Method for long-term performance prediction of high core rockfill dam and safety evaluation method thereof

Abstract

The application discloses a long-term performance prediction method and a safety evaluation method for high core wall rockfill dams, comprising: obtaining multi-scale high core wall rockfill dam surface deformation displacement monitoring data based on the Beidou navigation positioning system, space-based radar interferometry and unmanned aerial vehicle LiDAR; establishing a time sequence displacement three-dimensional fusion model of space-time integration coupling dynamic filtering according to the multi-scale high core wall rockfill dam surface deformation displacement monitoring data; jointly using large-scale geotechnical centrifuge testing, large-scale indoor material testing, large-scale fine numerical simulation, prototype monitoring and inversion analysis of field test data to study the deformation mechanism of the time sequence displacement three-dimensional fusion model under different stress conditions; generating a prediction model according to the deformation mechanism of the time sequence displacement three-dimensional fusion model under different stress conditions, and predicting the long-term performance of the high core wall rockfill dam by using the prediction model. The application can improve the safety guarantee level of high dams and large reservoirs.

Description

Technical Field

[0001] This invention relates to the field of high core rockfill dam technology, specifically to a method for predicting the long-term performance of high core rockfill dams and a method for evaluating their safety. Background Technology

[0002] High-core rockfill dams are one of the main types of high dams and large reservoirs both domestically and internationally. Statistics show that more than ten earth-core rockfill dams with heights exceeding 200 meters have been built abroad, the tallest being the Nurek Dam in the Soviet Union at 300 meters. With the rapid development of the national economy and the in-depth implementation of policies promoting efficient water resource utilization and green energy, my country's water conservancy and hydropower projects have developed rapidly, leading to an increasing number of high dams and large reservoirs, and the development of high-core rockfill dams has been rapid. The most representative earth-core rockfill dam before 2000 was the Xiaolangdi earth-core rockfill dam, built on a thick overburden layer, with a maximum height of 160 meters. In the 21st century, my country has successively built high earth-core rockfill dams such as the Pubugou (dam height 186 meters, maximum overburden thickness 78 meters), Nuozhadu (dam height 261.5 meters), Maoergai (dam height 186 meters, maximum overburden thickness approximately 50 meters), and Changheba (dam height 240 meters, maximum overburden thickness 50 meters). The Lianghekou core-wall rockfill dam, currently under construction, is 295 meters high; the Shuangjiangkou core-wall rockfill dam is 314 meters high. The planned Rumei core-wall rockfill dam is 315 meters high. These high core-wall rockfill dams, reaching 200-300 meters in height, are typical high dams and large reservoirs, playing an irreplaceable role in the effective regulation of water resources and ensuring the sustainable development of the national economy; however, their failure would have catastrophic consequences. Therefore, ensuring their safe operation is of paramount importance.

[0003] However, in actual operation, many projects have encountered problems. Many completed high-core rockfill dams have developed defects such as dam body cracks and excessive seepage, endangering dam safety. International examples include the Masjed-E-Soleyman Dam (177m) in Iran, the La Grande 2 Dam (160m) in Canada, and the Cougar Dam (158m) in the United States. Domestic examples include the Xiaolangdi, Pubugou, and Maoergai high-core rockfill dams, which have experienced problems such as large localized deformation, uneven deformation, and dam crest cracks after water impoundment. To ensure the long-term safety of high-core rockfill dams, it is urgent to develop scientific and reliable methods for long-term performance prediction and safety evaluation of these dams. Summary of the Invention

[0004] The purpose of this invention is to provide a method for predicting the long-term performance of high core rockfill dams and a method for evaluating their safety, so as to improve the safety assurance level of high dams and large reservoirs.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0006] This invention provides a method for predicting the long-term performance of high-core rockfill dams, comprising:

[0007] S1: Based on the BeiDou navigation and positioning system, space-based radar interferometry, and UAV LiDAR, acquire multi-scale surface deformation and displacement monitoring data of high core rockfill dams;

[0008] S2: Based on the surface deformation and displacement monitoring data of the multi-scale high-core rockfill dam, a three-dimensional fusion model of temporal displacement with integrated air-space-ground coupling spatiotemporal dynamic filtering is established.

[0009] S3: By combining large-scale geotechnical centrifuge testing, large-scale indoor material testing, large-scale refined numerical simulation, prototype monitoring and inversion analysis of field test data, the deformation mechanism of the time-series displacement three-dimensional fusion model under different stress conditions is studied;

[0010] S4: Based on the deformation mechanism under different stress conditions of the time-series displacement three-dimensional fusion model, a prediction model is generated, and the prediction model is used to predict the long-term performance of the high-core rockfill dam.

[0011] Optionally, the multi-scale high-core rockfill dam surface deformation and displacement monitoring data includes first monitoring data from the BeiDou navigation and positioning system, second monitoring data from UAV LiDAR, and third monitoring data from space-based radar interferometry.

[0012] S2 includes:

[0013] The first monitoring data was preprocessed using a BeiDou-based ground-based augmentation method to remove errors in the time-series displacement data of a high-core rockfill dam, resulting in time-series displacement data.

[0014] The deformation of various characteristic markers at different locations on the dam surface in the second monitoring data was obtained by using a high-precision deformation monitoring method for key components of a high-center wall rockfill dam based on high-density point clouds.

[0015] Based on the third monitoring data, a slow deformation inversion model of a high-core rockfill dam based on time-series InSAR is generated, and the optimal deformation equation is obtained based on the slow deformation inversion model of a high-core rockfill dam based on time-series InSAR.

[0016] The Kalman spatiotemporal dynamic filtering model is used to fuse the time-series displacement data, the deformation of various feature markers at different locations on the dam surface, and the optimal deformation equation, to obtain a three-dimensional fusion model of time-series displacement with integrated air-space-ground spatiotemporal dynamic filtering.

[0017] Alternatively, the method for eliminating time-series displacement data errors of high-core rockfill dams based on BeiDou ground-based augmentation includes:

[0018] Based on the first monitoring data, the geometric distance between the station and the satellite is obtained using the observation equation; wherein, the geometric distance includes relativistic effects, solid tides, antenna phase center, and phase rotation error;

[0019] The geometric distance is linearized to obtain the error equation;

[0020] The error equation is used to eliminate the time-series displacement data error of the high-core rockfill dam based on BeiDou ground-based augmentation, and the first monitoring data is obtained.

[0021] Alternatively, the observation equation is:

[0022]

[0023] The error equation is:

[0024] V = AδX + W

[0025] Among them, l p and dt and dT are the ionosphere-free combined pseudorange and phase observations, respectively; dt and dT are the receiver and satellite clock errors, respectively; c is the speed of light in vacuum; amb is the ionosphere-free combined ambiguity; m is the tropospheric projection function; ZTD is the zenith tropospheric delay; ε p and These represent observation noise and multipath error, respectively, where ρ is the geometric distance between the station and the satellite, including relativistic effects, solid tides, antenna phase center and phase rotation errors. A is the coefficient matrix, δX is the parameter to be estimated, including the three-dimensional coordinates of the station, receiver clock error, ionospheric ambiguity, and tropospheric zenith wet delay, and W is the constant term matrix.

[0026] Alternatively, high-precision deformation monitoring methods for key components of high-center-wall rockfill dams based on high-density point clouds include:

[0027] Multiple aerial photographs of the dam were taken using UAV lidar, resulting in several aerial images.

[0028] Based on the aforementioned aerial images, construct three-dimensional models of the dam at different times with centimeter-level precision.

[0029] Surface feature markers in the aforementioned aerial images were identified using deep learning image recognition technology.

[0030] Based on the three-dimensional models of the dam with centimeter-level precision at different times and the surface feature markers, the deformation of each feature marker at different locations on the dam surface is calculated.

[0031] Alternatively, the Kalman spatiotemporal dynamic filtering model is a spatiotemporal Kalman filtering model that combines Kriging interpolation with Kalman filtering.

[0032] The basic principle of the Kalman spatiotemporal dynamic filtering model is as follows:

[0033]

[0034] Among them, Z t (S) represents the observed value at position s at time t, a t Let ε be a state variable. t (S)~N(0,R1) represents the observation noise, Φ is the state transition matrix, and w t ~N(0, Q1) is the state noise, H is the spatial field describing spatial correlation, and H(S)' is the transpose of H(S).

[0035] Alternatively, S3 includes:

[0036] S31: Using centrifugal simulation technology, the evolution of unsaturated core wall pore pressure and deformation process under the condition of gradually increasing overburden load during the construction period of the rockfill dam;

[0037] S32: Conduct two-dimensional and three-dimensional hypergravity model tests with different modulus ratios of core wall and dam shell materials;

[0038] S33: Conduct centrifuge model tests on the interaction and deformation characteristics of the core wall and dam shell material under different water levels;

[0039] S34: Conduct centrifuge model tests on core-wall rockfill dam cracks and seismic failure modes.

[0040] Optionally, S31 includes: using a large geotechnical centrifuge to reveal the evolution process and deformation evolution law of the unsaturated core wall pore pressure of the centrifugal model during water level rise and fall;

[0041] Using a centrifuge vibration table, the failure mode of unsaturated core wall soil under seismic wave action was simulated to explore the evolution process of pore pressure and deformation evolution law of unsaturated core wall under seismic wave action.

[0042] A scaled-down model test of a highly saturated and unsaturated high-core rockfill dam under different stress states, involving wet-dry cycles, was conducted to reveal the evolution and mechanism of pore pressure evolution and deformation response of the unsaturated core wall. The test included: identifying the target high-core rockfill dam, selecting a suitable scale, and preparing a model. Tensiometers were embedded in the model to monitor changes in matrix suction within the core wall, earth pressure cells were embedded to monitor changes in earth pressure within the core wall, and laser displacement sensors were installed to monitor surface deformation of the rockfill dam. Wet-dry cycle conditions were controlled by the water inlet and outlet of the test chamber, connected by a hose and pump to control the water level. The test focused on the influence of different stress states on the deformation of the unsaturated core wall, the influence of wet-dry cycles on changes in pore pressure within the core wall, and the effect of deformation-pore pressure coupling.

[0043] S33 includes:

[0044] The simulation object of the high core wall rockfill dam was determined, and the model core wall was made using different moisture contents and gravel content. The model core wall was proportioned according to the common conditions of high core wall rockfill dams.

[0045] Thin-film pressure sensors were arranged at the bottom of the model core wall and on the bank slope. Pore pressure sensors, tension meters, and miniature TDRs were arranged along the elevation inside the core wall and on the dam axis. Pore pressure sensors were arranged along the upstream and downstream directions at several typical elevations on the dam cross section to obtain the model system.

[0046] The model system was gradually loaded in multiple stages until the centrifugal force reached about 200g. After each stage of loading was completed, the system was stabilized for a period of time, and the changes in soil pressure, pore pressure, and suction under each stage of load were recorded.

[0047] A membrane bag is placed upstream of the model system, and water is injected into the membrane bag to the design water level. The centrifuge is then loaded again until the preset centrifugal acceleration stabilizes. Changes in the seepage pressure and other observation indicators in the core wall are monitored. After the seepage pressure, suction and other indicators in the core wall tend to stabilize, the membrane bag is punctured and water is injected from upstream. Changes in various indicators in the model are monitored.

[0048] S34 includes:

[0049] Using the Xiaolangdi core wall rockfill dam as a prototype, a large-scale test model was made using a large geotechnical centrifuge to simulate construction, initial water storage, operation, and rapid water level changes. This model was used to reproduce and study the dynamic process of crack occurrence and development, so as to obtain complete key data on seepage, deformation, and stress, and provide a reliable basis and verification for long-term performance prediction and safety evaluation.

[0050] Using the Lianghekou core-wall rockfill dam as a prototype, a large-scale test model was constructed using a large geotechnical centrifuge and centrifuge shaking table. This model simulated the dam's acceleration response and seismic deformation under different levels of earthquakes, tested the dam's ultimate seismic resistance, studied the seismic failure mode and dynamic evolution process, and obtained complete seismic motion and seismic response data. This provides a basis and verification for the long-term performance prediction and safety evaluation of high core-wall rockfill dams with both static and dynamic safety.

[0051] Alternatively, the large geotechnical centrifuge monitors changes in matrix suction within the core wall by embedding tension gauges, monitors changes in soil pressure within the core wall by embedding soil pressure cells, monitors surface deformation of the rockfill dam by installing laser displacement sensors, and acquires digital images through a high-definition video monitoring system.

[0052] This invention also provides a safety evaluation method based on the above-mentioned long-term performance prediction method for high-core rockfill dams, the safety evaluation method comprising:

[0053] Based on the prediction results of the long-term performance of high core wall rockfill dams, and considering normal and emergency operating conditions during long-term operation, the evolution law of the long-term operating performance of dam construction materials is studied.

[0054] Based on the evolutionary patterns and the results of simulation analysis of actual engineering defects, safety evaluation indicators corresponding to the monitoring values ​​are obtained;

[0055] Safety evaluation is conducted using the aforementioned safety evaluation indicators.

[0056] The present invention has the following beneficial effects:

[0057] This invention considers both normal and extraordinary operating conditions (floods, earthquakes, sudden rises and falls in water levels, etc.) during long-term operation, establishes a predictive model, and constructs a long-term operational performance prediction and safety evaluation system based on deformation and deformation coordination dual control seepage and deformation multi-coupling, and static and dynamic dual safety. This system enables breakthroughs in supporting technologies, reveals core mechanisms, and advances in theoretical methods, significantly improving the accuracy and reliability of safety evaluation and enhancing the safety assurance level of high dams and large reservoirs. Attached Figure Description

[0058] Figure 1 This is a flowchart of the long-term performance prediction method for high core wall rockfill dams according to the present invention. Detailed Implementation

[0059] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0060] This invention provides a method for predicting the long-term performance of high-core rockfill dams, with reference to... Figure 1 As shown, it includes:

[0061] S1: Based on the BeiDou navigation and positioning system, space-based radar interferometry, and UAV LiDAR, acquire multi-scale surface deformation and displacement monitoring data of high core rockfill dams;

[0062] The multi-scale surface deformation and displacement monitoring data of the high-core rockfill dam of this invention includes first monitoring data from the BeiDou navigation and positioning system, second monitoring data from UAV LiDAR, and third monitoring data from space-based radar interferometry.

[0063] S2: Based on the surface deformation and displacement monitoring data of the multi-scale high-core rockfill dam, a three-dimensional fusion model of temporal displacement with integrated air-space-ground coupling spatiotemporal dynamic filtering is established.

[0064] Based on this, S2 includes:

[0065] S21: The first monitoring data is preprocessed using the error removal method for time-series displacement data of high-core rockfill dams based on BeiDou ground-based augmentation to obtain time-series displacement data;

[0066] Here, the error removal method for time-series displacement data of high-core rockfill dams based on BeiDou ground-based augmentation includes:

[0067] Based on the first monitoring data, the geometric distance between the station and the satellite is obtained using the observation equation; wherein, the geometric distance includes relativistic effects, solid tides, antenna phase center, and phase rotation error;

[0068] The geometric distance is linearized to obtain the error equation;

[0069] The observation equation is:

[0070]

[0071] The error equation is:

[0072] V = AδX + W

[0073] Among them, l p and dt and dT are the ionosphere-free combined pseudorange and phase observations, respectively; dt and dT are the receiver and satellite clock errors, respectively; c is the speed of light in vacuum; amb is the ionosphere-free combined ambiguity; m is the tropospheric projection function; ZTD is the zenith tropospheric delay; ε p and These represent observation noise and multipath error, respectively, where ρ is the geometric distance between the station and the satellite, including relativistic effects, solid tides, antenna phase center and phase rotation errors. A is the coefficient matrix, δX is the parameter to be estimated, including the three-dimensional coordinates of the station, receiver clock error, ionospheric ambiguity, and tropospheric zenith wet delay, and W is the constant term matrix.

[0074] The error equation is used to eliminate the time-series displacement data error of the high-core rockfill dam based on BeiDou ground-based augmentation, and the first monitoring data is obtained.

[0075] S22: Use a high-precision deformation monitoring method for key components of a high-center wall rockfill dam based on high-density point cloud to obtain the deformation of various characteristic markers at different locations on the dam surface in the second monitoring data;

[0076] High-precision deformation monitoring methods for key components of high-center-wall rockfill dams based on high-density point clouds include:

[0077] Multiple aerial photographs of the dam were taken using UAV lidar, resulting in several aerial images.

[0078] Based on the aforementioned aerial images, construct three-dimensional models of the dam at different times with centimeter-level precision.

[0079] Surface feature markers in the aforementioned aerial images were identified using deep learning image recognition technology.

[0080] Based on the three-dimensional models of the dam with centimeter-level precision at different times and the surface feature markers, the deformation of each feature marker at different locations on the dam surface is calculated.

[0081] S23: Based on the third monitoring data, generate a slow deformation inversion model of a high-core rockfill dam based on time-series InSAR, and obtain the optimal deformation equation based on the slow deformation inversion model of a high-core rockfill dam based on time-series InSAR.

[0082] Settlement monitoring of a high-core rockfill dam area was studied using the Small Baseline Set (SBAS) differential interferometry method. The SBAS technique sets thresholds for spatial and temporal baselines, and pairs SAR images into short-term temporal and spatial baseline interferometric pairs. Then, based on the minimum norm criterion of deformation rate, the singular value decomposition (SVD) method is used to obtain the small deformations of coherent targets along the radar line of sight. This method can reduce the influence of spatiotemporal decorrelation and obtain more accurate long-term series surface deformation.

[0083] The SBAS method combines all acquired SAR data into several sets, based on the principle that SAR images within a set have small baseline distances, while SAR images between sets have large baseline distances. The surface deformation time series for each small set can be easily obtained using the LS method, but it requires increasing the time sampling frequency within a single set to solve the deformation equation optimally.

[0084] S24: The Kalman spatiotemporal dynamic filtering model is used to fuse the time-series displacement data, the deformation of each feature marker at different locations on the dam surface, and the optimal deformation equation to obtain a three-dimensional fusion model of time-series displacement with integrated air-space-ground spatiotemporal dynamic filtering.

[0085] To explore the three-dimensional deformation evolution characteristics of high spatiotemporal resolution high-core rockfill dams, it is necessary to acquire multi-scale deformation data using multi-source data such as InSAR, BeiDou, and UAV LiDAR, and to achieve temporal displacement fusion using the Kalma spatiotemporal dynamic hidden wave model. The specific principle is as follows:

[0086] The Kalman spatiotemporal dynamic filtering model is a spatiotemporal Kalman filtering model that combines Kriging interpolation with Kalman filtering. By inputting InSAR spatial deformation and BeiDou point deformation data, it can couple and calculate the temporal displacement of a high-center-wall rockfill dam in three-dimensional space.

[0087] The basic principle of the Kalman spatiotemporal dynamic filtering model is as follows:

[0088]

[0089] Among them, Z t (S) represents the observed value at position s at time t, a t Let ε be a state variable. t (S)~N(0,R1) represents the observation noise, Φ is the state transition matrix, and w t ~N(0, Q1) is the state noise, H is the spatial field describing spatial correlation, and H(S)' is the transpose of H(S).

[0090] S3: By combining large-scale geotechnical centrifuge testing, large-scale indoor material testing, large-scale refined numerical simulation, prototype monitoring, and inversion analysis of field test data, the deformation mechanism of the time-series displacement three-dimensional fusion model under different stress conditions is studied.

[0091] Alternatively, S3 includes:

[0092] S31: Using centrifugal simulation technology, the evolution of unsaturated core wall pore pressure and deformation process under the condition of gradually increasing overburden load during the construction period of the rockfill dam;

[0093] S31 includes: using a large geotechnical centrifuge to reveal the evolution process and deformation evolution law of the unsaturated core wall pore pressure of a centrifugal model during water level rise and fall;

[0094] Using a centrifuge vibration table, the failure mode of unsaturated core wall soil under seismic wave action was simulated to explore the evolution process of pore pressure and deformation evolution law of unsaturated core wall under seismic wave action.

[0095] To reveal the evolution of pore pressure and deformation response in unsaturated core walls, this invention utilizes a scaled-down model test of a highly saturated and unsaturated high-core rockfill dam under different stress states, employing wet-dry cycles, to demonstrate the evolution and mechanism of pore pressure and deformation response in unsaturated core walls. The wet-dry cycle scaled-down model test of the highly saturated and unsaturated high-core rockfill dam under different stress states includes: identifying the target high-core rockfill dam, selecting a reasonable scale, preparing a high-core rockfill dam model, monitoring changes in matrix suction within the core wall by embedding tension gauges in the model, monitoring changes in soil pressure within the core wall by embedding earth pressure cells, and monitoring surface deformation of the rockfill dam by installing laser displacement sensors; controlling the wet-dry cycle conditions through the water inlet and outlet of the test chamber, with the inlet and outlet connected by a hose and pump to control the water level, focusing on the influence of different stress states on unsaturated core wall deformation, the influence of wet-dry cycles on core wall pore pressure changes, and the influence of deformation-pore pressure coupling.

[0096] S32: Conduct two-dimensional and three-dimensional hypergravity model tests with different modulus ratios of core wall and dam shell materials;

[0097] Based on the large-scale geotechnical centrifuge test platform, two-dimensional and three-dimensional hypergravity model tests were conducted with different modulus ratios of core wall and dam shell materials. The deformation and stress distribution laws of each zone inside the dam body under different modulus ratios were studied. The influence mechanism of modulus differences in each zone on the stress and deformation of the dam body was explored, and a quantitative calculation method for analyzing the influence of modulus ratio on the stress and deformation of the dam body was established.

[0098] S33: Conduct centrifuge model tests on the interaction and deformation characteristics of the core wall and dam shell material under different water levels;

[0099] S33 includes:

[0100] The simulation object of the high core wall rockfill dam was determined, and the model core wall was made using different moisture contents and gravel content. The model core wall was proportioned according to the common conditions of high core wall rockfill dams.

[0101] This invention selects the following three formulation schemes:

[0102] Serial Number describe Clay moisture content (%) Gravel content (%) M01 Dry + High Gravel Core Wall optimum moisture content -1 45 M02 Moist core wall with high gravel content Optimal moisture content +1.5 45 M03 Moist core wall with low gravel content Optimal moisture content +1.5 15

[0103] Thin-film pressure sensors were arranged at the bottom of the model core wall and on the bank slope. Pore pressure sensors, tension meters, and miniature TDRs were arranged along the elevation inside the core wall and on the dam axis. Pore pressure sensors were arranged along the upstream and downstream directions at several typical elevations on the dam cross section to obtain the model system.

[0104] The model system was gradually loaded in multiple stages until the centrifugal force reached about 200g. After each stage of loading was completed, the system was stabilized for a period of time, and the changes in soil pressure, pore pressure, and suction under each stage of load were recorded.

[0105] A membrane bag is placed upstream of the model system, and water is injected into the membrane bag to the design water level. The system is then loaded again in a centrifuge until the preset centrifugal acceleration stabilizes. Changes in indicators such as the seepage pressure inside the core wall are monitored. Once the seepage pressure and suction inside the core wall have stabilized, the membrane bag is punctured and water is injected from upstream. Changes in various indicators in the model are then monitored.

[0106] S34: Conduct centrifuge model tests on core-wall rockfill dam cracks and seismic failure modes.

[0107] S34 includes:

[0108] Using the Xiaolangdi core wall rockfill dam as a prototype, a large-scale test model was made using a large geotechnical centrifuge to simulate construction, initial water storage, operation, and rapid water level changes. This model was used to reproduce and study the dynamic process of crack occurrence and development, so as to obtain complete key data on seepage, deformation, and stress, and provide a reliable basis and verification for long-term performance prediction and safety evaluation.

[0109] Using the Lianghekou core-wall rockfill dam as a prototype, a large-scale test model was constructed using a large geotechnical centrifuge and centrifuge shaking table. This model simulated the dam's acceleration response and seismic deformation under different levels of earthquakes, tested the dam's ultimate seismic resistance, studied the seismic failure mode and dynamic evolution process, and obtained complete seismic motion and seismic response data. This provides a basis and verification for the long-term performance prediction and safety evaluation of high core-wall rockfill dams with both static and dynamic safety.

[0110] The large-scale geotechnical centrifuge of this invention monitors the changes in matrix suction within the core wall by embedding a tension meter, monitors the changes in soil pressure within the core wall by embedding a soil pressure cell, monitors the surface deformation of the rockfill dam by installing a laser displacement sensor, and acquires digital images through a high-definition video monitoring system.

[0111] S4: Based on the deformation mechanism under different stress conditions of the time-series displacement three-dimensional fusion model, a prediction model is generated, and the prediction model is used to predict the long-term performance of the high-core rockfill dam.

[0112] The fit of a prediction model depends on the ability of its factors to describe the actual problem. In terms of model construction and factor composition, the candidate factors for the prediction model consist of the influence components of the main environmental factors. It should include all important factors and exclude irrelevant ones. Specifically, the selection should be based on the load characteristics of the dam structure, material properties, geological conditions of the dam foundation, and construction conditions, guided by dam engineering theory.

[0113] The prediction model is:

[0114] δ(t)=δ H (t)+δ T (t)+δ θ (t)+δP (t)+δ C (t)

[0115] Among them, the model factors include the water level component δ H Factor selection and temperature component δ T The selection of factors for water level components includes upstream water level factor, downstream water level factor, previous water level factor, water level rise and fall rate factor, water level holding load factor, initial storage factor, and finite element water level factor.

[0116] As a water-retaining structure, the water pressure generated by the upstream water level of a dam is the primary cause of deformation in the dam body and foundation. This component is generally expressed as a polynomial of the water level: Where H(t) is the upstream water level depth, it is generally normalized using the maximum water level variation. Considering the spatial effect of water pressure, the power i of H(t) in the above formula can be 3 to 4.

[0117] When the downstream water level changes significantly, the displacement of the dam body and dam foundation is also affected by the downstream water level and pressure. The downstream water depth factor is generally expressed using a polynomial, with the highest power being the fourth power of the downstream water level.

[0118] Considering that uplift pressure affecting displacement in concrete dams, seepage pressure in earth-rock dams, and water temperature lag behind water level changes, the influence of the previous water level history should be taken into account, such as using the average water level over a previous period as the previous water level factor. Generally, multiple H values ​​are used. i-j The term is used to represent the lag effect, where i represents the duration of the lag effect, and j is the number of days prior to the average water level taken before the lag effect. In practical applications, several sets of i and j values ​​are often selected, such as i being 0 and j being 3, 7, 15, 30, 60, 120, etc., to represent the average water level 3, 7, 15, 30, 60, 120 days before the observation date; or i being 15 and j being 15 or 30 to represent the average water level 15 days before the observation date, and then 15 or 30 days prior.

[0119] For earth-rock dams, the rate of rise and fall of the water level upstream of the dam has a certain impact on the deformation of the dam body. The daily water level change rate and the square of the rate can be selected as the water level rise and fall rate factor. According to engineering practice, under the action of a sustained high water level, the deformation of dams (arch dams, gravity dams, earth-rock dams) will continue to increase with the time of action, producing an effect similar to creep. The interaction term of water level and time can be selected as a factor. If a more complex water level-time relationship is considered, an appropriate function of water level and time (such as a convolution integral) can also be added as a water level holding factor.

[0120] Since the deformation pattern of an earth-rock dam differs between the initial water level rise and subsequent re-flooding to the same level, the influence of the initial water level factor should be considered. The initial water level factor mainly includes the portion h of the new high water level exceeding the original highest water level when the water level is reached.D And consider the interaction between the excess portion and the water level itself.

[0121] Temperature components primarily reflect the periodic and seasonal annual variations in displacement, while also reflecting the impact of different actual temperature processes each year on displacement. The displacement of rockfill dams is less affected by the temperature field than that of concrete dams; however, in frigid regions, frost heave caused by sub-zero temperatures shows a significant correlation with measured temperatures. Since some dam sections contain concrete structures, the temperature effect should also be considered in their displacement. The following five types of temperature factors can be considered:

[0122] Temperature periodicity factor, air temperature factor and previous air temperature factor, measuring point temperature factor, convolution integral regression factor and frost heave factor.

[0123] Considering the annual periodicity of the temperature component (seasonal component), a periodic factor based on a trigonometric function of time is chosen to describe the common parts of the temperature component over the years. Generally, periodic factors with periods of six months or one year are selected, including sin(s), cos(s), sin²(s), and sin(s)cos(s), where s = 2πt' / 365, and t' is the length of time (in days) between the measurement date and the start date of the analysis. This factor can be used even when temperature observation data is scarce or incomplete, and practice has shown that it is effective in simulating seasonal variations.

[0124] The daily average temperature and the preceding average temperature have a direct impact on the dam's displacement. Temperature affects the boundary temperature, thus influencing the internal temperature field of the dam and consequently the temperature displacement component. The phase angle of temperature lag behind the boundary temperature varies at different locations, and the amplitude of the change generally decreases with increasing distance from the boundary. Factors such as the 1st to 3rd powers of the daily temperature and multiple values ​​of the form T are selected. i_j The term is represented by the preceding average temperature, where i represents the length of the lag effect and j is the number of days before the lag effect.

[0125] If there is a temperature measurement at or near the measuring point, the temperature at the measuring point can be selected as the 1st to 3rd power as the temperature factor of the measuring point.

[0126] To describe the influence of the interactions between water level and temperature, and temperature and time, on displacement, a "convolution integral regression factor" is established, consisting of an air-water level cross term, an air-time cross term, and an air-temperature periodic term cross term. The air-temperature term typically selects the daily air temperature, the 15-day average air temperature, or the 30-day average air temperature, forming an orthogonal polynomial with the water level, time, and periodic terms. One or more of these convolution integral regression factors can be selected based on the specific engineering conditions.

[0127] The problem of frost heave in dams in frigid regions can be addressed by setting up a set of frost heave factors, consisting of a "periodic function factor with a fixed period and a fixed initial angle" and a "temperature lag factor," to describe and simulate the frost heave phenomenon.

[0128] Regarding the time-dependent component δ θ Factor selection:

[0129] Time-dependent displacement is an irreversible quantity that develops in a certain direction over time. Due to the rheology of the rockfill, the creep of the dam foundation rock, and the influence of factors such as bedrock geological structure and dam cracks, the dam inevitably experiences displacement that changes over time, collectively referred to as time-dependent displacement. Time-dependent displacement generally has a curvilinear relationship with time, with the most commonly used factor being the linear factor t, the logarithmic formula (ln(1+t)), and the exponential formula (e^(-t)). -kt These three factors and their combinations can simulate the vast majority of time-dependent components. Sometimes, the second to third powers of time are also chosen as factors.

[0130] Furthermore, from a fitting perspective, a hyperbolic method can also be applied. S-type growth factor Multiple logarithmic exponential curves, multiple polynomials, broken line factors do not represent the non-monotonic changes in time caused by the construction period, dam reinforcement and repair construction, complex loading and unloading situations, or other reasons.

[0131] Regarding the rainfall component δ P Factor selection:

[0132] Rainfall can affect the displacement of rockfill dams, high slopes, etc., so the impact of rainfall should be considered, and the daily rainfall and previous rainfall factors should be selected. For previous rainfall, two types can be considered: total previous rainfall or average previous rainfall intensity.

[0133] Regarding the construction component δ C Factor selection:

[0134] For dams under construction, the current cross-sectional filling elevation, material zoning, and compaction density of the rockfill have a direct impact on the stress and displacement of the dam. In this case, the influence of construction components should be considered, such as using the 1st to 4th power of the filling elevation of the cross-section as a factor.

[0135] This invention also provides a safety evaluation method based on the above-mentioned long-term performance prediction method for high-core rockfill dams, the safety evaluation method comprising:

[0136] Based on the prediction results of the long-term performance of high core wall rockfill dams, S101 studies the evolution law of the long-term performance of dam construction materials, taking into account normal and emergency conditions during long-term operation.

[0137] S101 Based on the aforementioned evolutionary patterns and the results of simulation analysis of actual engineering defects, safety evaluation indicators corresponding to the monitoring values ​​are obtained;

[0138] The specific safety evaluation indicators include:

[0139] The evaluation focuses on deformation, including its total amount, degree of non-uniformity, and dynamic process. Settlement rate is used to evaluate the total deformation, deformation gradient to evaluate the degree of non-uniformity, and deformation rate and timeline to evaluate the dynamic process of deformation. Simultaneously, it integrates seismic safety evaluation concepts, considering long-term performance, and constructs a safety evaluation system based on deformation, stability, seepage prevention systems, and foundation safety.

[0140] S101 uses the aforementioned safety evaluation indicators to conduct a safety evaluation.

[0141] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting the long-term performance of a high-core rockfill dam, characterized in that, include: S1: Based on the BeiDou navigation and positioning system, space-based radar interferometry, and UAV LiDAR, acquire multi-scale surface deformation and displacement monitoring data of high core rockfill dams; S2: Based on the surface deformation and displacement monitoring data of the multi-scale high-core rockfill dam, a three-dimensional fusion model of temporal displacement with integrated air-space-ground coupling spatiotemporal dynamic filtering is established. S3: By combining large-scale geotechnical centrifuge testing, large-scale indoor material testing, large-scale refined numerical simulation, prototype monitoring and inversion analysis of field test data, the deformation mechanism of the time-series displacement three-dimensional fusion model under different stress conditions is studied; S4: Based on the deformation mechanism under different stress conditions of the time-series displacement three-dimensional fusion model, a prediction model is generated, and the prediction model is used to predict the long-term performance of the high-core rockfill dam. The multi-scale high-core rockfill dam surface deformation and displacement monitoring data includes the first monitoring data from the BeiDou navigation and positioning system, the second monitoring data from UAV LiDAR, and the third monitoring data from space-based radar interferometry. S2 includes: The first monitoring data was preprocessed using a BeiDou-based ground-based augmentation method to remove errors in the time-series displacement data of a high-core rockfill dam, resulting in time-series displacement data. The deformation of various characteristic markers at different locations on the dam surface in the second monitoring data was obtained by using a high-precision deformation monitoring method for key components of a high-center wall rockfill dam based on high-density point clouds. Based on the third monitoring data, a slow deformation inversion model of a high-core rockfill dam based on time-series InSAR is generated, and the optimal deformation equation is obtained based on the slow deformation inversion model of a high-core rockfill dam based on time-series InSAR. The Kalman spatiotemporal dynamic filtering model is used to fuse the time-series displacement data, the deformation of each feature marker at different locations on the dam surface, and the optimal deformation equation to obtain a three-dimensional fusion model of time-series displacement with integrated air-space-ground spatiotemporal dynamic filtering. S3 includes: S31: Using centrifugal simulation technology, the evolution of unsaturated core wall pore pressure and deformation process under the condition of gradually increasing overburden load during the construction period of the rockfill dam; S32: Conduct two-dimensional and three-dimensional hypergravity model tests with different modulus ratios of core wall and dam shell materials; S33: Conduct centrifuge model tests on the interaction and deformation characteristics of the core wall and dam shell material under different water levels; S34: Conduct centrifuge model tests on core-wall rockfill dam cracks and seismic failure modes.

2. The method for predicting the long-term performance of a high-core rockfill dam according to claim 1, characterized in that, The method for eliminating time-series displacement data errors of high-core rockfill dams based on BeiDou ground-based augmentation includes: Based on the first monitoring data, the geometric distance between the station and the satellite is obtained using the observation equation; wherein, the geometric distance includes relativistic effects, solid tides, antenna phase center, and phase rotation error; The geometric distance is linearized to obtain the error equation; The error equation is used to eliminate the time-series displacement data error of the high-core rockfill dam based on BeiDou ground-based augmentation, and the first monitoring data is obtained.

3. The method for predicting the long-term performance of a high-core rockfill dam according to claim 2, characterized in that, The observation equation is: The error equation is: in, and These are the pseudorange and phase observations for the ionosphere-free combination, respectively. and These are the receiver and the satellite clock bias, respectively. The speed of light in a vacuum. For ionosphere-free combined ambiguity, For tropospheric projection functions, For tropospheric delay in the zenith direction, and These are observation noise and multipath error, respectively. The geometric distance between the station and the satellite includes relativistic effects, solid tides, antenna phase center and phase rotation error, and , The coefficient matrix, The parameters to be estimated include the station's three-dimensional coordinates, receiver clock error, ionospheric-free combined ambiguity, and tropospheric zenith wet delay. It is a matrix of constant terms.

4. The method for predicting the long-term performance of a high-core rockfill dam according to claim 1, characterized in that, High-precision deformation monitoring methods for key components of high-center-wall rockfill dams based on high-density point clouds include: Multiple aerial photographs of the dam were taken using UAV lidar, resulting in several aerial images. Based on the aforementioned aerial images, construct three-dimensional models of the dam at different times with centimeter-level precision. Surface feature markers in the aforementioned aerial images were identified using deep learning image recognition technology. Based on the three-dimensional models of the dam with centimeter-level precision at different times and the surface feature markers, the deformation of each feature marker at different locations on the dam surface is calculated.

5. The method for predicting the long-term performance of a high-core rockfill dam according to claim 1, characterized in that, The Kalman spatiotemporal dynamic filtering model is a spatiotemporal Kalman filtering model that combines Kriging interpolation with Kalman filtering. The basic principle of the Kalman spatiotemporal dynamic filtering model is as follows: in, express t Time and location s Observations at that location For state variables, To observe the noise, Here is the state transition matrix. State noise, To describe the spatial field of spatial correlation, for The transpose of .

6. The method for predicting the long-term performance of a high-core rockfill dam according to claim 1, characterized in that, S31 includes: using a large geotechnical centrifuge to reveal the evolution process and deformation evolution law of the unsaturated core wall pore pressure of the centrifugal model during water level rise and fall; Using a centrifuge vibration table, the failure mode of unsaturated core wall soil under seismic wave action was simulated to explore the evolution process of pore pressure and deformation evolution law of unsaturated core wall under seismic wave action. A scaled-down model test of a highly saturated and unsaturated high-core rockfill dam under different stress states, involving wet-dry cycles, was conducted to reveal the evolution and mechanism of pore pressure evolution and deformation response of the unsaturated core wall. The test included: identifying the target high-core rockfill dam, selecting a suitable scale, and preparing a model. Tensiometers were embedded in the model to monitor changes in matrix suction within the core wall, earth pressure cells were embedded to monitor changes in earth pressure within the core wall, and laser displacement sensors were installed to monitor surface deformation of the rockfill dam. Wet-dry cycle conditions were controlled by the water inlet and outlet of the test chamber, connected by a hose and pump to control the water level. The test focused on the influence of different stress states on the deformation of the unsaturated core wall, the influence of wet-dry cycles on changes in pore pressure within the core wall, and the effect of deformation-pore pressure coupling. S33 includes: The simulation object of the high core wall rockfill dam was determined, and the model core wall was made using different moisture contents and gravel content. The model core wall was proportioned according to the common conditions of high core wall rockfill dams. Thin-film pressure sensors were arranged at the bottom of the model core wall and on the bank slope. Pore pressure sensors, tension meters, and miniature TDRs were arranged along the elevation inside the core wall and on the dam axis. Pore pressure sensors were arranged along the upstream and downstream directions at several typical elevations on the dam cross section to obtain the model system. The model system was gradually loaded in multiple stages until the centrifugal force reached about 200g. After each stage of loading was completed, the system was stabilized for a period of time, and the changes in soil pressure, pore pressure, and suction under each stage of load were recorded. A membrane bag is placed upstream of the model system, and water is injected into the membrane bag to the design water level. The centrifuge is then loaded again until the preset centrifugal acceleration stabilizes. Changes in the seepage pressure and other observation indicators in the core wall are monitored. After the seepage pressure, suction and other indicators in the core wall tend to stabilize, the membrane bag is punctured and water is injected from upstream. Changes in various indicators in the model are monitored. S34 includes: Using the Xiaolangdi core wall rockfill dam as a prototype, a large-scale test model was made using a large geotechnical centrifuge to simulate construction, initial water storage, operation, and rapid water level changes. This model was used to reproduce and study the dynamic process of crack occurrence and development, so as to obtain complete key data on seepage, deformation, and stress, and provide a reliable basis and verification for long-term performance prediction and safety evaluation. Using the Lianghekou core-wall rockfill dam as a prototype, a large-scale test model was constructed using a large geotechnical centrifuge and centrifuge shaking table. This model simulated the dam's acceleration response and seismic deformation under different levels of earthquakes, tested the dam's ultimate seismic resistance, studied the seismic failure mode and dynamic evolution process, and obtained complete seismic motion and seismic response data. This provides a basis and verification for the long-term performance prediction and safety evaluation of high core-wall rockfill dams with both static and dynamic safety.

7. The method for predicting the long-term performance of a high-core rockfill dam according to claim 6, characterized in that, The large geotechnical centrifuge monitors changes in matrix suction within the core wall by embedding tension gauges, monitors changes in soil pressure within the core wall by embedding soil pressure cells, monitors surface deformation of the rockfill dam by installing laser displacement sensors, and acquires digital images through a high-definition video monitoring system.

8. A safety evaluation method based on the long-term performance prediction method for high-core rockfill dams according to any one of claims 1-7, characterized in that, The safety evaluation method includes: Based on the prediction results of the long-term performance of high core wall rockfill dams, and considering normal and emergency operating conditions during long-term operation, the evolution law of the long-term operating performance of dam construction materials is studied. Based on the evolutionary patterns and the results of simulation analysis of actual engineering defects, safety evaluation indicators corresponding to the monitoring values ​​are obtained; Safety evaluation is conducted using the aforementioned safety evaluation indicators.

Images