Risk comprehensive analysis method based on multi-measurement item association and measurement point neighbor information

Abstract

The invention relates to a risk comprehensive analysis method based on multi-measurement-item association and measurement point neighbor information, and the method comprises the steps: S1, collecting the basic information of a foundation pit project, carrying out the digital and structured processing of the basic information, and determining the position relation between a foundation pit and a measurement point, and the construction influence range of the foundation pit project; s2, on the basis of a directed graph depth-first search method, the alarm data of different measurement items and measurement points in the S1 form a plurality of correlation analysis networks, wherein the measurement items and the measurement points included in each correlation analysis network have correlation; s3, performing comprehensive analysis of line node data on each correlation analysis network in the S2 by adopting directed graph depth-first search, and judging an early warning value of the correlation analysis network; and S4, performing normalization processing on the early warning value of each correlation analysis network based on distance weighting, and performing comprehensive judgment on the safety state of the foundation pit engineering. The method has the advantage that the rapid and automatic comprehensive judgment of the underground engineering monitoring data risk is realized by comprehensively analyzing the measuring point data in the correlation analysis network.

Description

Technical Field

[0001] The invention relates to the technical field of foundation pit engineering monitoring, and in particular to a comprehensive risk analysis method based on multi-measurement item association and measurement point neighbor information. Background Art

[0002] There are always signs of risks occurring during the construction process of foundation pit projects, which can usually be reflected through monitoring data of the structure itself and the surrounding environment. How to analyze and discover the real risk situation from the monitoring data requires professional and systematic analysis by monitoring personnel.

[0003] The analysis of traditional monitoring data mainly proposes alarm values ​​based on design and construction experience. When the monitoring data reaches the alarm value, a data alarm is issued. The person in charge of monitoring conducts a comprehensive evaluation based on his or her own experience and on-site inspection results.

[0004] However, with the rapid development of urban construction, foundation pit projects are constantly developing in the direction of large numbers, deep depths, large areas, and complex shapes. The difficulty of engineering safety control is increasing, and the risk assessment based on monitoring data is becoming increasingly demanding. The technical experience of the person in charge of monitoring is difficult to meet the requirements of traditional risk assessment methods. At the same time, with the continuous development of big data and machine learning algorithms, some scholars have proposed the use of machine learning analysis methods based on neural networks, etc., but such algorithms also require a large amount of experience data from similar engineering projects for model training in order to obtain an accurate analysis model. In fact, each engineering project has its own characteristics, and the risk influencing factors are also different. It is difficult to use an accurate model to intuitively analyze the data. Summary of the invention

[0005] The purpose of the present invention is to provide a comprehensive risk analysis method based on the association of multiple measurement items and measurement point neighbor information in light of the deficiencies in the above-mentioned prior art, and to establish a correlation analysis network between different measurement items and measurement points through a directed graph depth-first algorithm, so as to comprehensively consider the impact of the positional relationship between various monitoring data sources on the risk status of underground projects, and to achieve rapid and automatic comprehensive judgment of the risk of foundation pit engineering monitoring data by comprehensively analyzing the measurement point data in the correlation analysis network, and to remind on-site personnel to take corresponding measures in a timely manner in combination with different risk statuses to avoid the occurrence of risk events.

[0006] The purpose of the present invention is achieved by the following technical solutions:

[0007] A comprehensive risk analysis method based on multi-measurement item association and measurement point neighbor information, characterized in that the comprehensive risk analysis method comprises the following steps:

[0008] S1. Collect basic information of foundation pit engineering, digitize and structure the basic information, determine the positional relationship between the measuring point and the foundation pit and the construction impact range of the foundation pit engineering;

[0009] S2, based on the directed graph depth-first search method, the alarm data of different measurement items and measurement points in step S1 are organized into several association analysis networks, and the measurement items and measurement points contained in each of the association analysis networks are associated with each other;

[0010] S3, using directed graph depth-first search to conduct a comprehensive analysis of line node data for each of the association analysis networks in step S2, and determining the warning value of the association analysis network;

[0011] S4. Normalizing the warning values ​​of each of the associated analysis networks based on distance weighting to make a comprehensive judgment on the safety status of the foundation pit project.

[0012] Step S1 includes the following steps:

[0013] S1.1 Collect static information of foundation pit engineering and organize the static information into a structured data set;

[0014] S1.2 Collect the location information of the foundation pit project, connect the corner points of the foundation pit in sequence to form a closed contour M, obtain the coordinates P(x, y) of the measuring points of different measuring items, and determine the relationship P(x, y, T, S, O) between the measuring point P and the closed contour M, where T indicates whether the measuring point P is inside or outside the polygon M, S indicates the shortest perpendicular distance between the measuring point P and the polygon M, and O indicates the category attribute of the measuring point;

[0015] S1.3 According to different measuring point coordinates P i (x i ,y i ), P j (x j ,y j ), get the distance R between the measuring points ij The influence range radius R' of each measuring point is set according to the static information of the foundation pit project combined with the empirical information, and the correlation Q between the measuring points is determined by combining the distance between the measuring points and the influence range radius of each measuring point;

[0016] When the distance between the measuring points is less than or equal to the influence range radius, Q=1, and the measuring points are strongly correlated. When the distance between the measuring points is greater than the influence range radius, Q=0, and the measuring points are weakly correlated.

[0017] Step S2 includes the following steps:

[0018] S2.1 Collect monitoring and inspection data of each measuring point in the foundation pit project, compare each measuring point with the designed alarm value, assign the alarm status to the measuring point, extract the measuring points with the over-alarm status to obtain the set NP, if it does not exceed the alarm value, it is judged to be normal, and the normal measuring points are merged into the set AP;

[0019] S2.2 Select a measurement point P from the measurement point set NP i As the center point, the measurement point P is extracted based on the spatial position relationship. i The set of influence measurement points C n ; First, according to the influence radius R' of step S1.3, search for the adjacent measurement point set C of Pi n ', determine the measurement point set C n 'The position relationship between the measuring point and the foundation pit is T. When T = 1, select C in the measuring point set n The measuring point where Q=1 is taken as measuring point P i The set of influence measurement points C n ; When T = 0, it means that the measuring point is outside the foundation pit; select C in the measuring point set n 'In which Q = 1 and the corresponding measuring point and Pi are on the same side of the foundation pit and the measuring points in the foundation pit constitute the influencing measuring point set C n ;

[0020] S2.3 Determine the measurement point set C n Check whether there is an intersection with NP. If there is no intersection, the measuring point is considered to be an accidental alarm point and can be continuously monitored. If there is an intersection, go to step S2.4 for further judgment.

[0021] S2.4 with P i As the center point, use the directed graph depth-first search method to search the set C n Establish an association analysis network for the measurement points in the set C n Middle distance P i The nearest point P ci , mark the path of the measuring point; then use P ci Point is the starting point for searching except P i The closest measuring point outside the point is also marked with a path, and this process is repeated until the set C n All the measurement points in the i Analysis network of measuring points i ;

[0022] S2.5 Select the set NP except C n Point P outside the intersection j As the center point, repeat steps S2.2 to S2.4 to obtain P j Analysis network of measuring points j Until all the measuring points in the set NP are traversed, all monitoring data analysis network S is obtained N .

[0023] Step S3 includes the following steps:

[0024] S3.1 Take the analysis network S N A subnet S iPerform correlation analysis to obtain subnet risk analysis results E fi ;

[0025] S3.2 Get subnet S i The number of nodes Num, with the subnet Si center point P i The starting point is determined according to the path marked by the directed graph depth-first search method. If there is no alarm, the child node state value is assigned 0. If there is an alarm, the measurement point is further determined to be different from the starting point P. i Whether the attributes are consistent, if not, the child node status is assigned 1 / (Num-1), otherwise the child node status is assigned R i / (M*(Num-1));

[0026] S3.3 takes the current point as the starting point, repeats step S3.2 to determine the next measurement point marked on the path, and determines the state of the child node;

[0027] S3.4 Pair of subnets S i All nodes of the sub-node state judgment are performed, and the warning state of each sub-node is added to obtain the current sub-network correlation analysis warning value E fi ;

[0028] S3.5 Traversal Analysis Network S N All subnets in S j Perform correlation analysis and obtain the analysis result E fn (j=1...N).

[0029] Step S4 includes the following steps:

[0030] S4.1 Considering the impact of the distance between the measuring point and the foundation pit on the monitoring risk, the correlation analysis results are processed by distance weighting to obtain the analysis results of each subnet WE i , WE i =S i *E fi (i=1.2...n); for the analysis result WE i Normalize R i =(WE i -WE min ) / (WE max -WE min ), and obtain the normalized correlation analysis warning value R of each subnet i Among them, WE i represents the correlation analysis warning value of the i-th group of subnets, WE max , WE min Represents the association analysis warning value WE n The maximum and minimum values ​​of S i Subnet center point P iThe shortest perpendicular distance to the edge of the foundation pit polygon, R i Indicates the normalized correlation analysis warning value of each subnet;

[0031] S4.2 The correlation analysis warning value R obtained in step S4.1 i Make a judgment, R i The closer the value is to 1, the greater the risk weight of the subnet is. i The value interval distribution of R is classified and counted. i The number R with a value between s1 and 1.0 Num1 , R i The number of values ​​between s1 and s2 is R Num2 , the number of values ​​less than s2 is R Num3 ;

[0032] S4.3 By R Numi The value distribution of the project is used for comprehensive graded early warning. The project can be comprehensively judged as a red warning. When the project is in yellow warning state, other states can be considered as orange warning state.

[0033] In step S4, the value range of s1 is 0.8-0.9, the value range of s2 is 0.5-0.8, and R s1 The value range is 0.3-0.4, R s2 The value range is 0.8~1.0.

[0034] The advantages of the present invention are:

[0035] 1) The impact range during the foundation pit construction process is further considered, which can effectively avoid false alarms caused by measurement errors at different measuring points; it can quickly and comprehensively divide different correlation analysis networks, effectively avoiding repeated search work;

[0036] 2) Comprehensively consider the influence correlation between different measurement items and different measurement points, and the influence relationship between different measurement points and the distance between the foundation pit, conduct a comprehensive analysis of the data, and effectively improve the comprehensiveness of the engineering risk analysis;

[0037] 3) It relies on the basic information of a single project and various monitoring data, has few dependencies, and has a certain degree of versatility. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flow chart of the method of the present invention;

[0039] Figure 2 It is a flowchart of the relevance analysis network search based on the directed graph depth-first search in the present invention;

[0040] Figure 3This is a flowchart of comprehensive project analysis and early warning based on the correlation analysis network in the present invention. DETAILED DESCRIPTION

[0041] The features of the present invention and other related features are further described in detail below through embodiments in conjunction with the accompanying drawings to facilitate understanding by those skilled in the art:

[0042] Example: Figures 1 to 3 As shown, in this embodiment, the comprehensive risk analysis method based on multi-measurement item association and measurement point neighbor information is used to comprehensively analyze the monitoring data of the foundation pit project, and specifically includes the following steps:

[0043] S1.1) Collect static information such as excavation depth, excavation area, excavation division, engineering geological division, hydrogeological division, support plan, surrounding environment, monitoring items, monitoring points, and patrol points of the foundation pit project, and organize the information into a structured data set. The specific categories of static information can be selected according to the needs of different foundation pit projects.

[0044] S1.2) Collect the position information of the corner points of the foundation pit, connect the corner points of the foundation pit in sequence to form a closed contour M. For example, when the foundation pit is a polygon, form a closed polygon M, obtain the coordinates P (x, y) of the measuring points of different measuring items, and determine the relationship between the measuring points P and M (x, y, T, S, O), where T indicates whether the measuring point P is inside or outside the polygon M, S indicates the shortest perpendicular distance between the measuring point P and the polygon M, and O indicates the category attribute of the measuring point.

[0045] S1.3) According to different measuring point coordinates P i (x i ,y i ), P j (x j ,y j ), get the distance R between the measuring points ij , and set the influence range radius R' according to the static information of the data set in step 1.1) combined with the empirical information, and determine the correlation Q between the measuring points and the distance R between the measuring points ij ≤R', Q=1, indicating strong correlation between the measurement points, R ij When >R', Q=0, indicating a weak correlation between the measurement points.

[0046] S2.1) Collect monitoring and inspection data of each measuring point in the foundation pit project, compare each measuring point with the designed alarm value, assign the alarm status to the measuring point, extract the measuring points with over-alarm status to obtain the set NP, if it does not exceed the alarm value, it is judged to be normal, and the normal measuring points are merged into the set AP.

[0047] S2.2) Select a measurement point P from the measurement point set NP iAs the center point, the measurement point P is extracted based on the spatial position relationship. i The set of influence measurement points C n ; First, search for P according to the influence radius R' of S1.3) i The set of adjacent measuring points C n ', determine the measurement point set C n 'The position relationship between the measuring point and the foundation pit is T. When T = 1, select C in the measuring point set n The measuring point where Q=1 is taken as measuring point P i The set of influence measurement points C n ; When T = 0, it means that the measuring point is outside the foundation pit; select C in the measuring point set n 'In which Q = 1 and the corresponding measuring point and Pi are on the same side of the foundation pit and the measuring points in the foundation pit constitute the influencing measuring point set C n .

[0048] S2.3) Determine the set of influencing measuring points C n Check whether there is any intersection with the measuring point set NP. If there is no intersection, the measuring point can be considered as an accidental alarm point and can be continuously monitored. If there is an intersection, go to S2.4) for further judgment.

[0049] S2.4) P i As the center point, the directed graph depth-first search method is used to search the influence measurement point set C n The measurement points in the set C are used to establish an association analysis network. First, search for the influencing measurement point set C. n Middle distance P i The nearest point P ci , mark the path of the measuring point; then use P ci Point is the starting point for searching except P i The closest measuring point outside the point is also marked with a path, and this process is repeated until the set C n All the measurement points in the i Analysis network of measuring points i .

[0050] S2.5) Select the set NP except C n Point P outside the intersection j As the center point, repeat steps S2.2 to S2.4 to obtain P j Analysis network of measuring points j Until all the measuring points in the measuring point set NP are traversed, all monitoring data analysis network S is obtained N .

[0051] S3.1) Take the analysis network S N A subnet S i Perform correlation analysis to obtain subnet risk analysis results E fi .

[0052] S3.2) Get subnet S i The number of nodes Num, in subnet S i Center point P i The starting point is determined according to the path marked by the directed graph depth-first search method. If there is no alarm, the child node state value is assigned 0. If there is an alarm, the measurement point is further determined to be different from the starting point P. i Whether the attributes are consistent, if not, the child node status is assigned 1 / (Num-1), otherwise the child node status is assigned R i / (M*(Num-1)).

[0053] S3.3) Taking the current point as the starting point, repeat S3.2) to judge the next measuring point marked on the path and judge the state of the child node.

[0054] S3.4) for subnet S i All nodes of the sub-node state judgment are performed, and the warning state of each sub-node is added to obtain the current sub-network correlation analysis warning value E fi .

[0055] S3.5) Traverse the analysis network S N All subnets in S j Perform correlation analysis and obtain the analysis result E fn (j=1...N).

[0056] S4.1) Considering the impact of the distance between the measuring point and the foundation pit on the monitoring risk, the correlation analysis results are processed by distance weighting to obtain the analysis results of each subnet WE i , WE i =S i *E fi (i=1.2...n); for the analysis result WE i Normalize R i =(WE i -WE min ) / (WE max -WE min ), and obtain the normalized correlation analysis warning value R of each subnet i ;

[0057] Among them, WE i represents the correlation analysis warning value of the i-th group of subnets, WE max , WE min Represents the association analysis warning value WE n The maximum and minimum values ​​of S i Subnet center point P i The shortest perpendicular distance to the edge of the foundation pit polygon, R iIndicates the normalized association analysis warning value of each subnet.

[0058] S4.2) Correlation analysis warning value R obtained in S4.1) i Make a judgment, R i The closer the value is to 1, the greater the risk weight of the subnet is. i The value interval distribution of R is classified and counted. i The number R with a value between s1 and 1.0 Num1 , R i The number of values ​​between s1 and s2 is R Num2 , the number of values ​​less than s2 is R Num3 .

[0059] S4.3) Through R Numi The value distribution of the project is used for comprehensive graded early warning. The project can be comprehensively judged as a red warning. When the project is in yellow warning state, the project can be comprehensively judged as yellow warning, and other states can be considered as orange warning.

[0060] In this step, the recommended parameter values ​​are s1: 0.8-0.9, s2: 0.5-0.8, R s1 :0.3~0.4, R s2 :0.8~1.0.

[0061] Although the above embodiments have described in detail the concepts and embodiments of the present invention with reference to the accompanying drawings, ordinary technicians in this field can recognize that various improvements and changes can still be made to the present invention without departing from the scope of the claims, so they are not described one by one here.

Claims

1. A comprehensive risk analysis method based on multi-measurement item association and measurement point neighbor information, characterized by: The comprehensive risk analysis method includes the following steps: S1. Collect basic information of foundation pit engineering, digitize and structure the basic information, determine the position relationship between the measuring point and the foundation pit and the construction impact range of the foundation pit engineering; S2, based on the directed graph depth-first search method, the alarm data of different measurement items and measurement points in step S1 are organized into several association analysis networks, and the measurement items and measurement points contained in each of the association analysis networks are associated with each other; S3, using directed graph depth-first search to conduct a comprehensive analysis of line node data for each of the association analysis networks in step S2, and determining the warning value of the association analysis network; S4. Process the warning values ​​of each of the associated analysis networks based on distance weighting to make a comprehensive judgment on the safety status of the foundation pit project.

2. According to claim 1, a comprehensive risk analysis method based on multi-measurement item association and measurement point neighbor information is characterized by: Step S1 includes the following steps: S1.1 collects static information of foundation pit engineering, and organizes the static information into a structured data set; S1.2 collects location information of foundation pit engineering, connects the corner points of foundation pit in sequence, forms a closed contour M, obtains the coordinates P(x, y) of the measuring points of different measuring items, and determines the relationship P(x, y, T, S, O) between the measuring point P and the closed contour M, where T indicates whether the measuring point P is inside or outside the polygon M, S indicates the shortest perpendicular distance between the measuring point P and the polygon M, and O indicates the category attribute of the measuring point; S1.3 According to different measuring point coordinates P i (x i ,y i ), P j (x j ,y j ), get the distance R between the measuring points ij The influence range radius R' of each measuring point is set according to the static information of the foundation pit project combined with the empirical information, and the correlation Q between the measuring points is determined by combining the distance between the measuring points and the influence range radius of each measuring point; When the distance between the measuring points is less than or equal to the influence range radius, Q=1, and the measuring points are strongly correlated. When the distance between the measuring points is greater than the influence range radius, Q=0, and the measuring points are weakly correlated.

3. The comprehensive risk analysis method based on multi-measurement item association and measurement point neighbor information according to claim 1 is characterized by: Step S2 includes the following steps: S2.1 Collect monitoring and inspection data of each measuring point in the foundation pit project, compare each measuring point with the designed alarm value, assign the alarm status to the measuring point, extract the measuring points with the over-alarm status to obtain the set NP, if it does not exceed the alarm value, it is judged to be normal, and the normal measuring points are merged into the set AP; S2.2 Select a measurement point P from the measurement point set NP i As the center point, the measurement point P is extracted based on the spatial position relationship. i The set of influence measurement points C n ; First, according to the influence radius R' of step S1.3, search for the adjacent measurement point set C of Pi n ', determine the measurement point set C n 'The position relationship between the measuring point and the foundation pit is T. When T = 1, select C in the measuring point set n The measuring point where Q=1 is taken as measuring point P i The set of influence measurement points C n ; When T = 0, it means that the measuring point is outside the foundation pit; select C in the measuring point set n 'In which Q = 1 and the corresponding measuring point and Pi are on the same side of the foundation pit and the measuring points in the foundation pit constitute the influencing measuring point set C n ; S2.3 Determine the measurement point set C n Check whether there is an intersection with NP. If there is no intersection, the measuring point is considered to be an accidental alarm point and can be continuously monitored. If there is an intersection, go to step S2.4 for further judgment. S2.4 with P i As the center point, use the directed graph depth-first search method to search the set C n Establish an association analysis network for the measurement points in the set C n Middle distance P i The nearest point P ci , mark the path of the measuring point; then use P ci Point is the starting point for searching except P i The closest measuring point outside the point is also marked with a path, and this process is repeated until the set C n All the measurement points in the i Analysis network of measuring points i ; S2.5 Select the set NP except C n Point P outside the intersection j As the center point, repeat steps S2.2 to S2. S2.4 to obtain P j Analysis network of measuring points j Until all the measuring points in the set NP are traversed, all monitoring data analysis network S is obtained N .

4. The comprehensive risk analysis method based on multi-measurement item association and measurement point neighbor information according to claim 1 is characterized by: Step S3 includes the following steps: S3.1 Take the analysis network S N A subnet S i Perform correlation analysis to obtain subnet risk analysis results E fi ; S3.2 Get subnet S i The number of nodes Num, with the subnet Si center point P i The starting point is determined according to the path marked by the directed graph depth-first search method. If there is no alarm, the child node state value is assigned 0. If there is an alarm, the measurement point is further determined to be different from the starting point P. i Whether the attributes are consistent, if not, the child node status is assigned 1 / (Num-1), otherwise the child node status is assigned R i / (M*(Num-1)); S3.3 takes the current point as the starting point, repeats step S3.2 to judge the next measuring point of the path mark, and judges the state of the child node; S3.4 Pair of subnets S i All nodes of the sub-node state judgment are performed, and the warning state of each sub-node is added to obtain the current sub-network correlation analysis warning value E fi ; S3.5 Traversal Analysis Network S N All subnets in S j Perform correlation analysis and obtain the analysis result E fn (j=1...N).

5. The comprehensive risk analysis method based on multi-measurement item association and measurement point neighbor information according to claim 1 is characterized by: Step S4 includes the following steps: S4.1 Considering the impact of the distance between the measuring point and the foundation pit on the monitoring risk, the correlation analysis results are processed by distance weighting to obtain the analysis results of each subnet WE i , WE i =S i *E fi (i=1.2...n); for the analysis result WE i Normalize R i =(WE i -WE min ) / (WE max -WE min ), and obtain the normalized correlation analysis warning value R of each subnet i Among them, WE i represents the correlation analysis warning value of the i-th group of subnets, WE max , WE min Represents the association analysis warning value WE n The maximum and minimum values ​​of S i Subnet center point P i The shortest perpendicular distance to the edge of the foundation pit polygon, R i Indicates the normalized correlation analysis warning value of each subnet; S4.2 The correlation analysis warning value R obtained in step S4.1 i Make a judgment, R i The closer the value is to 1, the greater the risk weight of the subnet is. i The value interval distribution of R is classified and statistically analyzed. i The number R with a value between s1 and 1.0 Num1 , R i The number of values ​​between s1 and s2 is R Num2 , the number of values ​​less than s2 is R Num3 ; S4.3 By R Numi The value distribution of the project is used for comprehensive graded early warning. The project can be comprehensively judged as a red warning. The project can be comprehensively judged as a yellow warning, and other states can be considered as orange warning.

6. The comprehensive risk analysis method based on multi-measurement item association and measurement point neighbor information according to claim 5 is characterized by: In step S4, the value range of s1 is 0.8-0.9, the value range of s2 is 0.5-0.8, and R s1 The value range is 0.3~0.4, R s2 The value range is 0.8~1.0.

Images