Deep subway tunnel construction stratum deformation risk early warning method, device and equipment

Abstract

The invention relates to the field of deformation monitoring, in particular to a deep subway tunnel construction stratum deformation risk early warning method, device and equipment. The method comprises the steps of collecting multi-source detection data such as ground surface settlement, building settlement, a building inclination angle and an underground water level change value in real time, performing preprocessing to generate a standardized time sequence data set, and performing risk grade judgment by using a stratum deformation risk assessment model. The stratum deformation risk condition of the current construction area can be accurately and timely reflected, whether the early warning instruction is generated or not is rapidly determined based on the risk level signal, effective early warning of the deep subway tunnel construction stratum deformation risk is achieved, the accuracy of risk early warning is improved, precautionary measures can be taken in advance, construction safety is guaranteed, and the construction efficiency is improved. Therefore, potential loss caused by stratum deformation is reduced.

Description

Technical Field

[0001] The present application relates to the field of deformation monitoring, and in particular to a method, device and equipment for early warning of ground deformation risks during construction of deep subway tunnels. Background Art

[0002] Deep subway tunnel construction technology plays a crucial role in modern urban infrastructure development. It not only effectively alleviates urban traffic pressure but also offers broad prospects for the development and utilization of underground space. However, because subway tunnels are often located in complex geological environments, ground deformation is a frequent problem during construction, posing a threat to the safety of surrounding buildings and underground pipelines. Therefore, accurately monitoring and assessing ground deformation risks is crucial for ensuring construction safety and improving project efficiency.

[0003] Currently, to monitor ground deformation during deep subway tunnel construction, the industry typically deploys traditional surface settlement monitoring points, manually collecting data regularly to assess ground changes. However, this method has long data collection cycles and poor real-time performance, making it difficult to rapidly respond to ground deformation risks under complex geological conditions, resulting in poorly accurate risk warnings. Summary of the Invention

[0004] In order to improve the accuracy of risk warning, the present application provides a method, device and equipment for warning of stratum deformation risk in deep subway tunnel construction.

[0005] In the first aspect, the present application provides a method for early warning of ground deformation risk during construction of a deep subway tunnel, which adopts the following technical solutions: A method for early warning of ground deformation risk during deep subway tunnel construction, comprising: Real-time collection of multi-source detection data corresponding to the current construction area, including surface settlement, building settlement, building tilt angle, and groundwater level change; Preprocessing the multi-source detection data to generate a standardized time series data set; Inputting the standardized time series data set into a formation deformation risk assessment model to obtain a risk level signal; Based on the risk level signal, it is determined whether to generate an early warning instruction.

[0006] By adopting the above technical solution, multi-source detection data such as surface settlement, building settlement, building inclination angle and groundwater level change are collected in real time, and pre-processed to generate a standardized time series data set. The risk level is then determined using the stratum deformation risk assessment model. This can accurately and timely reflect the stratum deformation risk status of the current construction area, and then quickly determine whether to generate an early warning instruction based on the risk level signal. This achieves effective early warning of the stratum deformation risk in deep subway tunnel construction, improves the accuracy of risk warnings, and helps to take preventive measures in advance to ensure construction safety, thereby reducing potential losses caused by stratum deformation.

[0007] In one possible implementation, preprocessing the multi-source detection data to generate a standardized time series data set includes: Converting the building tilt angle into a radian-degree system to obtain a target building tilt angle; The groundwater level change value is corrected to obtain the target groundwater level change value, and the correction formula is H 目标 =H 初始 +K(P 采集 -P 基准 ), where H 目标 is the target groundwater level change value, H 初始 is the groundwater level change value, K is the geological permeability correction factor, P 采集 is the collected groundwater pressure value, P 基准 Calibrate the pressure reference value for groundwater; A segmented normalization strategy is adopted to perform normalization processing on the target detection building inclination angle and the target groundwater level change value to obtain processed target detection building inclination angle and target groundwater level change value; Unifying the dimensions of the ground surface settlement and the building settlement to obtain target ground surface settlement and target building settlement; Based on the processed target detection building inclination angle and target groundwater level change value, target surface settlement and target building settlement, a standardized time series dataset is generated.

[0008] In a possible implementation, before inputting the standardized time series data set into the formation deformation risk assessment model, the method further includes: Acquiring historical construction data corresponding to the current construction area, the historical construction data including historical multi-source detection data and historical stratum deformation data; Based on the historical construction data, a dynamic weight allocation network is trained. The dynamic weight allocation network assigns real-time weights to surface settlement, building inclination angle and groundwater level change according to the geological parameters and construction parameters of the current construction stage. Based on the trained dynamic weight allocation network, a stratum deformation risk assessment model is obtained.

[0009] In one possible implementation, training a dynamic weight allocation network based on the historical construction data includes: dividing the historical multi-source detection data into time-series subsets and constructing a dynamic weight allocation network, wherein the dynamic weight allocation network includes an input layer, an LSTM time-series encoding layer, and an attention weight allocation layer; The time series subset is input into the dynamic weight allocation network, and the historical stratum deformation data is used as a supervision label. The dynamic weight allocation network is trained through a loss function until the loss function converges to obtain a trained dynamic weight allocation network.

[0010] In one possible implementation, a formation deformation risk assessment model is obtained based on the trained dynamic weight distribution network, including: Constructing a risk assessment model, wherein the risk assessment model includes a convolutional layer, a Transformer encoding layer, and a fully connected layer; Based on the historical construction data, the risk assessment model is trained to obtain a target risk assessment model; Based on the trained dynamic weight distribution network and the target risk assessment model, a formation deformation risk assessment model is obtained.

[0011] In one possible implementation, the standardized time series data set is input into a formation deformation risk assessment model to obtain a risk level signal, including: Input the standardized time series data into the stratum deformation risk assessment model, and obtain the dynamic weights of the output surface settlement, building settlement, building inclination angle and groundwater level change value; Inputting the weighted monitoring data into the target risk assessment model, and obtaining a deformation probability value output by the target risk assessment model; A risk level signal corresponding to the current construction area is determined based on the deformation probability value.

[0012] In one possible implementation, determining whether to generate a warning instruction based on the risk level signal includes: Determining whether the risk level signal reaches a preset level threshold; If the risk equals that the signal does not reach the preset level threshold, no warning instruction is generated; If the risk level signal reaches the preset level threshold, a warning instruction is generated, and the construction state of the construction equipment is controlled based on the warning instruction, and the construction state is speed reduction or shutdown.

[0013] In a second aspect, the present application provides a ground deformation risk warning device for deep subway tunnel construction, which adopts the following technical solution: A ground deformation risk early warning device for deep subway tunnel construction, comprising: The acquisition module is used to collect multi-source detection data corresponding to the current construction area in real time. The multi-source detection data includes surface settlement, building settlement, building inclination angle and groundwater level change; A preprocessing module, configured to preprocess the multi-source detection data to generate a standardized time series data set; An input module, configured to input the standardized time series data set into a formation deformation risk assessment model to obtain a risk level signal; A determination module is used to determine whether to generate an early warning instruction based on the risk level signal.

[0014] In a third aspect, the present application provides an electronic device, which adopts the following technical solution: An electronic device, comprising: at least one processor; Memory; At least one application, wherein the at least one application is stored in a memory and configured to be executed by at least one processor, and the at least one application is configured to: execute the deep subway tunnel construction stratum deformation risk warning method as described in any one of the first aspects above.

[0015] In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium, comprising: a computer program stored therein that can be loaded by a processor and execute the method for early warning of ground deformation risk during deep subway tunnel construction as described in any one of the first aspects above.

[0016] In summary, this application has the following beneficial technical effects: By collecting multi-source detection data such as surface settlement, building settlement, building inclination angle, and groundwater level change in real time, and pre-processing it to generate a standardized time series data set, and then using the stratum deformation risk assessment model to determine the risk level, it can accurately and timely reflect the stratum deformation risk status of the current construction area, and then quickly determine whether to generate an early warning instruction based on the risk level signal. This realizes effective early warning of the stratum deformation risk in deep subway tunnel construction, improves the accuracy of risk warning, and helps to take preventive measures in advance to ensure construction safety, thereby reducing potential losses caused by stratum deformation. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of a method for early warning of ground deformation risk during construction of a deep subway tunnel provided by an embodiment of the present application; Figure 2 This is a block diagram of a device for early warning of ground deformation risk during construction of a deep subway tunnel provided by an embodiment of the present application; Figure 3 Schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0018] The following is combined with Figure 1 -Attached Figure 3 This application is described in further detail.

[0019] This specific embodiment is merely an explanation of the present application and is not a limitation of the present application. After reading this specification, those skilled in the art may make non-creative modifications to the present embodiment as needed, but as long as they are within the scope of the claims of the present application, they are protected by the patent law.

[0020] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0021] In this document, the term "and / or" simply describes a relationship between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " in this document, unless otherwise specified, generally indicates an "or" relationship between the related objects.

[0022] It should be noted that in the optional embodiments of the present application, the object information and other related data involved, when the embodiments in the present application are applied to specific products or technologies, need to obtain the permission or consent of the object, and the collection, use and processing of the relevant data need to comply with the relevant laws, regulations and standards of the relevant countries and regions. In other words, if the embodiments of the present application involve data related to the object, it needs to be obtained through the authorization and consent of the object, the authorization and consent of the relevant departments, and in compliance with the relevant laws, regulations and standards of the country and region. If personal information is involved in the embodiments, the acquisition of all personal information requires the consent of the individual. If sensitive information is involved, the separate consent of the information subject needs to be obtained. The embodiments also need to be implemented with the authorization and consent of the object.

[0023] The present application embodiment provides a method for early warning of ground deformation risk during construction of a deep subway tunnel. Figure 1 As shown, the method provided in the embodiment of the present application is performed by an electronic device, which can be a server or a terminal device, wherein the server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. The terminal device can be a smartphone, a tablet computer, a laptop computer, a desktop computer, etc., but is not limited thereto. The terminal device and the server can be directly or indirectly connected via wired or wireless communication, which is not limited in the embodiment of the present application. The method includes steps S101 to S104, wherein: Step S101: Collect multi-source detection data corresponding to the current construction area in real time.

[0024] Among them, multi-source detection data include surface settlement, building settlement, building tilt angle and groundwater level change.

[0025] Specifically, electronic equipment acquires real-time monitoring data through a sensor network deployed in the construction area (such as levels, total stations, stratified settlement meters, groundwater level meters, etc.). Specifically, it includes: Surface subsidence: Use a high-precision level to automatically scan surface monitoring points, and use laser ranging technology to record the elevation change of each point in real time; Building settlement / tilt angle: Install inclination sensors and static levels at key parts of the building (such as wall corners and column bases), and transmit data back to electronic equipment via wireless transmission modules; Groundwater level change value: Water level depth data is continuously collected through the pressure water level gauge buried in the groundwater level monitoring well and transmitted to the data acquisition terminal through the RS485 bus.

[0026] Step S102: pre-process the multi-source detection data to generate a standardized time series data set.

[0027] Among them, the preprocessing process can include denoising filtering, missing value filling and standardization conversion. Specifically, the sliding average filtering algorithm can be used to eliminate accidental errors (such as outliers caused by external vibration interference), and linear interpolation or historical data mean can be used for filling; at the same time, data of different dimensions (such as settlement in mm and water level change in m) are converted into dimensionless values in the range of [-1,1] through normalization formulas (such as Z-score standardization), and structured time series data sets are generated in timestamp order (such as CSV format tables, where each row corresponds to multi-dimensional data at a monitoring moment).

[0028] In this embodiment, preprocessing of multi-source detection data to generate a standardized time series data set includes: Convert the building tilt angle from radians to degrees to obtain the target building tilt angle; The groundwater level change value is corrected to obtain the target groundwater level change value. The correction formula is H 目标 =H 初始 +K(P 采集 -P 基准 ), where H 目标 is the target groundwater level change value, H 初始 is the groundwater level change value, K is the geological permeability correction factor, P 采集 is the collected groundwater pressure value, P 基准 Calibrate the pressure reference value for groundwater; The segmented normalization strategy is used to normalize the target detection building tilt angle and the target groundwater level change value to obtain the processed target detection building tilt angle and the target groundwater level change value; The dimensions of surface settlement and building settlement are unified to obtain target surface settlement and target building settlement. Based on the processed target detection building inclination angle, target groundwater level change value, target surface settlement and target building settlement, a standardized time series data set is generated.

[0029] After collecting the radian value output by the building tilt angle sensor (such as a MEMS tilt sensor), the conversion function in the current construction database can be used to multiply the radian value by (180 / π) to obtain the angle value, and the result is rounded to two decimal places to generate the target building tilt angle (such as converting from 0.0175rad to 1.0°), which is stored in the temporary data buffer area. Furthermore, the electronic device obtains the groundwater pressure value P through the pressure water level gauge of the groundwater level monitoring well. 采集 (Unit: kPa) and groundwater level change value H 初始(Unit: m), and call the pre-stored geological permeability correction factor K (which can be set according to actual conditions, such as 0.8 for sand layer and 1.2 for clay layer) and the groundwater calibration pressure reference value P 基准 (usually the initial pressure value before construction, such as 100kPa), then substitute the above groundwater level change value, geological permeability correction factor, collected groundwater pressure value and groundwater calibration pressure reference value into the correction formula H 目标 =H 初始 +K(P 采集 -P 基准 ), the target groundwater level change value H is calculated 目标 .

[0030] In view of the nonlinear characteristics of the tilt angle and groundwater level change, a piecewise normalization strategy is adopted to divide the data into a safe range (such as tilt angle < 0.5°, water level change < 0.3m) and a warning range (tilt angle ≥ 0.5°, water level change ≥ 0.3m). The minimum-maximum normalization is used for the safe range data: (mapped to the [0, 0.6] interval), and logarithmic normalization is used for the warning interval data: (Map to the interval [0.6, 1] to highlight the differences in high-risk data); output the processed target detection building inclination angle and target groundwater level change value (dimensionless value).

[0031] Furthermore, the surface settlement and building settlement can be unified in terms of dimensions. Specifically, the unit information can be obtained through sensor metadata or manual configuration files (e.g., the surface settlement meter is marked with "unit: mm"). If the building settlement unit is cm, it is multiplied by 10 (e.g., the original value "2.5 cm" is converted to "25 mm"). The 3σ principle is used to detect excessive data (e.g., a single-day settlement greater than 50 mm is considered abnormal) and the data is marked as missing values for subsequent filling.

[0032] Furthermore, the processed multi-source data can be integrated by time. Specifically, a structured data table can be created with timestamps (accurate to the minute, such as "2025-04-21 14:30") as the index. This table then stores columns containing the target surface settlement (mm), target building settlement (mm), processed tilt angle (dimensionless), and processed groundwater level change (dimensionless), resulting in a standardized time series dataset. Furthermore, the resulting standardized time series dataset can be saved in CSV format or a database table (such as the MySQL t_surface_deformation table) for periodic reading by the risk assessment model (e.g., updating the dataset every 10 minutes).

[0033] Step S103: input the standardized time series data set into the formation deformation risk assessment model to obtain a risk level signal.

[0034] The locally deployed stratum deformation risk assessment model is called, and the standardized time series dataset is cut into input samples according to time windows (such as the data of the previous 7 days). Each sample contains historical data from multiple monitoring cycles. The stratum deformation risk assessment model extracts features from the input data through trained parameters (such as the weight matrix of LSTM and the decision tree rules of random forest) and outputs the risk level signal of the current stratum deformation (such as a risk score of 0-100).

[0035] Specifically, before this embodiment, a process of establishing a stratum deformation risk assessment model may also be included, specifically: obtaining historical construction data corresponding to the current construction area, the historical construction data including historical multi-source detection data and historical stratum deformation data; Based on historical construction data, a dynamic weight allocation network is trained. The dynamic weight allocation network assigns real-time weights to surface settlement, building inclination angle, and groundwater level change according to the geological parameters and construction parameters of the current construction stage. Based on the trained dynamic weight distribution network, a formation deformation risk assessment model is obtained.

[0036] Among them, historical stratum deformation data: direct evidence of internal or surface deformation of the stratum (such as stratum displacement) obtained through geological exploration, structural monitoring and other means, is used to mark the real label of the risk level.

[0037] Multi-source detection data and historical stratum deformation data corresponding to subway tunnel construction are obtained from the current construction database to obtain historical construction data corresponding to the current construction area. Missing values in the historical multi-source detection data are filled (for example, using interpolation of adjacent time mean values) and outliers are eliminated (3σ principle). At the same time, geological parameters (such as soil type, moisture content, and permeability coefficient) and construction parameters (such as excavation speed, support type, and grouting pressure) are converted into numerical features (such as one-hot encoding support type).

[0038] Furthermore, an encoder-decoder structure can be used to receive geological parameters, construction parameters, and detection indicator features from the input layer. The middle layer learns the mapping relationship between parameters and weights through a fully connected layer, and the output layer generates real-time weights for each indicator (surface subsidence, tilt angle, groundwater level). The risk level (such as slight deformation, significant deformation, and dangerous deformation) in historical stratum deformation data is used as the supervision signal. A multi-task learning loss function can be used, and the weights can be iteratively updated using the Adam optimizer.

[0039] Furthermore, the trained dynamic weight distribution network is combined with the traditional risk assessment model (such as the analytic hierarchy process (AHP) or the fuzzy comprehensive evaluation model) to form a hybrid model architecture, thereby obtaining a formation deformation risk assessment model.

[0040] In this embodiment, a dynamic weight distribution network is established based on historical construction data and trained, including: Divide historical multi-source detection data into time-series subsets and construct a dynamic weight distribution network, which includes an input layer, an LSTM time-series encoding layer, and an attention weight distribution layer; The time series subset is input into the dynamic weight allocation network, and the historical stratum deformation data is used as the supervision label. The dynamic weight allocation network is trained through the loss function until the loss function converges to obtain the trained dynamic weight allocation network.

[0041] Extract time series from historical multi-source detection data (e.g., at a frequency of 1 hour), cut them into time series subsets according to fixed time windows (e.g., one subset per day), and construct a dynamic weight allocation network. Specifically, the construction process of the dynamic weight allocation network includes: input layer: receiving time series of multi-source detection data (e.g., surface settlement sequence St, building settlement sequence Wt, building tilt angle sequence Tt, groundwater level change sequence Ht), with the dimension of [time series length, feature number] (e.g., [100, 3]); LSTM time series encoding layer: using single-layer or multi-layer LSTM units (e.g., 128 neurons) to encode the time dependency of the input sequence and output the hidden state sequence h t ), capturing the temporal evolution of stratum deformation; Attention weight allocation layer: Based on the hidden state of LSTM output, the weight of the features at each moment is calculated through the scaled dot product attention mechanism, where the formula is: q is a learnable query vector, and the final output is the dynamic weight w = [ω1, ω2, ω3, ω4] (corresponding to the weights of surface settlement, building settlement, building tilt angle, and groundwater level change, respectively).

[0042] Furthermore, the time series subset is input into the dynamic weight allocation network, which outputs the dynamic weight w of each indicator, and then maps the historical formation deformation data into supervision labels. The mean square error (MSE) loss can be used: in, is the weighted predicted deformation, d i is the true deformation value. Furthermore, the network parameters are updated using a backpropagation algorithm (e.g., Adam optimizer, learning rate 0.001), and the loss value is verified every 100 batches. When the loss value no longer decreases after five consecutive epochs, the loss function is considered to have converged. At this point, the trained dynamic weight allocation network is obtained.

[0043] In this embodiment, a formation deformation risk assessment model is obtained based on the trained dynamic weight distribution network, including: Build a risk assessment model, which includes convolutional layers, Transformer encoding layers, and fully connected layers; Based on historical construction data, the risk assessment model is trained to obtain the target risk assessment model; Based on the trained dynamic weight distribution network and the target risk assessment model, a formation deformation risk assessment model is obtained.

[0044] Specifically, a neural network model can be built through a deep learning framework. More specifically, several convolution kernels (such as 3×3 size) are set to obtain a convolution layer, which can automatically extract the spatial features of the input data (such as the spatial distribution of monitoring points); a multi-head attention mechanism is introduced to obtain a Transformer encoding layer, which can capture the long-distance dependencies between different monitoring indicators in time series data (such as the correlation between surface subsidence and groundwater level changes); features are mapped to risk level values (such as 1-5 risk levels) through multiple layers of neurons (such as two layers) to obtain a fully connected layer.

[0045] Furthermore, the electronic device divides the historical construction data into a training set (80%) and a validation set (20%), and trains the model according to the following process: the historical multi-source detection data of the training set (such as surface subsidence, groundwater level, etc.) is input into the risk assessment model, and the risk level prediction value is output; the historical stratum deformation data (such as actual crack width, structural displacement, etc.) is used as the supervision label, and the loss between the predicted value and the label (such as mean square error or cross entropy) is calculated; the model parameters are adjusted through the back propagation algorithm (such as Adam optimizer), and the training is iterated until the validation set loss no longer decreases, and the parameters are saved as the target risk assessment model.

[0046] Furthermore, in the model inference stage, a dynamic weight distribution network is first used to assign real-time weights to surface settlement, building inclination angle, and groundwater level change values according to the geological parameters of the current construction stage (such as soil type and permeability coefficient) and construction parameters (such as excavation speed and support method) (such as 40% for surface settlement and 30% for inclination angle); the real-time weights are multiplied by the preprocessed multi-source detection data to generate a weighted feature vector; the weighted feature vector is input into the target risk assessment model, and after calculation through the convolution layer, Transformer layer and fully connected layer, the final risk level signal (such as "medium risk" or a specific risk value) is output.

[0047] In this embodiment, the standardized time series data set is input into the formation deformation risk assessment model to obtain a risk level signal, including: Input the standardized time series data into the ground deformation risk assessment model and obtain the dynamic weights of the output surface settlement, building settlement, building tilt angle and groundwater level change; The weighted monitoring data is input into the target risk assessment model, and the deformation probability value output by the target risk assessment model is obtained; based on the deformation probability value, the risk level signal corresponding to the current construction area is determined.

[0048] The trained dynamic weight allocation network (embedded in the stratum deformation risk assessment model) is called, and then the preprocessed standardized time series data set (including time series data such as surface settlement, building inclination angle, and groundwater level change) is input into the dynamic weight allocation network; the dynamic weight allocation network automatically calculates the real-time weight of each monitoring indicator (such as a weight of 0.4 for surface settlement, 0.1 for building settlement, 0.3 for building inclination angle, and 0.2 for groundwater level change) based on the geological parameters (such as soil type and permeability coefficient) and construction parameters (such as excavation depth and support strength) corresponding to the current input data; and extracts and saves the dynamic weight value of each indicator.

[0049] Furthermore, each monitoring indicator in the standardized time series data (surface subsidence, building subsidence, building tilt angle, and groundwater level change) is multiplied by the dynamic weights obtained in the first step to generate a weighted feature vector (e.g., surface subsidence * 0.4 + building subsidence * 0.1 + building tilt angle * 0.3 + groundwater level change * 0.2). This weighted feature vector is then input into the trained target risk assessment model (consisting of a convolutional layer, a Transformer encoding layer, and a fully connected layer). The convolutional layer extracts spatial features (e.g., abnormal patterns in the spatial distribution of monitoring points); the Transformer encoding layer captures long-term temporal dependencies between multiple indicators through a self-attention mechanism (e.g., the lagged correlation between groundwater level decline and surface subsidence); and the fully connected layer maps features into deformation probability values (e.g., 0.8 indicates an 80% probability of ground deformation).

[0050] Furthermore, a risk level threshold interval is preset (such as: low risk: deformation probability value <0.3; medium risk: 0.3≤deformation probability value <0.7; high risk: deformation probability value ≥0.7); the deformation probability value output in the second step is compared with the threshold and mapped to the corresponding risk level signal (such as "low risk", "medium risk" and "high risk").

[0051] Step S104: Determine whether to generate an early warning instruction based on the risk level signal.

[0052] Specifically, after receiving the risk level signal, it can be determined whether to generate an early warning instruction based on the risk level signal. If the risk level signal exceeds the yellow warning threshold (e.g., a risk score ≥ 60), a yellow warning is triggered, and an intermittent alarm is sounded through the sound and light alarm in the construction area, and a text message containing the risk location and level is sent to the on-site management personnel; if it exceeds the red warning threshold (e.g., a risk score ≥ 85), a red warning is triggered, and the emergency response procedure is automatically initiated, including suspending the shield machine excavation, cutting off the power supply to the high-risk area, and sending a risk report with pictures and text to the owner and supervision unit through the construction management platform.

[0053] Specifically, in this embodiment, determining whether to generate a warning instruction based on the risk level signal includes: Determining whether the risk level signal reaches a preset level threshold; If the risk equals that the signal has not reached the preset level threshold, no warning instruction will be generated; If the risk level signal reaches the preset level threshold, an early warning instruction is generated, and the construction status of the construction equipment is controlled based on the early warning instruction, and the construction status is speed reduction or shutdown.

[0054] The obtained risk level signal is compared with the threshold level threshold to determine whether the risk level signal of the current construction area reaches the preset level threshold. If the risk level signal reaches the preset level threshold, it is determined as "no warning requirement". At this time, the current monitoring state is maintained, no warning information is sent to the construction personnel or equipment, and new data continues to be collected and analyzed in real time. If the risk level signal reaches the preset level threshold, the preset warning template is matched according to the risk level (such as "medium risk" or "high risk") to generate instruction content (such as "medium risk: please reduce construction speed" and "high risk: stop work immediately"). Among them, the instruction includes control parameters (such as construction equipment operating speed threshold, shutdown signal coding). Instruction output: send visual / sound warnings to on-site personnel through the sound and light alarm system; send instructions to the construction equipment control system through the industrial bus (such as Modbus, CAN bus) or the Internet of Things interface (such as MQTT protocol) to adjust the equipment status. Specifically, speed reduction: modify the equipment operating speed parameter to below the safety threshold (such as the excavator excavation speed from 5m 3 / h down to 3m 3 / h); Shutdown: Send an emergency stop signal, cut off the power source of the equipment and lock the operating authority.

[0055] An embodiment of the present application provides a method for early warning of stratum deformation risk in deep subway tunnel construction. By real-time collection of multi-source detection data such as surface settlement, building settlement, building inclination angle, and groundwater level change, and preprocessing to generate a standardized time series data set, a stratum deformation risk assessment model is used to determine the risk level. This method can accurately and timely reflect the stratum deformation risk status of the current construction area, and then quickly determine whether to generate an early warning instruction based on the risk level signal, thereby achieving effective early warning of stratum deformation risks in deep subway tunnel construction, improving the accuracy of risk warnings, and helping to take preventive measures in advance to ensure construction safety, thereby reducing potential losses caused by stratum deformation.

[0056] The above embodiment introduces a method for warning of ground deformation risk in deep subway tunnel construction from the perspective of method flow. The following embodiment introduces a device for warning of ground deformation risk in deep subway tunnel construction from the perspective of a virtual module or virtual unit. Please see the following embodiment for details.

[0057] See also Figure 2 The deep subway tunnel construction ground deformation risk warning device 20 may specifically include: a collection module 201, a pre-processing module 202, an input module 203 and a determination module 204, wherein: A ground deformation risk early warning device 20 for deep subway tunnel construction, comprising: The acquisition module 201 is used to collect multi-source detection data corresponding to the current construction area in real time. The multi-source detection data includes surface settlement, building settlement, building tilt angle, and groundwater level change; A preprocessing module 202 is used to preprocess the multi-source detection data to generate a standardized time series data set; An input module 203 is used to input the standardized time series data set into the formation deformation risk assessment model to obtain a risk level signal; The determination module 204 is configured to determine whether to generate an early warning instruction based on the risk level signal.

[0058] In one possible implementation of the embodiment of the present application, the preprocessing module 202 is specifically configured to: Convert the building tilt angle from radians to degrees to obtain the target building tilt angle; The groundwater level change value is corrected to obtain the target groundwater level change value. The correction formula is H 目标 =H 初始 +K(P 采集 -P 基准 ), where H 目标 is the target groundwater level change value, H 初始is the groundwater level change value, K is the geological permeability correction factor, P 采集 is the collected groundwater pressure value, P 基准 Calibrate the pressure reference value for groundwater; The segmented normalization strategy is used to normalize the target detection building tilt angle and the target groundwater level change value to obtain the processed target detection building tilt angle and the target groundwater level change value; The dimensions of surface settlement and building settlement are unified to obtain target surface settlement and target building settlement. Based on the processed target detection building inclination angle, target groundwater level change value, target surface settlement and target building settlement, a standardized time series data set is generated.

[0059] In a possible implementation of the embodiment of the present application, the deep subway tunnel construction stratum deformation risk warning device 20 further includes: An acquisition module is used to obtain historical construction data corresponding to the current construction area. The historical construction data includes historical multi-source detection data and historical stratum deformation data; A training module is used to train a dynamic weight allocation network based on historical construction data. The dynamic weight allocation network assigns real-time weights to surface settlement, building inclination angle, and groundwater level change according to the geological parameters and construction parameters of the current construction stage. A module is obtained, which is used to obtain a stratum deformation risk assessment model based on the trained dynamic weight distribution network.

[0060] In one possible implementation of the embodiment of the present application, when the training module trains the dynamic weight distribution network based on historical construction data, it is specifically configured to: Divide historical multi-source detection data into time-series subsets and construct a dynamic weight distribution network, which includes an input layer, an LSTM time-series encoding layer, and an attention weight distribution layer; The time series subset is input into the dynamic weight allocation network, and the historical stratum deformation data is used as the supervision label. The dynamic weight allocation network is trained through the loss function until the loss function converges to obtain the trained dynamic weight allocation network.

[0061] In a possible implementation of the embodiment of the present application, when the module obtains the stratum deformation risk assessment model based on the trained dynamic weight distribution network, it is specifically used to: Build a risk assessment model, which includes convolutional layers, Transformer encoding layers, and fully connected layers; Based on historical construction data, the risk assessment model is trained to obtain the target risk assessment model; Based on the trained dynamic weight distribution network and the target risk assessment model, a formation deformation risk assessment model is obtained.

[0062] In one possible implementation of the embodiment of the present application, when the input module 203 inputs the standardized time series data set into the formation deformation risk assessment model to obtain the risk level signal, it is specifically configured to: Input the standardized time series data into the ground deformation risk assessment model and obtain the dynamic weights of the output surface settlement, building settlement, building tilt angle and groundwater level change; The weighted monitoring data is input into the target risk assessment model, and the deformation probability value output by the target risk assessment model is obtained; based on the deformation probability value, the risk level signal corresponding to the current construction area is determined.

[0063] In one possible implementation of the embodiment of the present application, when determining whether to generate a warning instruction based on the risk level signal, the determination module 204 is specifically configured to: Determining whether the risk level signal reaches a preset level threshold; If the risk equals that the signal has not reached the preset level threshold, no warning instruction will be generated; If the risk level signal reaches the preset level threshold, an early warning instruction is generated, and the construction status of the construction equipment is controlled based on the early warning instruction, and the construction status is speed reduction or shutdown.

[0064] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0065] See also Figure 3 , the embodiment of the present application also introduces an electronic device from the perspective of a physical device, such as Figure 3 As shown, Figure 3 The electronic device 300 shown includes a processor 301 and a memory 303. The processor 301 and the memory 303 are connected, for example, via a bus 302. Optionally, the electronic device 300 may further include a transceiver 304. It should be noted that in actual applications, the number of transceivers 304 is not limited to one, and the structure of the electronic device 300 does not constitute a limitation on the embodiments of the present application.

[0066] The processor 301 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor 301 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, and the like.

[0067] The bus 302 may include a path for transmitting information between the above components. The bus 302 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus. The bus 302 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0068] The memory 303 may be a ROM (Read Only Memory) or other type of static storage device that can store static information and instructions, a RAM (Random Access Memory) or other type of dynamic storage device that can store information and instructions, or an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory) or other optical disk storage, optical disk storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto.

[0069] The memory 303 is used to store application code for executing the solution of the present application, and the execution is controlled by the processor 301. The processor 301 is used to execute the application code stored in the memory 303 to implement the content shown in the above method embodiment.

[0070] Among them, electronic devices include but are not limited to: mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc., and can also be servers, etc. Figure 3 The electronic device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0071] An embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer-readable storage medium is run on a computer, the computer can execute the corresponding contents of the aforementioned method embodiment.

[0072] It should be understood that although the steps in the flowcharts of the accompanying drawings are shown in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some of the steps in the flowcharts of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0073] The above are only some of the implementation methods of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for early warning of ground deformation risk during deep subway tunnel construction, characterized in that: include: Real-time collection of multi-source detection data corresponding to the current construction area, including surface settlement, building settlement, building tilt angle, and groundwater level change; Preprocessing the multi-source detection data to generate a standardized time series data set; Inputting the standardized time series data set into a formation deformation risk assessment model to obtain a risk level signal; Based on the risk level signal, it is determined whether to generate an early warning instruction.

2. The method for early warning of ground deformation risk during deep subway tunnel construction according to claim 1 is characterized in that: The preprocessing of the multi-source detection data to generate a standardized time series data set includes: Converting the building tilt angle into a radian-degree system to obtain a target building tilt angle; The groundwater level change value is corrected to obtain the target groundwater level change value, and the correction formula is H 目标 =H 初始 +K(P 采集 -P 基准 ), where H 目标 is the target groundwater level change value, H 初始 is the groundwater level change value, K is the geological permeability correction factor, P 采集 is the collected groundwater pressure value, P 基准 Calibrate the pressure reference value for groundwater; A segmented normalization strategy is adopted to perform normalization processing on the target detection building inclination angle and the target groundwater level change value to obtain processed target detection building inclination angle and target groundwater level change value; Unifying the dimensions of the ground surface settlement and the building settlement to obtain target ground surface settlement and target building settlement; Based on the processed target detection building inclination angle and target groundwater level change value, target surface settlement and target building settlement, a standardized time series dataset is generated.

3. The method for early warning of ground deformation risk during deep subway tunnel construction according to claim 1 is characterized in that: Before inputting the standardized time series data set into the formation deformation risk assessment model, the method further includes: Acquiring historical construction data corresponding to the current construction area, the historical construction data including historical multi-source detection data and historical stratum deformation data; Based on the historical construction data, a dynamic weight allocation network is trained. The dynamic weight allocation network assigns real-time weights to surface settlement, building inclination angle and groundwater level change according to the geological parameters and construction parameters of the current construction stage. Based on the trained dynamic weight allocation network, a stratum deformation risk assessment model is obtained.

4. The method for early warning of ground deformation risk during deep subway tunnel construction according to claim 3 is characterized in that: Based on the historical construction data, a dynamic weight allocation network is trained, including: Dividing the historical multi-source detection data into time series subsets and constructing a dynamic weight distribution network, wherein the dynamic weight distribution network includes an input layer, an LSTM time series encoding layer, and an attention weight distribution layer; The time series subset is input into the dynamic weight allocation network, and the historical stratum deformation data is used as a supervision label. The dynamic weight allocation network is trained through a loss function until the loss function converges to obtain a trained dynamic weight allocation network.

5. The method for early warning of ground deformation risk during deep subway tunnel construction according to claim 4 is characterized in that: The stratum deformation risk assessment model is obtained based on the trained dynamic weight distribution network, including: Constructing a risk assessment model, wherein the risk assessment model includes a convolutional layer, a Transformer encoding layer, and a fully connected layer; Based on the historical construction data, the risk assessment model is trained to obtain a target risk assessment model; Based on the trained dynamic weight distribution network and the target risk assessment model, a formation deformation risk assessment model is obtained.

6. The method for early warning of ground deformation risk during deep subway tunnel construction according to claim 5, characterized in that: The standardized time series data set is input into a formation deformation risk assessment model to obtain a risk level signal, including: Input the standardized time series data into the stratum deformation risk assessment model, and obtain the dynamic weights of the output surface settlement, building settlement, building inclination angle and groundwater level change value; Inputting the weighted monitoring data into the target risk assessment model, and obtaining a deformation probability value output by the target risk assessment model; A risk level signal corresponding to the current construction area is determined based on the deformation probability value.

7. The method for early warning of ground deformation risk during deep subway tunnel construction according to claim 1, characterized in that: The determining whether to generate an early warning instruction based on the risk level signal includes: Determining whether the risk level signal reaches a preset level threshold; If the risk equals that the signal does not reach the preset level threshold, no warning instruction is generated; If the risk level signal reaches the preset level threshold, a warning instruction is generated, and the construction state of the construction equipment is controlled based on the warning instruction, and the construction state is speed reduction or shutdown.

8. A ground deformation risk warning device for deep subway tunnel construction, characterized in that: include: The acquisition module is used to collect multi-source detection data corresponding to the current construction area in real time. The multi-source detection data includes surface settlement, building settlement, building inclination angle and groundwater level change; A preprocessing module, configured to preprocess the multi-source detection data to generate a standardized time series data set; An input module, configured to input the standardized time series data set into a formation deformation risk assessment model to obtain a risk level signal; A determination module is used to determine whether to generate an early warning instruction based on the risk level signal.

9. An electronic device, characterized in that: The electronic device includes: at least one processor; Memory; At least one application, wherein the at least one application is stored in a memory and configured to be executed by at least one processor, and the at least one application is configured to: execute the deep subway tunnel construction stratum deformation risk early warning method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed in a computer, the computer is caused to execute the method for early warning of ground deformation risk during deep subway tunnel construction according to any one of claims 1 to 7.

Images