A shield cutter disengagement real-time sensing alarm system

Abstract

The present invention discloses a shield cutter disengagement sensing and alarm system, which relates to the technical field of shield construction, and includes a historical data processing module, a real-time data processing module, a shield cutter disengagement criterion and a self-learning processing module. The beneficial effects are: the present invention can realize real-time sensing and alarm of shield cutter disengagement based on data change trends according to the historical data characteristics of big data and the changes in real-time data streams, combined with the data change characteristics of abnormal cutter disengagement; the historical data characteristics, the real-time data stream characteristics and the cutter disengagement data characteristics are combined to realize independent sensing and judgment of cutter disengagement of multiple line data; the data change conditions are graded according to the health status and change conditions of the real-time data, and the algorithm hyperparameters are modified by forward feedback, so that the algorithm has the ability of self-adaptation and self-learning; thereby providing shield construction personnel with more accurate and timely sensing and alarm of shield cutter disengagement.

Description

Technical Field

[0001] The invention relates to the technical field of shield construction, in particular to a shield cutter disengagement instantaneous sensing and alarm system. Background Art

[0002] Shield machines are large, high-end equipment used to construct urban subways, municipal pipelines, water diversion projects, and tunnels through mountains and seas. With the continuous improvement of domestic shield machine design and manufacturing capabilities and the continuous reduction of costs in recent years, the domestic shield machine market has become prosperous. With the rapid and steady development of my country's economy, the number of tunnels constructed using the shield method is increasing. Compared with traditional construction methods, shield construction offers advantages such as a high degree of automation, high tunnel excavation accuracy, precise tunnel cross-sectional shape, reduced labor intensity, safer construction, and overall high construction efficiency. The advantages of using shield TBMs are particularly evident in underwater tunnels crossing rivers, lakes, or offshore areas, as well as mountain tunnels with low-quality surrounding rock.

[0003] Shield cutters are primarily installed on the cutterheads of earth pressure balance (EPB) and slurry balance (SBB) shields. They are located on either side of the cutterhead opening. The cutters penetrate the soil through their blades and heads, cutting the soil as the cutterhead rotates. These cutters are primarily suited for loose soils with particle sizes less than 40 mm, such as sand, gravel, and clay. Regarding cutter detachment during shield construction, EPB and SBB cutters operate in complex, alternating load-bearing tunnel faces. These alternating loads and sudden encounters with hard geology can easily lead to fatigue fracture or ultimate failure, resulting in frequent cutter detachment. Real-time monitoring of the health of all cutters in the tunnel face is difficult. Using oil pressure sensors to detect cutter detachment presents challenges such as high damage risk, high cost, and inability to accommodate existing shield machines. Consequently, cutter detachment is often difficult to detect. The falling of the tooth cutters will make excavation difficult, causing the torque of the cutterhead to increase, the propulsion speed to decrease, and the propulsion force to increase, resulting in increased energy consumption in the excavation process, increased wear of key components such as the main bearing, and reduced life of the shield machine. At the same time, it will increase the wear of other cutting tools on the disc, increase the consumption of lubricating grease, and easily induce other engineering accidents.

[0004] The traditional use of hydraulic sensors to detect cutter detachment is difficult to scale. The experience of shield tunneling operators and construction personnel in determining cutter detachment is subject to individual cognitive differences and cannot be monitored in real time. Furthermore, manual judgment is often only detected after an accident or a serious anomaly in tunneling. Therefore, rapid and accurate cutter detachment assessment and early warning are urgent needs for shield tunneling. To effectively address the cutter detachment problem during shield tunneling, a cutter detachment detection and alarm system was designed. This system detects and alarms cutter detachment in real time based on data change trends, combining historical data features, real-time data stream characteristics, and cutter detachment data characteristics. By combining historical data features, real-time data stream features, and cutter detachment data characteristics, it achieves independent cutter detachment detection and judgment for multiple lines of data. Data changes are classified based on the health and changes of real-time data, and algorithm hyperparameters are modified through feedback loops, enabling the algorithm to be adaptive and self-learning. This provides shield tunneling operators with more accurate and timely cutter detachment detection and alarms. Summary of the Invention

[0005] The purpose of the present invention is to provide a shield cutter disengagement instantaneous sensing alarm system in order to solve the above problems.

[0006] The present invention achieves the above-mentioned purpose through the following technical solutions:

[0007] A shield cutter disengagement sensing and alarm system comprises a historical data processing module, a real-time data processing module, a shield cutter disengagement criterion, a self-learning processing module, a shield cutter disengagement sensing and alarm system, a shield TBM big data platform, a data collector, and an on-site receiving end. The historical data processing module is electrically connected to the real-time data processing module, the real-time data processing module is electrically connected to the shield cutter disengagement criterion, the shield cutter disengagement criterion is electrically connected to the self-learning processing module, the self-learning processing module is electrically connected to the shield cutter disengagement sensing and alarm system, the shield cutter disengagement sensing and alarm system is electrically connected to the shield TBM big data platform, the data collector is electrically connected to the shield TBM big data platform, and the shield TBM big data platform is electrically connected to the on-site receiving end.

[0008] Preferably: the shield cutter disengagement real-time perception and alarm system is characterized in that: the historical data processing module will read the historical data information of all lines under construction, and combine the real-time data processing module, the shield cutter disengagement criterion, and the self-learning processing module to perceive and warn of cutter disengagement, and finally transmit the cutter disengagement information to the on-site receiving end through the shield TBM big data platform.

[0009] Preferably, the historical data processing module includes:

[0010] Historical data extraction: Multi-threaded historical data extraction is performed on all lines running on the shield TBM big data platform. A thread is opened for each shield tunneling data line. Historical data is extracted by line and the tunneling status, startup and stable segments, and short data segments are determined.

[0011] Historical data cleaning: Perform absolute quantity limit cleaning based on equipment characteristics and relative quantity cleaning based on continuous data patterns on historical data, remove startup segment data, and remove too short data segments;

[0012] Characteristic data analysis: Analyze the properties and distribution of historical data to obtain the slope, fluctuation, mean, and other characteristic data values ​​of propulsion force, propulsion speed, and cutter head torque parameters;

[0013] Shield status query: query the status of multiple-line shield machines and screen the shield machines in the excavation state for the next step;

[0014] Diameter type query: query the type and diameter information of the shield machine. For the earth pressure or slurry balance shield machines in excavation, establish an independent thread to further process the relevant data separately.

[0015] Preferably, the real-time data processing module includes:

[0016] Real-time data accumulation: real-time data accumulation for each excavation line;

[0017] Equivalent data detection: The accumulated data is tested for startup and stable segments, the number of point groups is detected, isolated point data is eliminated, and then the real-time data volume is tested to see if it meets the equivalent requirements;

[0018] Real-time data cleaning: perform absolute data cleaning on the accumulated real-time data every minute;

[0019] Sliding window processing: Smoothes the sliding window data and discards some data at the end of the sliding window to keep the valid ten-minute sliding window data;

[0020] Characteristic data analysis: Perform characteristic data processing on real-time data similar to historical data to obtain the values ​​of corresponding characteristic data such as the slope, fluctuation, and mean of each parameter.

[0021] Preferably, the shield cutter shedding criterion module includes:

[0022] Earth Pressure Balance Shield: Earth Pressure Balance Shield machine in the line under construction;

[0023] Propulsion speed change I: the propulsion speed suddenly decreases by 10%-15% based on the normal value of the line;

[0024] Propulsion force change I: The total propulsion force suddenly increases by 10% based on the normal value of the current route;

[0025] Cutter head torque change I: The cutter head torque suddenly decreases by 10% based on the normal value of the current line;

[0026] Shield cutter shedding criterion: programmed judgment rules for shield cutter shedding;

[0027] Slurry shield: a slurry shield machine used in the line under construction;

[0028] Propulsion speed change II: the propulsion speed suddenly decreases by 10%-15% based on the normal value of the line;

[0029] Propulsion force change II: The total propulsion force suddenly increases by 10% based on the normal value of the current route;

[0030] Cutter head torque change II: The cutter head torque suddenly decreases by 10% based on the normal value of the current line.

[0031] Preferably: the self-learning processing module includes:

[0032] Real-time data perception: Compare the data characteristics of real-time data with those of health data, and perceive the data deviation of individual data according to the deviation degree;

[0033] Data classification: Classify the perception results of single data items and process them differently for perception results of different levels; Feedback: For data with small deviation and no anomalies, a larger penalty coefficient is applied to the data, while for data with large deviation and no anomalies, a smaller penalty coefficient is applied to the data. For data with anomalies or sudden large deviations, a smaller or no feedback is applied to the data;

[0034] Modify algorithm hyperparameters: Modify the corresponding algorithm hyperparameters based on the forward feedback and adjust the size of the corresponding hyperparameters.

[0035] Preferably, the shield cutter disengagement sensing alarm system transmits the shield cutter disengagement sensing alarm to the shield TBM big data platform, and then pushes it to the on-site receiving terminal.

[0036] Preferably: the shield cutter disengagement perception alarm system performs the historical data processing module operation on all lines under construction, and performs the real-time data processing module operation on all excavation lines, compares the historical data with the real-time data to perform the shield cutter disengagement judgment operation, passes the judgment data to the self-learning processing module to adjust the algorithm parameters, and then performs the shield cutter disengagement perception alarm.

[0037] Preferably, on-site construction data is collected by the data collector and transmitted to the TBM big data platform for storage. On the TBM big data platform, the historical data processing module extracts historical data of the corresponding line, cleans it, and performs feature data analysis to obtain feature data characteristics, which are then transmitted to the cutter disengagement detection and alarm system. The real-time data processing module extracts real-time data of the corresponding line, cleans it, and performs feature data analysis to obtain feature data characteristics, which are then transmitted to the cutter disengagement detection and alarm system. The self-learning processing module extracts feature data of the current excavation status data and historical data of the corresponding line, and transmits them to the cutter disengagement detection and alarm system. The cutter disengagement detection and alarm system comprehensively processes the historical data, real-time data, and self-learning parameter information, performs cutter disengagement detection and judgment, and issues a warning, and transmits the warning status to the TBM big data platform. The TBM big data platform pushes the information to the on-site receiving terminal for on-site access.

[0038] Compared with the existing technology, the beneficial effects of the present invention are: the present invention can realize real-time perception and alarm of shield cutter shedding based on data change trends according to the historical data characteristics of big data and the changes in real-time data streams, combined with the data change characteristics of abnormal cutter shedding; combine historical data characteristics, real-time data stream characteristics and cutter shedding data characteristics to realize independent perception and judgment of cutter shedding of multiple line data; classify data changes according to the health status and changes of real-time data, and perform forward feedback to modify algorithm hyperparameters, so that the algorithm has the ability of self-adaptation and self-learning; and thus provide shield construction personnel with more accurate and timely perception and alarm of shield cutter shedding. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0040] Figure 1 This is a structural block diagram of a shield cutter disengagement instantaneous sensing and alarm system described in the present invention.

[0041] Figure 2 This is a flowchart of the working process of a shield cutter disengagement instantaneous sensing and alarm system described in the present invention.

[0042] The following are the descriptions of the reference numerals:

[0043] 1. Historical data processing module; 2. Real-time data processing module; 3. Shield cutter dropout judgment; 4. Self-learning processing module; 5. Shield cutter dropout detection alarm; 6. Historical data extraction; 7. Historical data cleaning; 8. Feature data analysis; 9. Shield status query; 10. Diameter type query; 11. Real-time data accumulation; 12. Equivalent data detection; 13. Real-time data cleaning; 14. Sliding window processing; 15. Feature data analysis; 16. Earth pressure balance shield; 17. Advancing speed change I; 18. Propelling force change I; 19. Cutterhead torque change I; 20. Shield cutter detachment criterion; 21. Slurry balance shield; 22. Advancing speed change II; 23. Propelling force change II; 24. Cutterhead torque change II; 25. Real-time data perception; 26. Data classification; 27. Feedback; 28. Modification of algorithm hyperparameters; 29. ​​Data collector; 30. Shield cutter detachment detection and alarm system; 31. TBM big data platform; 32. On-site receiving terminal. DETAILED DESCRIPTION

[0044] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", etc. are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, features defined as "first", "second", etc. may explicitly or implicitly include one or more of such features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0045] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be internal communication between two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood based on specific circumstances.

[0046] The present invention will be further described below in conjunction with the accompanying drawings:

[0047] like Figure 1-Figure 2As shown, a shield cutter disengagement perception and alarm system includes a historical data processing module 1, a real-time data processing module 2, a shield cutter disengagement criterion 3, a self-learning processing module 4, a shield cutter disengagement perception and alarm 5, a shield TBM big data platform 31, a data collector 29, and an on-site receiving terminal 32. The historical data processing module 1 is electrically connected to the real-time data processing module 2, the real-time data processing module 2 is electrically connected to the shield cutter disengagement criterion 3, the shield cutter disengagement criterion 3 is electrically connected to the self-learning processing module 4, the self-learning processing module 4 is electrically connected to the shield cutter disengagement perception and alarm 5, the shield cutter disengagement perception and alarm 5 is electrically connected to the shield TBM big data platform 31, the data collector 29 is electrically connected to the shield TBM big data platform 31, and the shield TBM big data platform 31 is electrically connected to the on-site receiving terminal 32.

[0048] The historical data processing module 1 will read the historical data information of all lines under construction, and combine with the real-time data processing module 2, the shield cutter fall-off criterion 3, and the self-learning processing module 4 to perceive and warn of cutter fall-off, and finally transmit the cutter fall-off information to the on-site receiving terminal 32 through the shield TBM big data platform 31.

[0049] The historical data processing module 1 specifically includes:

[0050] Historical Data Extraction 6: Perform multi-threaded historical data extraction on all lines running on the shield TBM big data platform 31. A thread is opened for each shield tunneling data line. Historical data is extracted by line and tunneling status, startup and stable segments, and short data segments are determined.

[0051] Historical data cleaning 7: Perform absolute quantity limit cleaning based on equipment characteristics and relative quantity cleaning based on continuous data patterns on historical data, remove startup segment data, and remove too short data segments;

[0052] Characteristic data analysis 8: Analyze the properties and distribution of historical data to obtain the slope, fluctuation, mean, and other characteristic data values ​​of propulsion force, propulsion speed, and cutter head torque parameters;

[0053] Shield status query 9: Query the status of multiple-line shield machines and perform further screening for shield machines in the excavation state;

[0054] Diameter type query 10: Query the type and diameter information of the shield machine. For the earth pressure or slurry balance shield machine in excavation, establish an independent thread to further process the relevant data separately.

[0055] The real-time data processing module 2 specifically includes:

[0056] Real-time data accumulation 11: Real-time data accumulation for each excavation line.

[0057] Equivalent data detection 12: Perform startup and stable segment detection on the accumulated data, detect the number of point groups, eliminate isolated point data, and then detect whether the real-time data volume meets the equivalent requirements;

[0058] Real-time data cleaning 13: Perform absolute data cleaning on the accumulated real-time data every one minute;

[0059] Sliding window processing 14: Smoothing the sliding window data, while discarding some data at the end of the sliding window, and keeping the valid ten-minute sliding window data;

[0060] Feature data analysis 15: Perform feature data processing on real-time data similar to that of historical data to obtain the values ​​of corresponding feature data such as the slope, fluctuation, and mean of each parameter.

[0061] The shield cutter shedding criterion module 3 specifically includes:

[0062] Earth Pressure Balance Shield 16: Earth Pressure Balance Shield Machine on the line under construction.

[0063] Propulsion speed change I17: The propulsion speed suddenly decreases by 10%-15% based on the normal value of the line;

[0064] Propulsion force change I18: The total propulsion force suddenly increases by 10% based on the normal value of the current route;

[0065] Cutter head torque change I19: The cutter head torque suddenly decreases by 10% based on the normal value of the current line;

[0066] Shield cutter shedding criterion 20: programmed judgment rules for shield cutter shedding;

[0067] Slurry shield 21: a slurry shield machine on the line under construction;

[0068] Propulsion speed change II22: The propulsion speed suddenly decreases by 10%-15% based on the normal value of the line;

[0069] Propulsion force change II23: The total propulsion force suddenly increases by 10% based on the normal value of the current route;

[0070] Cutter head torque change II24: The cutter head torque suddenly decreases by 10% based on the normal value of the current line.

[0071] The self-learning processing module 4 specifically includes:

[0072] Real-time data perception 25: Compare the data characteristics of real-time data with those of health data, and perceive the data deviation of individual data according to the deviation degree;

[0073] Data classification 26: Classify the perception results of a single data item and process the perception results of different levels differently;

[0074] Feedback 27: For data with small deviation and no anomalies, a larger penalty coefficient is applied to the data. For data with large deviation and no anomalies, a smaller penalty coefficient is applied to the data. For data with anomalies or sudden large deviations, a smaller or no feedback is applied to the data.

[0075] Modify algorithm hyperparameters 28: Modify the corresponding algorithm hyperparameters based on the forward feedback and adjust the size of the corresponding hyperparameters.

[0076] The above-mentioned shield cutter disengagement sensing and alarm system reads historical data, real-time data, and shield status. The shield cutter disengagement sensing and alarm system 30 transmits the shield cutter disengagement sensing and alarm 5 to the shield TBM big data platform 31, and then pushes it to the on-site receiving terminal 32.

[0077] The above-mentioned shield cutter disengagement sensing and alarm system specifically includes the shield cutter disengagement sensing and alarm system 30 operating the historical data processing module 1 for all under-construction lines, and operating the real-time data processing module 2 for all excavation lines, comparing the historical data and the real-time data to perform the shield cutter disengagement judgment 3, passing the judgment data to the self-learning processing module 4 to adjust the algorithm parameters, and then performing the shield cutter disengagement sensing and alarm.

[0078] The above-mentioned shield cutter disengagement real-time perception alarm system collects the on-site construction data by the data collector 29 and transmits it to the shield TBM big data platform 31 for storage. The shield TBM big data platform 31 uses two Intel (R) Xeon (R) CPU E5-2650 v4 @ 2.20GHz processors with 128GB of memory as cache servers; uses two Intel (R) Xeon (R) CPU E5-2698R v4 @ 2.20GHz server processors with 128GB of memory and a distributed storage cluster with a hard disk storage capacity of 80TB as data storage servers; uses two Intel (R) Xeon (R) CPU E5-2650 v4 @2.20GHz processor, equipped with 128GB of memory, serves as a network publishing server; on the shield TBM big data platform 31, the historical data processing module 1 extracts the historical data of the corresponding line and performs cleaning and feature data analysis to obtain feature data features and transmits them to the cutter disengagement perception alarm system 30; the real-time data processing module 2 extracts the real-time data of the corresponding line and performs cleaning and feature data analysis to obtain feature data features and transmits them to the cutter disengagement perception alarm system 30; the self-learning processing module 4 extracts the current excavation status data and feature data of the historical data of the corresponding line and transmits them to the cutter disengagement perception alarm system 30; the cutter disengagement perception alarm system 30 comprehensively processes the historical data information, real-time data information, and self-learning parameter information, performs cutter disengagement perception judgment and early warning, and transmits the early warning situation to the shield TBM big data platform 31. The shield TBM big data platform 31 pushes information to the on-site receiving terminal 32 for on-site access. The on-site receiving terminal 32 uses two Intel(R) Xeon(R) CPU E5-2650 v4 @2.20GHz processors and is equipped with 128GB of memory.

[0079] In summary:

[0080] The present invention can realize real-time perception and alarm of shield cutter shedding based on data change trends according to the historical data characteristics of big data and the changes in real-time data streams, combined with the data change characteristics of abnormal cutter shedding.

[0081] The present invention combines the characteristics of historical data, real-time data stream and serration shedding data to achieve independent real-time perception and judgment of serration shedding data for multiple lines.

[0082] The present invention classifies data changes according to the health status and changes of real-time data, and performs forward feedback to modify the algorithm hyperparameters, so that the algorithm has the ability of self-adaptation and self-learning; thereby providing shield construction personnel with more accurate and timely perception alarms when the shield cutter is disengaged.

[0083] Those skilled in the art should understand that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0084] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as the combination of the processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0085] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0086] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are performed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable device to implement the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0087] The basic principles, main features and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may be subject to various changes and improvements, and these changes and improvements shall fall within the scope of the invention claimed for protection.

Claims

1. A shield cutter disengagement sensing and alarm system, comprising a historical data processing module (1), a real-time data processing module (2), a shield cutter disengagement criterion module (3), a self-learning processing module (4), a shield cutter disengagement sensing and alarm module, a shield TBM big data platform (31), a data collector (29), and an on-site receiving terminal (32), characterized in that: The historical data processing module (1) is electrically connected to the real-time data processing module (2), the real-time data processing module (2) is electrically connected to the shield cutter fall-off criterion module (3), the shield cutter fall-off criterion module (3) is electrically connected to the self-learning processing module (4), the self-learning processing module (4) is electrically connected to the shield cutter fall-off sensing alarm module, the shield cutter fall-off sensing alarm module is electrically connected to the shield TBM big data platform (31), the data collector (29) is electrically connected to the shield TBM big data platform (31), and the shield TBM big data platform (31) is electrically connected to the on-site receiving terminal (32); The self-learning processing module (4) includes: Real-time data perception (25): Compare the data characteristics of real-time data with those of health data, and perceive the data deviation of individual data items according to the degree of deviation; Data classification (26): Classify the perception results of individual data items, and perform different processing on perception results of different levels; Feedback (27): For data with small deviation and no anomalies, a larger penalty coefficient is used for feedback. For data with large deviation and no anomalies, a smaller penalty coefficient is used for feedback. For data with anomalies or sudden large deviations, a smaller or no feedback is used. Modify algorithm hyperparameters (28): Modify the corresponding algorithm hyperparameters according to the forward feedback situation and adjust the size of the corresponding algorithm hyperparameters.

2. A shield cutter disengagement real-time sensing alarm system according to claim 1, characterized in that: The historical data processing module (1) reads the historical data information of all the lines under construction, and combines with the real-time data processing module (2), the shield cutter shedding judgment module (3), and the self-learning processing module (4) to sense and warn the cutter shedding, and finally transmits the cutter shedding information to the on-site receiving terminal (32) through the shield TBM big data platform (31).

3. The shield cutter disengagement real-time sensing alarm system according to claim 1, characterized in that: The historical data processing module (1) includes: historical data extraction (6): performing multi-threaded historical data extraction on all lines running on the shield TBM big data platform (31), opening a thread for each shield tunneling data line, extracting historical data according to the line and performing tunneling status judgment, startup segment and stable segment judgment, and too short data segment judgment; Historical data cleaning (7): Perform absolute quantity limit cleaning based on equipment characteristics and relative quantity cleaning based on continuous data patterns on historical data, remove startup segment data, and remove too short data segments; Characteristic data analysis (8): Analyze the properties of historical data and the data distribution to obtain the slope, fluctuation and mean of propulsion force, propulsion speed and cutter head torque parameters; Shield status query (9): query the status of multiple-line shield machines and perform the next step of screening for shield machines in the excavation state; Diameter type query (10): Query the type and diameter information of the shield machine. For the earth pressure or slurry balance shield machine in excavation, establish an independent thread to further process the relevant data separately.

4. The shield cutter disengagement real-time sensing alarm system according to claim 1, characterized in that: The real-time data processing module (2) includes: real-time data accumulation (11): accumulating real-time data for each excavation line; Equivalent data detection (12): Perform startup and stable segment detection and point group number detection on the accumulated data, remove isolated point data, and then detect whether the real-time data volume meets the equivalent requirements; Real-time data cleaning (13): Perform absolute data cleaning on the accumulated real-time data every one minute; Sliding window processing (14): Smoothing the sliding window data, discarding some data at the end of the sliding window, and keeping the valid ten-minute sliding window data; Characteristic data analysis (15): Perform characteristic data processing on real-time data similar to historical data to obtain the slope, fluctuation and mean of each parameter.

5. The shield cutter disengagement real-time sensing alarm system according to claim 1, characterized in that: The shield cutter fall-off criterion module (3) includes judgment rules for earth pressure balance shield (16) and slurry balance shield (21), and the judgment rules are: Propulsion speed change: the propulsion speed suddenly decreases by 10%-15% based on the normal value of the line; Propulsion force change: the total propulsion force suddenly increases by 10% based on the normal value of the current route; Cutter head torque change: The cutter head torque suddenly decreases by 10% based on the normal value of the current line; The change of propulsion speed, propulsion force and cutter head torque is used to realize the criterion of shield cutter shedding (20).

6. A shield cutter disengagement real-time sensing alarm system according to any one of claims 1 to 5, characterized in that: The historical data and the real-time data are compared to operate the shield cutter fall-off judgment module (3), and the judgment data is passed to the self-learning processing module (4) to adjust the algorithm hyperparameters, and then a sensing alarm is performed when the shield cutter falls off.

7. A shield cutter disengagement real-time sensing alarm system according to any one of claims 1 to 5, characterized in that: On-site construction data is collected by the data collector (29) and transmitted to the shield TBM big data platform (31) for storage. On the shield TBM big data platform (31), the historical data processing module (1) extracts the historical data of the corresponding line and performs cleaning and feature data analysis to obtain feature data features, which are then passed to the shield cutter disengagement perception alarm module; the real-time data processing module (2) extracts the real-time data of the corresponding line and performs cleaning and feature data analysis to obtain feature data features, which are then passed to the shield cutter disengagement perception alarm module; the self-learning processing module (4) extracts the current excavation state data and feature data of the historical data of the corresponding line and passes them to the shield cutter disengagement perception alarm module; the shield cutter disengagement perception alarm module comprehensively processes the historical data information, real-time data information, and self-learning parameter information, performs cutter disengagement perception judgment and early warning, and passes the early warning situation to the shield TBM big data platform (31), which pushes the information to the on-site receiving terminal (32) for on-site access.

Images