Multi-source data fusion and risk early warning system for reservoir dam safety monitoring

Abstract

This invention relates to the field of dam safety monitoring technology, specifically to a multi-source data fusion and risk early warning system for reservoir dam safety monitoring. The system includes: a data acquisition module, which acquires various monitoring data, temperature, and water level at each monitoring point on the dam at each moment within a preset time period prior to the current moment; and designates any monitoring point as a target monitoring point; a relative change analysis module, which obtains the relative change value of that monitoring data at a given moment for a monitoring point based on the ratio of the rate of change of a certain monitoring data at a given moment to that at a given moment for a monitoring point; an environmental analysis module, which analyzes the temperature and water level at each moment to obtain the degree of temperature and water level influence of each type of monitoring data; and a monitoring module, which acquires the normal relative change values ​​of various monitoring data at each monitoring point, then obtains the safety risk level of that monitoring data at the target monitoring point at the current moment, and subsequently monitors the target monitoring point. This application can effectively monitor the safety of dams.

Description

Technical Field

[0001] This invention relates to the field of dam safety monitoring technology, specifically to a multi-source data fusion and risk early warning system for reservoir dam safety monitoring. Background Technology

[0002] A reservoir dam is an important type of water conservancy infrastructure. It intercepts water flow to form a reservoir, thereby achieving water storage, regulation, and purposes such as flood control, water supply, and irrigation. The safe and stable operation of a reservoir dam is directly related to flood control safety, water supply security, and the safety of surrounding residents' property. Therefore, reservoir dam safety monitoring is crucial for the operation of reservoir dams.

[0003] Existing methods for monitoring the safety of reservoir dams primarily determine the presence of safety risks by monitoring whether various data points exceed normal ranges. However, some reservoir dams may exhibit abnormal safety risks even if the data changes within normal ranges, but the patterns of these changes are not normal. Traditional methods often fail to identify these trends, leading to delayed anomaly detection. Furthermore, environmental conditions such as temperature and water level can also cause changes in monitoring data. Traditional methods struggle to distinguish between data changes caused by environmental factors and actual dam structural anomalies, potentially resulting in inaccurate early warnings. Summary of the Invention

[0004] To address the aforementioned technical problems, the present invention aims to provide a multi-source data fusion and risk early warning system for reservoir dam safety monitoring. The specific technical solution adopted is as follows: One embodiment of the present invention provides a multi-source data fusion and risk early warning system for reservoir dam safety monitoring, the system comprising: The data acquisition module is used to acquire various monitoring data, temperature, and water level at each monitoring point on the dam at each time point within a preset time period before the current time; any monitoring point can be used as the target monitoring point. The relative change analysis module is used to obtain the rate of change of various monitoring data at each monitoring point at each time; to obtain the relative change value of that monitoring data at a monitoring point at a time based on the ratio of the rate of change of a certain monitoring data at a monitoring point to that of a target monitoring point at a time; and to form a sequence of relative change values ​​of that monitoring data at a monitoring point at each time within a preset time period by combining the relative change values ​​of that monitoring data at a monitoring point. The environmental analysis module is used to cluster data based on the temperature and water level at each time point within a preset time period at the target monitoring point. It then clusters data at each time point using the temperature as the clustering condition to obtain the temperature-water level distribution characteristic value and temperature distribution characteristic value for each type of monitoring data at the target monitoring point. Finally, it uses these characteristic values ​​to determine the degree of temperature influence for each type of monitoring data at the target monitoring point. Based on the degree of temperature influence for each type of monitoring data at each monitoring point, it obtains the degree of temperature influence for each type of monitoring data, and similarly, it obtains the degree of water level influence for each type of monitoring data. The monitoring module is used to obtain the environmental similarity of a monitoring data at a given moment based on the differences in temperature and water level between the target monitoring point and other monitoring points at a given moment, as well as the degree of influence of temperature and water level on a particular monitoring data. It also uses the environmental similarity of a monitoring data at different moments and the relative change value sequence of that monitoring data at a given monitoring point to obtain the normal relative change value of that monitoring data at that monitoring point. Furthermore, it uses the relative change value of a monitoring data at a given moment from other monitoring points and the normal relative change value of that monitoring data at other monitoring points to obtain the safety risk level of the target monitoring point at the given moment. Finally, it monitors the target monitoring point based on the safety risk level of various monitoring data at the target monitoring point at the given moment.

[0005] Preferably, obtaining the relative change value of the monitoring data at a given time point based on the ratio of the rate of change of a certain type of monitoring data at a given monitoring point to that of a target monitoring point at a given time point includes: The relative change value of a certain type of monitoring data at a given time point is obtained by dividing the rate of change of a certain type of monitoring data at a target monitoring point at that time point by the sum of the rate of change of that type of monitoring data at that target monitoring point and the hyperparameter.

[0006] Preferably, the temperature and water level at each moment within a preset time period of the target monitoring point are used as clustering conditions. Clustering is performed at each moment using the temperature as the clustering condition to obtain the temperature-water level distribution characteristic value and temperature distribution characteristic value for each type of monitoring data at the target monitoring point, including: The temperature and water level at each moment within a preset time period of the target monitoring point are used as clustering conditions to obtain different clusters. The median of a certain type of monitoring data at each moment of the target monitoring point in a cluster is used as the representative value of that type of monitoring data in that cluster. The reliability of the representative value of that type of monitoring data in that cluster is obtained by normalizing the sum of the standard deviation and hyperparameters of the monitoring data at each moment of the target monitoring point in a cluster. The absolute value of the difference between the representative values ​​of a certain type of monitoring data in two clusters is multiplied by the mean of the reliability values ​​of that type of monitoring data in the two clusters to obtain the representative difference between the two clusters. The mean of the representative differences between any two clusters is calculated and normalized to obtain the temperature-water level distribution characteristic value of that type of monitoring data at the target monitoring point. Similarly, the temperature-water level distribution characteristic value of each type of monitoring data at the target monitoring point can be obtained. Similarly, the temperature at each moment is used as a clustering condition to obtain the temperature distribution characteristic value of each type of monitoring data at the target monitoring point.

[0007] Preferably, the degree of temperature influence of each monitoring data point at the target monitoring point is obtained by utilizing the temperature-water level distribution characteristic value and temperature distribution characteristic value of each monitoring data point, including: The difference between the temperature-water level distribution characteristic value and the temperature distribution characteristic value of a certain type of monitoring data at the target monitoring point is compared with the temperature-water level distribution characteristic value of the same type of monitoring data at the target monitoring point and then normalized to obtain the degree of temperature influence of the same type of monitoring data at the target monitoring point.

[0008] Preferably, the temperature influence degree of each monitoring data point is obtained based on the temperature influence degree of each monitoring data point, including: The temperature influence of a monitoring data point is obtained by calculating the mean value of the temperature influence of that monitoring data point.

[0009] Preferably, the environmental similarity of the monitoring data at that moment is obtained based on the differences in temperature and water level between the target monitoring point at the current time and other monitoring points at a given time, and the degree of influence of temperature and water level on the monitoring data, including: The temperature weight of a monitoring data point is obtained by dividing its temperature influence by the sum of its temperature influence and water level influence. Similarly, the water level weight is obtained. The reciprocal of the temperature difference at that moment is obtained by adding the absolute values ​​of the temperature differences between the target monitoring point at the current moment and those at other monitoring points at a given moment to a hyperparameter. The reciprocal of the water level difference at that moment is obtained by adding the absolute values ​​of the water level differences between the target monitoring point at the current moment and those at other monitoring points at the same moment to a hyperparameter. The environmental similarity of the monitoring data at that moment is obtained by multiplying the normalized value of the reciprocal of the temperature difference at that moment by the temperature weight of the monitoring data point, and the product of the normalized value of the reciprocal of the water level difference at that moment by the water level weight of the monitoring data point.

[0010] Preferably, the normal relative change value of the monitoring data at a monitoring point is obtained by utilizing the environmental similarity of monitoring data at various times and the relative change value sequence of the monitoring data at a monitoring point, including: The weight of a monitoring data at a given moment is obtained by comparing the environmental similarity of a monitoring data at a given moment with the sum of the environmental similarity of the monitoring data at all moments. The weight of the monitoring data at each moment is then used to calculate the weighted average of the relative change values ​​of the monitoring data at a given monitoring point at each moment in the relative change value sequence, thus obtaining the normal relative change value of the monitoring data at that monitoring point.

[0011] Preferably, the safety risk level of the target monitoring point's monitoring data at the current moment is obtained based on the relative change value of a type of monitoring data at other monitoring points and the normal relative change value of the same type of monitoring data at other monitoring points, including: The absolute value of the difference between the relative change value of a certain type of monitoring data at each of the other monitoring points at the current moment and the normal relative change value of the same type of monitoring data at the corresponding other monitoring points is normalized to obtain the safety risk level of the target monitoring point at the current moment for that type of monitoring data.

[0012] Preferably, the target monitoring point is monitored based on the security risk level of various monitoring data at the current moment, including: Clustering is performed on the security risk level of a certain type of monitoring data at all historical moments for all monitoring points. With parameter k=3, three clusters are obtained. The average security risk level of each cluster is calculated. Based on the average security risk level from highest to lowest, the clusters corresponding to this type of monitoring data are categorized into high-risk, low-risk, and risk-free clusters. Similarly, high-risk, low-risk, and risk-free clusters are obtained for each type of monitoring data. If the security risk level of a certain type of monitoring data at a target monitoring point at the current moment belongs to either the high-risk or low-risk cluster, an anomaly warning is issued. If the security risk level of the same type of monitoring data at the target monitoring point at the current moment belongs to the risk-free cluster, no anomaly warning is issued.

[0013] The embodiments of the present invention have at least the following beneficial effects: This application obtains various monitoring data, temperature, and water level at each monitoring point on the dam at each time within a preset time period before the current time, and takes any monitoring point as the target monitoring point; then, it obtains the rate of change of various monitoring data at each time of each monitoring point, and then obtains the relative change value of that monitoring data at that time of the monitoring point based on the ratio of the rate of change of a monitoring data at a monitoring point to that of the target monitoring point at a time. Finally, a sequence of relative change values ​​of various monitoring data at each monitoring point can be obtained. Here, based on the continuity and transitivity between different monitoring locations of the reservoir dam, by calculating the relative change value of each monitoring data at other monitoring points relative to the target monitoring point, and then comparing it with the normal relative change value of other monitoring points relative to the target monitoring point, the abnormal performance of the data change trend of the target monitoring point can be analyzed. Compared with only comparing the differences in monitoring data values, it is more effective in showing the abnormal trend of data in the early stage of safety risks, which is conducive to improving the timeliness of safety risk warning. In the environmental analysis module, the impact of temperature and water level at various times on various monitoring data is considered. Based on the differences in monitoring data under different environmental conditions, the influence of each environmental factor on the monitoring data is analyzed, and the influence of temperature and water level on each monitoring data is obtained. Then, the environmental similarity of each monitoring data at each time is obtained. Then, the normal relative change value of the monitoring data at a monitoring point is obtained by using the environmental similarity of a monitoring data at each time and the relative change value sequence of the monitoring data at a monitoring point. This serves as the basis for anomaly identification and can effectively reduce the interference of non-reservoir dam structural anomalies on risk warning results, which is conducive to improving the accuracy of risk warning. Attached Figure Description

[0014] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the 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.

[0015] Figure 1 This is a system block diagram of a multi-source data fusion and risk early warning system for reservoir dam safety monitoring provided in an embodiment of the present invention. Detailed Implementation

[0016] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the multi-source data fusion and risk early warning system for reservoir dam safety monitoring proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0017] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0018] The following description, in conjunction with the accompanying drawings, details the specific scheme of the multi-source data fusion and risk early warning system for reservoir dam safety monitoring provided by this invention.

[0019] In this embodiment, the main application scenario of the present invention is to analyze various monitoring data of a reservoir dam to determine the safety of the reservoir, and to identify safety risks with abnormal trends by combining the correlation characteristics of the changing trends of different types of monitoring data at different locations on a normal reservoir dam and the environmental conditions of historical data, so as to improve the timeliness and accuracy of reservoir dam safety monitoring.

[0020] Please see Figure 1 The diagram illustrates a system block diagram of a multi-source data fusion and risk early warning system for reservoir dam safety monitoring provided by an embodiment of the present invention. The system includes the following modules: The data acquisition module is used to acquire various monitoring data, temperature and water level at each monitoring point on the dam at each time within a preset time period before the current time; any monitoring point can be used as the target monitoring point.

[0021] To compare the changes in monitoring data at different locations along the dam, multiple monitoring points are evenly set along the dam axis (the direction of the dam's extension). The interval between monitoring points can be adjusted according to the scale of the dam being monitored. Preferably, the interval between monitoring points in this application is set to 10m. For each monitoring point, it is necessary to monitor structural data related to the dam body and environmental condition data. The dam body structural data includes stress data monitored by stress gauges, strain data monitored by strain gauges, displacement data monitored by displacement gauges, and seepage pressure values ​​monitored by piezometers. All kinds of dam body structural data are collectively referred to as various monitoring data. Environmental condition data includes temperature and water level.

[0022] Therefore, various monitoring data, including temperature and water level, were obtained at each monitoring point on the dam at each moment within a preset time period prior to the current time. The preset time period was one month prior to the current time, with data collected every minute. The analysis focused on the patterns of data variation, using normal monitoring data from the preset time period (abnormal data was removed and supplemented using cubic spline interpolation). To facilitate comparative analysis of different types of monitoring data, all monitoring data, including temperature and water level, were normalized. All subsequent data used are normalized data. Furthermore, for ease of analysis and explanation, any given monitoring point was used as the target monitoring point for explanation.

[0023] The relative change analysis module is used to obtain the rate of change of various monitoring data at each monitoring point at each time; to obtain the relative change value of the monitoring data at a monitoring point at a time based on the ratio of the rate of change of a monitoring data at a monitoring point to that of a target monitoring point at a time; and to form a sequence of relative change values ​​of the monitoring data at a monitoring point at each time within a preset time period by combining the relative change values ​​of the monitoring data at a monitoring point at that time.

[0024] A reservoir dam is a continuous, integrated structure, so stress, deformation, and seepage are continuously distributed across different locations. This means that monitoring data changes at different locations within a normal dam are correlated. While the magnitude of these changes may differ, the direction of change is consistent, and the relative magnitudes of change are stable. However, an anomaly at a particular monitoring location disrupts this relative relationship between monitoring points. Therefore, it's possible to obtain the relative changes of various monitoring data from other points relative to the target monitoring point for anomaly monitoring.

[0025] The purpose of this application is to provide early warnings for monitoring data with abnormal trends. If the current monitoring data exceeds the normal threshold (the normal threshold is set by professionals based on experience), a risk warning will be issued directly, and the abnormal data type will be reported. For data that does not exceed the normal threshold, the following analysis will be performed.

[0026] First, based on the changing trend of each monitoring data point within its respective time period, the characteristic value of each data point at each moment is obtained. Since the data analysis process is similar for different types of dam structures, the analysis of one type of monitoring data is used as an example here.

[0027] The specific calculation model for the rate of change of a monitoring data point at a given time is as follows: , in, This represents the rate of change of the a-th type of monitoring data at the i-th time point of the u-th monitoring point. This represents the data value of the a-th type of monitoring data at the i-th time point of the u-th monitoring point. This indicates the time interval between adjacent monitoring times (here) ).

[0028] The influencing factors differ at different times, and the rate of change of data may vary. However, under normal conditions, the relative changes between different monitoring points are similar. Therefore, the relative change at each moment is obtained by comparing the rate of change of various monitoring data at different monitoring points with that of the target monitoring point at the same moment. The relative change value of that monitoring data at a given moment is obtained by comparing the rate of change of a certain type of monitoring data at a given monitoring point with that at the target monitoring point at a given moment.

[0029] Specifically, the relative change value of a monitoring data point at a given time is obtained by dividing the rate of change of a monitoring data point at a given time by the rate of change of the same monitoring data point at the target monitoring point at that time and the sum of the hyperparameters.

[0030] The specific calculation model for the relative change value of a monitoring point, a type of monitoring data, at a given time is as follows: , in, This represents the relative change value of the a-th type of monitoring data at the u-th monitoring point at the i-th time. This represents the rate of change of the a-th type of monitoring data at the i-th time point of the u-th monitoring point; This represents the rate of change of the a-th type of monitoring data at the i-th time point of the target monitoring point. These are hyperparameters used to ensure that the fractions are meaningful; they are set here. .

[0031] Similarly, for a type of monitoring data at a monitoring point, its relative change value relative to the target monitoring point at each moment can be obtained. Thus, by combining the relative change values ​​of a type of monitoring data at a monitoring point within a preset time period, a sequence of relative change values ​​for that type of monitoring data at that monitoring point can be formed. This allows the acquisition of relative change value sequences for various types of monitoring data at various monitoring points.

[0032] The environmental analysis module is used to cluster data based on the temperature and water level at each time point within a preset time period of the target monitoring point. It uses the temperature at each time point as the clustering condition to obtain the temperature-water level distribution characteristic value and temperature distribution characteristic value for each type of monitoring data at the target monitoring point. It then uses these characteristic values ​​to determine the degree of temperature influence for each type of monitoring data at the target monitoring point. Finally, it obtains the degree of temperature influence for each type of monitoring data at each monitoring point, and similarly, the degree of water level influence for each type of monitoring data.

[0033] Environmental conditions are one of the factors affecting the changes in various monitoring data. That is, the performance of normal monitoring data varies under different environmental conditions. In order to obtain the current normal data change pattern of the dam more accurately, it is necessary to combine the environmental conditions at different times and increase the weight of historical data that are closer to the environmental conditions at the current monitoring time, so as to obtain the normal data change pattern under the current environmental conditions.

[0034] Because different environmental factors affect different types of monitoring data for different reasons, the degree of influence may also vary. Therefore, it is necessary to analyze the degree of influence of each environmental condition on the monitoring data based on the changes in dam structure data under each environmental condition.

[0035] If the distribution of monitoring data is more concentrated in multiple categories when a certain environmental condition is used as a limiting condition compared to when it is used as an unrestricted condition, it indicates that the environmental condition has a greater impact on the data. If the data distribution is basically the same, it indicates that the environmental condition has a very small impact on the data. Therefore, based on the differences in data distribution under different conditions, we can analyze the degree of influence of each environmental factor on each monitoring data.

[0036] Since the dam structure data at different monitoring points are inherently different due to variations in stress and other factors, calculations need to be performed separately for each monitoring point. Here, we will take the target monitoring point as an example for analysis and explanation.

[0037] Using the temperature and water level at each time point within a preset time period as clustering conditions, and using the temperature at each time point as the clustering condition, clustering is performed on each time point to obtain the temperature-water level distribution characteristic value and temperature distribution characteristic value of each monitoring data of the target monitoring point.

[0038] Specifically, the temperature and water level at each moment within a preset time period of the target monitoring point are used as clustering conditions to obtain different clusters. The median of a type of monitoring data at each moment of the target monitoring point in a cluster is used as the representative value of that type of monitoring data in that cluster. The reliability of the representative value of that type of monitoring data in that cluster is obtained by normalizing the sum of the standard deviation and hyperparameters of the monitoring data at each moment of the target monitoring point in a cluster. The absolute value of the difference between the representative values ​​of a type of monitoring data in two clusters is multiplied by the mean of the reliability values ​​of that type of monitoring data in the two clusters to obtain the representative difference between the two clusters. The mean of the representative differences between any two clusters is calculated and normalized to obtain the temperature-water level distribution characteristic value of that type of monitoring data at the target monitoring point. Similarly, the temperature-water level distribution characteristic value of each type of monitoring data at the target monitoring point can be obtained. Similarly, the temperature at each moment is used as a clustering condition to obtain the temperature distribution characteristic value of each type of monitoring data at the target monitoring point.

[0039] For clustering based on temperature and water level, the k-means clustering method is used. The clustering parameter k is determined using the elbow method. Data points within the same cluster represent times when temperature and water level conditions are similar. The data distribution characteristics under temperature and water level conditions are analyzed based on the degree of difference in target data across different clusters. The representative value of a monitoring data point within a cluster is denoted as... , represents the representative value of the a-th monitoring data in the q-th cluster under the environmental conditions of temperature and water level.

[0040] If the target data within the same cluster is relatively concentrated, then the representative value is more representative of the overall level of the corresponding cluster. Therefore, the reliability of the representative value of each cluster is calculated based on the standard deviation of the target data in the same cluster.

[0041] The specific calculation model for the reliability of a cluster of monitoring data is as follows: , in, The subscript indicates the reliability of the representative value of the a-th type of monitoring data in the q-th cluster. This refers to the conditions under which temperature and water level are controlled; This represents the standard deviation of the a-th type of monitoring data in the q-th cluster. A smaller standard deviation indicates a more concentrated data distribution, meaning a higher level of reliability. These are hyperparameters used to ensure that the fractions are meaningful; they are set here. The norm function represents the normalization function.

[0042] The specific calculation model for the temperature-water level distribution characteristic values ​​of each type of monitoring data at the target monitoring point is as follows: , in, Let be the temperature-water level distribution characteristic value of the a-th type of monitoring data at the target monitoring point, representing the distribution characteristic value of this type of monitoring data under the conditions of temperature and water level; This represents the total number of pairwise combinations within the obtained clusters. Indicates the first In a cluster combination (here, the absolute value of the difference between the representative values ​​of the a-th monitoring data in the q-th and r-th cluster combinations), This indicates the reliability of the representative value of the a-th type of monitoring data in the q-th cluster. This indicates the reliability of the representative value of the a-th type of monitoring data in the r-th cluster. This represents the mean reliability of the cluster combination formed by the q-th cluster and the r-th cluster. Let represent the difference between the clusters formed by the q-th cluster and the r-th cluster, and norm denotes the normalization function.

[0043] Similarly, the temperature-water level distribution characteristic values ​​of each type of monitoring data at the target monitoring point can be obtained. Likewise, when using the temperature at each time moment as the clustering condition, the temperature distribution characteristic values ​​of each type of monitoring data at the target monitoring point under the temperature conditions can be obtained by following the method described above for obtaining the temperature-water level distribution characteristic values. This represents the temperature distribution characteristic value of the a-th type of monitoring data at the target monitoring point. Compared to The smaller the value, the more obvious the dispersion of data distribution under a single temperature condition, meaning that the temperature condition has a smaller impact on the target data. Therefore, the degree of influence of temperature conditions on the target data reflected by the target monitoring point data can be calculated.

[0044] Then, the temperature-water level distribution characteristic value and temperature distribution characteristic value of each monitoring data of the target monitoring point are used to obtain the degree of temperature influence of each monitoring data of the target monitoring point.

[0045] Specifically, the difference between the temperature-water level distribution characteristic value and the temperature distribution characteristic value of a certain type of monitoring data at the target monitoring point is compared with the temperature-water level distribution characteristic value of that type of monitoring data at the target monitoring point and normalized to obtain the degree of temperature influence of that type of monitoring data at the target monitoring point.

[0046] The specific calculation model for the degree of temperature influence of monitoring data at a target monitoring point is as follows: , in, This indicates the degree of temperature influence of the a-th type of monitoring data at the target monitoring point, where T represents temperature; The temperature-water level distribution characteristic value of the a-th type of monitoring data at the target monitoring point. Let be the temperature distribution characteristic value of the a-th type of monitoring data at the target monitoring point, and norm represent the normalization function.

[0047] Similarly, the temperature influence of the a-th type of monitoring data at each monitoring point can be obtained, and then the temperature influence of each type of monitoring data at each monitoring point can be obtained based on the temperature influence of each type of monitoring data. Specifically, the average value of the temperature influence of a type of monitoring data at each monitoring point is calculated to obtain the temperature influence of that type of monitoring data. Similarly, for water level, the water level influence of each type of monitoring data can be obtained.

[0048] The monitoring module is used to obtain the environmental similarity of a monitoring data at a given moment based on the differences in temperature and water level between the target monitoring point and other monitoring points at a given moment, as well as the degree of influence of temperature and water level on a particular monitoring data. It also uses the environmental similarity of a monitoring data at different moments and the relative change value sequence of that monitoring data at a given monitoring point to obtain the normal relative change value of that monitoring data at that monitoring point. Furthermore, it uses the relative change value of a monitoring data at a given moment from other monitoring points and the normal relative change value of that monitoring data at other monitoring points to obtain the safety risk level of the target monitoring point at the given moment. Finally, it monitors the target monitoring point based on the safety risk level of various monitoring data at the target monitoring point at the given moment.

[0049] The above obtained the degree of temperature influence and the degree of water level influence of each type of monitoring data. Furthermore, based on the difference between the temperature and water level of the target monitoring point at the current time and that of other monitoring points at a certain time, as well as the degree of temperature influence and the degree of water level influence of a type of monitoring data, the environmental similarity of that type of monitoring data at that time was obtained.

[0050] Specifically, the temperature weight of a monitoring data point is obtained by dividing its temperature influence by the sum of its temperature influence and water level influence. Similarly, the water level weight is obtained. The reciprocal of the sum of the absolute differences between the current temperature of the target monitoring point and the temperatures of other monitoring points at a given time is obtained by adding a hyperparameter to the sum of the absolute differences between the current temperature of the target monitoring point and the temperatures of other monitoring points at a given time. The reciprocal of the sum of the absolute differences between the current water levels of the target monitoring point and the water levels of other monitoring points at that time is obtained by adding a hyperparameter to the sum of the absolute differences between the current water levels of the target monitoring point and the water level weight of other monitoring points at that time. The environmental similarity of the monitoring data at that time is obtained by multiplying the normalized value of the reciprocal of the temperature difference at that time by the temperature weight of the monitoring data point, and the product of the normalized value of the reciprocal of the water level difference at that time by the water level weight of the monitoring data point.

[0051] A specific calculation model for the environmental similarity of monitoring data at a given time is as follows: , in, This represents the environmental similarity of the i-th type of monitoring data at the i-th time. This represents a hyperparameter used to ensure that the fraction is meaningful; it is set here. ; This indicates the number of monitoring points other than the target monitoring point; This represents the temperature at time i of the u-th monitoring point among the other monitoring points. This represents the temperature at the i-th time point of the target monitoring point. This represents the temperature weight of the a-th type of monitoring data. This is the reciprocal of the temperature difference at time i. The larger this value is, the higher the environmental similarity. This represents the water level at time i of the u-th monitoring point among the other monitoring points. This represents the water level at the i-th time point of the target monitoring point. This represents the water level weight of the a-th type of monitoring data. This is the reciprocal of the water level difference at time i. The larger this value is, the higher the environmental similarity.

[0052] Next, the normal relative change value of the monitoring data at a monitoring point is obtained by using the environmental similarity of the monitoring data at each time point and the relative change value sequence of the monitoring data at a monitoring point.

[0053] Specifically, the environmental similarity of a monitoring data at a certain moment is compared with the sum of the environmental similarities of the monitoring data at all moments to obtain the weight of the monitoring data at that moment; using the weight of the monitoring data at each moment, the relative change values ​​of the monitoring data at a certain monitoring point are weighted and averaged to obtain the normal relative change value of the monitoring data at that monitoring point.

[0054] The normal relative change values ​​of a type of monitoring data at each monitoring point can form a sequence. , This represents the normal relative change value of the a-th type of monitoring data at the first monitoring point. Indicates the first The normal relative change value of the a-th type of monitoring data at each monitoring point. The corresponding sequences can be obtained similarly for other types of monitoring data.

[0055] Therefore, the safety risk level of the target monitoring point at the current moment is obtained by comparing the relative change value of a type of monitoring data at other monitoring points with the normal relative change value of the same type of monitoring data at other monitoring points.

[0056] Specifically, the absolute value of the difference between the current relative change value of a certain type of monitoring data at each of the other monitoring points and the normal relative change value of the same type of monitoring data at the corresponding other monitoring points is normalized to obtain the current safety risk level of the target monitoring point's monitoring data. Comparing the relative change of normal monitoring data with the difference between the current target monitoring point's monitoring data and the changes of other monitoring points reflects the current safety risk level of the target monitoring point's monitoring data.

[0057] A specific calculation model for the security risk level of monitoring data at the current moment is as follows: , in, This indicates the degree of security risk of the a-th type of monitoring data at the target monitoring point at the current moment; This indicates the total number of monitoring points other than the target monitoring point. This represents the relative change value of the a-th type of monitoring data at the u-th monitoring point among other monitoring points at the current moment; This represents the normal relative change value of the a-th type of monitoring data at the u-th monitoring point; This represents the absolute value of the difference between the relative change value of the target monitoring point relative to the a-th monitoring data point among the u-th monitoring points at the current moment and the normal relative change characteristic value. The larger the absolute value, the higher the degree of safety risk. norm represents the normalization function.

[0058] Similarly, we can obtain the current level of security risk of various monitoring data at the target monitoring point, as well as the level of security risk of various monitoring data at different times in history. The same applies to other monitoring points.

[0059] The target monitoring point is monitored based on the security risk level of various monitoring data at the current moment. Specifically, the security risk level of a certain type of monitoring data at all historical moments of all monitoring points is clustered. With parameter k=3, three clusters are obtained. The average security risk level of each cluster is calculated, and the clusters corresponding to this type of monitoring data are divided into high-risk, low-risk, and risk-free clusters according to the average security risk level from largest to smallest. Similarly, high-risk, low-risk, and risk-free clusters are obtained for each type of monitoring data. If the security risk level of a certain type of monitoring data at the target monitoring point at the current moment belongs to the high-risk or low-risk cluster corresponding to that type of monitoring data, an anomaly warning is issued. If the security risk level of that type of monitoring data at the target monitoring point at the current moment belongs to the risk-free cluster corresponding to that type of monitoring data, no anomaly warning is issued.

[0060] It should be noted that for parameter k=3, this application effectively divides the risk level into three levels, which can be adjusted according to the actual situation. For a monitoring point, if the security risk level of a monitoring data belongs to either a high-risk cluster or a low-risk cluster, an anomaly warning is issued, and the type of monitoring data for which the anomaly warning is issued is reported. Furthermore, when determining the cluster to which a security risk level belongs, the distance between that security risk level and the cluster centers of the three clusters is used; the cluster with the smallest distance is the cluster to which that security risk level belongs.

[0061] In summary, this application analyzes the trend consistency characteristic values ​​of different monitoring points at each historical monitoring time based on the correlation between the changes in monitoring data from multiple monitoring points of a reservoir dam. Then, based on the similarity between each historical time and the current environmental conditions, the influence weight of the corresponding characteristic value at each monitoring time on the analysis results is adjusted to obtain a normal data change pattern model under the current environmental conditions. The safety risk of the reservoir dam is assessed based on the degree of deviation of the current monitoring data from the normal data change pattern model. Compared with traditional methods, this approach can detect early-stage data trend anomalies and reduce the impact of environmental changes on anomaly early warning, effectively improving the timeliness and accuracy of anomaly identification.

[0062] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0063] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0064] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-source data fusion and risk early warning system for reservoir dam safety monitoring, characterized in that, The system includes: The data acquisition module is used to acquire various monitoring data, temperature, and water level at each monitoring point on the dam at each time point within a preset time period before the current time; any monitoring point can be used as the target monitoring point. The relative change analysis module is used to obtain the rate of change of various monitoring data at each monitoring point at each time; to obtain the relative change value of that monitoring data at a monitoring point at a time based on the ratio of the rate of change of a certain monitoring data at a monitoring point to that of a target monitoring point at a time; and to form a sequence of relative change values ​​of that monitoring data at a monitoring point at each time within a preset time period by combining the relative change values ​​of that monitoring data at a monitoring point. The environmental analysis module is used to cluster data based on the temperature and water level at each time point within a preset time period at the target monitoring point. It then clusters data at each time point using the temperature as the clustering condition to obtain the temperature-water level distribution characteristic value and temperature distribution characteristic value for each type of monitoring data at the target monitoring point. Finally, it uses these characteristic values ​​to determine the degree of temperature influence for each type of monitoring data at the target monitoring point. Based on the degree of temperature influence for each type of monitoring data at each monitoring point, it obtains the degree of temperature influence for each type of monitoring data, and similarly, it obtains the degree of water level influence for each type of monitoring data. The monitoring module is used to obtain the environmental similarity of a monitoring data at a given moment based on the differences in temperature and water level between the target monitoring point and other monitoring points at a given moment, as well as the degree of influence of temperature and water level on a particular monitoring data. It also uses the environmental similarity of a monitoring data at different moments and the relative change value sequence of that monitoring data at a given monitoring point to obtain the normal relative change value of that monitoring data at that monitoring point. Furthermore, it uses the relative change value of a monitoring data at a given moment from other monitoring points and the normal relative change value of that monitoring data at other monitoring points to obtain the safety risk level of the target monitoring point at the given moment. Finally, it monitors the target monitoring point based on the safety risk level of various monitoring data at the target monitoring point at the given moment.

2. The multi-source data fusion and risk early warning system for reservoir dam safety monitoring according to claim 1, characterized in that, The step of obtaining the relative change value of the monitoring data at a given time point based on the ratio of the rate of change of a monitoring data point at a given time point to that of a target monitoring point at a given time point includes: The relative change value of a certain type of monitoring data at a given time is obtained by dividing the rate of change of a certain type of monitoring data at a given time point by the sum of the rate of change of the same type of monitoring data at the target monitoring point at that time point and the hyperparameter.

3. The multi-source data fusion and risk early warning system for reservoir dam safety monitoring according to claim 1, characterized in that, The method involves using the temperature and water level at each moment within a preset time period of the target monitoring point as clustering conditions, and using the temperature at each moment as the clustering condition to perform clustering for each moment to obtain the temperature-water level distribution characteristic value and temperature distribution characteristic value of each type of monitoring data at the target monitoring point, including: The temperature and water level at each moment within a preset time period of the target monitoring point are used as clustering conditions to obtain different clusters. The median of a certain type of monitoring data at each moment of the target monitoring point in a cluster is used as the representative value of that type of monitoring data in that cluster. The reliability of the representative value of that type of monitoring data in that cluster is obtained by normalizing the sum of the standard deviation and hyperparameters of the monitoring data at each moment of the target monitoring point in a cluster. The absolute value of the difference between the representative values ​​of a certain type of monitoring data in two clusters is multiplied by the mean of the reliability values ​​of that type of monitoring data in the two clusters to obtain the representative difference between the two clusters. The mean of the representative differences between any two clusters is calculated and normalized to obtain the temperature-water level distribution characteristic value of that type of monitoring data at the target monitoring point. Similarly, the temperature-water level distribution characteristic value of each type of monitoring data at the target monitoring point can be obtained. Similarly, the temperature at each moment is used as a clustering condition to obtain the temperature distribution characteristic value of each type of monitoring data at the target monitoring point.

4. The multi-source data fusion and risk early warning system for reservoir dam safety monitoring according to claim 1, characterized in that, The method of obtaining the degree of temperature influence of each monitoring data point at the target monitoring point by utilizing the temperature-water level distribution characteristic value and temperature distribution characteristic value of each monitoring data point includes: The temperature-water level distribution characteristic value and the temperature distribution characteristic value of a certain type of monitoring data at the target monitoring point are compared with the temperature-water level distribution characteristic value of the same type of monitoring data at the target monitoring point and then normalized to obtain the degree of temperature influence of the same type of monitoring data at the target monitoring point.

5. The multi-source data fusion and risk early warning system for reservoir dam safety monitoring according to claim 1, characterized in that, The process of obtaining the temperature influence degree of each monitoring data point based on the temperature influence degree of each monitoring data point includes: The temperature influence of a monitoring data point is obtained by calculating the mean value of the temperature influence of that monitoring data point.

6. The multi-source data fusion and risk early warning system for reservoir dam safety monitoring according to claim 1, characterized in that, The method of obtaining the environmental similarity of a monitoring data at a given time based on the differences in temperature and water level between the target monitoring point at the current time and other monitoring points at a given time, and the degree of influence of temperature and water level on a given monitoring data, includes: The temperature weight of a monitoring data point is obtained by dividing its temperature influence by the sum of its temperature influence and water level influence. Similarly, the water level weight is obtained. The reciprocal of the temperature difference at that moment is obtained by adding the absolute values ​​of the temperature differences between the target monitoring point at the current moment and those at other monitoring points at a given moment to a hyperparameter. The reciprocal of the water level difference at that moment is obtained by adding the absolute values ​​of the water level differences between the target monitoring point at the current moment and those at other monitoring points at the same moment to a hyperparameter. The environmental similarity of the monitoring data at that moment is obtained by multiplying the normalized value of the reciprocal of the temperature difference at that moment by the temperature weight of the monitoring data point, and the product of the normalized value of the reciprocal of the water level difference at that moment by the water level weight of the monitoring data point.

7. The multi-source data fusion and risk early warning system for reservoir dam safety monitoring according to claim 1, characterized in that, The method of obtaining the normal relative change value of the monitoring data at a monitoring point by utilizing the environmental similarity of monitoring data at various times and the relative change value sequence of the monitoring data at a monitoring point includes: The weight of a monitoring data at a given moment is obtained by comparing the environmental similarity of a monitoring data at a given moment with the sum of the environmental similarity of the monitoring data at all moments. The weight of the monitoring data at each moment is then used to calculate the weighted average of the relative change values ​​of the monitoring data at a given monitoring point at each moment in the relative change value sequence, thus obtaining the normal relative change value of the monitoring data at that monitoring point.

8. The multi-source data fusion and risk early warning system for reservoir dam safety monitoring according to claim 1, characterized in that, The method of obtaining the security risk level of the target monitoring point's monitoring data at the current moment based on the relative change value of a type of monitoring data at other monitoring points and the normal relative change value of the same type of monitoring data at other monitoring points includes: The absolute value of the difference between the relative change value of a certain type of monitoring data at each of the other monitoring points at the current moment and the normal relative change value of the same type of monitoring data at the corresponding other monitoring points is normalized to obtain the safety risk level of the target monitoring point at the current moment for that type of monitoring data.

9. The multi-source data fusion and risk early warning system for reservoir dam safety monitoring according to claim 1, characterized in that, The monitoring of the target monitoring point based on the security risk level of various monitoring data at the current time includes: Clustering is performed on the security risk level of a certain type of monitoring data at all historical moments for all monitoring points. With parameter k=3, three clusters are obtained. The average security risk level of each cluster is calculated. Based on the average security risk level from highest to lowest, the clusters corresponding to this type of monitoring data are categorized into high-risk, low-risk, and risk-free clusters. Similarly, high-risk, low-risk, and risk-free clusters are obtained for each type of monitoring data. If the security risk level of a certain type of monitoring data at a target monitoring point at the current moment belongs to either the high-risk or low-risk cluster, an anomaly warning is issued. If the security risk level of the same type of monitoring data at the target monitoring point at the current moment belongs to the risk-free cluster, no anomaly warning is issued.

Images