A method and system for optimizing disturbance parameters of shaft excavation based on big data

Abstract

The present application relates to the vertical shaft tunneling technical field, disclose a kind of vertical shaft tunneling disturbance parameter optimization method and system based on big data, by obtaining multiple information streams and evaluating its correlation degree to determine the cognitive degree of environment state, obtain environment response information and compare with pre-stored environment characteristic mode to determine the matching degree of environment characteristic, and the range of statistical information stream fluctuation and the frequency of abnormal event occurrence determine the degree of environment change intensity, comprehensive output a quantitative environment cognitive index.On this basis, according to the mode switching of environment cognitive index, dynamically adjust between the operation mode of pursuing efficiency and the exploration mode of giving priority to guaranteeing safety and actively acquiring information.The method effectively solves the problem that the existing technology under complex geological conditions, due to the failure of data analysis logic, the system makes wrong judgment and parameter adjustment instruction, and then deteriorates the downhole working condition, it is difficult to accurately understand the current environment and re-establish effective optimization strategy.

Description

Technical Field

[0001] This invention relates to the field of shaft excavation technology, and in particular to a method and system for optimizing shaft excavation disturbance parameters based on big data. Background Technology

[0002] In modern deep shaft excavation operations, advanced mechanized full-face tunneling systems are widely used. These systems are equipped with comprehensive sensor networks designed to intelligently optimize tunneling parameters. They continuously monitor various operational data, such as cutterhead speed, thrust, torque, main drive system vibration, drilling fluid density, viscosity, temperature, as well as annular pressure and flow rate. Initially, these systems perform exceptionally well in homogeneous and stable geological formations, establishing reliable correlations between machine parameters and geological responses. This allows the system to recommend optimized parameters to maximize tunneling efficiency and minimize energy consumption.

[0003] However, when tunneling systems encounter complex geological anomalies, such as highly fractured fault zones, large influxes of groundwater, and intermittent cavities, existing data analysis logic, primarily based on experience under stable geological conditions, often fails to accurately interpret drastic changes in sensor signals. This can lead to erroneous judgments and incorrect parameter adjustment commands, which in turn further worsen downhole conditions, resulting in a complex data stream that mixes the actual geological response with the consequences of erroneous interventions. In such situations, the system struggles to accurately understand the current environment and is unable to re-establish effective optimization strategies, posing a severe challenge to the safety and efficiency of tunneling operations.

[0004] Specifically, when a tunneling system abruptly transitions from a stable geological environment to a complex fault zone characterized by highly fractured rock, massive groundwater inflow, and intermittent cavities, its existing analytical logic becomes ineffective due to the drastic change in geological conditions. Furthermore, when the analytical logic fails and erroneous parameter adjustment commands have been issued, and these commands have been executed, further deteriorating downhole conditions and generating a complex information flow that mixes the true formation response with the consequences of erroneous system intervention, existing technologies struggle to effectively distinguish and eliminate the interfering information caused by erroneous intervention. Consequently, they are unable to quickly and re-establish a reliable parameter optimization strategy suitable for the new geological environment based solely on information reflecting the true characteristics of the current formation, thus failing to prevent further deterioration and guide subsequent correct operations. Simultaneously, existing technologies also struggle to ensure that during strategy reconstruction, the system can identify and avoid repeating previous erroneous intervention patterns that worsened the situation, thereby failing to achieve robust adaptation and optimization to complex and dynamic geological anomalies.

[0005] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0006] This invention provides a method for optimizing disturbance parameters in vertical shaft excavation based on big data. It aims to solve the problem that in existing deep vertical shaft excavation operations, when the excavation system encounters complex geological anomalies, the existing data analysis logic cannot accurately interpret drastic changes in sensor signals, leading to incorrect judgments and parameter adjustment instructions from the system, which in turn worsens the underground working conditions. It is difficult to accurately understand the current environment and re-establish an effective optimization strategy, posing a serious challenge to the safety and efficiency of the excavation operation.

[0007] The technical solution of this application is as follows:

[0008] In a first aspect, this application discloses a method for optimizing disturbance parameters in vertical shaft excavation based on big data, comprising the following steps:

[0009] Acquire multiple information streams and assess the degree of correlation between them to determine the level of awareness of the environmental state;

[0010] Acquire environmental response information and compare it with pre-stored environmental feature patterns to determine the degree of matching of environmental features;

[0011] The fluctuation range of statistical information flow and the frequency of abnormal events are used to determine the severity of environmental changes;

[0012] Based on the level of awareness of the environmental state, the degree of matching of environmental characteristics, and the severity of environmental changes, a quantitative environmental awareness index is output.

[0013] The mode is switched based on the environmental awareness index, and the modes include an efficiency-oriented operation mode and an exploration mode that prioritizes safety and actively acquires information.

[0014] Through this technical solution, this application can effectively integrate multi-source information, quantify the degree of environmental awareness, and dynamically adjust the tunneling mode accordingly, thereby achieving intelligent optimization of tunneling parameters under complex geological conditions, avoiding erroneous interventions caused by geological anomalies, and improving the safety and efficiency of tunneling operations.

[0015] Furthermore, after switching modes based on the environmental awareness index, the method also includes:

[0016] When in exploration mode, the limit range of operation parameters is generated based on the currently identified environmental features and the recorded restriction operation information;

[0017] Within the limits of the operating parameters, perform parameter adjustment operations and obtain environmental response information;

[0018] Record parameter adjustment operations and execution times, and based on the expected impact of parameter adjustment operations on equipment and environment, separate the signal components corresponding to parameter adjustment operations from environmental response information to obtain information reflecting the true characteristics of the environment;

[0019] Information reflecting the true characteristics of the environment is compared with pre-stored environmental feature patterns to identify current environmental characteristics and estimate environmental parameters.

[0020] Monitor the results of parameter adjustment operations. When a parameter adjustment operation leads to a deterioration of the environmental condition, record the parameter adjustment operation and the corresponding environmental conditions, and add the parameter adjustment operation to the restriction operation information.

[0021] Through this technical solution, this application can more cautiously adjust parameters in exploratory mode, and effectively avoid repeating mistakes by separating real environmental response signals, identifying environmental characteristics and estimating parameters, and recording operations that lead to deterioration, thereby gradually establishing a reliable optimization strategy in an uncertain environment.

[0022] More specifically, in some implementation schemes, the parameter adjustment operation and its execution time are recorded, and based on the expected impact of the parameter adjustment operation on the equipment and environment, the signal components corresponding to the parameter adjustment operation are separated from the environmental response information to obtain information reflecting the true characteristics of the environment, including:

[0023] Collect key sensor data from the tunneling machine and perform noise suppression and timestamp calibration on the key sensor data;

[0024] For each small-amplitude parameter adjustment operation, a causal response model is established, and the system response curve that should be generated on each sensor is predicted in real time based on the operation content and execution time.

[0025] The system response curve is separated from the key sensor data to obtain the preliminary formation response signal;

[0026] High-resolution time-frequency analysis was performed on the preliminary formation response signal to obtain the energy distribution of the preliminary formation response signal at different time points and frequencies;

[0027] Transient patterns are extracted from energy distribution and compared with a predefined geological event feature library to instantly identify geological micro-event types;

[0028] Geological event parameters are estimated based on the types of geological micro-events, and the types and parameters of geological micro-events are integrated into the current working condition description in order to adjust the uncertainty index.

[0029] Through this technical solution, this application can accurately separate the true response signal of the formation by means of refined signal processing and causal response model, and further identify the types and parameters of geological micro-events, thereby significantly improving the accuracy of understanding the true characteristics of the formation and providing a more reliable basis for subsequent parameter optimization.

[0030] Preferably, the result of parameter adjustment operations is monitored. When the parameter adjustment operation leads to a deterioration of the environmental condition, the parameter adjustment operation and the corresponding environmental conditions are recorded, and the parameter adjustment operation is added to the limiting operation information, including:

[0031] The sequence of operations prior to the environmental degradation was traced back, including system commands, command execution times, tunneling machine parameters, and mud circulation system status.

[0032] Retrieve sensor trust weights and calibration information during the execution of the operation sequence;

[0033] Based on the system's instructions, the execution time of the instructions, the tunneling machine's equipment parameters, the status of the mud circulation system, and the sensor trust weights and calibration information, the expected changes in sensor data should be recalculated under the condition of no internal sensor bias and only considering the system's own intervention.

[0034] Compare sensor data during the actual monitored deterioration period with the recalculated expected changes;

[0035] Based on the comparison results, identify the expected secondary effects caused by the system's own intervention during the deterioration process, the perception distortion caused by the internal bias of the sensor, and the anomalies caused by changes in the formation characteristics;

[0036] Based on the identification results, the main causes of the environmental degradation were determined;

[0037] When the primary cause is a change in formation characteristics that exceeds the system's adaptability, parameter adjustment operations and actual formation characteristic information are added to the constraint operation information.

[0038] Through this technical solution, this application can accurately identify the root causes of environmental degradation by retrospective analysis and multi-factor comparison, distinguish the effects of system intervention, sensor bias and changes in formation characteristics, thereby avoiding attributing system errors to the formation and effectively accumulating information that limits operation to prevent repeated mistakes.

[0039] Based on the above, this application further proposes a mode switching method based on the environmental awareness index, the steps of which include:

[0040] Real-time monitoring of multiple key information flows and continuous calculation of the rate of change of information flows within a short time window;

[0041] Set an early warning threshold for the rate of change;

[0042] When the rate of change exceeds the warning threshold, an emergency response signal is triggered.

[0043] When an emergency response signal is received, the system is forced to switch from efficiency-driven operation mode to exploration mode.

[0044] Through this technical solution, this application can achieve rapid response to sudden environmental changes or anomalies by real-time monitoring of the rate of change of information flow and setting early warning thresholds, and force a switch to exploration mode, thereby prioritizing safety and proactively acquiring information in emergency situations, effectively avoiding potential risks.

[0045] In one implementation, when parameter adjustments are made at the edge of a cavity, the parameter adjustment operation and execution time are recorded. Based on the expected impact of the parameter adjustment operation on the equipment and the environment, the signal components corresponding to the parameter adjustment operation are separated from the environmental response information to obtain information reflecting the true characteristics of the environment, including:

[0046] Real-time acquisition of key sensor data on cutter head torque, vibration, and annular pressure;

[0047] High-resolution time-frequency analysis is performed on key sensor data to generate time-frequency graphs;

[0048] In the time-frequency diagram, identify the signal patterns caused by the tunneling machine's own operation. The signal patterns include the delayed response, energy distribution, and duration after the operation.

[0049] In the time-frequency plot, identify transient signal patterns caused by micro-collapse or stress release of rock mass at the cavity edge. Transient signal patterns include energy surges, frequency shifts, and waveform shapes.

[0050] By comparing the time-frequency characteristics of the signal pattern and the transient signal pattern, the signal components generated by the tunneling machine's own operation are separated from the transient signal pattern;

[0051] The separated transient signal patterns are used as information reflecting the true characteristics of the environment.

[0052] Through this technical solution, this application can accurately distinguish between the tunneling machine's own operating signals and transient signals caused by micro-collapse or stress release of the rock mass at the cavity edge, based on the special geological conditions at the cavity edge, using high-resolution time-frequency analysis technology, thereby obtaining purer and more realistic information on the environmental characteristics of the cavity edge.

[0053] As an optional approach, the parameter adjustment operation and its execution time are recorded. Based on the expected impact of the parameter adjustment operation on the equipment and environment, the signal components corresponding to the parameter adjustment operation are separated from the environmental response information to obtain information reflecting the true characteristics of the environment, including:

[0054] The system collects data from multiple sensors in the tunnel boring machine cutterhead, propulsion system, and mud circulation system in real time. The sensor data includes vibration signals, pressure signals, torque signals, and flow signals.

[0055] Based on the type and intensity of the parameter adjustment operation, as well as the current status of the tunneling machine and the tunneling speed, calculate the instantaneous force distribution of the parameter adjustment operation on the contact area between the tunneling machine cutterhead and the stratum.

[0056] Based on the instantaneous force distribution, combined with the current acoustic propagation velocity, attenuation coefficient, and reflection coefficient of the formation medium, the time-domain response waveform that should be generated on multiple sensors after the force propagates in the formation is predicted.

[0057] The amplitude and phase of the time-domain response waveform are adjusted according to the water content, porosity, and degree of consolidation of the formation medium.

[0058] An adaptive separation method based on signal feature matching separates the adjusted predicted time-domain response waveform from real-time sensor data to obtain information reflecting the true characteristics of the formation.

[0059] Through this technical solution, this application can predict the response of parameter adjustment operations in different formation media by establishing a refined physical model, and accurately extract the true characteristics of the formation from complex sensor data by using an adaptive signal separation method, thereby improving the understanding and prediction capabilities of formation response.

[0060] Building upon the above, this application further proposes a mode-switching mechanism based on the environmental awareness index, including:

[0061] Real-time acquisition of environmental awareness index;

[0062] Based on the numerical range of the environmental awareness index, it is divided into an efficiency zone, an exploration zone, and a transition zone between the efficiency zone and the exploration zone.

[0063] A hysteresis interval for mode switching is set in the transition zone. The hysteresis interval has an upper threshold for switching from operation mode to exploration mode and a lower threshold for switching from exploration mode to operation mode. The upper threshold is higher than the lower threshold.

[0064] By introducing a hysteresis interval, this application effectively avoids the problem of frequent mode switching when the environmental perception index fluctuates slightly around the threshold, thereby improving the stability and reliability of the system operation.

[0065] Furthermore, mode switching is performed based on the environmental awareness index, including:

[0066] When the environmental awareness index gradually increases from the efficiency zone and reaches the upper limit threshold for the first time, the system switches from operation mode to exploration mode.

[0067] When the environmental awareness index gradually decreases from the exploration zone and reaches the lower limit threshold for the first time, the mode is switched from exploration mode to operation mode.

[0068] Within the hysteresis interval, when the environmental awareness index fluctuates between the upper and lower thresholds, the current operating mode remains unchanged.

[0069] Through this technical solution, this application ensures the smoothness and rationality of mode switching by establishing clear switching logic and hysteresis mechanism, avoiding unnecessary mode switching caused by instantaneous fluctuations in the environmental cognition index, and further improving the robustness of the system.

[0070] Secondly, this application also discloses a shaft excavation disturbance parameter optimization system based on big data, comprising:

[0071] The input end is used to acquire multiple information streams and evaluate the degree of correlation between the information streams to determine the level of awareness of the environmental state; it also acquires environmental response information and compares the environmental response information with pre-stored environmental feature patterns to determine the degree of matching of environmental features.

[0072] The statistics section is used to analyze the fluctuation range of information flow and the frequency of abnormal events in order to determine the severity of environmental changes.

[0073] The output end is used to output a quantitative environmental cognition index based on the level of awareness of the environmental state, the degree of matching of environmental characteristics, and the severity of environmental changes; the mode is switched according to the environmental cognition index, including an operation mode that pursues efficiency and an exploration mode that prioritizes safety and actively acquires information.

[0074] This application provides a system that can implement the above-mentioned method through this technical solution. Through modular design, it effectively integrates functions such as information acquisition, environmental assessment, index output and mode switching, providing an efficient and intelligent hardware platform for optimizing vertical shaft excavation disturbance parameters.

[0075] Beneficial effects

[0076] This application discloses a big data-based method for optimizing disturbance parameters in vertical shaft excavation. It determines the level of environmental awareness by acquiring multiple information streams and evaluating their correlation, acquiring environmental response information and comparing it with pre-stored environmental characteristic patterns to determine the degree of matching of environmental characteristics, and statistically analyzing the fluctuation range of information streams and the frequency of abnormal events to determine the severity of environmental changes. A quantitative environmental awareness index is then output. Based on this, a mode switching mechanism is implemented according to the environmental awareness index, dynamically adjusting between an efficiency-driven operation mode and an exploration mode that prioritizes safety and proactively acquires information. This method effectively solves the problem in existing technologies where, under complex geological conditions, data analysis logic failures lead to erroneous judgments and parameter adjustment commands, thus worsening downhole conditions and making it difficult to accurately understand the current environment and re-establish effective optimization strategies. By quantifying the environmental awareness index and switching modes, this application can flexibly adjust the tunneling strategy according to the complexity and uncertainty of the environment. When the environmental awareness is high and the changes are stable, efficiency is pursued, while when the environmental awareness is low and the changes are drastic, safety is prioritized and active exploration is carried out. This avoids the negative impact caused by blind intervention in complex geological anomalies, effectively distinguishes and removes interference information caused by erroneous intervention, and re-establishes the optimization strategy based only on information reflecting the true characteristics of the strata. This significantly improves the safety, efficiency and adaptability of tunneling operations to complex geological environments. Attached Figure Description

[0077] Figure 1 This is a flowchart of a method for optimizing vertical shaft excavation disturbance parameters based on big data, provided by an embodiment of the present invention.

[0078] Figure 2 This is a schematic diagram of a vertical shaft excavation disturbance parameter optimization system based on big data, provided in an embodiment of the present invention. Detailed Implementation

[0079] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0080] Reference Figure 1 , Figure 1 This is a flowchart of a method for optimizing vertical shaft tunneling disturbance parameters based on big data, provided by an embodiment of the present invention, including:

[0081] S11, acquire multiple information streams and evaluate the degree of correlation between the information streams to determine the level of awareness of the environmental state;

[0082] S12, acquire environmental response information and compare the environmental response information with pre-stored environmental feature patterns to determine the degree of matching of environmental features;

[0083] S13, Statistically analyze the fluctuation range of the information flow and the frequency of abnormal events to determine the severity of environmental changes;

[0084] S14. Based on the degree of awareness of the environmental state, the degree of matching of the environmental features, and the severity of the environmental change, output a quantitative environmental awareness index.

[0085] S15, switch modes according to the environmental cognition index, the modes include an operation mode that pursues efficiency and an exploration mode that prioritizes safety and actively acquires information.

[0086] In modern deep shaft excavation operations, although advanced mechanized full-face tunneling systems are equipped with comprehensive sensor networks and can intelligently optimize tunneling parameters in homogeneous strata, existing systems often fail when encountering complex geological anomalies due to limitations in their data analysis logic. This makes it difficult for the system to accurately interpret drastic changes in sensor signals, leading to erroneous parameter adjustment commands, further deteriorating downhole conditions, and creating a complex data stream that mixes real geological responses with the consequences of erroneous interventions. In this situation, the system struggles to accurately understand the current environment and cannot re-establish effective optimization strategies, posing a severe challenge to the safety and efficiency of tunneling operations.

[0087] To address this issue, this application proposes a big data-based method for optimizing disturbance parameters in vertical shaft excavation. This method involves acquiring multiple information streams and evaluating their correlation to determine the level of environmental awareness; acquiring environmental response information and comparing it with pre-stored environmental characteristic patterns to determine the degree of matching of environmental characteristics; and statistically analyzing the fluctuation range of information streams and the frequency of abnormal events to determine the severity of environmental changes. Based on the aforementioned level of awareness, matching degree, and severity, a quantified environmental awareness index is output, and mode switching is performed according to this index. The modes include an efficiency-oriented operation mode and an exploration mode that prioritizes safety and actively acquires information. This application aims to solve the problem that existing technologies, under complex geological conditions, cannot accurately distinguish between the actual response of the strata and the interference information generated by erroneous system interventions, thus hindering the rapid reconstruction of reliable parameter optimization strategies. This approach aims to achieve robust adaptation and optimization to complex and dynamic geological anomalies.

[0088] To better understand the method proposed in this application, the key terms involved will first be explained.

[0089] "Information flow" refers to the data sequence collected and transmitted in real time by various sensors and monitoring equipment during shaft excavation. This information flow can include, but is not limited to, cutterhead rotation speed, propulsion force, torque, main drive system vibration, drilling fluid density, viscosity, temperature, annular pressure, flow velocity, formation acoustic waves, seismic waves, and electromagnetic waves. Acquiring this information flow is fundamental to understanding the excavation environment and equipment status.

[0090] "The degree of environmental awareness" refers to the system's understanding and grasp of the current tunneling environment. This is not merely a simple accumulation of data, but a deeper understanding of the underlying geological conditions, equipment operating status, potential risks, and other in-depth information. The higher the degree of awareness, the more accurate the system's judgment of the environment, and the higher the reliability of its decisions.

[0091] "Environmental response information" refers to the feedback from the geological formation and surrounding environment to the tunneling operation during tunneling. This information can be physical quantities directly measured by sensors, or geological feature data that has been processed and analyzed.

[0092] "Pre-stored environmental feature patterns" refer to typical data patterns stored within the system that correspond to different geological conditions and environmental states. These patterns are established based on historical data, geological exploration data, and expert experience, and are used to compare and identify current environmental characteristics.

[0093] "The degree of matching of environmental features" refers to the similarity between the currently acquired environmental response information and the pre-stored environmental feature patterns. The higher the degree of matching, the closer the current environment is to the known patterns, and the more accurate the system's identification of the environment.

[0094] "Information flow fluctuation range" refers to the magnitude of change in information flow values ​​within a certain time window. A larger fluctuation range usually indicates a more unstable environment or equipment status.

[0095] "Frequency of abnormal events" refers to the number of times events deviating from normal operating conditions occur during the tunneling process. Abnormal events may include, but are not limited to, cutterhead jamming, groundwater inrush, ground subsidence, and equipment failure.

[0096] "The degree of drastic change in the environment" refers to the speed and intensity of changes in the tunneling environment or equipment status. The higher the degree of drastic change, the faster the system needs to respond and adjust.

[0097] The "Environmental Awareness Index" is a quantitative indicator that comprehensively reflects the level of awareness of the environmental state, the degree of matching with environmental characteristics, and the severity of environmental changes. This index is the core basis for the system to switch modes and optimize parameters.

[0098] "Efficiency-driven operation mode" refers to an operation strategy in which the system aims to maximize tunneling efficiency and minimize energy consumption when the environmental awareness index is high and the environment is relatively stable.

[0099] "The exploration mode that prioritizes safety and actively acquires information" refers to an operational strategy in which, when the environmental awareness index is low and the environment is complex and changeable, the system prioritizes operational safety and actively makes minor adjustments to parameters to acquire more environmental information.

[0100] The method proposed in this application first requires acquiring multiple information streams and assessing the correlation between these streams to determine the level of understanding of the environmental state. Information stream acquisition can be achieved in various ways. For example, mechanical parameters such as torque, speed, thrust, and vibration of the tunnel boring machine cutterhead can be collected in real time, as well as fluid parameters such as pressure, flow rate, density, and temperature of the mud circulation system. In addition, geophysical exploration data such as ground acoustic waves, seismic waves, and electromagnetic waves, as well as environmental monitoring data such as groundwater level and water quality, can be acquired. These information streams can be acquired through various sensors integrated on the tunnel boring machine, external monitoring equipment, and geological exploration data interfaces. After acquiring these information streams, the correlation between them needs to be assessed.

[0101] For example, methods such as cross-correlation analysis, Granger causality tests, and information entropy analysis can be used to quantify the mutual influence and dependencies between different information flows. By analyzing these correlations, it is possible to determine which information flows are independent and which are coupled, thereby gaining a more comprehensive understanding of the complexity and uncertainty of the current environmental state, and ultimately determining the level of awareness of the environmental state. For instance, when there is a strong positive correlation between the cutterhead torque and the propulsion force, and the vibration signal is stable, it may indicate that the formation is homogeneous and the system has a high level of awareness of the environment; conversely, when the correlation between multiple information flows is weak and the fluctuations are severe, it may indicate that the environment is complex and the level of awareness is low.

[0102] Secondly, it is necessary to acquire environmental response information and compare it with pre-stored environmental feature patterns to determine the degree of matching. Acquiring environmental response information can include real-time monitoring of the strata and surrounding environment feedback during tunneling. For example, acoustic responses within the strata can be acquired using stratum acoustic sensors, vibration characteristics of the strata and equipment can be acquired using vibration sensors, and annular pressure changes can be acquired using pressure sensors. This environmental response information is a direct feedback from the strata to tunneling disturbances. After acquiring this information, it needs to be compared with pre-stored environmental feature patterns. Pre-stored environmental feature patterns can be a database containing characteristic parameters of various typical geological conditions (such as homogeneous rock strata, fault zones, cavities, water-rich areas, etc.). Comparison methods can employ pattern recognition algorithms, such as support vector machines (SVM), neural networks, cluster analysis, etc. Through comparison, the similarity or distance between the current environmental response information and each pre-stored pattern can be calculated, thereby determining the degree of matching of environmental features. For example, when the environmental response information has a high degree of matching with the pre-stored "fault zone" pattern, it indicates that the current tunneling may have entered a fault zone area.

[0103] Secondly, it is necessary to statistically analyze the fluctuation range of the information flow and the frequency of abnormal events to determine the severity of environmental changes. The fluctuation range of the information flow can be obtained by calculating statistical quantities such as the difference between the maximum and minimum values, standard deviation, or coefficient of variation of each information flow within a certain time window. For example, a sliding time window can be set to continuously calculate the fluctuation amplitude of key parameters such as cutterhead torque and propulsion force within that window. The larger the fluctuation range, the more drastic the environmental change. The frequency of abnormal events can be identified by real-time monitoring of the information flow and using anomaly detection algorithms (such as statistical threshold-based or machine learning anomaly detection models) to identify events that deviate from normal operating modes. For example, a sudden drop in cutterhead torque or an abnormal increase in annular pressure can be identified as an abnormal event. By statistically analyzing the number of times these abnormal events occur per unit time, the frequency of abnormal events can be obtained. The larger the fluctuation range of the information flow and the higher the frequency of abnormal events, the more drastic the environmental change.

[0104] Finally, based on the level of awareness of the environmental state, the degree of matching of the environmental features, and the severity of the environmental changes, a quantified environmental awareness index is output, and mode switching is performed according to the environmental awareness index. The environmental awareness index can be calculated using methods such as weighted averaging, fuzzy logic reasoning, or machine learning models to integrate information from the three dimensions mentioned above. For example, weights can be assigned to each dimension, and then they can be linearly combined to obtain a comprehensive index. The numerical range of this index can be preset, for example, from 0 to 100. The higher the index, the more stable the environment and the clearer the awareness; the lower the index, the more complex the environment and the higher the uncertainty. After obtaining the environmental awareness index, the system will switch modes according to the index. The logic of mode switching can be set as follows: when the environmental awareness index is higher than a certain preset threshold, the system switches to an efficiency-oriented operation mode, in which the tunneling machine will operate with the goal of maximizing tunneling speed and minimizing energy consumption. For example, in homogeneous and stable strata, the environmental awareness index is high, and the system will select a parameter combination of high rotation speed and high thrust. When the environmental awareness index is lower than another preset threshold, the system switches to an exploration mode that prioritizes safety and actively acquires information. In this mode, the system will reduce the tunneling speed, decrease the range of parameter adjustments, and proactively perform small-scale parameter perturbations to obtain more environmental response information, thereby improving its understanding of complex environments. For example, when encountering fault zones or cavities, the environmental awareness index will decrease, and the system will switch to exploration mode to avoid potential risks and collect more geological information.

[0105] The big data-based method for optimizing disturbance parameters in vertical shaft tunneling proposed in this application outputs a quantified environmental awareness index by comprehensively assessing the level of awareness of the environmental state, the degree of matching of environmental characteristics, and the severity of environmental changes. Based on this index, the method intelligently switches operating modes. The core innovation of this method lies in its ability to dynamically assess the complexity and uncertainty of the tunneling environment and adjust the tunneling strategy according to the assessment results, rather than relying solely on pre-set experience models.

[0106] Compared with existing technologies, the advantages of this application are:

[0107] First, existing technologies often fail under complex geological conditions due to limitations in data analysis logic, making it difficult for the system to accurately interpret drastic changes in sensor signals and thus issuing incorrect parameter adjustment commands. This application constructs a quantitative "environmental cognition index" by introducing three dimensions: "degree of awareness of environmental conditions," "degree of matching of environmental characteristics," and "degree of drastic environmental changes." This index can more comprehensively and precisely reflect the true state of the current tunneling environment, thereby avoiding erroneous decisions caused by insufficient information or misjudgment in complex environments using traditional methods. For example, when a tunneling system enters a highly fractured fault zone, a traditional system may still attempt to maintain high-efficiency operation, thereby exacerbating the risk. However, the method in this application, through the assessment of the correlation of information flow, the comparison of environmental response information, and the statistics of information flow fluctuations and the frequency of abnormal events, can quickly identify the complexity and uncertainty of the environment, thereby reducing the environmental cognition index.

[0108] Secondly, existing technologies struggle to effectively distinguish and eliminate interference information caused by erroneous interventions when analytical logic fails and incorrect parameter adjustment commands are issued, leading to further deterioration of downhole conditions. Consequently, they cannot quickly and re-establish a reliable parameter optimization strategy suitable for new geological environments based solely on information reflecting the true characteristics of the current formation. This application addresses this by switching modes based on the environmental awareness index, providing an "efficiency-driven operation mode" and an "exploration mode prioritizing safety and proactively acquiring information." When the environmental awareness index decreases, the system can promptly switch from the efficiency-driven operation mode to the exploration mode. In exploration mode, the system prioritizes safety and proactively makes small-scale parameter adjustments to acquire more environmental information, thus avoiding the risks of blindly pursuing efficiency in uncertain environments. This mode-switching mechanism allows the system to adapt more flexibly to dynamically changing geological environments and provides a more reliable data foundation for subsequent parameter optimization. For example, in exploration mode, the system can make controlled, small-scale parameter adjustments and carefully analyze their actual impact on the environment, gradually establishing a new geological response model and avoiding repeating previous erroneous interventions.

[0109] In summary, this application significantly improves the adaptability, safety, and optimization capabilities of the shaft excavation system under complex geological conditions by constructing a multi-dimensional environmental cognitive assessment system and an intelligent mode switching mechanism, effectively solving the limitations of existing technologies in the face of complex and dynamic geological anomalies.

[0110] In some of the embodiments described above in this application, a mode switching mechanism based on an environmental awareness index is proposed to allow for a choice between an efficiency-driven operational mode and an exploration mode that prioritizes safety and proactively acquires information. However, in actual shaft excavation, simply switching to exploration mode cannot fully address the problem of safely and effectively adjusting parameters and acquiring information under unknown or uncertain geological conditions. Without a systematic learning and feedback mechanism, parameter adjustments in exploration mode may still lead to environmental degradation or even safety accidents, thus failing to truly optimize the disturbed parameters.

[0111] In response, this application further proposes that after the aforementioned mode switching based on the environmental awareness index, it also includes:

[0112] When in the exploration mode, the limitation range of the operation parameters is generated based on the currently identified environmental features and the recorded restriction operation information;

[0113] Within the limits of the operating parameters, perform parameter adjustment operations and obtain environmental response information;

[0114] Record the parameter adjustment operation and execution time, and based on the expected impact of the parameter adjustment operation on the equipment and environment, separate the signal component corresponding to the parameter adjustment operation from the environmental response information to obtain information reflecting the true characteristics of the environment;

[0115] The information reflecting the true characteristics of the environment is compared with the pre-stored environmental feature patterns to identify the current environmental characteristics and estimate environmental parameters.

[0116] The results of the parameter adjustment operation are monitored. When the parameter adjustment operation leads to a deterioration of the environmental condition, the parameter adjustment operation and the corresponding environmental conditions are recorded, and the parameter adjustment operation is added to the restriction operation information.

[0117] Specifically, when the system is in exploration mode, to ensure operational safety and effective information acquisition, it first dynamically generates a range of operational parameter limits based on the currently identified environmental characteristics and recorded operational restrictions. The identified environmental characteristics can be understood as information such as stratigraphic type, rock mass integrity, and groundwater occurrence status determined by the system through prior data analysis and comparison. Operational restrictions refer to parameter combinations or operational sequences that have been deemed potentially detrimental or unsafe in past operations. By integrating this information, a safe operational range can be defined, avoiding blind or high-risk parameter adjustments in unknown environments.

[0118] Within these constraints, the system will perform small, controlled parameter adjustments, such as fine-tuning the tunneling speed, cutterhead rotation speed, and mud pressure, while acquiring environmental response information in real time. This environmental response information typically includes real-time data from the tunneling machine cutterhead, propulsion system, mud circulation system, and ground monitoring sensors (such as vibration sensors, acoustic emission sensors, and pressure sensors). This data is crucial for evaluating the effectiveness of parameter adjustments and the actual state of the environment.

[0119] To accurately understand the true environmental impact of parameter adjustment operations, it is necessary to record the specific details and execution time of each operation. Subsequently, based on the expected impact of these adjustments on the tunneling equipment and its surrounding environment, the signal components directly caused by the parameter adjustments are separated from the acquired environmental response information. For example, a change in the cutterhead rotation speed of the tunneling machine directly causes specific changes in vibration frequency and torque; these are the equipment's own responses. By separating these expected signal components from the overall environmental response information, a purer and more accurate reflection of the true characteristics of the environment can be obtained, thus avoiding misinterpreting the equipment's own responses as changes in geological characteristics.

[0120] After acquiring information reflecting the true characteristics of the environment, it is compared with pre-stored environmental feature patterns. These pre-stored patterns are a database containing typical response patterns for different geological conditions, rock types, and hydrological characteristics. Through comparison, the system can identify current environmental characteristics, such as whether it has entered a new stratum, whether a fault zone exists, or whether it has encountered a water-rich area, and estimate relevant environmental parameters, such as rock mass strength and permeability coefficient.

[0121] During this process, the system continuously monitors the results of parameter adjustment operations. When a parameter adjustment operation is found to have deteriorated the environmental conditions, such as abnormal vibration, increased formation deformation, or a sudden increase in mud loss, the system will immediately record the parameter adjustment operation and its corresponding environmental conditions. These deteriorating operations will be added to the restricted operation information, forming a "blacklist" to prevent these dangerous operations from being performed again in similar environments in the future.

[0122] This application's solution effectively addresses the challenge of safely and efficiently acquiring formation information and optimizing operating parameters during shaft excavation in uncertain environments by establishing a closed-loop "exploration-learning-constraint" mechanism in exploratory mode. By limiting the operational range, the safety of exploratory adjustments is ensured; by precisely separating signal components, accurate understanding of the true characteristics of the environment is guaranteed; and through real-time monitoring and feedback, the identification and avoidance of hazardous operations are achieved, thereby continuously improving the operational strategy.

[0123] Through the aforementioned technical solution, this application can significantly improve the safety and adaptability of shaft excavation under complex geological conditions. It transforms the exploration mode from a passive, cautious approach to an active learning and optimization process, enabling the system to gain valuable experience from each parameter adjustment and convert it into actionable knowledge. This not only reduces the risk of accidents caused by unknown factors but also continuously optimizes excavation parameters, improves excavation efficiency, and provides more accurate decision support for subsequent excavation operations.

[0124] Recording the parameter adjustment operation and execution time, and based on the expected impact of the parameter adjustment operation on the equipment and environment, separating the signal component corresponding to the parameter adjustment operation from the environmental response information to obtain information reflecting the true characteristics of the environment can be achieved in the following manner.

[0125] First, key sensor data from the tunnel boring machine (TBM) is collected, and noise suppression and timestamp calibration are performed on this data. The key sensor data may include, but is not limited to, data on cutterhead torque, thrust, vibration, mud pressure, and flow rate. This data comprehensively reflects the TBM's operating status and its interaction with the geological formation. Noise suppression of the key sensor data aims to eliminate random errors introduced by sensor interference or environmental disturbances, improving data quality and signal-to-noise ratio. Timestamp calibration ensures that all sensor data are synchronized over time, providing an accurate time reference for subsequent causal analysis and signal separation.

[0126] Furthermore, a causal response model is established for each small-amplitude parameter adjustment operation, and the system response curves that should be generated on each sensor are predicted in real time based on the operation content and the execution time. The causal response model is constructed based on historical data and physical principles, and is used to describe the predictable impact of specific parameter adjustment operations (e.g., fine-tuning of cutterhead speed, slight increase in propulsion force) on the tunneling machine system (such as the cutterhead, propulsion system, and mud circulation system) and its direct environment (such as the tunnel face). Through this model, the theoretically expected signal changes, i.e., the system response curves, can be predicted on each key sensor based on the actual type, intensity, and precise execution time of the parameter adjustment operation.

[0127] Subsequently, the system response curve is separated from the key sensor data to obtain a preliminary formation response signal. This step is the core of signal separation. By subtracting the signal components caused by the predicted operation of the tunneling machine itself from the total sensor data actually collected, or by removing them through other signal processing techniques (such as adaptive filtering), the signal components mainly caused by changes in formation characteristics, i.e., the preliminary formation response signal, are extracted.

[0128] Based on this, high-resolution time-frequency analysis is performed on the preliminary stratigraphic response signal to obtain its energy distribution at different time points and frequencies. High-resolution time-frequency analysis (e.g., wavelet transform, Hilbert-Huang transform, etc.) can reveal the fine structure of the signal in time and frequency, thereby more accurately capturing transient events and non-steady-state characteristics that may be contained in the stratigraphic response signal. By analyzing the energy distribution, anomalous energy accumulations or changes within specific time windows and frequency ranges can be identified, which are often related to geological micro-events.

[0129] Next, transient patterns are extracted from the energy distribution and compared with a predefined geological event feature library to instantly identify the type of geological micro-event. The transient patterns refer to brief, non-periodic energy change patterns exhibited in the time-frequency diagram, such as sudden energy increases, frequency drifts, and specific waveform shapes. The geological event feature library pre-stores typical characteristic patterns in the time-frequency domain of various known geological micro-events (such as rock mass micro-collapse, stress release, small-scale fracture development, and localized weak interlayers). By rapidly comparing the extracted transient patterns with the feature library, the type of currently occurring geological micro-event can be identified in real time and accurately.

[0130] Finally, geological event parameters are estimated based on the described geological micro-event types, and these types and parameters are integrated into the current working condition description to adjust the uncertainty index. Once a geological micro-event type is identified, related geological event parameters can be further estimated based on its characteristics, such as the scale of micro-collapses, the intensity of stress release, and the density of fractures. This detailed geological event type and parameter information is integrated into the current tunneling working condition description, providing the system with a more comprehensive and accurate environmental understanding. By introducing this new and more specific stratigraphic information, the original environmental uncertainty index can be adjusted in real time, making it more accurately reflect the true complexity and risk level of the current strata.

[0131] The solution proposed in this application effectively distinguishes the signal components caused by the tunnel boring machine's own operation from the actual response signals of the formation through refined sensor data processing, system response prediction and separation based on causal models, and high-resolution time-frequency analysis and geological event pattern recognition. It is precisely this precise signal separation and instantaneous identification of geological micro-events that enables the system to acquire information reflecting the true characteristics of the environment, thereby avoiding misjudgments of formation characteristics caused by equipment operation interference.

[0132] The aforementioned technical solutions significantly enhance the ability to extract true geological characteristics from complex environmental response information, making the system more sensitive and accurate in identifying micro-geological events. This allows for more precise estimation of geological event parameters and real-time adjustment of the uncertainty index of environmental perception, providing a more reliable basis for subsequent mode switching and parameter optimization. This in-depth mining of true environmental characteristics helps to more effectively identify potential risks and optimize tunneling strategies in exploratory mode, thereby improving the safety and efficiency of shaft excavation.

[0133] In some embodiments described above in this application, a method is proposed that, in exploration mode, when a parameter adjustment operation is detected to cause environmental degradation, the operation is recorded and added to the restricted operation information. However, in its implementation, environmental degradation may be caused by various factors, such as secondary effects caused by system intervention, perception distortion caused by sensor internal biases, and anomalies caused by changes in formation characteristics. Failure to accurately distinguish these causes may lead to misjudgments of the operation, thereby unnecessarily limiting effective tunneling operations or failing to identify true formation risks in a timely manner.

[0134] In response, this application further proposes the following steps for monitoring the results of the aforementioned parameter adjustment operation: when the parameter adjustment operation leads to a deterioration of the environmental condition, recording the parameter adjustment operation and the corresponding environmental conditions, and adding the parameter adjustment operation to the restriction operation information, including:

[0135] The sequence of operations prior to the deterioration of the environmental conditions was traced back, including system-issued instructions, the execution time of the instructions, tunneling machine parameters, and the status of the mud circulation system.

[0136] Retrieve sensor trust weights and calibration information during the execution of the operation sequence;

[0137] Based on the instructions issued by the system, the execution time of the instructions, the parameters of the tunneling machine, the status of the mud circulation system, and the sensor trust weights and calibration information, the expected changes in sensor data should be recalculated under the condition of no internal sensor bias and only considering the system's own intervention.

[0138] Compare sensor data during the actual monitored deterioration period with the recalculated expected changes;

[0139] Based on the comparison results, identify the expected secondary effects caused by the system's own intervention during the deterioration process, the perception distortion caused by the internal bias of the sensor, and the anomalies caused by changes in the formation characteristics;

[0140] Based on the identification results, the main causes leading to the deterioration of the environmental condition were determined;

[0141] When the primary cause is a change in formation characteristics that exceeds the system's adaptability, the parameter adjustment operation and the actual formation characteristic information are added to the restriction operation information.

[0142] Specifically, tracing back the operational sequence prior to the environmental degradation refers to a complete record of all operational commands issued to the tunneling machine before the system detected the degradation, the execution time of these commands, real-time parameters of the tunneling machine equipment (e.g., cutterhead speed, thrust, torque), and the real-time status of the mud circulation system (e.g., mud flow rate, pressure, density). This information constitutes a detailed historical snapshot of the tunneling process and forms the basis for subsequent analysis.

[0143] The retrieval of sensor trust weights and calibration information during the execution of the operation sequence can be understood as obtaining the reliability assessment (trust weight) of each sensor's data and sensor calibration parameters for a period of time before deterioration occurs. Sensor trust weights can be determined based on historical data, sensor self-diagnostic results, or comparison with redundant sensors. Calibration information is used to correct systematic biases in sensor readings, ensuring data accuracy.

[0144] In practical applications, based on the instructions issued by the system, the execution time of the instructions, the parameters of the tunneling machine, the status of the mud circulation system, and the sensor trust weights and calibration information, the expected changes in sensor data under conditions of no internal sensor bias and considering only the system's own intervention are recalculated. This refers to using an accurate tunneling process simulation model or physical model, inputting historical operating data and equipment status, and combining it with calibrated sensor information to predict the data that the sensors should record under ideal conditions (i.e., no ground anomalies or sensor malfunctions). This expected change model considers the direct impact of tunneling machine operation on the environment; for example, increased thrust leads to increased ground pressure, and changes in cutterhead speed cause changes in vibration modes.

[0145] Furthermore, comparing the sensor data actually monitored during the deterioration period with the recalculated expected changes involves comparing the actual sensor data collected during the environmental deterioration period with the aforementioned predicted ideal sensor data one by one. This comparison can use statistical methods (such as mean squared error, correlation coefficient) or machine learning algorithms to quantify the differences.

[0146] Therefore, based on the comparison results, identifying the expected secondary effects caused by system intervention, perception distortion caused by internal sensor bias, and anomalies caused by changes in formation characteristics during deterioration refers to distinguishing anomalies from different sources by analyzing the deviation patterns between actual and expected data. For example, if there is a systematic deviation between actual and expected data on a specific sensor, and this is inconsistent with the sensor's low trust weight or calibration information, it may be attributed to internal sensor bias. If actual and expected data show changes on multiple sensors that are related to system instructions but exceed expected magnitudes, there may be secondary effects caused by system intervention. And if there are significant differences between actual and expected data that cannot be explained by system intervention or sensor bias, it is highly likely that changes in formation characteristics have occurred.

[0147] Specifically, determining the primary cause of the environmental degradation based on the identification results means, after distinguishing the three sources of anomalies mentioned above, using logical reasoning and priority judgment to determine which one is the root cause of the current environmental degradation. For example, if the anomaly caused by changes in geological characteristics is the most significant and cannot be explained, then it is identified as the primary cause.

[0148] In a preferred embodiment, when the primary cause is a change in formation characteristics that exceeds the system's adaptability, the parameter adjustment operation and the actual formation characteristic information are added to the restrictive operation information. This means that only when it is confirmed that the deterioration is caused by an uncontrollable change in the formation itself, and this change exceeds the current parameter adjustment capability of the tunneling system, will the parameter adjustment operation leading to the deterioration and its corresponding actual formation characteristic information be recorded and used as restrictive operation information to guide subsequent tunneling strategies.

[0149] This application's solution addresses the difficulty in accurately identifying the root causes of environmental degradation in exploratory mode by establishing a multi-dimensional backtracking and comparison mechanism. Through this technical solution, the accuracy and reliability of diagnosing environmental degradation in exploratory mode are significantly improved. Compared to the basic solution that only monitors degradation and records operations, this solution, by introducing multi-source information backtracking, sensor data calibration and trust assessment, and an anomaly separation mechanism based on model prediction, can accurately distinguish between degradation caused by internal system factors (such as secondary effects of operations and sensor malfunctions) and external factors (such as changes in formation characteristics). This precise attribution capability avoids misjudging harmless or controllable operations as dangerous and restricting them, thereby reducing unnecessary tunneling efficiency losses. Simultaneously, when it is confirmed that degradation is indeed caused by changes in formation characteristics exceeding the system's adaptability, the system can promptly and accurately add relevant operations and formation information to the restricted operation information, providing safer and more accurate guidance for subsequent tunneling, effectively improving the safety and intelligent decision-making level of the shaft tunneling process.

[0150] In some preferred embodiments, assuming that in the exploration mode of shaft excavation, after the tunnel boring machine performs a small-scale thrust adjustment operation, data from multiple sensors (e.g., formation pressure sensor, cutterhead torque sensor, vibration sensor) show abnormal fluctuations, indicating that the environmental conditions may be deteriorating. At this time, the system will initiate the aforementioned monitoring process:

[0151] First, the system will review the sequence of operations that led to this deterioration. This includes the system's command to "increase thrust by 5%", the precise execution time of the command, changes in equipment parameters such as the tunnel boring machine's cutterhead speed, thrust, and torque before and after the command was executed, as well as status data such as the flow rate and pressure of the mud circulation system.

[0152] Next, the system retrieves the trust weights and calibration information of all relevant sensors during the execution of this operation sequence. For example, a formation pressure sensor may have a slightly lower trust weight due to long-term use, or its calibration parameters may show a slight systematic deviation.

[0153] Then, the system will use its internal tunneling process simulation model, combined with the above-mentioned backtracked operation sequence and calibrated sensor information, to recalculate the expected change curves of these sensor data under ideal conditions where there are no formation anomalies or sensor failures.

[0154] The system then compares the sensor data actually monitored during the deterioration period with the recalculated expected changes.

[0155] By comparing the results, the system found that:

[0156] The actual data from the cutter head torque and vibration sensors deviated somewhat from the expected changes, but this deviation pattern was basically consistent with the secondary effects that should be caused by the increase in propulsion force of the system (e.g., a slight increase in torque and a slight change in vibration frequency), and was within an acceptable range.

[0157] The actual reading of a certain formation pressure sensor was significantly higher than expected, and this deviation pattern was consistent with the deviation pattern indicated by the sensor's low trust weight and historical calibration data. This suggests that there may be sensing distortion caused by internal biases within the sensor.

[0158] Another formation pressure sensor reading suddenly and sharply increased at a certain point in time, and this increase could not be explained by system operation or sensor bias. Its waveform characteristics were highly similar to transient signal patterns of formation micro-collapse or stress release. This was identified as an anomaly caused by changes in formation properties.

[0159] Based on the above identification results, the system determined that the main cause of this environmental degradation was "anomaly caused by changes in formation characteristics" (i.e., formation micro-collapse or stress release).

[0160] In some of the embodiments described above in this application, a mode-switching scheme based on the environmental awareness index was proposed. However, in actual shaft excavation operations, the environmental conditions may change suddenly and drastically. Relying solely on periodically calculated environmental awareness indices for mode switching may not be able to respond to these rapid changes in a timely manner, thus posing certain safety hazards or efficiency losses. Therefore, this application further proposes a more sensitive and mandatory mode-switching mechanism to cope with rapid dynamic changes in the environment.

[0161] The step of switching modes based on the environmental awareness index includes:

[0162] Real-time monitoring of multiple key information streams, and continuous calculation of the rate of change of the information streams within a short time window;

[0163] Set a warning threshold for the rate of change;

[0164] When the rate of change exceeds the warning threshold, an emergency response signal is triggered.

[0165] Upon receiving the emergency response signal, the system is forced to switch from the efficiency-driven operation mode to the exploration mode.

[0166] Specifically, real-time monitoring of multiple key information streams refers to the system continuously collecting data streams from multiple sources, including tunneling machine equipment, ground sensors, mud circulation systems, and environmental monitoring equipment. These information streams can include cutterhead torque, propulsion speed, vibration frequency, mud pressure, flow rate, gas concentration, ground stress, and groundwater seepage. Continuously calculating the rate of change of these information streams within a short time window means performing dynamic analysis on this real-time data. For example, algorithms such as moving average, exponential smoothing, or difference can be used to calculate the trend and speed of change within a preset short time window (e.g., a few seconds to tens of seconds). This rate of change can reflect the instantaneous dynamics of the environment or equipment status.

[0167] Setting warning thresholds for the rate of change can be understood as setting one or more critical values ​​for the rate of change of each key information flow based on historical data, expert experience, or safety regulations. These thresholds can be absolute values ​​or relative percentage changes. For example, a sudden increase of more than 20% in cutterhead torque within 5 seconds, or a drop of more than 15% in mud pressure within 10 seconds, may be considered abnormal. The purpose is to identify rapid dynamics that may indicate potential hazards or significant environmental changes.

[0168] In practical applications, when the rate of change exceeds the warning threshold, an emergency response signal is triggered. This emergency response signal can be a high-priority system interruption, an audible and visual alarm, or a data packet, the purpose of which is to immediately notify the system of a potential emergency.

[0169] Furthermore, upon receiving the emergency response signal, the system is forcibly switched from the efficiency-driven operation mode to the exploration mode. This forced switch means that the system will immediately interrupt the current efficiency-driven operation mode, such as reducing the tunneling speed, stopping certain high-risk operations, and entering the exploration mode, which prioritizes safety and actively acquires information. In exploration mode, the system will initiate additional sensor data acquisition, make minor parameter adjustments to probe the environment, or execute preset safety procedures to comprehensively assess the current environmental conditions.

[0170] This application's solution addresses the potential response lag issue in the basic approach that relies solely on environmental awareness indices for mode switching by introducing a real-time monitoring and early warning mechanism for the rate of change of key information flow. Through this technical solution, the application significantly improves the response speed and safety to sudden environmental changes during shaft excavation. Compared to mode switching based solely on environmental awareness indices, the real-time monitoring and forced switching mechanism based on the rate of change of information flow introduced in this application allows the system to more promptly and decisively shift from an efficiency-driven operation mode to a safety-priority exploration mode when facing rapidly evolving situations such as sudden geological changes or equipment failure precursors. This effectively avoids accidents that may result from slow response, reduces operational risks, and provides the system with valuable time to gather more information, assess risks, and develop more robust response strategies, thereby enhancing the intelligent safety management level of the entire excavation process.

[0171] This application further proposes that when parameter adjustments are performed at the edge of a cavity, the aforementioned parameter adjustment operation and execution time are recorded, and based on the expected impact of the parameter adjustment operation on the equipment and environment, the signal components corresponding to the parameter adjustment operation are separated from the environmental response information to obtain information reflecting the true characteristics of the environment, including:

[0172] Real-time acquisition of key sensor data on cutter head torque, vibration, and annular pressure;

[0173] High-resolution time-frequency analysis is performed on the key sensor data to generate a time-frequency graph;

[0174] In the time-frequency diagram, signal patterns caused by the tunneling machine's own operation are identified, including the delayed response, energy distribution, and duration after the operation.

[0175] In the time-frequency graph, transient signal patterns caused by micro-collapse or stress release of rock mass at the edge of the cavity are identified. These transient signal patterns include energy surges, frequency shifts, and waveform shapes.

[0176] By comparing the time-frequency characteristics of the signal pattern and the transient signal pattern, the signal components generated by the tunneling machine's own operation are separated from the transient signal pattern;

[0177] The separated transient signal patterns are used as information reflecting the true characteristics of the environment.

[0178] Specifically, real-time acquisition of key sensor data such as cutterhead torque, vibration, and annular pressure refers to deploying high-precision sensors in the contact area between the tunnel boring machine (TBM) cutterhead and the formation, as well as in the surrounding environment, to continuously acquire physical quantities closely related to the TBM's operating status and the formation's response. Cutterhead torque data reflects changes in the resistance to the cutterhead cutting the formation; vibration data reveals dynamic processes within the formation, such as stress release, microcrack propagation, or rock mass collapse; and annular pressure data can be used to monitor the interaction between the mud circulation system and the formation, especially at the edges of cavities, where abnormal fluctuations in annular pressure may indicate formation stability problems. These key sensor data form the basis for subsequent signal analysis, and their real-time acquisition and accuracy are crucial for timely detection of micro-events in the formation.

[0179] Furthermore, performing high-resolution time-frequency analysis on the key sensor data to generate a time-frequency graph involves using advanced signal processing techniques such as Short-Time Fourier Transform (STFT), Wavelet Transform, or Hilbert-Huang Transform to convert the acquired time-domain sensor data to the time-frequency domain. High-resolution time-frequency analysis can simultaneously reveal the distribution characteristics of the signal in both time and frequency, making it possible to identify transient and non-stationary events in complex signals. The generated time-frequency graph visually displays the changes of different frequency components over time, providing a visual basis for subsequent signal pattern recognition.

[0180] In the time-frequency graph, signal patterns caused by the tunneling machine's own operations are identified. These signal patterns include the delayed response, energy distribution, and duration after the operation. Tunneling machine operations, such as cutterhead speed adjustments, thrust changes, and mud pump start-up and shutdown, generate predictable signal patterns in the sensor data. These signal patterns typically have predictable delayed responses, meaning they appear on the sensors only after a certain period of time after the operation occurs; their energy distribution and duration are also relatively stable and can be learned and characterized using historical data or preset models. Identifying these known patterns helps distinguish equipment-related disturbances from the ground response.

[0181] Simultaneously, the time-frequency graph identifies transient signal patterns caused by micro-collapse or stress release of the rock mass at the cavity edge. These transient signal patterns include sudden energy increases, frequency shifts, and waveform shapes. Micro-collapse or stress release of the rock mass at the cavity edge is a sudden geological event, manifesting as transient signals in sensor data. These transient signals are characterized by a rapid increase in energy over a short period, possibly accompanied by rapid changes in frequency components (frequency shifts), and unique waveform shapes, such as shock waves or damped oscillations. These characteristics differ significantly from the signal patterns caused by the tunneling machine's own operation and are key to identifying the true characteristics of the formation.

[0182] Therefore, by comparing the time-frequency characteristics of the signal pattern and the transient signal pattern, the signal components generated by the tunnel boring machine's own operation can be separated from the transient signal pattern. Specifically, in the time-frequency diagram, there are significant differences between the signal pattern caused by the tunnel boring machine's own operation (e.g., its stable frequency components, predictable energy envelope) and the transient signal pattern caused by geological events at the cavity edge (e.g., sudden energy, broadband characteristics, nonlinear frequency drift). By accurately comparing and analyzing these differences, it is possible to identify which parts of the observed transient signal are generated by the tunnel boring machine's own operation (e.g., transient responses that may be caused by rapid parameter adjustments). Once identified, these transient signal components generated by the equipment's own operation can be effectively separated from the overall transient signal pattern, thereby revealing the microscopic transient response of the formation itself.

[0183] Ultimately, the separated transient signal patterns are used as information reflecting the true characteristics of the environment. This means that after eliminating interference from the tunnel boring machine's own operation, the obtained transient signal patterns can more accurately and directly reflect the true physical state and dynamic changes of the strata at the cavity edge, such as rock mass stability, stress concentration, and potential instability risks. This information is crucial for subsequent environmental feature identification, environmental parameter estimation, and optimization of tunneling disturbance parameters.

[0184] This application's solution effectively solves the problem of accurately distinguishing between equipment interference and the true response of the formation in traditional methods by employing high-resolution time-frequency analysis technology in the complex and sensitive area of ​​the cavity edge, combined with refined identification and comparison of the tunnel boring machine's own operating signal patterns and the transient signal patterns of geological events. Through this technical solution, this application can significantly improve the accuracy of identifying the true characteristics of the formation in environmental response information during shaft excavation, especially under complex geological conditions such as the cavity edge. Compared to relying solely on general signal separation methods, this solution, by introducing high-resolution time-frequency analysis and specifically identifying the differences between the tunnel boring machine's operating signals and geological transient signals, effectively avoids interference from equipment operation on the interpretation of formation information, thereby obtaining purer and more accurate information reflecting the true characteristics of the environment. This not only helps to more accurately identify current environmental characteristics and estimate environmental parameters, but also enables the timely detection of potential geological risks at the cavity edge, such as rock micro-collapse or stress release, providing a reliable basis for optimizing tunneling disturbance parameters, thus significantly improving the safety and efficiency of tunneling operations.

[0185] In some embodiments of this application, in exploration mode, by recording parameter adjustment operations and execution times, and based on the expected impact of the parameter adjustment operations on the equipment and environment, the signal components corresponding to the parameter adjustment operations are separated from the environmental response information to obtain information reflecting the true characteristics of the environment. However, in actual shaft excavation, the interaction between the tunneling machine and the stratum is extremely complex. Factors such as the heterogeneity, water content, porosity, and degree of consolidation of the stratum medium can significantly affect the propagation and attenuation of signals generated by the tunneling machine operation. If signal separation is performed solely based on a simple expected impact model, it may be impossible to accurately extract the signal components generated by the equipment's own operation, resulting in deviations in the information obtained reflecting the true characteristics of the environment. This, in turn, affects the accurate identification of the true characteristics of the stratum and the estimation of environmental parameters, and may ultimately limit the effectiveness of parameter optimization and safety assurance. To address this, this application further proposes a more refined signal separation method, aiming to accurately predict and adaptively separate equipment operation signals by comprehensively considering tunneling machine operating parameters, equipment status, and stratum medium characteristics, in order to obtain more realistic and reliable stratum response information.

[0186] The above records the parameter adjustment operation and execution time, and based on the expected impact of the parameter adjustment operation on the equipment and environment, separates the signal component corresponding to the parameter adjustment operation from the environmental response information to obtain information reflecting the true characteristics of the environment, including:

[0187] The system collects data from multiple sensors in the tunnel boring machine cutterhead, propulsion system, and mud circulation system in real time. The sensor data includes vibration signals, pressure signals, torque signals, and flow signals.

[0188] Based on the type and intensity of the parameter adjustment operation, as well as the current status of the tunneling machine and the tunneling speed, calculate the instantaneous force distribution of the parameter adjustment operation on the contact area between the tunneling machine cutterhead and the stratum.

[0189] Based on the instantaneous force distribution, combined with the acoustic propagation velocity, attenuation coefficient, and reflection coefficient of the current formation medium, the time-domain response waveform that should be generated on the multiple sensors after the force propagates in the formation is predicted.

[0190] The amplitude and phase of the time-domain response waveform are adjusted according to the water content, porosity, and degree of consolidation of the formation medium.

[0191] An adaptive separation method based on signal feature matching separates the adjusted predicted time-domain response waveform from real-time sensor data to obtain information reflecting the true characteristics of the formation.

[0192] Specifically, real-time data is collected from multiple sensors in the tunnel boring machine's cutterhead, propulsion system, and mud circulation system. The aim is to comprehensively capture key information such as mechanical vibration, pressure changes, torque fluctuations, and mud flow generated during parameter adjustments. This sensor data, including vibration, pressure, torque, and flow signals, can directly or indirectly reflect the dynamic interaction between the tunnel boring machine and the geological formation, providing initial data for subsequent signal separation and formation characteristic analysis.

[0193] The calculation of the instantaneous force distribution on the contact area between the tunneling machine cutterhead and the stratum, based on the type and intensity of the parameter adjustment operation, the current status of the tunneling machine, and the tunneling speed, can be understood as establishing a mechanical model of the tunneling machine operation and the contact with the stratum to accurately quantify the instantaneous mechanical effects of specific parameter adjustments (e.g., cutterhead rotation speed, propulsion force, mud pump pressure, etc.) at the cutterhead-stratum interface. The status of the tunneling machine (e.g., cutter wear, bearing clearance) and the tunneling speed directly affect the transmission and distribution of forces, and therefore need to be included in the calculation model to improve the accuracy of predictions.

[0194] In practical applications, based on the instantaneous force distribution and considering the acoustic propagation velocity, attenuation coefficient, and reflection coefficient of the current formation medium, predicting the time-domain response waveforms that the force should produce on the multiple sensors after propagating in the formation refers to simulating the propagation path and energy attenuation process of the tunnel boring machine's force in the formation using acoustic propagation theory and geoacoustic models. The acoustic propagation velocity, attenuation coefficient, and reflection coefficient of the formation medium are key parameters affecting the acoustic propagation characteristics; they determine the time delay, amplitude attenuation, and waveform distortion of the force signal when it reaches different sensors. By accurately predicting these time-domain response waveforms, a precise "fingerprint" can be provided for subsequently separating the equipment's own signal from actual sensor data.

[0195] Furthermore, the amplitude and phase of the time-domain response waveform are adjusted based on the water content, porosity, and degree of consolidation of the formation medium, with the aim of further improving the accuracy of the predicted waveform. The water content, porosity, and degree of consolidation of the formation medium are important factors affecting the formation's elastic modulus, density, and acoustic impedance. They significantly alter the propagation speed, attenuation characteristics, and reflection characteristics of sound waves in the medium, thereby affecting the amplitude and phase of the signal received by the sensor. By acquiring these formation parameters in real-time or near real-time and dynamically adjusting the predicted waveform, the prediction model can better adapt to the dynamic changes in the formation environment.

[0196] Finally, an adaptive separation method based on signal feature matching separates the adjusted predicted time-domain response waveform from the real-time acquired sensor data to obtain information reflecting the true characteristics of the formation. This method employs advanced signal processing techniques, such as Kalman filtering, independent component analysis (ICA), or deep learning, to adaptively separate the waveform from the actually acquired mixed sensor data, using the precisely adjusted predictive equipment operating signal as a reference. This method can identify and match specific time-frequency characteristics of the predicted signal, thereby effectively removing interference generated by the equipment's own operation and ultimately obtaining response information that purely reflects the true characteristics of the formation.

[0197] This application's solution addresses the problem of insufficient signal separation accuracy in complex geological environments using traditional methods by comprehensively considering tunneling machine operating parameters, equipment status, tunneling speed, and formation medium characteristics. Through this technical solution, the accuracy and reliability of separating information reflecting the true characteristics of the formation from environmental response information can be significantly improved. Compared to basic solutions that rely solely on anticipated effects for signal separation, this application, by meticulously considering tunneling machine operating parameters, equipment status, tunneling speed, and the acoustic and physical characteristics of the formation medium, can more accurately predict and adaptively remove signal interference generated by the equipment's own operation. Consequently, the acquired information on the true characteristics of the formation has higher fidelity, enabling more accurate identification of current environmental features and estimation of environmental parameters, thus providing a more solid data foundation for optimizing vertical shaft tunneling disturbance parameters. This high-precision information acquisition capability helps to discover potential geological risks earlier and more accurately in exploratory mode, optimize operating parameters, effectively avoid safety accidents or efficiency losses caused by misjudgment of formation conditions, and ultimately improve the safety and economy of the entire tunneling process.

[0198] In some of the embodiments described above in this application, a scheme for mode switching based on an environmental awareness index is proposed. However, in practical applications, if the environmental awareness index fluctuates frequently around a single switching threshold, it may lead to unnecessary and repeated switching between the operating mode and the exploration mode. Such frequent mode switching not only reduces the overall efficiency of tunneling operations but may also introduce additional safety risks due to the instability of system decision-making.

[0199] In response, this application further proposes steps for mode switching based on the environmental awareness index, including:

[0200] Real-time acquisition of environmental awareness index;

[0201] Based on the numerical range of the environmental awareness index, it is divided into an efficiency zone, an exploration zone, and a transition zone between the efficiency zone and the exploration zone;

[0202] A hysteresis interval for mode switching is set in the transition zone. The hysteresis interval has an upper threshold for switching from the operation mode to the exploration mode and a lower threshold for switching from the exploration mode to the operation mode. The upper threshold is higher than the lower threshold.

[0203] Specifically, the real-time acquisition of the environmental awareness index refers to the system continuously monitoring and calculating the quantitative indicators of the current environmental state. This index comprehensively reflects the level of awareness of the environmental state, the degree of matching of environmental characteristics, and the severity of environmental changes.

[0204] The efficiency zone typically corresponds to a low environmental awareness index, indicating a relatively stable and predictable environment. In this zone, the system tends to operate in an efficiency-oriented mode. The exploration zone corresponds to a high environmental awareness index, indicating increased environmental uncertainty or potential risks. In this zone, the system should prioritize safety and actively acquire information. The transition zone is the index range between the efficiency and exploration zones, representing a critical stage where environmental uncertainty is increasing or decreasing.

[0205] Furthermore, a hysteresis interval for mode switching is set within the transition zone to prevent frequent mode switching when the environmental awareness index fluctuates slightly within this zone. This hysteresis interval is defined as having an upper threshold for switching from operating mode to exploration mode and a lower threshold for switching from exploration mode to operating mode. The upper threshold is set higher than the lower threshold, thus forming a clear switching "buffer." When the environmental awareness index rises and reaches the upper threshold for the first time, the system switches from operating mode to exploration mode; conversely, when the environmental awareness index falls and reaches the lower threshold for the first time, the system switches from exploration mode back to operating mode. When the environmental awareness index fluctuates between the upper and lower thresholds, the system maintains its current mode, effectively preventing repeated mode switching.

[0206] This application's solution effectively addresses the frequent mode switching problem that can occur when the environmental awareness index fluctuates around a single threshold by introducing an efficiency zone, an exploration zone, and a transition zone, and setting a hysteresis interval, upper threshold, and lower threshold within the transition zone. Specifically, when the environmental awareness index gradually increases from a lower value (efficiency zone), the system remains in operating mode until the index reaches the upper threshold before switching to exploration mode. Conversely, when the environmental awareness index gradually decreases from a higher value (exploration zone), the system remains in exploration mode until the index drops to the lower threshold before switching back to operating mode. This mechanism ensures that the system maintains its current operating mode when the environmental awareness index is within the hysteresis interval between the upper and lower thresholds, avoiding unnecessary switching due to minor fluctuations. It is precisely this hysteresis characteristic that makes mode switching more stable and reliable.

[0207] Through the above technical solution, this application can significantly improve the stability of mode switching in the method for optimizing disturbance parameters in vertical shaft tunneling. By introducing a hysteresis interval and distinguishing between upper and lower thresholds, the problem of frequent mode switching when the environmental awareness index fluctuates near the critical value is effectively avoided, thereby reducing unnecessary system adjustments and operational interruptions. This not only improves the continuity and efficiency of tunneling operations, but also further ensures operational safety by ensuring a timely and stable switch to exploration mode when uncertainty increases. Compared with a single threshold switching scheme, the additional technical features of this application make the system decision-making more robust, reduce the risk of misjudgment, and thus optimize the overall tunneling process.

[0208] In some preferred embodiments, a specific example is given below. Assume the environmental awareness index ranges from 0 to 100. An efficiency zone can be defined as an environmental awareness index less than 40, and an exploration zone as an environmental awareness index greater than 60. The range between 40 and 60 is defined as a transition zone. Within the transition zone, an upper threshold of 55 and a lower threshold of 45 are set for mode switching.

[0209] Specifically, when the system is currently in an efficiency-oriented operating mode, and the environmental awareness index gradually increases from 30 to 40, 45, and 50, the system remains in operating mode. Only when the environmental awareness index further increases and reaches 55 for the first time will the system switch from operating mode to an exploration mode that prioritizes safety and actively acquires information.

[0210] Conversely, when the system is currently in exploration mode and the environmental awareness index gradually decreases from 70 to 60, 55, and 50, the system remains in exploration mode. Only when the environmental awareness index further decreases and reaches 45 for the first time will the system switch back to operation mode from exploration mode.

[0211] Within this hysteresis range, for example, if the environmental awareness index fluctuates between 46 and 55 and the system is currently in operating mode, it will remain in operating mode; if the system is currently in exploration mode, it will remain in exploration mode. This mechanism effectively prevents the system from repeatedly switching between the two modes when the environmental awareness index fluctuates slightly between 45 and 55, thus ensuring the stability of mode switching and the continuity of tunneling operations.

[0212] In some embodiments described above in this application, mode switching based on an environmental awareness index is proposed, and a hysteresis interval for mode switching is set to avoid frequent mode switching near the critical value of the environmental awareness index. However, in practical applications, simply setting a hysteresis interval may not be sufficient to completely eliminate potential mode oscillations or unnecessary mode switching caused by small fluctuations in the environmental awareness index within the hysteresis interval, especially when the index remains near the threshold for an extended period, in which case the system's behavior may be unstable and unpredictable.

[0213] In response, this application further proposes the aforementioned mode switching based on the environmental awareness index, including:

[0214] When the environmental awareness index gradually increases from the efficiency zone and reaches the upper limit threshold for the first time, the operation mode is switched to the exploration mode.

[0215] When the environmental awareness index gradually decreases from the exploration zone and reaches the lower limit threshold for the first time, the system switches from the exploration mode to the operation mode.

[0216] Within the hysteresis interval, when the environmental perception index fluctuates between the upper and lower thresholds, the current operating mode remains unchanged.

[0217] Specifically, when the environmental awareness index gradually increases from the efficiency zone and reaches the upper limit threshold for the first time, the system will switch from an efficiency-oriented operation mode to an exploration mode that prioritizes safety and actively acquires information. "Gradually increasing and reaching the upper limit threshold for the first time" means that the environmental awareness index shows a continuous upward trend and touches or exceeds the preset upper limit threshold for the first time, rather than experiencing brief, instantaneous fluctuations. Similarly, when the environmental awareness index gradually decreases from the exploration zone and reaches the lower limit threshold for the first time, the system will switch from exploration mode back to operation mode. This also requires the index to show a continuous downward trend and touch or fall below the preset lower limit threshold for the first time. The upper and lower thresholds together define a hysteresis interval, which serves as a buffer for mode switching. Within this hysteresis interval, when the environmental awareness index fluctuates between the upper and lower thresholds, the system will maintain its current mode, whether operation or exploration mode, thus avoiding frequent mode switching caused by small fluctuations in the index around the thresholds.

[0218] This application's solution introduces explicit mode-switching logic: mode switching only occurs when the environmental awareness index first crosses a preset upper or lower threshold, and the current mode remains unchanged within the hysteresis interval. This effectively solves the mode oscillation problem that may exist in the basic solution mentioned above. This mechanism ensures the stability of mode switching and avoids the system frequently switching back and forth between operating and exploration modes due to small fluctuations in the environmental awareness index. It is precisely because of this explicit switching rule that the system can operate more stably, reducing the system overhead and potential risks caused by unnecessary mode switching.

[0219] Through the above technical solution, this application can significantly improve the robustness and stability of mode switching. By setting clear switching conditions for the first time reaching the threshold and a strategy of keeping the mode unchanged within the hysteresis interval, the system can effectively avoid frequent mode switching caused by small fluctuations in the environmental awareness index near the critical value, thereby reducing system instability and lowering the additional computational burden and operational risks brought about by mode switching. As a result, the continuity and safety of tunneling operations are better guaranteed, and the practicality and reliability of the entire shaft tunneling disturbance parameter optimization method are improved.

[0220] In some preferred embodiments, it is assumed that the efficiency zone of the environmental awareness index is set to 0 to 0.5, the exploration zone to 0.7 to 1.0, and the transition zone to 0.5 to 0.7. Within the transition zone, the upper threshold for switching from operating mode to exploration mode is set to 0.65, and the lower threshold for switching from exploration mode to operating mode is set to 0.55. Specifically, when the system is currently in operating mode and the environmental awareness index gradually increases from 0.4 to 0.68, since 0.68 exceeds the upper threshold of 0.65 for the first time, the system will immediately switch from operating mode to exploration mode.

[0221] Conversely, when the system is currently in exploration mode and the environmental awareness index gradually decreases from 0.8 to 0.52, the system will immediately switch from exploration mode to operation mode because 0.52 is below the lower limit threshold of 0.55 for the first time.

[0222] If, while the system is in exploration mode, the environmental awareness index fluctuates from 0.7 to 0.6, and then to 0.63, since these values ​​are all within the hysteresis range (0.55 to 0.65) and do not cross the lower threshold of 0.55 for the first time, the system will remain in exploration mode until the index first falls below 0.55. This mechanism ensures the stability and predictability of mode switching.

[0223] refer to Figure 2 , Figure 2 This is a schematic diagram of a shaft excavation disturbance parameter optimization system based on big data, provided in an embodiment of the present invention, comprising:

[0224] The input end is used to acquire multiple information streams and evaluate the degree of correlation between the information streams to determine the level of awareness of the environmental state; acquire environmental response information and compare the environmental response information with pre-stored environmental feature patterns to determine the degree of matching of environmental features;

[0225] The statistics section is used to analyze the fluctuation range of the information flow and the frequency of abnormal events in order to determine the severity of environmental changes.

[0226] The output terminal is used to output a quantitative environmental cognition index based on the degree of cognition of the environmental state, the degree of matching of the environmental features, and the severity of the environmental changes; and to switch modes based on the environmental cognition index, the modes including an efficiency-oriented operation mode and an exploration mode that prioritizes safety and actively acquires information.

[0227] This system aims to address the problem in existing deep shaft tunneling operations where traditional systems, due to limitations in data analysis logic, cannot accurately interpret sensor signal changes when encountering complex geological anomalies, leading to erroneous judgments and operations, and further deteriorating downhole conditions. By setting up input, statistical, and output terminals, this system can systematically acquire and process multi-source information streams, comprehensively assess the level of environmental awareness, the degree of matching of environmental characteristics, and the severity of environmental changes, and then output a quantified environmental awareness index. Based on this index, the system can intelligently switch between an efficiency-oriented operation mode and an exploration mode that prioritizes safety and proactively acquires information. This avoids blind operation in uncertain environments and proactively adapts and optimizes tunneling strategies, effectively responding to complex and dynamic geological anomalies and ensuring the safety and efficiency of tunneling operations.

[0228] In some embodiments of this application, the specific details of acquiring multiple information streams, evaluating the correlation between information streams, acquiring environmental response information, and comparing environmental characteristic patterns have already been described in the above embodiments, and will not be repeated here. It should be emphasized that the input end of this application can be configured as a combination of a hardware interface module and a data preprocessing module. The hardware interface module can include various sensor interfaces, such as analog-to-digital converters, communication protocol interfaces such as CAN bus, Ethernet, etc., for receiving raw data from tunneling machines, geological exploration equipment, and environmental monitoring equipment in real time. The data preprocessing module can be configured to perform preliminary cleaning, format conversion, and time synchronization on the received raw data. For example, simple filtering algorithms can be used to remove high-frequency noise, or linear interpolation can be used to process missing data. This input end can be designed to preliminarily evaluate the correlation between information streams using preset rules or simple statistical methods. For example, by calculating the correlation coefficient of different sensor data within a fixed time window, or by using manually set thresholds to determine the strength of the correlation. Meanwhile, environmental response information can be obtained by directly reading sensor data, while the comparison with pre-stored environmental feature patterns can be done using a simple pattern matching algorithm based on distance metrics, such as Euclidean distance, to determine the degree of matching.

[0229] Furthermore, the specific details of the statistical information flow fluctuation range and the frequency of abnormal events have already been described in the above embodiments, and will not be repeated here. It should be emphasized that the statistical end of this application can be configured as a data analysis and processing unit. This unit can be designed to determine the fluctuation range by periodically calculating the difference between the maximum and minimum values ​​of the information flow within a preset time period, or simply calculating its standard deviation. Regarding the frequency of abnormal events, the statistical end can employ a detection method based on a fixed threshold. For example, when a sensor data point exceeds its historical average plus or minus three times the standard deviation, it is marked as an abnormal event and counted. These statistical results are used to preliminarily determine the severity of environmental changes.

[0230] Furthermore, the above embodiments have already described the specific details of outputting an environmental cognition index based on the level of awareness of the environmental state, the degree of matching of environmental features, and the severity of environmental changes, and switching modes based on this index, which will not be repeated here. It is important to emphasize that the output end of this application can be configured as a decision-making and control module. This module can be designed to integrate data from the input end and the statistical end through a simple weighted summation method to calculate the environmental cognition index. For example, a preset fixed weight can be assigned to the level of awareness, the degree of matching, and the severity, and then a linear combination can be performed. The logic for mode switching can be set to judge based on a preset fixed threshold. For example, when the environmental cognition index is higher than a certain fixed high threshold, the system switches to an efficiency-oriented operation mode; when it is lower than a certain fixed low threshold, it switches to an exploration mode. This output end can also include a human-computer interaction interface for displaying the current environmental cognition index and operation mode, and receiving operator instructions.

[0231] The big data-based vertical shaft excavation disturbance parameter optimization system proposed in this application significantly improves the adaptability, safety, and optimization capabilities of vertical shaft excavation systems under complex geological conditions through its unique system architecture, namely the collaborative work of the input, statistical, and output ends. Compared with existing technologies, this system can effectively solve the problems of misjudgment and operational degradation caused by the limitations of data analysis logic in traditional systems when encountering complex geological anomalies.

[0232] Specifically, existing technologies often rely on empirical models established under homogeneous and stable geological conditions. When the environment changes abruptly, they struggle to accurately interpret drastic sensor signals, leading to erroneous parameter adjustment commands. This system comprehensively acquires and evaluates multi-source information streams at the input end, precisely quantifies the severity of environmental changes at the statistical end, and generates a quantified environmental perception index at the output end. This multi-dimensional and dynamic evaluation mechanism enables the system to understand the true state of the current tunneling environment more comprehensively and precisely, avoiding erroneous decisions caused by insufficient information or misjudgment in complex environments, as is common with traditional methods. For example, when the tunneling system enters a highly fractured fault zone, a traditional system might still attempt to maintain high-efficiency operation, thus exacerbating the risk; while this system can quickly identify the complexity and uncertainty of the environment through its various functional ends, reducing the environmental perception index.

[0233] Furthermore, this system intelligently switches modes based on the environmental awareness index at the output end, providing both an efficiency-driven operation mode and an exploration mode that prioritizes safety and proactively acquires information. When the environmental awareness index decreases, the system can promptly switch from the efficiency-driven operation mode to the exploration mode. This mode-switching mechanism allows the system to adapt more flexibly to dynamically changing geological environments, prioritizing safety and proactively making minor parameter adjustments to obtain more environmental information. This avoids the risks of blindly pursuing efficiency in uncertain environments and provides a more reliable data foundation for subsequent parameter optimization. This contrasts sharply with the limitations of existing technologies, which struggle to distinguish between genuine stratigraphic responses and erroneous intervention information after analytical logic failures, thus hindering the rapid reconstruction of reliable optimization strategies. Through its intelligent mode switching, this system can gradually establish new geological response models, avoiding the repetition of previous erroneous interventions, thereby achieving robust adaptation and optimization to complex and dynamic geological anomalies.

[0234] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for optimizing disturbance parameters in vertical shaft excavation based on big data, characterized in that, include: Acquire multiple information streams and assess the degree of correlation between the information streams to determine the level of awareness of the environmental state; Acquire environmental response information and compare it with pre-stored environmental feature patterns to determine the degree of matching of environmental features; The fluctuation range of the information flow and the frequency of abnormal events are statistically analyzed to determine the severity of environmental changes. Based on the level of awareness of the environmental state, the degree of matching of the environmental features, and the severity of the environmental changes, a quantitative environmental awareness index is output. The mode is switched according to the environmental cognition index, multiple key information flows are monitored in real time, the rate of change of the information flow within a short time window is continuously calculated, and an early warning threshold for the rate of change is set. When the rate of change exceeds the early warning threshold, an emergency response signal is triggered. When the emergency response signal is received, the system is forced to switch from the efficiency-oriented operation mode to the exploration mode. The mode includes the efficiency-oriented operation mode and the exploration mode that prioritizes safety and actively acquires information. When in the exploration mode, the limitation range of the operation parameters is generated based on the currently identified environmental features and the recorded restriction operation information; Within the limits of the operating parameters, perform parameter adjustment operations and obtain environmental response information; The parameter adjustment operation and execution time are recorded. Based on the expected impact of the parameter adjustment operation on the equipment and environment, the signal components corresponding to the parameter adjustment operation are separated from the environmental response information to obtain information reflecting the true characteristics of the environment. Key sensor data of the tunneling machine are collected, and noise suppression and timestamp calibration are performed on the key sensor data. A causal response model is established for each small-amplitude parameter adjustment operation. Based on the operation content and the execution time, the system response curve that should be generated on each sensor is predicted in real time. The system response curve is separated from the key sensor data to obtain a preliminary formation response signal. High-resolution time-frequency analysis is performed on the preliminary formation response signal to obtain the energy distribution of the preliminary formation response signal at different time points and frequencies. Transient patterns are extracted from the energy distribution and compared with a predefined geological event feature library to instantly identify the geological micro-event type. Geological event parameters are estimated based on the geological micro-event type. The geological micro-event type and the geological event parameters are integrated into the current working condition description to adjust the uncertainty index. The information reflecting the true characteristics of the environment is compared with the pre-stored environmental feature patterns to identify the current environmental characteristics and estimate environmental parameters. The system monitors the results of parameter adjustment operations. When these operations lead to environmental degradation, the system records the parameter adjustment operation and the corresponding environmental conditions, and adds the operation to the restriction operation information. It then traces back the operation sequence preceding the environmental degradation, including system-issued commands, command execution times, tunneling machine parameters, and mud circulation system status. The system retrieves sensor trust weights and calibration information from the execution of the operation sequence. Finally, it analyzes the system-issued commands, command execution times, tunneling machine parameters, mud circulation system status, and sensor trust weights and calibration information. The system recalculates the expected changes in sensor data under conditions of no internal sensor bias and considering only the system's own intervention. It compares the sensor data during the actual deterioration period with the recalculated expected changes. Based on the comparison results, it identifies the expected secondary effects caused by the system's own intervention, the perception distortion caused by internal sensor bias, and the anomalies caused by changes in formation characteristics. Based on the identification results, it determines the main cause of the environmental state deterioration. When the main cause is changes in formation characteristics and these changes exceed the system's adaptability, the parameter adjustment operation and the actual formation characteristic information are added to the limiting operation information.

2. The method for optimizing vertical shaft excavation disturbance parameters based on big data according to claim 1, characterized in that, When parameter adjustments are made at the edge of a cavity, the parameter adjustment operation and execution time are recorded. Based on the expected impact of the parameter adjustment operation on the equipment and environment, the signal components corresponding to the parameter adjustment operation are separated from the environmental response information to obtain information reflecting the true characteristics of the environment, including: Real-time acquisition of key sensor data on cutter head torque, vibration, and annular pressure; High-resolution time-frequency analysis is performed on the key sensor data to generate a time-frequency graph; In the time-frequency diagram, signal patterns caused by the tunneling machine's own operation are identified, including the delayed response, energy distribution, and duration after the operation. In the time-frequency graph, transient signal patterns caused by micro-collapse or stress release of rock mass at the edge of the cavity are identified. These transient signal patterns include energy surges, frequency shifts, and waveform shapes. By comparing the time-frequency characteristics of the signal pattern and the transient signal pattern, the signal components generated by the tunneling machine's own operation are separated from the transient signal pattern; The separated transient signal patterns are used as information reflecting the true characteristics of the environment.

3. The method for optimizing vertical shaft excavation disturbance parameters based on big data according to claim 1, characterized in that, The process involves recording the parameter adjustment operation and its execution time, and based on the expected impact of the parameter adjustment operation on the equipment and environment, separating the signal components corresponding to the parameter adjustment operation from the environmental response information to obtain information reflecting the true characteristics of the environment, including: The system collects data from multiple sensors in the tunnel boring machine cutterhead, propulsion system, and mud circulation system in real time. The sensor data includes vibration signals, pressure signals, torque signals, and flow signals. Based on the type and intensity of the parameter adjustment operation, as well as the current status of the tunneling machine and the tunneling speed, calculate the instantaneous force distribution of the parameter adjustment operation on the contact area between the tunneling machine cutterhead and the stratum. Based on the instantaneous force distribution, combined with the acoustic propagation velocity, attenuation coefficient, and reflection coefficient of the current formation medium, the time-domain response waveform that should be generated on the multiple sensors after the force propagates in the formation is predicted. The amplitude and phase of the time-domain response waveform are adjusted according to the water content, porosity, and degree of consolidation of the formation medium. An adaptive separation method based on signal feature matching separates the adjusted predicted time-domain response waveform from real-time sensor data to obtain information reflecting the true characteristics of the formation.

4. The method for optimizing vertical shaft excavation disturbance parameters based on big data according to claim 1, characterized in that, The mode switching based on the environmental awareness index includes: Real-time acquisition of environmental awareness index; Based on the numerical range of the environmental awareness index, it is divided into an efficiency zone, an exploration zone, and a transition zone between the efficiency zone and the exploration zone; A hysteresis interval for mode switching is set in the transition zone. The hysteresis interval has an upper threshold for switching from the operation mode to the exploration mode and a lower threshold for switching from the exploration mode to the operation mode. The upper threshold is higher than the lower threshold.

5. The method for optimizing vertical shaft excavation disturbance parameters based on big data according to claim 4, characterized in that, The mode switching based on the environmental awareness index includes: When the environmental awareness index gradually increases from the efficiency zone and reaches the upper limit threshold for the first time, the operation mode is switched to the exploration mode. When the environmental awareness index gradually decreases from the exploration zone and reaches the lower limit threshold for the first time, the system switches from the exploration mode to the operation mode. Within the hysteresis interval, when the environmental perception index fluctuates between the upper and lower thresholds, the current operating mode remains unchanged.

6. A big data-based vertical shaft excavation disturbance parameter optimization system for performing the method of any one of claims 1-5, characterized in that, include: The input terminal is used to acquire multiple information streams and evaluate the degree of correlation between the information streams in order to determine the level of awareness of the environmental state. Acquire environmental response information and compare it with pre-stored environmental feature patterns to determine the degree of matching of environmental features; The statistics section is used to analyze the fluctuation range of the information flow and the frequency of abnormal events in order to determine the severity of environmental changes. The output end is used to switch modes according to the environmental cognition index, monitor multiple key information streams in real time, continuously calculate the rate of change of the information streams within a short time window, set a warning threshold for the rate of change, and trigger an emergency response signal when the rate of change exceeds the warning threshold. When the emergency response signal is received, the system is forced to switch from an efficiency-oriented operation mode to an exploration mode. The modes include an efficiency-oriented operation mode and an exploration mode that prioritizes safety and actively acquires information. When in the exploration mode, the limitation range of the operation parameters is generated based on the currently identified environmental features and the recorded restriction operation information; Within the limits of the operating parameters, perform parameter adjustment operations and obtain environmental response information; The parameter adjustment operation and execution time are recorded. Based on the expected impact of the parameter adjustment operation on the equipment and environment, the signal components corresponding to the parameter adjustment operation are separated from the environmental response information to obtain information reflecting the true characteristics of the environment. Key sensor data of the tunneling machine are collected, and noise suppression and timestamp calibration are performed on the key sensor data. A causal response model is established for each small-amplitude parameter adjustment operation. Based on the operation content and the execution time, the system response curve that should be generated on each sensor is predicted in real time. The system response curve is separated from the key sensor data to obtain a preliminary formation response signal. High-resolution time-frequency analysis is performed on the preliminary formation response signal to obtain the energy distribution of the preliminary formation response signal at different time points and frequencies. Transient patterns are extracted from the energy distribution and compared with a predefined geological event feature library to instantly identify the geological micro-event type. Geological event parameters are estimated based on the geological micro-event type. The geological micro-event type and the geological event parameters are integrated into the current working condition description to adjust the uncertainty index. The information reflecting the true characteristics of the environment is compared with the pre-stored environmental feature patterns to identify the current environmental characteristics and estimate environmental parameters. The system monitors the results of parameter adjustment operations. When these operations lead to environmental degradation, the system records the parameter adjustment operation and the corresponding environmental conditions, and adds the operation to the restriction operation information. It then traces back the operation sequence preceding the environmental degradation, including system-issued commands, command execution times, tunneling machine parameters, and mud circulation system status. The system retrieves sensor trust weights and calibration information from the execution of the operation sequence. Finally, it analyzes the system-issued commands, command execution times, tunneling machine parameters, mud circulation system status, and sensor trust weights and calibration information. The system recalculates the expected changes in sensor data under conditions of no internal sensor bias and considering only the system's own intervention. It compares the sensor data during the actual deterioration period with the recalculated expected changes. Based on the comparison results, it identifies the expected secondary effects caused by the system's own intervention, the perception distortion caused by internal sensor bias, and the anomalies caused by changes in formation characteristics. Based on the identification results, it determines the main cause of the environmental state deterioration. When the main cause is changes in formation characteristics and these changes exceed the system's adaptability, the parameter adjustment operation and the actual formation characteristic information are added to the limiting operation information.

Images