Drilling Trajectory Measurement System and Early Warning Method

Abstract

The present invention discloses a drilling trajectory measurement system and an early warning method. The measuring device includes a water-through cable drill rod, a drill bit, a micro-electromechanical measuring short section, a signal processing unit and an industrial control computer. The micro-electromechanical measuring short section and the drill bit are fixed at the end of the first section of the drill rod. The water-through cable drill rod can simultaneously realize the transmission of high-pressure flushing medium and the real-time transmission of measurement data. The water-through cable drill rod rotates at high speed, driving the drill bit to complete the drilling of the pile body. The micro-electromechanical measuring short section measures the drilling trajectory information in real time, and transmits it to the signal processing unit through a cable inserted in the cable hole of the drill rod. The signal processing unit then transmits the data to the industrial control computer by wireless transmission. The industrial control computer predicts the drilling trajectory information, thereby timely warning the position where the drilling trajectory deviates. The present invention can quickly and accurately capture the drilling trajectory information and three-dimensional position information of the pile body, providing important support for timely discovery and correction of the pile body posture.

Description

Technical Field

[0001] The present invention relates to a measurement system for detecting the borehole trajectory and a method for warning about the borehole trajectory. The present invention belongs to the technical field of soft foundation reinforcement in water conservancy and hydropower projects. Background Art

[0002] Before the construction of water conservancy and hydropower projects, it is necessary to treat the foundation, especially to reinforce some soft foundations. At present, the main method for reinforcing soft foundations is to implant gravel piles with a height of 10 meters to 50 meters into the formation, that is, to form holes by vibrating and punching with a vibroflot. While forming the holes by vibrating and punching, different particle sizes of sand and gravel are filled in. While vibrating and punching to form holes, filling sand and gravel and tamping sand and gravel at the same time to form gravel piles. During the process of vibrating and punching to form holes and filling sand and gravel materials, due to the influence of many factors such as different geological conditions, different hardnesses of the formation at different depths, the deviation of the downward vibrating and punching path of the vibroflot, and unstable filling materials, the phenomenon of the inclination of the vibro-compacted gravel piles often occurs, thus weakening the treatment effect of the vibro-compacted gravel piles on the foundation. Therefore, when treating soft foundations in water conservancy and hydropower projects, it is particularly necessary to detect and correct the pile body posture of the vibro-compacted gravel piles in a timely manner.

[0003] Generally, the method for detecting the pile body posture is to drill holes with a drill rig, and in combination with the borehole trajectory, invert the actual position of the borehole in the formation, judge whether the pile body in the deep formation has a large inclination / offset, and determine the rationality of the pile body posture.

[0004] Currently, for formations at a depth of thousands of meters, the drilling method is that the drill pipe does not rotate, and the motor drives the drill bit to rotate at a high speed. The drill bit drives the drill pipe to drill holes. Its advantages are that the drilling trajectory control is accurate and the obtained borehole trajectory is accurate; its disadvantages are: the drill bit is expensive and the construction cost is high. For formations at a depth of hundreds of meters (the pile height involved in the present invention is 70 - 100 meters), the commonly used drilling method is that the drill bit does not rotate, and the motor drives the drill pipe to rotate at a high speed. The high-speed rotating drill pipe drives the drill bit to drill and obtain the borehole trajectory. Its advantages are: low construction cost and convenient construction; its disadvantages are: 1. Since the drill pipe rotates at a high speed, the cable for transmitting data built in the drill pipe is easy to be wound and the cable is easily torn off; 2. Since during the drilling process, along with the downward drilling of the drill pipe, it is necessary to impact the formation with high-pressure water flow in real time. Therefore, the cable for transmitting data is interfered by high-pressure water flow, mud pulse, electromagnetic wave, etc., and the data transmission is unstable, the data transmission is not smooth, and it is greatly interfered; 3. Since the on-site construction environment is harsh, the data transmission cable is washed by high-pressure water flow for a long time and is extremely easy to be damaged, affecting the normal operation of the measuring device.

[0005] In short, currently in the technical field of soft foundation reinforcement in water conservancy and hydropower projects, due to the limitations of the borehole space, construction cost, and control method, it is impossible to obtain the borehole trajectory of hundreds of meters in real time and accurately, and it is impossible to judge the pile body posture in combination with the borehole trajectory. Summary of the Invention

[0006] With respect to the hundred-meter-level foundation reinforcement pile project, the purpose of the present invention is to provide a measurement system for detecting drilling trajectories which has stable and reliable operation and low cost.

[0007] Another object of the present invention is to provide an early warning method for drilling trajectory.

[0008] To achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a drilling trajectory measurement system, which is composed of several sections of water-through-cable drill rods that can be spliced, a motor, a drill bit, a micro-electromechanical measuring short section, a signal processing unit and an industrial control machine; the drill bit is fixedly connected to the first section of the water-through-cable drill rod, the motor is located on the ground, and the output shaft of the motor is connected to the last water-through-cable drill rod through a coupling to drive the water-through-cable drill rod to rotate, thereby advancing the drill bit to drill a hole; the signal processing unit is located on the ground, and the signal processing unit transmits data to the industrial control machine by wireless transmission; it is characterized in that:

[0009] The micro-electromechanical measuring short section is fixed in the water-through cable drill pipe of the first section, and the micro-electromechanical measuring short section includes a closed cylindrical metal shell and a MEMS sensor fixed in the shell, and the MEMS sensor includes a three-axis acceleration sensor, a three-axis gyroscope and a three-axis rotation speed sensor, so as to measure the position information of the water-through cable drill pipe of the first section, the yaw angle, pitch angle, roll angle of the drill pipe and the rotation speed of the drill pipe;

[0010] The data output end of the MEMS sensor is connected to the signal processing unit through an insulated wire inserted in the water and cable drill rod for data transmission;

[0011] The water and cable drill rod is a solid steel pipe with two longitudinal through holes, one for water and one for cable. A water pipe is passed through the water hole, and an insulated wire for data transmission is passed through the cable hole.

[0012] The diameter of the water hole should meet the following requirements:

[0013] d 通水孔 =1.05d 通水管

[0014]

[0015] d 通水管 ≤3d1

[0016] Where, d 通水孔 is the diameter of the water hole, in m; d 通水管 is the diameter of the water pipe, in m; Q is the high-pressure water flow, in m 3 / h;v maxis the maximum flow velocity of high-pressure water, unit: m / s; v min is the minimum flow velocity of high-pressure water, unit: m / s; d1 is the diameter of the water and cable passing drill pipe, unit: m;

[0017] A waterproof layer is provided between the insulated wire and the cable passing hole, and the diameter of the cable passing hole shall meet the following requirements:

[0018] d 通缆 = 1.05×(d 绝缘导线 + t 防水夹层 )

[0019] In the formula, d 通缆 is the diameter of the cable passing hole, unit: m; d 绝缘导线 is the diameter of the insulated wire, unit: m; t 防水夹层 is the thickness of the waterproof layer, unit: m.

[0020] The diameter of the water and cable passing drill pipe shall be greater than 80 mm, and the minimum distance between the water passing hole and the cable passing hole shall meet the following requirements:

[0021]

[0022] In the formula, d2 is the minimum distance between the water passing hole and the cable passing hole, unit: m; F is the maximum force borne by the drill pipe during rotation, unit: N; σ is the ultimate tensile strength of the steel used for the drill pipe, unit: MPa.

[0023] A drilling trajectory early warning method provided by the present invention includes the following steps:

[0024] S1. Construct a dataset of historical measurement information of the drilling trajectory;

[0025] Obtain the historical information data of the drilling trajectory measured by the three-axis acceleration sensor, three-axis gyroscope and three-axis rotation speed sensor that make up the MEMS sensor during drilling in different strata, and preprocess the historical measurement information dataset to form a dataset for drilling trajectory modeling;

[0026] This dataset contains indexes, dates, times, yaw angles, pitch angles, roll angles, drill pipe depths, and drill pipe rotation speed information under different stratum conditions;

[0027] S2. Construct an optimal drilling trajectory model;

[0028] S2.1: Cluster the drilling trajectory information data of the same stratum in the dataset constructed in step S1 into one category to form a new dataset;

[0029] S2.2: Organize the new dataset formed in step S2.1 into a time series, and use the drilling trajectory information in the first N minutes as the model input data and the drilling trajectory information in the N + 1 minute as the model output data;

[0030] S2.3: Construct a deep learning algorithm pool, traverse each algorithm model in the algorithm pool with the drill hole trajectory information data of each category, and perform deep learning to obtain multiple initial drill hole trajectory models corresponding to each clustering scenario;

[0031] S2.4: Use the AdaBoost data augmentation algorithm to enhance the multiple initial drill hole trajectory models corresponding to each clustering scenario established in step S2.3;

[0032] S2.5: Find the optimal drill hole trajectory model;

[0033] Take the drill hole trajectory information in the first N minutes of the drill hole trajectory information data of each category reorganized in step S2.2 as the model input data, and take the drill hole trajectory information in the (N + 1)-th minute as the model output data, and input them into the multiple initial drill hole trajectory models corresponding to each clustering scenario after the enhancement process in step S2.4. Evaluate the prediction results of different initial drill hole trajectory models through the relative error (R e ), root mean square error (RMSE) and mean absolute error (MAE), perform average weighting on the three indicators, and select the model with the smallest average weighted value as the optimal drill hole trajectory model under this clustering scenario; R e , RMSE and MAE calculation formulas are as follows:

[0034]

[0035]

[0036]

[0037] In the formula, the parameter y i represents the measured value of this feature, represents the predicted value of this feature, and n represents the number of samples of this feature;

[0038] S3. Drill hole trajectory warning;

[0039] S3.1: Obtain the latest drill hole trajectory information data measured by the MEMS sensor in real time;

[0040] S3.2: Preprocess the latest drill hole trajectory information data measured;

[0041] S3.3: Input the preprocessed latest drill hole trajectory information data into the optimal drill hole trajectory model of this category to predict the drill hole trajectory information at the next moment;

[0042] After preprocessing the latest drill hole trajectory information data, match the most similar historical data, and select the optimal drill hole trajectory model corresponding to the similar clustering result;

[0043] S3.4: Compare the prediction result with the trajectory information at the measured real time. When the relative error is greater than the threshold, an alarm is given. The threshold is 5% - 8% of the drill pipe length;

[0044] S3.5: Repeat steps S3.1 - S3.4 until the drilling trajectory early warning ends.

[0045] Among them, in step S2.1, the K - means algorithm is used to cluster and divide the data set.

[0046] The algorithm pool in step S2.3 includes common big data mining models such as recurrent neural network, convolutional neural network, self - attention mechanism, and deep reinforcement learning network. Brief Description of the Drawings

[0047] Figure 1 It is a schematic structural diagram of the drilling trajectory measurement system of the present invention;

[0048] Figure 2 It is a schematic top - view structural diagram of the water - and - cable - passing drill pipe of the present invention;

[0049] Figure 3 It is a side view of the water - and - cable - passing drill pipe of the present invention;

[0050] Figure 4 It is a schematic connection structure diagram of two adjacent water - and - cable - passing drill pipes of the present invention;

[0051] Figure 5 It is a schematic structural diagram of the first - section water - and - cable - passing drill pipe of the present invention;

[0052] Figure 6 It is a flowchart of the drilling trajectory early - warning method of the present invention. Detailed Embodiments

[0053] The structure and features of the present invention will be described in detail below with reference to the drawings and embodiments. It should be noted that various modifications can be made to the embodiments disclosed herein. Therefore, the embodiments disclosed in the specification should not be regarded as a limitation of the present invention, but only as examples of the embodiments, the purpose of which is to make the features of the present invention obvious.

[0054] Such as Figure 1As shown, the drilling trajectory measurement system disclosed in the present invention is composed of several sections of water-through cable drill pipes 1 that can be spliced, a motor 2, a drill bit 3, a micro-electromechanical measuring short section 4, a signal processing unit 5 and an industrial computer 6. The micro-electromechanical measuring short section 4 is installed in the first section of the water-through cable drill pipe 101, the drill bit 3 is fixedly connected to the first section of the water-through cable drill pipe 101, the motor 2 is located on the ground, and the output shaft of the motor 2 is connected to the water-through cable drill pipe 1 through a coupling, driving the water-through cable drill pipe 1 to rotate, thereby advancing the drill bit 3 to drill a hole. The signal processing unit 5 is located on the ground, and the data line of the micro-electromechanical measuring short section 4 is connected to the signal processing unit 5 through the spliced ​​water-through cable drill pipe 1, and the signal processing unit 5 transmits data to the industrial computer 6 by wireless transmission.

[0055] During the detection drilling, the motor 2 drives the water-through cable drill rod 1 to rotate at high speed, and the water-through cable drill rod 1 drives the drill bit 3 to rotate at high speed (the drill bit 3 is stationary relative to the drill rod) to drill the built pile body. While the water-through cable drill rod drives the drill bit to drill a hole in the pile body, the micro-electromechanical measuring short section 4 installed in the first section of the drill rod senses the trajectory information of the water-through cable drill rod, i.e., the drilling hole, such as the yaw angle, pitch angle, roll angle, rotation speed and depth of the drill rod in real time, and transmits the measurement data to the signal processing unit 5 located on the ground through the cable buried in the water-through cable drill rod 1. After the signal processing unit 5 processes the data, it transmits it to the industrial control computer 6 by wireless transmission. The industrial control computer 6 depicts the drilling trajectory based on the data measured in real time by the micro-electromechanical measuring short section 4, and predicts the drilling trajectory through the established data model. When it is found that the drilling trajectory deviates from the design, an alarm is given in time to correct / correct the pile body.

[0056] In order to reduce the resistance of drilling on the pile body and the need for heat dissipation, while the drill rod drives the drill bit to rotate at high speed to drill, it is necessary to introduce high-pressure water into the borehole for continuous high-pressure flushing. The present invention requires that while drilling, the inclination angle, position and other information of the drill rod be measured in real time by a micro-electromechanical measuring short section, and the measurement results are transmitted to the signal processing unit through a cable, and then the data is transmitted to the industrial computer. If the cable used to transmit data is immersed in high-pressure water flow and mud for a long time, it will not only affect the stability and reliability of data transmission, but also accelerate the damage of the cable; in addition, in actual engineering, the environment is harsh, and the cable for transmitting data often gets entangled, broken due to pulling, etc. Therefore, the present invention improves the drill rod structure so that it not only has the function of a drill rod, but also has the dual functions of passing water and passing electricity, and the water and electricity do not interfere with or affect each other. Figure 2 , Figure 3As shown in the figure, the water - and - cable - passing drill pipe 1 of the present invention is a solid steel pipe (such as a solid ZT590 steel pipe). Two longitudinal through - holes are opened on the drill pipe body. One is a water - passing hole 103 and the other is a cable - passing hole 104. A water - passing pipe 105 is inserted into the water - passing hole 103, and the water - passing pipe 105 is connected to an external high - pressure water source. An insulated wire 106 for transmitting data is inserted into the cable - passing hole 104.

[0057] To meet the requirements of high - pressure water flow velocity, the diameter of the water - passing hole 103 is:

[0058] d 通水孔 =1.05d 通水管

[0059]

[0060] d 通水管 ≤3d1

[0061] In the formula, d 通水孔 is the diameter of the water - passing hole, in m; d 通水管 is the diameter of the water - passing pipe, in m; Q is the high - pressure water flow rate, in m 3 / h; v max is the maximum velocity of the high - pressure water, in m / s; v min is the minimum velocity of the high - pressure water, in m / s; d1 is the diameter of the water - and - cable - passing drill pipe, in m.

[0062] An insulated wire 106 is inserted into the cable - passing hole 104, and a waterproof layer is wrapped outside the insulated wire 106 to fully ensure that the signal is not affected by the leakage of high - pressure water. The diameter of the cable - passing hole should meet the following requirements:

[0063] d 通缆 =1.05×(d 绝缘导线 +t 防水夹层 )

[0064] In the formula, d 通缆 is the diameter of the cable - passing hole, in m; d 绝缘导线 is the diameter of the insulated wire, in m; t 防水夹层 is the thickness of the waterproof layer, in m.

[0065] To ensure that the drill pipe has sufficient strength to meet the strength requirements of process technologies such as rotary drilling, sliding directional drilling, and compound drilling of the drill pipe, the diameter of the water - and - cable - passing drill pipe should be greater than 80 mm, and the minimum distance between the water - passing hole and the cable - passing hole of the drill pipe should meet the following requirements:

[0066]

[0067] Wherein, d2 is the minimum spacing between the water hole and the cable hole, in m; F is the maximum force borne by the drill pipe during rotation, in N; σ is the ultimate tensile strength of the steel used for the drill pipe, in MPa.

[0068] The present invention realizes independent transmission of high-pressure medium and signal transmission by arranging water holes and cable holes in the drill rod, and satisfies the real-time and accurate transmission of signals to the signal processing unit without affecting the passage of high-pressure medium. In addition, it is very difficult to arrange water holes and cable holes in a drill rod with limited diameter, which must meet the rigidity requirements of the drill rod itself and the minimum diameter requirements for the arrangement of high-pressure medium and wires. Therefore, it is very important to limit the aperture size of the water holes and cable holes through the above formula.

[0069] In order to realize the detection of the pile trajectory of 100 meters, the present invention comprises a plurality of sections of water- and cable-permeable drill rods 1 with a length of 6-8 meters. Figure 4 It is a connection mode of adjacent water-through cable drill rods, both ends of which are threaded with external threads, and the two adjacent sections of drill rods are quickly connected by locking nuts 107. Of course, the quick connection between the two adjacent sections of drill rods can also be achieved by welding male and female joints at both ends of the drill rod.

[0070] In order to reduce the resonance between drill rods, the present invention provides a shock absorbing material 108 between two adjacent sections of drill rods.

[0071] In a preferred embodiment of the present invention, the water pipe inserted into the water hole of the drill rod is a PE pipe, and the interface of the water pipe is connected with a PET heat shrink tube.

[0072] Figure 5 The schematic diagram of the structure of the first section of the water-through cable drill pipe of the present invention. As shown in the figure, at the front end of the first section of the water-through cable drill pipe 101, the drill bit 3 is fixedly connected to the water-through cable drill pipe by welding. A cavity is formed on the side wall of the first section of the water-through cable drill pipe 101, and the micro-electromechanical measuring short section 4 is fixed in the cavity by bolts 111. The micro-electromechanical measuring short section 4 is a closed cylindrical metal shell, and a mounting plate is fixed in the metal shell, and a MEMS sensor is welded on the mounting plate, and a power supply 110 is fixed. In a preferred embodiment of the present invention, the MEMS sensor includes a three-axis acceleration sensor, a three-axis gyroscope and a three-axis speed sensor, which can measure the position information of the drill pipe, the attitude information of the drill pipe and the rotation speed of the drill pipe, and the attitude information includes the yaw angle, pitch angle and roll angle of the drill pipe. The data output end of the MEMS sensor is connected to the insulated wire 106 that is inserted into the cable hole of the drill pipe, and the measured data is transmitted to the signal processing unit 5 fixed on the ground through the insulated wire 106.

[0073] To prevent the vibration generated by the high-speed rotation of the drill pipe from affecting the measurement accuracy of the MEMS measurement sub-section, a shock-absorbing material (such as high-impact polypropylene) 108 is wrapped outside the sealed cylindrical metal shell. To prevent high-pressure water from seeping in, a layer of waterproof material (such as JS waterproof latex) 109 is included outside the sealed metal shell. The JS waterproof latex selected in the present invention has good temperature adaptability and certain ductility, and can play a role in waterproofing, anti-seepage and buffer protection.

[0074] The signal processing unit 5 installed on the ground includes a shaping and filtering circuit, a microprocessor, and a wireless transmission module. After the data measured by the MEMS sensor is transmitted to the signal processing unit through the cable, it is processed by the shaping and filtering cable and then transmitted to the microprocessor. After the microprocessor classifies and stores the data, it is transmitted to the industrial control computer 6 through the wireless transmission module. After receiving the data sent by the signal processing unit, the industrial control computer processes, analyzes and models the data according to the attitude information and position information of the drill pipe, draws the drilling trajectory, and then draws the trajectory of the next period of drilling through the established drilling trajectory model. If the drilling trajectory of the next period does not meet the design requirements, an alarm will be given immediately and corrected by the construction personnel.

[0075] Figure 6 The flow chart of the method for warning the drilling trajectory by using the above drilling trajectory measuring device in the present invention is as shown in the figure. The specific method is as follows:

[0076] S1. Construct a data set of historical measurement information of the drilling trajectory.

[0077] S1.1: Obtain the historical information data of the drilling trajectory measured by the three-axis acceleration sensor, three-axis gyroscope and three-axis rotation speed sensor that make up the MEMS sensor when drilling in different strata;

[0078] S1.2: Since the data characteristics collected by different sensors are inconsistent, the data obtained by different sensors are merged with reference to the common time index to form a raw CSV file in a unified format;

[0079] S1.3: Delete the repeated features in the original CSV file, delete the abnormal data by the box plot method, and filter and denoise the data by using the first type of Chebyshev filter to remove the noise information generated by the complex environment, and form a data set for drilling trajectory modeling;

[0080] This data set contains 8 key information items: index, date, time, yaw angle, pitch angle, roll angle, drill pipe depth, and drill pipe rotation speed under different stratum conditions.

[0081] S2. Construct an optimal drilling trajectory model.

[0082] S2.1: Cluster the borehole trajectory information data of the same formation in the dataset constructed in step S1 into one category to form a new dataset.

[0083] To reduce the impact of data differences between different formations on modeling, re-cluster and partition the dataset, and cluster the borehole trajectory information data of the same formation into one category. The clustering and partitioning use the K-means algorithm, which has good robustness for large datasets. It is a simple and easy-to-implement clustering algorithm with fast calculation speed, suitable for large-scale datasets, and has good scalability for processing large amounts of data and high-dimensional datasets, and is applicable to various application scenarios.

[0084] S2.2: Organize the new dataset formed in step S2.1 into a time series, and use the borehole trajectory information in the first N minutes as the model input data, and the borehole trajectory information in the (N + 1)-th minute as the model output data.

[0085] A time series (or dynamic sequence) refers to a sequence formed by arranging the values of the same statistical indicator in the order of their occurrence time. For example, first arrange the data of each type of borehole trajectory information (i.e., the borehole trajectory information of each layer) in chronological order, and then use the trajectory information in the first 5 minutes as the model input data, and the trajectory information in the 6th minute as the model output data.

[0086] S2.3: Construct a deep learning algorithm pool, traverse each algorithm model in the algorithm pool with the data of each type of borehole trajectory information, and perform deep learning to obtain multiple initial borehole trajectory models corresponding to each type of clustering scenario.

[0087] The algorithm pool can include common big data mining models such as recurrent neural networks, convolutional neural networks, self-attention mechanisms, and deep reinforcement learning networks.

[0088] A borehole trajectory model is a model used to predict the future trajectory data characteristics of a borehole, which is obtained from the law of the change of the characteristics of the existing historical borehole trajectory data over time.

[0089] S2.4: Use the AdaBoost data augmentation algorithm to enhance the multiple initial borehole trajectory models corresponding to each type of clustering scenario established in step S2.3.

[0090] The AdaBoost algorithm is an ensemble learning method that can construct a strong classifier by combining multiple weak classifiers. It continuously adjusts the weights of data samples in an iterative manner, enabling the model to better learn the misclassified samples, thereby improving the accuracy of the overall result prediction.

[0091] Taking the recurrent neural network as an example, the recurrent neural network is generally composed of multiple stacked fully connected layers. Its advantage is that each update can consider the influence of historical inputs on the output, and the update speed is relatively fast. Its disadvantage is that when facing a complex drill - rock formation interaction relationship, there are problems such as low model - building accuracy and difficult convergence. Therefore, the AdaBoost data augmentation algorithm is used to enhance the recurrent neural network, thereby improving the prediction accuracy of the recurrent neural network.

[0092] S2.5: Search for the optimal borehole trajectory model

[0093] Take the borehole trajectory information in the first N minutes in the data of each type of borehole trajectory information re - organized in step S2.2 as the model input data, and take the borehole trajectory information in the (N + 1)-th minute as the model output data, and input them into the multiple initial borehole trajectory models corresponding to each type of clustering scenario after the enhancement process in step S2.4. Through the relative error (R e ), root - mean - square error (RMSE) and mean absolute error (MAE) to evaluate the prediction results of different initial borehole trajectory models, perform an average weighted calculation on the three indicators, and select the model with the smallest average weighted value as the optimal borehole trajectory model under this type of clustering scenario to predict the trajectory information at the future n - th moment. R e The calculation formulas for RMSE and MAE are as follows:

[0094]

[0095]

[0096]

[0097] In the formula, the parameter y i represents the measured value of this feature, represents the predicted value of this feature, and n represents the number of samples of this feature.

[0098] S3. Borehole trajectory warning.

[0099] S3.1: Real - time obtain the latest borehole trajectory information data measured by the MEMS sensor;

[0100] S3.2: Pre - process the measured latest borehole trajectory information data according to steps S1.2, S1.3, S2.1, and S2.2;

[0101] S3.3: Input the pre - processed measured latest borehole trajectory information data into the optimal borehole trajectory model of this type to predict the trajectory information of the borehole at the next moment.

[0102] After pre - processing the latest borehole trajectory information data, match the most similar historical data, and select the optimal borehole trajectory model corresponding to the similar clustering result.

[0103] S3.4: Compare the prediction result with the measured trajectory information at the real moment. When the relative error is greater than the threshold value, an alarm is given. The threshold value is 5% - 8% of the drill pipe length.

[0104] In a preferred embodiment of the present invention, the threshold value is 5% of the drill pipe length.

[0105] S3.5: Repeat steps S3.1 - S3.4 until the warning of the drilling trajectory ends.

[0106] As the drilling progresses forward, the data sequence continuously scrolls, thereby realizing real - time continuous warning of the new drilling trajectory.

[0107] The present invention uses a micro - electro - mechanical system measurement short joint to obtain the drilling trajectory, realizes the independent real - time transmission of the high - pressure flushing medium and the inclinometer signal through a water - and - cable - passing drill pipe, and realizes the real - time processing, analysis, interpretation and display of the drilling trajectory data through a signal receiving and interpreting terminal, senses the drilling trajectory, and gives a warning to the position where the drilling trajectory deflects. The present invention can quickly and accurately capture the drilling trajectory information and three - dimensional position information, has good prediction accuracy, provides an important technical support for timely discovering and correcting the pile body trajectory, and helps to overcome the problem of detecting the pile body attitude of a pile with a length of up to 100 meters.

Claims

1. A drilling trajectory measurement system, which consists of several sections of water-through and cable-through drill rods that can be spliced, a motor, a drill bit, a micro-electromechanical measuring short section, a signal processing unit and an industrial control computer; the drill bit is fixedly connected to the first section of the water-through and cable-through drill rod, the motor is located on the ground, and the output shaft of the motor is connected to the last water-through and cable-through drill rod through a coupling to drive the water-through and cable-through drill rod to rotate, thereby advancing the drill bit to drill a hole; the signal processing unit is located on the ground, and the signal processing unit transmits data to the industrial control computer by wireless transmission; it is characterized in that: The micro-electromechanical measuring short section is fixed in the water-through cable drill pipe of the first section, and the micro-electromechanical measuring short section includes a closed cylindrical metal shell and a MEMS sensor fixed in the shell, and the MEMS sensor includes a three-axis acceleration sensor, a three-axis gyroscope and a three-axis rotation speed sensor, so as to measure the position information of the water-through cable drill pipe of the first section, the yaw angle, pitch angle, roll angle of the drill pipe and the rotation speed of the drill pipe; The data output end of the MEMS sensor is connected to the signal processing unit via an insulated wire inserted in the water and cable drill rod for data transmission; The water and cable drill rod is a solid steel pipe with two longitudinal through holes, one for water and one for cable. A water pipe is passed through the water hole, and an insulated wire for data transmission is passed through the cable hole. The diameter of the water hole should meet the following requirements: d 通水孔 = 1.05d 通水管 d 通水管 ≤ 3d1 Where d 通水孔 is the diameter of the water passing hole, with the unit of m; d 通水管 is the diameter of the water passing pipe, with the unit of m; Q is the high-pressure water flow rate, with the unit of m 3 / h; v max is the maximum flow velocity of the high-pressure water, with the unit of m / s; v min is the minimum flow velocity of the high-pressure water, with the unit of m / s; d1 is the diameter of the water passing and cable passing drill pipe, with the unit of m; A waterproof interlayer is provided between the insulated wire and the cable hole, and the diameter of the cable hole shall meet the following requirements: d 通缆 = 1.05 × (d 绝缘导线 + t 防水夹层 ) where d 通缆 is the diameter of the through-cable hole, in m; d 绝缘导线 is the diameter of the insulated wire, in m; t 防水夹层 is the thickness of the waterproof layer, in m. The diameter of the water and cable drill rod should be greater than 80 mm, and the minimum spacing between the water hole and the cable hole should meet the following requirements: Where, d2 is the minimum distance between the water hole and the cable hole, unit: m; F is the maximum force that the drill pipe bears during rotation, in N; σ is the ultimate tensile strength of the steel used for the drill pipe, in MPa.

2. The drilling trajectory measurement system according to claim 1, characterized in that: The length of each section of the water and cable drill rod is 6-8m, and two adjacent sections of the drill rod are connected by a quick connector.

3. The drilling trajectory measurement system according to claim 2, wherein: Shock-absorbing materials are arranged between two adjacent sections of the water- and cable-passing drill rods.

4. The drilling trajectory measurement system according to claim 3, characterized in that: A cavity is arranged on the side wall of the first section of the water and cable permeable drill rod, and the micro-electromechanical measuring short section is fixed in the cavity.

5. The drilling trajectory measurement system according to claim 4, characterized in that: A layer of waterproof material is coated on the outside of the micro-electromechanical measuring short joint, and a shock-absorbing material is wrapped on the outside of the waterproof material.

6. A drilling trajectory early warning method using the drilling trajectory measurement system according to any one of claims 1 to 5, comprising the following steps: S1, constructing a data set of historical measurement information of drilling trajectory; Acquire the historical information data of the drilling trajectory measured by the triaxial acceleration sensor, triaxial gyroscope and triaxial rotation speed sensor constituting the MEMS sensor when drilling in different formations, pre-process the historical measurement information data set to form a data set for drilling trajectory modeling; The data set contains the index, date, time, yaw angle, pitch angle, roll angle, drill pipe depth, and drill pipe speed information under different formation conditions; S2, constructing an optimal drilling trajectory model; S2.1: clustering the drilling trajectory information data of the same stratum in the data set constructed in step S1 into one category to form a new data set; S2.2: Organize the new dataset formed in step S2.1 into a time series, use the drilling trajectory information in the first N minutes as the model input data, and use the drilling trajectory information in the (N + 1)-th minute as the model output data; S2.3: Construct a deep learning algorithm pool, traverse each algorithm model in the algorithm pool for each type of drilling trajectory information data, and perform deep learning to obtain multiple initial drilling trajectory models corresponding to each type of clustering scenario; S2.4: Use the AdaBoost data augmentation algorithm to augment the multiple initial drilling trajectory models corresponding to each type of clustering scenario established in step S2.3; S2.5: Find the optimal drilling trajectory model; Take the drilling trajectory information in the first N minutes in each type of drilling trajectory information data reorganized in step S2.2 as the model input data, and take the drilling trajectory information in the (N + 1)-th minute as the model output data, and input them into the multiple initial drilling trajectory models corresponding to each type of clustering scenario after the enhancement processing in step S2.

4. Through the relative error R e , root mean square error RMSE and mean absolute error MAE to evaluate the prediction results of different initial drilling trajectory models, perform average weighting on the three indicators, and select the model with the smallest average weighted value as the optimal drilling trajectory model under this type of clustering scenario; R e , the calculation formulas of RMSE and MAE are as follows: In the formula, parameter y i represents the measured value of this feature, represents the predicted value of this feature, and n represents the number of samples of this feature; S3. Drilling trajectory warning; S3.1: Obtain the latest drilling trajectory information data measured by the MEMS sensor in real time; S3.2: Preprocess the latest drilling trajectory information data measured; S3.3: Input the preprocessed latest drilling trajectory information data into the optimal drilling trajectory model of this type to predict the drilling trajectory information at the next moment; After preprocessing the latest drilling trajectory information data, match the most similar historical data, and select the optimal drilling trajectory model corresponding to the similar clustering result; S3.4: Compare the prediction result with the trajectory information at the measured real time. When the relative error is greater than the threshold, an alarm is issued. The threshold is 5% - 8% of the drill pipe length; S3.5: Repeat steps S3.1 - S3.4 until the drilling trajectory warning ends.

7. The drilling trajectory warning method according to claim 6, characterized in that: In step S2.1, the K-means algorithm is used to perform clustering division on the dataset.

8. The drilling trajectory warning method according to claim 6, characterized in that: The algorithm pool in step S2.3 includes common big data mining models such as recurrent neural network, convolutional neural network, self-attention mechanism, and deep reinforcement learning network.

9. The drilling trajectory warning method according to claim 6, wherein: The method for constructing the drilling trajectory historical measurement information dataset in step S1 is as follows: S1.1: Obtain the historical information data of the drilling trajectory measured by the three-axis acceleration sensor, three-axis gyroscope, and three-axis rotation speed sensor that make up the MEMS sensor when drilling in different strata; S1.2: Since the data characteristics collected by different sensors are inconsistent, refer to the common time index to merge the data obtained by different sensors to form a raw CSV file in a unified format; S1.3: Delete the duplicate features in the raw CSV file, delete the abnormal data by the box plot method, and use the first type of Chebyshev filter to filter and denoise the data to remove the noise information generated due to the complex environment, and form a dataset for drilling trajectory modeling.

Images