Dam safety cause analysis system and method based on multi-dimensional monitoring data

Abstract

This invention discloses a dam safety causal analysis system based on multi-dimensional monitoring data, relating to the field of water conservancy engineering safety. The system includes: a data acquisition module for acquiring multi-dimensional monitoring data of the dam; a data preprocessing module for preprocessing the multi-dimensional monitoring data; a causal inference module for acquiring real-time feature sets, causal structure models, and causal effect quantification results; and a decision support module for determining whether the dam has safety risks based on the real-time feature sets and the causal structure model, and for obtaining safety causal nodes and causal transmission paths based on the causal structure model and causal effect quantification results. The technical solution of this invention not only effectively distinguishes the causal relationships between various monitoring data, improving the accuracy of dam safety status determination results, but also obtains safety causal nodes and causal transmission paths affecting dam safety, enhancing the interpretability of the dam safety causal analysis results.

Description

Technical Field

[0001] This invention relates to the field of water conservancy engineering safety, and in particular to a system and method for analyzing the causes of dam safety based on multi-dimensional monitoring data. Background Technology

[0002] As the core facility of water conservancy projects, the safe and stable operation of dams is directly related to the realization of comprehensive benefits such as flood control, water supply, and power generation. Dam safety monitoring is a key means to ensure its long-term stable operation.

[0003] In existing technologies, the analysis of factors affecting dam safety mostly adopts traditional mathematical statistics methods such as multiple linear regression and time series analysis. These traditional mathematical statistics methods are used to calculate the degree of safety risk based on the dam's safety monitoring data. Alternatively, machine learning or deep learning methods such as support vector machines and long short-term memory networks are used to analyze the factors affecting dam safety.

[0004] However, traditional mathematical statistics methods cannot effectively distinguish causal relationships. When multiple influencing factors change simultaneously, spurious correlations are easily generated, leading to misjudgments of the dam's safety status. On the other hand, analysis methods based on support vector machines and long short-term memory networks have poor interpretability and cannot clearly identify the specific factors and action paths affecting the dam's status. Their feature importance analysis only reflects correlation and does not possess causality, thus limiting their ability to support engineering decision-making. Summary of the Invention

[0005] This invention provides a system, method, device, storage medium, and computer program product for analyzing the causes of dam safety based on multi-dimensional monitoring data, in order to solve the problems of large errors and poor interpretability in the analysis results of dam safety causes.

[0006] According to one aspect of the present invention, a dam safety cause analysis system based on multi-dimensional monitoring data is provided, comprising: a data acquisition module, a data preprocessing module, a causal inference module, and a decision support module connected in sequence; The data acquisition module is used to acquire multi-dimensional monitoring data of the dam; wherein, the multi-dimensional monitoring data includes structural response data and environmental monitoring data; The data preprocessing module is used to preprocess the multi-dimensional monitoring data; The causal inference module is used to obtain a real-time feature set based on the multi-dimensional monitoring data, a historical feature set based on historical monitoring data, a causal structure model based on the historical feature set, and a causal effect quantification result based on the causal structure model. The decision support module is used to determine whether the dam has a safety risk based on the real-time feature set and the causal structure model. When the dam is determined to have a safety risk, the module obtains the safety cause nodes and causal transmission paths based on the causal structure model and the causal effect quantification results.

[0007] The causal inference module is specifically used to acquire the temporal and correlation features of the multi-dimensional monitoring data, so as to construct a real-time feature set based on the temporal and correlation features; wherein, the temporal features include trend features, mutation features, periodic features and statistical features; the correlation features reflect the correlation between the structural response data and the environmental monitoring data.

[0008] The causal inference module is specifically used to obtain a causal structure model based on the historical feature set and the spatiotemporal adjacency matrix through a causal discovery algorithm based on spatiotemporal constraints; wherein, the spatiotemporal adjacency matrix is ​​related to the location of the monitoring equipment and the monitoring time interval corresponding to the historical monitoring data.

[0009] The causal inference module is specifically used to obtain the residuals of the processing variables through a gradient boosting tree and the residuals of the outcome variables through a random forest, so as to obtain the causal effect quantification results based on the residuals of the processing variables and the residuals of the outcome variables.

[0010] The decision support module is specifically used to obtain the predicted normal value of the structural response data through the causal structure model based on the environmental monitoring data, and to obtain the anomaly score of the multi-dimensional monitoring data based on the real-time monitoring value of the structural response data and the predicted normal value, so as to determine whether there is a safety risk to the dam based on the anomaly score.

[0011] The dam safety cause analysis system based on multi-dimensional monitoring data also includes a visualization application module; the visualization application module connects the data preprocessing module, the causal inference module and the decision support module, and is used to display the multi-dimensional monitoring data, the causal structure model, the causal effect quantification results, the safety cause nodes and the causal transmission path.

[0012] According to another aspect of the present invention, a method for analyzing the causes of dam safety based on multi-dimensional monitoring data is provided, comprising: The data acquisition module obtains multi-dimensional monitoring data of the dam; wherein, the multi-dimensional monitoring data includes structural response data and environmental monitoring data; The data preprocessing module preprocesses the multi-dimensional monitoring data; The causal inference module obtains a real-time feature set based on the multi-dimensional monitoring data, a historical feature set based on historical monitoring data, a causal structure model based on the historical feature set, and a causal effect quantification result based on the causal structure model. The decision support module determines whether the dam has a safety risk based on the real-time feature set and the causal structure model. When it determines that the dam has a safety risk, it obtains the safety cause nodes and causal transmission paths based on the causal structure model and the causal effect quantification results.

[0013] According to another aspect of the present invention, a device for analyzing the causes of dam safety based on multi-dimensional monitoring data is provided, comprising: A monitoring data acquisition module, configured within the data acquisition module, is used to acquire multi-dimensional monitoring data of the dam; wherein, the multi-dimensional monitoring data includes structural response data and environmental monitoring data; The preprocessing execution module, configured in the data preprocessing module, is used to preprocess the multi-dimensional monitoring data; The quantification result acquisition module is configured in the causal inference module and is used to acquire a real-time feature set based on the multi-dimensional monitoring data, acquire a historical feature set based on historical monitoring data, acquire a causal structure model based on the historical feature set, and acquire causal effect quantification results based on the causal structure model. The causal result acquisition module, configured in the decision support module, is used to determine whether the dam has a safety risk based on the real-time feature set and the causal structure model. When the dam is determined to have a safety risk, the module acquires the safety causal nodes and causal transmission paths based on the causal structure model and the causal effect quantification results.

[0014] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a processor to execute and implement the dam safety cause analysis method based on multi-dimensional monitoring data as described in any embodiment of the present invention.

[0015] According to another aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the dam safety cause analysis method based on multi-dimensional monitoring data as described in any embodiment of the present invention.

[0016] The technical solution of this invention involves a data acquisition module acquiring multi-dimensional monitoring data of a dam; a data preprocessing module preprocessing the multi-dimensional monitoring data; a causal inference module acquiring a real-time feature set based on the multi-dimensional monitoring data, a historical feature set based on historical monitoring data, a causal structure model based on the historical feature set, and a causal effect quantification result based on the causal structure model; and a decision support module determining whether the dam has a safety risk based on the real-time feature set and the causal structure model, and when a safety risk is determined, acquiring the safety causal nodes and causal transmission paths based on the causal structure model and the causal effect quantification results. This not only effectively distinguishes the causal relationships between various monitoring data, avoiding spurious associations arising from simultaneous changes in multiple influencing factors and improving the accuracy of dam safety status determination results, but also acquires the safety causal nodes and causal transmission paths affecting dam safety, enhancing the interpretability of the dam safety causal analysis results and greatly improving the support capability for engineering decision-making.

[0017] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of a dam safety cause analysis system based on multi-dimensional monitoring data provided in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of another dam safety cause analysis system based on multi-dimensional monitoring data provided in Embodiment 1 of the present invention; Figure 3 This is a schematic diagram of a causal directed acyclic graph provided in Embodiment 1 of the present invention; Figure 4 This is a flowchart of a method for analyzing the causes of dam safety based on multi-dimensional monitoring data, provided in Embodiment 2 of the present invention; Figure 5 This is a flowchart of another method for analyzing the causes of dam safety based on multi-dimensional monitoring data, provided in Embodiment 2 of the present invention; Figure 6This is a structural block diagram of a dam safety cause analysis device based on multi-dimensional monitoring data, provided in Embodiment 3 of the present invention. Detailed Implementation

[0020] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0021] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0022] Example 1 Figure 1 This is a schematic diagram of a dam safety cause analysis system based on multi-dimensional monitoring data provided in Embodiment 1 of the present invention, as shown below. Figure 1 As shown, the system includes a data acquisition module 100, a data preprocessing module 200, a causal inference module 300, and a decision support module 400 connected in sequence; the data acquisition module 100 is used to acquire multi-dimensional monitoring data of the dam; wherein, the multi-dimensional monitoring data includes structural response data and environmental monitoring data; the data preprocessing module 200 is used to preprocess the multi-dimensional monitoring data.

[0023] Specifically, the data acquisition module 100 can acquire different types of monitoring data through various monitoring devices deployed near the dam. Among them, structural response data can include displacement, seepage pressure, strain force, and uplift pressure. Displacement refers to the distance the dam body moves in three-dimensional space, which reflects the direct deformation result under load and can be acquired through monitoring devices such as GNSS (Global Navigation Satellite System) receivers, displacement gauges, crack gauges, and inclinometers. Seepage pressure refers to the pressure generated by water flowing in the dam body, which reflects the force of the seepage field on the dam and can be acquired through a piezometer.

[0024] Strain force refers to the deformation force occurring inside the dam, reflecting the stress state of the dam materials, and can be obtained through strain gauges; uplift pressure refers to the upward water pressure acting on the bottom of the dam foundation, reflecting the external load on the dam, and can be obtained through piezometers. Environmental monitoring data can include reservoir water level, rainfall, air temperature, and humidity; reservoir water level refers to the elevation of the water surface in front of the dam, which directly reflects the water storage capacity and can be obtained through water level gauges; rainfall refers to the depth of rainwater falling on the horizontal surface per unit time and can be obtained through rain gauges; air temperature refers to the temperature of the atmosphere and can be obtained through temperature sensors; humidity refers to the water vapor content in the air and can be obtained through humidity sensors.

[0025] Preprocessing can include data cleaning, spatiotemporal alignment, data denoising, and standardization. Data cleaning removes gross errors, missing values, and anomalous jumps from multi-dimensional monitoring data to ensure data integrity and accuracy; this can be achieved using adaptive thresholding methods. Spatiotemporal alignment matches different monitoring points and types of data based on the physical deployment locations (spatial coordinates) and monitoring time intervals to ensure spatiotemporal consistency. Data denoising filters out random noise while preserving the true patterns of data change; this can be performed using methods such as wavelet filtering and cluster analysis. Standardization normalizes the dataset after data cleaning, spatiotemporal alignment, and denoising to eliminate the influence of different data units and orders of magnitude, resulting in a standardized monitoring dataset.

[0026] The causal inference module 300 is used to obtain a real-time feature set based on the multi-dimensional monitoring data, a historical feature set based on historical monitoring data, a causal structure model based on the historical feature set, and a causal effect quantification result based on the causal structure model. The historical monitoring data has the same data type as the aforementioned multi-dimensional monitoring data, including structural response data (i.e., historical structural response data) and environmental monitoring data (i.e., historical environmental monitoring data). Each data type in the historical monitoring data includes continuous monitoring data over a relatively long period.

[0027] The feature set corresponding to multi-dimensional monitoring data (i.e., the implementation feature set) can be obtained through deep learning models. For example, nonlinear local features in the monitoring equipment network can be extracted through convolutional neural networks, such as the spatial distribution characteristics of stress concentration areas in dams and the spatial distribution characteristics of seepage anomalies. Bidirectional time-series features can be captured through bidirectional long short-term memory networks, such as the lag effect of seasonal changes in reservoir water level on dam deformation and periodic displacement caused by temperature loads. Based on this, multi-dimensional monitoring data can be transformed into quantifiable and analyzable feature indicators. The method for obtaining historical feature sets is the same as that for obtaining real-time feature sets.

[0028] Historical feature sets can serve as training samples for causal structure models, enabling causal structure learning. The causal structure model, after causal structure learning, is a qualitative and structured model, which can be represented as a directed acyclic graph (DAG). Nodes represent variables (e.g., reservoir water level, displacement, seepage flow), and arrows represent causal directions. Furthermore, the causal structure model can include multiple structural equations, each formally describing how each variable is determined by its parent node and random noise. Training the causal structure model can be achieved through the following methods: First, the aforementioned historical feature set is used as input data and initialized into a completely undirected initial undirected graph using a causal discovery algorithm (e.g., a time-series-based causal discovery algorithm). Here, the causal discovery algorithm refers to a method for inferring causal relationships between variables from observation data. The nodes of the initial undirected graph are the monitoring data of the aforementioned dimensions. Second, for each pair of adjacent nodes in the initial undirected graph, given their set of neighboring nodes, the Fisher's Z-test is used to test conditional independence, so that the node patterns conform to the physical laws of hydraulic engineering while deleting edges with no statistical correlation.

[0029] Then, using the V-structure rule and directional propagation rule, the remaining edges after the independence test are initially oriented, that is, the causal edges are oriented. At the same time, temporal constraints and engineering mechanism constraints are introduced. The temporal constraints prohibit variables that occur later in time from being causes of variables that occur earlier in time, and the engineering mechanism constraints prohibit causal orientation that violates the mechanism of hydraulic engineering. Based on this, the initial undirected graph is transformed into the final output causal directed acyclic graph representing the causal relationship of the factors affecting dam safety. The causal directed acyclic graph marks the causal transmission path between each variable, for example, rainfall → reservoir water level → seepage pressure → displacement.

[0030] Subsequently, the impact of various monitoring data on dam safety was analyzed based on a causal structure model to quantify the causal effect. First, treatment variables (e.g., reservoir water level and rainfall), outcome variables (e.g., displacement and seepage pressure), and confounding variables (e.g., temperature and dam operating years) that are related to both treatment and outcome variables were identified in the aforementioned causal directed acyclic graph to eliminate the interference of different variables on the estimation of causal effects. Second, regression models of treatment variables on confounding variables and regression models of outcome variables on confounding variables were fitted using machine learning models to obtain the residuals of treatment variables and outcome variables.

[0031] Then, a linear regression is performed on the treatment variable residuals using the residuals of the outcome variable. The regression coefficients are the average treatment effect (ATE) after correction. At the same time, the conditional average treatment effect (CATE) of each parent node on the child node is calculated, which is the effect value under different working conditions. The above average treatment effect and conditional average treatment effect are the quantitative results of the causal effect of the causal directed acyclic graph. It reflects the estimated value of the causal effect of the causal directed acyclic graph, that is, it reflects how much influence one variable has on another variable. It is a quantitative and calculable indicator value.

[0032] Optionally, in this embodiment of the invention, the causal inference module 300 is specifically used to acquire the temporal features and correlation features of the multi-dimensional monitoring data, so as to construct a real-time feature set based on the temporal features and the correlation features; wherein, the temporal features include trend features, mutation features, periodic features and statistical features; the correlation features reflect the correlation between the structural response data and the environmental monitoring data.

[0033] Specifically, trend characteristics refer to the current data change trend (e.g., upward trend, downward trend, stable trend, etc.), which can be obtained by sliding window mean or linear regression fitting slope; abrupt change characteristics refer to the sudden increase or decrease in the value of the current monitoring data, which can be obtained by Cumulative Sum (CUSUM) algorithm or derivative abrupt change detection method; periodic characteristics refer to the periodic changes of the current monitoring data such as daily, monthly, quarterly, etc., which can be obtained by Fourier transform or autocorrelation analysis; statistical characteristics refer to the statistical calculation results of the current monitoring data, such as mean, variance, extreme values, etc.

[0034] While obtaining the aforementioned time-series features, the correlation coefficients between structural response data (e.g., displacement) and environmental monitoring data (e.g., reservoir water level and rainfall) are calculated using Pearson correlation coefficient, Spearman rank correlation coefficient, covariance analysis, or other higher-order statistical methods. Strongly correlated variables are then selected. By integrating the aforementioned time-series features and correlation features, a dam safety feature set covering the spatiotemporal dimensions is formed. For example, the dam safety feature set can be represented in the form of a structured matrix, where each row represents a time point and each column represents a feature variable. Based on this, the real-time feature set constructed using time-series features and correlation features greatly improves the data representation capability of real-time dam monitoring data, ensuring the accuracy of subsequent dam safety status determination results.

[0035] Optionally, in this embodiment of the invention, the causal inference module 300 is specifically used to obtain a causal structure model based on the historical feature set and the spatiotemporal adjacency matrix through a causal discovery algorithm based on spatiotemporal constraints; wherein, the spatiotemporal adjacency matrix is ​​related to the location of the monitoring equipment and the monitoring time interval corresponding to the historical monitoring data.

[0036] Specifically, the monitoring time interval is the time interval between each data sampling by the monitoring device. A spatiotemporal adjacency matrix is ​​constructed based on the physical location of each monitoring device and the monitoring time interval. This spatiotemporal adjacency matrix is ​​then used as a constraint for the causal discovery algorithm, thus forming a causal discovery algorithm based on spatiotemporal constraints. This ensures that when selecting the condition set through the causal discovery algorithm, adjacent nodes in the spatiotemporal adjacency matrix are given priority, effectively reducing combinatorial explosion in high-dimensional cases and greatly improving the construction efficiency of the causal structure model.

[0037] Optionally, in this embodiment of the invention, the causal inference module 300 is specifically used to obtain the residuals of the processing variables through a gradient boosting tree and the residuals of the result variables through a random forest, so as to obtain the causal effect quantification result based on the residuals of the processing variables and the residuals of the result variables.

[0038] Specifically, Gradient Boosted Decision Trees (GBDT) are based on negative gradient iteration, approximating the residuals using the negative gradient of the loss function. This means the direction of the prediction error guides the construction of the new tree, thereby improving the accuracy of obtaining residuals for the processed variables. Furthermore, by allowing for a custom loss function, GBDT can adapt to diverse task requirements, enhancing the computational flexibility of residual processing. Random Forest (RF) not only achieves efficient parallel computation, improving the computational efficiency of residuals, but also uses out-of-bag error to evaluate model performance, improving the interpretability of residuals. Thus, by combining GBDT with Random Forest—a dual machine learning model—confounding biases are eliminated, significantly improving the accuracy of causal effect quantification results.

[0039] The decision support module 400 is used to determine whether there is a safety risk to the dam based on the real-time feature set and the causal structure model. When it is determined that there is a safety risk to the dam, it obtains the safety cause nodes and causal transmission paths based on the causal structure model and the causal effect quantification results.

[0040] Specifically, multi-dimensional monitoring data, as the currently collected real-time data, is input into a pre-trained causal structure model. The causal structure model then performs online inference on the real-time data to determine whether there is a safety risk to the dam. Different normal thresholds can be set for different types of monitoring data, and each monitoring data point is compared with its corresponding normal threshold. If none of the monitoring data points exceed the normal threshold, the dam is determined to have no safety risk. If at least one monitoring data point exceeds the normal threshold, the dam is determined to have a safety risk.

[0041] When determining that a dam poses a safety risk, a reverse search is performed using a causal directed acyclic graph. Starting from the node with the abnormal outcome variable, all its upstream parent nodes and indirect causal nodes are traced. Simultaneously, the average treatment effect and conditional average treatment effect obtained from the causal effect quantification results are combined to calculate the causal contribution of each upstream node to the abnormal outcome variable. The magnitude of the causal contribution reflects the degree of influence of the node on the dam's safety anomaly. Based on the ranking of causal contributions and combined with the actual mechanism of the water conservancy project, the core root cause node (i.e., the safety cause node) leading to the dam's safety anomaly can be accurately located. Furthermore, the causal transmission path of the anomaly can be determined through the core root cause node and the abnormal variable node, thus providing a clear basis for targeted treatment of dam safety anomalies.

[0042] Optionally, in this embodiment of the invention, the decision support module 400 is specifically used to obtain the predicted normal value of the structural response data through the causal structure model based on the environmental monitoring data, and to obtain the anomaly score of the multi-dimensional monitoring data based on the real-time monitoring value of the structural response data and the predicted normal value, so as to determine whether there is a safety risk to the dam based on the anomaly score.

[0043] Specifically, since the performance of structural response data may vary under different environmental monitoring data, and the causal structure model has learned the causal relationship between monitoring data (i.e. variables) based on historical monitoring data, it can predict the normal value of structural response data under different environmental monitoring data based on the current environmental monitoring data. That is, it obtains the predicted normal value of structural response data, then compares the real-time monitoring value with the predicted normal value, calculates the degree of deviation between the real-time monitoring value and the predicted normal value, and quantifies the degree of deviation as an anomaly score. The greater the degree of deviation, the higher the anomaly score, and the greater the probability of an anomaly.

[0044] If the anomaly score corresponding to each monitoring data point is less than or equal to the preset safety threshold, the dam is determined to have no safety risk. If the anomaly score corresponding to at least one monitoring data point is greater than the preset safety threshold, the dam is determined to have a safety risk. In this case, tracing is performed starting from the node with the greatest safety risk. Using the tracing method described in the above technical solution, the core root cause node leading to the dam safety anomaly is accurately located. At the same time, the causal transmission path of the anomaly is determined through the core root cause node and the node with the greatest safety risk. Compared with setting a static threshold, obtaining the predicted normal value of the structural response data through the causal structure model provides a more accurate threshold range for multi-dimensional monitoring data based on environmental factors, further improving the accuracy of the dam safety status detection results and avoiding misjudgment of the dam safety status.

[0045] like Figure 2 As shown, optionally, in this embodiment of the invention, the dam safety cause analysis system based on multi-dimensional monitoring data further includes a visualization application module 500; the visualization application module 500 is connected to the data preprocessing module 200, the causal inference module 300 and the decision support module 400, and is used to display the multi-dimensional monitoring data, the causal structure model, the causal effect quantification results, the safety cause nodes and the causal transmission path.

[0046] Specifically, the visualization application module 500 can dynamically display the aforementioned multi-dimensional monitoring data, causal structure models (including causal directed acyclic graphs), causal effect quantification results, safety cause nodes, and causal transmission paths through a visualization platform on the Web (World Wide Web) and mobile terminals. This enables the visualization and intuitive presentation of raw data and analysis results, while also supporting real-time querying and historical backtracking of monitoring data, providing a convenient operating interface and decision-making reference for managers who maintain dam safety.

[0047] Figure 3 This is a schematic diagram of the causal directed acyclic graph of the causal structure model obtained for training. The circular nodes represent the multi-dimensional monitoring data of the dam (i.e., various variables), including environmental variables (e.g., increased rainfall, changes in reservoir temperature), structural attribute variables (e.g., dam's service life, concrete aging), structural response variables (e.g., rising reservoir water level, increased seepage pressure, abnormal dam displacement), and safety warning variables (e.g., structural safety warning). The directed edges with arrows indicate the causal relationship and transmission direction between variables, with the starting point of the arrow being the "dependent variable" and the ending point being the "effect variable". Figure 3 The core causal transmission path is marked: increased rainfall → rising reservoir water level → increased seepage pressure → abnormal dam displacement → structural safety warning. This intuitively demonstrates the causal relationship and transmission law between various influencing factors and the safety status of the dam, reflecting the core output form of the causal structure learning of this invention.

[0048] The technical solution of this invention involves a data acquisition module acquiring multi-dimensional monitoring data of a dam; a data preprocessing module preprocessing the multi-dimensional monitoring data; a causal inference module acquiring a real-time feature set based on the multi-dimensional monitoring data, a historical feature set based on historical monitoring data, a causal structure model based on the historical feature set, and a causal effect quantification result based on the causal structure model; and a decision support module determining whether the dam has a safety risk based on the real-time feature set and the causal structure model, and when a safety risk is determined, acquiring the safety causal nodes and causal transmission paths based on the causal structure model and the causal effect quantification results. This not only effectively distinguishes the causal relationships between various monitoring data, avoiding spurious associations arising from simultaneous changes in multiple influencing factors and improving the accuracy of dam safety status determination results, but also acquires the safety causal nodes and causal transmission paths affecting dam safety, enhancing the interpretability of the dam safety causal analysis results and greatly improving the support capability for engineering decision-making.

[0049] Example 2 Figure 4This is a flowchart of a dam safety causal analysis method based on multi-dimensional monitoring data provided in Embodiment 2 of the present invention. This method can be executed by a dam safety causal analysis device based on multi-dimensional monitoring data. This device can be implemented in hardware and / or software, and can be configured within the dam safety causal analysis system based on multi-dimensional monitoring data described in Embodiment 1. Figure 4 As shown, the method includes: S101, The data acquisition module acquires multi-dimensional monitoring data of the dam; wherein, the multi-dimensional monitoring data includes structural response data and environmental monitoring data.

[0050] S102, The data preprocessing module preprocesses the multi-dimensional monitoring data.

[0051] S103. The causal inference module obtains a real-time feature set based on the multi-dimensional monitoring data, a historical feature set based on historical monitoring data, a causal structure model based on the historical feature set, and a causal effect quantification result based on the causal structure model.

[0052] S104. The decision support module determines whether the dam has a safety risk based on the real-time feature set and the causal structure model. When it determines that the dam has a safety risk, it obtains the safety cause nodes and causal transmission paths based on the causal structure model and the causal effect quantification results.

[0053] like Figure 5 As shown, the data acquisition module collects multi-dimensional monitoring data (i.e., raw data); the data preprocessing module preprocesses the collected multi-dimensional monitoring data; the causal inference module extracts features from the multi-dimensional monitoring data to construct a real-time feature set, which is then input into the causal structure model; prior to this, the causal inference module has constructed a historical feature set based on historical monitoring data, obtained a causal structure model based on the historical feature set, and obtained the causal effect quantification results based on the causal structure model.

[0054] Based on the aforementioned real-time feature set, the decision support module performs online causal reasoning through a causal structure model to determine whether the dam poses a safety risk. If the dam is determined to pose no safety risk, multi-dimensional monitoring data is collected for the next round of dam status assessment. If a safety risk is determined to exist, anomaly root cause tracing is triggered to obtain the safety causal nodes and causal transmission paths. The aforementioned safety causal nodes and causal transmission paths are output as analysis results to the visualization application module for display.

[0055] The technical solution of this invention involves a data acquisition module acquiring multi-dimensional monitoring data of a dam; a data preprocessing module preprocessing the multi-dimensional monitoring data; a causal inference module acquiring a real-time feature set based on the multi-dimensional monitoring data, a historical feature set based on historical monitoring data, a causal structure model based on the historical feature set, and a causal effect quantification result based on the causal structure model; and a decision support module determining whether the dam has a safety risk based on the real-time feature set and the causal structure model, and when a safety risk is determined, acquiring the safety causal nodes and causal transmission paths based on the causal structure model and the causal effect quantification results. This not only effectively distinguishes the causal relationships between various monitoring data, avoiding spurious associations arising from simultaneous changes in multiple influencing factors and improving the accuracy of dam safety status determination results, but also acquires the safety causal nodes and causal transmission paths affecting dam safety, enhancing the interpretability of the dam safety causal analysis results and greatly improving the support capability for engineering decision-making.

[0056] Example 3 Figure 6 This is a structural block diagram of a dam safety cause analysis device based on multi-dimensional monitoring data provided in Embodiment 3 of the present invention. This device can be configured in the dam safety cause analysis system based on multi-dimensional monitoring data described in Embodiment 1, and specifically includes: The monitoring data acquisition module 601 is configured in the data acquisition module and is used to acquire multi-dimensional monitoring data of the dam; wherein, the multi-dimensional monitoring data includes structural response data and environmental monitoring data; The preprocessing execution module 602 is configured in the data preprocessing module and is used to preprocess the multi-dimensional monitoring data; The quantification result acquisition module 603 is configured in the causal inference module and is used to acquire a real-time feature set based on the multi-dimensional monitoring data, acquire a historical feature set based on historical monitoring data, acquire a causal structure model based on the historical feature set, and acquire causal effect quantification results based on the causal structure model. The causal result acquisition module 604 is configured in the decision support module and is used to determine whether there is a safety risk to the dam based on the real-time feature set and the causal structure model. When it is determined that there is a safety risk to the dam, the module acquires the safety cause nodes and causal transmission paths based on the causal structure model and the causal effect quantification results.

[0057] The technical solution of this invention involves a data acquisition module acquiring multi-dimensional monitoring data of a dam; a data preprocessing module preprocessing the multi-dimensional monitoring data; a causal inference module acquiring a real-time feature set based on the multi-dimensional monitoring data, a historical feature set based on historical monitoring data, a causal structure model based on the historical feature set, and a causal effect quantification result based on the causal structure model; and a decision support module determining whether the dam has a safety risk based on the real-time feature set and the causal structure model, and when a safety risk is determined, acquiring the safety causal nodes and causal transmission paths based on the causal structure model and the causal effect quantification results. This not only effectively distinguishes the causal relationships between various monitoring data, avoiding spurious associations arising from simultaneous changes in multiple influencing factors and improving the accuracy of dam safety status determination results, but also acquires the safety causal nodes and causal transmission paths affecting dam safety, enhancing the interpretability of the dam safety causal analysis results and greatly improving the support capability for engineering decision-making.

[0058] The above-described apparatus can execute the dam safety causal analysis method based on multi-dimensional monitoring data provided in any embodiment of the present invention, and possesses the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the dam safety causal analysis method based on multi-dimensional monitoring data provided in any embodiment of the present invention.

[0059] Example 4 In some embodiments, the dam safety causal analysis method based on multi-dimensional monitoring data can be implemented as a computer program tangibly contained in a computer-readable storage medium, such as a storage unit. In some embodiments, part or all of the computer program can be loaded and / or installed on a heterogeneous hardware accelerator via ROM and / or a communication unit. When the computer program is loaded into RAM and executed by a processor, one or more steps of the dam safety causal analysis method based on multi-dimensional monitoring data described above can be performed. Alternatively, in other embodiments, the processor can be configured to perform the dam safety causal analysis method based on multi-dimensional monitoring data by any other suitable means (e.g., by means of firmware).

[0060] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0061] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0062] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0063] To provide interaction with a user terminal, the systems and techniques described herein can be implemented on a heterogeneous hardware accelerator, which includes: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user terminal; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user terminal provides input to the heterogeneous hardware accelerator. Other types of devices can also be used to provide interaction with the user terminal; for example, the feedback provided to the user terminal can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or haptic feedback); and input from the user terminal can be received in any form (including sound input, voice input, or haptic input).

[0064] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., client computers with graphical user interfaces or web browsers through which client computers can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0065] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0066] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0067] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A dam safety causal analysis system based on multi-dimensional monitoring data, characterized in that, include: The data acquisition module, data preprocessing module, causal inference module, and decision support module are connected sequentially. The data acquisition module is used to acquire multi-dimensional monitoring data of the dam; wherein, the multi-dimensional monitoring data includes structural response data and environmental monitoring data; The data preprocessing module is used to preprocess the multi-dimensional monitoring data; The causal inference module is used to obtain a real-time feature set based on the multi-dimensional monitoring data, a historical feature set based on historical monitoring data, a causal structure model based on the historical feature set, and a causal effect quantification result based on the causal structure model. The decision support module is used to determine whether the dam has a safety risk based on the real-time feature set and the causal structure model. When the dam is determined to have a safety risk, the module obtains the safety cause nodes and causal transmission paths based on the causal structure model and the causal effect quantification results.

2. The dam safety cause analysis system based on multi-dimensional monitoring data according to claim 1, characterized in that, The causal inference module is specifically used to acquire the temporal and correlation features of the multi-dimensional monitoring data, so as to construct a real-time feature set based on the temporal and correlation features; wherein, the temporal features include trend features, mutation features, periodic features and statistical features; the correlation features reflect the correlation between the structural response data and the environmental monitoring data.

3. The dam safety cause analysis system based on multi-dimensional monitoring data according to claim 1, characterized in that, The causal inference module is specifically used to obtain a causal structure model based on the historical feature set and the spatiotemporal adjacency matrix through a causal discovery algorithm based on spatiotemporal constraints; wherein, the spatiotemporal adjacency matrix is ​​related to the location of the monitoring equipment and the monitoring time interval corresponding to the historical monitoring data.

4. The dam safety cause analysis system based on multi-dimensional monitoring data according to claim 1, characterized in that, The causal inference module is specifically used to obtain the residuals of the processing variables through a gradient boosting tree and the residuals of the outcome variables through a random forest, so as to obtain the causal effect quantification results based on the residuals of the processing variables and the residuals of the outcome variables.

5. The dam safety cause analysis system based on multi-dimensional monitoring data according to claim 1, characterized in that, The decision support module is specifically used to obtain the predicted normal value of the structural response data through the causal structure model based on the environmental monitoring data, and to obtain the anomaly score of the multi-dimensional monitoring data based on the real-time monitoring value of the structural response data and the predicted normal value, so as to determine whether there is a safety risk to the dam based on the anomaly score.

6. The dam safety cause analysis system based on multi-dimensional monitoring data according to claim 1, characterized in that, The dam safety cause analysis system based on multi-dimensional monitoring data also includes a visualization application module; The visualization application module connects the data preprocessing module, the causal inference module, and the decision support module, and is used to display the multi-dimensional monitoring data, the causal structure model, the causal effect quantification results, the safety cause nodes, and the causal transmission path.

7. A method for analyzing the causes of dam safety based on multi-dimensional monitoring data, characterized in that, include: The data acquisition module obtains multi-dimensional monitoring data of the dam; wherein, the multi-dimensional monitoring data includes structural response data and environmental monitoring data; The data preprocessing module preprocesses the multi-dimensional monitoring data; The causal inference module obtains a real-time feature set based on the multi-dimensional monitoring data, a historical feature set based on historical monitoring data, a causal structure model based on the historical feature set, and a causal effect quantification result based on the causal structure model. The decision support module determines whether the dam has a safety risk based on the real-time feature set and the causal structure model. When it determines that the dam has a safety risk, it obtains the safety cause nodes and causal transmission paths based on the causal structure model and the causal effect quantification results.

8. A device for analyzing the causes of dam safety based on multi-dimensional monitoring data, characterized in that, include: A monitoring data acquisition module, configured within the data acquisition module, is used to acquire multi-dimensional monitoring data of the dam; wherein, the multi-dimensional monitoring data includes structural response data and environmental monitoring data; The preprocessing execution module, configured in the data preprocessing module, is used to preprocess the multi-dimensional monitoring data; The quantification result acquisition module is configured in the causal inference module and is used to acquire a real-time feature set based on the multi-dimensional monitoring data, acquire a historical feature set based on historical monitoring data, acquire a causal structure model based on the historical feature set, and acquire causal effect quantification results based on the causal structure model. The causal result acquisition module, configured in the decision support module, is used to determine whether the dam has a safety risk based on the real-time feature set and the causal structure model. When the dam is determined to have a safety risk, the module acquires the safety causal nodes and causal transmission paths based on the causal structure model and the causal effect quantification results.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the dam safety cause analysis method based on multi-dimensional monitoring data as described in claim 7.

10. A computer program product comprising a computer program that, when executed by a processor, implements the dam safety causation analysis method based on multi-dimensional monitoring data as described in claim 7.

Images