JP2023175273A - Support layer reaching determination device and support layer reaching determination method - Google Patents

Abstract

To provide a support layer reaching determination device and a support layer reaching determination method capable of highly accurately determining whether a support layer has been reached.SOLUTION: A support layer reaching determination device 30 comprises: a data acquisition unit 32 that acquires drilling recording data composed of a maximum depth of drilling and a value related to drilling resistance of a drilling machine for each preset measurement depth; a memory 31a that stores design soil quality data indicating a relation between the depth and soil quality based on a result of ground survey; an N value estimation unit 35 that calculates an estimated N value by using an N value estimation model including the drilling recording data and the design soil quality data in input elements; and a soil quality estimation unit 36 that estimates the soil quality by using a soil quality estimation model including the drilling recording data and the estimated N value in input elements.SELECTED DRAWING: Figure 3

Description

The present invention relates to a support layer arrival determination device and a support layer arrival determination method for determining whether excavation by an excavator has reached a support layer.

A construction method that supports the load of a structure is to drill a pile hole to the supporting layer, which is a hard stratum underground, and then use the pile hole to support the load of the structure through multiple piles driven into the ground. be. As such a construction method, an all-casing construction method is known, for example, as disclosed in Patent Document 1. In the all-casing construction method, the pile hole is formed by pressing the casing tube into the ground while swinging or rotating it, and then removing the earth and sand inside the casing tube to the ground using a hammer grab. When the casing tube reaches the support layer, a reinforcing cage or the like is placed inside the casing tube, and then concrete is poured while the casing tube is pulled out.

Japanese Patent Application Publication No. 2015-140548

By the way, before construction of a structure, a ground investigation is conducted to determine the depth of the supporting layer. Then, through the standard penetration test in this ground investigation, the N value, etc., which is an index indicating the hardness of the ground, is obtained. On the other hand, the geological structure is not necessarily the same between the location where the ground investigation was conducted and the location where the pile hole was excavated. Furthermore, it is not practical to conduct a ground investigation at each excavation location, considering the cost and effort involved. Therefore, there is a need for a technology that can determine with high accuracy whether or not the support layer has been reached at each excavation position. These requests are not limited to the all-casing construction method, but are common to construction methods that use an excavator to excavate to the supporting layer.

A support layer arrival determination device that solves the above problem is a support layer arrival determination device that determines whether excavation by an excavator has reached the support layer, and which determines the maximum depth of the excavation and the a data acquisition unit that acquires excavation record data consisting of values related to excavation resistance of an excavator; a design soil storage unit that stores design soil data indicating a relationship between depth and soil quality based on the results of a ground survey; an N value estimator that calculates an estimated N value using an N value estimation model that includes excavation record data and the design soil data as input elements; and a soil estimation unit that includes the excavation record data and the estimated N value as input elements. A soil quality estimation section that estimates soil quality using a model.

According to the present invention, it is possible to determine with high accuracy whether excavation by an excavator has reached the support layer.

1 is a diagram schematically showing a schematic configuration of an excavation system using an embodiment of a support layer arrival determination device. FIG. 2 is a diagram illustrating an example of a hardware configuration of an information processing device that functions as a support layer attainment determination device. It is a functional block diagram showing a support layer arrival determination device. It is a graph for explaining filtering processing for an elapsed time in which the maximum depth is equal to or less than the set value of the condition setting section. It is a graph for explaining the filtering process regarding the elapsed time in which the maximum depth is larger than the setting value of the condition setting section. (a) It is a graph which shows an example of the relationship between the maximum depth and the elapsed time before filtering processing, and (b) It is a graph which shows an example of the relationship between the maximum depth and the elapsed time after the filtering process. It is a flowchart which shows the flow of one embodiment of the support layer attainment determination method. FIG. 3 is a diagram illustrating a part of a display on a display device.

An embodiment of a support layer attainment determining device and a support layer attainment determining method will be described with reference to FIGS. 1 to 8. In addition, in this embodiment, the support layer attainment determination apparatus and the support layer attainment determination method will be described using as an example a case where a pile hole is excavated by the all-casing construction method.

As shown in FIG. 1, in the all-casing construction method, pile holes are formed in the ground 10 by a full-circumference rotating excavator 15. The all-round rotating excavator 15 includes a base device 16 , a holding device 17 , a rotating mechanism 18 , and a pushing mechanism 19 .

The base device 16 is a device for installing the full-circumference rotating excavator 15 on the ground 10. The holding device 17 is a device that supports the casing tube 20 that is press-fitted into the ground 10 from the outside. The holding device 17 supports the casing tube 20 so as to be rotatable about the central axis of the casing tube 20 and to be movable in the vertical direction.

The rotation mechanism 18 rotates the casing tube 20 around the central axis of the casing tube 20 by being supplied with power from the power device 21 . The pushing mechanism 19 pushes the casing tube 20 into the ground by being supplied with power from the power device 21 . The depth at which the lower end of the casing tube 20 pushed into the ground is located is referred to as the maximum depth D.

The power device 21 supplies, for example, hydraulic pressure to the all-round rotating excavator 15 as motive power. The power plant 21 is provided with an operation panel (not shown). The operator operates the all-round rotating excavator 15 by operating the operation panel. The full-circumference rotating excavator 15 pushes the casing tube 20 into the ground 10 with the pushing mechanism 19 while swinging or rotating the casing tube 20 with the rotating mechanism 18 according to the operator's operation. The earth and sand in the casing tube 20 is removed by a hammer grab (not shown) or the like.

In addition, in the all-round rotating excavator 15, when the length of the casing tube 20 is insufficient, a replenishment operation is performed in which a new casing tube is connected to the upper end of the casing tube 20.

The power plant 21 is equipped with a rotational torque measuring device 22 that measures the rotational torque of the rotating mechanism 18 and a pushing force measuring device 23 that measures the pushing force of the pushing mechanism 19. Further, the all-round rotating excavator 15 is equipped with a displacement measuring device 25. The displacement measuring device 25 measures the displacement of the casing tube 20 when the full-circumference rotating excavator 15 is driven. Each measuring device 22 , 23 , 25 performs measurement at a predetermined sampling period (for example, 1 sec) and outputs the measured value to the support layer attainment determination device 30 .

(Support layer attainment determination device)
As shown in FIG. 2, the support layer attainment determination device 30 is configured using an information processing device H10. 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. Note that this hardware configuration is just an example, and other hardware may be included.

The communication device H11 is an interface that establishes a communication path with other devices and executes data transmission and reception. The input device H12 is a device that receives input from a user or the like, and is, for example, a mouse, a keyboard, or the like. The display device H13 is a display, touch panel, or the like that displays various information. The storage device H14 is a storage unit that stores data and various programs for executing various functions of the support layer attainment determination device 30. Examples of the storage device H14 include ROM, RAM, hard disk, and the like.

The processor H15 controls each process in the support layer attainment determination device 30 using programs and data stored in the storage device H14. Examples of the processor H15 include a CPU, an MPU, and the like. This processor H15 expands a program stored in a ROM or the like into a RAM and executes various processes corresponding to various processing. For example, when the application program of the support layer attainment determination device 30 is started, the processor H15 operates a process that executes each process described below.

The processor H15 is not limited to performing software processing for all processes that it executes. For example, the processor H15 may include a dedicated hardware circuit (for example, an application-specific integrated circuit: ASIC) that performs hardware processing for at least part of the processing that it executes. That is, the processor H15 may be configured as follows.

(1) one or more processors that operate according to a computer program (software); (2) one or more dedicated hardware circuits that execute at least some of the various processes; or (3) a combination thereof. circuitry including
A processor includes a CPU and memory, such as RAM and ROM, where the memory stores program codes or instructions configured to cause the CPU to perform processing. Memory or computer-readable media includes any available media that can be accessed by a general purpose or special purpose computer.

As shown in FIG. 3, the support layer attainment determination device 30 includes a management section 31, a data acquisition section 32, an N-value learning section 33, a soil learning section 34, and an N-value estimating section as functional sections that function by executing various programs. 35, and a soil quality estimating section 36. The support layer attainment determination device 30 also includes a data storage section 41, an N-value estimation model storage section 42, and a soil estimation model storage section 43 as storage sections that store various data.

The management unit 31 manages handling of various data input to the support layer attainment determination device 30, execution of processing using the various data, and the like.
The data acquisition unit 32 acquires various data. The data acquisition section 32 includes a measurement value acquisition section 51, an elapsed time acquisition section 52, a filtering section 53, an integrated torque calculation section 54, an average value calculation section 55, and a standard deviation calculation section 56.

The measurement value acquisition unit 51 acquires the maximum depth D based on the displacement measured by the displacement measuring device 25. Furthermore, the measurement value acquisition unit 51 acquires the rotational torque T and the pushing force F at each measurement timing. In this embodiment, the measurement value acquisition unit 51 acquires the rotational torque T and the pushing force F for each predetermined measurement depth based on the maximum depth D. The measurement depth is set to 0.1 m, for example. For example, the measurement value acquisition unit 51 acquires, as the rotation torque T, a value that is the average of the measurement values of the rotation torque measuring device 22 during a period in which the maximum depth D has changed by the measurement depth. Further, the measured value acquisition unit 51 acquires, as the pushing force F, the average value of the measured values of the pushing force measuring device 23 during the period in which the maximum depth D changes by the measured depth.

The elapsed time acquisition unit 52 calculates the elapsed time t (= time required /measured depth).

The filtering unit 53 performs filtering processing on the elapsed time t acquired by the elapsed time acquisition unit 52 to remove abnormal values from the elapsed time t. The filtering unit 53 removes an abnormally large elapsed time t as an abnormal value, such as when the casing tube 20 is added while excavating the measured depth, for example.

The cumulative torque calculation unit 54 calculates the cumulative torque Ta when excavating by the measured depth. For example, the cumulative torque calculation unit 54 calculates the cumulative torque Ta by multiplying the rotational torque T and the elapsed time t.

The average value calculation unit 55 calculates the rotational torque average value T1 and the pushing force average for each of the rotational torque T, the pushing force F, and the cumulative torque Ta, using the values in the calculation target section up to the maximum depth D as a population. The value F1 and the cumulative torque average value Ta1 are calculated. The calculation target section is set to 1.0 m, for example. The average value calculation unit 55 calculates various average values T1, F1, and Ta1 for each measured depth.

The standard deviation calculation unit 56 calculates the rotational torque standard deviation T2, the pushing force standard deviation F2, and the cumulative torque for each of the rotational torque T, the pushing force F, and the cumulative torque Ta, using the values in the calculation target section as a population. Calculate the standard deviation Ta2. The standard deviation calculation unit 56 calculates various standard deviations T2, F2, and Ta2 for each measured depth.

The N value learning unit 33 generates an N value estimation model for calculating the estimated N value at the maximum depth D by performing machine learning using the N value teacher data stored in the data storage unit 41. The N-value learning section 33 stores the generated N-value estimation model in the N-value estimation model storage section 42 .

The soil quality learning unit 34 generates a soil quality estimation model for estimating the soil quality at the maximum depth D by performing machine learning using the soil quality training data stored in the data storage unit 41. The soil estimation section 36 stores the generated soil estimation model in the soil estimation model storage section 43 .

The N value estimation unit 35 calculates the estimated N value at the maximum depth D based on the design soil quality of the construction site, the excavation record at the excavation position, and the N value estimation model stored in the N value estimation model storage unit 42. .

The soil quality estimation unit 36 estimates the soil quality at the maximum depth D based on the excavation record at the excavation position, the estimated N value calculated by the N value estimation unit 35, and the soil quality estimation model stored in the soil quality estimation model storage unit 43. do.

(filtering section)
An example of filtering processing executed by the filtering unit 53 will be described. As described above, the filtering process is a process of removing abnormal values from the elapsed time t.

In the filtering process, the filtering unit 53 sets a removal target time and a condition setting section. The removal target time is the elapsed time t that is the target of removal in the filtering process. The condition setting section is a section for setting a determination value for determining whether or not the removal target time satisfies the abnormality condition based on the elapsed time t in the section. The condition setting section is set to 2.0 m, for example.

For elapsed time t where the corresponding maximum depth D is less than or equal to the set value of the condition setting section, the filtering unit 53 sets all the elapsed time t as the removal target time, and also sets the elapsed time t to the maximum depth D corresponding to the set value. Set the interval as the condition setting interval.

For example, as shown in FIG. 4, when the setting value of the condition setting section is 2.0 m, the filtering unit 53 sets the maximum depth D to 2.0 m for the elapsed time t when the maximum depth D is 2.0 m or less. The corresponding elapsed time t is set as a reference elapsed time ts, and all elapsed times t are set as removal target times, and determination values are set based on these elapsed times t.

For an elapsed time t for which the corresponding maximum depth D is greater than the setting value of the condition setting section, the filtering section 53 sets the removal target time and the condition setting section using the elapsed time t as the reference elapsed time ts.

Specifically, the filtering unit 53 sets the reference elapsed time ts and two elapsed times t before and after the reference elapsed time ts as the removal target times. The filtering unit 53 sets a predetermined depth interval, which is an interval based on a set value, up to the maximum depth D corresponding to the reference elapsed time ts as a condition setting interval.

For example, as shown in FIG. 5, if the setting value of the condition setting section is 2.0 m and the elapsed time t corresponding to the maximum depth of 5.0 m is the reference elapsed time ts, the filtering unit 53 A depth of 3.0 m to 5.0 m is set as a condition setting section, and a determination value is set based on the elapsed time t measured in that section.

As illustrated in FIGS. 4 and 5, the filtering unit 53 that has set the removal target time and the condition setting section calculates the median value tc of the elapsed time t in the condition setting section. Then, the filtering unit 53 sets an abnormal condition based on the median value tc. The abnormal condition is a condition in which the elapsed time t is abnormally long, regardless of changes in soil quality, such as when a casing tube is added. The abnormal condition is set to be equal to or longer than the elapsed time tb, which is three times the median value tc, for example, based on the results of verification conducted in advance. The filtering unit 53 removes the elapsed time t that satisfies the abnormal condition as an abnormal value. In FIGS. 4 and 5, the elapsed time t removed as an abnormal value is indicated by a black circle. Note that the filtering unit 53 may set, as the elapsed time corresponding to the elapsed time t removed as an abnormal value, a corrected elapsed time obtained by performing linear interpolation using the elapsed time t before and after the abnormal value.

The graph in FIG. 6(a) shows the relationship between the maximum depth D and the elapsed time t before the filtering process for examples in which casing tube addition work was performed when the maximum depth D was around 0 m, around 5 m, around 10 m, and around 15 m. It shows a relationship. The graph in FIG. 6(b) shows the relationship between the maximum depth D and the elapsed time t after filtering is performed on the elapsed time t shown in FIG. 6(a). As shown in FIGS. 6(a) and 6(b), it can be seen that the elapsed time t when the casing tube addition work was performed is removed as an abnormal value by the filtering process described above.

(Method for determining support layer reach)
As shown in FIG. 7, the support layer attainment determination method using the support layer attainment determination device 30 includes a pre-construction process, an investigation process, a pre-excavation process, and an excavation process. The pre-construction process is a process in which an N value estimation model and a soil quality estimation model are generated based on actual data. The investigation process is a process in which a ground investigation (for example, a boring investigation) is performed on the investigation position of the construction site where excavation is to be performed. The pre-excavation process is a process performed at the construction site before excavating each excavation position. The excavation process is a process performed during excavation of each excavation position.

(Pre-construction process)
In the pre-construction process, when a teacher data acquisition operation is performed using the input device H12 while an external storage medium or external terminal storing performance data is connected to the communication device H11, the management unit 31 acquires a teacher based on the performance data. A teacher data acquisition process for acquiring data is executed (step S1-1). Furthermore, when a learning operation is performed on the input device H12 after the teacher data acquisition process ends, the management unit 31 executes the learning process (step S1-2).

The performance data consists of various measured values obtained at past construction sites. Specifically, actual data includes design soil quality data, measured N value and actual soil quality for each measurement depth, and identification information that identifies the construction site and the excavation position at the construction site. The rotational torque T, pushing force F, maximum depth D, and elapsed time t before filtering are data associated with each other. Actual soil quality is classified into four types, for example, clay soil, sandy soil, gravel, and bedrock. The design soil data is data based on the results of a ground survey conducted at a predetermined survey position at the construction site. The design soil quality data is data indicating the actual soil quality at each depth at the survey location.

(Teacher data acquisition process)
In the teacher data acquisition process (step S1-1), the data acquisition unit 32 acquires actual excavation data based on actual data for each excavation position.

In acquiring the actual excavation data, the data acquisition unit 32 performs a filtering process on the elapsed time t, and also calculates the cumulative torque Ta, various average values T1, F1, Ta1, and standard deviations T2, F2, Ta2 for each measurement depth. I do. Then, the data acquisition unit 32 acquires, for the identification information, an elapsed time t after the filtering process, a rotational torque T corresponding to the elapsed time t, a pushing force F, a maximum depth D, an integrated torque Ta, and various averages. Actual excavation data is obtained by associating values T1, F1, Ta1 and standard deviations T2, F2, Ta2.

The data acquisition unit 32 creates training data for N values by associating measured N values corresponding to each maximum depth D with the actual excavation data and design soil data as correct data. Furthermore, the data acquisition unit 32 creates teacher data for soil quality by associating the actual soil quality corresponding to each maximum depth D with the actual excavation data and the measured N value as correct data. The data acquisition unit 32 outputs the created N-value training data and soil quality training data to the management unit 31. When the management unit 31 stores the N-value teacher data and the soil quality teacher data output by the data acquisition unit 32 in the data storage unit 41, it ends the teacher data acquisition process.

(Learning process)
The learning process (step S1-2) is a process in which an N-value estimation model and a soil quality estimation model are generated.

In the learning process, the N-value learning unit 33 generates an N-value estimation model by performing machine learning using the N-value teacher data stored in the data storage unit 41. Specifically, the N value learning unit 33 generates an N value estimation model by performing machine learning using actual excavation data and design soil data as an input layer and measured N values corresponding to the actual excavation data as an output layer. do. The N-value learning section 33 stores the generated N-value estimation model in the N-value estimation model storage section 42 .

In addition, in the learning process, the soil quality learning unit 34 generates a soil quality estimation model by performing machine learning using the soil quality teacher data stored in the data storage unit 41. Specifically, the soil quality learning unit 34 generates a soil quality estimation model by performing machine learning using the actual excavation data and the measured N value as an input layer and the actual soil quality corresponding to the actual excavation data as an output layer. The soil quality learning unit 34 stores the generated soil quality estimation model in the soil quality estimation model storage unit 43.

Various methods can be applied to the machine learning method in each learning section. Among them, to generate the N-value estimation model and soil quality estimation model, bagging, which is an ensemble learning method, LSTM (Long Short Term Memory), which is deep learning, and the process leading to prediction results and estimation results can be explained. GAM (Generalized Additive Model) and LIME (Local Interpretable Model Explanations), which are XAI (Explainable AI), are suitable. Further, CNN (Convolutional Neural Network), which is a deep learning method, is suitable for generating the soil quality estimation model. Once the N-value estimation model and the soil estimation model are stored, the management unit 31 ends the learning process. This completes the pre-construction process.

(Investigation process)
In the investigation process, after conducting a ground investigation at the investigation location at the construction site (step S2-1), design data is created based on the investigation results using an external terminal or the like (step S2-2). The design data is composed of design soil quality data that is the soil quality for each depth at the survey position and design N value data that is the N value for each depth. The design data is stored in an external terminal, external storage medium, etc.

(Pre-excavation process)
In the pre-excavation process, installation work is performed in which the all-round rotary excavator 15 and the power unit 21 are installed at the excavation position (step S3-1). Thereafter, wiring work is performed to connect the wiring of various measuring instruments such as the rotational torque measuring instrument 22, the pushing force measuring instrument 23, and the displacement measuring instrument 25 to the support layer arrival determination device 30 (step S3-2).

Further, various information such as design data at the construction site, identification information of the construction site, and identification information of the excavation position are input to the support layer arrival determination device 30 (step S3-3). The design data is input to the support layer attainment determination device 30 by performing a design data input operation on the input device H12 while an external storage medium or external terminal storing the design data is connected to the communication device H11. Ru. The identification information of each of the construction site and the excavation position is input to the support layer arrival determination device 30 by performing an identification information input operation on the input device H12. When the design data and identification information are input, the management unit 31 executes a registration process to register them in the memory 31a (step S3-4). The memory 31a that stores the design data functions as a design soil storage section.

(excavation process)
In the excavation process, the operator operates the operating panel to start excavation by the full-circumference rotary excavator 15. When excavation is started, an excavation start operation is performed on the input device H12. When the excavation opening operation is performed, the management unit 31 executes excavation record acquisition processing (step S4-1), estimation processing (step S4-2), and output processing (step S4-3), and then A determination is made as to whether the destination has been reached (step S4-4).

In the excavation record acquisition process (step S4-1), the data acquisition unit 32 acquires the rotational torque T, pushing force F, maximum depth D, and elapsed time t for each measured depth. In addition to filtering the elapsed time t, the data acquisition unit 32 calculates the integrated torque Ta, various average values T1, F1, Ta1, and standard deviations T2, F2, Ta2 for each measurement depth. The data acquisition unit 32 then obtains the rotational torque T, the pushing force F, the maximum depth D, the elapsed time t after filtering processing, the integrated torque Ta, and various average values T1, F1, Ta1 and standard deviations T2, F2, Excavation record data associated with Ta2 is acquired for each measurement depth. The data acquisition unit 32 outputs the acquired excavation record data to the management unit 31. The management unit 31 stores the excavation record data output by the data acquisition unit 32 in the memory 31a in association with the design data stored in the memory 31a.

In the estimation process (step S4-2), the management unit 31 outputs the latest excavation record data and design soil data to the N value estimation unit 35. The N-value estimation unit 35 inputs the excavation record data and the design soil data as input elements to the N-value estimation model stored in the N-value estimation model storage unit 42. Then, the N value estimation unit 35 calculates the value output to the output layer of the N value estimation model as the estimated N value at the maximum depth D. Furthermore, the management unit 31 outputs the latest excavation record data to the soil quality estimation unit 36. The soil estimation unit 36 inputs the excavation record data and the estimated N value calculated by the N value estimation unit 35 as input elements to the soil estimation model stored in the soil estimation model storage unit 43. Then, the soil quality estimation unit 36 estimates the soil quality corresponding to the value output to the output layer of the soil quality estimation model as the actual soil quality at the maximum depth D.

In the output process (step S4-3), the management unit 31 outputs the information obtained through the excavation record acquisition process and the estimation process to the display device H13 or the like together with the design data.
FIG. 8 illustrates part of the display on the display device H13. The display device H13 displays the design N value, estimated N value, rotational torque T, integrated torque Ta, required time (time required to excavate to each maximum depth), elapsed time t after filtering process, pushing force F, Design soil quality, estimated soil quality, etc. are displayed. Here, it is preferable that the design N value and the estimated N value are displayed in the same graph. This makes it easier to understand the changes in the design N value and the estimated N value, and also makes it easier to compare the design N value and the estimated N value. Further, it is preferable that the designed soil quality and estimated soil quality are displayed side by side. This makes it easier to compare the design soil quality and estimated soil quality.

In the determination of reaching the supporting layer (step S4-4), it is determined whether the maximum depth D has reached the supporting layer based on the estimated N value and the estimated soil quality. If it is determined that the maximum depth D has not reached the support layer, excavation by the all-round rotary excavator 15 is continued, and (step S4-1) to (step S4-4) are repeated. On the other hand, if it is determined that the maximum depth D has reached the support layer, the operation panel is operated to end the excavation by the all-round rotary excavator 15, and an operation to end the excavation is performed using the input device H12. When the excavation end operation is performed, the management unit 31 erases various identification information, design data, and excavation record data stored in the memory 31a.

When the excavation by the all-round rotating excavator 15 is completed, a reinforcing cage is built inside the casing tube 20, and then concrete placement using a tremie pipe is performed while pulling out the casing tube 20. . This will build a pile at the excavation location. When the pile is constructed, the wiring of various measuring instruments is removed from the support layer arrival determination device 30, and then a pre-excavation process and an excavation process are performed for the next excavation position. Note that the support layer arrival determination device 30 may be configured so that registration of design data is omitted in the second and subsequent excavation pre-processes.

The effects of this embodiment will be explained.
(1) In the support layer attainment determination device 30, an estimated N value is calculated using an N value estimation model generated using actual excavation data and design soil data as an input layer and a measured N value as an output layer. In addition, soil quality is estimated using a soil quality estimation model generated using actual excavation data and measured N values as an input layer and actual soil quality as an output layer. Therefore, since the accuracy of the estimated N value and the estimated soil quality is increased, it is possible to determine with high accuracy whether the maximum depth D has reached the supporting layer.

(2) By performing a filtering process on the elapsed time t acquired for each measured depth, abnormal values of the elapsed time t, such as when the casing tube 20 is added, can be removed. As a result, the accuracy of the estimated N value and estimated soil quality can be further improved.

(3) In the filtering process, a determination value is set based on the median value tc of the elapsed time t in the condition setting section. Thereby, it is possible to set a determination value that is robust against an abnormal value of the elapsed time t. As a result, even if the condition setting interval includes an abnormal value of the elapsed time t, the abnormal value can be removed more reliably.

(4) By setting a corrected elapsed time obtained by linearly interpolating the elapsed time t before and after the abnormal value as a substitute for the abnormal value, it is estimated based on the N value and soil quality at the maximum depth D corresponding to the abnormal value with high accuracy. be able to.

(5) The excavation record data includes average values T1, F1, Ta1 and standard deviations T2, F2, Ta2 of the rotational torque T, pushing force F, and integrated torque Ta. According to this configuration, the N value and the soil quality can be estimated by taking into account changes in excavation resistance as excavation progresses.

This embodiment can be modified and implemented as follows. This embodiment and the following modified examples can be implemented in combination with each other within a technically consistent range.
- In the above embodiment, the support layer arrival determination device 30 and the support layer arrival determination method are applied to the all-casing construction method using the all-round rotating excavator 15. Not limited to this, the support layer arrival determination device 30 and the support layer arrival determination method can also be applied to other construction methods that can obtain the excavation resistance of an excavator, such as the earth auger construction method. .

- The filtering unit 53 may filter the elapsed time t based on a preset judgment value (fixed value) instead of the judgment value set based on the condition setting section. The determination value in this case is preferably set for each depth based on design data, for example.

- The support layer attainment determination device 30 includes a management section 31, a data acquisition section 32, an N-value estimation model storage section 42, a soil estimation model storage section 43, an N-value estimation section 35, and a soil estimation section 36. Bye. Therefore, the supporting layer attainment determination device 30 is configured to import learned N-value estimation models and soil estimation models from external storage media or external terminals that store these models and store them in the respective storage units 42 and 43. There may be.

- In the above embodiment, various models were generated by supervised learning using the measured N value and actual soil quality as correct data. The N-value estimation model is not limited to this, and may be generated by unsupervised learning using actual excavation data and design soil data as input layers. Further, the soil quality estimation model may be generated by unsupervised learning using actual excavation data and the output results of the N-value estimation model as input layers. Further, in such a configuration, it is preferable that the management unit 31 stores various identification information, design data, and excavation record data stored in the memory 31a in the data storage unit 41 at the end of the estimation process. Thereby, each of the learning units 33 and 34 can further perform machine learning using the excavation record data, so that the accuracy of various models can be improved.

- When XAI, which can be explained as a machine learning method, is used, input elements with a high degree of contribution may be displayed on the display device H13. According to such a configuration, it is possible to grasp which input element has a high degree of contribution depending on the situation at the time.

- The support layer attainment determination device 30 may have a selection unit that selects a machine learning method. The selection unit selects a machine learning method according to the situation at the time and various settings, for example, according to the design data at the time, the degree of contribution of input elements, etc.

- The support layer reaching determination device 30 may include a determination unit that determines whether the maximum depth has reached the support layer based on the estimated N value and the estimated soil quality. The determination unit can be configured, for example, by a determination model generated by machine learning in which the estimated N value and the estimated soil quality are used as an input layer, and whether or not the support layer has been reached is used as an output layer. Note that the input layer of the determination model may include resistance data and design soil data.

- In the above embodiment, the filtering unit 53 sets the reference elapsed time ts and the two elapsed times t before and after the reference elapsed time ts as the removal target times. The present invention is not limited to this, and the filtering unit 53 may set the latest value of the elapsed time t as the reference elapsed time and the removal target time.

- In the above embodiment, the data acquisition unit 32 acquires excavation record data for each measured depth, that is, using timing based on a change in the maximum depth D as the measurement timing. However, the present invention is not limited to this, and the data acquisition unit 32 may acquire excavation record data every certain period of time, with a certain period of time as the measurement timing. The certain period of time is set to 10 seconds, for example.

When acquiring excavation record data at regular time intervals, the data acquisition unit 32 calculates the average value of the measurement values of the rotational torque measuring device 22 over a certain period of time as the rotational torque T, and the average value of the measurement values of the pushing force measuring device 23. is obtained as the pushing force F. Furthermore, the data acquisition unit 32 acquires a value based on the fixed time and the maximum depth D (=fixed time/amount of change in the maximum depth D in the fixed time) as the elapsed time t.

Further, in the case where excavation record data is acquired at regular intervals, when work such as adding a casing tube is performed, a large amount of excavation record data is acquired near the depth at the time of the work. Therefore, for example, when a plurality of excavation record data are acquired in a predetermined depth section (for example, 0.1 m), the filtering section 53 sets a threshold value for the number of data in the elapsed time t in the depth section and performs filtering processing. It is preferable to carry out.

H10...information processing device, H11...communication device, H12...input device, H13...display device, H14...storage device, H15...processor, 10...ground, 15...all-round rotating excavator, 16...base device, 17...holding Device, 18... Rotating mechanism, 19... Pushing mechanism, 20... Casing tube, 21... Power device, 22... Rotating torque measuring device, 23... Pushing force measuring device, 25... Displacement measuring device, 30... Support layer arrival determination device, 31... Management section, 31a... Memory, 32... Data acquisition section, 33... N value learning section, 34... Soil quality learning section, 35... N value estimation section, 36... Soil quality estimation section, 41... Data storage section, 42... N Value estimation model storage section, 43... Soil estimation model storage section, 51... Measured value acquisition section, 52... Elapsed time acquisition section, 53... Filtering section, 54... Integrated torque calculation section, 55... Average value calculation section, 56... Standard Deviation calculation section.

Claims (6)

A support layer arrival determination device that determines that excavation by an excavator has reached the support layer,
a data acquisition unit that acquires excavation record data consisting of the maximum depth of excavation and a value related to excavation resistance of the excavator at each preset measurement timing;
a design soil storage unit that stores design soil data indicating a relationship between depth and soil quality based on the results of a ground survey;
an N value estimation unit that calculates an estimated N value using an N value estimation model that includes the excavation record data and the design soil data as input elements;
A support layer reaching determination device, comprising: a soil quality estimating unit that estimates soil quality using a soil quality estimation model that includes the excavation record data and the estimated N value as input elements. The data acquisition unit includes:
an elapsed time acquisition unit that acquires the elapsed time required to excavate a unit depth;
a filtering unit that performs filtering processing on the elapsed time,
The support layer arrival determination device according to claim 1, wherein the elapsed time after the filtering process is acquired as a component of the excavation record data. The filtering section includes:
setting a reference elapsed time and setting a predetermined section up to the maximum depth corresponding to the reference elapsed time as a condition setting section;
setting an abnormal condition based on the median elapsed time in the condition setting interval;
The support layer attainment determining device according to claim 2, wherein when the reference elapsed time satisfies the abnormal condition, the reference elapsed time is removed as an abnormal value. The filtering section includes:
setting the reference elapsed time and the elapsed time before and after the reference elapsed time as the removal target time;
The support layer arrival determination device according to claim 3, wherein the removal target time that satisfies the abnormal condition is removed as an abnormal value. The excavator is a full-circumference rotating excavator having a rotation mechanism that rotates a casing tube and a pushing mechanism that pushes the casing tube into the ground,
The data acquisition unit includes:
As values related to the excavation resistance of the excavator, the rotational torque of the rotating mechanism, the pushing force of the pushing mechanism, and the integrated torque based on the rotating torque are obtained at each measurement timing, and the rotating torque and the pushing force are obtained. , and for each of the integrated torques, a predetermined section up to the maximum depth is set as a calculation target section, and an average value and a standard deviation are obtained using the values in the calculation target section as a population. The support layer attainment determination device according to any one of the items. A support layer arrival determination method for determining that excavation by an excavator has reached the support layer using a support layer arrival determination device, the method comprising:
The support layer attainment determination device includes:
a step of acquiring excavation record data consisting of the maximum depth of the excavation and a value related to the excavation resistance of the excavator at each preset measurement timing;
a step of storing design soil data indicating the relationship between depth and soil quality based on the results of a ground survey;
calculating an estimated N value using an N value estimation model that includes the excavation record data and the design soil data as input elements;
a step of estimating soil quality using a soil quality estimation model including the excavation record data and the estimated N value as input elements.

Images