selected time period, and the standard deviation of the values of the predictors selected within the selected time period, respectively.
4. The method for forecasting the medium-and-long-term runoff considering the dispatching influence of the hydraulic engineering according to claim 1, wherein in S4, the method for determining the optimal values of the error penalty parameter C, the kernel parameter sigma and the insensitive loss coefficient epsilon in the SVR model by adopting the particle swarm optimization algorithm comprises the following steps:
s401, determining the value ranges of the three parameters of C, sigma and epsilon;
s402, initializing a particle swarm, namely setting the scale, the iteration times, the position and the speed of the particle swarm;
s403, determining a fitness evaluation function, and calculating a fitness value of each particle by using the fitness function;
s404, comparing the fitness of the current position of each particle with the fitness of the historical best position of each particle, and determining the current best position
S405, comparing the fitness of the current optimal position of each particle with the fitness of the current optimal position of the whole group to determine the current optimal position;
s406, updating the particle speed and the particle position;
s407, judging whether an end condition is met, if not, turning to S403; if the global optimal solution is satisfied, outputting a global optimal solution, wherein the global optimal solution at the moment is the optimal solution of three parameters, namely C, sigma and epsilon; the end condition is that the number of iterations reaches an upper limit.
5. The method for forecasting the medium-and-long-term runoff considering the scheduling influence of the hydraulic engineering according to claim 4, wherein S403 specifically comprises the steps of selecting a certainty coefficient as a fitness evaluation function, and calculating the fitness value of each particle by using the following formula as an evaluation function:
where DC is the deterministic coefficient, yc(i) To predict value, yo(i) In order to be the actual value of the measurement,
is the average value of measured values of years, and n is the number of years of the training set sample.
6. The method for forecasting the medium-and-long-term runoff considering the dispatching influence of the hydraulic engineering according to claim 4, wherein S406 specifically comprises:
the particle velocity is updated according to the following equation:
vi+1＝wvi+c1r1(localbest-xi)+c2r2(globalbest-xi)
the particle position is updated according to the following formula:
xi+i＝xi+vi+1
wherein i represents the number of iterations, xiIndicating the position of the particle at the i-th iteration, viRepresenting the velocity of the particle at the i-th iteration, r1，r2Two random numbers, c, expressed between (0, 1)1，c2The expression is a speed-increasing factor, the value of which is generally 2, w is a dynamic weight factor, and the value range of the factor is [0.4, 0.9 ]]。
7. The method for forecasting the medium and long term runoff considering the dispatching influence of the hydraulic engineering according to claim 6, wherein the dynamic weighting factor w is dynamically updated according to the following formula:
wherein, winiAnd wendThe initial value and the final value of the dynamic weight factor are respectively 0.9 and 0.4, G is the iteration number of the PSO algorithm, and G is the current iteration number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910114775.4A CN109919362B (en) | 2019-02-14 | 2019-02-14 | Medium-and-long-term runoff forecasting method considering hydraulic engineering scheduling influence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910114775.4A CN109919362B (en) | 2019-02-14 | 2019-02-14 | Medium-and-long-term runoff forecasting method considering hydraulic engineering scheduling influence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109919362A CN109919362A (en) | 2019-06-21 |
| CN109919362B true CN109919362B (en) | 2021-06-04 |
Family
ID=66961539
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910114775.4A Expired - Fee Related CN109919362B (en) | 2019-02-14 | 2019-02-14 | Medium-and-long-term runoff forecasting method considering hydraulic engineering scheduling influence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109919362B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111445085A (en) * | 2020-04-13 | 2020-07-24 | 中国水利水电科学研究院 | Medium-and-long-term runoff forecasting method considering influence of medium-and-large-sized reservoir engineering water storage |
| CN112818608A (en) * | 2021-02-22 | 2021-05-18 | 南方电网调峰调频发电有限公司 | Medium-and-long-term runoff forecasting method based on improved particle swarm optimization algorithm and support vector machine |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2732477B1 (en) * | 1995-03-31 | 1997-06-27 | Ecole Nationale Superieure Agr | METHOD AND SYSTEM FOR DETERMINING A FLOW OF WATER OUTSIDE A WATER BASIN, METHOD FOR DETERMINING DATA THEREOF |
| CN107274031A (en) * | 2017-06-29 | 2017-10-20 | 华中科技大学 | The hydrologic forecasting method and system of a kind of coupled neural network and VIC distributed hydrological model |
| CN108647425B (en) * | 2018-05-07 | 2019-05-21 | 中国水利水电科学研究院 | K-means high flow or low flow time forecasting procedure based on particle group optimizing |
| CN108876021B (en) * | 2018-05-31 | 2020-06-02 | 华中科技大学 | Medium-and-long-term runoff forecasting method and system |
-
2019
- 2019-02-14 CN CN201910114775.4A patent/CN109919362B/en not_active Expired - Fee Related
Non-Patent Citations (2)
| Title |
|---|
| Wavelet ANN based monthly runoff forecast;Huijun Xu;《Applied Mechanics and Materials》;20131231;第421卷;全文 * |
| 水库中长期水文预报模型研究;冯小冲;《中国优秀硕士学位论文全文数据库（工程科技Ⅱ辑）》;20110515(第5期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109919362A (en) | 2019-06-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Tegegne et al. | Comparison of hydrological models for the assessment of water resources in a data-scarce region, the Upper Blue Nile River Basin | |
| Song et al. | Differences in extremes and uncertainties in future runoff simulations using SWAT and LSTM for SSP scenarios | |
| CN111665575B (en) | Medium-and-long-term rainfall grading coupling forecasting method and system based on statistical power | |
| CN111104981A (en) | Hydrological prediction precision evaluation method and system based on machine learning | |
| Wu et al. | A framework for assessing compound drought events from a drought propagation perspective | |
| CN111445085A (en) | Medium-and-long-term runoff forecasting method considering influence of medium-and-large-sized reservoir engineering water storage | |
| CN111553394B (en) | Reservoir water level prediction method based on cyclic neural network and attention mechanism | |
| WO2022110582A1 (en) | Runoff forecasting method based on baseflow separation and artificial neural network model | |
| CN112883558B (en) | Hydrological model parameter time-varying form construction method | |
| CN109919362B (en) | Medium-and-long-term runoff forecasting method considering hydraulic engineering scheduling influence | |
| CN113361742A (en) | Hydrologic simulation-based regional comprehensive drought identification method | |
| CN112287299B (en) | Quantitative attribution method, device and system for river health change | |
| CN111626006A (en) | Quantitative identification method for multivariate evaluation on runoff change sensitivity | |
| CN115238947A (en) | Social and economic exposure degree estimation method for drought, waterlogging and sudden turning event under climate change | |
| CN114357737B (en) | Agent optimization calibration method for time-varying parameters of large-scale hydrologic model | |
| Kumar et al. | Multi-model evaluation of catchment-and global-scale hydrological model simulations of drought characteristics across eight large river catchments | |
| CN117094516B (en) | Urban group month living water demand prediction method based on fixed effect model | |
| CN113836807B (en) | River and lake ecological flow forecasting and early warning method based on entropy method and long-term and short-term memory neural network | |
| Tadić et al. | Mapping of maximum snow load values for the 50-year return period for Croatia | |
| CN113487069B (en) | Regional flood disaster risk assessment method based on GRACE daily degradation scale and novel DWSDI index | |
| CN115186874A (en) | Hydrological drought prediction method in changing environment | |
| CN115859416A (en) | Average wind speed profile prediction method based on building wind load extreme value equivalence | |
| CN108108860A (en) | A kind of four steps coupling MEDIUM OR LONG RANGE HYDROLOGIC FORECAST METHOD | |
| Liu et al. | Meteorological drought forecasting using Markov Chain model | |
| CN113222235B (en) | Natural gas index prediction method and system based on diffusion and synthesis indexes |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20210604 |