Method for monitoring and early warning of hydraulic structure settlement in water conservancy projects

Abstract

The present invention relates to the technical field of building settlement monitoring, and more specifically to a method for monitoring and early warning the settlement of hydraulic structures used in water conservancy projects. The present invention clusters environmental interference data of various dimensions, and based on the clustering results of the environmental interference data of various dimensions, obtains the direct factors affecting the inclination data of each dimension. Furthermore, by analyzing the abnormality of the correlation between any two dimensions of environmental interference data at the corresponding time of the inclination data, combined with the degree of change in the inclination data, the degree of abnormality of the environmental interference data of each dimension at any moment is obtained. Furthermore, the training weights of an SVM model are obtained, and compensation correction is performed on the inclination data to complete settlement prediction. By considering the gain effects between multiple types of environmental interference data, the model can better handle interference in complex environments, thereby improving the robustness of the model in practical applications and the accuracy of settlement monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of building settlement monitoring, and in particular to a method for monitoring and early warning the settlement of hydraulic structures used in water conservancy projects. Background Art

[0002] With the rapid development of urbanization and construction technology, the safety and stability of building structures are receiving increasing attention. Especially in large, high-rise or long-term construction projects, structural settlement has become an important monitoring indicator, directly related to the safety and service life of the building. The existing settlement warning method for hydraulic structures usually involves placing inclinometer sensors at the edges and corners of the building to monitor the tilt angle of the building in real time, and using the combination of BIM technology and the Internet of Things to uniformly monitor the angle changes of several settlement monitoring points. However, due to various environmental and weather influences, it usually leads to affected interference data for hydraulic structures. Therefore, it is usually necessary to use a support vector machine (SVM) model to compensate for the monitoring data of the inclinometer.

[0003] Support vector machine models for compensating inclinometer monitoring data typically use the inclinometer monitoring data and multiple simultaneous environmental disturbance data sets as training data sets, then perform compensation corrections based on the trained model. Conventional training of support vector machines (SVMs) is one-on-one. However, in complex environmental disturbances, multiple disturbance data sets often appear simultaneously, and some disturbances exhibit additive effects on hydraulic structures. This can lead to inaccurate compensation models derived from conventional training methods. Summary of the Invention

[0004] In order to solve the above technical problems, the purpose of the present invention is to provide a method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects.

[0005] According to a first aspect of an embodiment of the present invention, a method for monitoring and early warning the settlement of hydraulic structures in a water conservancy project is provided, wherein the technical solution adopted is as follows:

[0006] Collect inclination data of hydraulic structures and multi-dimensional environmental interference data;

[0007] Clustering the environmental interference data of each dimension to obtain environmental interference data clusters, analyzing the uniformity of the influence of the environmental interference data clusters on the inclination data, and analyzing the distribution of the environmental interference data within the environmental interference data clusters to obtain a direct factor of influence of the environmental interference data of each dimension on the inclination data;

[0008] Analyze the abnormality of the correlation between the environmental interference data of any two dimensions at the time corresponding to the inclination data, and combine the degree of change between the inclination data at any moment and the inclination data at adjacent moments to obtain the degree of environmental interference abnormality of the environmental interference data of each dimension at any moment;

[0009] According to the direct influencing factors and the abnormal degree of the environmental interference, the weight of the environmental interference data of each dimension when participating in the support vector machine (SVM) model training at any time is obtained to complete the compensation correction of the inclination data;

[0010] Based on the compensated inclination data, the settlement amount is predicted and settlement monitoring is completed.

[0011] In some embodiments of the present invention, clustering the environmental interference data of each dimension to obtain environmental interference data clusters may further include:

[0012] Analyzing the continuity of the environmental interference data at corresponding moments in the environmental interference data cluster to obtain a plurality of continuous interference environment data groups;

[0013] The step of analyzing the uniformity of the influence of the environmental interference data cluster on the tilt angle data includes:

[0014] The uniformity of the influence of each continuous interference environment data group in the environmental interference data cluster on the tilt angle data is analyzed.

[0015] In some embodiments of the present invention, analyzing the uniformity of the influence of each continuous interference environment data group in the environmental interference data cluster on the tilt angle data includes:

[0016] Analyze the discrete degree of the inclination data corresponding to each continuous interference environment data group in the environmental interference data cluster, and combine the amplitude of the environmental interference data in other dimensions at the corresponding moment to obtain the uniformity of the influence of each continuous interference environment data group in the environmental interference data cluster on the inclination data.

[0017] In some embodiments of the present invention, analyzing the distribution of the environmental interference data within the environmental interference data cluster includes:

[0018] Calculating the normalized intra-cluster sum of squares of the environmental interference data cluster, and obtaining the data volume of any continuous interference environment data group within the environmental interference data cluster, to obtain the distribution of the environmental interference data within the environmental interference data cluster;

[0019] The analysis of the uniformity of the influence of the environmental interference data clusters on the inclination data and the distribution of the environmental interference data within the environmental interference data clusters to obtain the direct factors of the influence of the environmental interference data of each dimension on the inclination data includes:

[0020] The influence uniformity is weighted by the data amount, and combined with the intra-cluster sum of squares to obtain a direct factor of influence of the environmental interference data of each dimension on the inclination data.

[0021] In some embodiments of the present invention, analyzing the abnormality of the correlation between the environmental interference data in any two dimensions at the time corresponding to the tilt data includes:

[0022] Analyze the correlation factors between environmental interference data in any two dimensions;

[0023] Define the time window corresponding to any moment;

[0024] The Pearson correlation coefficient between the environmental interference data of any two dimensions in the time window at any moment is obtained, and combined with the correlation factor, the abnormality of the correlation between the environmental interference data of any two dimensions in the time window at any moment is obtained.

[0025] In some embodiments of the present invention, analyzing the correlation factor between any two dimensions of environmental interference data includes:

[0026] Segmenting the inclination data to obtain inclination change segments;

[0027] The Pearson correlation coefficient between any two dimensions of environmental interference data within the corresponding time of the tilt angle change segment is calculated, and the correlation factor between any two dimensions of environmental interference data is obtained by combining the amplitude of the tilt angle data within the tilt angle change segment.

[0028] In some embodiments of the present invention, segmenting the tilt angle data to obtain tilt angle change segments includes:

[0029] Determining the time series consisting of the tilt angle data, where the determination criterion is whether the tilt angle data is 0;

[0030] If it is 0, it stops, forming an inclination change section;

[0031] If it is not 0, continue to judge the next inclination data;

[0032] Repeat this operation to obtain multiple consecutive tilt angle change segments.

[0033] In some embodiments of the present invention, the degree of change between the tilt angle data at any moment and the tilt angle data at adjacent moments includes:

[0034] Calculate the mean of the inclination data within the time window corresponding to any moment;

[0035] The mean value and the inclination data in the inclination variation segment are linearly normalized to obtain normalized inclination data, which represents the degree of variation between the inclination data at any moment and the inclination data at adjacent moments.

[0036] In some embodiments of the present invention, collecting the inclination data of hydraulic structures and multi-dimensional environmental interference data includes:

[0037] Collect the inclination data and multi-dimensional environmental interference data of hydraulic structures at the current moment, as well as the inclination data and multi-dimensional environmental interference data within one month before the current moment;

[0038] The tilt angle data and the multi-dimensional environmental interference data are subjected to standardized preprocessing to obtain standardized tilt angle data and standardized multi-dimensional environmental interference data.

[0039] In some embodiments of the present invention, clustering the environmental interference data of each dimension to obtain environmental interference data clusters includes:

[0040] K-means clustering is performed on the environmental interference data of each dimension, and the optimal number of clusters is determined by the elbow method to obtain the environmental interference data clusters.

[0041] Compared with the prior art, the hydraulic structure settlement monitoring and early warning method for water conservancy projects provided by the present invention has the following beneficial effects:

[0042] The present invention performs cluster analysis on environmental interference data of various dimensions, and obtains the direct factors of influence of environmental interference data of various dimensions on the current hydraulic structure based on the clustering results; then obtains the change segment based on the inclination data, and obtains the correlation factors between the environmental interference data in combination with the actual shaking of the hydraulic structure, and further obtains the authenticity and abnormality evaluation of the environmental interference data of each dimension based on the performance of the correlation coefficient between each environmental interference data; finally, it is placed in the SVM model to optimize the compensation model and improve the accuracy of compensation. By considering the gain effect between multiple environmental interference data, the model can better handle interference in complex environments, and improve the robustness of the model in practical applications; by analyzing the correlation between environmental interference data and inclination data, and considering the authenticity and abnormality of environmental interference data, the SVM model is optimized and the accuracy of settlement monitoring is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0044] Figure 1 A schematic diagram of the basic flow of a method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects provided by one embodiment of the present invention. DETAILED DESCRIPTION

[0045] To further illustrate the technical means and effectiveness of the present invention in achieving its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, describes in detail the specific implementation, structure, features, and effectiveness of the method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects proposed by the present invention. In the following description, different references to "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.

[0046] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this invention pertains. Terms such as "comprises," "comprising," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a circuit structure, article, or device comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such article or device. In the absence of further limitations, the phrase "comprising a ..." to define an element does not preclude the presence of other identical elements in the article or device comprising the element.

[0047] The following describes in detail a method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects provided by the present invention in conjunction with the accompanying drawings.

[0048] See also Figure 1 , which shows the basic process of a method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects provided by an embodiment of the present invention.

[0049] like Figure 1 As shown, an embodiment of the present invention provides a method for monitoring and early warning the settlement of hydraulic structures in a water conservancy project, which specifically includes:

[0050] S100: Collect inclination data of hydraulic structures and multi-dimensional environmental interference data.

[0051] Based on the structure and other information of the hydraulic structure to be monitored, monitoring points are selected and an inclinometer is installed at each monitoring point, with a preset zero value. (This implementation is prior art and will not be described in detail here.) Continuous, real-time monitoring results from the inclinometers at multiple monitoring points are obtained to obtain the inclination data of the hydraulic structure. At the same time, multi-dimensional environmental interference data, such as wind speed and local vibration, is obtained for each monitoring point within its local area. Ultimately, each inclination data point corresponds to a set of environmental interference data.

[0052] And obtain the inclination data and multi-dimensional environmental interference data within one month before the current moment as a reference data set.

[0053] Furthermore, the tilt angle data and the multi-dimensional environmental interference data are pre-processed for standardization to obtain standardized tilt angle data and standardized multi-dimensional environmental interference data. It should be noted that, unless otherwise specified, the tilt angle data and the multi-dimensional environmental interference data refer to the standardized tilt angle data and the standardized multi-dimensional environmental interference data.

[0054] S200: Clustering the environmental interference data of each dimension to obtain environmental interference data clusters, analyzing the uniformity of the impact of the environmental interference data clusters on the inclination data, and analyzing the distribution of the environmental interference data within the environmental interference data clusters to obtain the direct factors affecting the inclination data of each dimension.

[0055] When monitoring the settlement of hydraulic structures, environmental factors, such as water impact and excessive wind speed, often significantly interfere with the results. These interferences typically cause immediate changes, rather than long-term changes such as soil deposition. Furthermore, these interferences are not noise, but rather have a real impact on the inclinometer's monitoring results. Environmental interference can significantly affect the effectiveness of support vector machines (SVMs) in compensating for inclination data.

[0056] The occurrence of environmental interference is seasonal and regular. Therefore, for different environmental interference data, it is necessary to first analyze the occurrence pattern of a certain environmental interference data in the current area, and then pre-set the sensitivity of each environmental interference data to the monitoring data of the inclinometer.

[0057] Based on the above analysis, in an embodiment of the present invention, by clustering the environmental interference data of each dimension to obtain environmental interference data clusters, analyzing the uniformity of the impact of the environmental interference data clusters on the inclination data, and analyzing the distribution of the environmental interference data within the environmental interference data clusters, the direct factors affecting the inclination data of each dimension are obtained. Further, the following methods are included:

[0058] First, cluster the environmental interference data of each dimension to obtain the environmental interference data clusters. The specific implementation method is: perform K-means clustering on the environmental interference data of each dimension, determine the optimal number of clusters by the elbow method, and obtain the environmental interference data of each dimension. Environmental interference data clusters. These clusters represent the frequently occurring amplitude ranges of environmental interference data for that dimension within a month. The inclination data at multiple moments corresponding to the environmental interference data within a cluster reflects the extent to which the inclination data is affected when that environmental interference occurs.

[0059] Then, the continuity of the environmental interference data corresponding to the time in the environmental interference data cluster is analyzed to obtain multiple continuous interference environment data groups. Each environmental interference data in the environmental interference data cluster is judged at adjacent moments. If the moments corresponding to the environmental interference data adjacent to the environmental interference data are adjacent, it is determined to be a group of continuous interference environmental data. After traversing all environmental interference data in the environmental interference data cluster, the first multiple groups of continuous interference environment data in an environmental interference data cluster; at the same time, for environmental interference data whose adjacent environmental interference data correspond to moments that do not exist adjacent moments, the environmental interference data at the left and right adjacent moments corresponding to the interference environment data (that is, environmental interference data that are not in the same environmental interference data cluster) are combined with the interference environment data to form a continuous interference environment data group; through the above two methods, each environmental interference data in the environmental interference data cluster is divided into multiple continuous interference environment data groups, and finally each continuous interference environment data group has at least two environmental interference data.

[0060] Then, analyze the uniformity of the influence of the environmental interference data cluster on the inclination data. Further, analyze the uniformity of the influence of each continuous interference environmental data group on the inclination data in the environmental interference data cluster. The specific implementation method is: analyze the discrete degree of the inclination data corresponding to each continuous interference environmental data group in the environmental interference data cluster, and combine the amplitude of the environmental interference data of other dimensions at the corresponding moment to obtain the uniformity of the influence of each continuous interference environmental data group on the inclination data in the environmental interference data cluster. Construct the first The first dimension of environmental interference data The first The calculation formula for the uniformity of the impact of a continuous interference environment data group on the tilt data is:

[0061]

[0062] Where, Indicates the The first dimension of environmental interference data The first The uniformity of the impact of a continuous interference environment data set on the inclination data; Indicates the The first dimension of environmental interference data The first The first of the continuous interference environment data sets The amplitude of the inclination data corresponding to the interference environment data; Indicates the The first dimension of environmental interference data The first The mean amplitude of the inclination data corresponding to all interference environment data in a continuous interference environment data group; Indicates the The first dimension of environmental interference data The first The number of interference environment data in a continuous interference environment data group; Indicates that except the current The remaining dimensional environmental interference data except the dimensional environmental interference data; Indicates the The first dimension of environmental interference data The first The first of the continuous interference environment data sets The interference environment data corresponds to the The amplitude of the environmental interference data in each dimension; Indicates preset hyperparameters, set here , can be adjusted by yourself; Expressed as a natural constant An exponential function with base .

[0063] Indicates the first The amplitude of the inclination data corresponding to the continuous interference environment data group is the same as the amplitude of the inclination data in the environmental interference data cluster. The difference between the amplitude mean values of the inclination data corresponding to the continuous interference environment data groups represents the first The smaller the discrete degree of the inclination data corresponding to the continuous interference environment data set, the closer the actual changes of the inclinometers in the continuous interference environment data set are. Is the change in the inclinometer monitoring data corresponding to the continuous interference environment data group mainly caused by the The dimensional environment interferes with the data, so here we need to adjust the weight based on the amplitude of the interference data environment of the other dimensional environments at the same time, that is, through right Weighted and normalized The closer the amplitude of the environmental interference data in each dimension is to 1, the smaller the weight assigned. The normalized index changes faster and can better widen the gap between larger and smaller values. The gap can be further widened by presetting hyperparameters.

[0064] Then, the distribution of environmental interference data within the environmental interference data cluster is analyzed. For each environmental interference data cluster, the larger the intra-cluster sum of squares is, and the more data there is in the environmental interference data cluster, the more frequently and concentratedly the environmental interference data appears within the amplitude range corresponding to the environmental interference data cluster in the current water conservancy project environment. Therefore, the specific implementation method is to calculate the intra-cluster sum of squares of the environmental interference data cluster after normalization, that is, for the first The first dimension of environmental interference data The environmental interference data clusters are linearly normalized, and after obtaining the centroid position of the environmental interference data clusters, the normalized intra-cluster sum of squares SSE (SSE is an indicator to measure the difference between the sample points in the cluster and their centroid. The smaller the SSE value, the denser the sample points in the cluster and the better the aggregation effect) of the environmental interference data clusters are calculated. This feature is a well-known technology. In addition, the data volume of any continuous interference environment data group in the environmental interference data clusters is obtained to obtain the distribution of environmental interference data in the environmental interference data clusters.

[0065] Finally, based on the uniformity of the impact of the environmental interference data clusters on the inclination data and the distribution of the environmental interference data within the environmental interference data clusters, the direct factor of the impact of the environmental interference data of each dimension on the inclination data is obtained. Furthermore, the uniformity of the impact is weighted by the amount of data, and combined with the sum of squares within the cluster, the direct factor of the impact of the environmental interference data of each dimension on the inclination data is obtained. The calculation formula for the direct factor of the impact of environmental interference data in each dimension on the inclination data is:

[0066]

[0067] Where, Indicates the Direct factors affecting the inclination data from environmental interference data in each dimension; Indicates the The first dimension of environmental interference data The first The uniformity of the impact of a continuous interference environment data set on the inclination data; Indicates the The first dimension of environmental interference data The first The amount of data in a continuous interference environment data group; Indicates the The first dimension of environmental interference data The normalized intra-cluster sum of squares of each environmental interference data cluster; Indicates the The first dimension of environmental interference data The number of continuous interference environment data groups in each environmental interference data cluster; Indicates the The number of environmental interference data clusters for each dimension of environmental interference data; Represents the activation function.

[0068] Since the structure of water conservancy projects usually does not shake easily, short-term interference should not show obvious interference. Moreover, for sensors used to detect environmental parameters, the environment is relatively harsh and easily interfered with, which leads to noise interference in the data. Therefore, according to the first The amount of data in a continuous interference environment data group The larger the value, the longer the continuous interference duration. right Weighted, The larger the value, the higher the corresponding weight; The larger the value is, the higher the current environmental interference data cluster is in terms of quantity and aggregation effect than the other environmental interference data clusters. It further indicates that the current environmental interference data appears relatively frequently within the amplitude range corresponding to the environmental interference data cluster and has a more direct impact on the monitored building. Therefore, the environmental interference data of this dimension has a greater direct factor on the inclination data.

[0069] S300: Analyze the abnormality of the correlation between the environmental interference data of any two dimensions at the time corresponding to the inclination data, and combine the degree of change between the inclination data at any moment and the inclination data at its adjacent moments to obtain the abnormality of the environmental interference data of each dimension at any moment.

[0070] In more complex environmental disturbances, multiple dimensions of environmental disturbance data often appear simultaneously, and some of the environmental disturbances have a positive effect on hydraulic structures, meaning that the sum of the two is greater than the sum of the two. Therefore, by analyzing the abnormality of the correlation between any two dimensions of environmental disturbance data at the corresponding time of the inclination data, and combining the degree of change between the inclination data at any moment and the inclination data at adjacent moments, the degree of abnormality of the environmental disturbance data in each dimension at any moment can be obtained. Further analysis includes:

[0071] First, analyze the correlation factor between any two dimensions of environmental interference data. The specific implementation method is: segment the inclination data to obtain inclination change segments; the segmentation method can be to judge the time series composed of the inclination data, and the judgment standard is whether the inclination data is 0; if it is 0, stop to form an inclination change segment; if it is not 0, continue to judge the next inclination data; repeat this operation to obtain multiple continuous inclination change segments, each inclination change segment corresponds to the actual shaking of its hydraulic structure. Calculate the Pearson correlation coefficient between any two dimensions of environmental interference data within the time corresponding to the inclination change segment, and combine the amplitude of the inclination data in the inclination change segment to obtain the correlation factor between any two dimensions of environmental interference data. Construct the first The calculation formula for the correlation factor between the environmental interference data of each dimension is:

[0072]

[0073] Where, Indicates the The correlation factors between environmental interference data of different dimensions; Indicates the In the inclination change section Pearson coefficient between environmental interference data of different dimensions; Indicates the The mean amplitude of all inclination data in an inclination change segment; Indicates the total number of inclination change segments; represents the linear normalization function.

[0074] When calculating the correlation factor between environmental interference data, the larger the inclination data corresponding to the current inclination change segment is, the more obvious the corresponding interference is, and the Pearson coefficient obtained at this time will be more realistic. Therefore, by right Weighted, and The larger it is, the greater the weight.

[0075] The logic of correlation factor analysis is: since the current environmental interference is to obtain whether different parameters have a gain-in-force effect on hydraulic structures, the correlation analysis cannot be based entirely on the environmental interference data in the entire historical data. Instead, the corresponding environmental interference data under the change of inclination data should be analyzed.

[0076] Then, define the time window corresponding to any moment. Specifically, for the first At a certain moment, obtain environmental interference data for five minutes before and after each adjacent moment (because the sampling frequency of environmental interference data is usually not high, and this type of interference usually lasts for a long time, it is necessary to select in minutes).

[0077] Then, the Pearson correlation coefficient between the environmental interference data of any two dimensions in the time window at any moment is obtained, and combined with the correlation factor, the abnormality of the correlation between the environmental interference data of any two dimensions in the time window at any moment is obtained.

[0078] Finally, the abnormality of the correlation between any two dimensions of environmental interference data at the time corresponding to the inclination data is combined with the degree of change between the inclination data at any moment and the inclination data at adjacent moments to obtain the degree of environmental interference abnormality of each dimension of environmental interference data at any moment. The degree of change between the inclination data at any moment and the inclination data at adjacent moments includes: calculating the mean of the inclination data within the time window corresponding to any moment; and linearly normalizing the mean with the inclination data within the inclination change segment to obtain normalized inclination data, which represents the degree of change between the inclination data at any moment and the inclination data at adjacent moments.

[0079] Build the The environmental interference data of the dimensions is The time window of the moment The abnormality calculation formula of the correlation between the environmental interference data of the dimensions is:

[0080]

[0081] Where, Indicates the The environmental interference data of the dimensions is The time window of the moment The abnormality of the correlation between the environmental interference data of the dimensions; Indicates the The environmental interference data of the dimensions is The time window of the moment Pearson coefficient between environmental interference data of different dimensions; Indicates the The correlation factors between environmental interference data of different dimensions; Indicates the The environmental interference data of the dimensions is The average value of the inclination angle data in the time window corresponding to the moment is linearly normalized with the inclination angle data in the aforementioned multiple (more than 3) inclination angle change segments. The obtained normalized inclination angle data represents the first The degree of change between the inclination data at a certain moment and the inclination data at adjacent moments; The number of dimensions representing the environmental interference data, Indicates taking the absolute value.

[0082] Indicates the The time window corresponding to the moment The difference between the Pearson coefficient between the environmental interference data of the two dimensions and the correlation factor between the environmental interference data of the two dimensions, the greater the difference, the greater the deviation between the two environmental interference data in the current time window and the performance represented by the long time series, and there may be unreliable components between the corresponding data; the more obvious the shaking of the hydraulic structure, the more appropriate the gain combination performance should be between the corresponding environmental interference in the current time window. Therefore, based on this effect, the degree of authenticity obtained at this time is expanded, that is, The larger the value, the more abnormal performance there is, and the weight of its subsequent participation in filtering is relatively low. However, if the instability of the hydraulic structure is weaker at this time, then regardless of whether the correlation is true or not, due to the effect of this weight, the abnormality of its actual performance will be weaker, avoiding possible interference caused by the randomness of the environment.

[0083] S400: According to the direct influencing factors and the degree of abnormal environmental interference, the weight of the environmental interference data of each dimension participating in the support vector machine (SVM) model training at any time is obtained to complete the compensation correction of the inclination data.

[0084] According to the direct influencing factors and the abnormal degree of environmental interference, the weight of each dimension of environmental interference data participating in the support vector machine SVM model training at any time is obtained. The environmental interference data of the dimensions is The weight calculation formula for participating in the support vector machine SVM model training at time is:

[0085]

[0086] Where, Indicates the The environmental interference data of the dimensions is The weight involved in the support vector machine SVM model training at time; Indicates the The environmental interference data of the dimensions is The time window of the moment The abnormality of the correlation between the environmental interference data of the dimensions; Indicates the Direct factors affecting the inclination data from environmental interference data in each dimension; represents the linear normalization function.

[0087] The unusual nature of the correlations between environmental disturbances The larger the value, the smaller the weight, and the more it affects the direct factor The larger the value, the greater its weight. The two correct each other's constraints, and after normalization, the interference terms between the environmental interference data of each dimension are obtained. Therefore, it is worth noting that the linear normalization range here is the weight range obtained at each moment for all environmental interference data.

[0088] At this point, the weight of each environmental interference data is directly added as a label to the loss model of the support vector machine (SVM), thereby completing the optimization of the compensation model and compensating and correcting the inclination data.

[0089] S500: Based on the compensated and corrected inclination data, the settlement amount is predicted and settlement monitoring is completed.

[0090] Based on compensated inclination data, data is collected regularly and analyzed to determine the trend and degree of building tilt. Based on inclination changes and known geometric relationships, a mathematical model is developed to predict settlement and monitor settlement. For example, if an inclinometer is installed on a dam, the horizontal displacement and, therefore, settlement can be estimated based on the inclination changes and the dam's height.

[0091] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0092] 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 the differences from other embodiments.

Claims

1. A method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects, characterized in that: The method comprises: Collect inclination data of hydraulic structures and multi-dimensional environmental interference data; Clustering the environmental interference data of each dimension to obtain environmental interference data clusters, analyzing the uniformity of the influence of the environmental interference data clusters on the inclination data, and analyzing the distribution of the environmental interference data within the environmental interference data clusters to obtain a direct factor of influence of the environmental interference data of each dimension on the inclination data; Analyze the abnormality of the correlation between the environmental interference data of any two dimensions at the time corresponding to the inclination data, and combine the degree of change between the inclination data at any moment and the inclination data at adjacent moments to obtain the degree of environmental interference abnormality of the environmental interference data of each dimension at any moment; According to the direct influencing factors and the abnormal degree of the environmental interference, the weight of the environmental interference data of each dimension when participating in the support vector machine (SVM) model training at any time is obtained to complete the compensation correction of the inclination data; According to the compensated and corrected inclination data, the settlement amount is predicted and settlement monitoring is completed; Clustering the environmental interference data of each dimension to obtain environmental interference data clusters, followed by: Analyzing the continuity of the environmental interference data at corresponding moments in the environmental interference data cluster to obtain a plurality of continuous interference environment data groups; The step of analyzing the uniformity of the influence of the environmental interference data cluster on the tilt angle data includes: Analyzing the uniformity of the influence of each continuous interference environment data group in the environmental interference data cluster on the tilt angle data; Analyzing the uniformity of the influence of each continuous interference environment data group in the environmental interference data cluster on the tilt angle data includes: Analyze the discrete degree of the inclination data corresponding to each continuous interference environment data group in the environmental interference data cluster, and combine the amplitude of the environmental interference data in other dimensions at the corresponding moment to obtain the uniformity of the influence of each continuous interference environment data group in the environmental interference data cluster on the inclination data.

2. The method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects according to claim 1, characterized in that: Analyzing the distribution of the environmental interference data within the environmental interference data cluster includes: Calculating the normalized intra-cluster sum of squares of the environmental interference data cluster, and obtaining the data volume of any continuous interference environment data group within the environmental interference data cluster, to obtain the distribution of the environmental interference data within the environmental interference data cluster; The analysis of the uniformity of the influence of the environmental interference data clusters on the inclination data and the distribution of the environmental interference data within the environmental interference data clusters to obtain the direct factors of the influence of the environmental interference data of each dimension on the inclination data includes: The influence uniformity is weighted by the data amount, and combined with the intra-cluster sum of squares to obtain a direct factor of influence of the environmental interference data of each dimension on the inclination data.

3. The method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects according to claim 1, characterized in that: Analyzing the abnormality of the correlation between the environmental interference data in any two dimensions at the time corresponding to the tilt data includes: Analyze the correlation factors between environmental interference data in any two dimensions; Define the time window corresponding to any moment; The Pearson correlation coefficient between the environmental interference data of any two dimensions in the time window at any moment is obtained, and combined with the correlation factor, the abnormality of the correlation between the environmental interference data of any two dimensions in the time window at any moment is obtained.

4. The method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects according to claim 3, characterized in that: Analyze the correlation factors between any two dimensions of environmental interference data, including: Segmenting the inclination data to obtain inclination change segments; The Pearson correlation coefficient between any two dimensions of environmental interference data within the corresponding time of the tilt angle change segment is calculated, and the correlation factor between any two dimensions of environmental interference data is obtained by combining the amplitude of the tilt angle data within the tilt angle change segment.

5. The method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects according to claim 4, characterized in that: Segmenting the inclination data to obtain inclination change segments includes: Determining the time series consisting of the tilt angle data, where the determination criterion is whether the tilt angle data is 0; If it is 0, it stops, forming an inclination change section; If it is not 0, continue to judge the next inclination data; Repeat this operation to obtain multiple consecutive tilt angle change segments.

6. The method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects according to claim 5, characterized in that: The degree of change between the inclination data at any moment and the inclination data at adjacent moments includes: Calculate the mean of the inclination data within the time window corresponding to any moment; The mean value and the inclination data in the inclination variation segment are linearly normalized to obtain normalized inclination data, which represents the degree of variation between the inclination data at any moment and the inclination data at adjacent moments.

7. The method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects according to claim 1, characterized in that: Collect inclination data of hydraulic structures and multi-dimensional environmental interference data, including: Collect the inclination data and multi-dimensional environmental interference data of hydraulic structures at the current moment, as well as the inclination data and multi-dimensional environmental interference data within one month before the current moment; The tilt angle data and the multi-dimensional environmental interference data are subjected to standardized preprocessing to obtain standardized tilt angle data and standardized multi-dimensional environmental interference data.

8. The method for monitoring and early warning the settlement of hydraulic structures in water conservancy projects according to claim 1, characterized in that: Clustering the environmental interference data of each dimension to obtain environmental interference data clusters, including: K-means clustering is performed on the environmental interference data of each dimension, and the optimal number of clusters is determined by the elbow method to obtain the environmental interference data clusters.

Images