Deep foundation pit engineering risk monitoring and early warning method and system based on Internet of Things

Abstract

The invention belongs to the technical field of risk monitoring, and discloses a deep foundation pit engineering risk monitoring and early warning method and system based on the Internet of Things, and the method comprises the steps: constructing a foundation pit twinborn model, setting corresponding monitoring nodes based on the foundation pit twinborn model, carrying out the data collection of the monitoring nodes in a target deep foundation pit engineering, and obtaining a corresponding engineering monitoring data set; carrying out abnormal point identification and correction on the obtained engineering monitoring data set, and carrying out data synchronization on the corrected engineering monitoring data to obtain corresponding time sequence data; inputting the data into a pre-constructed parameter prediction model to obtain corresponding periodic prediction data; performing risk assessment on the target deep foundation pit project based on the obtained period prediction data to obtain a corresponding risk state, generating a corresponding risk early warning level based on the risk state, and performing early warning feedback; according to the method, the accuracy and efficiency of comprehensive monitoring and early warning of the deep foundation pit engineering risk are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of risk monitoring, and more specifically, to a risk monitoring and early warning method and system for deep foundation pit engineering based on the Internet of Things. Background Art

[0002] In the process of modern urban construction, with the increasing tension of land resources, the development and utilization of underground space is becoming more and more in-depth. As a key link in underground construction, the number and scale of deep foundation pit projects have shown a significant growth trend. However, due to the involvement of many complex factors such as geological conditions, surrounding environment, and construction technology, there are many risks in the construction and use of deep foundation pit projects, such as foundation pit collapse, settlement of surrounding buildings, rupture of underground pipelines, etc. These risks will not only delay the project progress, increase the construction cost, but also endanger the lives of personnel and cause serious social impacts.

[0003] Most traditional risk monitoring methods for deep foundation pit projects rely on manual regular measurements, which have problems such as low data collection frequency, large errors, and inability to obtain data in real time. Manual measurement can only be carried out at specific time points, making it difficult to capture the instantaneous abnormal changes of the foundation pit during construction. Once an unexpected situation occurs during the interval between two measurements, it is very difficult to detect it in time and take effective measures. At the same time, manual reading and recording of data are easily affected by subjective factors, resulting in doubts about the accuracy of the data.

[0004] Some projects have introduced some simple monitoring instruments, but these instruments usually have single functions and can only monitor one or a few parameters, unable to comprehensively reflect the overall situation of the foundation pit. Moreover, there is no effective data fusion and collaborative working mechanism between different types of monitoring instruments, and the monitoring data is scattered, making it difficult to conduct comprehensive analysis, resulting in one-sidedness in the assessment of foundation pit risks and unable to provide a comprehensive and reliable basis for engineering decision-making.

[0005] In terms of risk early warning, existing early warning systems often rely on simple threshold judgments and do not fully consider the dynamic changes during the construction process of deep foundation pit projects and the mutual relationships between various parameters. When the monitoring data approaches or exceeds the set threshold, there may be a large number of false alarms or missed alarms. In addition, after discovering risks, traditional early warning systems lack effective information transmission and feedback mechanisms, unable to convey early warning information to relevant personnel in time, nor can they provide specific guiding suggestions for subsequent risk handling, resulting in engineering managers being difficult to make accurate decisions quickly and delaying the best opportunity for risk handling.

[0006] In view of this, the present invention proposes a risk monitoring and early warning method and system for deep foundation pit engineering based on the Internet of Things to solve the above problems. Summary of the Invention

[0007] To overcome the above-mentioned defects of the prior art and to achieve the above object, the present invention provides the following technical solutions:

[0008] An Internet of Things-based risk monitoring and early warning method for deep foundation pit engineering, comprising:

[0009] Step 1: Construct a twin model of the foundation pit, and set corresponding monitoring nodes based on it. Collect data from the monitoring nodes in the target deep foundation pit project by using Internet of Things terminals with different functions pre-deployed, and obtain a corresponding set of engineering monitoring data;

[0010] Step 2: Identify abnormal points in the obtained set of engineering monitoring data, and correct the corresponding abnormal points. After the correction is completed, synchronize the engineering monitoring data corresponding to different Internet of Things terminals to obtain corresponding time series data; and input it into a pre-constructed parameter prediction model to obtain corresponding periodic prediction data;

[0011] Step 3: Based on the obtained periodic prediction data, and combined with the collected set of engineering monitoring data, conduct a risk assessment on the target deep foundation pit project to obtain a corresponding risk status, generate a corresponding risk early warning level based on the risk status, and conduct early warning feedback.

[0012] Further, the process of constructing a twin model of the foundation pit and setting corresponding monitoring nodes based on it includes:

[0013] Obtain the basic engineering information corresponding to the deep foundation pit project, and based on the basic engineering information, obtain the BIM data and 3DGIS data corresponding to the target deep foundation pit project;

[0014] Input the corresponding BIM data and 3DGIS data into the corresponding BIM professional software and GIS professional software respectively; obtain corresponding physical models and numerical models; superimpose the corresponding physical models and digital models to obtain corresponding BIM models and 3DGIS models and conduct model integration to obtain a corresponding twin foundation pit model;

[0015] Conduct numerical simulation based on the twin foundation pit model, obtain high-frequency risk areas in the corresponding deep foundation pit project at the current construction stage, and set a number of data monitoring points in the target high-frequency risk areas based on the pre-set monitoring distribution requirements.

[0016] Further, the process of collecting data from the monitoring nodes in the target deep foundation pit project to obtain a corresponding set of engineering monitoring data includes:

[0017] Set data collection nodes, and the data collection nodes are composed of a number of Internet of Things sensing terminals with different functions;

[0018] Deploy the data acquisition nodes into the set monitoring nodes, collect data at the corresponding monitoring nodes based on them, and upload the data to a pre-constructed monitoring terminal; the monitoring terminal aggregates the collected data according to a pre-set collection period and uploads the data to obtain a corresponding engineering monitoring data set; the engineering monitoring data set is composed of engineering monitoring data corresponding to a number of different engineering monitoring parameters.

[0019] Further, the process of preprocessing the obtained data set to obtain a corresponding time series data set includes:

[0020] Obtain the engineering monitoring data corresponding to the corresponding engineering monitoring parameters; construct a two-dimensional rectangular coordinate system with time as the abscissa and the engineering monitoring parameters as the ordinate, and map the obtained engineering monitoring parameters into the corresponding two-dimensional rectangular coordinate system to obtain a corresponding parameter change curve;

[0021] Obtain the parameter average value corresponding to the engineering monitoring parameters within the corresponding collection period based on the parameter change curve and respectively obtain the Mahalanobis distance between each data point in the corresponding parameter change curve and the corresponding parameter average value;

[0022] Set a distance threshold, and combine with a pre-set constraint condition to determine whether the corresponding data point is an abnormal point; if the corresponding Mahalanobis distance is less than the distance threshold and satisfies the constraint condition, no other operations are performed; if the Mahalanobis distance is not less than the distance threshold or does not satisfy the constraint condition, data correction is performed on the corresponding data point based on the linear interpolation algorithm, and the corrected data point replaces the original data point in the parameter change curve; if the Mahalanobis distance is not less than the distance threshold and does not satisfy the constraint condition, the corresponding data point is removed;

[0023] Based on the above data correction process, perform data correction on the engineering monitoring data corresponding to all the corresponding engineering monitoring parameters, and synchronize the corrected engineering monitoring data of different engineering monitoring parameters based on a pre-set reference time series to obtain the time series data corresponding to the corresponding engineering monitoring parameters and perform statistics to obtain a corresponding time series data set.

[0024] Further, the formula for defining the constraint condition is: ; where represents the curve slope from the data point to the data point ; represents the average value of the slope differences of adjacent data points to the data point ; represents the total number of data points in the parameter change curve; represents the data point +1 to the data point The slope of the curve of 2; m represents the index of the data points within the parameter change curve;

[0025] The mathematical formula for time synchronization is: ; where represents the parameter value of the data point after time synchronization at time, and respectively represent the parameter values corresponding to and times within the corrected reference change curve; and ; represents the index of the th time within the reference time series; and respectively represent the and rd times of the time series corresponding to the corrected parameter change curve, being the time index within the time series corresponding to the corrected parameter change curve.

[0026] Furthermore, the process of obtaining the corresponding period prediction data includes:

[0027] Construct a parameter prediction model, input the obtained time series data set into the constructed parameter prediction model, obtain the corresponding model data result, and based on it, obtain the period prediction data for the next acquisition period;

[0028] Among them, the construction process of the parameter prediction model includes:

[0029] The backbone network of the parameter prediction model is an improved multi-level neural network, and a two-stage attention mechanism is introduced; the basic framework of the improved multi-level neural network includes an input layer, a feature layer, and an output layer;

[0030] The input layer is used to receive the input time series data and perform data difference operation on it to obtain the corresponding input sequence data;

[0031] The feature layer consists of an encoding unit and a decoding unit;

[0032] The encoding unit assigns weights to the corresponding input sequence data based on the introduced attention mechanism and performs sequence update to obtain the corresponding updated sequence data , and encodes the updated sequence data based on the built-in Bi-LSTM, while capturing the forward and backward dependencies of the time series;

[0033] The decoding unit is used to receive the encoded updated sequence data, update the hidden layer state and the hidden layer state weight in the corresponding decoding unit based on the encoded updated sequence data, and perform weighted summation and decoding on the encoded updated sequence data received by the decoding unit based on the updated hidden layer state and the hidden layer state weight, and obtain the corresponding prediction result based on the weighted summation and decoding;

[0034] The input layer is used to receive the obtained prediction result and map it to the same data dimension as the original time series data;

[0035] Construct a training data set;

[0036] Construct a multi-layer neural network, and initialize the network parameters of the multi-layer neural network by using the WOA algorithm; after the initialization is completed, divide the training data set into several training batches, and input the corresponding parameter prediction model in sequence based on time for model training until the loss function converges in several consecutive training batches, then save the model parameters.

[0037] Further, the formula for weight assignment is:

[0038] ; where represents the weight of the input sequence data at time t1, , and both represent weight matrices; and respectively represent the hidden layer state and the cell state in the LSTM layer of the encoding unit at time; represents the operation process of the corresponding LSTM layer; represents the input sequence data; represents the hyperbolic tangent function; e is the matrix dimension of the weight matrix;

[0039] Obtain the corresponding prediction result ; and represent weight matrices; and represent bias terms; and represent the hidden layer state and the cell state at the decoding stage at time t1

[0040] The formula for updating the hidden layer state is: ; where represents the activation function operation; represents the state vector obtained by processing the hidden layer state and the cell state at time t1-1 through the activation function softmax;

[0041] The update formula for the hidden layer state weights is as follows:

[0042] ; 、 and represent the representation weight matrix in the decoding stage; and represent the hidden layer state and cell state at the decoding stage at time t1 - 1; represents the matrix dimension.

[0043] Furthermore, the process of obtaining the risk status of the target deep foundation pit project based on the obtained periodic prediction data and performing risk early warning includes:

[0044] Obtain the time series data corresponding to the current acquisition period and the adjacent historical acquisition periods, and combine the obtained periodic prediction data to obtain the parameter change rate and cumulative change amount corresponding to the corresponding factory monitoring parameters;

[0045] Construct an early warning risk framework, and construct a corresponding fuzzy membership function based on the construction stage where the current deep foundation pit is located; the early warning risk framework consists of three sub - intervals, and the sub - intervals include a normal interval, an abnormal interval, and a fault interval;

[0046] Obtain the parameter change rate and cumulative change amount corresponding to the engineering monitoring parameters, and combine them with the fuzzy membership function and the boundary points of each sub - interval to obtain the probability assignment values corresponding to the corresponding parameter change rate and cumulative change amount, and denote them as and ; g ∈ ; respectively represent the normal interval, the abnormal interval, and the fault interval; and respectively represent the probability assignment values of the parameter change rate and cumulative change amount for the sub - interval g;

[0047] Fuse the probability assignment values corresponding to the corresponding parameter change rate and cumulative change amount to obtain the corresponding fusion parameter , is the fusion coefficient, which is used to represent the probability assignment value that the engineering monitoring parameter belongs to the sub - interval ; represents the data fusion operation based on the D - S evidence algorithm;

[0048] Obtain the fusion coefficients corresponding to each engineering monitoring parameter, and perform data merging based on the improved D - S evidence algorithm to obtain the corresponding membership distribution value of the deep foundation pit project ;

[0049] Obtain the sub-interval corresponding to the maximum membership distribution value of the deep foundation pit project, and determine whether there is a risk for the corresponding deep foundation pit project based on it; if the sub-interval is a normal interval, no other operations are performed; if the sub-interval is an abnormal or faulty interval, mark the abnormal or faulty interval as the risk status of the corresponding deep foundation pit project; obtain the fusion coefficients corresponding to different engineering monitoring parameters, and obtain the parameter scores corresponding to the corresponding engineering monitoring parameters based on the preset scoring interval;

[0050] Obtain the parameter weights corresponding to each engineering monitoring data under different risk factors, and perform weighted summation on the corresponding parameter scores based on them to obtain the risk assessment scores corresponding to the corresponding risk factors, obtain the risk factor corresponding to the highest risk assessment score, and output it as the risk cause;

[0051] According to the deep foundation pit safety assessment requirements, the risk assessment scores correspond to different abnormal or faulty warning levels, and feedback them to the corresponding monitoring terminals. The monitoring terminals are based on the built-in alarm devices and perform different types of sound and light warnings in combination with the abnormal or faulty warning levels; at the same time, the monitoring terminals generate corresponding alarm information and perform warning feedback based on the pre-constructed WeChat public account.

[0052] Further, the process of obtaining the parameter weights corresponding to each engineering monitoring data under different risk factors includes:

[0053] Obtain the historical time series data set corresponding to the corresponding risk factor trigger, and respectively obtain the feature vectors corresponding to each engineering monitoring parameter based on the historical time series data set, and obtain the correlation degree between different engineering monitoring parameters based on them; construct the corresponding correlation matrix based on the correlation degree ; where, represents the engineering monitoring parameter and the engineering monitoring parameter The correlation degree between; and both represent the indexes of the engineering monitoring parameters under the corresponding risk factor, and ;

[0054] Based on the correlation matrix, obtain the initial weight corresponding to the corresponding engineering monitoring parameter , where, , E is an integer, representing the total number of categories of engineering monitoring parameters under the corresponding risk factor;

[0055] Perform standardization processing on the obtained correlation matrix to obtain the corresponding correction matrix ; where, ; is the matrix element of the correction matrix, used to represent the correlation degree after standardization processing;

[0056] Obtain corresponding weight correction coefficients based on the correction matrix ; where ; used to represent the parameter deviation of engineering monitoring parameters

[0057] And based on it, perform weight correction on the obtained initial weights to obtain corresponding corrected weights; the formula for weight correction is: ; represents the engineering monitoring parameter of the corrected weight

[0058] Perform weighted summation on the obtained corrected weights and initial weights to obtain corresponding combined weights, and use them as the parameter weights of engineering monitoring parameters under corresponding risk factors

[0059] Obtain all known types of risk factors, and based on the above process of obtaining parameter weights, assign weights to engineering monitoring parameters under different risk factors

[0060] The risk monitoring and early warning system for deep foundation pit engineering based on the Internet of Things includes:

[0061] A data acquisition module, used to construct a twin model of the foundation pit, and based on it, set corresponding monitoring nodes, and collect data from the monitoring nodes in the target deep foundation pit project by using pre-deployed Internet of Things terminals with different functions to obtain a corresponding engineering monitoring data set

[0062] A data processing module, used to identify abnormal points in the obtained engineering monitoring data set, and correct the corresponding abnormal points. After the correction is completed, synchronize the engineering monitoring data corresponding to different Internet of Things terminals to obtain corresponding time series data; and input it into a pre-constructed parameter prediction model to obtain corresponding periodic prediction data

[0063] A data analysis module, based on the obtained periodic prediction data, and combined with the collected engineering monitoring data set, conduct a risk assessment on the target deep foundation pit project to obtain a corresponding risk status

[0064] A data feedback module, and generate a corresponding risk warning level based on the risk status, and conduct warning feedback

[0065] The technical effects and advantages of the risk monitoring and early warning method and system for deep foundation pit engineering based on the Internet of Things in the present invention:

[0066] 1. Risk assessment is carried out based on periodic prediction data. By calculating the parameter change rate and cumulative change amount, combined with the fuzzy membership function and D-S evidence algorithm for data fusion and merging, the engineering membership coefficient is determined. According to the principle of the largest membership degree, the risk status is judged. For abnormal or faulty situations, parameter scoring and weighted summation are performed to obtain the risk assessment score, the risk cause is determined and output. At the same time, according to the risk warning level, acoustic and optical warnings, WeChat official account feedback, and color marking of the foundation pit twin model are carried out, which is convenient for the staff to timely grasp the engineering risk situation, take corresponding measures, and ensure the safety of the deep foundation pit project.

[0067] 2. The initial weight is determined by obtaining the correlation degree of the eigenvector comparison of the engineering monitoring data. Through standardization processing and entropy value calculation, the weight correction coefficient is obtained, and then the combined weight is obtained as the parameter weight. It can reasonably evaluate according to the importance difference of different engineering monitoring parameters, make the risk assessment result more in line with the actual situation, and improve the scientificity and reliability of the risk assessment. Description of the Drawings

[0068] Figure 1 It is a schematic diagram of the risk monitoring and warning method for deep foundation pit projects based on the Internet of Things of the present invention;

[0069] Figure 2 It is a schematic diagram of the risk monitoring and warning system for deep foundation pit projects based on the Internet of Things of the present invention. Detailed Embodiment

[0070] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative work shall fall within the protection scope of the present invention.

[0071] Embodiment 1 Please refer to Figure 1 As shown, the risk monitoring and warning method for deep foundation pit projects based on the Internet of Things described in this embodiment includes:

[0072] Step 1: Construct a foundation pit twin model, and set corresponding monitoring nodes based on it. Based on the pre-deployed Internet of Things terminals with different functions, data collection is carried out on the monitoring nodes in the target deep foundation pit project to obtain the corresponding engineering monitoring data set;

[0073] Step 2: Identify the abnormal points in the obtained engineering monitoring data set, and correct the corresponding abnormal points. After the correction is completed, synchronize the engineering monitoring data corresponding to different Internet of Things terminals to obtain the corresponding time series data; and input it into the pre-constructed parameter prediction model to obtain the corresponding periodic prediction data;

[0074] Step 3: Based on the obtained cycle prediction data and combined with the collected engineering monitoring data set, conduct a risk assessment on the target deep foundation pit project, obtain the corresponding risk status, generate a corresponding risk warning level based on the risk status, and give a warning feedback;

[0075] It should be further noted that in the specific implementation process, the process of constructing the foundation pit twin model and setting corresponding monitoring nodes based on it includes:

[0076] Obtain the foundation engineering information corresponding to the deep foundation pit project. The foundation engineering information includes relevant data such as the design drawings, geological exploration reports, and surrounding environment conditions corresponding to the corresponding deep foundation pit project. Among them, the foundation engineering information is obtained by relevant staff through on-site exploration before the construction of the target deep foundation pit project;

[0077] Obtain the current construction stage of the target deep foundation pit project, and based on the foundation engineering information, obtain the BIM data and 3DGIS data corresponding to the target deep foundation pit project; the BIM data includes geological exploration data, support structure design models, underground pipeline models, construction progress plans, etc.; the 3DGIS data refers to the geographical space data related to the target deep foundation pit project, such as terrain data, deep foundation pit images, etc.;

[0078] Input the corresponding BIM data and 3DGIS data into the corresponding BIM professional software and GIS professional software respectively; obtain the corresponding physical model and numerical model; superimpose the corresponding physical model and digital model to obtain the corresponding BIM model and 3DGIS model;

[0079] Furthermore, integrate the obtained BIM model and 3DGIS model to obtain the corresponding twin foundation pit model; the twin foundation pit model can be used to quickly locate the detailed engineering situation of the deep foundation pit project and can also be used for the construction simulation of the deep foundation pit project;

[0080] Conduct numerical simulation based on the foundation pit twin model, obtain the high-frequency risk areas in the corresponding deep foundation pit project at the current construction stage, and set several data monitoring points in the target high-frequency risk areas based on the pre-set monitoring distribution requirements; the monitoring distribution requirements refer to deploying monitoring points at key parts such as the support structure (such as retaining piles, support beams, etc.), surrounding soil, adjacent buildings, and underground pipelines in the corresponding deep foundation pit project according to factors such as the shape, size, geological conditions, and surrounding environment of the deep foundation pit; for example: arranging inclinometers and stress monitoring points at certain intervals on the retaining structure in the deep foundation pit project, and arranging displacement monitoring points and pore water pressure monitoring points in a certain grid pattern in the surrounding soil of the foundation pit, etc., to ensure that monitoring data can be obtained comprehensively and accurately;

[0081] It should be further noted that in the specific implementation process, the process of collecting data from the monitoring nodes in the target deep foundation pit project to obtain the corresponding engineering monitoring data set includes:

[0082] Set up data collection nodes, which are composed of several Internet of Things sensing terminals with different functions. The Internet of Things sensing terminals include, but are not limited to, displacement sensing terminals, stress sensing terminals, water level sensing terminals, etc.;

[0083] Furthermore, deploy the data collection nodes to the set monitoring nodes, and based on them, collect data at the corresponding monitoring nodes and upload it to the pre-constructed monitoring terminal; the monitoring terminal summarizes the collected data according to the pre-set collection period and uploads it to obtain the corresponding engineering monitoring data set; the engineering monitoring data set is composed of the engineering monitoring data corresponding to several different engineering monitoring parameters; among them, each engineering monitoring parameter has a corresponding Internet of Things sensing terminal; and there is at least one monitoring terminal in a high-frequency risk area.

[0084] It should be further noted that in the specific implementation process, the process of preprocessing the obtained data set to obtain the corresponding time series data set includes:

[0085] Taking a certain engineering monitoring data as an example, obtain the engineering monitoring data corresponding to the corresponding engineering monitoring parameter; construct a two-dimensional rectangular coordinate system of time with respect to the engineering monitoring parameter, and map the obtained engineering monitoring parameter into the corresponding two-dimensional rectangular coordinate system to obtain the corresponding parameter change curve;

[0086] Based on the parameter change curve, obtain the parameter average value corresponding to the engineering monitoring parameter within the corresponding collection period and respectively obtain the Mahalanobis distance between each data point in the corresponding parameter change curve and the corresponding parameter average value ; where represents the Mahalanobis distance corresponding to the data point , represents the parameter value of the data point in the parameter change curve; represents the inverse matrix of the covariance matrix corresponding to the parameter change curve; represents the transpose operation;

[0087] Set a distance threshold, and determine whether the corresponding data point is an outlier in combination with the preset constraint conditions; if the corresponding Mahalanobis distance is less than the distance threshold and the constraint conditions are met, it indicates that the corresponding data point is not an outlier, then no other operations are performed; if the Mahalanobis distance is not less than the distance threshold or the constraint conditions are not met, it indicates that the corresponding data point is an outlier, then the corresponding outlier is corrected based on the linear interpolation algorithm, and the corrected data point replaces the original data point in the parameter change curve; if the Mahalanobis distance is not less than the distance threshold and the constraint conditions are not met, then the corresponding data point is removed;

[0088] Among them, the formula for defining the constraint conditions is: ; In the formula, represents the data point to the data point curve slope; represents the average value of the slope differences of adjacent data points to the data point ; represents the total number of data points in the parameter change curve; represents the data point +1 to the data point 2 curve slope; m represents the index of the data point in the parameter change curve;

[0089] Based on the above data correction process, data correction is performed on the engineering monitoring data corresponding to all the corresponding engineering monitoring parameters, and time synchronization is performed on the corrected engineering monitoring data of different engineering monitoring parameters based on the preset reference time series to obtain the time series data corresponding to the corresponding engineering monitoring parameters and count them to obtain the corresponding time series data set; to eliminate the differences caused by the data collection frequencies of different Internet of Things sensing terminals;

[0090] Among them, the mathematical formula for time synchronization is: ; In the formula, represents the parameter value of the data point after time synchronization at the moment, and respectively represent the parameter values corresponding to the and moments in the corrected reference change curve; and ; represents the index of the th moment in the reference time series; and respectively represent the and th moments of the time series corresponding to the corrected parameter change curve, is the moment index within the time series corresponding to the corrected parameter change curve;

[0091] It should be further noted that in the specific implementation process, the process of obtaining the corresponding cycle prediction data includes:

[0092] Construct a parameter prediction model, input the obtained time series data set into the constructed parameter prediction model, obtain the corresponding model data result, and obtain the cycle prediction data for the next acquisition cycle based on it;

[0093] Among them, the construction process of the parameter prediction model includes:

[0094] The backbone network of the parameter prediction model is an improved multi-level neural network, and a two-stage attention mechanism is introduced; the basic framework of the improved multi-level neural network includes an input layer, a feature layer, and an output layer;

[0095] The input layer is used to receive the input time series data, perform data difference operation on it, and obtain the corresponding input sequence data; to obtain a stationary time series data that meets the model input; the formula for performing data difference operation is: ; where represents the order difference of the data point corresponding to the th moment, represents the time lag operator; represents the parameter value of the data point corresponding to the th moment in the time series data; represents the combination number; used to represent the influence of the observed value at the time node in the order data difference operation on the observed value at the current th moment; 0, is used to represent the index of the time lag term in the

[0096] The feature layer consists of an encoding unit and a decoding unit;

[0097] The encoding unit assigns weights to the corresponding input sequence data based on the introduced attention mechanism, and performs sequence update to obtain the corresponding updated sequence data , and encodes the updated sequence data based on the built-in Bi-LSTM, while capturing the forward and backward dependencies of the time series, so as to extract richer feature information; among them, the Bi-LSTM is composed of two LSTM layers in parallel, one processes the forward part of the input sequence data, and calculates sequentially from the first element to the last element of the input sequence data; the other processes its reverse part, and calculates in reverse order from the last element to the first element of the input sequence data; the corresponding output results are merged at each time step or at the end of the input sequence data to enhance the model's ability to capture bidirectional context;

[0098] The formula for weight assignment is:

[0099] ; in the formula, represents the weight of the input sequence data at time t1, , and all represent weight matrices; and respectively represent the hidden layer state and cell state in the LSTM layer at time; represents the operation process of the corresponding LSTM layer; represents the input sequence data; represents the hyperbolic tangent function; e is the matrix dimension of the weight matrix;

[0100] The decoding unit is used to receive the encoded updated sequence data, update the hidden layer state and the hidden layer state weight in the corresponding decoding unit based on it, and perform weighted summation and decoding on the encoded updated sequence data received by the decoding unit based on the updated hidden layer state and the hidden layer state weight, and obtain the corresponding prediction result based on it ; and represent weight matrices; and represent bias terms; and represent the hidden layer state and cell state in the decoding stage at time t1;

[0101] The formula for updating the hidden layer state is: ; in the formula, represents the activation function operation; represents the state vector obtained by processing the hidden layer state and cell state at time t1-1 through the activation function softmax;

[0102] The formula for updating the hidden layer state weight is:

[0103] ; , and represent the representation weight matrix in the decoding stage; and represent the hidden layer state and cell state in the decoding stage at time t1 - 1; represents the matrix dimension;

[0104] The input layer is used to receive the obtained prediction result and map it to the same data dimension as the original time - series data; that is, the inverse data difference operation process;

[0105] Define the loss function of the parameter prediction model as ; where, and respectively represent the index and total number of samples in the training data set, and respectively represent the actual result and expected result of the sample, represents the model parameters; represents the learning efficiency factor;

[0106] The training process of the corresponding parameter prediction model:

[0107] Construct a training data set, which is composed of deep foundation pit engineering risk data in the normal state within several adjacent acquisition periods;

[0108] Construct a multi - layer neural network, and initialize the network parameters of the multi - layer neural network based on the WOA algorithm. The initialization setting refers to obtaining the optimal network parameters based on the WOA algorithm and applying them to improve the prediction accuracy of the model. The network parameters include parameters such as weight matrices and bias terms;

[0109] After the initialization is completed, divide the training data set into several training batches, and input them into the corresponding parameter prediction model in sequence based on time for model training until the loss function converges under several consecutive training batches, then save the model parameters.

[0110] It should be further noted that in the specific implementation process, the process of obtaining the risk status of the target deep foundation pit project based on the obtained cycle prediction data and performing risk early warning includes:

[0111] Taking any engineering monitoring parameter as an example, obtain the time - series data corresponding to the current acquisition cycle and adjacent historical acquisition cycles, and combine the obtained cycle prediction data to obtain the parameter change rate and cumulative change amount corresponding to the corresponding factory monitoring parameter;

[0112] Construct an early warning risk framework and construct a corresponding fuzzy membership function based on the current construction stage of the deep foundation pit; among them, the early warning risk framework consists of three sub-intervals, and the sub-intervals include a normal interval, an abnormal interval, and a fault interval; the fuzzy membership function includes the probability distribution functions corresponding to each sub-interval; among them, in the actual application process, the construction stages of different deep foundation pit projects are also different, so the present invention does not specifically limit the probability distribution function;

[0113] Obtain the parameter change rate and cumulative change amount corresponding to the engineering monitoring parameters, and combine them with the fuzzy membership function and the boundary points of each sub-interval to obtain the probability distribution values corresponding to the corresponding parameter change rate and cumulative change amount, and record them respectively as and ; g ∈ ; respectively represent the normal interval, the abnormal interval, and the fault interval; and respectively represent the probability distribution values of the parameter change rate and the cumulative change amount for the sub-interval g;

[0114] Perform data fusion on the probability distribution values corresponding to the corresponding parameter change rate and cumulative change amount to obtain the corresponding fusion parameter , is the fusion coefficient, which is used to represent the probability distribution value that the engineering monitoring parameter belongs to the sub-interval ; represents the data fusion operation based on the D-S evidence algorithm;

[0115] Furthermore, obtain the fusion coefficients corresponding to each engineering monitoring parameter, and perform data merging based on the improved D-S evidence algorithm to obtain the corresponding membership distribution value of the deep foundation pit project ;

[0116] In the formula, and respectively represent the fusion coefficients corresponding to the first and the th factory monitoring parameters; is the total number of categories of engineering monitoring parameters; represents the conflict factor; represents the early warning interval, including the abnormal interval and the fault interval;

[0117] Obtain the sub-interval corresponding to the maximum membership distribution value of the deep foundation pit project, and determine whether there is a risk for the corresponding deep foundation pit project based on it; if the sub-interval is a normal interval, do not perform any other operations; if the sub-interval is an abnormal or faulty interval, mark the abnormal or faulty interval as the risk status of the corresponding deep foundation pit project; furthermore, obtain the fusion coefficients corresponding to different engineering monitoring parameters, and obtain the parameter scores corresponding to the corresponding engineering monitoring parameters based on the pre-set scoring intervals; for example: assume that the scoring interval corresponding to the abnormal interval of a certain engineering monitoring parameter is ; then the parameter score of the corresponding engineering monitoring parameter is ; represents the probability distribution value (i.e., the fusion coefficient) that the corresponding engineering monitoring parameter belongs to the abnormal interval;

[0118] Furthermore, obtain the parameter weights corresponding to each engineering monitoring data under different risk factors, and perform weighted summation on the corresponding parameter scores based on them to obtain the risk assessment scores corresponding to the corresponding risk factors, obtain the risk factor corresponding to the highest risk assessment score, and output it as the risk cause;

[0119] Further, according to the deep foundation pit safety assessment requirements, the risk assessment scores are corresponding to different abnormal or faulty early warning levels, and are fed back into the corresponding monitoring terminals. Furthermore, the monitoring terminals are based on the built-in alarm devices and combine the abnormal or faulty early warning levels to perform different types of sound and light warnings; at the same time, the monitoring terminals generate corresponding alarm information and perform warning feedback based on the pre-constructed WeChat public account. The alarm information includes relevant information such as the geographical location, risk cause, and early warning time of the corresponding monitoring terminal in the corresponding deep foundation pit project;

[0120] At the same time, color marking is performed at the corresponding positions in the corresponding foundation pit twin model based on the corresponding abnormal or faulty early warning levels. The corresponding staff can view the color-marked foundation pit twin model through the early warning notification of the WeChat public account to quickly understand the engineering status of the current deep foundation pit project;

[0121] It should be further noted that in the specific implementation process, the process of obtaining the parameter weights corresponding to each engineering monitoring data under different risk factors includes:

[0122] Taking a certain risk factor as an example, obtain the historical time series data set corresponding to the trigger of the corresponding risk factor, and respectively obtain the feature vectors corresponding to each engineering monitoring parameter based on the historical time series data set, and obtain the correlation degree between different engineering monitoring parameters based on them; the correlation degree is the ratio after the normalization processing of the corresponding feature vectors;

[0123] Based on the correlation degree, construct the corresponding correlation matrix ; where, Indicating the correlation degree between engineering monitoring parameters and engineering monitoring parameters between; and both represent the indexes of engineering monitoring parameters under corresponding risk factors, and ;

[0124] Based on the correlation matrix, obtain the initial weight corresponding to the corresponding engineering monitoring parameter wherein, in the formula, , where E is an integer representing the total number of categories of engineering monitoring parameters under corresponding risk factors; , E is an integer, representing the total number of categories of engineering monitoring parameters under corresponding risk factors;

[0125] Meanwhile, perform a normalization process on the obtained correlation matrix to obtain a corresponding correction matrix ; where, ; is the matrix element of the correction matrix, used to represent the correlation degree after normalization processing. The normalization processing is a prior art and will not be elaborated in detail in the present invention;

[0126] Furthermore, based on the correction matrix, obtain a corresponding weight correction coefficient ; where, ; used to represent the parameter deviation of the engineering monitoring parameter;

[0127] And based on it, perform weight correction on the obtained initial weight to obtain a corresponding corrected weight; the formula for weight correction is: ; represents the corrected weight of the engineering monitoring parameter ;

[0128] Perform weighted summation on the obtained corrected weight and initial weight to obtain a corresponding combined weight, and use it as the parameter weight of the engineering monitoring parameter under the corresponding risk factor;

[0129] Furthermore, obtain all known types of risk factors, and based on the above process of obtaining parameter weights, assign weights to the engineering monitoring parameters under different risk factors.

[0130] By collecting and analyzing engineering monitoring data in real time, the present invention can timely discover potential risks and take corresponding measures to ensure the safety and stability of deep foundation pit engineering; at the same time, the method also has a high level of automation and intelligence, which can reduce the labor intensity of staff and improve work efficiency.

[0131] Embodiment 2 Please refer to Figure 2 shown. For the parts not described in detail in this embodiment, refer to the description content of Embodiment 1. Provide an Internet of Things-based risk monitoring and early warning system for deep foundation pit engineering, including:

[0132] A data acquisition module, configured to construct a twin model of the foundation pit, set corresponding monitoring nodes based on it, and collect data of the monitoring nodes in the target deep foundation pit project by using pre-deployed Internet of Things terminals with different functions, so as to obtain a corresponding set of engineering monitoring data;

[0133] A data processing module, configured to identify abnormal points in the obtained set of engineering monitoring data, correct the corresponding abnormal points, and after the correction is completed, synchronize the engineering monitoring data corresponding to different Internet of Things terminals to obtain corresponding time series data; and input it into a pre-constructed parameter prediction model to obtain corresponding periodic prediction data;

[0134] A data analysis module, based on the obtained periodic prediction data, and combining the collected set of engineering monitoring data, conducts a risk assessment on the target deep foundation pit project to obtain a corresponding risk status;

[0135] A data feedback module, which generates a corresponding risk warning level based on the risk status and conducts warning feedback;

[0136] Each module is connected by wired and / or wireless means to realize data transmission between modules.

[0137] Embodiment 3 This embodiment publicly provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, it realizes the operation mode of the above-mentioned Internet of Things-based deep foundation pit engineering risk monitoring and warning method and system.

[0138] Since the electronic device introduced in this embodiment is the electronic device used to implement the Internet of Things-based deep foundation pit engineering risk monitoring and warning method and system in the embodiments of the present application, based on the Internet of Things-based deep foundation pit engineering risk monitoring and warning method and system introduced in the embodiments of the present application, those skilled in the art can understand the specific implementation manner and various change forms of the electronic device in this embodiment. Therefore, the specific implementation of how this electronic device realizes the method in the embodiments of the present application will not be described in detail here. As long as those skilled in the art implement the electronic device used in the Internet of Things-based deep foundation pit engineering risk monitoring and warning method and system in the embodiments of the present application, it falls within the scope of protection of the present application.

[0139] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to obtain a formula that is closest to the real situation. The preset parameters and threshold selection in the formulas are set by those skilled in the art according to the actual situation.

[0140] The above are only the preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions falling within the concept of the present invention belong to the protection scope of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements should also be regarded as within the protection scope of the present invention.

Claims

1. A risk monitoring and early warning method for deep foundation pit engineering based on the Internet of Things, characterized in that: include: Step 1: Construct a twin model of the foundation pit, and set up corresponding monitoring nodes based on it. Collect data from the monitoring nodes in the target deep foundation pit project based on pre-deployed IoT terminals with different functions to obtain the corresponding engineering monitoring data set; Step 2: Identify abnormal points in the acquired engineering monitoring data set and correct the corresponding abnormal points. After the correction is completed, synchronize the engineering monitoring data corresponding to different IoT terminals to obtain the corresponding time series data; And input it into the pre-built parameter prediction model to obtain the corresponding period prediction data; Step 3: Based on the obtained periodic prediction data and in combination with the collected engineering monitoring data set, a risk assessment is conducted on the target deep foundation pit project to obtain the corresponding risk status, and based on the risk status, a corresponding risk warning level is generated, and warning feedback is provided.

2. The method for monitoring and early warning of deep foundation pit engineering risks based on the Internet of Things according to claim 1 is characterized in that: The process of constructing the foundation pit twin model and setting corresponding monitoring nodes based on it includes: Acquire the foundation engineering information corresponding to the deep foundation pit project, and acquire the BIM data and 3DGIS data corresponding to the target deep foundation pit project based on the foundation engineering information; Input the corresponding BIM data and 3DGIS data into the corresponding BIM professional software and GIS professional software respectively; obtain the corresponding physical model and numerical model; superimpose the corresponding physical model and digital model to obtain the corresponding BIM model and 3DGIS model and integrate the models to obtain the corresponding twin foundation pit model; Based on the foundation pit twin model, numerical simulation is performed to obtain high-frequency risk areas in the corresponding deep foundation pit project under the current construction stage, and a number of data monitoring points are set in the target high-frequency risk areas based on pre-set monitoring distribution requirements.

3. The method for monitoring and early warning of deep foundation pit engineering risks based on the Internet of Things according to claim 2 is characterized in that: The process of collecting data from monitoring nodes in the target deep foundation pit project and obtaining the corresponding engineering monitoring data set includes: Setting up a data collection node, wherein the data collection node is composed of a number of IoT sensor terminals with different functions; The data collection node is deployed in the set monitoring node, and data is collected at the corresponding monitoring node based on it, and uploaded to the pre-built monitoring terminal; the monitoring terminal summarizes the collected data according to the pre-set collection cycle and uploads the data to obtain the corresponding engineering monitoring data set; the engineering monitoring data set is composed of engineering monitoring data corresponding to several different engineering monitoring parameters.

4. The method for monitoring and early warning of deep foundation pit engineering risks based on the Internet of Things according to claim 3 is characterized in that: The process of preprocessing the obtained data set to obtain the corresponding time series data set includes: Acquire engineering monitoring data corresponding to corresponding engineering monitoring parameters; construct a two-dimensional rectangular coordinate system of time with respect to the engineering monitoring parameters, and map the obtained engineering monitoring parameters into the corresponding two-dimensional rectangular coordinate system to obtain corresponding parameter change curves; Based on the parameter change curve, the parameter average value corresponding to the engineering monitoring parameter in the corresponding acquisition period is obtained , and obtain the Mahalanobis distance between each data point in the corresponding parameter change curve and the corresponding parameter average value respectively; Set a distance threshold, and determine whether the corresponding data point is an outlier based on the pre-set constraints; if the corresponding Mahalanobis distance is less than the distance threshold and meets the constraints, no other operations are performed; if the Mahalanobis distance is not less than the distance threshold or does not meet the constraints, the corresponding data point is corrected based on the linear interpolation algorithm, and the corrected data point replaces the original data point in the parameter change curve; if the Mahalanobis distance is not less than the distance threshold and does not meet the constraints, the corresponding data point is eliminated; Based on the above data correction process, the engineering monitoring data corresponding to all engineering monitoring parameters are corrected, and based on the pre-set reference time series, the engineering monitoring data corresponding to different engineering monitoring parameters after data correction are synchronized, and the time series data corresponding to the corresponding engineering monitoring parameters are obtained and counted to obtain the corresponding time series data set.

5. The method for monitoring and early warning of deep foundation pit engineering risks based on the Internet of Things according to claim 3 is characterized in that: The formula for defining the constraints is: ; In the formula, Represents data points To data point The slope of the curve; Representation and data points The average of the slope differences of adjacent data points; Indicates the total number of data points in the parameter variation curve; Represents data points To data point 2, the slope of the curve; Represents the index of the data point in the parameter variation curve; The mathematical formula for time synchronization is: ; In the formula, Indicated in The parameter value of the data point after time synchronization at the moment, and Respectively represent the reference change curve after correction and The parameter value corresponding to the moment; and ; Indicates the first The index of a moment; and They represent the time series of the corrected parameter change curves. and a moment, It is the time index in the time series corresponding to the corrected parameter change curve.

6. The method for monitoring and early warning of deep foundation pit engineering risks based on the Internet of Things according to claim 5 is characterized in that: The process of obtaining the corresponding cycle forecast data includes: Construct a parameter prediction model, and input the obtained time series data set into the constructed parameter prediction model to obtain the corresponding model data results, and based on the results, obtain the period prediction data for a future collection period; The construction process of the parameter prediction model includes: The backbone network of the parameter prediction model is an improved multi-layer neural network, and a two-stage attention mechanism is introduced; the basic framework of the improved multi-layer neural network includes an input layer, a feature layer and an output layer; The input layer is used to receive input time series data and perform data difference operation on it to obtain corresponding input sequence data; The feature layer consists of an encoding unit and a decoding unit; The encoding unit assigns weights to the corresponding input sequence data based on the introduced attention mechanism, and performs sequence updates to obtain corresponding updated sequence data , and encode the update sequence data based on the built-in Bi-LSTM, while capturing the previous and next dependencies of the time series; The decoding unit is used to receive the encoded update sequence data, and based on the updated hidden layer state and the hidden layer state weight in the corresponding decoding unit, and based on the updated hidden layer state and the hidden layer state weight, the encoded update sequence data received by the decoding unit is weighted and decoded to obtain a corresponding prediction result; The input layer is used to receive the obtained prediction results and map them to the same data dimension as the original time series data; Build a training data set; Construct a multi-level neural network, and initialize the network parameters of the multi-level neural network based on the WOA algorithm; after the initialization is completed, divide the training data set into several training batches, and input the corresponding parameter prediction model in sequence based on the time before and after for model training, until the loss function converges under several consecutive training batches, then save the model parameters.

7. The method for monitoring and early warning of deep foundation pit engineering risks based on the Internet of Things according to claim 6 is characterized in that: The formula for assigning weights is: ; In the formula, represents the weight of the input sequence data at time t1, , and Both represent weight matrices; and Respectively represent the coding unit in Hidden layer state and cell state in the LSTM layer at this moment; Represents the operation process of the corresponding LSTM layer; Represents input sequence data; represents the hyperbolic tangent function; e is the matrix dimension of the weight matrix; Get the corresponding prediction results ; and represents the weight matrix; and represents the bias term; and express Hidden layer state and cell state at the moment decoding stage The update formula for the hidden layer state is: ; In the formula, Represents the activation function operation; express The state vector obtained by processing the hidden layer state and cell state at each moment through the activation function softmax; The update formula for the hidden layer state weight is: ; , and represents the weight matrix of the decoding stage; and Represents the hidden layer state and cell state at the decoding stage at time t1-1; Represents the matrix dimensions.

8. The method for monitoring and early warning of deep foundation pit engineering risks based on the Internet of Things according to claim 7 is characterized in that: The process of obtaining the risk status of the target deep foundation pit project based on the obtained periodic prediction data and conducting risk warning based on the risk status includes: Obtain the time series data corresponding to the current collection cycle and the adjacent historical collection cycles, and obtain the parameter change rate and cumulative change amount corresponding to the corresponding factory monitoring parameters in combination with the obtained cycle prediction data; Constructing an early warning risk framework, and constructing a corresponding fuzzy membership function based on the current construction stage of the deep foundation pit; the early warning risk framework is composed of three sub-intervals, including a normal interval, an abnormal interval, and a fault interval; Obtain the parameter change rate and cumulative change amount corresponding to the engineering monitoring parameters, and combine them with the fuzzy membership function and the boundary points of each sub-interval to obtain the probability distribution values ​​corresponding to the corresponding parameter change rate and cumulative change amount, and record them as and ; g∈ ; They represent normal interval, abnormal interval and fault interval respectively; and They represent the probability distribution values ​​of the parameter change rate and the cumulative change for the subinterval g respectively; The corresponding parameter change rate and the probability distribution value corresponding to the cumulative change are fused to obtain the corresponding fusion parameter , is the fusion coefficient; represents the data fusion operation based on DS evidence algorithm; Obtain the fusion coefficient corresponding to each engineering monitoring parameter, and merge the data based on the improved DS evidence algorithm to obtain the corresponding membership distribution value of the corresponding deep foundation pit project ; Obtain the sub-interval corresponding to the maximum membership distribution value corresponding to the deep foundation pit project, and determine whether the corresponding deep foundation pit project has risks based on it; if the sub-interval is a normal interval, no other operations are performed; if the sub-interval is an abnormal or faulty interval, the abnormal or faulty interval is marked as the risk status of the corresponding deep foundation pit project; obtain the fusion coefficient corresponding to different engineering monitoring parameters, and obtain the parameter score corresponding to the corresponding engineering monitoring parameter based on the pre-set scoring interval; Obtain the parameter weights corresponding to the monitoring data of each project under different risk factors, and perform weighted summation on the corresponding parameter scores based on them to obtain the risk assessment scores corresponding to the corresponding risk factors, obtain the risk factor corresponding to the highest risk assessment score, and output it as the risk cause; According to the safety assessment requirements of deep foundation pits, the risk assessment scores are mapped to different abnormal or fault warning levels and fed back to the corresponding monitoring terminals. The monitoring terminals use built-in alarm devices and perform different types of sound and light warnings in combination with the abnormal or fault warning levels. At the same time, the monitoring terminals generate corresponding alarm information and provide warning feedback based on a pre-built WeChat public account.

9. The method for monitoring and early warning of deep foundation pit engineering risks based on the Internet of Things according to claim 7 is characterized in that: The process of obtaining the parameter weights corresponding to the monitoring data of each project under different risk factors includes: Obtain the historical time series data set corresponding to the corresponding risk factor that triggers the corresponding risk factor, and obtain the feature vectors corresponding to each engineering monitoring parameter based on the historical time series data set, and obtain the correlation between different engineering monitoring parameters based on the feature vectors; construct the corresponding correlation matrix based on the correlation ;in, Indicates engineering monitoring parameters and engineering monitoring parameters The degree of correlation between and Both represent the index of engineering monitoring parameters under the corresponding risk factors, and ; Obtain corresponding engineering monitoring parameters based on the association matrix The corresponding initial weight , where ,E is an integer, indicating the total number of categories of engineering monitoring parameters under the corresponding risk factor; The obtained correlation matrix is ​​standardized to obtain the corresponding correction matrix ;in, ; is the matrix element of the correction matrix, which is used to represent the correlation degree after standardization; Obtaining corresponding weight correction coefficients based on the correction matrix ;in, ; Used to indicate the parameter deviation of engineering monitoring parameters; Based on this, the obtained initial weight is modified to obtain the corresponding modified weight; the formula for weight modification is: ; Indicates engineering monitoring parameters The correction weight of The obtained revised weight and initial weight are weighted and summed to obtain the corresponding combined weight, which is used as the parameter weight of the engineering monitoring parameter under the corresponding risk factor; All known types of risk factors are obtained, and based on the acquisition process of the above parameter weights, weights are assigned to engineering monitoring parameters under different risk factors.

10. A deep foundation pit engineering risk monitoring and early warning system based on the Internet of Things, which is used to implement the deep foundation pit engineering risk monitoring and early warning method based on the Internet of Things as claimed in any one of claims 1 to 9, characterized in that: include: The data acquisition module is used to construct a twin model of the foundation pit, set up corresponding monitoring nodes based on it, and collect data from the monitoring nodes in the target deep foundation pit project based on pre-deployed IoT terminals with different functions to obtain the corresponding engineering monitoring data set; The data processing module is used to identify abnormal points in the acquired engineering monitoring data set and correct the corresponding abnormal points. After the correction is completed, the engineering monitoring data corresponding to different IoT terminals are synchronized to obtain the corresponding time series data; And input it into the pre-built parameter prediction model to obtain the corresponding period prediction data; The data analysis module conducts risk assessment on the target deep foundation pit project based on the obtained period prediction data and in combination with the collected engineering monitoring data set to obtain the corresponding risk status; The data feedback module generates a corresponding risk warning level based on the risk status and provides warning feedback.

Images