A Dynamic Assessment Method and System for Reservoir Dam Safety Status Based on Multi-Source Data Fusion
By utilizing the seepage finite element model and reinforcement material test data in reservoir dam safety monitoring, a time-varying decay function of the permeability coefficient was established, and multi-source data fusion back analysis was performed. This solved the problem of dynamic assessment of the seepage state of the reinforced dam and enabled quantitative and physically based safety state determination.
Patent Information
- Application Number
- CN202611407774.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-11
- Publication Date
- 2026-11-10
AI Technical Summary
Existing technologies for reservoir dam safety monitoring lack long-term historical monitoring data, making it difficult to conduct continuous, quantitative, and dynamic assessments of the seepage status of reinforced dams with clear physical evidence. Furthermore, the changes in the permeability performance of reinforced structures are not fully utilized.
Using the seepage finite element model from the reinforcement design phase and the permeability test data of the reinforcement materials, a time-varying decay function of the permeability coefficient is established. Through multi-source data fusion and back analysis, the equivalent permeability coefficient of the reinforcement influence parameter zone is obtained. Combined with mass conservation constraints and image cross-validation, the safety status of the dam is determined.
Without the need for long-term historical data accumulation, dynamic and quantitative assessment of the seepage status of reinforced dams has been achieved, improving the timeliness and reliability of safety status identification and reducing the impact of single monitoring anomalies on assessment conclusions.
Smart Images

Figure CN123069786A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the technical field, specifically to a method and system for dynamic assessment of the safety status of reservoir dams based on multi-source data fusion. Background Technology
[0002] Reservoir dams are affected by factors such as reservoir water level changes, rainfall infiltration, material property evolution, and changes in local seepage channels during long-term operation, resulting in significant time-varying seepage states. Current dam safety monitoring typically obtains operational information through methods such as water level, rainfall, piezometers, weirs, dam surface inspections, or image monitoring, and combines this information with monitoring index thresholds, historical statistical patterns, or finite element analysis of seepage to determine the dam's safety status. While the level of monitoring automation has improved and multi-source data acquisition conditions are relatively well-established, for reservoirs that have recently undergone reinforcement, the newly added or upgraded monitoring facilities have a short operational history and often lack long-term historical monitoring data reflecting the normal state after reinforcement, making it difficult to directly establish a reliable statistical baseline.
[0003] On the other hand, the permeability of reinforced structures such as grouting bodies and cutoff walls changes over time during the consolidation phase after construction. Using fixed parameters or thresholds for evaluation can easily confuse normal material evolution with abnormal seepage changes. Currently, monitoring data, finite element analysis results, and dam surface images are often analyzed separately, and the influence of environmental factors such as water level and rainfall on seepage status, as well as the physical correlation between piezometric head, total seepage flow, and dam surface wetting, is not fully utilized. Therefore, how to conduct continuous, quantitative, and physically grounded dynamic assessments of the seepage status of reinforced dams in the absence of long-term historical baselines remains a problem that needs to be solved in reservoir dam safety monitoring. Summary of the Invention
[0004] In response, this application provides a method, apparatus, and storage medium for dynamic assessment of the safety status of reservoir dams based on multi-source data fusion, so as to at least partially solve the above-mentioned technical problems.
[0005] This application provides a method for dynamic assessment of the safety status of reservoir dams based on multi-source data fusion, including the following steps: Obtain the seepage finite element model established during the dam reinforcement design phase, set the upstream water level boundary and rainfall infiltration boundary of the seepage finite element model as time-varying input interfaces, and divide the dam body into reinforcement influence parameter area and non-reinforcement area in the seepage finite element model; Based on the permeability test data of the reinforcement materials in the reinforcement design data, a time-varying decay function of the permeability coefficient of each reinforcement influence parameter region is established. The envelope width is determined by combining the discreteness of the permeability test data, and the expected evolution envelope of the permeability coefficient of each reinforcement influence parameter region is generated. In each data acquisition cycle, the real-time collected water level data and rainfall data are input into the seepage finite element model as boundary conditions, and the measured water head of the piezometer is used as the observation value. The equivalent permeability coefficient of each of the reinforcement influence parameter zones is solved by back analysis to obtain the actual value of the equivalent permeability coefficient of each of the reinforcement influence parameter zones at the current time. The actual value of the equivalent permeability coefficient is compared with the corresponding expected evolution envelope, and the normalized deviation of each reinforcement influence parameter region is calculated. The current safety status of the dam is determined based on the normalized deviation.
[0006] In one possible embodiment, dividing the dam body into reinforcement influence parameter zones and non-reinforced zones includes, according to the actual construction content of the reinforcement project, dividing the areas corresponding to different reinforcement measures within the reinforcement influence range into independent reinforcement influence parameter zones, wherein the number of reinforcement influence parameter zones does not exceed the number of piezometers minus 1; the permeability coefficient of the non-reinforced zone is taken from the recommended value in the geological survey data and kept fixed, and is not involved in the back analysis solution.
[0007] In one possible embodiment, the time-varying decay function of the permeability coefficient adopts an exponential decay form, which includes two parameters: long-term stability ratio and decay rate. The long-term stability ratio is determined by the ratio of the stable period test value to the initial test value in the long-term indoor permeability test, and the decay rate is calculated by the time required for the permeability coefficient to decrease to half of the initial value. The envelope width is obtained by multiplying the coefficient of variation of the permeability test results of multiple groups of specimens in the same batch with the output value of the time-varying decay function.
[0008] In one possible embodiment, the inverse analysis of the equivalent permeability coefficient of each of the reinforcement influence parameter zones includes minimizing the weighted sum of squares of the difference between the model-calculated head and the measured head at each piezometer as the objective function; the observation weight of each piezometer is allocated according to its sensitivity to the change in permeability coefficient of the reinforcement influence parameter zone, and the sensitivity is obtained by applying a set proportional perturbation to the permeability coefficient of the reinforcement influence parameter zone and calculating the proportion of the change in head of each piezometer to the total change in head of all piezometers.
[0009] In one possible embodiment, the mass conservation constraint condition set during the inverse analysis solution process includes ensuring that the relative deviation between the total downstream seepage flow calculated by the seepage finite element model under the parameters obtained from the inverse analysis and the measured flow of the measuring weir does not exceed the allowable relative error, wherein the allowable relative error is determined based on the measurement accuracy of the measuring weir.
[0010] In one possible embodiment, determining the current safety status of the dam based on the normalized deviation includes taking the maximum value of the absolute value of the normalized deviation of each reinforcement influence parameter zone as a comprehensive evaluation index; when the maximum value is less than a first threshold, it is determined to be in a normal state; when the maximum value is not less than the first threshold and less than a second threshold, it is determined to be in a state of concern; when the maximum value is not less than the second threshold, it is determined to be in a state of alert; the second threshold is determined according to the reservoir project level.
[0011] In one possible embodiment, the method further includes an image cross-validation step: acquiring images of the downstream dam face, calculating the saturation difference between the corresponding regions of the current image and the reference image in the color space, marking connected regions where the saturation difference exceeds a judgment threshold as wet regions; spatially comparing the spatial location of the wet regions with the seepage outflow location calculated by the seepage finite element model under the current actual value of the equivalent permeability coefficient, and upgrading the current safety status of the dam when the deviation between the two exceeds a set distance.
[0012] In one possible embodiment, the judgment threshold is determined by on-site calibration: multiple images are collected under known wet and known dry conditions on the dam surface, the distribution of saturation difference under the two conditions is statistically analyzed, and the value with the smallest cross-misjudgment rate between the two distributions is taken as the judgment threshold.
[0013] In one possible embodiment, the inverse analysis solution is solved iteratively using the Levenberg-Marquardt method. In each iteration, a small perturbation is applied to the permeability coefficient of each of the reinforcement influence parameter regions, and the seepage calculation is rerun to obtain the sensitivity matrix. The search range of the permeability coefficient of each of the reinforcement influence parameter regions is subject to physically reasonable upper and lower bound constraints. The iteration termination condition is that the relative change of the objective function between two adjacent iterations is less than the convergence threshold or the number of iterations reaches the set upper limit.
[0014] In another aspect, this application also provides a dynamic assessment system for the safety status of reservoir dams based on multi-source data fusion, comprising: The model configuration module is used to obtain the seepage finite element model established during the dam reinforcement design phase, set the upstream water level boundary and rainfall infiltration boundary of the seepage finite element model as time-varying input interfaces, and divide the dam body into reinforcement influence parameter area and non-reinforcement area in the seepage finite element model. The envelope generation module is used to establish a time-varying decay function of the permeability coefficient of each of the reinforcement influence parameter regions based on the permeability test data of the reinforcement materials in the reinforcement design data, determine the envelope width in combination with the discreteness of the permeability test data, and generate the expected evolution envelope of the permeability coefficient of each of the reinforcement influence parameter regions. The inverse analysis solution module is used to input the real-time collected water level data and rainfall data into the seepage finite element model as boundary conditions in each data acquisition cycle, and use the measured water head of the piezometer as the observation value to perform inverse analysis to solve the equivalent permeability coefficient of each of the reinforcement influence parameter zones, so as to obtain the actual value of the equivalent permeability coefficient of each of the reinforcement influence parameter zones at the current time. The safety status determination module is used to compare the actual value of the equivalent permeability coefficient with the corresponding expected evolution envelope, calculate the normalized deviation of each reinforcement influence parameter region, and determine the current safety status of the dam based on the normalized deviation.
[0015] This application addresses the issues of the lack of long-term historical monitoring baselines for reservoir dams after reinforcement and the continuous change in the permeability of reinforcement materials during the consolidation process. It utilizes existing seepage finite element models and permeability test data of reinforcement materials from the reinforcement design phase to establish a projected evolution envelope that reflects the time-dependent changes in normal permeability in the reinforcement-affected parameter zones. During operation, real-time water level and rainfall data are used as model boundary conditions. Piezometric head measurements are used for back-analysis of the current equivalent permeability coefficient in each reinforcement-affected parameter zone. Simultaneously, measured flow rates from a measuring weir are used to apply mass conservation constraints to the back-analysis results. The safety status is determined based on the deviation of the actual equivalent permeability coefficient from the projected evolution envelope and its changing trend. Verification is further performed by considering the spatial consistency between the downstream dam surface wet zone and the model seepage outflow zone. This allows the safety assessment benchmark to change synchronously with the normal performance evolution of the reinforcement materials, forming dynamic evaluation results with clear physical basis without the need for long-term historical data accumulation. It also reduces the impact of abnormal single monitoring quantities, local fitting deviations, or the model not including local seepage channels on the reliability of the assessment conclusions, and improves the timeliness and credibility of identifying the seepage safety status of the reinforced dam. Attached Figure Description
[0016] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0017] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of a dynamic assessment method for the safety status of a reservoir dam based on multi-source data fusion, provided in an embodiment of this application.
[0018] Figure 2This is a schematic diagram of the back analysis of the equivalent permeability coefficient under multi-source monitoring constraints provided in the embodiments of this application.
[0019] Figure 3 A schematic diagram showing the evolution of the equivalent permeability coefficient in the reinforcement influence parameter zone provided in the embodiments of this application.
[0020] Figure 4 This is a schematic diagram illustrating the spatial cross-verification of the wet zone and seepage outflow zone on the dam surface, provided in an embodiment of this application.
[0021] Figure 5 This is a schematic diagram of the structure of a dynamic assessment system for the safety status of a reservoir dam based on multi-source data fusion, provided in an embodiment of this application.
[0022] Figure 6 This is a schematic diagram of the structure of a device provided in an embodiment of this application. Detailed Implementation
[0023] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0024] This embodiment provides a dynamic assessment method for the safety status of reservoir dams based on multi-source data fusion, applicable to small earth-rock dams that have completed reinforcement construction. The prerequisites for implementing this method typically include: a finite element analysis model of dam seepage has been established during the dam reinforcement design phase; after reinforcement construction, seepage monitoring instruments such as piezometers and weirs have been installed according to the design scheme; and the dam site is equipped with an automatic water level monitoring system, an automatic rainfall monitoring system, and a downstream dam surface image automatic monitoring system, all of which are in normal operating condition. The reinforcement design data also includes indoor permeability test results for the reinforcement materials.
[0025] The following detailed description, in conjunction with specific embodiments, illustrates the implementation process of the dynamic assessment method for the safety status of reservoir dams based on multi-source data fusion described in this application. It should be noted that this embodiment is merely for explaining this application and not for limiting the scope of protection of this application. Conventional adjustments or substitutions made by those skilled in the art to each step without departing from the concept of this application should be included within the scope of protection of this application.
[0026] like Figure 1 As shown in the figure, this application discloses a schematic diagram of a dynamic assessment method for the safety status of a reservoir dam based on multi-source data fusion, including the following method steps: Step S1: Obtain the seepage finite element model and set the time-varying input interface and parameter partitioning. Specifically, obtain the seepage finite element model established during the dam reinforcement design phase, set the upstream water level boundary and rainfall infiltration boundary of the seepage finite element model as time-varying input interfaces, and divide the dam body into reinforcement influence parameter zone and non-reinforcement zone in the seepage finite element model.
[0027] In some embodiments, a seepage finite element model established during the dam reinforcement design phase is obtained from reinforcement design outcome data. This model is used during the design phase to calculate the seepage field distribution of the reinforced dam body, including the geometry of the reinforced dam body, mesh generation, and initial assignment of the material permeability coefficient. For small earth-rock dams with a height of 15 to 30 meters, a typical two-dimensional profile model usually contains 3,000 to 8,000 triangular or quadrilateral finite element elements, and the computational domain covers the dam foundation, dam body, and reinforcement structure.
[0028] Specifically, the boundary conditions of the model are adapted for runtime. The upstream water level boundary is changed from a fixed water level value in the design phase to a time-varying input interface, receiving real-time reservoir water level data collected by the automatic water level monitoring system. Similarly, the rainfall infiltration boundaries at the dam crest and downstream slope are also changed to time-varying input interfaces, receiving rainfall intensity data from the automatic rainfall monitoring system and multiplying it by an infiltration reduction factor to convert it into actual infiltration intensity. The infiltration reduction factor is taken from the field infiltration test results of the dam body soil in the reinforcement design report; for example, the infiltration reduction factor for loam-filled dams is typically between 0.3 and 0.6. The downstream drainage outlet surface maintains a free infiltration boundary condition, i.e., the water head equals the elevation of the node. The governing equations for solving the seepage finite element model are the unsteady seepage continuity equations: in The total head is measured in meters (m). , , respectively, are the horizontal and vertical permeability coefficients, in m / s; The specific water storage coefficient is dimensional (1 / m). Let be time, in units of s.
[0029] In one possible implementation, the aforementioned unsteady seepage equations are discretized in time using an implicit backward Euler scheme, with the time step set to 1 / 10 of the monitoring data acquisition cycle (i.e., a time step of 36 minutes when the acquisition cycle is 6 hours). Spatial discretization employs the Galerkin weighted residual method, using six-node triangular quadratic elements, and the overall stiffness matrix is stored using one-dimensional variable bandwidth. Nonlinear iteration uses Picard iteration, with a convergence tolerance set at a head variation not exceeding 1 × 10⁻⁻⁴. 4 m. The permeability coefficient in the unsaturated zone is reduced according to the relationship curve between matrix suction and permeability coefficient (obtained from soil-water characteristic curve tests), and the reduction factor is the measured value.
[0030] In one embodiment, the installation coordinates of all newly added piezometers and the cross-sectional location of the measuring weir are marked in the model. The marking method involves determining the node number in the finite element mesh closest to the instrument installation coordinates; the head value of this node is then extracted as the model output during subsequent calculations. The cross-section corresponding to the measuring weir location is used to calculate the integral value of the seepage flow at that cross-section.
[0031] In one embodiment, based on the actual construction content of the reinforcement project, the dam body is divided into reinforcement-affected parameter zones and non-reinforced zones in the seepage finite element model. Specifically, the reinforcement-affected parameter zones refer to areas where the original permeability has been altered by reinforcement measures, such as curtain grouting zones and concrete cutoff wall zones. Each of these is designated as an independent parameter zone, and the permeability coefficient within each parameter zone is considered uniform and used as a parameter to be determined in subsequent back analysis. The number of reinforcement-affected parameter zones does not exceed the number of piezometers minus one to ensure that the back analysis problem satisfies the well-posedness condition. For example, if the dam body has four piezometers, then the reinforcement-affected parameter zones can be divided into a maximum of three. The permeability coefficient of the non-reinforced zone is taken from the recommended value in the geological survey data and remains fixed in all subsequent calculations, not participating in the back analysis solution.
[0032] The model adaptation process is performed once during system initialization, and the adapted model is directly called during subsequent runtime. Since mesh generation, material partitioning, and initial parameter assignment are completed during the model building phase of the reinforcement design phase, runtime adaptation only requires modifying the boundary condition type and marking observation points, without the need for remodeling.
[0033] Step S2: Establish a time-varying decay function for the permeability coefficient and generate the expected evolution envelope. Specifically, based on the permeability test data of the reinforcement material in the reinforcement design data, establish a time-varying decay function for the permeability coefficient of each reinforcement influence parameter region, determine the envelope width by combining the discreteness of the permeability test data, and generate the expected evolution envelope of the permeability coefficient of each reinforcement influence parameter region.
[0034] After the reinforcement construction is completed, the permeability coefficients of the newly constructed structures, such as the grout and the cutoff wall, gradually decrease with cement hydration and material consolidation, stabilizing after several months. This physical process means that the permeability coefficient in the reinforcement-affected parameter zone is not constant in the initial stage. Using indoor permeability test data of various reinforcement materials stored in the reinforcement design data, a time-varying decay function of the permeability coefficient is established for each reinforcement-affected parameter zone, and the expected evolution envelope is generated by combining the discreteness of the test data. The expected evolution envelope refers to the upper and lower bounds of the permeability coefficient change over time in each reinforcement-affected parameter zone, representing the range within which the permeability coefficient should fall under the premise of normal reinforcement quality.
[0035] Specifically, in one embodiment, for the first For a reinforcement-affected parameter zone, with the completion time of reinforcement construction as the time origin, the expected median permeability coefficient changes over time as follows: in For the first The initial permeability coefficient of each reinforcement influence parameter area at the time of construction completion, in m / s, is taken from the measured value of on-site water injection test or indoor permeability test at the time of construction completion in the reinforcement design report. It is a dimensionless time-varying decay function, which characterizes the proportion of change of the permeability coefficient relative to the initial value.
[0036] The time-varying decay function adopts an exponential decay form: in The time is calculated from the completion of construction, in the dimension d (days). The long-term stability ratio, physically defined as the ratio of the final stable value of the permeability coefficient to the initial value, ranges from 0.05 to 0.5 and is determined by the ratio of the stable period test value to the initial test value in a long-term indoor permeability test. The decay rate is the time required for the permeability coefficient to decrease to half of its initial value. The conversion yields, i.e. .
[0037] As an example, for the parameter zone of cement grout, indoor tests showed that the permeability coefficient decreased to about 40% of the initial value after 28 days, and decreased to 12% to 15% of the initial value after 180 days, and then tended to stabilize. Based on this, the following can be fitted... , Approximately 22 days, corresponding to 1 / d.
[0038] In one embodiment, the envelope width is determined by the coefficient of variation of the penetration test results from multiple groups of specimens in the same batch. Specifically, let the first... The reinforcement materials corresponding to the reinforcement influence parameter areas at the same age total [number] The coefficient of variation is obtained by calculating the ratio of the standard deviation to the mean of the permeability coefficient test values of the group of specimens. The half-width of the envelope is: For cement grouting bodies, The value is typically between 0.2 and 0.4. Alternatively, when the number of test pieces in the same batch is less than 5 groups, it can be determined based on engineering experience. .
[0039] In some embodiments, an upper and lower bound for the expected evolutionary envelope is generated: Among them, the upper bound and the lower bound constitute the first... The reinforcement influence parameter region at time... The expected evolution envelope means that, under normal reinforcement quality, the equivalent permeability coefficient of the region should fall within this range.
[0040] Therefore, each reinforcement influence parameter region has a predicted evolution envelope that changes over time, serving as a comparison benchmark in subsequent safety status determination. All inputs for this step come from existing experimental data in the design documentation, requiring no additional experiments or long-term observations; it can be generated immediately after the instrument is put into operation. This is the key to solving the problem of no historical baseline after reinforcement: replacing the traditional statistical baseline with the material consolidation law, providing an evaluation reference from a physical mechanism perspective.
[0041] Step S3 involves inputting real-time monitoring data into the seepage finite element model for back analysis to solve the equivalent permeability coefficient. Specifically, in each data acquisition cycle, the real-time collected water level data and rainfall data are input into the seepage finite element model as boundary conditions. The measured water head from the piezometer is used as the observed value to perform back analysis on the equivalent permeability coefficient of each of the reinforcement influence parameter zones, thereby obtaining the actual value of the equivalent permeability coefficient of each of the reinforcement influence parameter zones at the current moment.
[0042] This step is performed after acquiring a complete set of monitoring data for each data acquisition cycle (typically 4 to 6 hours for small reservoirs using automatic acquisition systems).
[0043] Figure 2 This diagram illustrates the back analysis of the equivalent permeability coefficient under multi-source monitoring constraints provided in this application embodiment. As shown, the automatic water level monitoring system acquires the upstream water level and uses it as the time-varying upstream water level boundary. The automatic rainfall monitoring system acquires rainfall data and forms a time-varying rainfall infiltration intensity, which is then input into the seepage finite element model as the rainfall infiltration boundary. The model divides the dam body into a reinforcement influence parameter zone and a non-reinforced zone. The permeability coefficient of the non-reinforced zone is fixed, while the permeability coefficient of the reinforcement influence parameter zone is used as the parameter to be determined. Piezometers are installed within the dam body, and the measured water head is used as the observed value. During the back analysis of the seepage finite element model, the objective function is to minimize the weighted sum of squares of the differences between the model-calculated water head and the measured water head at each piezometer. Simultaneously, a mass conservation constraint is applied based on the downstream measured flow rate obtained from the weir monitoring system, requiring that the relative deviation between the model-calculated downstream total seepage flow rate and the weir measured flow rate not exceed the allowable relative error. Finally, the actual value of the equivalent permeability coefficient at the current moment for each reinforcement influence parameter zone is output.
[0044] Specifically, in one embodiment, real-time collected water level data is input into the seepage finite element model as the upstream water level boundary condition, and real-time collected rainfall data, after infiltration reduction, is input into the model as the rainfall infiltration boundary condition. Seepage calculations are then performed to obtain the calculated hydraulic head values at each piezometer location node. The specific value of the calculated head depends on the current input boundary conditions and the permeability coefficient values of each region in the model.
[0045] In one possible implementation, an inverse analysis objective function is constructed; specifically, let there be a total... The branch piezometer has parameters to be determined, which are vectors composed of the equivalent permeability coefficients of each reinforcement influence parameter zone. , To increase the number of areas affecting parameters, the weighted sum of squares of the differences between the calculated head and the measured head at each piezometer is used as the objective function: in For parameter vectors After running the seepage finite element model at the first Calculated head extracted from the location of the branch piezometer, dimensionless (m); For the first The measured head at the same moment on the branch piezometer, in meters (m); For the first The observation weights of the branch manometers. Inverse analysis solves this by finding the weights that make the branch manometers more sensitive to pressure changes. The parameter vector that takes the minimum value .
[0046] The observation weight of each piezometer is allocated according to its sensitivity to changes in the permeability coefficient of the reinforcement-affected parameter zone. Specifically, during the model adaptation phase, the permeability coefficient of each reinforcement-affected parameter zone is increased by a set percentage (e.g., 10%) based on its current value, the seepage calculation is rerun, and the change in head at each piezometer is recorded. The weights are calculated as the proportion of the change in head at each piezometer to the total change in head at all piezometers. Piezometers located closer to the reinforced area and showing a more significant response to changes in permeability coefficient receive a higher weight, while those farther away receive a lower weight. This weight is calculated once during system initialization and then fixed.
[0047] In one embodiment, a mass conservation constraint is set; specifically, the downstream total seepage flow rate calculated by the seepage finite element model under the parameters obtained from the back analysis is... The measured flow rate of the measuring weir is obtained by integrating the flow velocity of each unit at the downstream infiltration surface. The readings are directly obtained from the water measuring weir system. The constraint condition requires that the relative deviation between the two measurements not exceed the allowable relative error. : Allowable relative error The accuracy is determined based on the measurement precision of the weir. For example, for a right-angled triangular thin-walled weir, the measurement precision is typically ±5% to ±10%. The value is set to 0.05 to 0.10. This constraint ensures that the parameters obtained from the back analysis meet the piezometric head matching requirements locally, while also maintaining the overall flow balance, thus avoiding a situation where the head fit at a single point is good but the total flow is distorted.
[0048] In one embodiment, the inverse analysis solution employs the Levenberg-Marquardt method for iterative solving. Specifically, this method belongs to the standard solution algorithm for nonlinear least squares problems, introducing a damping factor based on the Gauss-Newton method to balance convergence speed and stability. In each iteration, a small perturbation (e.g., increasing the permeability coefficient by 1% from the current value) is applied to the permeability coefficient of each reinforcement-affected parameter region, and the seepage finite element calculation is rerun. The sensitivity matrix (i.e., the Jacobian matrix) is obtained by dividing the change in head by the parameter perturbation. ).for Each parameter area requires additional execution in each iteration. Finite element method calculation.
[0049] In one possible implementation, the inverse analysis solver is implemented as follows: Input is A dimensional vector is: in To increase the number of areas affected by the reinforcement parameter, the value is set to [value]. to ( (Number of pressure testing tubes), unit for each element is... The vector dimension is .
[0050] The iterative state variable is the current parameter vector. (dimension) ), and the current objective function value (Scalar, unit is) ).
[0051] In each iteration, for the th Apply parameters to the region The tiny perturbation, namely: Rerun the seepage finite element model and calculate the Jacobian matrix elements: The dimension of this matrix is Damping factor Initial selection Calculate the gain ratio after each iteration: in This is an approximation for a linearized model. The update rules are as follows: when hour, Halve ( ); when hour, double( ).
[0052] The iteration terminates when both of the following conditions are met: (a) The relative change of the objective function between two consecutive iterations: (b) The relative change of the parameter vector between two consecutive iterations: The output is the optimal parameter vector. (dimension) The unit is and the corresponding minimum value of the objective function .
[0053] In some embodiments, the search range for the permeability coefficient of each reinforcement influence parameter zone is set with physically reasonable upper and lower bounds: the lower bound is taken as the permeability coefficient of a complete and dense rock mass (e.g., ...). m / s), the upper limit is taken as the permeability coefficient of the dam fill material (e.g., m / s), (m / s) to prevent physically unreasonable parameter values from occurring during iteration. The iteration termination condition is that the relative change of the objective function between two consecutive iterations is less than the convergence threshold (e.g., m / s). (or the number of iterations reaches the set limit (e.g., 50 times)).
[0054] As an example, a small dam has two reinforcement influence parameter zones and four piezometers, with a model of approximately 5000 elements. At a certain moment, the upstream water level is 42.5m, and the measured water heads from the four piezometers are 38.2, 36.8, 35.1, and 34.0m, respectively. Using initial parameters... m / s Starting from m / s, the objective function converges after 7 iterations, yielding the following results. m / s The flow rate is m / s, and the residual difference between the calculated and measured head of each piezometer is less than 0.15 m. The relative deviation between the calculated seepage flow rate and the measured flow rate of the measuring weir is 4.3%, which meets the constraints. The single back analysis run time is approximately 20 seconds.
[0055] Optionally, after the back analysis is completed, the actual equivalent permeability coefficient of each reinforcement influence parameter zone at the current time is output. This value is recorded and stored to form a time series.
[0056] Optionally, if the residual of the objective function in a back analysis is too large (e.g., more than 3 times the historical residual mean), the result is marked as an outlier and not included in subsequent trend analysis, while triggering a data quality check.
[0057] Step S4: Compare the actual value of the equivalent permeability coefficient with the expected evolution envelope and determine the safety status. Specifically, compare the actual value of the equivalent permeability coefficient with the corresponding expected evolution envelope, calculate the normalized deviation of each reinforcement influence parameter region, and determine the current safety status of the dam based on the normalized deviation.
[0058] Figure 3 This diagram illustrates the evolution of the equivalent permeability coefficient in the reinforcement influence parameter zone, as provided in this embodiment. As shown, the completion time of reinforcement construction is taken as the starting point, with the horizontal axis representing time and the vertical axis representing the equivalent permeability coefficient in the reinforcement influence parameter zone. Based on the permeability test data of the reinforcement material, an expected evolution median value is established, which decays over time and gradually stabilizes. Combined with the dispersion of the test data, an expected evolution envelope is formed around this median value to characterize the expected range of change in the equivalent permeability coefficient under normal reinforcement quality conditions. The actual value of the equivalent permeability coefficient is obtained through back-analysis for each data acquisition cycle, forming an actual evolution trajectory over time. At the current moment, the corresponding data point represents the actual value of the equivalent permeability coefficient. Its relative positional relationship with the expected evolution median and the expected evolution envelope provides a basis for subsequent safety status determination.
[0059] In one embodiment, for each reinforcement influence parameter zone, the actual value of the equivalent permeability coefficient obtained from the aforementioned back analysis is used. The normalized deviation is calculated by comparing the actual value of the equivalent permeability coefficient with the previously generated expected evolution envelope. The normalized deviation refers to the degree to which the actual value deviates from the median of the expected evolution envelope. It is a dimensionless index obtained after normalization based on the envelope width. Molecular weight is m / s minus m / s = m / s. This indicates that the actual value of the equivalent permeability coefficient falls within the expected evolution envelope range. This indicates that the range is exceeded.
[0060] In one embodiment, the maximum value among the normalized absolute values of the normalized deviations of each reinforcement influence parameter region is taken as the comprehensive evaluation index. The current safety status of the dam shall be determined according to the following criteria: when When the value is less than the first threshold, it is considered a normal state; when... When the value is not less than the first threshold and less than the second threshold, it is determined to be in a state of concern; when When the value is not less than the second threshold, the system is considered to be in a state of alert. The first threshold is set to 1.0, which is the boundary of the expected evolution envelope. The second threshold is determined according to the reservoir project's grade: 1.5 for Grade IV projects and 2.0 for Grade V projects, which can be based on the differentiated safety margin requirements for different grades of projects in the current reservoir dam safety management regulations.
[0061] Continuing with the previous example, if the actual value of the equivalent permeability coefficient of the curtain grouting zone at a certain moment is... m / s, at which moment the expected evolution envelope median is m / s, envelope half-width is m / s, then the normalized deviation is It falls within the expected evolution envelope, which is a normal state.
[0062] In one embodiment, within a set period after the reinforcement construction is completed (e.g., 6 months), the material consolidation process is physically irreversible, and the permeability coefficient should show a continuous decreasing trend. If the actual value of the equivalent permeability coefficient shows a continuous increasing trend over multiple consecutive data collection periods (e.g., 3 or more consecutive periods), even if the actual value of the equivalent permeability coefficient is still within the expected evolution envelope, the current safety status of the dam will be judged as a state of concern, and a trend warning will be issued. The physical basis for this judgment is that the increase in permeability coefficient during the consolidation period indicates that there are external factors that offset the consolidation effect, such as seepage erosion removing grouting material or structural interface detachment.
[0063] Optionally, when the reservoir's operating water level is higher than the normal storage level, the increased hydraulic gradient may temporarily overestimate the equivalent permeability coefficient obtained from the back analysis. In this case, the upper bound of the expected evolution envelope is multiplied by a relaxation factor, which is the square of the ratio of the current water level to the normal storage level. This method is based on Darcy's law, which states that seepage flow is approximately proportional to hydraulic gradient: a rise in water level causes an increase in hydraulic gradient. Under the condition of constant seepage flow, the magnitude of the overestimation of the equivalent permeability coefficient obtained from the back analysis is approximately proportional to the increase in hydraulic gradient, which in turn is proportional to the increase in head difference, which is approximately proportional to the square of the water level ratio.
[0064] In one embodiment, optionally, as an auxiliary verification method, an automatic image monitoring system is used to independently verify the calculation results of the seepage finite element model. A downstream dam surface image taken under dry, sunny conditions is used as a reference image. The saturation difference between each pixel region of the currently acquired image and the reference image is calculated in the HSV color space. When seepage occurs on the dam surface, the saturation of the wet area is significantly higher than that of the dry area. Connected regions with saturation differences exceeding a threshold are marked as wet areas.
[0065] The judgment threshold is determined through on-site calibration. More than 10 images are collected under known wet and known dry conditions on the dam surface, respectively. The distribution characteristics of the saturation difference under the two conditions are statistically analyzed, and the value with the smallest cross-misjudgment rate between the two distributions is taken as the judgment threshold, for example, a value of 0.15.
[0066] Figure 4 This diagram illustrates the spatial cross-validation of the wet zone and seepage outflow zone on the dam surface provided in this embodiment. As shown, the downstream dam surface image automatic monitoring system acquires images of the dam surface and uses camera calibration parameters to convert the pixel coordinates of the wet zone in the image into actual coordinates of the dam surface. The seepage finite element model calculates the seepage outflow zone on the dam surface based on the actual value of the current equivalent permeability coefficient. Using the actual location of the dam surface, such as the dam toe, as a spatial reference, the wet zone and the seepage outflow zone are spatially compared, and their spatial deviation is determined. When the spatial deviation does not exceed a set distance, the spatial state of the image recognition result is considered consistent with the spatial state of the model calculation result; when the spatial deviation exceeds a set distance, the current safety state of the dam is upgraded; when the model predicts no outflow but the image shows the presence of a wet zone, the current safety state of the dam is also upgraded, so as to use independent dam surface image information to cross-validate the model calculation results.
[0067] Specifically, after converting the pixel coordinates of the wetted area to the actual coordinates of the dam surface using camera calibration parameters, a spatial comparison is made with the seepage outflow location calculated by the seepage finite element model under the current actual value of the equivalent permeability coefficient. If the spatial deviation is within 3 meters, it indicates that the model matches the actual state; if the deviation exceeds a set distance (e.g., 3 meters), or if the model predicts no outflow but the image shows a wetted area, the dam's current safety status is upgraded. Discrepancies are usually caused by the existence of local seepage channels not included in the model, such as cracks or animal burrows, requiring notification to management personnel for on-site verification. The image verification results are not involved in the aforementioned back-analysis calculation process and serve only as independent verification information for the safety status assessment.
[0068] Optionally, a monitoring system equipped with night vision can use the near-infrared channel to identify wet areas on the dam surface at night.
[0069] It should be noted that, in the above steps, the acquisition of the seepage finite element model and the setting of the time-varying input interface and parameter partitions are executed once during the system initialization phase, and the establishment of the time-varying decay function of the permeability coefficient and the generation of the expected evolution envelope are executed once after the acquisition of material test data and then stored. Afterwards, the cyclic operation phase begins: in each data acquisition cycle, three calculations are completed sequentially: back analysis solution, safety status assessment, and image cross-validation, outputting the current safety status level and normalized deviation value. All calculations are automatically run on the on-site industrial control computer or a remote server. For typical small dam models (5000 to 8000 elements, 2 to 3 reinforcement influence parameter zones), a single cycle calculation takes no more than 2 minutes.
[0070] Using the above method, under the condition that the reinforced piezometer and weir have just been put into use and there are no historical monitoring records, an evaluation benchmark is constructed using the existing finite element model and material test data from the reinforcement design stage. The safety judgment is based on whether the actual evolution trajectory of the equivalent permeability coefficient conforms to the physical laws of material consolidation. This allows for a quantitative assessment of the dam's safety status without waiting for years of data accumulation. The multi-source monitoring data serve different physical roles, including boundary condition input, parameter back-analysis constraints, mass conservation verification, and spatial cross-validation. The evaluation criteria have clear physical meanings that engineers can directly understand.
[0071] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of a dynamic assessment system for the safety status of a reservoir dam based on multi-source data fusion, provided in an embodiment of this application. Figure 5 As shown, the system includes: The model configuration module 501 is used to obtain the seepage finite element model established during the dam reinforcement design stage, set the upstream water level boundary and rainfall infiltration boundary of the seepage finite element model as time-varying input interfaces, and divide the dam body into reinforcement influence parameter area and non-reinforcement area in the seepage finite element model. The envelope generation module 502 is used to establish a time-varying decay function of the permeability coefficient of each of the reinforcement influence parameter regions based on the permeability test data of the reinforcement material in the reinforcement design data, determine the envelope width in combination with the discreteness of the permeability test data, and generate the expected evolution envelope of the permeability coefficient of each of the reinforcement influence parameter regions. The inverse analysis solution module 503 is used to input the real-time collected water level data and rainfall data into the seepage finite element model as boundary conditions in each data acquisition cycle, and use the measured water head of the piezometer as the observation value to perform inverse analysis to solve the equivalent permeability coefficient of each of the reinforcement influence parameter zones, so as to obtain the actual value of the equivalent permeability coefficient of each of the reinforcement influence parameter zones at the current time. The safety status determination module 504 is used to compare the actual value of the equivalent permeability coefficient with the corresponding expected evolution envelope, calculate the normalized deviation of each reinforcement influence parameter zone, and determine the current safety status of the dam based on the normalized deviation.
[0072] Each processing unit and / or module in the embodiments of this application can be implemented by an analog circuit that implements the functions described in the embodiments of this application, or by software that executes the functions described in the embodiments of this application.
[0073] Please see Figure 6 It shows a schematic diagram of the structure of an electronic device according to an embodiment of this application, which can be used to implement... Figure 1The method in the illustrated embodiment. (As shown) Figure 6 As shown, the electronic device may include: The system includes at least one processor 601, at least one network interface 604, a user interface 603, a memory 605, and at least one communication bus 602. The communication bus 602 is used to enable connection and communication between the components. The user interface 603 may include buttons, and optionally include a standard wired or wireless interface. The network interface 604 may include, but is not limited to, a Bluetooth module, an NFC module, a Wi-Fi module, etc.
[0074] The processor 601 may include one or more processing cores and connect to various parts within the electronic device through various interfaces and lines. It implements various functions and data processing of the electronic device by running or executing instructions, programs, code sets, or instruction sets stored in the memory 605, and by accessing data in the memory 605. Optionally, the processor 601 may be implemented using at least one hardware form of DSP, FPGA, or PLA. The processor 601 may also integrate one or more combinations of CPU, GPU, and modem. The CPU is mainly used to handle the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display on the screen; and the modem is used for wireless communication. It is understood that the modem may not be integrated into the processor 601, but may be implemented through a separate chip.
[0075] Memory 605 may include random access memory (RAM) or read-only memory (ROM). Optionally, memory 605 includes a non-transitory computer-readable medium for storing instructions, programs, code, code sets, or instruction sets. Memory 605 may be divided into a program storage area and a data storage area, wherein the program storage area may be used to store instructions for implementing an operating system, instructions for implementing at least one function (such as touch functionality, audio playback functionality, image playback functionality, etc.), and instructions for implementing the foregoing method embodiments; the data storage area may be used to store data involved in the relevant method embodiments. Memory 605 may also be at least one storage device located remotely from processor 601. Figure 6 As shown, the memory 605, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and program instructions.
[0076] In particular, the methods and / or embodiments in this application can be implemented as computer software programs. For example, the embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. When the computer program is executed by processor 601, it performs the functions defined in the methods of this application.
[0077] Another embodiment of this application provides a computer-readable storage medium having computer program instructions stored thereon, which can be executed by a processor to implement the methods and / or technical solutions of any one or more embodiments of this application described above.
[0078] The computer-readable storage medium may include, but is not limited to, any type of disk, including floppy disks, optical disks, DVDs, CD-ROMs, microdrives, as well as magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, DRAMs, VRAMs, flash memory devices, magnetic cards or optical cards, nanosystems (including molecular memory ICs), or any type of medium or device suitable for storing instructions and / or data.
[0079] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.
Claims
1. A method for dynamic assessment of the safety status of reservoir dams based on multi-source data fusion, characterized in that, include: Obtain the seepage finite element model established during the dam reinforcement design phase, set the upstream water level boundary and rainfall infiltration boundary of the seepage finite element model as time-varying input interfaces, and divide the dam body into reinforcement influence parameter area and non-reinforcement area in the seepage finite element model; Based on the permeability test data of the reinforcement materials in the reinforcement design data, a time-varying decay function of the permeability coefficient of each reinforcement influence parameter region is established. The envelope width is determined by combining the discreteness of the permeability test data, and the expected evolution envelope of the permeability coefficient of each reinforcement influence parameter region is generated. In each data acquisition cycle, the real-time collected water level data and rainfall data are input into the seepage finite element model as boundary conditions, and the measured water head of the piezometer is used as the observation value. The equivalent permeability coefficient of each of the reinforcement influence parameter zones is solved by back analysis to obtain the actual value of the equivalent permeability coefficient of each of the reinforcement influence parameter zones at the current time. The actual value of the equivalent permeability coefficient is compared with the corresponding expected evolution envelope, and the normalized deviation of each reinforcement influence parameter region is calculated. The current safety status of the dam is determined based on the normalized deviation.
2. The method according to claim 1, characterized in that, The division of the dam body into reinforcement influence parameter zones and non-reinforcement zones includes, based on the actual construction content of the reinforcement project, dividing the areas corresponding to different reinforcement measures within the reinforcement influence range into independent reinforcement influence parameter zones, wherein the number of reinforcement influence parameter zones does not exceed the number of piezometers minus 1; the permeability coefficient of the non-reinforcement zone is taken from the recommended value in the geological survey data and kept fixed, and is not involved in the back analysis solution.
3. The method according to claim 1, characterized in that, The time-varying decay function of the permeability coefficient adopts an exponential decay form and includes two parameters: long-term stability ratio and decay rate. The long-term stability ratio is determined by the ratio of the stable period test value to the initial test value in the long-term indoor permeability test. The decay rate is calculated by converting the time required for the permeability coefficient to decrease to half of the initial value. The envelope width is obtained by multiplying the coefficient of variation of the permeability test results of multiple groups of specimens in the same batch with the output value of the time-varying decay function.
4. The method according to claim 1, characterized in that, The back-analysis solution for the equivalent permeability coefficient of each of the reinforcement influence parameter zones includes minimizing the weighted sum of squares of the difference between the model-calculated head and the measured head at each piezometer as the objective function; the observation weight of each piezometer is allocated according to its sensitivity to the change in permeability coefficient of the reinforcement influence parameter zone, and the sensitivity is obtained by applying a set proportional perturbation to the permeability coefficient of the reinforcement influence parameter zone and calculating the proportion of the change in head of each piezometer to the total change in head of all piezometers.
5. The method according to claim 4, characterized in that, The mass conservation constraint conditions set in the inverse analysis solution process include that the relative deviation between the total downstream seepage flow calculated by the seepage finite element model under the parameters obtained from the inverse analysis and the measured flow of the measuring weir does not exceed the allowable relative error, and the allowable relative error is determined according to the measurement accuracy of the measuring weir.
6. The method according to claim 1, characterized in that, The step of determining the current safety status of the dam based on the normalized deviation includes taking the maximum value of the absolute value of the normalized deviation of each reinforcement influence parameter zone as a comprehensive evaluation index; when the maximum value is less than a first threshold, it is determined to be in a normal state; when the maximum value is not less than the first threshold and less than a second threshold, it is determined to be in a state of concern; when the maximum value is not less than the second threshold, it is determined to be in a state of alert; the second threshold is determined according to the reservoir project level.
7. The method according to claim 1, characterized in that, It also includes an image cross-validation step: acquiring images of the downstream dam surface, calculating the saturation difference between the corresponding regions of the current image and the reference image in the color space, marking connected regions where the saturation difference exceeds the judgment threshold as wet regions; comparing the spatial location of the wet regions with the seepage outflow location calculated by the seepage finite element model under the current actual value of the equivalent permeability coefficient, and upgrading the current safety status of the dam when the deviation between the two exceeds a set distance.
8. The method according to claim 7, characterized in that, The judgment threshold is determined through on-site calibration: multiple images are collected under known wet and known dry conditions on the dam surface, the distribution of saturation difference under the two conditions is statistically analyzed, and the value with the smallest cross-misjudgment rate between the two distributions is taken as the judgment threshold.
9. The method according to claim 1, characterized in that, The inverse analysis solution is solved using the Levenberg-Marquardt iterative method. In each iteration, a small perturbation is applied to the permeability coefficient of each of the reinforcement influence parameter regions, and the seepage calculation is rerun to obtain the sensitivity matrix. The search range of the permeability coefficient of each of the reinforcement influence parameter regions is subject to physically reasonable upper and lower bound constraints. The iteration terminates when the relative change of the objective function between two adjacent iterations is less than the convergence threshold or the number of iterations reaches the set upper limit.
10. A dynamic assessment system for the safety status of reservoir dams based on multi-source data fusion, characterized in that, include: The model configuration module is used to obtain the seepage finite element model established during the dam reinforcement design phase, set the upstream water level boundary and rainfall infiltration boundary of the seepage finite element model as time-varying input interfaces, and divide the dam body into reinforcement influence parameter area and non-reinforcement area in the seepage finite element model. The envelope generation module is used to establish a time-varying decay function of the permeability coefficient of each of the reinforcement influence parameter regions based on the permeability test data of the reinforcement materials in the reinforcement design data, determine the envelope width in combination with the discreteness of the permeability test data, and generate the expected evolution envelope of the permeability coefficient of each of the reinforcement influence parameter regions. The inverse analysis solution module is used to input the real-time collected water level data and rainfall data into the seepage finite element model as boundary conditions in each data acquisition cycle, and use the measured water head of the piezometer as the observation value to perform inverse analysis to solve the equivalent permeability coefficient of each of the reinforcement influence parameter zones, so as to obtain the actual value of the equivalent permeability coefficient of each of the reinforcement influence parameter zones at the current time. The safety status determination module is used to compare the actual value of the equivalent permeability coefficient with the corresponding expected evolution envelope, calculate the normalized deviation of each reinforcement influence parameter region, and determine the current safety status of the dam based on the normalized deviation.
