Search

CN-122020118-A - Deep foundation pit multi-parameter intelligent monitoring method, system, medium and product based on Internet of things

CN122020118ACN 122020118 ACN122020118 ACN 122020118ACN-122020118-A

Abstract

The application provides a multi-parameter intelligent monitoring method, a system, a medium and a product of a deep foundation pit based on the Internet of things, and relates to the technical field of deep foundation pit monitoring, wherein the method comprises the steps of obtaining multi-type monitoring data of the deep foundation pit, determining a cross-physical quantity conduction relation among monitoring parameters in the data, and constructing a heterogeneous association diagram structure according to the relation; the method comprises the steps of carrying out feature extraction on data on a graph structure to obtain time sequence change features and conduction influence features, carrying out dynamic weight disturbance operation on associated edges according to the time sequence change features and the conduction influence features to obtain weight stability indexes, marking the associated edges with the indexes larger than a preset stability threshold in the graph structure as effective conducting edges, determining a stable conduction path set according to connectivity of the effective conducting edges in the graph structure, and when a target conduction path meeting collaborative risk triggering conditions exists in the set, determining that monitoring nodes in the path are in a multi-parameter collaborative risk state, and carrying out risk response intervention operation on the path according to the risk state.

Inventors

  • WANG WEI

Assignees

  • 烟台金达测绘有限公司

Dates

Publication Date
20260512
Application Date
20260122

Claims (10)

  1. 1. A deep foundation pit multi-parameter intelligent monitoring method based on the Internet of things is characterized by comprising the following steps: under the condition that multi-type monitoring data of a deep foundation pit are obtained, determining a cross-physical quantity conduction relation among monitoring parameters in the multi-type monitoring data, and constructing a heterogeneous association graph structure according to the cross-physical quantity conduction relation, wherein the heterogeneous association graph structure comprises monitoring nodes and influence factor nodes, the monitoring nodes are in one-to-one correspondence with the monitoring parameters, the influence factor nodes are in one-to-one correspondence with external influence events of the deep foundation pit, and association edges included in the heterogeneous association graph structure represent physical conduction paths among the monitoring nodes and between the monitoring nodes and the influence factor nodes; Carrying out space-time conduction characteristic extraction on the multi-type monitoring data on the heterogeneous association graph structure to obtain time sequence change characteristics of the monitoring nodes and conduction influence characteristics of the associated edges, and carrying out dynamic weight disturbance operation on the associated edges according to the time sequence change characteristics and the conduction influence characteristics to obtain weight stability indexes of the associated edges; marking the association edges with the weight stability index larger than a preset stability threshold value in the heterogeneous association graph structure as effective conduction edges, and determining a stable conduction path set according to the connectivity of the effective conduction edges in the heterogeneous association graph structure; When a target conduction path meeting a collaborative risk triggering condition exists in the stable conduction path set, determining that monitoring nodes in the target conduction path are in a multi-parameter collaborative risk state, and executing risk response intervention operation on the target conduction path according to the multi-parameter collaborative risk state, wherein the collaborative risk triggering condition is that the total number of the monitoring nodes in the target conduction path is larger than a preset node number threshold value, and the target conduction path takes an influence factor node as a path starting node.
  2. 2. The method according to claim 1, wherein in the case of acquiring multi-type monitoring data of a deep foundation pit, determining a cross-physical-quantity conduction relationship between monitoring parameters in the multi-type monitoring data, and constructing a heterogeneous association graph structure according to the cross-physical-quantity conduction relationship, specifically comprising: acquiring the multi-type monitoring data, wherein the multi-type monitoring data comprises the displacement monitoring data, the stress monitoring data, the water pressure monitoring data and the inclination monitoring data; acquiring the distribution of the supporting layers of the deep foundation pit, the connection relation of the enclosure walls, the influence range of the dewatering well and the monitoring position information of each monitoring parameter; Performing physical quantity type matching analysis on the multi-type monitoring data to determine a cross-physical quantity conduction relationship between different types of monitoring parameters; determining a structural force transmission path among all the monitoring parameters according to the distribution of the supporting layers, the connection relation of the enclosure walls and the influence range of the precipitation well, and determining the monitoring space distance among all the monitoring parameters according to the monitoring position information; Performing mapping operation on the monitored space distance and a preset distance influence function to obtain a distance weight coefficient representing the space proximity degree; Generating a physical conduction constraint matrix according to the cross-physical quantity conduction relation, the structural force transmission path and the distance weight coefficient, wherein element values in the physical conduction constraint matrix represent conduction intensity coefficients among different types of monitoring parameters; And constructing the heterogeneous association graph structure according to the physical conduction constraint matrix, the monitoring node and the influence factor node.
  3. 3. The method according to claim 2, wherein said constructing the heterogeneous association graph structure from the physical conduction constraint matrix, the monitoring node and the influence factor node comprises: Acquiring construction influence events and environment influence events of all the monitoring parameters; creating the monitoring nodes for the monitoring parameters, creating construction factor nodes for the construction influence events, and creating environment factor nodes for the environment influence events; Performing effectiveness comparison analysis on the conduction intensity coefficient and a preset conduction threshold value to determine two monitoring parameters with the conduction intensity coefficient larger than the preset conduction threshold value as a monitoring parameter pair with an effective conduction relation; Determining a construction influence space range according to the event type and the construction position of the construction influence event, and determining an environment influence space range according to the event type and the action area of the environment influence event; performing first spatial position matching on the construction influence spatial range and the monitoring position information to obtain a first matching result, and determining monitoring nodes with the monitoring position information in the construction influence spatial range as construction influenced nodes according to the first matching result; Determining the construction conduction influence degree according to the monitoring position information of the construction affected node and the monitoring position information of the corresponding construction factor node; performing second spatial position matching on the environmental influence spatial range and the monitoring position information to obtain a second matching result, and determining a monitoring node of which the monitoring position information is positioned in the environmental influence spatial range as an environmental influenced node according to the second matching result; determining the influence degree of environmental conduction according to the monitoring position information of the affected node and the monitoring position information of the corresponding environmental factor node; And constructing the heterogeneous association graph structure according to the monitoring parameter pairs, the construction affected nodes, the construction conduction influence degree, the environment affected nodes and the environment affected nodes.
  4. 4. The method according to claim 1, wherein the performing space-time conduction feature extraction on the heterogeneous association graph structure on the multi-type monitoring data to obtain a time sequence variation feature of the monitoring node and a conduction influence feature of the association edge specifically includes: Acquiring a plurality of time sequence data fragments of the monitoring node in a preset historical time window; inputting the time sequence data fragments into a time sequence feature coding network to acquire time sequence change features of the monitoring nodes; Determining a first change characteristic of a source node connected with a target associated edge and a second change characteristic of a target node connected with the target associated edge from the time sequence change characteristics, and determining a cross-correlation function of the source node and the target node according to the first change characteristic and the second change characteristic; Determining a time sequence correlation coefficient between the source node and the target node according to the cross correlation function, and determining a conduction delay parameter according to the peak time position of the cross correlation function and the time sequence correlation coefficient; determining a conduction activity index between the source node and the target node according to the time sequence correlation coefficient and the conduction delay parameter, and performing coupling analysis on the conduction activity index and a target initial edge weight of the target associated edge to obtain a conduction response intensity value of the associated edge; Acquiring a first physical quantity type of the monitoring parameter corresponding to the source node and a second physical quantity type of the monitoring parameter corresponding to the target node, and inquiring a cross-physical quantity conduction attenuation factor combined with the physical quantity type corresponding to the first physical quantity type and the second physical quantity type from a preset physical quantity type mapping table; And carrying out modulation operation on the conduction response intensity value and the cross-physical quantity conduction attenuation factor to obtain the conduction influence characteristic.
  5. 5. The method according to claim 4, wherein the performing a dynamic weight perturbation operation on the associated edge according to the time sequence variation characteristic and the conduction influence characteristic to obtain the weight stability index of the associated edge specifically includes: Determining a first liveness index of the source node according to the time sequence change characteristics, and determining a second liveness index of the target node according to the time sequence change characteristics; determining an edge activation factor of the target associated edge according to the first liveness index and the second liveness index, and generating a disturbance amplitude coefficient of the target associated edge according to the edge activation factor and the conduction influence characteristic; generating a plurality of groups of random disturbance vectors according to a preset disturbance sampling strategy, and performing scaling operation on the plurality of groups of random disturbance vectors and the disturbance amplitude coefficient to obtain a plurality of groups of target disturbance amounts; Superposing the multiple groups of target disturbance quantities with the target initial edge weights respectively to obtain multiple groups of disturbance edge weights, and updating the multiple groups of disturbance edge weights into the heterogeneous associated graph structure to determine graph topology connectivity variation when edge weight disturbance occurs in the heterogeneous associated graph structure; Performing sensitivity comparison analysis on the topological connectivity variation of the graph and a preset connectivity sensitivity threshold to obtain a sensitivity comparison result; When the topological connectivity variation of the graph is determined to be larger than the connectivity sensitivity threshold according to the sensitivity comparison result, a sensitive connectivity marker is distributed to the corresponding disturbance edge weight, and when the topological connectivity variation of the graph is determined to be smaller than or equal to the connectivity sensitivity threshold according to the sensitivity comparison result, a stable connectivity marker is distributed to the corresponding disturbance edge weight; And determining the weight stability index of the target associated edge according to the number of the disturbance edge weights with the stable connectivity marks in the multiple groups of disturbance edge weights.
  6. 6. The method of claim 1, wherein the marking the association edge in the heterogeneous association graph structure with the weight stability index greater than a preset stability threshold as an effective conduction edge, and determining the stable conduction path set according to connectivity of the effective conduction edge in the heterogeneous association graph structure, specifically includes: performing stability comparison analysis on the weight stability index of the associated edge and the preset stability threshold to mark the associated edge with the weight stability index larger than the preset stability threshold as an effective conducting edge, and marking the associated edge with the weight stability index smaller than or equal to the preset stability threshold as an unstable edge; Removing all the unstable edges from the heterogeneous association graph structure to obtain an effective conduction sub-graph; performing a connected component detection operation in the effective conductive sub-graph to determine a plurality of connected sub-regions that are independent of each other; taking each influence factor node in the node set of each connected sub-region as the path starting node, and taking each monitoring node in the node set of each connected sub-region as the path ending node, and executing path searching operation to obtain candidate conduction paths; And determining a stable conduction path from the candidate conduction paths according to the weight stability index of the effective conduction edge in the candidate conduction paths, and generating the stable conduction path set.
  7. 7. The method according to claim 6, wherein when there is a target conductive path in the stable conductive path set that satisfies a collaborative risk triggering condition, determining that a monitoring node in the target conductive path is in a multi-parameter collaborative risk state, and performing a risk response intervention operation on the target conductive path according to the multi-parameter collaborative risk state, specifically comprising: acquiring the total number of monitoring nodes and the total number of influence factor nodes included in each stable conduction path in the stable conduction path set; Determining stable conduction paths of which the total number of the monitoring nodes is larger than the preset node number threshold value and the total number of the influence factor nodes is larger than or equal to the preset influence factor number threshold value as target conduction paths meeting the collaborative risk triggering condition; acquiring real-time monitoring data of each monitoring node in the target conduction path, and determining the data deviation degree of each monitoring node in the target conduction path according to the real-time monitoring data; Determining whether each monitoring node in the target conduction path is in the multi-parameter collaborative risk state according to the data deviation degree; under the condition that each monitoring node in the target conduction path is in the multi-parameter cooperative risk state, determining a path cooperative risk index according to the data deviation degree and the position information of each monitoring node in the target conduction path; under the condition that the target conduction path is determined to be in an early warning observation risk level according to the path collaborative risk index, an early warning observation task is executed on the target conduction path; executing an encryption collaborative monitoring task on the target conduction path under the condition that the target conduction path is determined to be in an encryption monitoring risk level according to the path collaborative risk index; and executing a risk blocking task on the target conduction path under the condition that the target conduction path is determined to be at an emergency intervention risk level according to the path cooperative risk index.
  8. 8. A deep foundation pit multiparameter intelligent monitoring system comprising one or more processors and a memory coupled to the one or more processors, the memory for storing computer program code comprising computer instructions that the one or more processors invoke to cause the deep foundation pit multiparameter intelligent monitoring system to perform the method of any of claims 1-7.
  9. 9. A computer readable storage medium comprising instructions which, when run on a deep foundation pit multiparameter intelligent monitoring system, cause the deep foundation pit multiparameter intelligent monitoring system to perform the method of any one of claims 1-7.
  10. 10. A computer program product, characterized in that the computer program product, when run on a deep foundation pit multiparameter intelligent monitoring system, causes the deep foundation pit multiparameter intelligent monitoring system to perform the method according to any one of claims 1-7.

Description

Deep foundation pit multi-parameter intelligent monitoring method, system, medium and product based on Internet of things Technical Field The application relates to the technical field of deep foundation pit monitoring, in particular to a deep foundation pit multi-parameter intelligent monitoring method, system, medium and product based on the Internet of things. Background The deep foundation pit engineering is used as a key link for urban underground space development, and safety monitoring in the construction process has important significance for preventing collapse accidents. In the deep foundation pit excavation process, parameters such as displacement, stress, water pressure and the like of the periphery of a foundation pit can present complicated dynamic change characteristics due to the mutual coupling influence of factors such as soil stress redistribution, groundwater seepage, supporting structure deformation and the like. The multi-parameter coupling change phenomenon often causes hidden accumulation of dangerous situations under the condition that all monitoring parameters are not obviously out of limit, when single-point data are obviously abnormal, the safety margin of a foundation pit is greatly reduced, and even accidents such as collapse or sedimentation of surrounding buildings are caused when serious accidents occur. In the related art, in order to solve the technical problems, an intelligent deep foundation pit monitoring mode based on the Internet of things is provided, namely, real-time automatic acquisition of various physical parameters such as displacement, inclination, axial force, pore water pressure and the like is realized by arranging various sensor networks around the foundation pit, and trend analysis is carried out on data of each monitoring point by utilizing a time sequence prediction model. Specifically, a long-period memory network is adopted to carry out modeling learning on the historical data of each measuring point, the development trend of each parameter in a future period is predicted, and early warning is triggered when the predicted value exceeds a set threshold value, so that the acquisition frequency of monitoring data and the timeliness of early warning are improved to a certain extent. However, by adopting the intelligent monitoring mode of the deep foundation pit based on the internet of things, only the numerical variation trend of each monitoring point is independently predicted, and systematic risk accumulation effects caused by small variation of a plurality of monitoring parameters within a safety threshold range are difficult to effectively identify, so that even if the predicted numerical values of each measuring point are in the safety range, the deep foundation pit is still likely to have sudden dangerous situations due to accumulation of hidden association variation among a plurality of parameters, and further the reliability of multi-parameter cooperative early warning of the deep foundation pit in the related technology is poor. Disclosure of Invention The application provides a deep foundation pit multi-parameter intelligent monitoring method, a system, a medium and a product based on the Internet of things, which are used for improving the reliability of multi-parameter collaborative early warning of a deep foundation pit. In a first aspect, the application provides a multi-parameter intelligent monitoring method for a deep foundation pit based on the Internet of things, which is applied to the multi-parameter intelligent monitoring system for the deep foundation pit, and the method comprises the steps of determining a cross-physical quantity conduction relation between monitoring parameters in multi-type monitoring data under the condition that the multi-type monitoring data of the deep foundation pit are acquired, and constructing a heterogeneous association graph structure according to the cross-physical quantity conduction relation, wherein the heterogeneous association graph structure comprises monitoring nodes and influence factor nodes, the monitoring nodes are in one-to-one correspondence with the monitoring parameters, the influence factor nodes are in one-to-one correspondence with external influence events of the deep foundation pit, and association edges included in the heterogeneous association graph structure represent physical conduction paths between the monitoring nodes and the influence factor nodes; carrying out space-time conduction characteristic extraction on multi-type monitoring data on a heterogeneous association graph structure to obtain time sequence change characteristics of monitoring nodes and conduction influence characteristics of associated edges, carrying out dynamic weight disturbance operation on the associated edges according to the time sequence change characteristics and the conduction influence characteristics to obtain weight stability indexes of the associated edges, marking the associated edges with t