A Method for Quantifying Overtopping Risk of Dams in Cascade Reservoir Groups Based on the CRORM Model
By constructing the CRORM model and combining Monte Carlo simulation and the DB-IWHR model, a quantitative analysis of the overtopping risk of dams in a cascade reservoir group was achieved. This solved the problem that existing technologies cannot quantify the impact of uncertain factors on overtopping risk and revealed the risk transmission pattern among cascade reservoirs.
Patent Information
- Application Number
- CN202611183161.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-05
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies are insufficient to quantitatively analyze the impact of uncertainties on the overtopping risk of cascade reservoir groups, and cannot reveal the complex relationship of risk transmission between cascade reservoirs. Traditional methods cannot achieve risk quantification.
A CRORM model was constructed, and Monte Carlo simulation and DB-IWHR model were combined. The entire process of reservoir overtopping identification, dam break flood analysis and downstream flood regulation was calculated using MATLAB software. The risk transmission law under different working conditions was analyzed, and the impact of upstream reservoir status on downstream overtopping risk was quantified.
It has achieved the step-by-step quantification of the overtopping risk of a cascade reservoir group and the analysis of the risk transmission process, reduced the deviation between the risk quantification results and the actual engineering conditions, and revealed the risk transmission law between cascade reservoirs.
Smart Images

Figure CN123115641A_ABST
Abstract
Description
Technical Field
[0001] This invention provides a quantitative method for assessing the overtopping risk of cascade reservoir groups based on numerical simulation technology. First, it clarifies the impact of uncertainties on overtopping risk. Second, it employs a combination of Monte Carlo stochastic simulation and flood control calculations, along with the DB-IWHR dam-break model, to construct a cascade reservoir group dam overtopping risk model (CRORM) based on the cascading failure process of upstream flood peak, inter-district flow, and inflow. Then, it uses MATLAB software to perform calculations for the entire process of "reservoir overtopping identification, dam-break flood analysis, and downstream reservoir flood control," quantifying the impact of upstream reservoir status on downstream overtopping risk. Finally, it analyzes different operating conditions using the controlled variable method to reveal the risk transmission patterns among cascade reservoirs. This invention belongs to the field of safety and risk analysis in water conservancy and hydropower engineering, specifically involving numerical simulation and other technologies. Background Technology
[0002] Against the backdrop of the accelerated transformation of the national energy structure, the deep integration of water conservancy projects with digital and intelligent technologies has become a mainstream trend in the industry. Cascade hydropower development, as an important component of clean energy, has experienced rapid growth. Cascade reservoirs, as crucial national strategic infrastructure, play an irreplaceable role in watershed water resource allocation, flood control and disaster reduction, and energy supply. Their safe operation is directly related to the safety of people's lives and property, sustainable economic and social development, and ecological stability within the basin. my country's cascade water conservancy and hydropower projects are mostly distributed in the high mountain and canyon areas of southwest China, facing complex natural conditions such as high altitude, strong earthquakes, and steep slopes. Frequent extreme weather events and the intertwined risks of earthquakes, floods exceeding standard levels, and bank instability make the working environment of cascade reservoir groups extremely complex. Overtopping, as one of the main factors leading to reservoir failure, has made its risk prevention and control a key issue in the safety management of water conservancy projects.
[0003] Existing research on the quantification of overtopping risk in cascade reservoir groups primarily relies on qualitative analysis, making it difficult to quantitatively analyze the impact of uncertainties on overtopping risk or reveal the complex relationships of risk transmission between cascades. With technological advancements, this field has achieved a leap from qualitative to quantitative analysis: by analyzing uncertain factors such as natural inflow floods, flood discharge capacity, flood regulation capacity, and downstream flood discharge capacity, numerical simulation technology is used to complete the entire process calculation, obtaining core data on overtopping probability and dam-break floods. This enables the quantification of overtopping risk, reveals the risk transmission patterns of cascading overtopping in cascade reservoir groups, and outputs risk quantification results that closely reflect actual operating conditions. This overcomes the technical limitations of relying solely on qualitative analysis and failing to accurately quantify overtopping risk.
[0004] Considering the need for a method to quantify the risk of dam overtopping in cascade reservoir groups, and addressing the limitations of traditional methods that only consider a single factor and cannot achieve risk quantification, this invention proposes a method for quantifying the risk of dam overtopping in cascade reservoir groups based on the CRORM model. First, combining Monte Carlo simulation and the DB-IWHR model, a CRORM model is constructed based on the cascading failure process of "upstream flood peak - interval flow - inflow". Then, MATLAB software is used to calculate the entire process of dam overtopping in the cascade reservoir group. Finally, through analysis of different operating conditions, the transmission law of overtopping risk in the cascade reservoir group is revealed.
[0005] The technical challenge of this invention lies in connecting the uncertain random sampling, time-by-time flood regulation calculation, upstream reservoir overtopping judgment, dam break flood calculation, downstream inflow flood construction, and overtopping failure probability calculation in the same cascade calculation chain, and making the discharge flow process of the previous cascade reservoir a component of the inflow flood of the next cascade reservoir in the calculation. Summary of the Invention
[0006] This invention addresses the limitations of traditional single-reservoir, qualitative analysis models, which are ill-suited to the complex working environment of cascade reservoir groups in high mountain and canyon areas. It constructs a "CRORM" model that couples random variable sampling, step-by-step flood regulation calculations, overtopping failure identification, and dam-break flood process calculations. This model enables the step-by-step quantification of overtopping risk in cascade reservoir groups and the analysis of risk transmission processes. By simulating different operating conditions, it obtains the conditional overtopping probability, overall overtopping probability, and risk transmission relationships between cascade reservoirs at each level. First, the uncertainty of factors affecting overtopping failure is analyzed from the perspective of flood. By parameterizing key random variables and setting appropriate probability distributions, the flood peak is fitted with a P-III type distribution, and the error parameters and discharge correction coefficients are described by a normal distribution, laying the foundation for risk quantification. Second, flood control calculation and Monte Carlo simulation techniques are integrated. Random samples are generated using Latin hypercube sampling and substituted into the water balance equation to simulate the dynamic response of the reservoir. The evolution of the earth-rock dam break flood is analyzed using the DB-IWHR model. Overtopping failure is judged through system function, and the risk rate is estimated by failure probability. The entire process of overtopping calculation for cascade reservoir group dams, including "reservoir overtopping discrimination, dam break flood analysis, and downstream reservoir flood control", is realized using MATLAB software. Finally, by analyzing different operating conditions: uncertainty of flood discharge capacity, uncertainty of inflow flood size, uncertainty of flood control capacity, increase of downstream flood discharge capacity, and joint increase of downstream flood discharge capacity and flood control capacity, and early warning, the risk transmission law between cascade reservoirs is revealed. The main process consists of the following three steps: uncertainty factor analysis, flood control calculation and Monte Carlo simulation, and CRORM model construction.
[0007] 1. Analysis of uncertainties regarding dam overtopping in a cascade reservoir group:
[0008] The external uncertainties of cascade reservoirs include earthquakes, floods exceeding standard levels, and bank instability. The internal uncertainties include flood discharge capacity, emergency measures, dam type combinations, flood control storage capacity, and initial regulation water level. From the perspective of floods, the uncertainties affecting overtopping failure include: uncertainty of natural inflow floods, uncertainty of flood control initial regulation water level, uncertainty of water level-storage capacity curve, uncertainty of downstream flow, and uncertainty of dam type. Key random variables are parameterized: the peak flow of natural inflow floods is distributed using the P-III type, the initial regulation water level is uniformly distributed, the error of the water level-storage capacity curve and the discharge correction coefficient follow a normal distribution, and the dam type is set according to the failure mode of "overtopping failure of earth-rock dams and non-overtopping failure of concrete dams", clarifying the physical meaning and value boundaries of the variables.
[0009] 2. Flood Control Calculation and Monte Carlo Simulation:
[0010] Based on the water balance equation, overtopping is determined by calculating total inflow, outflow, and net inflow flood volumes. The criterion is that the cumulative net inflow flood volume exceeds the initial flood control capacity. Monte Carlo simulation is used, with a set number of simulations to meet accuracy requirements. Latin hypercube sampling is employed to extract samples of key random variables. These samples are substituted into the flood control calculation equation to simulate reservoir water level changes and record the highest water level. Upstream earth-rock dam breach floods are obtained using the DB-IWHR model, and downstream inflow floods are the superposition of upstream outflow floods and inter-regional floods. Failure is determined based on the system function, and the calculated failure probability is used as an estimate of the overtopping failure probability.
[0011] 3. CRORM Model Construction:
[0012] The CRORM model, programmed using MATLAB, integrates overtopping detection, dam-break flood analysis, and downstream flood regulation. The calculation process is as follows: scaling up typical floods to deduce different frequency processes → extracting random variables and performing flood regulation calculations → combining downstream inflow floods → repeating simulations, identifying and recording overtopping events → calculating risk rates. Through analysis of different operating conditions, the risk transmission patterns between cascade reservoirs are revealed.
[0013] The relationships between the various computational stages of the CRORM model:
[0014] For each random sample, the extracted uncertainty parameters are substituted into the time-by-time flood control calculation, and the flood overflow judgment results are used to determine whether the upstream reservoir has experienced flood overflow failure. The downstream reservoir inflow flood is constructed based on whether the upstream reservoir has experienced a breach: if no breach has occurred, it consists of the upstream reservoir outflow flood and the interval flood; if a breach has occurred, the time of breach is used as the switching point, the upstream reservoir normal outflow flood is included before the breach, and the upstream reservoir breach flood is included after the breach; the normal outflow flood and the breach flood are included according to their respective time segments and are not superimposed repeatedly in the same calculation period, while the interval flood is superimposed once in each calculation period.
[0015] Samples of downstream reservoir overtopping failures under both upstream and downstream reservoir overtopping failure conditions were collected separately. The corresponding conditional overtopping failure probabilities were calculated, and based on the total probability relation, the overall overtopping failure probability of the downstream reservoir was obtained by weighting the probabilities of the two upstream reservoir states. For multi-stage reservoirs, the discharge flow process of the preceding stage reservoir was considered as part of the inflow flood of the following stage reservoir. Flood control calculations, overtopping detection, and probability statistics were repeatedly performed along the flow direction to quantify the transmission, superposition, or blockage of risks between cascade reservoirs.
[0016] The various calculation stages are closely interconnected: random samples are the input for flood control calculations, flood control results are the basis for overtopping judgment, overtopping judgment results and dam type are the control conditions for determining the upstream reservoir discharge process, the upstream reservoir discharge process and the interval flood together form the downstream reservoir inflow flood, and the downstream flood control and overtopping judgment results are used to calculate the conditional overtopping failure probability and the overall overtopping failure probability; the output of the previous stage constitutes the input or control condition of the next stage, thus forming the overall calculation process of the CRORM model.
[0017] Compared with the prior art, the present invention has the following advantages:
[0018] First, it realizes the analysis of uncertainty in the quantification of overtopping risk of cascade reservoir groups. It breaks through the technical limitations of existing technologies that only consider uncertainty in a single and scattered manner and often use fixed value simplification. It selects appropriate probability distributions for each uncertainty to complete parameter calibration and quantification, takes into account the influence between various uncertainties, avoids the defects of traditional methods that quantify in isolation and ignore interaction, and reduces the deviation between the risk quantification results and the actual engineering conditions.
[0019] Secondly, most existing technologies only conduct independent overtopping risk calculations for a single reservoir, neglecting the risk transmission, superposition, and evolution characteristics between cascade reservoirs via floods. This invention takes a holistic system perspective of a cascade reservoir group. Its core advantage lies in considering the risk relationship between the normal and overtopping states of upstream reservoirs and changes in their overtopping conditions on downstream reservoirs, characterizing the risk amplification effect caused by flood superposition, and overcoming the shortcomings of traditional methods in quantifying the overall and accurate overtopping risk of a cascade reservoir group. This approach is more in line with the actual risk prevention and control needs of watershed engineering. Attached Figure Description
[0020] The key processes and results involved in this invention are visually illustrated in the following figures:
[0021] Figure 1 A schematic diagram of the “CRORM” model: This diagram summarizes the model’s calculation process, including flood input, state assessment, and risk output.
[0022] Figure 2 Analysis of the overtopping risk rates of HZY and CHB considering the uncertainty of flood control reservoir capacity. Detailed Implementation
[0023] Based on the above, the following uses the Houziyan-Changheba (HZY-CHB) cascade reservoir system as an example to illustrate the specific implementation process of the CRORM model. This implementation method requires the combination of hydrological data, reservoir and dam characteristic parameters, and the MATLAB software environment, but the application of this invention is not limited to this example.
[0024] Step 1. Uncertainty Factor Analysis
[0025] Located in a high mountain and canyon area, the HZY-CHB cascade reservoirs face uncertainties in external factors, including: the probability, intensity, and extent of earthquakes are uncertain, potentially causing damage or even collapse of the HZY earth-rock dam; floods exceeding standard levels occur in extreme scenarios with a return period of 5,000-10,000 years, and the inflow peak of the HZY reservoir follows a P-III type distribution with an average value of 3,322 m³ / s; and bank instability can easily lead to landslide dams or surges, exacerbating the risk of overtopping. The uncertainties of internal factors are manifested in the following ways: the flood discharge capacity is corrected for deviations by the normal distribution correction coefficient, and the flood discharge capacity of HZY does not meet the specifications when it drops to 77%; emergency measures are subject to the presence or absence of early warning, and early warning can buy an additional 3 hours of flood control time; HZY is an earth-rock dam that will collapse if it overflows, and CHB has a flood control capacity of 415 million cubic meters; the flood control capacity is divided into three levels, with the starting water level for HZY being 1802-1845.41m and for CHB being 1600-1690m, and excessively high water levels will reduce the flood control capacity.
[0026] Step 1.1 Parameterization of Uncertainty Regarding Overtopping Risk of a Single Reservoir
[0027] The uncertainty parameterization of the risk of overtopping of a single reservoir is as follows: The randomness of natural inflow floods is characterized by the probability density function of type P-III distribution (1), and quantified by shape, scale, and location parameters (2)-(4). The statistical parameters are consistent with the factors of super-standard floods and are determined by the method of moments from the historical flood peak flow samples of HZY reservoir: =3322 m 3 / s, C v =0.26, C s =1.30; the starting water level for flood control follows a uniform distribution. The fluctuation range of HZY Reservoir is 1802-1845.41m, and that of CHB Reservoir is 1600-1690m. The fluctuation range between the actual value and the expected value is directly related to the loss of flood control storage capacity.
[0028] The probability density function of the P-III type distribution is expressed as follows:
[0029] (1)
[0030] In the formula, ; It is a gamma function; , and The shape, scale, and location parameters of the P-III type distribution are represented as follows:
[0031] (2)
[0032] (3)
[0033] (4)
[0034] In the formula, This is the average of the peak flow series. The deviation coefficient reflects the degree of dispersion of the peak flow series; The skewness coefficient reflects the degree of asymmetry in the probability distribution of peak flood flow.
[0035] The water level-storage capacity curve is set as a piecewise linear function according to formula (5). The reference water level of HZY reservoir is 1802m, and the storage capacity coefficients are p1=0.49, p2=-5.98, and p3=57.69. The error between the actual and theoretical curves is quantified by formula (7), and the error parameters follow a standard normal distribution. The discharge flow is quantified by formula (8), and the discharge correction coefficient is adjusted according to a normal distribution. The broad-crested weir coefficients of HZY reservoir are C1=1.43, C2=0.94, and m=0.8. The dam type is set as the failure mode parameters of an earth-rock dam, i.e., overtopping and collapse, and a corresponding relationship related to the overtopping risk is established.
[0036] In the engineering design phase, the curve interpolation capacity relationship is expressed as:
[0037] (5)
[0038] (6)
[0039] In the formula, For reservoir Real-time storage capacity For reservoir Water level at any time; The relationship between water level and reservoir capacity. , Let be the water level and reservoir capacity values for the i-th topographic control point; , This represents the water level and reservoir capacity values for the next adjacent control point.
[0040] During reservoir operation, at a given water level H, the actual reservoir capacity... It follows a normal distribution, and its mean is the reservoir capacity corresponding to the design water level-reservoir capacity relationship. .
[0041] (7)
[0042] In the formula, This is the actual water level-reservoir capacity relationship function. For the ideal design water level-reservoir capacity relationship function, The error parameter between the actual storage capacity and the ideal storage capacity follows a standard normal distribution.
[0043] The discharge flow rate is expressed as:
[0044] (8)
[0045] (9)
[0046] In the formula, It is the outflow. It refers to the reservoir's water storage capacity. This is the discharge correction factor; The curve showing the relationship between reservoir water level and discharge capacity is shown in m. w B is the spillway weir flow coefficient. a For the net width of the flood discharge channel, H a Where g is the elevation of the spillway crest, and g is the acceleration due to gravity. for The inverse function, assumed to follow a normal distribution. , The standard deviation is determined based on the design and actual discharge data of different reservoirs.
[0047] Step 2. Flood Control Calculation and Monte Carlo Simulation for a Single Reservoir
[0048] Step 2.1 Execution of Flood Control Calculation Procedure
[0049] Step 2.1.1 Initial State Setting
[0050] This embodiment selects the historical flood event of September 1981 as the typical flood event. Based on the ratio of the natural inflow peak discharge to the typical flood peak discharge of each randomly selected sample, the typical flood event is amplified by the same factor to form the corresponding sample's natural inflow flood event. The flood event lasts 110 hours, with a calculation time step of 1 hour.
[0051] The initial reservoir water levels of HZY and CHB reservoirs were determined by corresponding random samples or specific calculation conditions. The initial breach width, final breach elevation, and initiation flow velocity of HZY reservoir were used for the DB-IWHR dam-break flood calculation after overtopping, and were not used as initial conditions for normal flood control calculations.
[0052] Step 2.1.2 Calculation of flood regulation for cascade reservoirs
[0053] The 110-hour simulation period is discretized into 110 time periods, Δt=1 hour. The trapezoidal integral method is used to transform equation (10) into the water balance equation (11) to calculate the water storage changes of the two reservoirs in each time period, ensuring the accuracy of numerical calculation.
[0054] (10)
[0055] (11)
[0056] In the formula, and They are respectively and The amount of water stored at any given time. and The calculation period is divided into inbound and outbound flow processes. .
[0057] Step 2.1.3 Calculation of flood volume in cascade reservoirs
[0058] Calculate the total outflow of water from HZY Reservoir according to formula (12). Based on the cumulative discharge volume of the flood discharge plan, the flood discharge volume is... ; Calculate the total inflow of water into HZY according to formula (13) Accumulated volume of floods exceeding standard levels, and flow rate of floods exceeding standard levels. ; Calculate the net inflow of CHB, the cumulative amount of HZY discharge, inter-regional flood and dam-break flood according to formula (14); the cumulative value of net inflow of floodwater into the reservoir. It is the difference between the total inflow and the total outflow of water.
[0059] (12)
[0060] (13)
[0061] (14)
[0062] Step 2.1.4 Overshoot Judgment
[0063] When the total inflow of floodwater Greater than the total outflow from the reservoir At that time, flood risk begins to accumulate in the reservoir. If, during a flood event, there is a certain moment... This results in the cumulative value of net inflow floodwater. Exceeding the initial flood control capacity of the reservoir At this point, the reservoir overflows and transfers the accumulated risk to the next cascade. Therefore, when a reservoir encounters a flood exceeding the standard, the upstream reservoir is judged to have failed due to overflow by formula (15). The initial flood control capacity of HZY reservoir is determined based on its water level, and the initial flood control capacity of CHB reservoir is 415 million cubic meters. After overflow, it is handled according to the failure mode of the corresponding dam type.
[0064] (15)
[0065] Step 2.2 Monte Carlo Simulation Implementation
[0066] Step 2.2.1 Simulation Parameter Setting
[0067] In this embodiment, we assume a random variable They are statistically independent, and their probability density functions are respectively The function for determining the top of the flood is:
[0068] (16)
[0069] In the formula, X1 represents the natural inflow peak discharge, with a P-III type distribution; X2 represents the error in the water level-reservoir capacity relationship. X3 is the discharge correction coefficient ξ, which is normally distributed. X4 represents the initial water level H0 of the reservoir, which is uniformly distributed; and H dam This is the elevation of the dam crest. This represents the highest reservoir water level obtained through time-period flood control calculations. When Z≤0, i.e., H... max ≥H dam When this occurs, a top overflow failure is determined.
[0070] First, random sampling is used to obtain the random quantile values of each random variable. Given N samplings, the function corresponding to the quantile value of each sampling variable is Z. i Z i ≤0 indicates that the dam overtopping, and there are M Z values in N iterations. i ≤0, therefore, the probability of the dam overtopping can be calculated using the following formula:
[0071] (17)
[0072] (18)
[0073] In the formula, Indicated by Sampling failure probability The estimated value is obtained by using the Latin hypercube sampling method to generate N=100,000 random variable samples, and the ratio of the number of failed samples to the total number of samples is used as the estimated value of the flood top failure probability.
[0074] The random variables for HZY Reservoir include the natural inflow peak discharge, the water level-reservoir capacity error, the discharge correction coefficient, and the initial regulation water level; the random variables for CHB Reservoir include the water level-reservoir capacity error, the discharge correction coefficient, and the initial regulation water level. The physical meaning, probability distribution, and parameter determination methods for each random variable are set according to the aforementioned content in this step.
[0075] Step 2.2.2 Simulate the execution of the steps.
[0076] The Latin hypercube sampling technique was used to extract N=100000 sets of random variable samples to make the cumulative distribution interval of each random variable layered and covered to reduce the bias caused by sampling aggregation. Each set of samples was substituted into the flood control calculation equation to simulate the dynamic changes of the water level of the two reservoirs and record the highest water level in front of the dam. The flood of the HZY reservoir dam failure was calculated using the DB-IWHR model. After the HZY reservoir overtopping failure, the dam failure flood process was calculated by the wide-crested weir flow formula (19)-(20), the breach section flow velocity formula (21)-(22), the water level-storage capacity relationship formula (23)-(25), and the water balance formula (26)-(28). The inflow flood of the CHB reservoir is the superposition of the HZY downstream flood and the interval flood with a constant value of 2000 m³ / s. According to the system function (16), if Zi≤0, the dam overtopping failure is determined. The dam crest elevation of the HZY reservoir is recorded as H. dam,HZY The elevation of the CHB reservoir dam crest is denoted as H. dam,CHB Both are used as input parameters for the model and are directly read from the dam crest elevation in the corresponding reservoir engineering design data. In this embodiment, the highest reservoir water level reaching or exceeding the dam crest elevation is used as the criterion for overtopping failure, and no additional tolerance range for judgment error is set. Then, according to formula (17), M is the number of failure samples, N=100000; the failure frequency is calculated as the estimated value of overtopping failure probability, and the result is retained to 6 decimal places.
[0077] The flow rate at the breach section is calculated using the broad-crested weir formula:
[0078] (19)
[0079] (20)
[0080] In the formula, m b m q These represent the flow coefficient and lateral contraction coefficient of the broad-crested weir, respectively. H is the upstream water level of the reservoir above the breach, and z is the elevation of the bottom of the breach. B is the weir crest width, and C is the combined flow coefficient, with a theoretical value of 1.7m. 1 / 2 / s.
[0081] When water flows from the reservoir into the breach, there will be a drop in water level, from H to h. Assuming the flow velocity in the reservoir and the head loss at the inlet section are negligible, then:
[0082] (twenty one)
[0083] (twenty two)
[0084] In the formula, It is the ratio of total head to actual water depth.
[0085] Based on historical cross-sectional data of the river channel, on-site measurements, and remote sensing techniques, the relationship between reservoir capacity and water level was obtained:
[0086] (twenty three)
[0087] In the formula, W is the reservoir capacity, H is the reservoir water level elevation, and H r This is the baseline water level. Expressed as:
[0088] (twenty four)
[0089] (25)
[0090] In the formula, p1, p2, and p3 are the fitting coefficients of the water level-reservoir capacity relationship curve, respectively.
[0091] The breach discharge is determined by the loss of reservoir capacity per unit time, therefore
[0092] (26)
[0093] In the formula, To calculate the change in reservoir capacity over a period of time, This represents the change in reservoir water level over a given period of time. To calculate the time step, q ru This refers to the inbound flow.
[0094] According to the law of conservation of mass, the water balance equation is obtained:
[0095] (27)
[0096] (28)
[0097] In the formula, m is the flow coefficient, which is dimensionless.
[0098] Step 3. CRORM Model Construction and Application
[0099] Step 3.1 Numerical simulation of dam-break floods in a cascade reservoir group
[0100] Step 3.1.1 Solving for the flow rate at the breach cross-section
[0101] HZY is an earth-rock dam. After the dam overflows and breaks, the DB-IWHR model is used to calculate the dam-break flood hydrograph. The flow rate at the breach section is calculated according to the broad-crested weir formula (19) (20). The theoretical value of the comprehensive flow coefficient C is 1.7 m¹ / ² / s. CHB is set as a concrete dam. When the dam overflows, the flow rate is calculated using the broad-crested weir flow formula. The breach does not develop.
[0102] In this embodiment, the input parameters of the DB-IWHR model are as follows:
[0103]
[0104] Step 3.1.2 Solving for the flow velocity at the breach cross-section
[0105] The flow velocity at the breach section of HZY Reservoir was calculated according to formulas (21) and (22). The inflow velocity and head loss at the inlet section were ignored. The water level dropped from H to h, and the starting flow velocity was 3 m / s. HZY Reservoir began to breach at 0:00. The starting flow velocity of the breach scour was 3.0 m / s. The breach flood peak was reached at 22:59, with a peak flow of 17685.18 m³ / s. The flow velocity dropped to below 3.0 m / s at 13:00 on the third day, and the dam breach process ended. The dam breach lasted for 61 hours.
[0106] Step 3.1.3 Solving the water balance equation
[0107] The continuous water level-storage capacity fitting relationship of HZY Reservoir adopts equations (23)-(24), the derivative of the water level-storage capacity relationship with respect to water level adopts equation (25), and the breach flow and water balance relationship adopts equations (26)-(28). The reservoir water level and breach flow at each time are solved according to the preset time step. This step directly calls the water level-storage capacity relationship that has been established before, and does not repeat the establishment of the water level-storage capacity relationship.
[0108] Step 3.2 Establishing the CRORM Model
[0109] Step 3.2.1 Principle of Calculating Overtopping Risk of Cascade Reservoirs
[0110] The one-way propagation of floods means that the downstream reservoir's state does not feed back to change the upstream reservoir's calculation results, but it does not mean that the upstream and downstream flooding events are independent in a probabilistic sense. The flooding failure probability of HZY reservoir is calculated by equation (29); equations (30) and (31) are sample estimates of the flooding failure probability of the downstream reservoir under the conditions of flooding and non-flooding of the upstream reservoir, respectively; equation (32) weights the probabilities under the two conditions according to the upstream state probability to obtain the overall flooding failure probability of CHB reservoir.
[0111] Taking a two-reservoir system connected in series, with A being the upstream reservoir and B being the downstream reservoir, as an example, what is the probability of overtopping failure of the upstream reservoir A? Its mathematical expression is:
[0112] (29)
[0113] In the formula, This indicates that upstream reservoir A has experienced overtopping failure. The value of is independent of the state of downstream reservoir B and is not affected by the state of downstream reservoir B. The conditional flooding probability of downstream reservoir B under the two states of flooding and non-flooding of upstream reservoir A is calculated according to equations (30) and (31), respectively.
[0114] (30)
[0115] (31)
[0116] In the formula, This represents the number of samples where upstream reservoir A experienced overtopping failure. The downstream reservoir B experienced overtopping failure. The upstream reservoir A did not experience overtopping failure. This represents the number of samples where both upstream and downstream reservoirs experienced overtopping failure. This represents the number of samples where upstream reservoir A did not experience overtopping failure while downstream reservoir B did.
[0117] However, the failure probability of downstream reservoir B The condition of upstream reservoir A will be affected. The conditional probabilities of upstream reservoir A under the two conditions of flooding and non-flooding are weighted according to equation (32) to obtain the overall flooding failure probability of downstream reservoir B:
[0118] (32)
[0119] In the formula, Let A be the probability of the upstream reservoir A operating normally. This represents the probability of failure of downstream reservoir B under the condition that upstream reservoir A experiences overtopping failure. This represents the probability of failure of downstream reservoir B under normal operating conditions of upstream reservoir A.
[0120] Step 3.2.2 The process of upstream risk being transmitted to downstream reservoirs
[0121] In this embodiment, the inter-reservoir flood flow between HZY Reservoir and CHB Reservoir is taken as 2000 m³ / s throughout the entire simulation period. When HZY Reservoir does not experience overtopping failure, the inflow flood into CHB Reservoir is formed by the superposition of the normal outflow flood from HZY Reservoir and the inter-reservoir flood. When HZY Reservoir experiences overtopping failure, the normal outflow flood from HZY Reservoir is used before the dam breach, and the dam breach flood from HZY Reservoir is used after the dam breach. The natural inflow flood from HZY Reservoir has already been reflected in the corresponding outflow flood process through flood regulation calculations or dam breach calculations, and will not be superimposed again.
[0122] When upstream reservoir A experiences overtopping failure, i.e. The process of flooding occurs when, before the dam breach, the normal outflow from upstream reservoir A and the inter-regional flood overlap to form the downstream inflow. After the dam breach, the flood from upstream reservoir A breaching and the inter-regional flood overlap to form the downstream inflow. The normal outflow and the dam breach flood are switched at the moment of the dam breach. The flood process is represented as follows:
[0123] (33)
[0124] In the formula, This refers to the inflow flood process into downstream reservoir B. This is a normal flood discharge process from upstream reservoir A. This is a regional flood process. This refers to the flood event caused by the breach of an upstream dam. and These represent the time of the upstream reservoir dam failure and the time when the flood process ended.
[0125] When upstream reservoir A is operating normally When it occurs, its flood control capacity This will play a role in regulating and intercepting floodwaters. The normal outflow floodwaters and the inter-regional floodwaters will overlap to form the inflow flood process of downstream Reservoir B. The flood process is represented as follows:
[0126] (34)
[0127] Step 3.2.3 Model Programming Implementation and Calculation Process
[0128] The CRORM model was implemented using MATLAB R2016a. By programming the dynamic changes in reservoir flood control, risk analysis methods were applied to the software program to calculate risk probabilities. The specific calculation process is as follows:
[0129] 1) By using typical flood hydrographs and employing the same magnification method, flood processes of different frequencies can be derived;
[0130] 2) Determine the total number of random simulations and the dam type of the current cascade reservoirs. Using Latin hypercube sampling, extract the natural inflow peak flow, water level-storage capacity relationship error, discharge correction coefficient, and starting water level. Based on the sampling results, perform flood control calculations to obtain the state of the upstream reservoir and the flood discharge process.
[0131] 3) Based on whether the upstream reservoir has overtossed, the downstream inflow flood process is combined: the overtoss group combines the upstream pre-dam breach flood and the dam breach flood, and is calculated using the DB-IWHR model; the non-overtoss group uses the upstream normal outflow flood process. Both groups are superimposed with the interval flood process.
[0132] 4) Repeat steps 2-3 until the set total number of simulations N=100,000 is reached. After completing the time-period flood control calculation for each random sample, when the highest reservoir water level satisfies: H max ≥H dam, Or, at a certain moment, the cumulative net inflow of floodwater exceeds the initial flood control capacity. If the flooding occurs, it is determined that a flooding failure has occurred. The time of flooding, the highest reservoir water level, and the corresponding sample parameters are recorded, and the number of flooding failure samples is accumulated.
[0133] 5) Calculate the overtopping risk of the upstream reservoir and the overtopping risk of the downstream reservoir under different conditions of the upstream reservoir;
[0134] Formulas (1) to (28) are used for the simulation of uncertainty of natural inflow floods, the calculation of water level-storage capacity and discharge relationship, the calculation of water balance, the overtopping discrimination, the Monte Carlo probability estimation and the calculation of dam-break flood process, respectively, thereby obtaining the reservoir water level process, discharge flow process, overtopping discrimination result and dam-break occurrence time corresponding to each group of random samples. Formulas (29) to (32) classify the random samples according to the overtopping discrimination result, calculate the downstream overtopping failure probability under the two states of overtopping and non-overtopping of the upstream reservoir, and multiply the two overtopping failure probabilities by the occurrence probability of the corresponding state of the upstream reservoir and add them together to obtain the overall overtopping failure probability of the downstream reservoir. Formulas (33) to (34) construct the downstream reservoir inflow flood process according to the overtopping discrimination result of the upstream reservoir; when the upstream overtopping breach occurs, the normal outflow flood before the dam breach and the dam-break flood after the dam breach are used as the boundary, and superimposed with the interval flood to form the downstream reservoir inflow flood. The constructed downstream reservoir inflow flood process is further used as the input for the downstream reservoir's time-by-time flood regulation calculation. The flood regulation results are used for downstream overtopping identification and probability statistics. For multi-stage reservoirs, the discharge flow process of the previous stage reservoir is used as a component of the inflow flood of the next stage reservoir, and the above calculation is repeated along the flow direction. Thus, the output of the previous calculation stage constitutes the input and state criteria of the next calculation stage. The formulas are sequentially linked according to the order of random sampling, flood regulation calculation, overtopping identification, flood process construction, downstream flood regulation, and probability synthesis to form the overall CRORM calculation method.
[0135] Step 3.3 Model Validation and Application
[0136] Step 3.3.1 Model Validation
[0137] Using the flood control process of the "Xiaogangjian-Yibaidao" cascade landslide dams in the Wenchuan earthquake as the verification object, the CRORM model was compared with the DB-IWHR-2018 model. The differences in reservoir capacity simulation were small, the initial and final times of reservoir water level simulation were basically consistent, and the error in the middle period was <5%, thus verifying the reliability of the model.
[0138] Step 3.3.2 Operating Condition: Risk Analysis of the Uncertainty Impact of Flood Control Reservoir Capacity
[0139] To analyze the uncertainty impact on flood control storage capacity, this experiment used the HZY and CHB cascade system as the research object. The total number of simulations was uniformly set to N=100,000, with a constant flood flow rate of 2000 m³ / s, and the frequency of inflow floods was controlled within the 5000-year to 10000-year return period range. The initial reservoir water level was used to characterize the variation of flood control storage capacity. Flood control calculations and overtopping risk assessment were completed using the CRORM model. The results of each set of parameters were analyzed one by one, and the transmission law of overtopping risk in the cascade reservoir group was summarized. The results are as follows:
[0140] Parameter setting 1: HZY initial reservoir water level 1802m is the dead water level, CHB initial reservoir water level 1690m is the flood limit water level, upstream flood discharge capacity is 77% of the design value, and downstream flood discharge capacity is 100%.
[0141] Simulation results: No overtopping failure samples were found in HZY Reservoir, and the estimated probability of overtopping failure of HZY Reservoir is 0; under the condition that HZY Reservoir does not experience overtopping, the estimated probability of overtopping failure of CHB Reservoir is 0.
[0142] Analysis: Under these parameter settings, the upstream reservoir water level drops to the dead water level, and the flood control storage capacity is ample. Even if the upstream flood discharge capacity is damaged and can only reach 77% of the design value, the sufficient flood control storage capacity can still effectively accommodate excess floodwater, making up for the deficiencies caused by insufficient flood discharge capacity and avoiding the risk of upstream reservoir overflow from the source.
[0143] Parameter setting 2: The initial water level of HZY reservoir is 1802m, which is the dead water level; the initial water level of CHB reservoir is 1690m, which is the flood control limit water level; and the flood discharge capacity of both upstream and downstream reservoirs is 100% of the design value.
[0144] Simulation results: The estimated probability of overtopping failure of HZY reservoir is 0; under the condition that HZY reservoir does not overtopping, the estimated probability of overtopping failure of CHB reservoir is 0.
[0145] Analysis: The upstream reservoir's water level is at a dead water level, and its flood discharge facilities are operating at full capacity, effectively regulating and controlling floodwaters and smoothly releasing floodwaters downstream. The downstream reservoir can safely receive upstream inflows and floods within a constant range, and the overall risk is under control.
[0146] Parameter setting 3: The initial water level of HZY reservoir is 1835m, which is the flood control limit level; the initial water level of CHB reservoir is 1690m, which is the flood control limit level; the upstream flood discharge capacity is 77% of the design value; and the downstream flood discharge capacity is 100%.
[0147] Simulation results: The estimated probability of overtopping failure of HZY Reservoir is 0.1120. Under the condition of overtopping failure of HZY Reservoir, the estimated probability of overtopping failure of CHB Reservoir is 1; under the condition that HZY Reservoir does not experience overtopping, the estimated probability of overtopping failure of CHB Reservoir is 0. Based on the aforementioned conditional probability relationship, the estimated overall probability of overtopping failure of CHB Reservoir is 0.1120.
[0148] Analysis: Compared to parameter setting 1, the upstream reservoir water level rises to the flood control limit level, which significantly reduces the available flood control capacity. Coupled with the unfavorable condition of insufficient flood discharge capacity, the reservoir cannot effectively store excess floodwater, and the risk of overflowing the reservoir increases significantly.
[0149] Parameter setting 4: The initial reservoir water level of HZY is 1835m, which is the flood control limit level; the initial reservoir water level of CHB is 1690m, which is the flood control limit level; and the upstream and downstream flood discharge capacity is 100% of the design value.
[0150] Simulation results: The estimated probability of overtopping failure of HZY reservoir is 0; under the condition that HZY reservoir does not experience overtopping, the estimated probability of overtopping failure of CHB reservoir is 3 × 10⁻ 5 Since the HZY reservoir did not experience overtopping under this operating condition, the estimated overall overtopping failure probability of the CHB reservoir is also 3×10⁻. 5 .
[0151] Analysis: As the water level in the upstream reservoir rises, the downstream discharge needs to be increased to ensure its own safety. The superposition of upstream floodwaters and inter-regional floodwaters exacerbates the flood control load on the downstream reservoirs. Even if the upstream reservoirs do not overflow, the risk of downstream overflow will still increase slightly.
[0152] Parameter setting 5: HZY initial reservoir water level 1835m is the flood control limit water level, CHB initial reservoir water level 1650m is the dead water level, upstream flood discharge capacity is 77% of the design value, and downstream flood discharge capacity is 100%.
[0153] Simulation results: The estimated probability of overtopping failure of HZY Reservoir remains 0.1120. Under the condition of overtopping and breaching of HZY Reservoir, the estimated probability of overtopping failure of CHB Reservoir is 1; under the condition that HZY Reservoir does not experience overtopping, the estimated probability of overtopping failure of CHB Reservoir is 0. Based on the aforementioned conditional probability relationship, the estimated overall probability of overtopping failure of CHB Reservoir is 0.1120.
[0154] Analysis: Under the premise that the upstream water level and flood discharge capacity remain unchanged, the downstream reservoir water level is lowered from the flood limit level to the dead water level, the flood control capacity is expanded, the flood holding capacity is significantly improved, and the risk of flood transmitted from the upstream is effectively resisted.
[0155] Parameter setting 6: The initial reservoir water level of HZY is 1835m, which is the flood control limit level; the initial reservoir water level of CHB is 1650m, which is the dead water level; and the upstream and downstream flood discharge capacity is 100% of the design value.
[0156] Simulation results: The estimated probability of overtopping failure of HZY reservoir is 0; under the condition that HZY reservoir does not experience overtopping, the estimated probability of overtopping failure of CHB reservoir is approximately 2 × 10⁻ 5 Since the HZY reservoir did not experience overtopping under this operating condition, the estimated overall overtopping failure probability of the CHB reservoir is approximately 2 × 10⁻⁻⁻⁶. 5 .
[0157] Analysis: This parameter setting further verifies that after the downstream reservoir actively releases its flood control capacity, it can stably receive upstream discharge and superimposed floods within the 5000-year to 10000-year flood range, effectively controlling the risk of overtopping. If the initial water level of CHB is further lowered to 1600m, the flood control capacity will be further increased, and downstream risks will be managed at a deeper level.
[0158] Parameter setting 7: The initial water level of HZY reservoir is 1845.41m, which is the check water level. The initial water level of CHB reservoir fluctuates normally in the range of 1650~1690m. The flood discharge capacity of both upstream and downstream is 100% of the design value.
[0159] Simulation results: The estimated probability of overtopping failure of HZY Reservoir is 0.99635.
[0160] Analysis: This parameter is set to the extreme risk condition, with the upstream reservoir water level close to the check level and the remaining available flood control capacity almost exhausted. Even if the flood discharge facilities operate at full capacity, they cannot accommodate extreme floods, and the probability of the reservoir overflowing is close to 100%.
[0161] Based on the above parameter settings, the risk transmission pattern among cascade reservoirs was analyzed: In the sample of the HZY reservoir experiencing a roof breach, before the dam breach, the inflow flood into the CHB reservoir was formed by the superposition of the normal outflow flood from the HZY reservoir and the inter-reservoir flood; after the dam breach, the inflow flood into the CHB reservoir was formed by the superposition of the HZY reservoir breach flood and the inter-reservoir flood. The normal outflow flood and the breach flood were calculated separately according to different time stages and were not superimposed at the same time.
[0162] The CHB Reservoir has a storage capacity of 1.015 billion cubic meters below its flood control limit, with a usable flood regulation capacity of 415 million cubic meters, only about 30 million cubic meters more than the HZY Reservoir. Therefore, when the HZY Reservoir experiences a breach and overflow, the CHB Reservoir's capacity to regulate the combined effects of the dam-break flood and inter-regional floods is limited, making it difficult to withstand high-intensity superimposed floods. This could easily trigger a chain reaction of flood overflow risks, further amplifying the overall risk to the basin.
Claims
1. A method for quantifying the overtopping risk of cascade reservoir groups dams based on the CRORM model, characterized in that, Includes the following steps: S1: Obtain typical flood processes, water level-storage capacity relationships, discharge capacity, dam crest elevation, dam type, initial regulating water level, and inter-reservoir flood data for each level of reservoir in the cascade reservoir group; determine the natural inflow peak flow, water level-storage capacity relationship error, discharge correction coefficient, and initial regulating water level as random variables for overtopping risk calculation; and set failure modes according to dam type, such as overtopping of earth-rock dam (i.e., determining the occurrence of breach and formation of a breach) and overtopping of concrete dam (i.e., flow passing through but the breach not developing); S2, random samples are generated by Latin hypercube sampling based on the probability distribution of each random variable; the typical flood process is amplified by the same ratio based on the ratio of the extracted natural inflow peak flow to the typical flood peak flow to obtain the natural inflow flood process corresponding to each random sample. S3. Substitute each random sample into the water balance equation to perform flood control calculations for each time period, calculate the reservoir capacity, reservoir water level and outflow of each level of reservoir, and determine whether overtopping will occur based on the relationship between the highest reservoir water level and the dam crest elevation or the relationship between the cumulative value of net inflow flood and the initial flood control reservoir capacity. When the upstream reservoir is an earth-rock dam and overtopping occurs, the overtopping is treated as a breach and the DB-IWHR model is used to calculate the breach development and dam-break flood process; when the upstream reservoir is a concrete dam and overtopping occurs, the broad-crested weir flow formula is used to calculate the overtopping flow process, and the breach does not develop. S4. Construct the downstream reservoir inflow flood process according to the unidirectional propagation direction of the flood: Before the upstream reservoir overflows or the dam breaks, superimpose the normal outflow flood from the upstream reservoir with the inter-regional flood; after the upstream reservoir dam breaks, superimpose the upstream reservoir dam break flood with the inter-regional flood; continue to substitute the downstream reservoir inflow flood process into the flood regulation calculation and perform overflow judgment; S5. Repeat steps S2-S4 until the preset number of simulations is reached. Count the number of upstream reservoir overflow samples and the number of downstream reservoir overflow failure samples under the two conditions of upstream reservoir overflow and no overflow. Calculate the upstream reservoir overflow failure probability and the downstream reservoir overflow failure probability under the conditions of upstream reservoir overflow and no overflow. S6, multiply the two probabilities of overtopping failure under the above conditions by the probabilities of overtopping and not overtopping of the upstream reservoir respectively, and add them together to obtain the overall probability of overtopping failure of the next cascade reservoir; when the upstream reservoir is an earth-rock dam, the state in which it overtops is recorded as the overtopping failure state. S7. Take the next-level reservoir as the current-level reservoir, and combine its outflow process with the flood in the adjacent section as the inflow flood of the next-next-level reservoir. Repeat steps S3-S6 along the water flow direction until the overtopping failure probability calculation of each reservoir in the cascade reservoir group is completed, and obtain the quantitative results of risk transmission, superposition or blockage.
2. The method for quantifying the overtopping risk of cascade reservoir group dams according to claim 1, characterized in that, The probability distribution of the random variable is determined as follows: The natural inflow peak discharge is characterized by a P-III type distribution, with its probability density function, shape parameter, scale parameter, and location parameter as follows: (1); (2); (3); (4); In the formula, x is the natural inflow peak flow. , For gamma function, , and These are the shape, scale, and location parameters of the P-III type distribution. This is the average of the peak flow series. The coefficient of variation is... The skewness coefficient is used, the starting water level adopts a uniform distribution within the given upper and lower limits, the water level-reservoir capacity relationship error adopts a normal distribution with a mean of 0, and the discharge correction coefficient adopts a normal distribution with a mean of 1.
3. The method for quantifying the overtopping risk of cascade reservoir group dams according to claim 1, characterized in that, The water level-reservoir capacity relationship and the discharge flow rate are determined by the following formula: (5); (6); (7); (8); (9); In the formula, For reservoir Real-time storage capacity For reservoir Water level at any time; The relationship between water level and reservoir capacity. , and , For water levels and reservoir capacity at adjacent topographic control points; This is the actual water level-reservoir capacity relationship function. For the ideal design water level-reservoir capacity relationship function, This refers to the storage capacity error parameter; It is the outflow. It refers to the reservoir's water storage capacity. This is the discharge correction factor; The curve showing the relationship between reservoir water level and discharge capacity is shown in m. w B is the spillway flow coefficient. a For the net width of the flood discharge channel, H a Where g is the elevation of the spillway crest, and g is the acceleration due to gravity. for The inverse function of is assumed to follow a normal distribution.
4. The method for quantifying the overtopping risk of cascade reservoir group dams according to claim 1, characterized in that, The time-by-time flood control calculation, flood volume calculation, and overtopping determination are performed according to the following formula: (10); (11); (12); (13); (14); (15); In the formula, and They are respectively and The amount of water stored at any given time. and The calculation period is divided into inbound and outbound flow processes. ;Qᵢ n Q o ᵤ t ΔQ and V represent the cumulative inflow, cumulative outflow, and cumulative net inflow flood volume, respectively. r (T0) represents the initial flood control capacity; when the cumulative net inflow of floodwater at any time reaches or exceeds the initial flood control capacity, it is determined that the reservoir has overflowed.
5. The method for quantifying the overtopping risk of cascade reservoir group dams according to claim 1, characterized in that, In Monte Carlo simulations, the overtopping discrimination function, the estimated overtopping failure probability, and the number of simulations are determined by the following formula: (16); (17); (18); In the formula, H dam This is the elevation of the dam crest. This is the highest reservoir water level obtained through time-by-time flood control calculations; A flooding failure is determined when the function value is not greater than 0; N is the total number of samplings, M is the number of flooding failure samples, and P̂ f This is the estimated probability of failure at the top of the roof.
6. The method for quantifying the overtopping risk of cascade reservoir group dams according to claim 1, characterized in that, The flow rate and velocity at the breach section in the DB-IWHR model are determined by the following formula: (19); (20); (21); (22); In the formula, m b m q λ represents the flow coefficient and lateral contraction coefficient of the broad-crested weir, respectively; H is the upstream water level of the reservoir upstream of the breach; z is the elevation of the bottom of the breach; B is the weir crest width; C is the comprehensive flow coefficient; h is the actual water depth at the breach cross-section; v is the flow velocity at the breach cross-section; and λ is the flow velocity at the breach cross-section. h It is the ratio of total head to actual water depth.
7. The method for quantifying the overtopping risk of cascade reservoir group dams according to claim 1, characterized in that, The continuous water level-storage capacity relationship and water balance relationship in the DB-IWHR model are determined by the following formula: (23); (24); (25); (26); (27); (28); In the formula, W is the reservoir capacity, H is the reservoir water level elevation, and H r Using the reference water level as the baseline, p1, p2, and p3 are the fitting coefficients of the water level-reservoir capacity relationship curve, respectively. To calculate the change in reservoir capacity over a period of time, This represents the change in reservoir water level over a given period of time. To calculate the time step, q ru The inflow rate is m, where m is the flow coefficient, which is dimensionless. Based on the water level-storage capacity relationship and mass conservation relationship, the reservoir water level, breach velocity, and breach flow are calculated hourly to obtain the dam-break flood process of the upstream earth-rock dam.
8. The method for quantifying the overtopping risk of cascade reservoir group dams according to claim 1, characterized in that, For a series system consisting of upstream reservoir A and downstream reservoir B, the overtopping failure probabilities of the upstream and downstream reservoirs are determined by the following formula: (29); (30); (31); (32); In the formula, and This indicates that upstream reservoir A and downstream reservoir B have experienced overtopping failures, respectively. The upstream reservoir A did not experience overtopping failure. This represents the number of overtopping failure samples from upstream reservoir A. This represents the number of samples from both upstream and downstream reservoirs that experienced overtopping failures. The sample size is the number of upstream reservoir A that has not overflowed while downstream reservoir B has overflowed. The overall overflow failure probability of downstream reservoir B is obtained by weighting the conditional probabilities of upstream reservoir A under the two states of overflow and non-overflow.
9. The method for quantifying the overtopping risk of cascade reservoir group dams according to claim 1, characterized in that, The inflow flood process of downstream reservoir B is constructed according to the state of upstream reservoir A as follows: (33); (34); In the formula, This refers to the inflow flood process into downstream reservoir B. This is a normal flood discharge process from upstream reservoir A. This is a regional flood process. This refers to the flood event caused by the breach of an upstream dam. and These are the times when the upstream reservoir dam breaks and when the flood process ends, respectively; normal outflow flood and dam-break flood are calculated according to different time stages and are not superimposed at the same time.
10. The method for quantifying the overtopping risk of cascade reservoir group dams according to any one of claims 1 to 9, characterized in that, The CRORM model is implemented by a computer program, and executes the following steps sequentially from upstream to downstream: flood process generation, random variable sampling, flood regulation calculation, overtopping discrimination, dam-break flood calculation, downstream inflow flood combination, and failure probability statistics. For cascade reservoir groups of three or more levels, the normal outflow flood or dam-break flood of the previous level reservoir is combined with the flood of the adjacent section as the inflow flood of the next level reservoir, and the next level reservoir is taken as the current cascade reservoir. Steps S3-S6 are repeated until the overtopping risk quantification of all cascade reservoirs is completed. The risk transmission quantification results under different operating conditions are obtained by changing at least one parameter among the following: flood discharge capacity, inflow flood scale, starting water level or flood regulation capacity, downstream flood discharge capacity, and early warning and pre-discharge time.
