Reservoir real-time regulation and storage characteristic-based inconsistent design flood deduction method

Abstract

The invention belongs to the technical field of water conservancy projects, and discloses a reservoir real-time regulation and storage characteristic-based inconsistent design flood deduction method, which comprises the following steps of: determining an effective flood regulation storage capacity and a flood regulation storage capacity of a reservoir in a preset time period before a flood peak according to an actually measured flood peak sequence of a drainage basin, and determining an effective reservoir index sequence; deducing design floods in different recurrence periods under the non-consistency condition by taking the design floods as influence factors; and carrying out rationality analysis on calculation of the inconsistent design flood, and determining the inconsistent design flood of the watershed. According to the method, the effective reservoir indexes are established by combining the operation state and the flood control water level of the reservoir before the flood peak comes, the real-time regulation and storage capacity of the reservoir before the session flood comes is quantified, the non-consistent design flood model is effectively integrated, powerful reference is provided for the basin non-consistent design flood remarkably affected by reservoir regulation and storage and the applicability of the basin non-consistent design flood, and the method has good application prospects. And potential safety hazards of flood control objects and engineering design caused by unreasonable reservoir index selection are reduced.

Description

Technical Field

[0001] The invention discloses a method for calculating a non-consistent design flood based on real-time regulation and storage characteristics of a reservoir, and belongs to the technical field of water conservancy engineering. Background Art

[0002] Floods are one of the most destructive natural disasters, causing unprecedented damage to the natural environment, social economy and even life in the basin. Flood frequency analysis is the most widely used technology for characterizing flood risks and assessing engineering safety. It has important reference value for basin water resources management and flood control goals, and is an important prerequisite for the construction and operation of water conservancy projects.

[0003] Traditional flood frequency analysis methods are usually based on the "consistency" assumption, which assumes that the statistical characteristics of flood sequences remain stable over a long period of time. At present, there are more than 60,000 reservoirs larger than 15m in the world, and their operation can redistribute the temporal and spatial changes of runoff, thereby regulating the runoff state. Under the influence of this human activity, more than 50% of the world's river runoff patterns have undergone inconsistent changes, resulting in great challenges for traditional flood frequency analysis methods. In practice, with the construction and operation of reservoir groups, the inconsistency characteristics of flood sequences affected by reservoir regulation have become increasingly significant, making it impossible for traditional flood frequency analysis methods to accurately reflect the current real flood situation in the basin, nor to provide a reliable design flood basis for the construction and operation of water conservancy projects.

[0004] Under the influence of reservoir group construction, the inconsistency of flood sequences is mainly reflected in two aspects: first, the regulating function of reservoirs changes the natural distribution process of runoff, causing changes in the time and size of flood peaks; second, with the increase in the number of reservoirs and the expansion of capacity, the overall flood regulation capacity of the basin is enhanced, resulting in changes in the statistical characteristics of flood sequences. Therefore, it is expected to establish a connection between reservoir regulation and floods through reservoir indicators to reflect the impact of reservoir regulation on flood situations. However, the existing reservoir indicators all consider the characteristic water level or characteristic storage capacity of the reservoir. In actual operation, some characteristic water levels of the reservoir (such as dead water level) do not play a role in regulating floods, and the characteristic storage capacity of the reservoir, as a fixed value, cannot effectively establish a connection with the dynamic water inflow, and cannot truly reflect the impact of the reservoir's real-time regulation on the flood situation when the flood comes. The mechanism of the existing indicator construction shows that its correlation with floods is weak, which will lead to the inability of the calculation of inconsistent design floods to truly reflect the characteristics of basin floods, thereby bringing risks to basin flood control and water conservancy project construction. Summary of the invention

[0005] The purpose of the present invention is to provide a method for calculating a non-consistent design flood to solve the technical problems in the prior art that the physical basis of reservoir indicators is weak and the relationship between the real-time regulation and storage operation of the reservoir and the flood peak cannot be reflected. To achieve the above purpose, the present invention proposes a method for calculating a non-consistent design flood based on the real-time regulation and storage characteristics of the reservoir. The specific scheme is as follows:

[0006] A method for calculating non-consistent design flood based on real-time regulation and storage characteristics of reservoirs, comprising:

[0007] Step 1: Determine the effective reservoir index sequence based on the effective flood control storage capacity and flood control storage capacity of the reservoir in the preset period before the peak of the flood in the field in the hydrological data;

[0008] Step 2: deriving the inconsistent design flood based on the effective reservoir index sequence.

[0009] Preferably, the step 1 specifically includes:

[0010] Obtain the flood peak sequence and occurrence time measured in the basin based on hydrological data;

[0011] Extracting the effective flood control storage capacity and flood control storage capacity of the reservoir in the preset period before each flood peak in the flood peak sequence according to the occurrence time;

[0012] Determine the effective reservoir index according to the ratio of the effective flood control storage capacity in the preset period before each flood peak to the flood control storage capacity;

[0013] An effective reservoir index sequence is determined based on the multiple effective reservoir indexes.

[0014] Preferably, step 2 specifically includes:

[0015] According to the effective reservoir index sequence, the inconsistent design flood is deduced by the source tracing reconstruction method or the time-varying moment-reliability method.

[0016] Preferably, the step 2 specifically includes:

[0017] Step 2.1, according to the effective reservoir index sequence, the inconsistent design flood is deduced respectively by the source tracing reconstruction method and the time-varying moment-reliability method, and two deduction results are obtained;

[0018] Step 2.2: Determine the inconsistent design flood based on the two derivation results.

[0019] Preferably, the step 2.2 specifically includes:

[0020] The uncertainty assessment is performed on the inconsistent design flood results derived by the source tracing reconstruction method and the time-varying moment-reliability method respectively;

[0021] An optimal non-consistent design flood derivation method is determined based on the uncertainty assessment results.

[0022] Preferably, the step 2.2 specifically includes:

[0023] After the non-uniform changes occur, the consistent design flood representing the current conditions of the basin is derived;

[0024] The inconsistent design flood results derived by the source tracing reconstruction method and the time-varying moment-reliability method are compared with the consistent design flood results respectively to determine the optimal method for deriving the inconsistent design flood.

[0025] Preferably, the consistent design flood is derived, specifically including:

[0026] Performing inconsistent variation diagnosis on the flood peak data in the hydrological data;

[0027] Determining a consistent sequence representing the current conditions of the watershed according to the diagnosis results of the inconsistent variation diagnosis;

[0028] A consistent design flood is deduced based on the consistent sequence.

[0029] Preferably, deducing a consistent design flood according to the consistent sequence specifically includes:

[0030] Performing multi-scale analysis on the consensus sequence to determine the optimal distribution line type of the consensus sequence;

[0031] estimating distribution parameters of the optimal distribution line shape;

[0032] A frequency curve is drawn according to the optimal distribution line type and the distribution parameters, and a consistent design flood is deduced according to the frequency curve.

[0033] Preferably, the non-consistent design flood is derived according to the time-varying moment-reliability method, specifically including:

[0034] The non-consistent design flood is derived based on the time-varying moment and equal reliability method or the average design life level method.

[0035] Preferably, before deducing the inconsistent design flood according to the time-varying moment and equal reliability method or the average design life level method, the method further includes:

[0036] The optimal distribution type of the time-varying moment is determined by two evaluation criteria: Akaike Information Criterion and Schwarz Bayesian Criterion.

[0037] Beneficial effects: The present invention combines the operating status of the reservoir before the arrival of a flood and the flood limit water level, establishes a connection between the remaining storage capacity of the reservoir and the flood peak, constructs an effective reservoir indicator that can reflect the real-time dynamic scheduling of the reservoir, quantifies the storage capacity of the reservoir before the arrival of a flood, and reflects the impact of the reservoir on the downstream flood peak. It can be effectively and flexibly combined with the inconsistent design flood model to improve the calculation accuracy of the design flood caused by inconsistent changes in the flood sequence caused by the reservoir, and reduce the safety hazards of flood control objects and engineering design caused by unreasonable selection of reservoir indicators. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a schematic diagram of characteristic water level and characteristic storage capacity of the reservoir of the present invention;

[0039] Figure 2 It is a flow chart of a method for calculating a non-consistent design flood in an embodiment of the present invention;

[0040] Figure 3 It is the Mann-Kendall mutation test result of the flood peak sequence measured by Guide Hydrological Station from 1960 to 2020 in the embodiment of the present invention;

[0041] Figure 4 It is the Pettitt mutation test result of the flood peak sequence measured by Guide Hydrological Station from 1960 to 2020 in the embodiment of the present invention;

[0042] Figure 5 It is the design flood with different return periods of Guide Hydrological Station under the consistent status quo conditions based on the traditional hydrological frequency analysis framework in the embodiment of the present invention;

[0043] Figure 6 is the time series of the effective reservoir index ERI of the Longyangxia Reservoir in the upper reaches of the Yellow River in the embodiment of the present invention;

[0044] Figure 7 A schematic diagram of statistical fitting of the effective reservoir index ERI sequence and the flood peak sequence of Guide Hydrological Station in an embodiment of the present invention;

[0045] Figure 8 It is a schematic diagram comparing the sequence reconstructed by the source tracing reconstruction method using ERI as an influencing factor and the flood peak sequence measured by Guide Hydrological Station in an embodiment of the present invention;

[0046] Fig. 9 Design floods under non-consistent conditions with different return periods obtained based on the source tracing reconstruction method in the embodiment of the present invention;

[0047] Fig.10 It is an optimal inconsistency quantile curve diagram based on time-varying moment with ERI as a covariate in an embodiment of the present invention;

[0048] Fig.11The embodiment of the present invention is based on the ER and ADLL reliability concepts respectively, and uses ERI as a covariate under different return periods for non-consistent design floods;

[0049] Fig.12 Schematic diagram of comparison between non-consistent design flood and consistent design flood under two different return periods of source tracing and reconstruction method and time-varying moment-reliability method in an embodiment of the present invention, wherein (a) is the source tracing and reconstruction method, and (b) is the time-varying moment-reliability method;

[0050] Fig.13 The Bootstrap method is used in the embodiment of the present invention to evaluate the uncertainty of inconsistent design floods under different return periods, where (a) is the source reconstruction method and (b) is the time-varying moment-reliability method. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with specific implementation methods. It should be understood that the specific implementation methods described here are only used to explain the present invention and do not limit the scope of protection of the present invention.

[0052] The existing reservoir index has a weak physical basis and cannot reflect the relationship between the real-time regulation and storage operation of the reservoir and the flood peak. The present invention proposes an effective reservoir index, denoted as ERI. It is specifically an effective reservoir index that combines the effective flood regulation storage capacity of the reservoir itself and the flood control limit water level, such as Figure 1 The present invention defines effective flood control storage capacity as: the storage capacity between the reservoir check flood level and the reservoir real-time water level. Specifically, the expression formula of the effective reservoir index is as follows (1):

[0053]

[0054] Among them, ERC is the effective flood control storage capacity; RC is the flood control storage capacity.

[0055] The present invention uses ERI as a flood inconsistency influencing factor and quantifies the reservoir's ability to regulate floods according to the remaining storage capacity of the reservoir before the flood peak, thereby improving the accuracy of calculating the designed flood under inconsistent conditions, reducing the risk of project operation, and ensuring the safety of flood control targets.

[0056] The following two embodiments specifically illustrate the method for deriving water supply based on inconsistency design of effective reservoir indicators of the present invention.

[0057] Embodiment 1:

[0058] Step 1: Determine the effective reservoir index sequence based on the effective flood control storage capacity and flood control storage capacity in the preset period before the peak of the flood in the field in the hydrological data;

[0059] Furthermore, step 1 specifically includes: obtaining the flood peak sequence and occurrence time measured in the basin according to the hydrological data;

[0060] Extract the effective flood control storage capacity and flood control storage capacity of each reservoir in the preset period before the flood peak in the flood peak sequence according to the occurrence time; determine the effective reservoir index according to the ratio of the effective flood control storage capacity to the flood control storage capacity in the preset period before each flood peak; determine the effective reservoir index sequence according to multiple effective reservoir indicators. The preset period can be adjusted by technicians in this field according to the actual project.

[0061] Specifically, in this embodiment, historical data of the basin are collected and sorted, specifically historical observation data of flood peaks in the basin and reservoir operation are collected. For example, the measured flood peak data and the time of historical flood peaks at the Guide Hydrological Station from 1960 to 2020 are obtained. Taking the physical causes of inconsistent changes in flood peaks as the starting point, the reservoir operation criteria are clarified, the effective flood control storage capacity and flood control storage capacity are determined, and thus the effective reservoir indicators are determined.

[0062] This embodiment specifically includes: according to the flood peak occurrence time, find the effective flood control storage capacity of the reservoir one day before the flood peak occurs; and calculate the flood control storage capacity in combination with the flood limit water level of the reservoir operation. The effective reservoir index sequence is obtained by the ratio of the effective flood control storage capacity to the flood control storage capacity.

[0063] Among them, this embodiment also includes performing consistent design flood deduction.

[0064] Furthermore, the flood peak data in the hydrological data are diagnosed for inconsistent variation; based on the diagnostic results of the inconsistent variation diagnosis, a consistent sequence that meets the current conditions is determined; and based on the consistent sequence, a consistent design flood is derived.

[0065] Furthermore, a multi-scale analysis is performed on the consistency sequence to determine the optimal distribution line type of the consistency sequence; the distribution parameters of the optimal distribution line type are estimated; a frequency curve is drawn according to the optimal distribution line type and the distribution parameters, and the consistency design flood is deduced according to the frequency curve.

[0066] Specifically, the inconsistent variation of flood peak data is diagnosed. Based on a variety of inconsistent diagnosis methods, the inconsistent variation caused by the operation of water conservancy projects on the flood peak sequence is clarified, including trend test and mutation test.

[0067] The trend tests specifically include: Mann-Kendall (MK) trend test, Pearson correlation coefficient test and Breusch-Panan (BP) variance test;

[0068] Mutation tests specifically include: Mann-Kendall (MK) mutation test, Pettitt mutation test and two-sample T test.

[0069] Find the consistent sequence of flood peaks under current conditions. Based on the above variation diagnosis of flood peak data, a variety of inconsistency diagnosis methods are used to find the consistent flood peak sequence under current conditions from the period after the flood peak data mutates, in preparation for subsequent rationality analysis.

[0070] The traditional hydrological frequency analysis framework is established, and the consistent flood peak sequence under the above-mentioned current conditions is calculated, and the design flood analysis is performed on the consistent sequence. This mainly includes the determination of distribution linearity and the estimation of distribution parameters.

[0071] The linear distribution includes: multi-scale screening distribution linearity. Make full use of various distribution line types suitable for hydrology, including Weibull, Gamma, Gumbel, Lognormal, Logistic and Gev distributions, and build a library of candidate distribution line types. The information of the candidate distribution line types is shown in Table 1.

[0072] Table 1 Alternative probability distribution information

[0073]

[0074] The adaptability of each line type is measured by the goodness of fit test criterion, and the Kolmogorov-Smirnov (KS) test and Nash efficiency coefficient NSE are used. QQ The accuracy of different line types is evaluated by using three indicators: , root mean square error (RMSE). The accuracy evaluation indicators are not limited to the above three. This embodiment uses the above three commonly used indicators.

[0075] a) Kolmogorov-Smirnov (KS) test

[0076] The KS test determines the fitting effect by the difference between the empirical cumulative frequency distribution of the hydrological series and the theoretical overall distribution. The test statistic is shown in the following formula (2):

[0077]

[0078] Where F0(x) is the theoretical cumulative probability density, F n (x) is the empirical cumulative probability density of the sample sequence, D n is the KS test statistic, is the critical value under the significance level α, and n is the length of the measured sequence. Then, based on the significance level α, we accept the null hypothesis F n (x) = F0(x).

[0079] Empirical frequency F n(x) Calculated according to the empirical frequency formula proposed by Gringorten:

[0080]

[0081] Where i is the order of the observations in ascending order.

[0082] b) NSE QQ Evaluation Criteria

[0083] The Nash efficiency coefficient (NSE) between the theoretical quantile and the empirical quantile is QQ ) as the evaluation criteria, as shown in the following formula (4):

[0084]

[0085] In the formula, is the sample x i The quantile of the theoretical distribution F0(x) corresponding to the empirical frequency of i is the measured flood peak sequence value, is the mean of the measured flood peak sequence.

[0086] c) Root mean square error (RMSE)

[0087] RMSE can effectively evaluate the extreme value phenomenon in the simulation, as shown in the following formula (5):

[0088]

[0089] At the same time, in order to estimate the distribution parameters of the line type, the linear moment method is used in this embodiment to estimate the distribution parameters. It should be understood that the present invention is not limited to this method, and the maximum likelihood method, moment method, Bayesian method, etc. can be selected.

[0090] Step 2: Derivation of the inconsistent design flood based on the effective reservoir index sequence.

[0091] According to the effective reservoir index sequence, the inconsistent design flood is deduced by the source reconstruction method or the time-varying moment-reliability method.

[0092] Based on the above effective reservoir index sequence, the inconsistent design flood is deduced. The present invention uses two inconsistent design flood deduction methods, namely the source tracing reconstruction method and the time-varying moment-reliability method, to respectively deduce the inconsistent design flood of the basin after being affected by the reservoir. It should be understood that those skilled in the art can select a suitable deduction method according to the actual project.

[0093] Specifically, in this embodiment, two methods, namely, source tracing reconstruction method and time-varying moment-reliability method, are used to deduce the inconsistent design flood respectively.

[0094] Specifically, step 2.1, according to the effective reservoir index sequence, the inconsistent design flood is deduced by the source reconstruction method and the time-varying moment-reliability method respectively, and two deduction results are obtained;

[0095] The following is an introduction to the source reconstruction and time-varying moment-reliability method:

[0096] (1) The source reconstruction method is based on the physical mechanism of the effect of influencing factors on the hydrological sequence. It seeks the source function that causes the hydrological sequence to change, removes the influencing factors of the hydrological sequence according to the source function, and thus realizes the conversion of the inconsistent sequence into a new consistent reconstructed sequence. This method assumes that the hydrological sample Y is composed of all influencing factors X, and the product model expression is as follows:

[0097]

[0098] In the formula, Y is the hydrological sample, X is i is the i-th influencing factor of hydrological sample Y, f i (X i ) is the impact factor X i The source function of , n is the total number of all influencing factors that constitute the hydrological sample Y.

[0099] After removing all or the main inconsistency change influencing factors X, the remaining part is the consistent reconstruction sequence RS, as shown in the following formula:

[0100]

[0101] Where m is the number of influencing factors that undergo inconsistent changes.

[0102] Combined with the framework of consistent design flood, the distribution linearity and distribution parameters of the consistent reconstruction sequence RS are determined to obtain the traceability reconstruction sequence design value RS P Next, determine the value of the influencing factor under the condition of non-consistent change, and return it to the traceability reconstruction design value RS according to formula (8): P , then the hydrological design value Y under non-consistent conditions can be obtained P , as shown in the following formula:

[0103]

[0104] Where: RS P is the design value of the source reconstruction sequence obtained by traditional hydrological frequency analysis based on the reconstruction sequence RS, X i,design is the value of the influencing factor of non-uniform change under design conditions

[0105] (2) The basic principle of hydrological frequency analysis based on the time-varying moment method is that the distribution line shape of the hydrological sequence is the same, while the parameters will no longer be the same constant, but a function of the covariate, which will change with the change of the covariate. In particular, the generalized additive model (GAMLSS) with position, scale and shape parameters proposed based on the time-varying moment method realizes the flexible application of the time-varying moment method. The invention combines the GAMLSS model with a series of methods based on the concept of reliability (time-varying moment-reliability method) to deduce the hydrological design value under non-consistent conditions.

[0106] This embodiment uses the distribution line type selected by the above assessment as the candidate distribution for design flood derivation. In order to screen the optimal distribution type of the GAMLSS model, the Akaike Information Criterion (AIC) and the Schwarze Bayesian Criterion (SBC) are used to determine the optimal non-consistent model. The smaller the AIC and SBC values, the better the model fit.

[0107] Next, flood frequency analysis is performed based on the GAMLSS model results and combined with reliability. Frequency analysis is one of the most effective means of assessing flood risks. In reality, "recurrence period" is usually used instead of "frequency" to characterize the magnitude of floods. However, when using the time-varying moment method, if the flood design value of the time-varying distribution is determined based on the traditional concept of recurrence period, there will be a different design value every year. In this regard, this embodiment uses "engineering reliability" instead of recurrence period to clarify the possibility of flood events exceeding the design service life of the project, that is:

[0108]

[0109] Where: ns represents the non-consistent condition; T1-T2 is the design service life of the water conservancy project, that is, from T1 to T2; p t is the probability of exceeding that varies with time; θ t is a vector of distribution parameters that vary with time; F y (y p |θ t ) is the distribution function of the hydrological variable y.

[0110] Based on "engineering reliability", the "equal reliability" (ER) method and the "average design life level" (ADLL) method are used in combination with the GAMLSS model to deduce the non-consistent flood design value. The equal reliability method and the average design life level method are briefly introduced below.

[0111] ①Equal reliability method (ER)

[0112] Under the condition that the hydrological sequence meets the consistency, the reliability of the design value of the hydrological sequence under the return period of T years is expressed as Indicates that:

[0113]

[0114] In the formula, is the reliability of the design value under consistency conditions.

[0115] Under the condition of water level sequence inconsistency, during the engineering design life period T1-T2, the reliability of the project is the same as formula (9). Although the "T-year return period" based on the consistency condition is no longer applicable in theory, the hydrological design value y under the "T-year return period" can still be deduced based on the concept of equal reliability. p ,according to So we have the following formula:

[0116]

[0117] According to the equal reliability concept, the concept of return period under consistent conditions is still applicable. Combined with the design service life of the project, the hydrological design value under inconsistent conditions can be obtained according to the above formula.

[0118] ②Average Design Life Level Method (ADLL)

[0119] Similarly, some scholars proposed the average design life level method (ADLL) and defined the annual average reliability index as follows:

[0120]

[0121] Where: is the annual average reliability.

[0122] The study redefined the probability distribution of peak flow within the engineering design service life T1-T2 as shown in the following formula:

[0123]

[0124] Where: Essentially, the flood peak is within the design service life T1-T2 of the project. The weighted mixed distribution.

[0125] Assumptions Then the following formula can be obtained:

[0126]

[0127] Based on the above formula, the flood design values ​​of the project under different recurrence periods within the design life can be derived.

[0128] Step 2.2: Determine the inconsistent design flood based on the two derivation results.

[0129] Specifically, in this embodiment, the optimal non-consistent design flood model is selected, the accuracy of the non-consistent design flood derivation results obtained by the above different models is evaluated, and the non-consistent design flood derivation model suitable for the characteristics of the watershed is selected.

[0130] Uncertainty assessments are performed on the inconsistent design flood results derived by the source reconstruction method and the time-varying moment-reliability method respectively; and the optimal inconsistent design flood derivation method is determined based on the uncertainty assessment results.

[0131] Specifically, uncertainty assessment of non-consistent design floods: Using the Bootstrap method, the significance level α = 0.05 was selected, and the obtained non-consistent design floods were sampled 5000 times (the specific number of samplings can be adjusted according to actual conditions), and the corresponding design flood 95% confidence interval was obtained based on the sampling results of the flood design value.

[0132] Specifically, based on the two derivation results, determining the inconsistent design flood also includes: deducing the consistent design flood under the current conditions; comparing the inconsistent design flood results deduced by the source reconstruction method and the time-varying moment-reliability method with the consistent design flood results under the current conditions, and determining the derivation method of the optimal inconsistent design flood.

[0133] That is, based on the design flood value of the consistent sequence under the inferred current conditions, the calculation results of different inconsistent design flood models are compared with it. By observing the degree of deviation from the design flood of the consistent sequence under different recurrence periods, the rationality of the design flood results of different models is evaluated. There is no need to construct the measured consistent sequence multiple times and calculate its design flood, only different models of inconsistent design flood calculations are needed for comparison.

[0134] By comparing with the consistent design flood and conducting uncertainty analysis, the optimal non-uniform design flood model suitable for the basin is screened.

[0135] Embodiment 2:

[0136] In recent years, as one of the important hydropower development bases in China, the upper reaches of the Yellow River have basically formed a cascade hydropower station pattern. The construction of hydropower projects has an impact on the evolution of hydrological elements. One of the impacts is that the flood peak sequence downstream of the project has undergone significant inconsistent changes. As the "dragon head" reservoir in the upper reaches of the Yellow River, Longyangxia is a large-scale controlled reservoir with multi-year regulation capabilities. Its operation has a significant impact on the natural flood peak situation. Tangnaihai and Guide Hydrological Stations serve as the inlet and outlet stations of the Longyangxia Reservoir, and monitor the changes in flood situations upstream and downstream of Longyangxia before and after the operation. Therefore, with regard to the operation of the Longyangxia Reservoir and the measured flood peaks of the Tangnaihai and Guide Hydrological Stations as the data basis, the technical solution of the present invention is further specifically described through Example 1 and the accompanying drawings. This embodiment is specifically as follows Figure 2 shown.

[0137] S1: Collect historical flood data from Tangnaihai and Guide hydrological stations from 1960 to 2020, as well as operation data of Longyangxia Reservoir since its completion in 1986, such as water level-reservoir capacity relationship curve, daily average water level of the reservoir, etc.

[0138] S2: Diagnosis of non-consistent variation of flood peak sequence. The MK trend test, Pearson correlation coefficient test and BP variance test methods were used to test the trend of the flood peak sequence of the two hydrological stations. The results are shown in Table 2. It can be seen that the mean and variance of the flood peak sequence of the upstream Tangnaihai hydrological station did not show a significant trend, while the mean of the flood peak sequence of the Guide hydrological station showed a significant downward trend.

[0139] Table 2 Trend test of flood peak series in Tangnaihai and Guide

[0140]

[0141] Note: -4.275 and -0.543 indicate significant trend changes.

[0142] Since the measured flood peak sequence in Guide showed a significant trend change, the Pettitt test, MK mutation test and two-sample T test were used to perform mutation tests on the sequence. The results are as follows: Figure 3 , Figure 4 And as shown in Table 3.

[0143] Table 3 Two-sample T test results

[0144] years N1 / a N2 / a p-value 1986 27 34 <![CDATA[2.98*10 -12 ]]>

[0145] Note: 2.98*10 -12 Indicates a significant trend change.

[0146] Figure 3The results show that when the significance level α = 0.05, the UF and UB curves have only one intersection point (1986), and the intersection point is within the confidence interval, which means that 1986 is the time when the flood peak sequence undergoes a mutation.

[0147] Figure 4 The results show that the maximum value of the test method statistic occurred in 1986, and the corresponding probability of occurrence is 1.28×10 -7 , if the significance test is less than α=0.05, it is considered that the peak sequence underwent a significant mutation in 1986.

[0148] In Table 3, in 1986, the two-sample T test P = 2.98 × 10 -12 <α=0.05, so the double sample T believes that the measured flood peak sequence in Guide suddenly changed in 1986.

[0149] The trend and mutation test results show that the measured flood peak sequence at Guide Station has a significant downward trend and a mutation occurred in 1986. Considering that Longyangxia Reservoir was built in 1986 and the sluice gate was closed to store water in October 1986, it is believed that the operation of Longyangxia Reservoir is the main reason for the inconsistent changes in the flood peak sequence of Guide Hydrological Station.

[0150] S3: Based on the variation points of the S2 flood sequence, find the consistent flood peak sequence under the current conditions. After diagnosis by the inconsistency test method, it was finally confirmed that the measured flood peak sequence of Guide Hydrological Station from 1988 to 2018 represents the consistent sequence under the current conditions.

[0151] S4: Establish a traditional hydrological frequency analysis framework and use the traditional hydrological frequency to deduce the design flood for the consistent series from 1988 to 2018. Combined with Table 4, all alternative distributions pass the KS test. According to NSE QQ The maximum and the minimum RMSE are finally considered to be the optimal distribution. The distribution parameters are: μ = 746.394, σ = 245.383, ε = 0.775

[0152] Table 4 ERI traceability reconstruction sequence distribution test results

[0153]

[0154]

[0155] Based on the Weibull optimal distribution, the consistent flood design value under the current conditions is obtained, such as Figure 5 As shown, the design value for once in a hundred years is 2510m 3 / s.

[0156] S5: Establish effective reservoir index. First, combine the flood control limit water level of Longyangxia Reservoir to calculate the flood control storage capacity (RC). And according to the flood peak time of Guide Station, find the Longyangxia operating water level one day before the flood peak, so as to obtain the effective flood control storage capacity (ERC). According to formula 1, the effective reservoir index sequence (ERI) is obtained, as follows: Figure 6 The effective reservoir indicator series plotted as shown.

[0157] Figure 6 The effective reservoir index sequence is plotted. It can be seen that since the construction and operation of Longyangxia Reservoir in 1986, ERI has shown irregular ups and downs, indicating that the index is constantly changing with the different operating conditions of the reservoir.

[0158] S6: Based on the effective reservoir indicators, derive the inconsistent design flood.

[0159] (1) The effective reservoir index (ERI) is used as the source reconstruction method of the influencing factor; the ERI series and the flood peak series are statistically fitted, such as Figure 7 As shown. The source function is: y = 2553.25 × e^(-0.3014 × ERI)

[0160] After eliminating the influence of the reservoir according to Formula 7, the reconstructed sequence can be constructed, as follows: Figure 8 As shown in Table 5, it can be seen that the trend of the reconstructed sequence weakened and the mutation point in 1986 disappeared. After the trend test (Table 5), no significant trend was found.

[0161] Table 5 Trend test results of ERI reconstruction sequence

[0162]

[0163] Based on the traditional hydrological frequency framework constructed by S4, the design flood calculation is performed on the consistent ERI reconstruction sequence, combined with the KS test, NSE QQ , RMSE, it can be obtained that Gumbel is the optimal distribution (Table 6), and the distribution parameters are: μ = 2272, σ = 612.

[0164] Table 6 ERI traceability reconstruction sequence distribution test results

[0165]

[0166]

[0167] Based on the Gumbel optimal distribution, the ERI index in the current stage is used as the design value, and the design flood value under the non-consistent condition is obtained by combining Formula 8, as follows: Fig. 9 As shown, the design value for once in a hundred years is 2670m 3 / s.

[0168] (2) Effective Reservoir Index (ERI) is used as a covariate to derive the non-consistent design flood using the time-varying moment-reliability method

[0169] In recent years, the GAMLSS model containing position, scale and shape parameters has made the application of the time-varying moment method more flexible. At present, most studies on non-consistent design floods usually use time t as a covariate, and believe that the parameters of the hydrological series distribution change with the change of time t. However, if only time t is used as a covariate to explore the changes in flood sequences, it obviously lacks practical physical meaning, and the results are not in line with reality and cannot be used in the future. Therefore, the invention will use the GAMLSS model based on ERI as a covariate and combine it with the reliability method to deduce the non-consistent design flood value.

[0170] Table 7 shows the GAMLSS model fitting results of each alternative distribution with ERI as the covariate under four scenarios. Among them, the AIC and SBC values ​​of the Lognormal distribution under scenario 4 are the smallest among all scenarios and all distributions, which are 881.6 and 887.9 respectively. Therefore, when both parameters change with ERI, the Lognormal model has the best fitting effect as the alternative distribution model.

[0171] Table 7 Statistical results of alternative distributions with ERI as covariate

[0172]

[0173]

[0174] Table 8 shows the corresponding distribution parameters according to the optimal distribution model with ERI as the covariate. The fitting results of the theoretical quantiles of the optimal distribution inconsistency model and the measured flood peak sequence are plotted, as shown in Table 8. Fig.10 shown.

[0175] Table 8 Statistics of optimal distribution parameters of ERI covariates

[0176]

[0177] Fig.10 It can be seen that the optimal non-consistent model with ERI as a covariate simulates the flood peaks changing year by year, and the change trend is similar to the measured flood peak sequence. ERI as a covariate can effectively fit the non-consistent changes in flood peaks caused by water conservancy project construction, and has a better fitting effect.

[0178] Based on ERI as the optimal distribution model of the covariate (ERI: Lognormal distribution g(μ):ERI, g(σ):1), the ER and ADLL reliability methods are used to calculate the non-consistent design flood. The results are as follows: Fig.11 shown.

[0179] S7: Design flood plausibility assessment.

[0180] (1) Comparison of the flood design values ​​based on the consistent sequence and the inconsistent flood design values ​​based on the source reconstruction method and the time-variable moment-reliability method. Fig.12 It can be seen that when the return period is small, the non-consistent design flood obtained by the source reconstruction method and the time-varying moment-reliability method with ERI as the influencing factor is slightly larger than the measured consistent flood peak sequence design value, that is, the calculation of the non-consistent design flood method is safer. As the return period increases, the deviation gradually decreases. When the return period is once in 100 years, the deviation of the source reconstruction method is only 160m 3 / s. It is worth noting that the time-varying moment-reliability method using ERI as a covariate is equal to the consistent sequence flood design value under a return period of 100 years, which can better fit the flood situation of the current basin conditions.

[0181] Table 9 Design floods with corresponding return periods

[0182]

[0183] Compared with the consistent sequence design flood, it can be seen that the ERI index can effectively reflect the impact of the reservoir's real-time regulation and storage on the flood situation. When the source reconstruction method removes the ERI effect, the impact of the reservoir on the flood can be effectively removed, thereby obtaining real and effective restored data. When the moment-reliability method considers the ERI index, the derived design flood under inconsistent conditions is more in line with the current design flood form.

[0184] (2) The uncertainty of the calculation results of the non-consistent design flood model is evaluated based on the Bootstrap method. The results are as follows: Fig.13 It can be seen that the uncertainty of the design value of the non-consistent design flood increases with the increase of the return period, but the uncertainty based on the ERI source tracing reconstruction method is relatively small, and the growth rate of uncertainty is also relatively small; in the time-varying moment-reliability method, the uncertainty intervals of the non-consistent design flood of the ADLL and ER methods are particularly similar.

[0185] Table 10 summarizes the uncertainty intervals of the design flood calculated by different models. It can be seen that the uncertainty of the source reconstruction method is the smallest overall, and the uncertainty interval is only 770m when the 100-year flood event occurs. 3 / s. The time-varying moment method has a large uncertainty in the basin as a whole. The uncertainty interval of the 100-year return period for the inconsistency design flood value derived by using ER and ADLL reliability is 1110m 3 / s、1100m 3 / s.

[0186] Table 10 Uncertainty intervals of design floods under different return periods (m 3 / s)

[0187] 10 20 50 100 Source Reconstruction 420 530 670 770 ERI_ER 640 760 970 1110 ERI_ADLL 620 750 960 1100

[0188] After analyzing the rationality of two inconsistent design flood models, it is found that the ERI index reflecting the real-time regulation and storage characteristics of the reservoir can meet a variety of inconsistent design flood methods, making different inconsistent design flood methods have good applicability in this basin. Specifically, the difference in design flood values ​​between the source reconstruction method and the time-varying moment-reliability method considering the single influencing factor of ERI is less than 6% under different return periods. When the return period is small, the design flood values ​​of the source reconstruction method and the time-varying moment-reliability method are slightly larger than the current status of the basin. However, as the return period increases, the design floods of the source reconstruction method and the time-varying moment-reliability method considering the ERI index effectively represent the design flood value of the current status of the basin. Under a return period of 100 years, the design flood value of the time-varying moment-reliability method is equal to the current design flood of the basin, which can provide a reasonable reference for the design and operation of water conservancy projects and ensure the safe and stable operation of water conservancy projects under inconsistent conditions. From the perspective of Bootstrap uncertainty, the uncertainty interval of the calculation results of the source reconstruction method is generally small, and its results are more reliable than those of the time-varying moment-reliability method. Therefore, overall, the source reconstruction method based on ERI is more suitable for the design flood calculation of this basin. When the basin flood changes inconsistently, this result can provide theoretical reference and guidance for the design flood calculation of this basin.

[0189] The present invention firstly pays attention to the actual operation of water conservancy projects before the flood peak arrives, locates the remaining storage capacity of the reservoir according to the time of flood occurrence, and thus constructs the ERI sequence for annual floods. According to ERI, the reservoir's storage capacity for flood peaks can be judged, thereby estimating the impact of the reservoir on the downstream flood peak situation. Nowadays, with a large number of water conservancy projects already under construction or in preparation, reservoirs cause more than 50% of the world's river runoff to undergo inconsistent changes. Based on the ERI index that reflects the real-time storage characteristics of reservoirs, the design flood frequency analysis of inconsistent changes in hydrological conditions caused by the operation of water conservancy projects can be effectively solved.

[0190] The above are only several embodiments of the present invention and are not intended to limit the present invention in any form. Although the present invention is disclosed as above in the form of a preferred embodiment, it is not intended to limit the present invention. Any technician familiar with the profession, without departing from the scope of the technical solution of the present invention, using the above disclosed technical content to make slight changes or modifications are equivalent to equivalent implementation cases and fall within the scope of the technical solution.

Claims

1. A method for calculating non-consistent design flood based on real-time regulation and storage characteristics of reservoirs, characterized in that: include: Step 1: Determine the effective reservoir index sequence based on the effective flood control storage capacity and flood control storage capacity of the reservoir in the preset period before the peak of the flood in the field in the hydrological data; Step 2: deriving the inconsistent design flood based on the effective reservoir index sequence.

2. The method for calculating the non-consistent design flood according to claim 1, characterized in that: The step 1 specifically includes: Obtain the flood peak sequence and occurrence time measured in the basin based on hydrological data; Extracting the effective flood control storage capacity and flood control storage capacity of the reservoir in the preset period before each flood peak in the flood peak sequence according to the occurrence time; Determine the effective reservoir index according to the ratio of the effective flood control storage capacity in the preset period before each flood peak to the flood control storage capacity; An effective reservoir index sequence is determined based on the multiple effective reservoir indexes.

3. The method for calculating the non-consistent design flood according to claim 1, characterized in that: Step 2 specifically includes: According to the effective reservoir index sequence, the inconsistent design flood is deduced by the source tracing reconstruction method or the time-varying moment-reliability method.

4. The method for calculating the non-consistent design flood according to claim 1, characterized in that: The step 2 specifically includes: Step 2.1, according to the effective reservoir index sequence, the inconsistent design flood is deduced respectively by the source tracing reconstruction method and the time-varying moment-reliability method, and two deduction results are obtained; Step 2.2: Determine the inconsistent design flood based on the two derivation results.

5. The method for calculating the non-consistent design flood according to claim 4, characterized in that: The step 2.2 specifically includes: The uncertainty assessment is performed on the inconsistent design flood results derived by the source tracing reconstruction method and the time-varying moment-reliability method respectively; An optimal non-consistent design flood derivation method is determined based on the uncertainty assessment results.

6. The method for calculating the non-consistent design flood according to claim 4, characterized in that: The step 2.2 specifically includes: After the non-uniform changes occur, the consistent design flood representing the current conditions of the basin is derived; The inconsistent design flood results derived by the source tracing reconstruction method and the time-varying moment-reliability method are compared with the consistent design flood results respectively to determine the optimal method for deriving the inconsistent design flood.

7. The method for calculating the non-consistent design flood according to claim 6, characterized in that: Develop consistent design floods, including: Performing inconsistent variation diagnosis on the flood peak data in the hydrological data; Determining a consistent sequence representing the current conditions of the watershed according to the diagnosis results of the inconsistent variation diagnosis; A consistent design flood is deduced based on the consistent sequence.

8. The method for calculating the non-consistent design flood according to claim 7, characterized in that: Deducing a consistent design flood according to the consistent sequence specifically includes: Performing multi-scale analysis on the consensus sequence to determine the optimal distribution line type of the consensus sequence; estimating distribution parameters of the optimal distribution line shape; A frequency curve is drawn according to the optimal distribution line type and the distribution parameters, and a consistent design flood is deduced according to the frequency curve.

9. The method for calculating the non-consistent design flood according to claim 3 or 4, characterized in that: The non-consistent design flood is derived according to the time-varying moment-reliability method, specifically including: The non-consistent design flood is derived based on the time-varying moment and equal reliability method or the average design life level method.

10. The method for calculating the non-consistent design flood according to claim 9, characterized in that: Before the non-consistent design flood is derived based on the time-varying moment and equal reliability method or the average design life level method, it also includes: The optimal distribution type of the time-varying moment is determined by two evaluation criteria: Akaike Information Criterion and Schwarz Bayesian Criterion.

Images