JP7169244B2 - Control information output device and control information output method - Google Patents

Description

The present invention relates to a control information output device and a control information output method for a shield excavator.

2. Description of the Related Art Conventionally, a shield excavator for excavating a natural ground is used for constructing a tunnel or the like. The operator who operates the shield excavator operates the direction of excavation while monitoring the construction environment (soil quality, water pressure, etc.) of the site where the shield excavator is to excavate, according to excavation instructions that indicate the direction of excavation planned in advance. is doing.

By constructing tunnels at various sites, operators have gained experience in controlling shield excavators in response to changes in the construction environment, and have improved their proficiency. A skilled operator with improved skills will be able to apply controls that correspond to the current construction environment at the construction site when controlling the shield excavator at the construction site during excavation, and apply knowledge of control in similar construction environments in the past. are doing. However, in the case of an operator who has worked in a small number of construction sites, it is not possible to properly control the shield excavator in a construction environment that has never been experienced with poor experience and basic operating knowledge.

That is, there is a problem that the accuracy and safety of the design of the tunnel to be excavated varies depending on the operator's skill in controlling the shield excavator. In order to solve this problem, there is a configuration in which the shield excavator is automatically operated based on measurement data indicating the rotation state of the cutter of the shield excavator and the propulsion state of the propulsion jack during excavation (see, for example, Patent Document 1).

In addition, artificial intelligence (AI) is generally used to analyze human emotions. In a method using AI, for example, a learned model is created by executing machine learning using training data that associates human facial expressions (input) with emotions (output) corresponding to the facial expressions. . By inputting human facial expressions into this trained model, it is possible to estimate the emotions that the facial expressions mean, and to perform human emotion analysis.

JP-A-07-71189

In Patent Literature 1 described above, the operation of the shield excavator by a skilled operator cannot be sufficiently reproduced. That is, since the control is performed by comparing the measured data with the set value, unlike the control based on the experience of a skilled operator, the control corresponding to the ever-changing construction environment at the site is appropriately performed. However, it was difficult to say that the accuracy and safety of the design of excavated tunnels would improve.

On the other hand, it is conceivable to apply the AI method described above to the operation of a shield excavator. For example, a trained model can be created by executing machine learning using training data that associates the excavation conditions (input) obtained from a shield excavator with the operations (output) of a skilled operator corresponding to the excavation conditions. create. Desired operation of the shield excavator can be estimated by inputting data indicating the current excavation situation into this learned model. By performing an operation in line with the operation estimated by the learned model, even an inexperienced operator can perform an operation similar to that of a skilled operator.

Incidentally, the excavation instruction document describes numerical targets (hereinafter also referred to as numerical items) for operation items such as azimuth angles and pitching angles that should be maintained or aimed during excavation work. In the AI-learned model, the input is the excavation situation including the numerical items described in the excavation instruction sheet, and learning is performed with the ideal operation as the output in the input. It is possible to infer operations.

However, excavation instructions may include comments in addition to the numerical items described above. In the comments, for example, items that the operator should give priority to in operating the shield excavator, such as "I want you to follow the instructed azimuth angle with certainty" are described. In such a case, a skilled operator may examine the numerical items described in the excavation instruction sheet and the contents of the comments, give priority to the contents of the comments, and allow deviation from the numerical target. For example, in order to precisely match the azimuth angle described in the comment, an operation may be performed that allows the specified pitching angle to deviate more than usual.

When using operation guidance by AI as described above, only the numerical items described in the excavation instructions are taken into account, and the content of comments cannot be taken into account. In other words, there are cases in which the content of desirable operations estimated by AI does not match the intuition of a skilled operator who performs operations in consideration of comments. Alternatively, since the AI cannot consider the contents of the comments, there is a possibility that the contents of the operation that cannot be said to be appropriate in view of the comments are inferred. Such comments, unlike numerical items, are not necessarily written in drilling instructions. Also, comments are not always quantitative like numerical items. For this reason, even if AI learns the contents of comments in the same way as numerical items, there is a high possibility that it will not be able to predict desirable operations well, and it is easy to make AI make predictions that take into account the contents of comments. is not.

SUMMARY OF THE INVENTION The present invention has been made in view of such circumstances, and an object of the present invention is to provide a control information output device capable of performing control according to priority items in operation of a shield excavator, and a control information output device. to provide a method.

In order to solve the above-described problems, a control information output device according to an embodiment of the present invention includes a priority information acquisition unit that acquires priority information indicating priority in excavation work, and a situation in which a shield excavator excavates. Inputting the measured data to a measured data acquisition unit that acquires the measured data, and to a trained model that has learned a corresponding relationship between the excavation situation and the operations of the shield excavator that give priority to the items indicated in the priority information. and an output unit configured to output information indicating the operation target estimated by the estimation unit.

Further, in the control information output device according to one embodiment of the present invention, the priority information acquisition unit acquires the priority information indicating one of the plurality of priority items, and the estimation unit includes an excavation situation, Using the learned model selected according to the priority information from among the learned models that have learned the correspondence relationship between each of the plurality of priorities in the shield excavator and the operation that gives priority to each priority. It is characterized by estimating the operation target.

Further, in the control information output device according to one embodiment of the present invention, each time the priority information acquisition unit acquires the priority information, the estimation unit acquires the priority information acquired from each of the learned models. is characterized by selecting the learned model according to .

Further, in the control information output device of one embodiment of the present invention, the priority is at least one of no priority, azimuth angle, pitching angle, amount of deviation from planned line, and clearance. .

A control information output method according to an embodiment of the present invention includes a priority information acquisition step in which the priority information acquisition unit acquires priority information indicating priority in excavation work; A measurement data acquisition step of acquiring measurement data obtained by measuring an excavation situation, and learning in which the estimation unit learns a correspondence relationship between the excavation situation and an operation of the shield excavator prioritizing the items indicated in the priority information. an estimation step of estimating an operation target, which is an operation target of the shield excavator, by inputting the measurement data into the finished model; and an output unit outputting information indicating the operation target estimated by the estimation unit. and an outputting step.

As described above, according to the present invention, it is possible to perform control according to priority items in the operation of the shield excavator.

BRIEF DESCRIPTION OF THE DRAWINGS It is a schematic block diagram which shows the structural example of the shield excavator 10 to which the control information output device 30 of embodiment is applied. 3 is a block diagram showing a configuration example of a control information output device 30 of the embodiment; FIG. 4 is a flowchart showing an operation example of the control information output device 30 of the embodiment;

A control information output device according to an embodiment will be described below with reference to the drawings.

FIG. 1 is a schematic configuration diagram showing a configuration example of a shield excavator 10 to which a control information output device 30 of this embodiment is applied. FIG. 1(a) is a schematic side view of the shield excavator 10, and FIG. 1(b) is a front view of a shield jack 20 for propelling the shield excavator 10. As shown in FIG.
As shown in FIG. 1(a), the shield excavator 10 constructs a primary lining S by assembling segments at the rear of a cylindrical skin plate 11 using an erector (not shown), while digging the ground. It is a mechanism for excavating. In the shield excavator 10, a chamber 12 is provided behind an annular face plate type cutter 16 having a cutter bit 15. As shown in FIG. A plurality of soil pressure gauges D are installed on the side wall inside the chamber 12 . The earth pressure gauge D measures the pressure of mud in the chamber 12 (controlled earth pressure).
A mud material 14 is injected into the chamber 12 from a mud material injection pipe 13 . The excavated soil accumulated in the chamber 12 is mixed with the mud-making soil material 14 by stirring with a mixing blade (not shown) and converted into mud.
The screw conveyor 17 discharges the mud in the chamber 12 to the conveyor 18 through the discharge gate G. Then, the conveyor 18 carries the mud discharged from the screw conveyor 17 out of the tunnel via the conveyor 19 . The gantry M supports a screw conveyor 17 and conveyors 18 and 19 .

Moreover, as shown in FIG.1(b), multiple shield jacks 20 are provided so that the inner periphery of the skin plate 11 may be enclosed, and are arrange|positioned between the skin plate 11 and a segment. When the shield jack 20 is propelled (extended) by hydraulic operation, the surface of the skin plate 11 is pushed and the shield excavator 10 is propelled.

FIG. 2 is a block diagram showing a configuration example of the control information output device 30 of the embodiment.
The control information output device 30 includes a priority information acquisition section 31 , a measurement data acquisition section 32 , an estimation section 33 , an output section 34 and a learned model storage section 35 .

The priority information acquisition unit 31 acquires priority information. The priority information is information indicating matters that are prioritized when the shield excavator 10 performs excavation work.
Priority information is, for example, normal, that is, no priority. If there is no priority, priority is given to returning the position of the shield excavator 10 to the position according to the tunnel plan alignment without giving priority to specific items among the numerical items indicated in the excavation instruction.
The priority information is, for example, the azimuth angle. The azimuth angle is the horizontal angle of the current attitude of the shield excavator 10 . For example, when there is an existing structure in a position close to it in the horizontal direction, an operation that gives priority to the azimuth angle is required so as not to affect the structure. When giving priority to the azimuth angle, adherence to the azimuth angle indicated in the excavation instruction is given priority over other numerical items.

The priority information is, for example, the pitching angle. The pitching angle is the vertical angle of the attitude of the shield excavator 10 . When the pitching angle is prioritized, adherence to the pitching angle specified in the excavation instructions is prioritized over other numerical items.
The priority information is, for example, recovery of planned linear deviation. The planned linear deviation is the deviation (amount of deviation) between the tunnel planned line indicated in the excavation instruction sheet and the position of the shield excavator 10 . When priority is given to recovering the planned alignment deviation, priority is given to returning the position of the shield excavator 10 to the position in line with the tunnel planned alignment.
Priority information is, for example, tail clearance. Tail clearance is the distance between the inner side of skin plate 11 and the outer surface of the segment. For example, in a curved section where the shield excavator 10 changes the direction of excavation and progresses, if the alignment excavated by the shield excavator 10 deviates from the alignment of the assembled segment, the inner peripheral surface of the skin plate 11 and the outer periphery of the segment In order to avoid chipping or breaking the segment due to collision with the surface, operation that prioritizes tail clearance is required. If tail clearance is prioritized, adherence to the tail clearance specified in the excavation instructions is prioritized over other numerical items.

The measurement data acquisition unit 32 acquires measurement data. The measurement data is data obtained by measuring the excavation conditions of the shield excavator 10 . The measurement data are, for example, the planned linear deviation amount, the azimuth angle, the pitching angle, the tail clearance, etc., which are shown as an example of the priority information described above. Of course, the measurement data may be data obtained by measuring at least the excavation situation of the shield excavator 10, and includes the total propulsion force of the shield jack 20, the earth pressure and water pressure of the ground and excavated soil, the internal pressure of the chamber 12, and the work pressure. Data obtained by measuring the injection amount of the mud material 14 or the like may be used.

The learned model storage unit 35 stores learned models.
The learned model is a model that has learned the corresponding relationship between the excavation situation and the operation of the shield excavator 10 . The excavation status is information about excavation, for example, measurement data acquired by the measurement data acquisition unit 32 during excavation. The learned model is, for example, learning data in which excavation situations (inputs in learning data) and operations performed on the shield excavator 10 (outputs in learning data) are associated with each other, which are collected in various excavation operations. is generated by machine learning. When data indicating unlearned excavation situations is input to the learned model, the learned model extracts similar learned excavation situations from the learned data. Then, the learned model estimates the operation associated with the extracted learned excavation situation as the operation corresponding to the unlearned excavation situation.

The learned model storage unit 35 stores a plurality of learned models (normal time operation model 350, azimuth angle priority operation model 351, pitching angle priority operation model 352, planned linear deviation recovery operation model 353, and A clearance priority operation model 354) is provided.

The normal operation model 350 is a model that has learned the correspondence relationship between the excavation situation and the normal operation, that is, when there is no priority. The operation when there is no priority is an operation performed without giving priority to specific numerical items with respect to the numerical items described in the excavation instruction sheet. The normal operation model 350 is created, for example, by executing machine learning using learning data created based on the results of operations performed without prioritizing any of the numerical items described in the excavation instruction sheet. be done. For example, when the degree of divergence between each numerical item described in the excavation instruction sheet and each item when the shield excavator 10 actually excavates is all less than a predetermined range, it is determined that the operation has no priority. be done. The degree of divergence can be calculated using any index that is generally used as an indicator of the degree of divergence, such as the simple addition average value of the absolute value of the difference between the numerical target and the actual value, the representative value, the variance, and the standard deviation. An index may be used.

The azimuth angle priority operation model 351 is a model that has learned the correspondence relationship between the excavation situation and the operation when the azimuth angle is given priority. The azimuth angle priority operation model 351 performs machine learning using, for example, learning data created based on operations performed with the azimuth angle given priority over other numerical items described in the excavation instruction sheet. Created by For example, when the degree of divergence of the actual azimuth angle from the numerical target of the azimuth angle described in the drilling instruction is smaller than the degree of divergence of other items, the azimuth angle is prioritized over the other numerical items. operation.

The pitching angle priority operation model 352 is a model that has learned the correspondence relationship between the excavation situation and the operation when the pitching angle is prioritized. The pitching angle priority operation model 352 performs machine learning using, for example, learning data created based on operations performed with the pitching angle having priority over other numerical items described in the excavation instruction sheet. Created by For example, when the degree of deviation of the actual pitching angle from the numerical target of the pitching angle described in the drilling instruction is smaller than the degree of deviation of other items, the pitching angle is prioritized over the other numerical items. operation.

The planned linear deviation recovery operation model 353 is a model that has learned the correspondence relationship between the excavation situation and the operation when priority is given to the planned linear deviation recovery. The planned linear deviation recovery operation model 353 performs machine learning using learning data created based on operations in which planned linear deviation recovery is prioritized over other numerical items described in the excavation instruction sheet. Created by running For example, when the degree of divergence of the alignment actually excavated by the shield excavator 10 from the planned alignment described in the excavation instruction sheet is smaller than the degree of divergence of other items, the planned alignment deviation recovery is higher than the other numerical items. It is determined that the operation was performed with priority.

The clearance-prioritized operation model 354 is a model that has learned the correspondence relationship between the excavation situation and the operation when priority is given to clearance. The clearance priority operation model 354 is created by executing machine learning using learning data created based on operations performed in which the clearance is prioritized over other numerical targets described in the excavation instructions. be. For example, if the degree of deviation of the clearance result from the numerical target described in the excavation instruction sheet is smaller than the degree of deviation of other items, it is determined that the clearance is given priority over the other numerical targets. .

The estimation unit 33 estimates the operation target using a learned model as a machine learning technique. A trained model is a model stored in the trained model storage unit 35 . The operation target is information indicating the target of operation in the shield excavator 10 , and is, for example, target values such as the position of the point of force applied to the shield excavator 10 and the rotational speed of the screw conveyor 17 . For example, when the position of the force applied to the shield excavator 10 is estimated by the estimation unit 33 as the operation target, the operator moves the position of the force applied to the shield excavator 10 to the position of the force applied to the shield excavator 10 as the operation target. An operation to select the shield jack 20 is performed so as to do.

The estimation unit 33 selects a learned model according to the priority items corresponding to the priority information.
The estimation unit 33 selects the normal operation model 350 when the priority corresponding to the priority information is no priority. The estimation unit 33 selects the azimuth angle priority operation model 351 when the priority corresponding to the priority information is the azimuth angle. The estimation unit 33 selects the pitching angle priority operation model 352 when the priority corresponding to the priority information is the pitching angle. The estimation unit 33 selects the planned linear deviation recovery operation model 353 when the priority corresponding to the priority information is the planned linear deviation recovery. The estimation unit 33 selects the clearance priority operation model 354 when the priority corresponding to the priority information is clearance.
The estimating unit 33 estimates an output obtained by inputting measurement data to the selected trained model as an operation target in the measurement data.

The estimation unit 33 may select a learned model each time the priority information acquisition unit 31 acquires priority information. For example, while the shield excavator 10 is performing excavation work, if the condition of the ground or the attitude of the shield excavator 10 changes, the operator changes the operation priority. be. In such a case, the priority information corresponding to the changed priority is acquired by the priority information acquisition unit 31 . The estimation unit 33 reselects the learned model based on the priority information newly acquired by the priority information acquisition unit 31 . Thereby, even when the priority is changed, the estimation unit 33 can estimate the operation target according to the changed priority.

The output unit 34 outputs information indicating the operation target estimated by the estimation unit 33 . For example, when the operation target is the rotational speed of the screw conveyor 17 , the output unit 34 outputs information indicating the operation target to a drive control unit (not shown) that controls the rotation of the screw conveyor 17 . The drive control unit controls the rotation of the screw conveyor 17 based on the information indicating the operation target (rotational speed of the screw conveyor 17) acquired from the control information output device 30. FIG. Thereby, automatic control of the shield excavator 10 can be performed.

Alternatively, the output unit 34 may output information indicating the operation target to a terminal device (not shown) provided in the operation room of the shield excavator 10, for example. The terminal device displays the operation target acquired from the control information output device 30 on the display section of the terminal device. By displaying the operation target, it is possible to provide the operator with operation guidance (guidance).

FIG. 3 is a flow chart showing an operation example of the control information output device 30 of the embodiment.
First, the control information output device 30 acquires priority information by the priority information acquisition unit 31 (step S10). The priority information is input to an input unit (not shown) of the control information output device 30 by an operator's input operation, for example. The input unit outputs the acquired priority information to the priority information acquisition unit 31 . The operator determines the priority according to, for example, comments written in the excavation instruction sheet, or according to the conditions of the ground being excavated and the shield excavator 10 being excavated.
Next, the control information output device 30 acquires measurement data by the measurement data acquisition section 32 (step S11). The measurement data is data that is measured regarding excavation and is data that is input to the trained model when estimating the effort target.

Next, the control information output device 30 selects a learned model according to the priority information from the plurality of learned models stored in the learned model storage unit 35 by the estimation unit 33 (step S12). The estimation unit 33 estimates the output obtained by inputting the measured data to the selected trained model as the operation target for the measured data (step S13). Then, the control information output device 30 outputs information indicating the operation target through the output unit 34 (step S14).

As described above, in the control information output device 30 of the embodiment, the estimator 33 selects a learned model according to the priority, and estimates the operation target using the selected learned model. As a result, the control information output device 30 of the embodiment can estimate the operation target according to the priority item, and can perform control according to the priority item using the estimation result.

Further, in the control information output device 30 of the embodiment, the priority information acquisition unit 31 acquires priority information indicating one of the plurality of priority items, and the estimation unit 33 acquires the priority information stored in the learned model storage unit 35. The operation goal is estimated using a learned model selected from among the plurality of learned models obtained according to the priority information. As a result, in the control information output device 30 of the embodiment, even when it is necessary to selectively use a plurality of priorities depending on the situation, it is possible to estimate the operation target according to each priority. It has the same effect as the effect.

Further, in the control information output device 30 of the embodiment, the estimation unit 33 selects a learned model corresponding to the priority information from each of the learned models each time the priority information is acquired by the priority information acquisition unit 31. . As a result, in the control information output device 30 of the embodiment, even if it becomes necessary to change the priority items due to changes in the ground conditions during excavation work, The operation target can be estimated, and the same effects as those described above can be obtained.

Also, the priority is at least one of no priority, azimuth angle, pitching angle, planned line deviation recovery, and clearance. As a result, in the control information output device 30 of the embodiment, each of no priority item, azimuth angle, pitching angle, planned line deviation recovery, and clearance can be set as priority items, and an operation target corresponding to each priority item can be set. By estimating, the same effects as those described above can be obtained.

All or part of the processing performed by the control information output device 30 in the above-described embodiment may be realized by a computer. In that case, a program for realizing this function may be recorded in a computer-readable recording medium, and the program recorded in this recording medium may be read into a computer system and executed. It should be noted that the "computer system" referred to here includes hardware such as an OS and peripheral devices. The term "computer-readable recording medium" refers to portable media such as flexible discs, magneto-optical discs, ROMs and CD-ROMs, and storage devices such as hard discs incorporated in computer systems. Furthermore, "computer-readable recording medium" means a medium that dynamically retains a program for a short period of time, like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. It may also include something that holds the program for a certain period of time, such as a volatile memory inside a computer system that serves as a server or client in that case. Further, the program may be for realizing a part of the functions described above, or may be capable of realizing the functions described above in combination with a program already recorded in the computer system. It may be implemented using a programmable logic device such as FPGA.

Although the embodiment of the present invention has been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and design and the like are included within the scope of the gist of the present invention.

DESCRIPTION OF SYMBOLS 10... Shield excavator, 20... Shield jack, 30... Control information output device, 31... Priority information acquisition part, 32... Measurement data acquisition part, 33... Estimation part, 34... Output part, 35... Learned model storage part, 350 Normal time operation model 351 Azimuth angle priority operation model 352 Pitching angle priority operation model 353 Planned linear deviation recovery operation model 354 Clearance priority operation model

Claims (5)

a priority information acquisition unit that acquires priority information indicating priority in excavation work;
a measurement data acquisition unit that acquires measurement data obtained by measuring an excavation situation of the shield excavator;
By inputting the measurement data into a learned model that has learned the correspondence relationship between the excavation situation and the operation of the shield excavator that gives priority to the items indicated in the priority information, the target of the operation of the shield excavator is obtained. an estimating unit for estimating a certain operational target;
an output unit that outputs information indicating the operation target estimated by the estimation unit;
A control information output device comprising: The priority information acquisition unit acquires the priority information indicating one of the plurality of priorities,
The estimating unit selects, according to the priority information, from among learned models that have learned correspondence relationships between excavation situations and operations that give priority to each priority for each of the plurality of priorities in the shield excavator. estimating the operational goal using the learned model that has been trained;
2. The control information output device according to claim 1, wherein: The estimation unit selects the learned model according to the acquired priority information from each of the learned models each time the priority information acquisition unit acquires the priority information.
3. The control information output device according to claim 2, wherein: The priority is at least one of no priority, azimuth angle, pitching angle, amount of deviation from the planned line, and clearance.
4. The control information output device according to any one of claims 1 to 3, characterized in that: a priority information acquisition step in which the priority information acquisition unit acquires priority information indicating priority in excavation work;
a measurement data acquisition step in which the measurement data acquisition unit acquires measurement data obtained by measuring the excavation situation of the shield excavator;
The estimating unit inputs the measurement data to a learned model that has learned the correspondence relationship between the excavation situation and the operation of the shield excavator that gives priority to the items indicated by the priority information, so that the shield excavator an estimating step of estimating an operational target, which is a target of the operation;
an output step in which an output unit outputs information indicating the operation target estimated by the estimation unit;
A control information output method characterized by comprising:

Images