Vertical shaft foundation pit inrush early warning method based on pore pressure monitoring and multi-data fusion

Abstract

The invention discloses a vertical shaft foundation pit inrush early warning method based on pore pressure monitoring and multi-data fusion, and the method comprises the following steps: carrying out the real-time monitoring of a plurality of monitoring sites at the bottom of a vertical shaft foundation pit, obtaining the pore pressure value of each monitoring site, and carrying out the preprocessing; when the pore pressure value reaches a preset accumulation threshold value, a reminding signal is sent out; the change rate of the pore pressure value is calculated in real time, and when the change rate is positive and continuously rises, it is judged that the phase is a pore pressure accumulation phase; and when in the pore pressure accumulation stage and the pore pressure value drops beyond the dynamically corrected pore pressure sudden drop threshold value within the set time, judging that the sudden surge occurrence moment occurs, and sending out an early warning signal. The method has the advantages of being high in real-time performance, high in early warning precision, easy and convenient to operate and the like, an excellent early warning effect is achieved through multi-parameter cooperative verification, the safety and stability of vertical shaft foundation pit construction can be effectively guaranteed, cost is low, popularization is easy, and the method is suitable for sudden surge monitoring and early warning of various vertical shaft foundation pit projects.

Description

Technical Field

[0001] The present invention relates to the technical field of geotechnical engineering monitoring, and in particular to a shaft foundation pit sudden burst early warning method based on pore pressure monitoring and multi-data fusion. Background Art

[0002] During vertical shaft excavation construction, sudden water inrush is a common and highly detrimental geological hazard. These inrushes are typically caused by the sudden release of pore water pressure in the soil at the bottom of the excavation, leading to soil instability, water and sand inrush, and in severe cases, pit collapse, resulting in casualties and property damage.

[0003] Current methods for predicting and warning sudden surge accidents have certain flaws. The empirical formula method, based on empirical formulas derived from historical engineering data, has limited applicability and low prediction accuracy. The geological exploration method acquires geological information through drilling and geophysical exploration, but this is costly and difficult to update in real time. The numerical simulation method uses methods such as finite element analysis to simulate the stress and seepage fields around the foundation pit, but the model establishment is complex, computationally intensive, and difficult to achieve real-time warning. The on-site monitoring method uses monitoring equipment to monitor parameters such as soil stress and pore water pressure in real time. However, existing technologies only use single pore pressure data changes to warn of sudden surges, without dynamically adjusting the pore pressure drop threshold based on multi-dimensional parameters such as soil displacement, groundwater level changes, and excavation rate. This results in a high rate of false alarms and delayed warnings. Summary of the Invention

[0004] To address the above technical issues, the present invention proposes a method for warning sudden surges in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion. This method addresses the problems of inaccurate calculation of critical conditions for sudden surges in different vertical shaft foundation pits, difficulty in predicting the timing of sudden surges, and inability to take timely emergency measures. By monitoring the accumulation and sudden drop characteristics of pore pressure values at the bottom of the foundation pit in real time, and integrating pore pressure monitoring with a collaborative mechanism analysis of multi-sensor application and multi-parameter dynamic correction thresholds, accurate early warning of sudden surge accidents is achieved, significantly improving warning accuracy. This method has the advantages of strong real-time performance, high accuracy, and ease of operation.

[0005] In order to achieve the above object, the technical solution of the present invention is as follows:

[0006] A method for early warning of sudden outbursts in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion includes the following steps:

[0007] Conduct real-time monitoring of several monitoring points at the bottom of the vertical shaft foundation pit, obtain the pore pressure value of each monitoring point and perform pre-processing;

[0008] When the pore pressure value reaches the preset accumulation threshold, a warning signal is issued;

[0009] The rate of change of the pore pressure value is calculated in real time. When the rate of change is positive and continues to rise, it is determined to be the pore pressure accumulation stage. When it is in the pore pressure accumulation stage and the pore pressure value drops within the set time and exceeds the dynamically corrected pore pressure sudden drop threshold, it is determined to be the moment of sudden surge and an early warning signal is issued.

[0010] Preferably, several monitoring sites are evenly arranged at the bottom of the shaft pit.

[0011] Preferably, a pressure sensor with wireless transmission function is used to monitor the monitoring site in real time, and the measurement range of the pressure sensor is 0 to 500 kPa, the accuracy is ±0.1 kPa, and the sampling frequency is 1 time / second.

[0012] Preferably, the pore pressure values are preprocessed, including low-pass filtering to remove high-frequency noise and smoothing to eliminate abnormal values.

[0013] Preferably, the preset accumulation threshold is set according to design specifications and engineering experience.

[0014] Preferably, the set time and pore pressure sudden drop threshold are dynamically adjusted by analyzing the correlation between soil layer type, excavation depth, soil displacement rate, groundwater level change rate and excavation rate and sudden surge accidents through a machine learning model.

[0015] Preferably, the reminder signal and the early warning signal are both sent to the relevant construction personnel through one or more combinations of an audible and visual alarm device, a text message notification, a mobile application, or a remote monitoring platform push.

[0016] Preferably, the method further comprises the following steps:

[0017] Obtain historical pore pressure values and sudden surge accident data to build a prediction data set;

[0018] Training a preset surge prediction model based on the prediction data set to obtain a trained surge prediction model;

[0019] Integrate the trained surge prediction model into the early warning system to predict the probability of surge;

[0020] When the predicted surge probability exceeds the preset threshold, the early warning system automatically issues a warning signal.

[0021] Preferably, model parameters are optimized during training by grid search and cross validation.

[0022] Preferably, evaluation indicators are determined, and the prediction performance of the surge prediction model is evaluated based on the evaluation indicators. The evaluation indicators include accuracy, precision, recall, and F1 score.

[0023] Based on the above technical solution, the beneficial effects of the present invention are as follows: the present invention solves the difficult problems of inaccurate calculation of critical conditions for sudden surges under different vertical shaft foundation pits, difficult prediction of sudden surge moments, and inability to take emergency measures in a timely manner. The present invention takes pore pressure monitoring as the core, monitors the changes in pore water pressure at the bottom of the foundation pit in real time, and dynamically corrects the pore pressure drop threshold value in combination with parameters such as soil displacement rate and groundwater level change rate, accurately determines the time of sudden surge, and solves the problems of poor geological adaptability of traditional early warning methods due to fixed thresholds and low early warning accuracy under different geological conditions, avoiding false alarms and missed alarms; the system realizes fully automated early warning through multi-parameter collaborative verification (pore pressure + displacement + water level), without manual intervention, and can be widely used in vertical shaft foundation pit projects under complex geological conditions such as subways, mines, and water conservancy; it is low in cost and does not require complex numerical simulation, is suitable for small and medium-sized projects, and has significant economy, wide applicability and good promotion value. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 The present invention is a flow chart of a method for early warning of sudden surge in a vertical shaft foundation pit based on pore pressure monitoring and multi-data fusion in one embodiment;

[0025] Figure 2 The present invention is a graph showing the trend of pore pressure value changes in a shaft foundation pit sudden burst warning method based on pore pressure monitoring and multi-data fusion in one embodiment. DETAILED DESCRIPTION

[0026] The technical solutions in the embodiments of the present invention will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present invention.

[0027] like Figure 1 、 2 As shown, this embodiment provides a method for warning sudden surges in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion. With pore pressure monitoring as the core, it monitors the changes in pore water pressure at the bottom of the foundation pit in real time, and dynamically corrects the set time and pore pressure sudden drop threshold value in combination with parameters such as soil displacement rate and groundwater level change rate, accurately determines the time of sudden surge, solves the problems of poor geological adaptability of traditional warning methods due to fixed thresholds and low warning accuracy under different geological conditions, and avoids false alarms and missed alarms; the system realizes fully automated warning through multi-parameter collaborative verification (pore pressure + displacement + water level) without manual intervention. The specific operation includes the following steps:

[0028] Step 100: Real-time monitoring of several monitoring points at the bottom of the vertical shaft foundation pit, obtaining the pore pressure value of each monitoring point and performing pre-processing;

[0029] Step 200: When the pore pressure value reaches a preset accumulation threshold, a reminder signal is issued;

[0030] Step 300: Calculate the rate of change of the pore pressure value in real time. When the rate of change is positive and continues to rise, it is determined to be the pore pressure accumulation stage. When it is in the pore pressure accumulation stage and the pore pressure value drops within the set time and exceeds the dynamically corrected pore pressure sudden drop threshold, it is determined to be the moment of sudden surge and an early warning signal is issued.

[0031] In this embodiment, multiple pressure sensors with wireless transmission capabilities are evenly distributed at the bottom of the vertical shaft foundation pit. Additional pressure sensors can be deployed at key locations based on the ground conditions of the vertical shaft foundation pit, such as locations with thinner aquicludes or poorer soil conditions. These pore pressure sensors monitor the pore water pressure at the bottom of the foundation pit in real time and transmit the pore pressure monitoring data to a central processing unit. The central processing unit uses the pore pressure monitoring data as a benchmark and synchronously receives auxiliary data from the soil displacement sensor and the groundwater level monitor for preprocessing to eliminate interference factors in the data; the central processing unit analyzes the changing trend of the pore pressure value in real time, identifies the accumulation characteristics of the pore pressure value, and when it reaches the preset accumulation threshold (such as 100kPa), it marks the time point as the pore pressure too high reminder moment and immediately issues a reminder signal; sets the pore pressure sudden drop threshold and performs dynamic correction at the same time. If the pore pressure value drops by more than the pore pressure sudden drop threshold (such as 50kPa) within the set time (such as within 10 seconds), it is determined to be the moment of sudden surge and an early warning signal is immediately issued (the system independently monitors the sudden drop characteristics of the pore pressure value. Even if the pore pressure does not reach the preset accumulation threshold, as long as the sudden drop amplitude and time meet the dynamic threshold conditions, the early warning signal will be triggered). See Figure 2 1 represents the preset accumulation threshold of the pore pressure value (e.g., 100 kPa). Upon reaching this threshold, a warning signal is issued, alerting relevant construction personnel to prepare for emergency response. The preset accumulation threshold is determined primarily based on design specifications and engineering experience. 2 represents the pore pressure accumulation stage during real-time monitoring. 3 represents the pore pressure sudden drop stage during real-time monitoring. During this stage, if the pore pressure value drops by more than the dynamically corrected pore pressure sudden drop threshold (e.g., 50 kPa) within a short period of time (e.g., 10 seconds), an early warning signal is immediately issued. The set time and pore pressure sudden drop threshold are dynamically adjusted by analyzing the correlation between soil layer type, excavation depth, soil displacement rate, groundwater level change rate, excavation rate, and sudden surge accidents using a machine learning model. 4 represents the node at which the pore pressure sudden drop stage is determined, at which an early warning signal is immediately issued.

[0032] Among them, 1) determining the set time: it is mainly determined based on the time of sudden drop in pore pressure due to sudden burst in the foundation pit according to the corresponding relationship determined in historical data. It will also vary depending on the soil type, excavation depth, soil displacement rate, groundwater level change rate, and the correlation between excavation rate and sudden burst accidents.

[0033] For a set time, the following features are extracted from historical data: the duration of the sudden drop in pore pressure when a sudden surge accident occurs (such as the interval from the beginning of the sudden drop to the occurrence of the accident); the distribution characteristics of the sudden drop time under soil layer type, excavation depth, soil displacement rate, groundwater level change rate and excavation rate; and the correlation with the sudden drop amplitude.

[0034] Time series classification models (such as LSTM, random forest time classifier) or statistical methods (such as percentile method) are used to analyze the critical time threshold of pore pressure drop in historical sudden surge accidents.

[0035] The trained "set time" model is integrated into the early warning system to dynamically adjust the time conditions for sudden drop judgment.

[0036] 2) Determine the pore pressure drop threshold: Dynamically modify the pore pressure drop threshold based on soil displacement rate, groundwater level change rate and excavation rate. The pore pressure drop threshold T d The calculation formula is as follows:

[0037] T d =T0(1+αD+βW+γE)

[0038] Where: T0 is the basic threshold, D is the soil displacement rate, W is the groundwater level change rate, E is the excavation rate, and α, β, and γ are correction coefficients.

[0039] The process of determining the correction coefficient includes the following steps: obtaining historical data, which includes soil layer type, excavation depth, soil displacement rate D, groundwater level change rate W, excavation rate E and corresponding sudden burst accident labels; first determining the engineering correspondence based on the soil layer type, excavation depth, sudden burst accident label, etc., and then feature encoding and normalizing the historical data, and dividing it into training set, validation set and test set; constructing a multivariate linear regression or random forest regression model, inputting the historical data into the multivariate linear regression or random forest regression model, outputting correction coefficients α, β, and γ, and optimizing the model parameters through grid search and cross-validation during the training process.

[0040] Both reminder signals and early warning signals can be issued at the construction site through sound and light alarm devices; both reminder signals and early warning signals can be pushed through SMS notifications, mobile applications or remote monitoring platforms, reminding construction personnel to monitor the reminder location in real time and take appropriate preventive measures.

[0041] Specifically, the pressure sensor uses a high-precision pressure sensor with a measurement range of 0 to 500 kPa, an accuracy of ±0.1 kPa, and a sampling frequency of 1 time per second. The pressure sensor is connected to a power supply and has wireless transmission capabilities. Before construction of the vertical shaft excavation, a soil sampler can be used to remove a long column of soil sample, which can then be placed back into the device.

[0042] In one embodiment of a shaft foundation pit sudden burst warning method based on pore pressure monitoring and multi-data fusion, a process for preprocessing the pore pressure value is also provided, which includes the following steps: the central processing unit performs low-pass filtering on the collected data to remove high-frequency noise; smoothes the filtered data to eliminate abnormal values; and stores the processed data in a database for subsequent analysis and warning.

[0043] In one embodiment, a method for early warning of sudden surge in a vertical shaft foundation pit based on pore pressure monitoring and multi-data fusion also provides an early warning optimization process, which includes the following steps:

[0044] Step 41: Collect historical data, pore pressure monitoring records, and sudden surge accident reminders and warning records of various vertical shaft foundation pit construction projects to construct a prediction data set;

[0045] Step 411: Clean the original data of historical construction, including pore pressure values, time points, soil layer types, excavation depths, groundwater levels, time of sudden surge accidents, pore pressure variation characteristics, and accident consequences, to remove noise and outliers, and normalize the data to ensure that data with different characteristics are at the same level.

[0046] Step 412: Divide the prediction data set into a training set, a validation set, and a test set for model training and evaluation;

[0047] Step 42: extracting characteristics such as the accumulation rate, sudden drop amplitude, and change trend of the pore pressure value;

[0048] Step 421: extract key features from the raw data, such as the accumulation rate of the pore pressure value, the sudden drop amplitude of the pore pressure value, the change trend of the pore pressure value (increase, decrease, fluctuation), and the combined features of the soil layer type and excavation depth;

[0049] Step 422: Use a feature selection algorithm (correlation-based feature selection, recursive feature elimination, etc.) to select the features that have the greatest impact on sudden surge warning.

[0050] Step 43: Use the random forest model to train the data, adjust the tree depth and number of nodes, and optimize the model performance;

[0051] Step 431: Select a machine learning model suitable for time series data and classification tasks. Random Forest can handle high-dimensional data and has good generalization ability.

[0052] Step 432: Use the training set data to train the preset surge prediction model and adjust hyperparameters (such as tree depth, number of nodes, etc.) to optimize model performance;

[0053] Step 433: Use the validation set data to evaluate the model performance to avoid overfitting;

[0054] Step 44: Evaluate the model on the test set, achieving an accuracy of 95% and a recall of 90%. Further optimize the model parameters to improve the model's prediction accuracy.

[0055] Step 441: Evaluate the performance of the model using metrics such as accuracy and recall. The accuracy should reach 95% and the recall should reach 90%. Focus on the recall to ensure that the model can identify as many sudden surge events as possible.

[0056] Step 442: Optimize model parameters through cross-validation, grid search, etc.

[0057] Step 443: Introducing ensemble learning methods (such as Boosting and Bagging) to further improve the prediction accuracy of the surge prediction model;

[0058] Step 45: Integrate the trained surge prediction model into the early warning system to monitor the pore pressure in real time and predict the surge risk. When the probability of surge exceeds 80%, the system automatically issues a warning signal.

[0059] Step 451: Integrate the trained machine learning model into the real-time warning system as the core algorithm for sudden surge warning;

[0060] Step 452: collecting pore pressure monitoring data in real time and inputting it into a surge prediction model for prediction;

[0061] Step 453: Determine whether to issue an early warning signal based on the surge probability output by the surge prediction module. When the surge probability exceeds 80%, the system automatically issues an early warning signal.

[0062] Step 46: Regularly retrain the model using new monitoring data to ensure that the surge prediction model can adapt to changes in the construction environment.

[0063] It should be understood that, although the various steps in the above flow chart are shown in sequence as indicated by the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be performed in other orders. Moreover, at least a portion of the steps in the above flow chart may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily to be performed in sequence, but can be performed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0064] The foregoing description is merely a preferred embodiment of the method for early warning of sudden surges in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion disclosed herein and is not intended to limit the scope of protection of the embodiments of this specification. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the embodiments of this specification shall be included within the scope of protection of the embodiments of this specification.

Claims

1. A method for early warning of sudden surge in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion, characterized by: The steps include: Conduct real-time monitoring of several monitoring points at the bottom of the vertical shaft foundation pit, obtain the pore pressure value of each monitoring point and perform pre-processing; When the pore pressure value reaches the preset accumulation threshold, a warning signal is issued; The rate of change of the pore pressure value is calculated in real time. When the rate of change is positive and continues to rise, it is determined to be the pore pressure accumulation stage. When it is in the pore pressure accumulation stage and the pore pressure value drops within the set time and exceeds the dynamically corrected pore pressure sudden drop threshold, it is determined to be the moment of sudden surge and an early warning signal is issued.

2. The method for early warning of sudden surge in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion according to claim 1 is characterized in that: Several monitoring points are evenly distributed at the bottom of the shaft pit.

3. The method for early warning of sudden surge in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion according to claim 1 is characterized in that: A pressure sensor with wireless transmission function is used to monitor the monitoring site in real time. The measurement range of the pressure sensor is 0-500kPa, the accuracy is ±0.1kPa, and the sampling frequency is 1 time / second.

4. The method for early warning of sudden surge in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion according to claim 1 is characterized in that: The pore pressure values are preprocessed, including low-pass filtering to remove high-frequency noise and smoothing to eliminate abnormal values.

5. The method for early warning of sudden surge in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion according to claim 1 is characterized in that: The preset accumulation threshold is set according to design specifications and engineering experience.

6. The method for early warning of sudden surge in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion according to claim 1 is characterized in that: The set time and pore pressure sudden drop threshold are dynamically adjusted by analyzing the correlation between soil layer type, excavation depth, soil displacement rate, groundwater level change rate and excavation rate and sudden surge accidents through a machine learning model.

7. The method for early warning of sudden surge in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion according to claim 1 is characterized in that: The reminder signal and the early warning signal are both used to remind relevant construction personnel through one or more combinations of sound and light alarm devices, SMS notifications, mobile applications or remote monitoring platform push.

8. The method for early warning of sudden surge in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion according to claim 1 is characterized in that: The following steps are also included: Obtain historical pore pressure values and sudden surge accident data to build a prediction data set; Training a preset surge prediction model based on the prediction data set to obtain a trained surge prediction model; Integrate the trained surge prediction model into the early warning system to predict the probability of surge; When the predicted surge probability exceeds the preset threshold, the early warning system automatically issues a warning signal.

9. The method for early warning of sudden surge in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion according to claim 8 is characterized in that: During the training process, model parameters are optimized through grid search and cross-validation.

10. The method for early warning of sudden surge in vertical shaft foundation pits based on pore pressure monitoring and multi-data fusion according to claim 8, characterized in that: Determine evaluation indicators and evaluate the prediction performance of the surge prediction model based on the evaluation indicators. The evaluation indicators include accuracy, precision, recall rate, and F1 score.

Images