Anomaly diagnosis system, anomaly diagnosis method, and anomaly diagnosis program

Abstract

This invention provides an anomaly diagnosis system, an anomaly diagnosis method, and an anomaly diagnosis program that prevent mechanical problems occurring in shield tunneling machines, regardless of the operator's skill level. [Solution] The abnormality diagnosis system comprises a management device 30 and a diagnostic device 40. The diagnostic device 40 includes a normality determination model in which excavation information acquired during excavation by the shield tunneling machine 10 is used as an explanatory variable and the normality of the equipment equipped by the shield tunneling machine 10 is used as the objective variable. The normality represents the degree of similarity to the data set of excavation information when the equipment to be evaluated is operating normally. The control unit of the management device 30 acquires the normality from the normality determination model by inputting the excavation information into the normality determination model. The control unit of the management device 30 determines whether the normality acquired from the normality determination model meets a threshold, and outputs an alarm if the normality does not meet the threshold.

Description

Technical Field

[0001] The present invention relates to an abnormality diagnosis system, an abnormality diagnosis method, and an abnormality diagnosis program in the shield method.

Background Art

[0002] When constructing a tunnel, the shield method using a shield tunneling machine may be adopted (for example, Patent Document 1). The shield tunneling machine excavates the ground in front while covering the excavated inner peripheral surface by assembling segments.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] During the excavation work, many mechanical troubles occurring in the shield tunneling machine slightly appear as abnormal noises, vibrations, odors, etc. as precursors. Conventionally, preventive inspections have been carried out to prevent mechanical troubles, but the accuracy relied on the experience and five senses of skilled workers, so there was a part that depended on the operator. Therefore, a technology for preventing mechanical troubles occurring in the shield tunneling machine without depending on the skill level of the operator is desired.

Means for Solving the Problems

[0005] An abnormality diagnosis system that solves the above problems is an abnormality diagnosis system comprising a control unit that determines whether or not there is an abnormality in the shield tunneling machine, wherein the control unit inputs the excavation information acquired when the shield tunneling machine excavates as an explanatory variable and the normality of the equipment installed in the shield tunneling machine as an objective variable to obtain the normality from the normality determination model, determines whether or not the normality obtained from the normality determination model meets a threshold, and outputs an alarm if the normality does not meet the threshold. [Effects of the Invention]

[0006] According to the present invention, mechanical problems occurring in shield tunneling machines can be prevented regardless of the operator's skill level. [Brief explanation of the drawing]

[0007] [Figure 1] Figure 1 is an overall schematic diagram of the anomaly diagnosis system. [Figure 2] Figure 2 is an explanatory diagram of the hardware configuration of the embodiment. [Figure 3] Figure 3 is a block diagram of the anomaly diagnosis system. [Figure 4] Figure 4 is a flowchart of the anomaly diagnosis process. [Figure 5] Figure 5 is a flowchart of the underground fire detection process. [Figure 6] Figure 6 is a flowchart of the ground vibration prediction process. [Modes for carrying out the invention]

[0008] An embodiment of the anomaly diagnosis system, anomaly diagnosis method, and anomaly diagnosis program will be described below with reference to Figures 1 to 6. In this embodiment, the anomaly diagnosis system notifies workers of anomalies that occur in the shield tunneling machine during tunneling work to form a tunnel using the shield method, or anomalies that occur in the surrounding tunnel.

[0009] (Overall structure) As shown in Figure 1, the shield tunneling machine 10 excavates the ground while sequentially assembling segments SG1 with erectors 10A to form a lining T1 that covers the inner surface of the excavation. The shield tunneling machine 10 is equipped with shield jacks 10B as propulsion devices. The shield jacks 10B obtain forward propulsion force from the reaction force from the segments SG1 that make up the lining T1.

[0010] The shield tunneling machine 10 used in the shield tunneling method is equipped with a skin plate 11, a bulkhead 12, a cutter 13, a cutter motor 14, an additive injection pipe 15, a screw conveyor 16, a belt conveyor 17, a control device 20, and the like.

[0011] The skin plate 11 is a cylindrical steel member that forms the outer shell of the shield tunneling machine 10. The partition wall 12 is provided on the skin plate 11. The partition wall 12 partitions the chamber SP1 in the front part of the skin plate 11.

[0012] The cutter 13 is equipped with a plurality of bits 13A facing forward in the direction of excavation. The cutter 13 excavates into the ground by rotating the plurality of bits 13A. The plurality of bits 13A include a copy cutter, which is an extendable bit 13A. The cutter motor 14 is the drive source for rotating the cutter 13. The driving force of the cutter motor 14 is transmitted to the cutter 13 via a support arm.

[0013] The additive injection pipe 15 is a tubular member for injecting a foaming agent, which is a foamy additive. The foaming agent is obtained by foaming a liquid additive in a foaming device (not shown). The tip of the additive injection pipe 15 is located in front of the cutter 13. The foaming agent is injected onto the faceplate of the cutter 13.

[0014] The excavated soil, drilled by the cutter 13, is mixed with a foaming agent by the rotation of the cutter 13, thereby increasing its fluidity. The excavated soil mixed with the foaming agent then flows into chamber SP1 as a mixed soil.

[0015] The screw conveyor 16 takes in the mixed soil that has flowed into the chamber SP1 into the working space on the rear side of the partition wall 12. The belt conveyor 17 conveys the mixed soil discharged from the gate of the screw conveyor 16 further to the rear shaft side.

[0016] The control device 20 is a computer system for controlling the operations of each part of the shield tunneling machine 10. The control device 20 is, for example, a PLC (Programmable Logic Controller). The control device 20 collects the measurement results of various sensors included in a sensor group 21 (see FIG. 3) provided in the shield tunneling machine 10 described later.

[0017] The abnormality diagnosis system includes a management device 30 and a diagnosis device 40. The management device 30 and the diagnosis device 40 are installed, for example, in a central control room CT1 on the ground. The management device 30 is a computer system for remotely managing the tunneling operation by the shield tunneling machine 10 from the ground. The management device 30 remotely controls the operation of the shield tunneling machine 10 by communicating with the control device 20 of the shield tunneling machine 10. The management device 30 acquires information regarding the operation status of the shield tunneling machine 10 from the control device 20. The information regarding the operation status of the shield tunneling machine 10 includes the measurement results of various sensors provided in the shield tunneling machine 10 and command values included in the control signals from the control device 20 to each part of the equipment provided in the shield tunneling machine 10. The management device 30 displays the information regarding the operation status transmitted from the control device 20 on a display or the like.

[0018] The diagnosis device 40 is a computer system for evaluating the presence or absence of abnormalities in the equipment provided in the shield tunneling machine 10, the presence or absence of a fire in the tunnel, the magnitude of vibrations generated during the tunneling operation, etc., based on the information regarding the operation status of the shield tunneling machine 10. The diagnosis device 40 outputs the evaluation results of each evaluation item to the management device 30 in response to the input of the information regarding the operation status of the shield tunneling machine 10 from the management device 30. The management device 30 issues an alarm according to the evaluation results of the diagnosis device 40.

[0019] In addition, in the present embodiment, when the shield tunneling machine 10 tunnels, the vibration generated is measured by an operator on the ground using the first vibration sensor 50. The ground vibration information regarding the vibration on the ground measured by the operator using the first vibration sensor 50 is input to the management device 30 by wireless communication from the first vibration sensor 50 or by an input operation by the operator. The ground vibration information measured by the first vibration sensor 50 is used for calibration of a vibration prediction model 43 (see FIG. 3) described later provided in the diagnostic device 40, the details of which will be described later.

[0020] (Hardware Configuration Example) FIG. 2 is a hardware configuration example of an information processing device H10 that functions as the management device 30 and the like. The information processing device H10 includes a communication device H11, an input device H12, a display device H13, a storage device H14, and a processor H15. The control device 20 and the diagnostic device 40 may also adopt the following hardware configuration example. Note that this hardware configuration is an example, and other hardware may be included.

[0021] The communication device H11 is an interface for establishing a communication path for transmitting and receiving data to and from other devices. The communication device H11 is, for example, a network interface, a wireless interface, or the like.

[0022] The input device H12 is a device that receives inputs from users and the like. The input device H12 is, for example, a mouse, a keyboard, or the like. The display device H13 is a display or the like that displays various information. The information processing device H10 may have a touch panel that also serves as the input device H12 and the display device H13.

[0023] The storage device H14 is a device that stores data and various programs for executing the various functions of the management device 30. Examples of the storage device H14 include a ROM, a RAM, a hard disk, and the like.

[0024] The processor H15 controls each process in the information processing unit H10 using programs and data stored in the memory device H14. Examples of processor H15 include CPUs and MPUs. This processor H15 executes various processes corresponding to various operations by loading programs stored in ROM, etc., into RAM.

[0025] The processor H15 is not limited to performing all of its operations through software processing. For example, the processor H15 may include dedicated hardware circuits (e.g., application-specific integrated circuits: ASICs) that perform hardware processing for at least some of the operations it performs. In other words, the processor H15 may be configured as follows:

[0026] [1] One or more processors that operate according to a computer program (software) [2] One or more dedicated hardware circuits that perform at least some of the various processes, [3] Circuits that include combinations of these. The H15 processor includes the CPU and memory such as RAM and ROM. Memory stores program code or instructions configured to cause the CPU to perform processing. Memory, or computer-readable media, includes any media accessible and usable by a general-purpose or dedicated computer.

[0027] (Sensor group 21) As shown in Figure 3, the shield tunneling machine 10 is equipped with a sensor group 21 consisting of multiple sensors. The measurement results from the sensor group 21 are aggregated in the control device 20 and then transmitted from the control device 20 to the management device 30.

[0028] The sensor group 21 includes a ground condition sensor 21A, an equipment condition sensor 21B, an odor sensor 21C, a temperature sensor 21D, a second vibration sensor 21E, and the like. The ground condition sensor 21A is a sensor that acquires ground information regarding the properties of the ground to be excavated during the excavation of the shield tunneling machine 10. The ground information is a physical quantity that correlates with the properties of the ground to be excavated.

[0029] The ground condition sensor 21A is a sensor that measures physical quantities correlated with the properties of the soil contained in the ground, such as an earth pressure sensor that measures the pressure of the mixed soil flowing into chamber SP1, a torque sensor for the screw conveyor 16, and a rotation speed meter for the screw conveyor 16. The ground condition sensor 21A is a sensor that measures physical quantities correlated with the compaction by the surrounding ground or the frictional force with the surrounding ground, such as a total thrust meter for the shield tunneling machine 10. The ground condition sensor 21A is a sensor that measures physical quantities correlated with soil conditions such as the appearance rate of gravel and boulders contained in the ground, such as a torque sensor for the cutter 13. The ground condition sensor 21A may also be a sensor that detects the opening degree of the gate of the screw conveyor 16, which is adjusted according to the properties of the soil contained in the ground.

[0030] The equipment status sensor 21B is a sensor that acquires equipment information regarding the status of various equipment installed in the shield tunneling machine 10 during the tunneling operation of the shield tunneling machine 10. The equipment information is a physical quantity that correlates with the operating status of the various equipment installed in the shield tunneling machine 10. For example, the equipment information is measured values ​​such as current value, voltage value, torque value, and rotational speed of the various equipment installed in the shield tunneling machine 10.

[0031] The odor sensor 21C measures the amount of gas produced in the tunnel during the tunneling operation of the shield tunneling machine 10, either when a tunnel fire occurs or as a precursor to a tunnel fire. The amount of gas measured by the odor sensor 21C is an example of odor information.

[0032] The temperature sensor 21D measures the temperature inside the tunnel during the tunneling operation of the shield tunneling machine 10. The temperature sensor 21D is, for example, a thermal camera. The temperature inside the tunnel measured by the temperature sensor 21D is an example of temperature information.

[0033] The second vibration sensor 21E measures vibrations generated by the tunneling operation of the shield tunneling machine 10 on the shield tunneling machine 10. The vibration information measured by the second vibration sensor 21E on the shield tunneling machine 10 is an example of underground vibration information.

[0034] (Diagnostic device 40) As shown in Figure 3, the diagnostic device 40 includes multiple normality determination models 41, a fire detection model 42, and a vibration prediction model 43. All of these models are pre-trained models that have been trained using a dataset. Depending on their application, these models can employ forms such as one-class support vector machines (OCSVM), random forests, and neural networks.

[0035] The normality determination model 41 is a machine learning model that uses excavation information acquired during excavation by the shield tunneling machine 10 as an explanatory variable and the normality of the equipment installed in the shield tunneling machine 10 as an objective variable. For example, the normality determination model 41 outputs the normality of the equipment of the shield tunneling machine 10 from the output layer by inputting the excavation information into the input layer.

[0036] The "normality" of the equipment of the shield tunneling machine 10 represents the degree of similarity to the equipment's normal operating state. In other words, the normality represents the degree of similarity to the data set of tunneling information when the equipment being evaluated is operating normally. For example, the higher the equipment's normality, the closer the equipment's operating state is to a normal state.

[0037] The excavation information input to the normality determination model 41 includes ground information relating to the properties of the ground to be excavated and equipment information relating to the status of the equipment installed in the shield tunneling machine 10. The equipment information is not limited to measured values ​​measured by the equipment status sensor 21B, but may also be command values ​​included in the control signals from the control device 20 to each part of the equipment installed in the shield tunneling machine 10.

[0038] The excavation information input as explanatory variables includes, for example, the measured and controlled values ​​(target values) of the pressure of the excavated mixed soil. The excavation information includes, for example, the total thrust, the speed and pressure of the shield jack 10B, and the stroke amount. The excavation information includes, for example, the rolling and pitching amounts related to the attitude of the shield tunneling machine 10, the bending angles in the horizontal and vertical directions, and the stroke amount of the bending jack. The excavation information includes, for example, the rotational speed, rotational angle, rotational direction, torque, and pressure of the cutter 13. The excavation information includes, for example, the bearing temperature of the cutter 13, the temperature of the cutter motor 14, the additive injection rate, the amount of additive injected, the jack pressure and slewing angle of the erector 10A. The excavation information includes, for example, the rotational instruction value, rotational speed, screw pressure, earth pressure in the screw, and rotational direction of the screw conveyor 16. Excavation information includes, for example, the gate opening of the screw conveyor 16, the pressure and stroke amount of the copy cutter, the grease supply cycle time and lubrication pressure, the amount of lubrication supplied to the tail seal, the bit wear amount, the backfill injection rate and injection amount, various current values, various voltage values, etc. Excavation information may also include, for example, the torque and rotational speed of the rotating part in a detection device equipped with a rotating part that is rotated by the mixed soil in chamber SP1. The torque and rotational speed of the rotating part serve as indicators of the fluidity of the mixed soil in chamber SP1.

[0039] For example, the diagnostic device 40 includes a separate normality determination model 41 for each piece of equipment of the shield tunneling machine 10 whose operating status is to be evaluated. As an example, the diagnostic device 40 includes a separate normality determination model 41 for evaluating the normality of each piece of equipment, such as the erector 10A, shield jack 10B, cutter 13, screw conveyor 16, copy cutter drive unit, and grease supply device. The tunneling information used as explanatory variables in each normality determination model 41 can be selected for each piece of equipment of the shield tunneling machine 10 to be evaluated.

[0040] The normality determination model 41 may, for example, calculate the similarity of each drilling information input as an explanatory variable to a set of drilling information data under normal operating conditions of the equipment being evaluated, and calculate the normality using a function with the similarity of each drilling information as a variable. In this case, the similarity of each drilling information may be weighted in the function.

[0041] The training method for the normality determination model 41 may be supervised learning. In this case, the drilling information is labeled as either normal or abnormal as training data, and the normality determination model 41 is trained on this data. Alternatively, if, for example, there is a small amount of training data in which abnormalities occur in the equipment being evaluated, unsupervised learning may be used. The drilling information used as training data for the normality determination model 41 includes both ground information and equipment information.

[0042] The normality determination model 41 may output a normality threshold for determining whether the target equipment is normal or abnormal. For example, if the normality determination model 41 is trained using supervised learning, the normality threshold can be output using a method such as a Support Vector Machine (SVM). Alternatively, if the normality determination model 41 is trained using unsupervised learning, the normality threshold can be output using a method such as OCSVM.

[0043] The fire detection model 42 is a machine learning model that determines the presence or absence of an underground fire based on fire detection information acquired during the excavation of the shield tunneling machine 10. For example, the fire detection model 42 outputs the presence or absence of an underground fire from the output layer by inputting fire detection information into the input layer. The fire detection information is, for example, odor information measured by the odor sensor 21C or temperature information measured by the temperature sensor 21D, or both. The learning method for the fire detection model 42 may be supervised learning such as SVM or unsupervised learning such as OCSVM. The fire detection information used as training data for the fire detection model 42 includes odor information, temperature information, or both.

[0044] The vibration prediction model 43 is a machine learning model that predicts ground vibration information related to vibrations measured on the surface based on underground vibration information related to vibrations measured by the second vibration sensor 21E installed on the shield tunneling machine 10. For example, the vibration prediction model 43 outputs predicted values ​​of ground vibration information from the output layer by inputting underground vibration information into the input layer.

[0045] One example of a learning method for the vibration prediction model 43 is supervised learning. In this case, the vibration prediction model 43 is trained by associating underground vibration information measured by the second vibration sensor 21E with ground vibration information measured by the first vibration sensor 50 as training data.

[0046] The vibration prediction model 43 may output a threshold value for determining whether or not a large vibration is occurring on the ground, based on the predicted values ​​of ground vibration information. For example, the above threshold value can be output using a method such as OCSVM.

[0047] (Management device 30) As shown in Figure 3, the control device 30 is configured to communicate with the control device 20 and the diagnostic device 40 of the shield tunneling machine 10. The control device 30 may also be configured to communicate with the first vibration sensor 50 via wireless communication or other means.

[0048] The control device 30 comprises a control unit 31, a storage unit 32, an input unit 33, and an output unit 34. The control unit 31 controls the operation of each part of the control device 30. The storage unit 32 stores programs and data necessary for the control unit 31 to perform various processes. The input unit 33 is an input device for operating the shield tunneling machine 10 and a device for the user to input various information to the control device 30. The output unit 34 includes a display that shows information related to the operating status of the shield tunneling machine 10, and a speaker that outputs sound.

[0049] The control unit 31 functions as an abnormality determination unit 31A, a fire determination unit 31B, a vibration prediction unit 31C, etc., by executing an abnormality diagnosis program. The abnormality determination unit 31A inputs the excavation information acquired during the excavation of the shield tunneling machine 10 to the normality determination model 41, thereby obtaining the normality of the equipment to be evaluated from the normality determination model 41. The abnormality determination unit 31A may also obtain a threshold for the normality of the equipment to be evaluated, along with the normality, from the normality determination model 41 to determine whether the equipment to be evaluated is normal or abnormal. Alternatively, the threshold may be stored in the storage unit 32 in advance.

[0050] The abnormality determination unit 31A determines whether the normality level obtained from the normality determination model 41 exceeds a threshold. If the normality level exceeds the threshold, the abnormality determination unit 31A outputs an equipment abnormality alarm to notify that the equipment being evaluated is in an abnormal state. The equipment abnormality alarm may be output, for example, from the output unit 34 as an image, sound, or both.

[0051] The equipment abnormality alarm output by the abnormality detection unit 31A includes, for example, information that can identify the equipment where the abnormality is occurring. In this case, the equipment abnormality alarm outputs information that can identify the equipment where the abnormality is occurring as the inspection target. In addition to the inspection target, the equipment abnormality alarm may also include a list of possible mechanical troubles. The list of mechanical troubles may be output by the normality determination model 41, or data that associates the equipment to be evaluated with the list of possible mechanical troubles may be stored in the storage unit 32 in advance and read from the storage unit 32.

[0052] The fire detection unit 31B inputs fire detection information acquired during the tunneling of the shield tunneling machine 10 to the fire detection model 42, thereby obtaining a determination result from the fire detection model 42 regarding the presence or absence of an underground fire. Based on the determination result of the presence or absence of an underground fire obtained from the fire detection model 42, the fire detection unit 31B outputs a fire alarm to indicate the occurrence of an underground fire. The fire alarm may be output, for example, from the output unit 34 as an image, sound, or both. The fire alarm may include, for example, information about the location of the fire.

[0053] The vibration prediction unit 31C inputs underground vibration information acquired during the excavation of the shield tunneling machine 10 into the vibration prediction model 43, thereby obtaining predicted values ​​for surface vibration information from the vibration prediction model 43. The vibration prediction unit 31C may also acquire a threshold value for the predicted surface vibration information, along with a normality score, from the vibration prediction model 43 to determine whether or not large vibrations are occurring on the surface. Alternatively, this threshold value may be stored in advance in the storage unit 32.

[0054] The vibration prediction unit 31C determines whether the predicted value of the ground vibration information obtained from the vibration prediction model 43 exceeds a threshold. If the predicted value of the ground vibration information exceeds the threshold, the vibration prediction unit 31C outputs a vibration alarm indicating that there is a high probability that large vibrations are occurring on the ground.

[0055] Furthermore, the vibration prediction unit 31C may perform calibration of the vibration prediction model 43 using the measured values ​​of ground vibration information measured on the surface using the first vibration sensor 50 and the measured values ​​of underground vibration information measured on the shield tunneling machine 10 using the second vibration sensor 21E. Calibration here refers to the process of adjusting the vibration prediction model 43 according to the properties of the ground to be excavated. As an example, the calibration of the vibration prediction model 43 is a fine-tuning of the vibration prediction model 43 using the learning data of underground vibration information acquired when the shield tunneling machine 10 performs excavation work on the ground to be excavated, and the measured values ​​of ground vibration information.

[0056] For example, the vibration prediction model 43 may be calibrated at each excavation site, or it may be calibrated at predetermined excavation distances within the same site.

[0057] By calibrating the vibration prediction model 43, the model 43 is adjusted according to the properties of the ground being excavated. Therefore, the accuracy of the ground vibration information predicted by the vibration prediction model 43 can be improved.

[0058] (Methods for diagnosing abnormalities) As shown in Figure 4, the abnormality detection unit 31A performs the abnormality diagnosis process in steps S11 to S16 at predetermined time steps while the shield tunneling machine 10 is excavating.

[0059] First, the abnormality determination unit 31A acquires excavation information from the control device 20, such as the measurement results from the ground condition sensor 21A and the equipment condition sensor 21B, and the command values ​​included in the control signal from the control device 20 to the shield tunneling machine 10, when the shield tunneling machine 10 is excavating (step S11).

[0060] Next, the abnormality determination unit 31A inputs the acquired excavation information into the normality determination model 41 (step S12). At this time, the abnormality determination unit 31A inputs multiple sets of excavation information, each set for the equipment to be evaluated, into each normality determination model 41 provided in the diagnostic device 40.

[0061] Next, the abnormality determination unit 31A obtains the normality of the equipment to be evaluated from the normality determination model 41 (step S13). Furthermore, the abnormality determination unit 31A obtains a normality threshold from the normality determination model 41 (step S14). At this time, the abnormality determination unit 31A may obtain a normality threshold that has been previously stored from the storage unit 32.

[0062] Next, the abnormality determination unit 31A determines whether the normality level obtained from the normality determination model 41 exceeds a threshold (step S15). If the normality level does not exceed the threshold (step S15: YES), the output unit 34 outputs an equipment abnormality alarm (step S16). After that, the abnormality diagnosis process ends. If the normality level does exceed the threshold (step S15: NO), the abnormality diagnosis process ends.

[0063] (Method for detecting underground fires) As shown in Figure 5, the fire detection unit 31B executes the underground fire detection process in steps S21 to S25 at predetermined time steps while the shield tunneling machine 10 is excavating. The underground fire detection process is performed in parallel with the abnormality diagnosis process of the shield tunneling machine 10.

[0064] First, the fire determination unit 31B acquires fire determination information from the control device 20 during the excavation of the shield tunneling machine 10, which includes at least one of the odor information measured by the odor sensor 21C and the temperature information measured by the temperature sensor 21D (step S21).

[0065] Next, the fire determination unit 31B inputs the acquired fire determination information into the fire determination model 42 (step S22). Then, the fire determination unit 31B obtains the determination result of whether or not there is an underground fire from the fire determination model 42 (step S23).

[0066] The fire detection unit 31B then outputs a fire alarm to indicate the occurrence of an underground fire, based on the determination result obtained from the fire detection model 42. Specifically, if the fire detection unit 31B obtains a determination result from the fire detection model 42 indicating that an underground fire has occurred (step S24: YES), it outputs a fire alarm from the output unit 34 (step S25). After that, the underground fire detection process ends. Also, if the fire detection unit 31B obtains a determination result from the fire detection model 42 indicating that no underground fire has occurred (step S24: NO), it ends the underground fire detection process.

[0067] (Method for predicting ground vibrations) Prior to ground vibration prediction processing, the vibration prediction unit 31C performs the following calibration processing at the start of excavation by the shield tunneling machine 10 or at any arbitrary timing during excavation.

[0068] During the calibration process, the vibration prediction unit 31C acquires learning-type underground vibration information measured by the second vibration sensor 21E installed on the shield tunneling machine 10 from the control device 20 when the shield tunneling machine 10 is excavating. It then acquires the measured values ​​of ground vibration information measured by the first vibration sensor 50 from the first vibration sensor 50 or through input from the operator. The vibration prediction unit 31C then inputs the acquired learning-type underground vibration information and the measured values ​​of ground vibration information as learning data into the vibration prediction model 43, thereby performing calibration of the vibration prediction model 43.

[0069] During the calibration process, the measured ground vibration information obtained by the first vibration sensor 50 may include vibrations caused by disturbances other than those originating from the shield tunneling machine 10. In this case, ground vibration information containing vibrations caused by disturbances may be excluded from the training data. Examples of vibrations caused by disturbances include vibrations caused by passing vehicles.

[0070] As shown in Figure 6, the vibration prediction unit 31C executes the ground vibration prediction processing in steps S31 to S36 at predetermined time steps after the calibration of the vibration prediction model 43 is completed and while the shield tunneling machine 10 is excavating. The ground vibration prediction processing is performed in parallel with the abnormality diagnosis processing of the shield tunneling machine 10 equipment and the underground fire detection processing.

[0071] In the ground vibration prediction process, the vibration prediction unit 31C acquires underground vibration information measured by the second vibration sensor 21E installed on the shield tunneling machine 10 from the control device 20 during the tunneling of the shield tunneling machine 10 (step S31). Next, the vibration prediction unit 31C inputs the acquired underground vibration information into the vibration prediction model 43 (step S32).

[0072] Next, the vibration prediction unit 31C obtains predicted values ​​of ground vibration information from the vibration prediction model 43 (step S33). Furthermore, the vibration prediction unit 31C obtains threshold values ​​of predicted ground vibration information from the vibration prediction model 43 (step S34). At this time, the vibration prediction unit 31C may obtain threshold values ​​of predicted ground vibration information that have been previously stored from the storage unit 32.

[0073] Next, the vibration prediction unit 31C determines whether the predicted value of the ground vibration information obtained from the vibration prediction model 43 exceeds a threshold (step S35). If the predicted value of the ground vibration information exceeds the threshold (step S35: YES), the output unit 34 outputs a vibration warning (step S36). After that, the ground vibration prediction process ends. If the predicted value of the ground vibration information does not exceed the threshold (step S35: NO), the ground vibration prediction process ends.

[0074] (Effects of the embodiment) (1) According to this embodiment, the abnormality determination unit 31A inputs excavation information to the normality determination model 41, thereby obtaining a normality score, which is an index value indicating whether or not the equipment to be evaluated is in a normal operating state. In other words, it is possible to determine whether or not the equipment of the shield tunneling machine 10 is in a normal state based on the excavation information. By comparing the normality score with a threshold, it is possible to prevent mechanical troubles that occur in the shield tunneling machine 10, regardless of the skill level of the operator.

[0075] (2) The abnormality determination unit 31A inputs ground information regarding the properties of the ground to be excavated and equipment information regarding the equipment status of the shield tunneling machine 10 as excavation information to the normality determination model 41. As a result, the normality determination model 41 outputs the normality, which is the objective variable, using the ground information and equipment information as explanatory variables. If we were to try to calculate the normality of the shield tunneling machine 10's equipment using only the equipment information, it would be difficult to distinguish whether the fluctuations in the equipment information are due to an abnormality or to fluctuations in the properties of the ground. In this regard, by using both ground information and equipment information as explanatory variables, it is possible to obtain the normality of the shield tunneling machine 10's equipment calculated based on the correlation between the equipment status and the properties of the ground.

[0076] (3) The fire determination unit 31B inputs fire determination information into the fire determination model 42 and obtains a determination result from the fire determination model 42 regarding the presence or absence of an underground fire. This makes it possible to determine whether or not there is an underground fire associated with the tunneling work of the shield tunneling machine 10.

[0077] (4) The vibration prediction unit 31C inputs underground vibration information into the vibration prediction model 43, obtains predicted values ​​of surface vibration information from the vibration prediction model 43, and determines the magnitude of vibrations propagating to the surface based on the predicted values ​​of surface vibration information. In conventional shield tunneling methods, surface workers periodically monitored vibrations propagating to the surface in order to prevent vibrations generated during the excavation work of the shield tunneling machine 10 from adversely affecting the surrounding environment. In contrast, according to this embodiment, vibrations propagating to the surface can be monitored for the entire working period of the shield tunneling machine 10, or for a longer period than with conventional methods.

[0078] Furthermore, vibration measurements by workers on the ground may include vibrations caused by external disturbances such as those from passing vehicles. Conventionally, if ground workers experienced increased ground vibrations due to external disturbances, they had to take measures such as recording the cause in a report. In this regard, by outputting predicted ground vibration information based on the underground vibration information measured by the second vibration sensor 21E installed on the shield tunneling machine 10, it is possible to eliminate external disturbances that may occur during ground vibration measurements.

[0079] (5) Prior to the ground vibration prediction processing, the vibration prediction unit 31C performs calibration on the vibration prediction model 43 using the measured values ​​of the underground vibration information and ground vibration information obtained when the shield tunneling machine 10 performed drilling work in the ground to be excavated as learning data. As a result, the vibration prediction model 43 is adjusted according to the properties of the ground to be excavated, thereby improving the accuracy of the ground vibration information prediction.

[0080] (Example of change) This embodiment can be implemented with the following modifications. This embodiment and the following modifications can be combined with each other to the extent that they do not contradict each other technically.

[0081] If the vibration prediction model 43 can provide predicted values ​​of ground vibration information with sufficient accuracy, the calibration process for the vibration prediction model 43 may be omitted. The ground vibration prediction processing by the vibration prediction unit 31C may be omitted.

[0082] The process of detecting an underground fire by the fire determination unit 31B may be omitted. The excavation information input to the abnormality determination unit 31A may include at least one of the ground information and equipment information. For example, if the normality determination model 41 evaluates equipment of the shield tunneling machine 10 such as the erector 10A and the grease supply device, whose operating state is less affected by the properties of the ground, then only equipment information may be used as an input parameter. Alternatively, if the normality determination model 41 evaluates equipment whose operating state fluctuates depending on the properties of the ground, then multiple pieces of excavation information that correlate with both the properties of the ground and the operating state of the equipment may be used as input parameters. Even in this case, by comparing the normality obtained from the normality determination model 41 with a threshold, mechanical troubles occurring in the shield tunneling machine 10 can be prevented without depending on the skill level of the operator.

[0083] The diagnostic device 40 may, instead of having a configuration that includes a normality determination model 41 for each piece of equipment of the shield tunneling machine 10, include a normality determination model 41 that outputs the normality of multiple pieces of equipment. In other words, multiple normality determination models 41 for outputting the normality of each piece of equipment of the shield tunneling machine 10 may be integrated. Also, the normality determination model 41, the fire detection model 42, and the vibration prediction model 43 may be configured as an integrated machine learning model. Alternatively, the normality determination model 41 and the fire detection model 42 may be integrated, or the normality determination model 41 and the vibration prediction model 43 may be integrated, or the fire detection model 42 and the vibration prediction model 43 may be integrated.

[0084] The normality assessment model 41 may also use fire detection information such as odor information and temperature information, as well as predicted values ​​of underground vibration information or surface vibration information, as tunneling information. In this case, abnormality diagnoses of the shield tunneling machine 10 can be performed by including elements such as odor, temperature, and vibration.

[0085] • In the fire detection model 42, for example, various types of excavation information and normal status may also be used as fire detection information. In this case, the occurrence of an underground fire caused by a malfunction in the shield tunneling machine 10 can be diagnosed more accurately.

[0086] The vibration prediction model 43 may use information other than underground vibration information as input parameters, such as various types of excavation information and normality levels. In this case, it is possible to more accurately predict the increase in vibration caused by abnormalities in the shield tunneling machine 10. For example, parameters corresponding to ground conditions other than underground vibration information may be input as input parameters to the vibration prediction model 43. In this case, the vibration prediction model 43 may be calibrated using the ground condition parameters used as input parameters.

[0087] In the anomaly diagnosis system, the management device 30 and the diagnostic device 40 may be implemented as a single device. The management device 30 may be implemented as a single device or may be distributed across multiple devices or subsystems. Similarly, the diagnostic device 40 may be implemented as a single device or may be distributed across multiple devices or subsystems. [Explanation of symbols]

[0088] CT1...Central control room, H10...Information processing device, H11...Communication device, H12...Input device, H13...Display device, H14...Storage device, H15...Processor, SG1...Segment, SP1...Chamber, T1...Cladding body, 10...Shield tunneling machine, 10A...Erector, 10B...Shield jack, 11...Skin plate, 12...Bulkhead, 13...Cutter, 13A...Bit, 14...Cutter motor, 15...Additive injection pipe, 16...Screw conveyor, 17...Be Belt conveyor, 20...control device, 21...sensor group, 21A...ground condition sensor, 21B...equipment condition sensor, 21C...odor sensor, 21D...temperature sensor, 21E...second vibration sensor, 30...management device, 31...control unit, 31A...abnormality determination unit, 31B...fire determination unit, 31C...vibration prediction unit, 32...storage unit, 33...input unit, 34...output unit, 40...diagnostic device, 41...normality determination model, 42...fire determination model, 43...vibration prediction model, 50...first vibration sensor.

Claims

1. An abnormality diagnosis system comprising a control unit for determining whether or not there is an abnormality in a shield tunneling machine, The control unit, The excavation information acquired during the excavation of the shield tunneling machine is used as an explanatory variable, and the normality of the equipment installed in the shield tunneling machine is used as the objective variable. By inputting the excavation information into this normality determination model, the normality is obtained from the normality determination model. The normality level obtained from the normality determination model is determined to satisfy the threshold, An alarm is output if the aforementioned normality does not meet the threshold. An anomaly diagnosis system.

2. The aforementioned excavation information includes ground information relating to the properties of the ground to be excavated, and equipment information relating to the status of the shield tunneling machine's equipment. The normality determination model uses the ground information and the equipment information as explanatory variables and outputs the normality, which is the objective variable. The control unit inputs the ground information and the equipment information into the normality determination model to obtain the normality from the normality determination model. The abnormality diagnosis system according to claim 1.

3. The control unit, A fire determination model determines the presence or absence of an underground fire based on fire determination information that includes at least one of odor information and temperature information inside the mine. By inputting the fire determination information into the fire determination model, a determination result for the presence or absence of an underground fire is obtained from the fire determination model. Based on the above determination result, an alarm is output to indicate the occurrence of the underground fire. An abnormality diagnosis system according to claim 1 or 2.

4. The control unit, Based on the underground vibration information measured by the vibration sensor installed in the shield tunneling machine, the underground vibration information is input into a vibration prediction model that predicts ground vibration information measured on the surface, thereby obtaining a predicted value of the ground vibration information from the vibration prediction model. Determine whether the predicted value exceeds the threshold, An alarm is output when the predicted value exceeds the threshold. An abnormality diagnosis system according to claim 1 or 2.

5. The control unit, The vibration prediction model is calibrated using the in-tunnel vibration information measured by the vibration sensor when the shield tunneling machine performs excavation work in the ground to be excavated, and the measured values ​​of the above-ground vibration information, as training data. By inputting the underground vibration information into the vibration prediction model after performing the calibration described above, the predicted value of the above-ground vibration information is obtained from the vibration prediction model. The abnormality diagnosis system according to claim 4.

6. An anomaly diagnosis method for determining whether or not there is an anomaly in a shield tunneling machine, using an anomaly diagnosis system equipped with a control unit, The control unit, The excavation information acquired during the excavation of the shield tunneling machine is used as an explanatory variable, and the normality of the equipment installed in the shield tunneling machine is used as the objective variable. By inputting the excavation information into this normality determination model, the normality is obtained from the normality determination model. The normality level obtained from the normality determination model is determined to satisfy the threshold, An alarm is output if the aforementioned normality does not meet the threshold. Methods for diagnosing abnormalities.

7. An anomaly diagnosis program that uses an anomaly diagnosis system equipped with a control unit to determine whether or not there is an anomaly in a shield tunneling machine, The control unit, The excavation information acquired during the excavation of the shield tunneling machine is used as an explanatory variable, and the normality of the equipment installed in the shield tunneling machine is used as the objective variable. By inputting the excavation information into this normality determination model, the normality is obtained from the normality determination model. The normality level obtained from the normality determination model is determined to satisfy the threshold, This system functions as a means to output an alarm when the normality level does not meet the threshold. An anomaly diagnosis program.

Images