While-drilling detection device and detection method for anchoring depth of side slope anchor rod

Abstract

The invention relates to the technical field of slope protection, in particular to a slope anchor rod anchoring depth while-drilling detection device and method.The slope anchor rod anchoring depth while-drilling detection device comprises a rod body, the rod body is arranged between a drill bit and a drill rod, and a mounting cavity is formed in the middle of the rod body; according to the invention, changes and characteristics of various drilling parameters are synchronously analyzed during drilling, and a recognition model between the while-drilling parameters and the stratum information is established by using a feedforward neural network, so that the stratum information of a drilling area is intelligently judged and recognized; a database system is combined to realize computer autonomous decision-making of the anchoring depth of the anchor rod, and the decision-making is uploaded to a feedback APP in real time, so that stratum lithology and geologic structure information of a drilling area is automatically identified in real time, the actual stratum condition of a current drill hole is analyzed according to an intelligent decision-making system, and the anchoring depth of the anchor rod is re-evaluated and designed. And the stratum condition of the anchoring area meets the requirement.

Description

Technical Field

[0001] The invention relates to the technical field of slope protection, and in particular to a device and method for detecting the anchoring depth of a slope anchor rod while drilling. Background Art

[0002] Highway and railway projects involve soil and mountain excavation, forming a large number of slopes. Protecting the slopes and maintaining their long-term stability are one of the key and difficult points of the project. Slope support refers to the support, reinforcement and protection measures taken on the slopes to ensure the safety of the slopes and their environment. Anchor rods are usually used to anchor the slopes. This process has the advantages of high utilization of mechanical equipment, fast construction speed and good protection effect. The anchoring effect of this process depends on the integrity of the soil layer at the designed anchoring depth. The anchor rod has a better anchoring effect in the soil layer with relatively complete rock mass and good strength.

[0003] At present, the anchoring depth is generally determined based on the stratigraphic information of the previous geological survey, and the acquisition of stratigraphic information is usually designed by coring with a borehole. The in-situ coring technology has a relatively high drilling cost and has strict requirements on the number and quality of coring. However, this method connects the positions of the same stratigraphic information at different detection points through a smooth curve to obtain a slope stratigraphic information map, and the judgment of stratigraphic information is still inaccurate. The cost of using geological radar is high, and the reliability of the identification results is affected by the complex geological environment and the detection depth. These factors may cause complex reflection and diffraction of the signal, and reduce the reliability of the measurement results of the stratigraphic information. Inaccurate stratigraphic information may cause the anchor position of the anchor rod to be in a cavity or soft soil layer, which greatly reduces the anchoring effect. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a device and method for detecting the anchoring depth of slope anchor rods while drilling. By synchronously analyzing the changes and characteristics of various drilling parameters during drilling, a recognition model between the drilling parameters and the formation information is established using a feedforward neural network to realize intelligent identification of the formation information in the drilling area. In combination with the database system, the computer makes autonomous decisions on the anchoring depth of the anchor rod and uploads it to the feedback APP in real time, thereby automatically and in real time identifying the formation lithology and geological structure information of the drilling area. According to the intelligent decision-making system, the actual formation conditions of the current drilling are analyzed, the anchoring depth of the anchor rod is re-evaluated and designed, and it is ensured that the formation conditions in the anchoring area meet the requirements.

[0005] The objective of the present invention is achieved through the following technical solutions: A device for detecting the anchoring depth of a slope anchor rod while drilling comprises a rod body, wherein the rod body is arranged between a drill bit and a drill rod, a mounting cavity is arranged in the middle of the rod body, and a sensor group, a data transmission module, a power supply, a data processing module and an intelligent decision-making system are arranged in the mounting cavity. The sensor group is used to collect drilling parameters, which include drilling depth, drilling torque, drilling speed, slurry pressure and slurry concentration; The data transmission module is used to transmit the drilling data collected by the sensor group to the data processing system; The data processing system is used to preprocess and clean the drilling data, and extract the main features of the formation information from the preprocessed and cleaned drilling data; The intelligent decision-making system is used to accurately identify different stratum information according to the correspondence between the extracted main features and the stratum information, and realizes autonomous decision-making on the anchor bolt anchoring depth in combination with a database system for storing a large amount of stratum information and anchor bolt anchoring depth related data; The power supply is used to supply power to the sensor group and the data transmission module.

[0006] Further, the sensor group includes a displacement sensor, a torque sensor, a slurry pressure sensor and a slurry concentration sensor; The displacement sensor is used to collect real-time data of drilling depth and drilling speed during the drilling process; The torque sensor is used to collect torque change data of the drill rod during the drilling process; The slurry pressure sensor is used to collect slurry pressure change data during the drilling process; The slurry concentration sensor is used to collect slurry concentration change data during the drilling process.

[0007] Furthermore, the installation cavity is divided by a partition to form a first cavity, a second cavity and a third cavity that are (evenly distributed along the circumference and) independent of each other, the sensor group is arranged in the first cavity, the power supply is arranged in the second cavity, and the data transmission module is arranged in the third cavity.

[0008] Furthermore, dampers are provided between the sensor group and the first cavity, between the power supply and the second cavity, and between the data transmission module and the third cavity. The dampers protect the sensor group and other electronic components in the measurement while drilling device, are installed around the sensor (the power supply and the data transmission module are installed in the same way), and can effectively reduce vibration transmission under working conditions.

[0009] Furthermore, the sensor group, power supply, and data transmission module are all fixed in the installation cavity by epoxy resin bonding.

[0010] Furthermore, threads are respectively provided at both ends of the rod body, and the rod body is connected to the drill rod and the drill bit through the threads respectively.

[0011] Furthermore, the rod body is made of a rigid material.

[0012] Furthermore, the partition is a steel plate.

[0013] A method for detecting anchoring depth based on a slope anchor bolt anchoring depth while drilling detection device comprises the following steps: S1. Use feedforward neural network technology to establish a recognition model of drilling depth, drilling speed, slurry pressure and slurry concentration and different formation information: establish a training set and a test set based on a large amount of formation information and anchor bolt anchoring depth related data stored in the database system, where the training set is 70%-80% of the initial data, which is used for recognition model training; the test set is the remaining initial data, which is used to test the accuracy of the recognition model; Define the training set as the input layer of the neural network, define the formation information as the output layer of the neural network, and add a hidden layer; Select an activation function for each neuron, input the training set data into the neural network, calculate the weighted input of each hidden layer neuron and generate the output through the activation function; repeat the same steps for each hidden layer until the output of the output layer is calculated; Calculate the weighted input of the output layer and generate the final network output through the activation function; S2. Use the loss function to measure the difference between the network output and the actual label. The formula of the loss function is: In the formula, represents the mean square error, represents the true value, represents the predicted value, Indicates the number of samples; S3. Based on the loss function, the difference between the network output and the actual label is measured, and the recognition model is trained through the training set: the error is back-propagated back to the network to calculate the gradient of each node. Based on the chain rule, the partial derivatives of each node on each training sample are calculated layer by layer, and the calculated gradient is used to update the weight and bias of each node connection, and the performance of the neural network is gradually optimized to obtain the training recognition model: including the following steps: First, calculate the error of the output layer: In the formula, is the loss function, It is Tier The input of each neuron (weighted sum); Indicates activation value A small change in Indicates Tier The activation value of a neuron, that is, the output of the neuron after being processed by the activation function; Secondly, recursively calculate the error of the previous layer: In the formula, It is The number of neurons in the layer, Is the connection Layer neurons and Layer The weight of the neuron, is the activation function, is the derivative of the activation function.

[0014] Then calculate the gradients of weights and biases: In the formula, It is Layer The output of a neuron.

[0015] Finally update the weights and biases: In the formula, is the learning rate; Indicates l Tier j The bias of a neuron A small change in Represents the bias of the jth neuron in the lth layer of the neural network; S4, bringing the test set into the training recognition model and obtaining the prediction accuracy of the training recognition model; When the prediction accuracy is less than 95%, return to step S3; when the prediction accuracy is greater than 95%, a trained recognition model is obtained; S5. Input the drilling parameters into the trained recognition model, and output the formation information corresponding to the drilling data.

[0016] The beneficial effects of the present invention are: The present invention synchronously analyzes the changes and characteristics of various drilling parameters during drilling, uses a feedforward neural network to establish a recognition model between the drilling parameters and the formation information, realizes intelligent identification of the formation information in the drilling area, combines with the database system to realize computer autonomous decision-making on the anchoring depth of the anchor rod, and uploads it to the feedback APP in real time, thereby automatically and in real time identifying the formation lithology and geological structure information of the drilling area, analyzing the actual formation conditions of the current drilling according to the intelligent decision-making system, re-evaluating and designing the anchoring depth of the anchor rod, and ensuring that the formation conditions in the anchoring area meet the requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A three-dimensional diagram of a device for detecting the anchoring depth of a slope anchor rod while drilling in an embodiment of the present invention; Figure 2 It is a bottom view of the device for detecting the anchoring depth of the slope anchor bolt while drilling; Figure 3 It is a rear view of the device for detecting the anchoring depth of the slope anchor bolt while drilling; In the figure, 1. rod body; 2. data transmission module; 3. power supply; 4. displacement sensor; 5. torque sensor; 6. slurry pressure sensor; 7. partition; 8. damper. DETAILED DESCRIPTION

[0018] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0019] See also Figure 1-Figure 3 , the present invention provides a technical solution: Embodiment 1: like Figure 1-Figure 3 As shown, a device for detecting the anchoring depth of a slope anchor rod while drilling comprises a rod body 1, wherein the rod body 1 is arranged between a drill bit and a drill rod, wherein a mounting cavity is arranged in the middle of the rod body 1, wherein a sensor group, a data transmission module 2, a power supply 3, a data processing module and an intelligent decision-making system are arranged in the mounting cavity, The sensor group is used to collect drilling parameters, which include drilling depth, drilling torque, drilling speed, slurry pressure and slurry concentration; The data transmission module 2 is used to transmit the drilling data collected by the sensor group to the data processing system; it adopts efficient and stable communication technology and has remote control function and fault tolerance function.

[0020] The data processing system is used to preprocess and clean the drilling data, and extract the main features of the formation information from the preprocessed and cleaned drilling data; The intelligent decision-making system is used to establish training sets and test sets according to the correspondence between the extracted main features and the formation information, and use the feedforward neural network method to establish a mapping set of while-drilling parameter change characteristics and formation information feature parameters, so as to achieve the effect of accurately identifying different formation information, and combine with a database system for storing a large amount of formation information and anchor bolt anchoring depth related data to realize autonomous decision-making on anchor bolt anchoring depth, and upload it to the feedback APP in real time.

[0021] The power supply 3 is used to supply power to the sensor group and the data transmission module 2. The power supply 3 is integrated with a battery management system (BMS), which is responsible for monitoring parameters such as battery status, temperature, current and voltage, and can automatically cut off the power supply 3 and issue an early warning.

[0022] The sensor group includes a displacement sensor 4, a torque sensor 5, a slurry pressure sensor 6 and a slurry concentration sensor; The displacement sensor 4 is used to collect real-time data of drilling depth and drilling speed during the drilling process; The torque sensor 5 is used to collect torque change data of the drill rod during the drilling process; The slurry pressure sensor 6 is used to collect slurry pressure change data during the drilling process; The slurry concentration sensor is used to collect slurry concentration change data during the drilling process.

[0023] like Figure 1 As shown, the installation cavity is divided by a partition 7 to form a first cavity, a second cavity and a third cavity which are (evenly distributed along the circumference and) independent of each other, the sensor group is arranged in the first cavity, the power supply 3 is arranged in the second cavity, and the data transmission module 2 is arranged in the third cavity.

[0024] A damper 8 is provided between the sensor group and the first cavity, between the power supply 3 and the second cavity, and between the data transmission module 2 and the third cavity. The damper 8 protects the sensor group and other electronic components in the measurement while drilling device, and is installed around the sensor (the power supply 3 and the data transmission module 2 are installed in the same way), and can effectively reduce vibration transmission under working conditions.

[0025] The sensor group, power supply 3 and data transmission module 2 are all fixed in the installation cavity by epoxy resin bonding.

[0026] The two ends of the rod body 1 are respectively provided with threads, and the rod body 1 is connected to the drill rod and the drill bit through the threads.

[0027] The rod body 1 is made of rigid material.

[0028] The partition 7 is a steel plate, wherein the width of the steel plate is smaller than the diameter of the rod body 1, and is installed on the left and right sides of the rod body 1, dividing the inside of the rod body 1 into three spaces, and the sensor circuit can be arranged by punching small holes in the steel plate.

[0029] When the drill rod and drill bit are drilling, the downhole detection device is drilled along with the drilling, and the downhole parameters are collected in real time through the sensor group. The sensor group transmits the collected downhole parameters to the data processing module through the data transmission module 2 for data preprocessing and cleaning, and extracts the main features of the formation information from the downhole data after preprocessing and cleaning.

[0030] Afterwards, the intelligent decision-making system established training sets and test sets based on the correspondence between the extracted main features and the formation information, and used the feedforward neural network method to establish a mapping set of while-drilling parameter change characteristics and formation information feature parameters, so as to achieve the effect of accurately identifying different formation information, and combined with the database system to realize autonomous decision-making on the anchor bolt anchoring depth, and upload it to the feedback APP in real time.

[0031] Embodiment 2: A method for detecting anchoring depth based on a slope anchor bolt anchoring depth while drilling detection device comprises the following steps: S1. Use feedforward neural network technology to establish a recognition model of drilling depth, drilling speed, slurry pressure and slurry concentration and different formation information: establish a training set and a test set based on a large amount of formation information and anchor bolt anchoring depth related data stored in the database system, where the training set is 70%-80% of the initial data, which is used for recognition model training; the test set is the remaining initial data, which is used to test the accuracy of the recognition model; Define the training set as the input layer of the neural network, define the formation information as the output layer of the neural network, and add a hidden layer; Select an activation function for each neuron, input the training set data into the neural network, calculate the weighted input of each hidden layer neuron and generate the output through the activation function; repeat the same steps for each hidden layer until the output of the output layer is calculated; Calculate the weighted input of the output layer and generate the final network output through the activation function; S2. Use the loss function to measure the difference between the network output and the actual label. The formula of the loss function is: In the formula, represents the mean square error, represents the true value, represents the predicted value, Indicates the number of samples; S3. Based on the loss function, the difference between the network output and the actual label is measured, and the recognition model is trained through the training set: the error is back-propagated back to the network to calculate the gradient of each node. Based on the chain rule, the partial derivatives of each node on each training sample are calculated layer by layer, and the calculated gradient is used to update the weight and bias of each node connection, and the performance of the neural network is gradually optimized to obtain the training recognition model: including the following steps: First, calculate the error of the output layer: In the formula, is the loss function, It is Tier The input of each neuron (weighted sum); Indicates activation value A small change in Indicates Tier The activation value of a neuron, that is, the output of the neuron after being processed by the activation function; Secondly, recursively calculate the error of the previous layer: In the formula, It is The number of neurons in the layer, Is the connection Layer neurons and Layer The weight of the neuron, is the activation function, is the derivative of the activation function.

[0032] Then calculate the gradients of weights and biases: In the formula, It is Layer The output of a neuron.

[0033] Finally update the weights and biases: In the formula, is the learning rate; Indicates l Tier j The bias of a neuron A small change in Represents the bias of the jth neuron in the lth layer of the neural network; the bias is updated in combination with the learning rate so that it is continuously optimized to reduce the value of the loss function.

[0034] S4, bringing the test set into the training recognition model and obtaining the prediction accuracy of the training recognition model; When the prediction accuracy is less than 95%, return to step S3; when the prediction accuracy is greater than 95%, a trained recognition model is obtained; S5. Input the drilling parameters into the trained recognition model, and output the formation information corresponding to the drilling data.

[0035] The present invention synchronously analyzes the changes and characteristics of various drilling parameters during drilling, uses a feedforward neural network to establish a recognition model between the drilling parameters and the formation information, realizes intelligent identification of the formation information in the drilling area, combines with the database system to realize computer autonomous decision-making on the anchoring depth of the anchor rod, and uploads it to the feedback APP in real time, thereby automatically and in real time identifying the formation lithology and geological structure information of the drilling area, analyzing the actual formation conditions of the current drilling according to the intelligent decision-making system, re-evaluating and designing the anchoring depth of the anchor rod, and ensuring that the formation conditions in the anchoring area meet the requirements.

[0036] The above is only a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used in various other combinations, modifications and environments, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art shall not deviate from the spirit and scope of the present invention, and shall be within the scope of protection of the claims attached to the present invention.

Claims

1. A device for detecting the anchoring depth of a slope anchor while drilling, characterized in that: It includes a rod body, a data processing module and an intelligent decision-making system. The rod body is arranged between the drill bit and the drill rod. A mounting cavity is arranged in the middle of the rod body. A sensor group, a data transmission module and a power supply are arranged in the mounting cavity. The sensor group is used to collect drilling parameters, which include drilling depth, drilling torque, drilling speed, slurry pressure and slurry concentration; The data transmission module is used to transmit the drilling data collected by the sensor group to the data processing system; The data processing system is used to preprocess and clean the drilling data, and extract the main features of the formation information from the preprocessed and cleaned drilling data; The intelligent decision-making system is used to accurately identify different stratum information according to the correspondence between the extracted main features and the stratum information, and realizes autonomous decision-making on the anchor bolt anchoring depth in combination with a database system for storing a large amount of stratum information and anchor bolt anchoring depth related data; The power supply is used to power the sensor group and the data transmission module; The installation cavity is divided into a first cavity, a second cavity and a third cavity which are independent of each other by a partition, the sensor group is arranged in the first cavity, the power supply is arranged in the second cavity, and the data transmission module is arranged in the third cavity.

2. The device for detecting the anchoring depth of a slope anchor bolt while drilling according to claim 1, characterized in that: The sensor group includes a displacement sensor, a torque sensor, a slurry pressure sensor and a slurry concentration sensor; The displacement sensor is used to collect real-time data of drilling depth and drilling speed during the drilling process; The torque sensor is used to collect torque change data of the drill rod during the drilling process; The slurry pressure sensor is used to collect slurry pressure change data during the drilling process; The slurry concentration sensor is used to collect slurry concentration change data during the drilling process.

3. The device for detecting the anchoring depth of a slope anchor bolt while drilling according to claim 1, characterized in that: Dampers are respectively arranged between the sensor group and the first cavity, between the power supply and the second cavity, and between the data transmission module and the third cavity.

4. The device for detecting the anchoring depth of a slope anchor bolt while drilling according to claim 1, characterized in that: The sensor group, power supply and data transmission module are all fixed in the installation cavity by epoxy resin bonding.

5. The device for detecting the anchoring depth of a slope anchor bolt while drilling according to claim 1, characterized in that: The two ends of the rod body are respectively provided with threads, and the rod body is connected to the drill rod and the drill bit through the threads respectively.

6. The device for detecting the anchoring depth of a slope anchor bolt while drilling according to claim 1, characterized in that: The rod body is made of rigid material.

7. The device for detecting the anchoring depth of a slope anchor bolt while drilling according to claim 3, characterized in that: The partition is a steel plate.

8. A method for detecting anchoring depth based on the device for detecting anchoring depth while drilling according to any one of claims 1 to 7, characterized in that: The steps include: S1. Use feedforward neural network technology to establish a recognition model of drilling depth, drilling speed, slurry pressure and slurry concentration and different formation information: establish a training set and a test set based on a large amount of formation information and anchor bolt anchoring depth related data stored in the database system, where the training set is 70%-80% of the initial data, which is used for recognition model training; the test set is the remaining initial data, which is used to test the accuracy of the recognition model; Define the training set as the input layer of the neural network, define the formation information as the output layer of the neural network, and add a hidden layer; Select an activation function for each neuron, input the training set data into the neural network, calculate the weighted input for each hidden layer neuron and generate output through the activation function; Repeat the same steps for each hidden layer until the output of the output layer is calculated; Calculate the weighted input of the output layer and generate the final network output through the activation function; S2. Use the loss function to measure the difference between the network output and the actual label. The formula of the loss function is: In the formula, represents the mean square error, represents the true value, represents the predicted value, Indicates the number of samples; S3. Based on the loss function, the difference between the network output and the actual label is measured, and the recognition model is trained through the training set: the error is back-propagated back to the network to calculate the gradient of each node. Based on the chain rule, the partial derivatives of each node on each training sample are calculated layer by layer, and the calculated gradient is used to update the weight and bias of each node connection, and the performance of the neural network is gradually optimized to obtain the training recognition model: including the following steps: First, calculate the error of the output layer: In the formula, is the loss function, It is Tier The input of each neuron; Indicates activation value A small change in Indicates Tier The activation value of a neuron, that is, the output of the neuron after being processed by the activation function; Secondly, recursively calculate the error of the previous layer: In the formula, It is The number of neurons in the layer, Is the connection Layer neurons and Layer The weight of the neuron, is the activation function, is the derivative of the activation function; Then calculate the gradients of weights and biases: In the formula, It is Layer The output of a neuron; Finally update the weights and biases: In the formula, is the learning rate; Expressing the l Tier j The bias of a neuron A small change in Represents the bias of the jth neuron in the lth layer of the neural network; S4, bringing the test set into the training recognition model and obtaining the prediction accuracy of the training recognition model; When the prediction accuracy is less than 95%, return to step S3; when the prediction accuracy is greater than 95%, a trained recognition model is obtained; S5. Input the drilling parameters into the trained recognition model, and output the formation information corresponding to the drilling data.

Images