Shale gas drill rod pose less-sensing monitoring method based on metamorphic theory

Abstract

The invention discloses a shale gas drill rod pose less-sensing monitoring method based on a metamorphic theory, and the method comprises the steps: constructing a multi-section linear body model, simplifying a shale gas exploitation drill rod into a multi-section continuum, and simulating the interaction force between the drill rod and the geology; the pose information of the multi-section linear body model is described through three-dimensional metamorphic constraint, the pose information of all turning points is subjected to correlation expression through an adjacent matrix, one-by-one fitting is conducted through the multi-section equal-curvature model, and numerical solution of the overall form theoretical value of the directional drill rod is achieved; constructing a point location detection system to obtain a spatial position of a drill rod connection point, judging whether the spatial position is a true value of a turning point or not, and constructing an isotropic graph mechanical network model by taking a theoretical value obtained by the multi-section equal curvature model as input and a true value obtained by the point location detection system as output; and the calculated sensor data is substituted into the isotropic graph mechanical network model, so that form data of the directional mining drill rod can be realized by using extremely little incomplete sensing data.

Description

Technical Field

[0001] The present invention relates to a perception and monitoring method for continuum robots, and particularly to a less-sensor monitoring method for the posture of shale gas drill pipes based on metamorphic theory. Background Art

[0002] With the development of science and technology, robot technology has become a part of human life and work, and plays an increasingly important role in various aspects. As an important unconventional natural gas resource, shale gas occupies an increasingly important position in the global energy structure. Shale gas resources are rich, widely distributed, and have huge reserves, which can effectively increase natural gas supply; compared with coal and oil, shale gas is a relatively clean energy source, emitting less carbon dioxide and other pollutants after combustion. And with the progress of extraction technologies such as hydraulic fracturing, by developing its own shale gas resources, a country can reduce its dependence on external energy sources and enhance the security of energy supply. During the current shale gas extraction process, in order to improve efficiency, ensure safety, and optimize drilling operations, various sensor technologies are used to monitor and control the system. Commonly used sensor technologies include: sensors such as pressure, temperature, flow rate, and acoustic waves. These sensors are usually integrated into the data acquisition system of the drilling platform and transmit data to the ground control room in real time so that operators can monitor the drilling situation in real time and make necessary adjustments.

[0003] Although these sensor technologies play a crucial role in shale gas extraction, they also have some disadvantages and limitations, mainly including the following aspects:

[0004] 1. Environmental adaptability, the downhole environment is usually under high temperature and high pressure, and is accompanied by strong corrosion of minerals and wear, vibration, and impact during the extraction process.

[0005] 2. Measurement accuracy and reliability, the shale gas reservoir requires sensors to have high-precision resolution and accuracy, and long-term stable operation will cause the performance of the sensors to gradually drift.

[0006] 3. Data transmission and processing, electromagnetic interference and environmental noise downhole may affect the transmission of sensor signals, resulting in data loss or errors, and long-distance transmission will cause the weakening of the electrical signal intensity.

[0007] 4. Size and weight issues, due to space and design limitations, sensors need to be installed inside the drill pipe or bit, which limits the size and weight of the sensors. Summary of the Invention

[0008] To solve the limitation problems of sensors in shale gas extraction, the present invention provides a less-sensor monitoring method for the posture of shale gas drill pipes based on metamorphic theory. This method can realize the real-time dynamic perception of the posture of the directional drill pipe under the condition of few sensors (under-sensing).

[0009] The present invention provides the following technical solutions:

[0010] A method for less-sensor monitoring of the pose of a shale gas drill pipe based on the metamorphic cell theory, the method comprising the following steps:

[0011] Step S1, first simplify the shale gas extraction drill pipe into a multi-segment continuum, and adopt a multi-segment linear body model under the stretching action of a flexible rope-like structure. This model can describe the drill pipes with different stiffness characteristics, and can simulate the interaction force between the drill pipe and the geology by applying radial forces with different magnitudes and directions at the end of the drill pipe.

[0012] Step S2, based on the above multi-segment linear body model, use three-dimensional metamorphic cell constraints to describe the pose information of the turning points of the drill pipe under different stiffness characteristics and different geological conditions. Express the pose information of each turning point based on the base coordinates through the adjacency matrix, and fit each segment with a multi-segment equal-curvature model to realize the numerical solution of the multi-segment equal-curvature model under the action of three-dimensional metamorphic cell constraints.

[0013] Step S3, construct a real-time point position detection system for the drill pipe connection points, obtain the spatial position of the connection points and determine whether they are correct turning points, and calibrate the spatial position of the correct connection points as the true position of the turning points.

[0014] The point position detection system is a system for monitoring the state and parameters of specific points. It conducts real-time monitoring and data collection on the target points through sensors and other devices to achieve precise control and management of the target points. It is mainly used in step S4 to output the true value data of the collected sample points.

[0015] Step S4, use the theoretical value solved by the multi-segment equal-curvature model as the input, and the true value obtained by the real-time point position detection system as the output to construct a less-sensor prediction model for the directional pose of the shale gas extraction drill pipe. And use the Kolmogorov-Arnold network (KAN) to replace the multi-layer perceptron (MLP) in the equivariant graph mechanics network model to strengthen the learning and training of the equivariant graph mechanics network model, so as to realize the less-sensor monitoring of the directional pose of the shale gas extraction drill pipe under three-dimensional metamorphic cell constraints.

[0016] Compared with the prior art, the beneficial effects of the present invention are:

[0017] The present invention simplifies the drill pipe for shale gas extraction by using a multi-segment linear body model with the stretching effect of a flexible rope-like structure. The relative pose of the simplified model is described by three-dimensional variable cell constraints. The theoretical values of the connection points are solved through a multi-segment equal-curvature model and regarded as reference data. The pose information collected at the turning points is used as the training set. The data set is used as the input and the training set is used as the output and substituted into the constructed equivariant graph mechanics network model. And KAN is used instead of MLP in the equivariant graph mechanics network model to strengthen the learning and training of the equivariant graph mechanics network model, so as to realize the few-sensor monitoring of the directional pose of the drill pipe for shale gas extraction with variable cell constraints. Few-sensor means a state of under-sensing. In the case of partial failure or damage of the sensor, real-time dynamic monitoring of the directional drill pipe that generates the variable cell constraint effect is carried out. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is the flow chart of the monitoring method of the present invention;

[0019] Figure 2 is the force diagram of the radial node force and the axial node force;

[0020] Figure 3 is the three-dimensional variable cell constraint state diagram;

[0021] Figure 4 is the schematic diagram of the equal-curvature model of a single-section drill pipe;

[0022] Figure 5 is the structure diagram of KAN;

[0023] Figure 6 is the structure diagram of the equivariant graph mechanics network model.

[0024] Figure 7 is the geometric equivalent message passing graph. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a 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 of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0026] As Figure 1 shown, this embodiment provides a few-sensor monitoring method for the pose of a shale gas drill pipe based on the variable cell theory, including the following steps:

[0027] First, the drill pipe for shale gas extraction is simplified into a continuous multi-segment body, and a multi-segment linear body model with the stretching effect of a flexible rope-like structure is used to describe the rope tension of different magnitudes, and its formula expression is:

[0028]

[0029] m i u tt (x, t) = F e+1 -F e +f i (x, t)

[0030] F e = k e (a i -a j )

[0031] {α} = [K] -1 {F}

[0032] [K]{α} = {F}

[0033]

[0034] Where: x is the spatial variable, t is the time variable, u tt (x, t) is the acceleration condition varying with time, u xx (x, t) is the degree of "bending" in space, q(x) is a function of the spatial variable, u t (n i , t) is the rate varying with time, n i is the spatial position parameter, U(t) is a function related to the time variable t, m i is the element mass, F e is the axial nodal force, f i (x, t) is the external force, k e is the axial stiffness, a i and a j are the displacements of the two end points respectively, {a} is the nodal displacement vector, [K] is the global stiffness matrix, [k e is the element stiffness matrix, {F} is the nodal force vector, E is the elastic modulus, S is the cross-sectional area, L e is the length element.

[0035] This model can describe the mining drill pipes with different stiffness characteristics, and the interaction force F between the drill pipe and the geology can be simulated by applying nodal radial forces F with different magnitudes and directions at the end of the drill pipe r , as t shown Figure 2 .

[0036] As Figure 3As shown in the figure, the three-dimensional variable cell constraint is used to depict the pose of the drill pipe at the turning points under different stiffness characteristics and different geological conditions. During the mining process, due to obstacles in the path, path adjustment is required, resulting in the deflection of the drill pipe under the combined action of radial force and axial force, and then forming a curved path. The three-dimensional variable cell constraint is used to describe the current relative pose state. The drill pipe has multi-degree-of-freedom deflection in three-dimensional space. Among them, the second, third, fifth, and eighth drill pipes deflect in the X direction in the XZ plane, the fourth drill pipe deflects in the Y direction in the YZ plane to avoid obstacles, the sixth drill pipe deflects in the -Y direction, and the seventh drill pipe deflects in the Z direction in the YZ plane. The adjacency matrix is used to describe the pose information of adjacent drill pipes. For the convenience of explanation in this paper, the poses of adjacent drill pipes in the same state are omitted. The matrix expression is:

[0037]

[0038] In the formula, α, β, and γ are the angles of deflection in the X, Y, and Z axis directions respectively. Combining with the fixed length of the drill pipe, the displacement of the drill pipe in each direction can be calculated.

[0039] As Figure 4 shown, the pose information of the turning point of the drill pipe is converted into a homogeneous coordinate transformation matrix and brought into the multi-segment equal-curvature model;

[0040]

[0041] In the formula: α i is the bending angle, β i is the azimuth angle, L i is the arc length of the segment, Z and Y are the fixed rotations around {0,0,0} and {0,1,0} respectively, k i is the bending radius of the line segment, is the homogeneous transformation matrix of the position on the equal-curvature segment, and l is the arc length of the equal-curvature segment.

[0042] As Figure 5 shown, the Kolmogorov-Arnold network (KAN) is used to replace the multi-layer perceptron (MLP) in the equivariant graph mechanics network model because KAN has higher performance phenotypes in function fitting, accuracy, and parameter efficiency, more prominent function expression and equivariance, and stronger interpretability and scientific application value. The KAN formula is:

[0043]

[0044] In the formula: x is the input vector, Φ is the feature mapping function, q is the dimension of the feature space, n is the number of neurons in the hidden layer, and p is the number of neurons in the output layer.

[0045] As Figure 6As shown, the sensor values after optimized denoising processing are input into the optimized equivariant graph mechanics network model to obtain the pose information of the directional drill pipe in real time. The iterative process of the equivariant graph mechanics network model is as follows:

[0046]

[0047] In the formula: is the interaction force, is the generalized position, is the generalized velocity, is the generalized acceleration, is the position, is the velocity, is the node feature, e ji is the edge feature, Z is the input, σ w is an arbitrary multi-layer perceptron (MLP) with parameter w, is the Euclidean equivariant function.

[0048] Taking the pose information of the key points analyzed by the equal-curvature model as the output and the position information of the sample points directly collected by the sensor as the output, iterative prediction is carried out. The key point data collected is input into the trained equivariant graph mechanics network model, so as to realize the real-time dynamic monitoring of the directional drill pipe that generates variable cell constraint effects under the condition of few sensors.

[0049] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for less - sensor monitoring of the posture of a shale gas drill pipe based on the metamorphic cell theory, characterized in that: S1. Construct a multi - segment linear body model under the action of stretching like a flexible rope, simplify the shale gas extraction drill pipe into a multi - segment continuum, and simulate the interaction force between the drill pipe and the geology; S2. Use three - dimensional metamorphic cell constraints to describe the posture information of the multi - segment linear body model, and use the adjacency matrix to correlate and express the posture information of each turning point. Fit one by one through the multi - segment equal - curvature model to achieve the numerical solution of the theoretical value of the overall shape of the directional drill pipe; S3. Construct a point - position detection system to obtain the spatial position of the drill pipe connection point, judge whether the calibrated spatial position is the true value of the turning point. Take the theoretical value obtained by the multi - segment equal - curvature model as the input and the true value obtained by the point - position detection system as the output to construct an equivariant graph mechanics network model improved by KAN; S4. Substitute the sensor numerical values after optimized denoising processing into the equivariant graph mechanics network model, and then it is possible to realize the shape data of the directional extraction drill pipe with very few incomplete sensing data.

2. The method for less - sensor monitoring of the posture of a shale gas drill pipe based on the metamorphic cell theory according to claim 1, characterized in that: In step S1, for the multi - segment linear body model, its formula expression is: m i u tt (x,t) = F e+1 -F e +f i (x,,t) F e = k e (a i - a j ) {α} = [K] -1 {F} [K]{a}=F} where: x is the spatial variable, t is the time variable, u tt (x,t) is the acceleration varying with time, u xx (x, t) is the degree of "bending" in space, q(x) is a function of the spatial variable, u t (n i , t) is the rate varying with time, n i is the spatial position parameter, U(t) is a function related to the time variable t, m i is the element mass, F e is the axial nodal force, f i (x, t) is the external force, k e is the axial stiffness, a i and a j are the displacements of the two end points respectively, {a} is the nodal displacement vector, [K] is the global stiffness matrix, [k e is the element stiffness matrix, {F} is the nodal force vector, E is the elastic modulus, S is the cross-sectional area, L e is the length element.

3. The method for less - sensor monitoring of the posture of a shale gas drill pipe based on the metamorphic cell theory according to claim 1, characterized in that: In step S2, use the adjacency matrix to describe the posture information of adjacent drill pipes, and the matrix expression is: Where α, β, and γ are the angles of deflection in the X, Y, and Z axis directions respectively. Combining with the fixed length of the drill pipe, the displacement of the drill pipe in each direction can be calculated.

4. The method for less - sensor monitoring of the posture of a shale gas drill pipe based on the metamorphic cell theory according to claim 1, characterized in that: In step S2, convert the posture information of the turning point of the drill pipe into a homogeneous coordinate transformation matrix and substitute it into the multi - segment equal - curvature model; Where: α i is the bending angle, β i is the azimuth angle, L i is the arc length of the segment, Z and Y are the fixed rotations about {0, 0, 0} and {0, 1, 0} respectively, k i is the bending radius of the line segment, is the homogeneous transformation matrix of the position on the constant curvature segment, and l is the arc length of the constant curvature segment.

5. The method for less-sensor monitoring of the posture of a shale gas drill pipe based on metamorphic cell theory according to claim 1, characterized in that: The iterative process of the equivariant graph mechanics network model is: where: is the interaction force, is the generalized position, is the generalized velocity, is the generalized acceleration, is the position, is the velocity, is the node feature, e ji is the edge feature, Z is the input, σ w is an arbitrary multi-layer perceptron with parameter w, FK is an Euclidean equivariant function.

6. The method for less-sensor monitoring of the posture of a shale gas drill pipe based on metamorphic cell theory according to claim 5, characterized in that: Take the posture information of the key points analyzed by the equal - curvature model as the output and the position information of the sample points directly collected by the sensor as the output for iterative prediction. Input the collected key - point data into the trained equivariant graph mechanics network model to realize the real - time dynamic monitoring of the directional drill pipe with metamorphic cell constraints under less - sensor conditions.

7. The method for less-sensor monitoring of the posture of a shale gas drill pipe based on the metamorphic cell theory according to claim 1, wherein: In step S3, the point - position detection system is a system for monitoring the state and parameters of specific points. It realizes real - time monitoring and data collection of the target points through sensors to achieve precise control and management of the target points, and is used to output the real - value data of the collected sample points in step S4.

Images