Deep foundation pit monitoring system and monitoring method

Abstract

The invention relates to the technical field of deep foundation pit monitoring, in particular to a deep foundation pit monitoring system and method. According to the deep foundation pit monitoring method, the condition information of the deep foundation pit is reflected more comprehensively by collecting multi-directional monitoring factor data, and the comprehensiveness of the data is improved; unified management and analysis are performed on the data based on the cloud monitoring terminal, so that the practicability is improved; the constructed deep foundation pit monitoring index body generation model and the constructed monitoring factor monitoring optimization model can generate a specific deep foundation pit monitoring scheme according to specific conditions and monitoring factor data of the deep foundation pit, and effective monitoring of the deep foundation pit is achieved; through abnormal identification of the monitoring factors, the human cost investment is reduced, and the accuracy of common identification of the monitoring factors is improved; the constructed deep foundation pit deformation rule map can generate a deep foundation pit monitoring scheme according to a deep foundation pit monitoring index body and a monitoring factor abnormity identification result, and the functionality of the deep foundation pit monitoring scheme is improved.

Description

Technical Field

[0001] The invention relates to the technical field of deep foundation pit monitoring, and in particular to a deep foundation pit monitoring system and a monitoring method. Background Art

[0002] With the acceleration of urbanization, the scale of construction projects is getting larger and larger, and the forms of construction are becoming more and more diverse, which requires the safety and quality of construction projects to be guaranteed. In this context, the quality assurance of deep foundation pit projects has received increasing attention. Deep foundation pit projects have the characteristics of large area, deep depth, complex geological conditions, and sensitive surrounding environment. The stability and safety issues in their construction process are becoming increasingly prominent. In order to ensure the safe construction of deep foundation pit projects, reduce quality accidents and potential accidents, and ensure the safety of surrounding buildings, underground pipelines and roads, deep foundation pit monitoring technology came into being. This technology uses scientific and systematic monitoring methods to grasp various data during the foundation pit excavation process, provide timely and accurate feedback information for construction, thereby optimizing the design and construction plan and ensuring the smooth progress of the foundation pit project.

[0003] For example, patent number CN116336994A discloses a deep foundation pit monitoring method and system with early warning, which relates to the field of intelligent monitoring technology. The method includes: based on the deep foundation pit structural information collected by the image acquisition device, deploying the automatic monitoring and acquisition instrument, monitoring the deep foundation pit, establishing a three-dimensional simulated deep foundation pit based on the internal shape of the deep foundation pit, the external shape of the deep foundation pit and the deep foundation pit structural information, evaluating the state of the deep foundation pit, obtaining multiple deep foundation pit evaluation results, inputting the foundation pit collapse model to perform collapse analysis, outputting the disaster analysis results, providing early warning for potential disasters in the deep foundation pit, and emergency alarm for sudden events in the deep foundation pit, thereby solving the technical problem that the existing technology lacks control over deep foundation pit monitoring, which ultimately makes it impossible to provide early warning for deep foundation pit collapse, and realizes the rational and precise control over deep foundation pit monitoring, thereby achieving the technical effect of early warning for deep foundation pit collapse.

[0004] However, the system has the following problems: 1. It is unable to quickly and accurately determine the monitoring points and monitoring frequencies of various monitoring types in the target deep foundation pit project; 2. It has low practicality and lacks different monitoring scheme recommendations for different deep foundation pit deformation conditions, making it difficult to accurately predict deep foundation pit monitoring parameters.

[0005] In view of this, the present application proposes a deep foundation pit monitoring system and a monitoring method to solve the above problems. Summary of the invention

[0006] The purpose of the present invention is to solve at least one of the technical problems existing in the prior art and to provide a deep foundation pit monitoring system and a monitoring method.

[0007] To achieve the above object, the technical solution adopted by the present invention is as follows: a deep foundation pit monitoring method, including a monitoring method of a cloud monitoring terminal, the steps of which include the following: S1. Use neural network algorithm to construct deep foundation pit index generation model, deep foundation pit deformation law map, monitoring factor anomaly recognition model and deep foundation pit monitoring optimization model; S2. Acquire monitoring factor data of deep foundation pit, and generate deep foundation pit index body using deep foundation pit index body generation model according to the monitoring factor data to obtain deep foundation pit index body; the monitoring factor data of deep foundation pit includes vertical displacement and horizontal displacement of slope top, groundwater level, inclinometer, internal force of anchor cable, vertical displacement of surrounding tunnels, surrounding surface settlement, and deformation of surrounding pipelines; S3. Based on the monitoring factor data and the deep foundation pit indicators, the monitoring factor anomaly identification model is used to identify the monitoring factor anomaly and obtain the monitoring factor anomaly identification result; S4. According to the abnormal identification results of the monitoring factors and the deep foundation pit indicators, the deep foundation pit monitoring optimization model is used to perform monitoring optimization of the monitoring factors to obtain a monitoring optimization strategy for the monitoring factors, wherein the monitoring optimization strategy for the monitoring factors includes the locations of several points to be monitored and the frequencies to be monitored; S5. Using the deep foundation pit deformation law atlas, the monitoring factor monitoring optimization strategy is adjusted to obtain the adjusted monitoring factor monitoring optimization strategy; S6. Based on the abnormal identification results of deep foundation pit indicators and monitoring factors, a deep foundation pit monitoring plan is generated using the deep foundation pit deformation law map to obtain deep foundation pit monitoring suggestions; S7, sending the abnormal identification results of monitoring factors, the monitoring optimization strategy of the adjusted monitoring factors and the deep foundation pit monitoring suggestions to the terminal; The deep foundation pit indicator generation model is constructed based on the N-GAN-Attention-LSTM algorithm; the monitoring factor monitoring anomaly recognition model is constructed based on the RNN-GNN-MLP algorithm; and the monitoring factor monitoring optimization model is constructed based on the DQN algorithm.

[0008] Further, step S1 includes collecting a number of deep foundation pit monitoring factors and deep foundation pit monitoring scheme knowledge, and a number of historical monitoring factor monitoring data, and performing preprocessing to obtain a number of preprocessed deep foundation pit monitoring factors and deep foundation pit monitoring scheme knowledge, and a number of preprocessed historical monitoring factor monitoring data; Performing data dimension reduction on several pre-processed historical monitoring factor monitoring data to obtain several dimension-reduced historical monitoring factor monitoring data and a set of key monitoring factor indicators; Based on the historical monitoring data of several reduced-dimensional monitoring factors, a deep foundation pit index generation model is constructed using a neural network algorithm, and several deep foundation pit indexes are generated; Based on several pre-processed deep foundation pit monitoring factors and deep foundation pit monitoring scheme knowledge, a deep foundation pit deformation law map is constructed using language processing algorithms; Based on several dimensionality-reduced historical monitoring factor monitoring data and several deep foundation pit indicators, a neural network algorithm is used to build a monitoring factor anomaly recognition model and generate several historical monitoring factor anomaly recognition results; Based on several historical monitoring factor anomaly identification results and several deep foundation pit indicators, a monitoring factor optimization model is constructed using a reinforcement learning algorithm.

[0009] Furthermore, based on several pre-processed deep foundation pit monitoring factors and deep foundation pit deformation knowledge, a deep foundation pit deformation law map is constructed using a language processing algorithm, which includes: Using a pre-trained named entity extraction model, extracting several monitoring factor monitoring named entities and several monitoring scheme named entities of the pre-processed deep foundation pit monitoring factors and deep foundation pit monitoring scheme knowledge; Using a pre-trained entity relationship extraction model, extracting monitoring factor monitoring entity relationships between a number of monitoring factor monitoring named entities and monitoring scheme entity relationships between a number of monitoring scheme named entities; According to several monitoring factor monitoring named entities, several monitoring factor monitoring entity relationships, several monitoring scheme named entities and several monitoring scheme entity relationships, a deep foundation pit deformation law atlas is constructed.

[0010] Further, step S2 includes: performing data dimensionality reduction on the acquired monitoring factor data according to the key monitoring factor indicator set to obtain the reduced-dimensional monitoring factor data; Input the reduced-dimensional monitoring factor data into the deep foundation pit index generation model to extract N real-time key dimension features; According to the preset attention weights, N key dimension features are weighted concatenated to obtain weighted concatenated features; According to the weighted splicing features, monitoring factor labels are predicted to obtain a number of monitoring factor labels, and based on the number of monitoring factor labels, deep foundation pit indicator bodies are generated to obtain deep foundation pit indicator bodies.

[0011] Further, step S3 includes: inputting the reduced-dimensional monitoring factor data and the deep foundation pit indicator body into the monitoring factor anomaly recognition model; Extract the data features of the monitoring factor data after dimensionality reduction and the features of the deep foundation pit index body; The data features of the monitoring factor data after dimensionality reduction and the features of the deep foundation pit indicator body are fused to obtain real-time fusion features; According to the real-time fusion features, the abnormal monitoring factors are identified and the abnormal monitoring factor prediction labels are obtained, that is, the abnormal monitoring factor identification results.

[0012] Further, step S4 includes: extracting a number of matched historical monitoring factor monitoring optimization experiences in the experience playback pool of the monitoring factor monitoring optimization model according to the deep foundation pit indicator; According to several historical monitoring factor monitoring optimization experiences and real-time monitoring factor abnormality identification results, the model parameters of the monitoring factor monitoring optimization model are updated to obtain updated model parameters; According to the updated model parameters, the monitoring factor monitoring optimization model is used to perform monitoring factor monitoring optimization, and a number of execution actions are obtained. Based on the number of execution actions, a monitoring factor monitoring optimization strategy is obtained.

[0013] Further, step S5 includes: searching and matching the action named entities corresponding to the execution actions in the monitoring factor monitoring optimization strategy in the monitoring factor monitoring named entities of the deep foundation pit deformation law map to obtain a number of matching monitoring factor monitoring named entities; According to a number of matching monitoring factor monitoring named entities and the monitoring factor monitoring entity relationship between the matching monitoring factor monitoring named entities and other monitoring factor monitoring named entities, the monitoring factor monitoring optimization strategy is adjusted to obtain an adjusted monitoring factor monitoring optimization strategy.

[0014] Further, step S6 includes: searching and matching the monitoring scheme named entities in the deep foundation pit deformation law map according to the monitoring factor named entities corresponding to the real-time monitoring factor labels in the deep foundation pit indicator body and the abnormal named entities corresponding to the real-time monitoring factor abnormal prediction labels in the monitoring factor abnormal identification results, to obtain a number of matching monitoring scheme named entities; According to a number of matching monitoring scheme named entities and the monitoring scheme entity relationship between the matching monitoring scheme named entity and other monitoring scheme named entities, a deep foundation pit monitoring scheme is generated to obtain a deep foundation pit monitoring suggestion.

[0015] A deep foundation pit monitoring method, including a terminal monitoring method, the steps of which include the following: S100, receiving the deep foundation pit monitoring factor data collected by the deep foundation pit monitoring instrument, and uploading it to the cloud monitoring terminal; S200, receiving the abnormal identification results of monitoring factors, the adjusted monitoring optimization strategy of monitoring factors, and the deep foundation pit monitoring suggestions from the cloud monitoring terminal, and generating alarm information according to the abnormal identification results of monitoring factors, and visualizing the alarm information, the abnormal identification results of monitoring factors, and the deep foundation pit monitoring suggestions; S300, generating a monitoring instruction based on the monitoring optimization strategy of the adjusted monitoring factors, and sending the monitoring instruction to the deep foundation pit monitoring instrument.

[0016] A deep foundation pit monitoring system comprises a cloud monitoring end, a terminal and a plurality of deep foundation pit monitoring instruments, wherein the cloud monitoring end is respectively connected with terminal signals, and the terminal is respectively connected with deep foundation pit monitoring instrument signals; the cloud monitoring end comprises a model building unit, a deep foundation pit indicator body generating unit, a monitoring factor abnormality identifying unit, a monitoring factor monitoring optimization unit, an optimization strategy adjusting unit, a monitoring suggestion and alarm generating unit and a data sending unit which are connected in sequence.

[0017] Compared with the prior art, the technical solution of this application has the following beneficial effects: 1. This application improves the comprehensiveness of data by collecting multi-dimensional monitoring factor data to more comprehensively reflect the deep foundation pit status information; it improves practicality by unified management and analysis of data based on the cloud monitoring terminal; the constructed deep foundation pit monitoring indicator generation model and monitoring factor monitoring optimization model can generate specific deep foundation pit monitoring plans according to the specific conditions of the deep foundation pit and monitoring factor data, thereby realizing effective monitoring of the deep foundation pit; by identifying abnormal monitoring factors, it reduces the labor cost investment and improves the accuracy of normal identification of monitoring factors; the constructed deep foundation pit deformation law map can generate a deep foundation pit monitoring plan according to the deep foundation pit monitoring indicator and monitoring factor abnormal identification results, thereby improving its functionality.

[0018] 2. The deep foundation pit monitoring method of this application can not only monitor the state of the deep foundation pit, but also predict possible engineering problems, issue early warnings, and avoid accidents. In addition, the deep foundation pit monitoring method can also improve the accuracy and efficiency of construction, that is, it can quickly and accurately determine the monitoring points and monitoring frequencies of various monitoring types in the target deep foundation pit project, and provide strong technical support for the development and utilization of underground space. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a structural schematic diagram of a deep foundation pit monitoring method of a cloud monitoring terminal in a preferred embodiment of the present invention; Figure 2 It is a structural schematic diagram of a deep foundation pit monitoring method of a terminal in a preferred embodiment of the present invention; Figure 3 It is a schematic structural diagram of a deep foundation pit monitoring system in a preferred embodiment of the present invention.

[0020] Figure numerals: 1. cloud monitoring terminal; 2. terminal; 3. deep foundation pit monitoring instrument. DETAILED DESCRIPTION

[0021] The technical solutions in the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0022] Reference Figure 1As shown, a preferred embodiment of the present invention is a deep foundation pit monitoring method, including a monitoring method of a cloud monitoring terminal 1, and the steps include the following: S1. Use neural network algorithm to construct deep foundation pit index generation model, deep foundation pit deformation law map, monitoring factor anomaly recognition model and deep foundation pit monitoring optimization model, which includes; S101, collecting a number of deep foundation pit monitoring factors and deep foundation pit monitoring scheme knowledge, and a number of historical monitoring factor monitoring data, and performing preprocessing to obtain a number of preprocessed deep foundation pit monitoring factors and deep foundation pit monitoring scheme knowledge, and a number of preprocessed historical monitoring factor monitoring data; The monitoring factor data of the deep foundation pit include vertical displacement and horizontal displacement of the slope top, groundwater level, inclinometer, internal force of anchor cables, vertical displacement of surrounding tunnels, surrounding surface settlement, deformation of surrounding pipelines, etc.; The deep foundation pit monitoring scheme knowledge includes deep foundation pit knowledge related to monitoring factors and deep foundation pit monitoring scheme related deep foundation pit knowledge, wherein the deep foundation pit knowledge related to monitoring factors includes monitoring factor definitions, abnormal deep foundation pit deformations caused by monitoring factors, monitoring factor monitoring rules, etc., and the deep foundation pit monitoring scheme related deep foundation pit knowledge includes abnormal monitoring factor ranges, deep foundation pit monitoring scheme rules, etc.; Preprocessing includes noise removal, data screening, data cleaning and normalization of the collected data in order to eliminate noise interference, remove duplicate and erroneous data, and eliminate data magnitude differences to improve data quality; S102, performing data dimensionality reduction on a number of pre-processed historical monitoring factor monitoring data to obtain a number of dimensionality-reduced historical monitoring factor monitoring data and a set of key monitoring factor indicators; The principal component analysis (PCA) method is used to perform matrix transformation on several pre-processed historical monitoring factor monitoring data to obtain the corresponding historical monitoring factor monitoring data matrix X = [x1, x2, ... x p ,...,x n ] T , where x p is the row vector of the p-th pre-processed historical monitoring factor monitoring data, p is the row vector indicator, and n is the total number of pre-processed historical monitoring factor monitoring data; T is the matrix transposition symbol; the initial row vector of the historical monitoring factor monitoring matrix is ​​the pre-processed historical monitoring factor monitoring data, and the initial column vector of the historical monitoring factor data matrix is ​​the historical monitoring factor index data, and the historical monitoring factor index data includes monitoring factor indicators including the vertical displacement and horizontal displacement of the slope top, groundwater level, inclinometer, anchor cable internal force, surrounding tunnel vertical displacement, surrounding surface settlement, surrounding pipeline deformation, etc., and each index data corresponds to the data value of each column in the historical monitoring factor data; The historical monitoring factor data matrix is ​​standardized to obtain the corresponding standardized historical monitoring factor data matrix; the formula is: X´= (X-μ) / σ, where X´ is the standardized historical monitoring factor data matrix; X is the processed historical monitoring factor data matrix; μ is the mean of the historical monitoring factor data matrix; σ is the variance of the historical monitoring factor data matrix; Obtain the covariance matrix of the standardized historical monitoring factor matrix, and obtain the corresponding candidate principal component matrix based on the standardized historical monitoring factor matrix and the covariance matrix; the candidate row vector of the candidate principal component matrix is ​​the pre-processed historical monitoring factor data, and the candidate column vector of the candidate principal component matrix is ​​the historical monitoring factor candidate indicator data; the formula is: D=n -1 YY T =P(n -1 X´X´ T ) T =E T (n -1 X´X´ T )E, where D is the covariance matrix of the historical monitoring factor data matrix after standardization; Y is the candidate principal component matrix; P is the conversion matrix; E is the unit eigenvector matrix; n is the total number of historical monitoring factors after preprocessing; Y=PX´, where Y is the candidate principal component matrix; P is the conversion matrix; X´ is the historical monitoring factor data matrix after standardization; The candidate principal component matrix Y = [y1, y2, ...y m ,...,y M ] Several candidate column vectors y' of the top 90% of the cumulative contribution rate of variance l As the corresponding principal component column vectors, the reduced-dimensional historical monitoring factor data matrix Y'=[y'1, y'2, ...y' m ,...,y K ]; the key row vector of the historical monitoring factor data matrix after dimensionality reduction is the historical monitoring factor data after dimensionality reduction, and the data index corresponding to the key column vector of the historical monitoring factor data matrix after dimensionality reduction is the key monitoring factor index, and the key monitoring factor index set is obtained; The formula is:

[0023] In the formula, G is the cumulative contribution rate of variance; m is the mth candidate principal component y m The variance of ; m is the indicator of the candidate principal component; M is the total number of candidate principal components; K is the total number of principal components; Perform inverse matrix transformation on the data matrix of historical monitoring factors after dimension reduction to obtain several historical postoperative monitoring factors after dimension reduction; S103, constructing a deep foundation pit indicator generation model based on a number of historical monitoring factor monitoring data after dimensionality reduction using a neural network algorithm, and generating a number of deep foundation pit indicators; The deep foundation pit indicator generation model includes N key dimension feature extraction modules based on the Generative Adversarial Networks (GAN) algorithm, an attention weight generation module based on the Attention mechanism, and a deep foundation pit indicator generation model based on the Long Short-Term Memory (LSTM) network, where N is the total number of key dimensions, including the deep foundation pit grade dimension, the deep foundation pit deformation difficulty dimension, the monitoring factor risk dimension, and the monitoring status dimension. The key dimension feature extraction module includes a generator and a discriminator. The goal of the generator is to generate sufficiently real data, while the goal of the discriminator is to distinguish between real data and data generated by the generator. In the key dimension feature extraction, GAN can be used to generate data that matches the real data features to achieve feature extraction of key dimensions. The attention weight generation module uses the preset attention weights to achieve splicing of key dimension features and integrate scattered features. The LSTM network is a special recurrent neural network. The special cell state design makes it perform well in processing and predicting time series data. It can capture long-term dependencies and is suitable for data with time correlation. It is used to predict spliced ​​data features. S104, constructing a deep foundation pit deformation law map using a language processing algorithm based on a number of pre-processed deep foundation pit monitoring factors and deep foundation pit monitoring scheme knowledge, which includes: S1041, using a pre-trained named entity extraction model, extracting a number of monitoring factor monitoring named entities and a number of monitoring scheme named entities of the pre-processed deep foundation pit monitoring factors and deep foundation pit monitoring scheme knowledge; S1042, using a pre-trained entity relationship extraction model to extract monitoring factor monitoring entity relationships between a number of monitoring factor monitoring named entities and monitoring scheme entity relationships between a number of monitoring scheme named entities; S1043, constructing a deformation law atlas of a deep foundation pit according to a number of monitoring factor monitoring named entities, a number of monitoring factor monitoring entity relationships, a number of monitoring scheme named entities, and a number of monitoring scheme entity relationships; S105, constructing a monitoring factor abnormality recognition model using a neural network algorithm based on a number of historical monitoring factor monitoring data after dimensionality reduction and a number of deep foundation pit indicators, and generating a number of historical monitoring factor abnormality recognition results; The monitoring factor anomaly recognition model includes a data feature extraction module based on the Recurrent Neural Network (RNN) algorithm, a graph feature extraction module based on the Graph Neural Network (GNN) algorithm, and a fusion prediction module based on the Multi-Layer Perceptron (MLP) algorithm. The RNN network is a special neural network architecture that allows information to flow in time series data and can process and predict sequence data. It is used to extract monitoring factor data features. The GNN network is a neural network specifically used to process graph data, where the graph consists of nodes and edges. It is used to extract graph features of deep foundation pit indicators, including node information of deep foundation pit indicator labels and edge information between deep foundation pit indicator labels. The MLP network is a fully connected layer, which is characterized by all neurons between layers being interconnected. It is used to fuse different features to obtain corresponding fusion features, improve data representation capabilities, and perform feature fusion. Based on the fusion features, abnormal monitoring factor identification is performed. S106. Based on several historical monitoring factor abnormality identification results and several deep foundation pit indicators, a monitoring factor monitoring optimization model is constructed using a reinforcement learning algorithm; S2. Acquire monitoring factor data of deep foundation pit, and generate deep foundation pit index body using deep foundation pit index body generation model according to the monitoring factor data to obtain deep foundation pit index body, which includes: S201, performing data dimensionality reduction on the acquired monitoring factor data according to the key monitoring factor indicator set to obtain the reduced-dimensional monitoring factor data; S202, inputting the dimensionality-reduced monitoring factor data into a deep foundation pit index generation model to extract N real-time key dimensional features; S203, performing weighted concatenation on N key dimension features according to preset attention weights to obtain weighted concatenated features; S204, predicting monitoring factor labels according to the weighted splicing features to obtain a number of monitoring factor labels, and generating a deep foundation pit indicator body according to the number of monitoring factor labels to obtain a deep foundation pit indicator body; S3. Based on the monitoring factor data and deep foundation pit indicators, the monitoring factor anomaly identification model is used to identify the monitoring factor anomaly, and the monitoring factor anomaly identification results are obtained, which include: S301, inputting the reduced-dimensional monitoring factor data and deep foundation pit index into a monitoring factor anomaly recognition model; S302, extracting data features of the monitoring factor data after dimensionality reduction and deep foundation pit indicator body features; S303, performing feature fusion on the data features of the monitoring factor data after dimensionality reduction and the deep foundation pit indicator body features to obtain real-time fusion features; S304: According to the real-time fusion features, abnormal monitoring factor identification is performed to obtain abnormal monitoring factor prediction labels, that is, abnormal monitoring factor identification results, which include: S4. According to the abnormal identification results of the monitoring factors and the deep foundation pit indicators, the deep foundation pit monitoring optimization model is used to perform monitoring optimization of the monitoring factors to obtain a monitoring optimization strategy for the monitoring factors, wherein the monitoring optimization strategy for the monitoring factors includes the locations of several points to be monitored and the frequencies to be monitored; S401. Extracting several matched historical monitoring factor monitoring optimization experiences from the experience playback pool of the monitoring factor monitoring optimization model according to the deep foundation pit indicator body; S402, updating the model parameters of the monitoring factor monitoring optimization model according to several historical monitoring factor monitoring optimization experiences and real-time monitoring factor abnormality identification results to obtain updated model parameters; S403, according to the updated model parameters, using the monitoring factor monitoring optimization model, performing monitoring factor monitoring optimization, obtaining a number of execution actions, and obtaining a monitoring factor monitoring optimization strategy according to the number of execution actions; S5. Using the deep foundation pit deformation law atlas, the monitoring factor monitoring optimization strategy is adjusted to obtain the adjusted monitoring factor monitoring optimization strategy, which includes: S501, searching and matching the action named entities corresponding to the execution actions in the monitoring factor monitoring optimization strategy in the monitoring factor monitoring named entities of the deep foundation pit deformation law map to obtain a number of matching monitoring factor monitoring named entities; S502, adjusting the monitoring factor monitoring optimization strategy according to a number of matching monitoring factor monitoring named entities and the monitoring factor monitoring entity relationship between the matching monitoring factor monitoring named entity and other monitoring factor monitoring named entities, to obtain an adjusted monitoring factor monitoring optimization strategy; S6. Based on the abnormal identification results of deep foundation pit indicators and monitoring factors, a deep foundation pit monitoring plan is generated using the deep foundation pit deformation law map to obtain deep foundation pit monitoring suggestions, which include: S601, searching and matching the monitoring scheme named entities in the deep foundation pit deformation law atlas according to the monitoring factor named entities corresponding to the real-time monitoring factor labels in the deep foundation pit indicator body and the anomaly named entities corresponding to the real-time monitoring factor anomaly prediction labels in the monitoring factor anomaly recognition results, to obtain a number of matching monitoring scheme named entities; S602, generating a deep foundation pit monitoring plan according to a plurality of matching monitoring plan named entities and a monitoring plan entity relationship between the matching monitoring plan named entity and other monitoring plan named entities, and obtaining a deep foundation pit monitoring suggestion; S7. Send the abnormal identification results of monitoring factors, the adjusted monitoring optimization strategy of monitoring factors and the deep foundation pit monitoring suggestions to the terminal.

[0024] like Figure 2 As shown, a deep foundation pit monitoring method includes a monitoring method of terminal 2, and the steps include the following: S100, receiving the deep foundation pit monitoring factor data collected by the deep foundation pit monitoring instrument 3, and uploading it to the cloud monitoring terminal 1; S200, receiving the abnormal identification results of monitoring factors, the adjusted monitoring optimization strategy of monitoring factors, and the deep foundation pit monitoring suggestions from the cloud monitoring terminal 1, and generating alarm information according to the abnormal identification results of monitoring factors, and visualizing the alarm information, the abnormal identification results of monitoring factors, and the deep foundation pit monitoring suggestions; S300, generating a monitoring instruction according to the monitoring optimization strategy of the adjusted monitoring factors, and sending the monitoring instruction to the deep foundation pit monitoring instrument 3.

[0025] like Figure 3 As shown, a deep foundation pit monitoring system includes a cloud monitoring terminal 1, a terminal 2 and a plurality of deep foundation pit monitoring instruments 3, wherein the cloud monitoring terminal 1 is respectively connected to the terminal 2 for signal connection, and the terminal 2 is respectively connected to the deep foundation pit monitoring instrument 3 for signal connection; Terminal 2 is used to receive the deep foundation pit monitoring factor data collected by the deep foundation pit monitoring instrument and upload it to the cloud monitoring terminal; at the same time, it receives the monitoring factor abnormality identification results, the adjusted monitoring factor monitoring optimization strategy and the deep foundation pit monitoring suggestions from the cloud monitoring terminal, and generates alarm information according to the monitoring factor abnormality identification results, and visualizes the alarm information, the monitoring factor abnormality identification results and the deep foundation pit monitoring suggestions; generates monitoring instructions according to the adjusted monitoring factor monitoring optimization strategy, and sends the monitoring instructions to the deep foundation pit monitoring instrument; Deep foundation pit monitoring instrument 3 is used to collect deep foundation pit monitoring factor data and send it to the terminal; at the same time, it receives the real-time monitoring factor monitoring instructions sent by the terminal and executes the real-time monitoring factor monitoring instructions The cloud monitoring terminal 1 includes a model building unit, a deep foundation pit indicator generating unit, a monitoring factor abnormality identifying unit, a monitoring factor monitoring optimization unit, an optimization strategy adjusting unit, a monitoring suggestion and alarm generating unit and a data sending unit which are connected in sequence.

[0026] The present invention discloses a deep foundation pit monitoring system and a monitoring method. By collecting monitoring factor data and combining them with deep foundation pit monitoring indicators, the system can more comprehensively reflect the deep foundation pit status information, thereby improving the comprehensiveness of the data. The system can uniformly manage and analyze the data based on a cloud monitoring terminal, thereby improving the practicality. The system can construct a deep foundation pit monitoring indicator generation model and a monitoring factor monitoring optimization model, thereby generating a specific deep foundation pit monitoring plan according to the specific conditions of the deep foundation pit and the monitoring factor data, thereby realizing effective monitoring of the deep foundation pit. By identifying abnormalities of the monitoring factors, the system can reduce the investment in manpower costs and improve the accuracy of identifying normal monitoring factors. The system can construct a deep foundation pit deformation law atlas, thereby generating a deep foundation pit monitoring plan according to the abnormal identification results of the deep foundation pit monitoring indicators and the monitoring factors, thereby improving its functionality.

[0027] Under the premise that no conflict occurs, those skilled in the art may freely combine and superimpose the above-mentioned additional technical features.

[0028] It is to be understood that the present invention is described by some embodiments, and it is known to those skilled in the art that various changes or equivalent substitutions may be made to these features and embodiments without departing from the spirit and scope of the present invention. In addition, under the teachings of the present invention, these features and embodiments may be modified to adapt to specific circumstances and materials without departing from the spirit and scope of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the scope of protection of the present invention.

Claims

1. A deep foundation pit monitoring method, characterized in that: The monitoring method including the cloud monitoring terminal includes the following steps: S1. Use neural network algorithm to construct deep foundation pit index generation model, deep foundation pit deformation law map, monitoring factor anomaly recognition model and deep foundation pit monitoring optimization model; S2. Acquire monitoring factor data of deep foundation pit, and generate deep foundation pit index body using deep foundation pit index body generation model according to the monitoring factor data to obtain deep foundation pit index body; the monitoring factor data of deep foundation pit includes vertical displacement and horizontal displacement of slope top, groundwater level, inclinometer, internal force of anchor cable, vertical displacement of surrounding tunnels, surrounding surface settlement, and deformation of surrounding pipelines; S3. Based on the monitoring factor data and the deep foundation pit indicators, the monitoring factor anomaly identification model is used to identify the monitoring factor anomaly and obtain the monitoring factor anomaly identification result; S4. According to the abnormal identification results of the monitoring factors and the deep foundation pit indicators, the deep foundation pit monitoring optimization model is used to perform monitoring optimization of the monitoring factors to obtain a monitoring optimization strategy for the monitoring factors, wherein the monitoring optimization strategy for the monitoring factors includes the locations of several points to be monitored and the frequencies to be monitored; S5. Using the deep foundation pit deformation law atlas, the monitoring factor monitoring optimization strategy is adjusted to obtain the adjusted monitoring factor monitoring optimization strategy; S6. Based on the abnormal identification results of deep foundation pit indicators and monitoring factors, a deep foundation pit monitoring plan is generated using the deep foundation pit deformation law map to obtain deep foundation pit monitoring suggestions; S7, sending the abnormal identification results of monitoring factors, the monitoring optimization strategy of the adjusted monitoring factors and the deep foundation pit monitoring suggestions to the terminal; The deep foundation pit indicator generation model is constructed based on the N-GAN-Attention-LSTM algorithm; the monitoring factor monitoring anomaly recognition model is constructed based on the RNN-GNN-MLP algorithm; and the monitoring factor monitoring optimization model is constructed based on the DQN algorithm.

2. The deep foundation pit monitoring method according to claim 1, characterized in that: The step S1 includes collecting a number of deep foundation pit monitoring factors and deep foundation pit monitoring scheme knowledge, and a number of historical monitoring factor monitoring data, and performing preprocessing to obtain a number of preprocessed deep foundation pit monitoring factors and deep foundation pit monitoring scheme knowledge, and a number of preprocessed historical monitoring factor monitoring data; Performing data dimension reduction on several pre-processed historical monitoring factor monitoring data to obtain several dimension-reduced historical monitoring factor monitoring data and a set of key monitoring factor indicators; Based on the historical monitoring data of several reduced-dimensional monitoring factors, a deep foundation pit index generation model is constructed using a neural network algorithm, and several deep foundation pit indexes are generated; Based on several pre-processed deep foundation pit monitoring factors and deep foundation pit monitoring scheme knowledge, a deep foundation pit deformation law map is constructed using language processing algorithms; Based on several dimensionality-reduced historical monitoring factor monitoring data and several deep foundation pit indicators, a neural network algorithm is used to build a monitoring factor anomaly recognition model and generate several historical monitoring factor anomaly recognition results; Based on several historical monitoring factor anomaly identification results and several deep foundation pit indicators, a monitoring factor optimization model is constructed using a reinforcement learning algorithm.

3. The deep foundation pit monitoring method according to claim 2, characterized in that: According to several pre-processed deep foundation pit monitoring factors and deep foundation pit deformation knowledge, a deep foundation pit deformation law map is constructed using a language processing algorithm, which includes: Using a pre-trained named entity extraction model, extracting several monitoring factor monitoring named entities and several monitoring scheme named entities of the pre-processed deep foundation pit monitoring factors and deep foundation pit monitoring scheme knowledge; Using a pre-trained entity relationship extraction model, extracting monitoring factor monitoring entity relationships between a number of monitoring factor monitoring named entities and monitoring scheme entity relationships between a number of monitoring scheme named entities; According to several monitoring factor monitoring named entities, several monitoring factor monitoring entity relationships, several monitoring scheme named entities and several monitoring scheme entity relationships, a deep foundation pit deformation law atlas is constructed.

4. The deep foundation pit monitoring method according to claim 3, characterized in that: The step S2 comprises: performing data dimensionality reduction on the acquired monitoring factor data according to the key monitoring factor indicator set to obtain the reduced-dimensional monitoring factor data; Input the reduced-dimensional monitoring factor data into the deep foundation pit index generation model to extract N real-time key dimension features; According to the preset attention weights, N key dimension features are weighted concatenated to obtain weighted concatenated features; According to the weighted splicing features, monitoring factor labels are predicted to obtain a number of monitoring factor labels, and based on the number of monitoring factor labels, deep foundation pit indicator bodies are generated to obtain deep foundation pit indicator bodies.

5. The deep foundation pit monitoring method according to claim 4, characterized in that: The step S3 comprises: inputting the reduced-dimensional monitoring factor data and the deep foundation pit indicator into the monitoring factor abnormality recognition model; Extract the data features of the monitoring factor data after dimensionality reduction and the features of the deep foundation pit index body; The data features of the monitoring factor data after dimension reduction and the features of the deep foundation pit indicator body are fused to obtain real-time fusion features; According to the real-time fusion features, the monitoring factor anomaly identification is performed to obtain the monitoring factor anomaly prediction label, that is, the monitoring factor anomaly identification result.

6. The deep foundation pit monitoring method according to claim 5, characterized in that: The step S4 comprises: extracting a number of matched historical monitoring factor monitoring optimization experiences in the experience playback pool of the monitoring factor monitoring optimization model according to the deep foundation pit indicator; Based on several historical monitoring factor monitoring optimization experiences and real-time monitoring factor abnormality identification results, the model parameters of the monitoring factor monitoring optimization model are updated to obtain updated model parameters; According to the updated model parameters, the monitoring factor monitoring optimization model is used to perform monitoring factor monitoring optimization, and a number of execution actions are obtained. Based on the number of execution actions, a monitoring factor monitoring optimization strategy is obtained.

7. The deep foundation pit monitoring method according to claim 6, characterized in that: The step S5 comprises: searching and matching the action named entities corresponding to the execution actions in the monitoring factor monitoring optimization strategy in the monitoring factor monitoring named entities of the deep foundation pit deformation law map to obtain a plurality of matching monitoring factor monitoring named entities; According to a number of matching monitoring factor monitoring named entities and the monitoring factor monitoring entity relationship between the matching monitoring factor monitoring named entities and other monitoring factor monitoring named entities, the monitoring factor monitoring optimization strategy is adjusted to obtain an adjusted monitoring factor monitoring optimization strategy.

8. The deep foundation pit monitoring method according to claim 7, characterized in that: The step S6 comprises: searching and matching the monitoring scheme named entities in the deep foundation pit deformation law map according to the monitoring factor named entities corresponding to the real-time monitoring factor labels in the deep foundation pit indicator body and the abnormal named entities corresponding to the real-time monitoring factor abnormal prediction labels in the monitoring factor abnormal identification results, to obtain a plurality of matching monitoring scheme named entities; According to a number of matching monitoring scheme named entities and the monitoring scheme entity relationship between the matching monitoring scheme named entity and other monitoring scheme named entities, a deep foundation pit monitoring scheme is generated to obtain a deep foundation pit monitoring suggestion.

9. A deep foundation pit monitoring method, characterized in that: A monitoring method for a terminal includes the following steps: S100, receiving the deep foundation pit monitoring factor data collected by the deep foundation pit monitoring instrument, and uploading it to the cloud monitoring terminal; S200, receiving the abnormal identification results of monitoring factors, the adjusted monitoring optimization strategy of monitoring factors, and the deep foundation pit monitoring suggestions from the cloud monitoring terminal, and generating alarm information according to the abnormal identification results of monitoring factors, and visualizing the alarm information, the abnormal identification results of monitoring factors, and the deep foundation pit monitoring suggestions; S300, generating a monitoring instruction based on the monitoring optimization strategy of the adjusted monitoring factors, and sending the monitoring instruction to the deep foundation pit monitoring instrument.

10. A deep foundation pit monitoring system, characterized in that: It includes a cloud monitoring end, a terminal and several deep foundation pit monitoring instruments. The cloud monitoring end is connected to the terminal signal respectively, and the terminal is connected to the deep foundation pit monitoring instrument signal; the cloud monitoring end includes a model construction unit, a deep foundation pit indicator body generation unit, a monitoring factor abnormality identification unit, a monitoring factor monitoring optimization unit, an optimization strategy adjustment unit, a monitoring suggestion and alarm generation unit and a data sending unit which are connected in sequence.

Images