Unfavorable geological foundation pit automatic monitoring analysis and early warning management system and method

Abstract

The invention relates to an unfavorable geology foundation pit automatic monitoring analysis and early warning management system and an unfavorable geology foundation pit automatic monitoring analysis and early warning management method. Comprising a rock-soil database module, a field soil texture image acquisition robot, a soil body model module, a foundation pit design model module, an intelligent monitoring module, a data operation center, a calling and display module, a human-computer interaction interface, an early warning and alarm system module and a consultation decision module. According to the system, various indexes of the foundation pit can be accurately monitored in real time, potential safety hazards can be found in time, and disasters such as foundation pit collapse, mud burst and water burst are effectively prevented, so that the safety of constructors and surrounding buildings is guaranteed; all-weather and uninterrupted data acquisition can be carried out, personal errors are reduced, and the accuracy and reliability of monitoring data are improved; meanwhile, the system supports remote control, the monitoring period is flexibly adjusted according to actual conditions, and the monitoring efficiency is further improved.

Description

Technical Field

[0001] The present invention relates to the technical field of foundation pit construction monitoring, and in particular to an automated monitoring and analysis and early warning management system and method for foundation pits with poor geology. Background Art

[0002] Foundation pit monitoring is related to the safety of building foundation pits and the protection of the surrounding environment of foundation pits. Therefore, the quality of monitoring is of vital importance, and the results must be reliable, the technology must be advanced, and the economy must be reasonable. However, for deep foundation pit projects in special soils and corrosive environments such as expansive soils, collapsible loess, red clay, frozen soil, saline soil, and highly sensitive soft soil, traditional monitoring methods are discrete, incomplete, and non-intuitive. Moreover, due to the influence of the construction process, the synchronization between monitoring and construction is often poor, and monitoring information cannot be provided in a timely manner. How to conduct high-frequency and real-time observations of foundation pits to understand the feedback of comprehensive information during the entire construction stage of the foundation pit has become an important problem in the quality and efficiency of foundation pit monitoring.

[0003] At present, the automation and intelligence levels of some foundation pit monitoring facilities are low, and they are unable to realize the intelligent layout of foundation pit monitoring network, the identification and effect analysis of adverse geological conditions in foundation pits, and the prediction of foundation pit monitoring data to optimize subsequent construction procedures. Therefore, there is an urgent need for a platform that can be widely used in the analysis and early warning management of foundation pit monitoring data in adverse geological environments to solve the above problems. Summary of the invention

[0004] The present invention aims to solve the deficiencies of the prior art and to provide an automated monitoring and analysis and early warning management system and method for foundation pits with poor geology.

[0005] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: an automatic monitoring, analysis and early warning management system for bad geological foundation pits, including a geotechnical database module, an on-site soil image acquisition robot, a soil model module, a foundation pit design model module, an intelligent monitoring module, a data operation center, a call and display module, a human-computer interaction interface, an early warning and alarm system module, and a consulting and decision-making module;

[0006] The geotechnical database module collects soil particle vector diagrams, geotechnical classifications, hydrogeological parameters, physical parameters and mechanical parameters to form a conventional geotechnical database. Through artificially supervised target learning, training forms a packaged module with data storage, recognition and vectorization functions, namely a digital geotechnical database.

[0007] The on-site soil image acquisition robot forms points with precise elevation and coordinates by checking the coordinates with the on-site control points. It forms a point cloud stereo model with elevation and soil image superposition information by taking three-dimensional oblique photography of the excavation surface. It performs image recognition by calling the digital geotechnical database and outputs a soil layer parameter information module with soil layer top elevation, hydrological conditions, soil physical and mechanical properties parameter information.

[0008] The soil model module automatically fits the top and bottom elevations of each soil layer in the point information into a virtual surface based on the drilling sampling information in the geological survey report. The two virtual surfaces are separated by three-dimensional soil layers. The soil layer models are constructed from top to bottom to form an initialized soil model, which is used to generate an initialized foundation pit design model and compare the survey values ​​with the actual soil layer parameter changes.

[0009] The foundation pit design model module performs finite element simulation on the construction conditions of the initial soil model generated by the geological survey report according to the foundation pit support design plan, earthwork excavation conditions, and precipitation plan to form a foundation pit design initialization model. The calculation results are used by the intelligent monitoring module to lay out the monitoring network and determine the monitoring method, expected monitoring point values, and alarm values;

[0010] The data operation center is connected with the soil layer parameter information generation module, soil model module, and foundation pit design model module for soil physical and mechanical properties and seepage stability analysis, foundation pit support structure deformation and strength analysis, and soil excavation and foundation pit dewatering on surrounding buildings, pipelines, and roads. It is also used for monitoring point simulation values, monitoring point measured values ​​and simulation values, and early warning value comparative relationship analysis; the data operation center is connected with the call and display module and the consulting decision module for manual interaction to form subsequent working condition monitoring values ​​and trend predictions, correction plans, error analysis and risk assessment, and emergency response plans;

[0011] The calling and display module is mainly composed of the chart data I / O display module, which is connected to the data operation center and is used to call all the data in the data operation center, and upload, download, store, and access models and monitoring data according to user customization; the calling and display module sets access permissions; the calling and display module automatically generates various pictures, tables, documents, and reports for manual processing, review, and archiving; through the connection with the human-computer interaction interface, it displays the virtual construction conditions of the project, the actual construction conditions of the project, various tables and data processing diagrams, various parameters and early warning conditions;

[0012] The early warning and alarm system module is used to send out alarm signals when soil parameters change significantly, monitoring data and simulation prediction data exceed alarm values, or monitoring data exceeds the warning values ​​set by the project, and transmit them to the human-computer interaction interface. The consulting decision-making module calls the data computing center to form optimization plans and emergency plans, and transmits the information through the local area network to the preset management personnel's mobile phone, triggering the project emergency response procedure.

[0013] In particular, the data calculation center presets an exit interface reminder after every 10,000 calculations of a single simulation, which lasts for 1 minute and allows the user to choose to exit the calculation. If no choice is made, the calculation will continue. After each calculation, a rough result of the data trend will be obtained, and reasonable suggestions will be given according to the database experience value for the subsequent adjustment of the manual interaction technical parameters.

[0014] An automated monitoring, analysis and early warning management method for bad geological foundation pits, using an automated monitoring, analysis and early warning management system for bad geological foundation pits, comprises the following steps:

[0015] S1. Geotechnical database module training to form a digital geotechnical database;

[0016] S2, the on-site soil image acquisition robot collects information and calls the digital geotechnical database to generate soil layer parameter information;

[0017] S3, soil model module generates soil model M according to geological survey report i , i=0, 1, 2, soil layer parameter information in S2 and soil model M i Information versus compliance;

[0018] When there is a bad geological interlayer that has not been explored by geological survey, or the soil layer parameters do not match the soil quality within the range of more than two meters thick in the initialized soil model, or the physical and mechanical parameter threshold of the soil layer identification is greatly different from the value in the geological survey report and is not within the allowable error range, it is "No", triggering the early warning mechanism of the early warning and alarm system module, sending early warning and alarm information to the project supervisor through the local area network and transmitting it to the human-computer interaction interface. After manual confirmation and processing, the soil model M i Make adjustments on

[0019] When other non-critical adjustments occur in the soil layer, that is, the top or bottom elevation of the soil layer fluctuates within the allowable error range, the answer is "yes". At this time, in the soil model M i After adjustment, the soil model M i+1 Replaces the previous version of soil model M i , forming an iteratively updated soil model M i+1 , participate in subsequent calculations;

[0020] S4, foundation pit design model module according to the foundation pit support design, excavation and dewatering plan, and soil model M i Conduct finite element simulation to generate foundation pit design model K j , j = 1, 2…;

[0021] S5. Through the relevant theoretical knowledge, specifications, standards, monitoring methods and costs of foundation pit monitoring, foundation pit design model K jand intelligent deployment of monitoring network and monitoring method using human-computer interaction interface;

[0022] S6. Deploy sensors at monitoring points;

[0023] S7. Monitoring points are networked and fed back to the foundation pit design model K j On the top, correct the monitoring point information and generate monitoring data;

[0024] S8, foundation pit design model K j Corresponding monitoring point data prediction, and transmit the prediction data to the data operation center;

[0025] S9. Check the consistency of the monitoring data with the predicted data. If the deviation is outside the allowable error range, it is "No", triggering the early warning mechanism of the early warning and alarm system module, sending early warning and alarm information to the project supervisor through the local area network and transmitting it to the human-computer interaction interface; if the deviation is within the allowable error range, it is "Yes". At this time, check the consistency of the monitoring data with the early warning value. If the deviation is outside the allowable error range, it is "No", triggering the early warning mechanism of the early warning and alarm system module, sending early warning and alarm information to the project supervisor through the local area network and transmitting it to the human-computer interaction interface; if the deviation is within the allowable error range, it is "Yes", and transmitting it to the data operation center;

[0026] S10, the calling and display module and the consulting decision-making module call the data of the data operation center, and interact through the human-computer interaction interface to form subsequent working condition monitoring values ​​and trend predictions, correction plans, error analysis and risk assessment, and emergency response plans.

[0027] In particular, the training steps of the geotechnical database module are as follows:

[0028] A1. Data acquisition: collect a large number of soil particle vector diagrams through the image acquisition system and perform preprocessing;

[0029] A2, perform local feature analysis and cluster analysis on the data collected in A1 through a feature analyzer, and form a corresponding relationship between the image and the target;

[0030] A3. Assign values ​​to the target, including category, physical and mechanical parameters;

[0031] A4. Form a test version of geotechnical database;

[0032] A5. Input new pictures for N tests and call the test version of the geotechnical database for identification;

[0033] A6: Manually supervise the target learning and form the final digital geotechnical database, which will be updated and returned to A4.

[0034] In particular, in A3, after the assignment, the filtering SVM principle is applied for optimization, that is: , where ω is the hyperplane normal vector, b is the bias term, and x i is the eigenvector, y i is the class label, usually +1 or -1.

[0035] In particular, the steps for the intelligent monitoring module to deploy the monitoring network and determine the monitoring method are as follows:

[0036] B1. Summarize and develop a monitoring database based on relevant theoretical knowledge, specifications, standards, monitoring methods and costs, basic components of foundation pit monitoring, operation procedures, precautions, calculation methods, risk assessment methods, and typical cases;

[0037] B2. Through logical programming, the AdaBoost ensemble learning method is used to combine multiple weak learners to update the weights of correctly classified samples and realize automatic feature selection for multi-classification problems. The algorithm is expressed as follows: and , where: t is the learner number, n represents the total number of weak learners, , , Represent the weight, prediction function, and error rate of the t-th learner respectively;

[0038] Call the foundation pit design model and monitoring database, identify similar typical engineering cases, and automatically deploy monitoring networks and monitoring methods based on the characteristic data points of the foundation pit design model in accordance with the location, method and quantity specified in the specification. After economic analysis, select the economically reasonable monitoring plan, and intelligently deploy the monitoring network and monitoring method V1.0;

[0039] B3. Through the human-computer interaction interface, adjust the points and monitoring methods of the generated monitoring network, and intelligently deploy the monitoring network and monitoring method V2.0;

[0040] B4. Deploy sensors at monitoring points;

[0041] B5. After the monitoring network is arranged, the monitoring points are fed back to the foundation pit design model through the network, and after modifications are made to the model, the monitoring point information is corrected, and the monitoring network and monitoring method V3.0 are intelligently arranged. During this period, selective manual intervention can also be carried out through the human-computer interaction interface.

[0042] In particular, the early warning and alarm system module works as follows:

[0043] C1. By default, the initialized soil model generated by the geological survey report is used as the true value of the soil layer parameters;

[0044] Manually set soil layer parameter values ​​and error tolerances;

[0045] Compare the on-site soil quality with the model information. If the difference is greater than the allowable error value, a Class A warning is triggered.

[0046] The monitoring data is compared with the emergency alarm value. If it is greater than the allowable error value, a first-level warning is triggered and transmitted to the human-computer interaction interface;

[0047] C2. Foundation pit design model K j The corresponding monitoring point simulation value is used as the monitoring data target value;

[0048] Manually set monitoring data target values ​​and error tolerances;

[0049] The monitoring data is compared with the predicted data. If the difference is greater than the allowable error value, a Class B warning is triggered;

[0050] The monitoring data is compared with the emergency alarm value. If it is greater than the allowable error value, a first-level warning is triggered and transmitted to the human-computer interaction interface;

[0051] C3. The specification stipulates the foundation pit warning value corresponding to the foundation pit grade as the monitoring data warning limit;

[0052] Manually set monitoring data warning limits and error tolerances;

[0053] The monitoring data is compared with the warning value. If it is greater than the allowable error value, a Class C warning is triggered;

[0054] The monitoring data is compared with the emergency alarm value. If it is greater than the allowable error value, a first-level warning is triggered and transmitted to the human-computer interaction interface;

[0055] C4, corresponding to 1.25 times of the foundation pit warning value, as the emergency response alarm threshold;

[0056] Manually set emergency response alarm thresholds and error tolerances;

[0057] The monitoring data is compared with the emergency alarm value. If it is greater than the allowable error value, a first-level warning is triggered and transmitted to the human-computer interaction interface;

[0058] C5. After the information in the above steps C1, C2, C3 and C4 is transmitted to the human-computer interaction interface, the control room personnel are reminded to pay special attention, and the data operation center is called through the consulting decision-making module to form optimization plans and emergency plans. The information is sent to the project supervisor through the local area network to send early warning and alarm information, triggering the project emergency response procedure.

[0059] The beneficial effects of the present invention are as follows: the automated monitoring, analysis and early warning management system for poor geological foundation pits provided by the present invention is optimized by applying the filtering SVM principle, and through artificially supervised target learning, a packaging module with data storage, recognition and vectorization functions is trained, namely a digital geotechnical database; the AdaBoost integrated learning method is adopted to update the weights of correctly classified samples, and automatic feature selection for multi-classification problems is easily realized, forming a human-computer interactive intelligent deployment monitoring network and determining monitoring methods; and numerical simulation and intelligent networked monitoring are integrated to form a virtual / real three-dimensional visualized intelligent inspection and intelligent sub-item graded early warning and alarm system in the construction stage of foundation pit engineering.

[0060] In view of this, the system can monitor various indicators of the foundation pit in real time and accurately, such as horizontal displacement, vertical displacement, deep displacement, cracks, internal force of the support structure, soil pressure, groundwater level, etc., timely discover potential safety hazards, and effectively prevent the occurrence of disasters such as foundation pit collapse and sudden mud and water gushing, thereby ensuring the safety of construction personnel and surrounding buildings. Compared with traditional manual monitoring methods, the automated monitoring system can collect data around the clock and uninterruptedly, reduce human errors, and improve the accuracy and reliability of monitoring data. At the same time, the system supports remote control and flexibly adjusts the monitoring cycle according to actual conditions to further improve monitoring efficiency. The system has functions such as data storage, analysis, and display, and displays monitoring data on terminal devices in the form of charts, reports, curves, etc., which is convenient for construction management personnel to view and analyze at any time, fully understand the status of foundation pit construction, and improve construction efficiency and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 This is a workflow diagram of the automatic monitoring, analysis and early warning management system for bad geological foundation pits of the present invention;

[0062] Figure 2 It is a flow chart of geotechnical database module training of the present invention;

[0063] Figure 3 It is a working mechanism diagram of intelligently laying out a monitoring network and determining a monitoring method of the present invention;

[0064] Figure 4 It is the module workflow diagram of the early warning and alarm system of the present invention;

[0065] The following is a detailed description of the embodiments of the present invention with reference to the accompanying drawings. DETAILED DESCRIPTION

[0066] The present invention will be further described below in conjunction with embodiments:

[0067] like Figure 1-Figure 4As shown, an automated monitoring, analysis and early warning management system for foundation pits with poor geological conditions includes a geotechnical database module, an on-site soil image acquisition robot, a soil model module, a foundation pit design model module, an intelligent monitoring module, a data operation center, a calling and display module, a human-computer interaction interface, an early warning and alarm system module, and a consulting and decision-making module.

[0068] The geotechnical database module collects soil particle vector diagrams, geotechnical classifications, hydrogeological parameters, physical parameters and mechanical parameters to form a conventional geotechnical database. Through artificially supervised target learning, it is trained to form a packaged module with data storage, recognition and vectorization functions, namely the digital geotechnical database.

[0069] The on-site soil image acquisition robot forms points with precise elevations and coordinates by calibrating the coordinates with the on-site control points. It forms a point cloud stereo model with elevation and soil image superposition information through three-dimensional oblique photography of the excavation surface. It performs image recognition by calling the digital geotechnical database and outputs a soil layer parameter information module with soil layer top elevation, hydrological conditions, and soil physical and mechanical property parameter information.

[0070] The soil model module automatically fits the top and bottom elevations of each soil layer in the point information into a virtual surface based on the drilling sampling information in the geological survey report. The two virtual surfaces are separated by a three-dimensional soil layer. The soil layer models are formed from top to bottom to form an initial soil model, which is used to generate the initial foundation pit design model and compare the survey value with the actual soil layer parameter changes; the soil layer and hydrogeological parameter information output by the on-site soil image acquisition robot are compared with the initial soil model. When there is an unfavorable geological interlayer that has not been explored by the geological survey, or the soil layer parameters do not match the soil in the initial soil model within a range of more than 2 meters thick, or the soil layer identification physical and mechanical parameter threshold is significantly different from the geological survey report value, the early warning mechanism is triggered. After manual confirmation and processing, adjustments are made to the soil model; when other non-critical adjustments occur to the soil layer, such as the floating elevation of the top or bottom of the soil layer, adjustments are made to the soil model. The adjusted soil model will replace the previous version of the soil model to form an iteratively updated soil model and participate in subsequent calculations.

[0071] The foundation pit design model module performs finite element simulation on the construction conditions of the initialized soil model generated by the geological survey report according to the foundation pit support design plan, earth excavation conditions, and precipitation plan to form a foundation pit design initialization model. The calculation results are used by the intelligent monitoring module to lay out the monitoring network and determine the monitoring method, expected monitoring point values, and alarm values.

[0072] The intelligent monitoring module is used to intelligently lay out the monitoring network and determine the monitoring method. Through the theoretical knowledge, specifications, standards, monitoring methods and costs, and typical cases related to foundation pit monitoring, the monitoring database is summarized and developed; through logical programming, the foundation pit design model is called to intelligently lay out the monitoring network and monitoring methods; through the human-computer interaction interface, the generated monitoring network (points) and monitoring methods are adjusted. In the process of setting sensors at the monitoring points, due to the inconvenient construction problems and installation deviation problems that may be encountered, after the monitoring network is arranged, the monitoring points are networked and fed back to the foundation pit design model, and after the model is modified, it is transmitted to the data operation center to form the final monitoring network and monitoring method.

[0073] The data operation center is connected with the soil layer parameter information generation module, soil model module, and foundation pit design model module, and is used for soil physical and mechanical properties and seepage stability analysis, foundation pit support structure deformation and strength analysis, and soil excavation, foundation pit dewatering on surrounding buildings, pipelines, and roads. It is also used for monitoring point simulation values, monitoring point measured values ​​and simulation values, and early warning value comparative relationship analysis; the data operation center is connected with the call and display module and the consulting decision module for manual interaction to form subsequent working condition monitoring values ​​and trend predictions, correction plans, error analysis and risk assessment, and emergency response plans; due to the large discreteness of the geotechnical engineering constitutive model, data may not converge due to special circumstances. The data operation center presets a reminder to exit the interface after every 10,000 operations of a single simulation, which lasts for 1 minute. You can choose to exit the operation. If you do not choose, the calculation will continue; after each operation, a rough result of the data trend will be obtained, and reasonable suggestions will be given according to the database experience value for the subsequent adjustment of the technical parameters of manual interaction.

[0074] The calling and display module is mainly based on the chart data I / O display module, which is connected to the data operation center. It is used to call all the data in the data operation center and upload, download, store, and review models and monitoring data according to user customization; the calling and display module sets the review authority; the calling and display module automatically generates various pictures, tables, documents, and reports for manual processing, review, and archiving; by connecting with the human-computer interaction interface, it displays the project's virtual construction conditions, the project's actual construction conditions, various tables and data processing diagrams, various parameters and early warning conditions.

[0075] The early warning and alarm system module is used to send out alarm signals when soil parameters change significantly, monitoring data and simulation prediction data exceed alarm values, or monitoring data exceeds the warning values ​​set by the project, and transmit them to the human-computer interaction interface. The consulting decision-making module calls the data computing center to form optimization plans and emergency plans, and transmits the information through the local area network to the preset management personnel's mobile phone, triggering the project emergency response procedure.

[0076] like Figure 1As shown, a method for automatic monitoring, analysis and early warning management of bad geological foundation pits is performed using an automatic monitoring, analysis and early warning management system for bad geological foundation pits, comprising the following steps:

[0077] S1. Geotechnical database module training to form a digital geotechnical database;

[0078] S2, the on-site soil image acquisition robot collects information and calls the digital geotechnical database to generate soil layer parameter information;

[0079] S3, soil model module generates soil model M according to geological survey report i , i=0, 1, 2, soil layer parameter information in S2 and soil model M i Information versus compliance;

[0080] When there is a bad geological interlayer that has not been explored by geological survey, or the soil layer parameters do not match the soil quality within the range of more than two meters thick in the initialized soil model, or the physical and mechanical parameter threshold of the soil layer identification is greatly different from the value in the geological survey report and is not within the allowable error range, it is "No", triggering the early warning mechanism of the early warning and alarm system module, sending early warning and alarm information to the project supervisor through the local area network and transmitting it to the human-computer interaction interface. After manual confirmation and processing, the soil model M i Make adjustments on

[0081] When other non-critical adjustments occur in the soil layer, that is, the top or bottom elevation of the soil layer fluctuates within the allowable error range, the answer is "yes". At this time, in the soil model M i After adjustment, the soil model M i+1 Replaces the previous version of soil model M i , forming an iteratively updated soil model M i+1 , participate in subsequent calculations;

[0082] S4, foundation pit design model module according to the foundation pit support design, excavation and dewatering plan, and soil model M i Conduct finite element simulation to generate foundation pit design model K j , j = 1, 2…;

[0083] S5. Through the relevant theoretical knowledge, specifications, standards, monitoring methods and costs of foundation pit monitoring, foundation pit design model K j and intelligent deployment of monitoring network and monitoring method using human-computer interaction interface;

[0084] S6. Deploy sensors at monitoring points;

[0085] S7. Monitoring points are networked and fed back to the foundation pit design model K j On the top, correct the monitoring point information and generate monitoring data;

[0086] S8, foundation pit design model K j Corresponding monitoring point data prediction, and transmit the prediction data to the data operation center;

[0087] S9. Check the consistency of the monitoring data with the predicted data. If the deviation is outside the allowable error range, it is "No", triggering the early warning mechanism of the early warning and alarm system module, sending early warning and alarm information to the project supervisor through the local area network and transmitting it to the human-computer interaction interface; if the deviation is within the allowable error range, it is "Yes". At this time, check the consistency of the monitoring data with the early warning value. If the deviation is outside the allowable error range, it is "No", triggering the early warning mechanism of the early warning and alarm system module, sending early warning and alarm information to the project supervisor through the local area network and transmitting it to the human-computer interaction interface; if the deviation is within the allowable error range, it is "Yes", and transmitting it to the data operation center;

[0088] S10, the calling and display module and the consulting decision-making module call the data of the data operation center, and interact through the human-computer interaction interface to form subsequent working condition monitoring values ​​and trend predictions, correction plans, error analysis and risk assessment, and emergency response plans.

[0089] like Figure 2 As shown in the figure, the training steps of the geotechnical database module are as follows:

[0090] A1. Data collection: a large number of soil particle vector diagrams are collected and preprocessed through the image acquisition system; the soil particle vector diagram contains multi-dimensional information such as soil particle size, dimension, shape, color, etc., which is convenient for data collection;

[0091] A2: Perform local feature analysis and cluster analysis on the data collected in A1 through a feature analyzer, and form a corresponding relationship between the image and the target; local feature analysis can extract key local features from each image through a computer vision algorithm, and convert the extracted features into a numerical vector form; cluster analysis, apply a clustering algorithm to group feature vectors, and identify soil particle groups with similar features; the connection between a specific type of soil particle and its image can be established based on the clustering results, and this connection helps to better understand the characteristics of different types of soil particles and their potential application value;

[0092] A3. Assign values ​​to the target, including category, physical and mechanical parameters; after assignment, apply the filtering SVM principle for optimization, namely: , where ω is the hyperplane normal vector, b is the bias term, and x i is the eigenvector, y i is the class label, usually +1 or -1;

[0093] A4. Form a test version of geotechnical database;

[0094] A5. Input new pictures for N tests and call the test version of the geotechnical database for identification;

[0095] A6: Manually supervise the target learning and form the final digital geotechnical database, which will be updated and returned to A4.

[0096] like Figure 3 As shown, the steps for the intelligent monitoring module to lay out the monitoring network and determine the monitoring method are as follows:

[0097] B1. Summarize and develop a monitoring database based on relevant theoretical knowledge, specifications, standards, monitoring methods and costs, basic components of foundation pit monitoring, operation procedures, precautions, calculation methods, risk assessment methods, and typical cases;

[0098] B2. Through logical programming, the AdaBoost (Adaptive Boosting) ensemble learning method is used to combine multiple weak learners (such as small decision trees) to update the weights of correctly classified samples and realize automatic feature selection for multi-classification problems. The algorithm is expressed as follows: and , where: t is the learner number, n represents the total number of weak learners, , , Represent the weight, prediction function, and error rate of the t-th learner respectively;

[0099] Call the foundation pit design model and monitoring database, identify similar typical engineering cases, and automatically deploy monitoring networks and monitoring methods based on the characteristic data points of the foundation pit design model in accordance with the location, method and quantity specified in the specification. After economic analysis, select the economically reasonable monitoring plan, and intelligently deploy the monitoring network and monitoring method V1.0;

[0100] B3. Through the human-computer interaction interface, adjust the points and monitoring methods of the generated monitoring network, and intelligently deploy the monitoring network and monitoring method V2.0;

[0101] B4. Deploy sensors at monitoring points;

[0102] B5. After the monitoring network is arranged, the monitoring points are fed back to the foundation pit design model through the network, and after modifications are made to the model, the monitoring point information is corrected, and the monitoring network and monitoring method V3.0 are intelligently arranged. During this period, selective manual intervention can also be carried out through the human-computer interaction interface.

[0103] like Figure 4 As shown, the working steps of the early warning and alarm system module are as follows:

[0104] C1. By default, the initialized soil model generated by the geological survey report is used as the true value of the soil layer parameters;

[0105] Manually set soil layer parameter values ​​and error tolerances;

[0106] Compare the on-site soil quality with the model information. If the difference is greater than the allowable error value, a Class A warning is triggered.

[0107] The monitoring data is compared with the emergency alarm value. If it is greater than the allowable error value, a first-level warning is triggered and transmitted to the human-computer interaction interface;

[0108] C2. Foundation pit design model K j The corresponding monitoring point simulation value is used as the monitoring data target value;

[0109] Manually set monitoring data target values ​​and error tolerances;

[0110] The monitoring data is compared with the predicted data. If the difference is greater than the allowable error value, a Class B warning is triggered;

[0111] The monitoring data is compared with the emergency alarm value. If it is greater than the allowable error value, a first-level warning is triggered and transmitted to the human-computer interaction interface;

[0112] C3. The specification stipulates the foundation pit warning value corresponding to the foundation pit grade as the monitoring data warning limit;

[0113] Manually set monitoring data warning limits and error tolerances;

[0114] The monitoring data is compared with the warning value. If it is greater than the allowable error value, a Class C warning is triggered;

[0115] The monitoring data is compared with the emergency alarm value. If it is greater than the allowable error value, a first-level warning is triggered and transmitted to the human-computer interaction interface;

[0116] C4, corresponding to 1.25 times of the foundation pit warning value, as the emergency response alarm threshold;

[0117] Manually set emergency response alarm thresholds and error tolerances;

[0118] The monitoring data is compared with the emergency alarm value. If it is greater than the allowable error value, a first-level warning is triggered and transmitted to the human-computer interaction interface;

[0119] C5. After the information in the above steps C1, C2, C3 and C4 is transmitted to the human-computer interaction interface, the control room personnel are reminded to pay special attention, and the data operation center is called through the consulting decision-making module to form optimization plans and emergency plans. The information is sent to the project supervisor through the local area network to send early warning and alarm information, triggering the project emergency response procedure.

[0120] The automated monitoring, analysis and early warning management system for foundation pits with poor geological conditions provided by the present invention is optimized by applying the filtering SVM principle, and through artificially supervised target learning, a packaging module with data storage, recognition and vectorization functions is trained, namely, a digital geotechnical database; the AdaBoost integrated learning method is adopted to update the weights of correctly classified samples, and automatic feature selection for multi-classification problems is easily realized, forming a human-computer interactive intelligent deployment monitoring network and determining monitoring methods; and numerical simulation and intelligent networked monitoring are integrated to form a virtual / real three-dimensional visualized intelligent inspection and intelligent item-by-item graded early warning and alarm system in the construction stage of foundation pit engineering.

[0121] In view of this, the system can monitor various indicators of the foundation pit in real time and accurately, such as horizontal displacement, vertical displacement, deep displacement, cracks, internal force of the support structure, soil pressure, groundwater level, etc., timely discover potential safety hazards, and effectively prevent the occurrence of disasters such as foundation pit collapse and sudden mud and water gushing, thereby ensuring the safety of construction personnel and surrounding buildings. Compared with traditional manual monitoring methods, the automated monitoring system can collect data around the clock and uninterruptedly, reduce human errors, and improve the accuracy and reliability of monitoring data. At the same time, the system supports remote control and flexibly adjusts the monitoring cycle according to actual conditions to further improve monitoring efficiency. The system has functions such as data storage, analysis, and display, and displays monitoring data on terminal devices in the form of charts, reports, curves, etc., which is convenient for construction management personnel to view and analyze at any time, fully understand the status of foundation pit construction, and improve construction efficiency and quality.

[0122] The above is an exemplary description of the present invention. Obviously, the specific implementation of the present invention is not limited to the above-mentioned method. As long as various improvements are made using the method concept and technical solution of the present invention, or they are directly applied to other occasions without improvement, they are all within the protection scope of the present invention.

Claims

1. An automated monitoring, analysis and early warning management system for bad geological foundation pits, characterized in that: It includes geotechnical database module, on-site soil image acquisition robot, soil model module, foundation pit design model module, intelligent monitoring module, data operation center, call and display module, human-computer interaction interface, early warning and alarm system module, and consulting and decision-making module; The geotechnical database module collects soil particle vector diagrams, geotechnical classifications, hydrogeological parameters, physical parameters and mechanical parameters to form a conventional geotechnical database. Through artificially supervised target learning, training forms a packaged module with data storage, recognition and vectorization functions, namely a digital geotechnical database. The on-site soil image acquisition robot forms points with precise elevation and coordinates by checking the coordinates with the on-site control points. It forms a point cloud stereo model with elevation and soil image superposition information by taking three-dimensional oblique photography of the excavation surface. It performs image recognition by calling the digital geotechnical database and outputs a soil layer parameter information module with soil layer top elevation, hydrological conditions, soil physical and mechanical properties parameter information. The soil model module automatically fits the top and bottom elevations of each soil layer in the point information into a virtual surface based on the drilling sampling information in the geological survey report. The two virtual surfaces are separated by three-dimensional soil layers. The soil layer models are constructed from top to bottom to form an initialized soil model, which is used to generate an initialized foundation pit design model and compare the survey values ​​with the actual soil layer parameter changes. The foundation pit design model module performs finite element simulation on the construction conditions of the initial soil model generated by the geological survey report according to the foundation pit support design plan, earthwork excavation conditions, and precipitation plan to form a foundation pit design initialization model. The calculation results are used by the intelligent monitoring module to lay out the monitoring network and determine the monitoring method, expected monitoring point values, and alarm values; The data operation center is connected with the soil layer parameter information generation module, soil model module, and foundation pit design model module for soil physical and mechanical properties and seepage stability analysis, foundation pit support structure deformation and strength analysis, and soil excavation and foundation pit dewatering on surrounding buildings, pipelines, and roads. It is also used for monitoring point simulation values, monitoring point measured values ​​and simulation values, and early warning value comparative relationship analysis; the data operation center is connected with the call and display module and the consulting decision module for manual interaction to form subsequent working condition monitoring values ​​and trend predictions, correction plans, error analysis and risk assessment, and emergency response plans; The calling and display module is mainly composed of the chart data I / O display module, which is connected to the data operation center and is used to call all the data in the data operation center, and upload, download, store, and access models and monitoring data according to user customization; the calling and display module sets access permissions; the calling and display module automatically generates various pictures, tables, documents, and reports for manual processing, review, and archiving; through the connection with the human-computer interaction interface, it displays the virtual construction conditions of the project, the actual construction conditions of the project, various tables and data processing diagrams, various parameters and early warning conditions; The early warning and alarm system module is used to send out alarm signals when soil parameters change significantly, monitoring data and simulation prediction data exceed alarm values, or monitoring data exceeds the warning values ​​set by the project, and transmit them to the human-computer interaction interface. The consulting decision-making module calls the data computing center to form optimization plans and emergency plans, and transmits the information through the local area network to the preset management personnel's mobile phone, triggering the project emergency response procedure.

2. The automatic monitoring, analysis and early warning management system for unfavorable geological foundation pits according to claim 1 is characterized in that: The data calculation center presets an exit interface reminder after every 10,000 calculations of a single simulation, which lasts for 1 minute. You can choose to exit the calculation. If you do not choose to exit, the calculation will continue. After each calculation, a rough result of the data trend will be obtained, and reasonable suggestions will be given according to the database experience value for the subsequent adjustment of the manual interaction technical parameters.

3. A method for automated monitoring, analysis and early warning management of unfavorable geological foundation pits, using the automated monitoring, analysis and early warning management system of unfavorable geological foundation pits according to claim 2, characterized in that: The following steps are involved: S1. Geotechnical database module training to form a digital geotechnical database; S2, the on-site soil image acquisition robot collects information and calls the digital geotechnical database to generate soil layer parameter information; S3, soil model module generates soil model M according to geological survey report i , i=0, 1, 2, soil layer parameter information in S2 and soil model M i Information versus compliance; When there is a bad geological interlayer that has not been explored by geological survey, or the soil layer parameters do not match the soil quality within the range of more than two meters thick in the initialized soil model, or the physical and mechanical parameter threshold of the soil layer identification is greatly different from the value in the geological survey report and is not within the allowable error range, it is "No", triggering the early warning mechanism of the early warning and alarm system module, sending early warning and alarm information to the project supervisor through the local area network and transmitting it to the human-computer interaction interface. After manual confirmation and processing, the early warning and alarm information is generated in the soil model M. i Make adjustments on When other non-critical adjustments occur in the soil layer, that is, the top or bottom elevation of the soil layer fluctuates within the allowable error range, the answer is "yes". At this time, in the soil model M i After adjustment, the soil model M i+1 Replaces the previous version of soil model M i , forming an iteratively updated soil model M i+1 , participate in subsequent calculations; S4, foundation pit design model module according to the foundation pit support design, excavation and dewatering plan, and soil model M i Conduct finite element simulation to generate foundation pit design model K j , j = 1, 2…; S5. Through the relevant theoretical knowledge, specifications, standards, monitoring methods and costs of foundation pit monitoring, foundation pit design model K j and intelligent deployment of monitoring network and monitoring method using human-computer interaction interface; S6. Deploy sensors at monitoring points; S7. Monitoring points are networked and fed back to the foundation pit design model K j On the top, correct the monitoring point information and generate monitoring data; S8, foundation pit design model K j Corresponding monitoring point data prediction, and transmit the prediction data to the data operation center; S9. Check the consistency of the monitoring data with the predicted data. If the deviation is outside the allowable error range, it is "no", triggering the early warning mechanism of the early warning and alarm system module, sending early warning and alarm information to the project supervisor through the local area network and transmitting it to the human-computer interaction interface; if the deviation is within the allowable error range, it is "yes". At this time, check the consistency of the monitoring data with the early warning value. If the deviation is outside the allowable error range, it is "no", triggering the early warning mechanism of the early warning and alarm system module, sending early warning and alarm information to the project supervisor through the local area network and transmitting it to the human-computer interaction interface; if the deviation is within the allowable error range, it is "yes", and transmitting it to the data operation center; S10, the calling and display module and the consulting decision-making module call the data of the data operation center, and interact through the human-computer interaction interface to form subsequent working condition monitoring values ​​and trend predictions, correction plans, error analysis and risk assessment, and emergency response plans.

4. The method for automatic monitoring, analysis and early warning management of unfavorable geological foundation pits according to claim 3 is characterized in that: The training steps of the geotechnical database module are as follows: A1. Data acquisition: collect a large number of soil particle vector diagrams through the image acquisition system and perform preprocessing; A2, perform local feature analysis and cluster analysis on the data collected in A1 through a feature analyzer, and form a corresponding relationship between the image and the target; A3. Assign values ​​to the target, including category, physical and mechanical parameters; A4. Form a test version of geotechnical database; A5. Input new pictures for N tests and call the test version of the geotechnical database for identification; A6: Manually supervise the target learning and form the final digital geotechnical database, which will be updated and returned to A4.

5. The method for automatic monitoring, analysis and early warning management of unfavorable geological foundation pits according to claim 4 is characterized in that: In A3, after the assignment, the filtering SVM principle is applied for optimization, namely: , where ω is the hyperplane normal vector, b is the bias term, and x i is the eigenvector, y i is the class label, usually +1 or -1.

6. The method for automatic monitoring, analysis and early warning management of unfavorable geological foundation pits according to claim 3 is characterized in that: The steps for the intelligent monitoring module to lay out the monitoring network and determine the monitoring method are as follows: B1. Summarize and develop a monitoring database based on relevant theoretical knowledge, specifications, standards, monitoring methods and costs, basic components of foundation pit monitoring, operation procedures, precautions, calculation methods, risk assessment methods, and typical cases; B2. Through logical programming, the AdaBoost ensemble learning method is used to combine multiple weak learners to update the weights of correctly classified samples and realize automatic feature selection for multi-classification problems. The algorithm is expressed as follows: and , where: t is the learner number, n represents the total number of weak learners, , , Represent the weight, prediction function, and error rate of the t-th learner respectively; Call the foundation pit design model and monitoring database, identify similar typical engineering cases, and automatically deploy monitoring networks and monitoring methods based on the characteristic data points of the foundation pit design model in accordance with the location, method and quantity specified in the specification. After economic analysis, select the economically reasonable monitoring plan, and intelligently deploy the monitoring network and monitoring method V1.0; B3. Through the human-computer interaction interface, adjust the points and monitoring methods of the generated monitoring network, and intelligently deploy the monitoring network and monitoring method V2.0; B4. Deploy sensors at monitoring points; B5. After the monitoring network is arranged, the monitoring points are fed back to the foundation pit design model through the network, and after modifications are made to the model, the monitoring point information is corrected, and the monitoring network and monitoring method V3.0 are intelligently arranged. During this period, selective manual intervention can also be carried out through the human-computer interaction interface.

7. The method for automatic monitoring, analysis and early warning management of unfavorable geological foundation pits according to claim 3 is characterized in that: The working steps of the early warning and alarm system module are as follows: C1. By default, the initialized soil model generated by the geological survey report is used as the true value of the soil layer parameters; Manually set soil layer parameter values ​​and error tolerances; Compare the on-site soil quality with the model information. If the difference is greater than the allowable error value, a Class A warning is triggered. The monitoring data is compared with the emergency alarm value. If it is greater than the allowable error value, a first-level warning is triggered and transmitted to the human-computer interaction interface; C2. Foundation pit design model K j The corresponding monitoring point simulation value is used as the monitoring data target value; Manually set monitoring data target values ​​and error tolerances; The monitoring data is compared with the predicted data. If the difference is greater than the allowable error value, a Class B warning is triggered; The monitoring data is compared with the emergency alarm value. If it is greater than the allowable error value, a first-level warning is triggered and transmitted to the human-computer interaction interface; C3. The specification stipulates the foundation pit warning value corresponding to the foundation pit grade as the monitoring data warning limit; Manually set monitoring data warning limits and error tolerances; The monitoring data is compared with the warning value. If it is greater than the allowable error value, a Class C warning is triggered; The monitoring data is compared with the emergency alarm value. If it is greater than the allowable error value, a first-level warning is triggered and transmitted to the human-computer interaction interface; C4, corresponding to 1.25 times of the foundation pit warning value, as the emergency response alarm threshold; Manually set emergency response alarm thresholds and error tolerances; The monitoring data is compared with the emergency alarm value. If it is greater than the allowable error value, a first-level warning is triggered and transmitted to the human-computer interaction interface; C5. After the information in the above steps C1, C2, C3 and C4 is transmitted to the human-computer interaction interface, the control room personnel are reminded to pay special attention, and the data operation center is called through the consulting decision-making module to form optimization plans and emergency plans. The information is sent to the project supervisor through the local area network to send early warning and alarm information, triggering the project emergency response procedure.

Images