CN110134006B - Complex borehole trajectory optimization method based on improved multi-target particle swarm optimization - Google Patents

Complex borehole trajectory optimization method based on improved multi-target particle swarm optimization Download PDF

Info

Publication number
CN110134006B
CN110134006B CN201910410779.7A CN201910410779A CN110134006B CN 110134006 B CN110134006 B CN 110134006B CN 201910410779 A CN201910410779 A CN 201910410779A CN 110134006 B CN110134006 B CN 110134006B
Authority
CN
China
Prior art keywords
particles
algorithm
actual
particle
sinφ
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Fee Related
Application number
CN201910410779.7A
Other languages
Chinese (zh)
Other versions
CN110134006A (en
Inventor
沙林秀
李文燕
张奇志
李琳
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xian Shiyou University
Original Assignee
Xian Shiyou University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xian Shiyou University filed Critical Xian Shiyou University
Priority to CN201910410779.7A priority Critical patent/CN110134006B/en
Publication of CN110134006A publication Critical patent/CN110134006A/en
Application granted granted Critical
Publication of CN110134006B publication Critical patent/CN110134006B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/0205—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric not using a model or a simulator of the controlled system
    • G05B13/024—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric not using a model or a simulator of the controlled system in which a parameter or coefficient is automatically adjusted to optimise the performance

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Feedback Control In General (AREA)

Abstract

The method comprises the steps of (1) setting parameters of a multi-target particle swarm algorithm MOPSO, (2) initializing a population; (3) calculating an objective function value, (4) updating the position and the speed of each generation of particles; (5) carrying out mutation operation on the particles; (6) calculating the objective function value of each particle in the population; (7) updating the individual optimum is a process from the beginning of iteration to the current optimum position of the algorithm, (8) sorting a non-dominating set nd, (9) sorting non-inferior solutions in an external file in the MOPSO in a descending order according to objective function values; (10) adopting a truncation method to delete the following redundant individuals; (11) global optimization; (12) obtaining an optimal solution set optimized by an algorithm, namely, the actual measurement length and the actual control torque of the well track are relatively optimal; the invention realizes the multi-target borehole trajectory parameter optimization under the actual drilling condition, improves the drilling success rate, reduces the drilling cost and lays the theoretical decision-making foundation.

Description

Complex borehole trajectory optimization method based on improved multi-target particle swarm optimization
Technical Field
The invention relates to the technical field of wellbore trajectory optimization, in particular to a complex wellbore trajectory optimization method based on an improved multi-objective particle swarm algorithm.
Background
With the increasing change of oil and gas exploration from inland to deep sea, desert and other areas and the increasing number of unconventional, deep water, deep layer, polar region and other oil and gas fields, the drilling technology, the while-drilling data acquisition technology, the comprehensive interpretation method of logging data and the intelligent optimization algorithm which are suitable for the oil and gas exploration are also developed in a long way. In addition, the requirements for well pattern arrangement during oil field development are increasing day by day, and the scanning and collision prevention between wells are receiving more and more attention. Secondly, in order to further increase the yield of the oil field, the development of thin oil reservoirs enters a new stage, the requirement on the position precision of the well track is higher, and the development is developing towards a deeper target. Therefore, it is important to optimize and precisely control the trajectory of the well bore during the drilling process. The method is used for realizing effective, real-time and quick optimization of the well track, and is a premise for realizing accurate control of the well track, improving the target hitting rate and reducing the drilling risk.
The well track optimization is to determine a well track optimization objective function meeting the process requirements on the basis of determining a technical route, a construction scheme and a wellhead position before construction, and to preferably select well track parameters such as a tool face angle, an inclination angle, a curvature radius, a deflecting point range and the like under the constraint conditions of a drilling tool, a stratum and the like so as to achieve the purposes of improving the drilling success rate and saving the drilling cost. However, the existing optimization of the well track mainly takes single-target optimization as a main point, and the result of optimizing the three-dimensional well track often has a large deviation from the actual drilling requirement, for example, a patent with the patent number of 201710132117.9 discloses a 'complex well track optimization method based on a fast adaptive quantum genetic algorithm', and the optimization method only optimizes the length of the well track and cannot really guide the actual drilling process, so that the traditional single-target optimization cannot meet the actual acquisition of well track parameters, and a plurality of optimization targets need to be considered, so that an oil layer is drilled more accurately.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide a complex borehole trajectory Optimization method based on an improved Multi-objective Particle Swarm Algorithm, wherein the improved Multi-objective Particle Swarm Algorithm (Multi-objective Particle Swarm Optimization Algorithm) is adopted to improve the selection modes of the individual optimal position and the global optimal position and apply a disturbance mutation operator to the particles, so that the generation of local optimal position, external set, non-dominant set selection mode and the like are avoided, and the Algorithm improves the Particle Swarm Algorithm to optimize the Multi-objective problem.
In order to achieve the purpose, the technical scheme of the invention is realized as follows:
the complex borehole trajectory optimization method based on the improved multi-objective particle swarm optimization algorithm comprises the following steps of:
(1) and setting parameters of the multi-target particle swarm algorithm MOPSO, including the maximum value and the maximum value of the dynamic inertia weight, an acceleration factor, the population scale and the maximum iteration GEN.
(2) Initializing a population, wherein the population comprises the speed and initial position of particles, the position according with constraint conditions, an external archive, an individual optimum and a global optimum; azimuth angle, inclination angle, dog-leg angle, curvature radius, actual measured length of each well section, actual vertical depth, actual control torque and casing length; wherein the initial positions of the particles in the population are randomly generated, the initial positions of the population are assigned to the initial positions which meet the constraint condition, and the individual optimal position and the global optimal position are set as the particles.
(3) When the initial position of the population meets constraint conditions, calculating an objective function value, wherein the constraint conditions comprise a value range of an independent variable, a casing length range, a target vertical well Depth non-negative constraint range, a dog leg angle range in a stratum and non-negative constraints in the optimization problem of the actual complex well track, the non-negative constraints mean that the actual Measurement Depth and the vertical Depth cannot be negative, selecting 12 geometric parameters of the well track for optimization, and realizing that the actual Measurement Depth TMD (TMD) of a dual target to be optimized and the actual Control Torque TCT (TCT) reach relative optimization, and the geometric parameters comprise a deflecting point Depth, a well inclination angle and an azimuth angle;
wherein the objective function can be defined as:
obj_function＝min{TMD、TCT}
wherein: TMD ═ Dkop+D1+D2+D3+D4+D5+HD
TCT＝T1+T2+T3+T4+T5+T6+T7
s.t.xmin≤x≤xmax (1)
TVDmin≤TVD≤TVDmin
Cimin≤Ci≤Cimax(i＝1,2,3)
Ds＞0(s＝1,2,3,4,5)
In the formula (1)
Figure BDA0002062744670000031
I.e. the feasible solution space R12The system consists of 12-dimensional decision vectors X, namely parameters to be optimized; TMD and TCT are optimization objective functions, and the units are ft and N.ft respectively; i number of stages designed for casing, TVDmax，TVDminRespectively the upper and lower limits of the actual vertical depth of the well track.
The calculation formula of each section of the borehole trajectory is defined as:
Figure BDA0002062744670000041
D2＝(Dd-Dkop-D1×(sinφ1-sinφ0)/(φ1-φ0))/cos(φ1) (3)
Figure BDA0002062744670000042
D4＝(DB-Dd-D3×(sinφ2-sinφ1)/(φ2-φ1))/cos(φ2) (5)
Figure BDA0002062744670000043
the total actual measured length is:
TMD＝Dk+D1+D2+D3+D4+D5+HD (7)
wherein D1 and D5 are the first and second ramp sections, theta1～θ2: inclination at two measuring pointsOblique angle;
Figure BDA0002062744670000044
the azimuth is shown for two measuring points.
The actual measurement curve length increment of the deflection increasing section is as follows:
Figure BDA0002062744670000045
in the formula (8), r is a curvature radius,
Figure BDA0002062744670000046
the incremental calculation of the curve segment of equation (8) in three-dimensional coordinates can be defined as:
Figure BDA0002062744670000047
Figure BDA0002062744670000048
Figure BDA0002062744670000049
the stress F1-F7 of each well section is calculated as follows:
F7＝0 (13)
F6＝F7+Bw·HDcosφ3＝Bw·HDcosφ3 (14)
F5＝F6+Bw·D5(sinφ3-sinφ2)/(φ3-φ2) (15)
F4＝F5+Bw·D4cosφ2 (16)
F3＝F4+Bw·D3(sinφ2-sinφ1)/(φ2-φ1) (17)
F2＝F3+Bw·D2cosφ1 (18)
F1＝F2+Bw·D1(sinφ1-sinφ0)/(φ1-φ0) (19)
calculating torques T1-T7 corresponding to F1-F7 according to the following calculation formula:
T1＝μrw·Dksinφ0 (20)
Figure BDA0002062744670000051
T3＝μrw·D2sinφ1 (22)
Figure BDA0002062744670000052
T5＝μrw·D4sinφ2 (24)
Figure BDA0002062744670000053
T7＝μrw·HD·sinφ3 (26)
the total actual control torque is:
TCT＝T1+T2+T3+T4+T5+T6+T7 (27)
the length calculation formula of each section of the casing is as follows:
C1＝Dk+D1*sin(θ1)/θ1 (28)
Figure BDA0002062744670000054
Figure BDA0002062744670000061
the actual sag depth of each segment is calculated as follows:
Figure BDA0002062744670000062
Figure BDA0002062744670000067
TVDdrop＝D3*(sin(rad*θ2)-sin(rad*θ1))/((θ2-θ1)*rad) (33)
Figure BDA0002062744670000063
Figure BDA0002062744670000064
the total actual vertical depth is:
Figure BDA0002062744670000065
the meanings and the value ranges of the parameters in the formulae (2) to (36) are shown in table 2. And (3) calculating the actual measurement lengths D1-D5, the actual stresses F1-F7 and the corresponding actual control torques T1-T7 of the trajectories of the well sections by the formulas (2) to (25). Calculating the target functions TMD and TCT by the formulas (1) to (36), storing the optimal solution meeting the constraint condition, temporarily storing the first generation of optimal feasible solution and optimal target function value meeting the constraint condition as the global optimal solution and global target function, calculating the target adaptability values TMD and TCT according to the formula (1) along with the increase of the iteration times, and recording the current optimal solution.
(4) Updating the position and speed of each generation of particles according to the formula (37) and the formula (38), and if the particles exceed the boundary in the process, adopting boundary constraint; the velocity update mode of the particles is shown in formula (37).
Figure BDA0002062744670000066
In the formula (37), vpj(t) is the speed of the particle p in the jth dimension at time t; psowc1And psowc2Is a positive acceleration constant; r is1j(t),r2j(t) is the interval [0,1]]The random number generated in (1); yp represents the current optimal position of the particle p;
Figure BDA0002062744670000071
finding optimal positions for all particles in the current population; psow is the introduced inertial weight.
The position of the particle is updated in a matrix manner, and the velocity of the particle is updated in a manner shown in formula (38).
xp(t+1)＝xp(t)+vp(t+1) (38)
In the formula (38), xp(t) is the position of the particle p at time t, xpj(t) is the position of the particle p in dimension j at time t, and xi～U(xmin，xmax)。
(5) And carrying out mutation operation on the particles, and adopting a mutation operator to act on the MOPSO so as to guide the flight of the particles, improve the ability of the population to jump out of the local optimum and strengthen the local and global search strength. After the algorithm is searched for a period of time, the number of individuals participating in mutation is reduced, local development is carried out, and mutation operators are adopted to apply disturbance to the particles, so that the particles are prevented from falling into local optimum; for particle p, the positional variation of the particle is:
xp＝xp+16*varsig*(1-randr)*vp(39) in the formula (39), varsig ═ 1 indicates whether the direction of movement of the particles after variation is the same as the original direction of movement; 16 is a coefficient that enables the particles to jump out of the local optimum position; v. ofpIs the mutation probability.
Figure BDA0002062744670000072
In the formula (40), ite is the current algebra of the algorithm, and GEN is the maximum iteration number of the algorithm.
Generating a random number randr in an interval [0,1] for each particle in the population, and carrying out mutation operation on the particles when the randr is less than randp, otherwise, not carrying out mutation; in addition, when the particles are varied, it is necessary to define the particles that exceed the boundary of the domain on the boundary.
(6) Aiming at the particles meeting the constraint conditions, calculating an objective function value of each particle in the population according to each target of each particle; and (5) if the constraint condition is not satisfied for the particles which do not satisfy the constraint condition for four times, redistributing the speed and the position of the particles according to the steps (4) and (5), and restricting the variation and the boundary of the particles.
(7) Updating the individual optimal algorithm, iterating from the beginning to the current optimal position, and if the current position x dominates the individual extreme value position xpThen update to current position x.
(8) And sequencing the non-dominating set nd, after local optimal updating is carried out on each particle in the population, storing non-inferior solutions in algorithm iteration, searching for optimal individuals by adopting a non-dominating set algorithm, and quickly sequencing the non-dominating set by adopting a multi-target domination relationship.
(9) With the iteration, each group of non-inferior solutions is compared with the solution of the current non-dominating set one by one, and then the non-inferior solutions in the external files in the MOPSO are arranged in a descending order according to the objective function values, so that the calculation complexity of the algorithm is reduced.
(10) And dynamically controlling the scale of the external files by adopting a dynamic congestion distance method, and when the external set ex exceeds a set range, deleting the subsequent redundant individuals by adopting a truncation method.
(11) And (3) global optimization, namely selecting a particle from the foremost part of the external files after descending sorting as the global optimization, guiding the particle swarm to continuously search for a better solution, and ensuring the distributivity of the algorithm.
(12) if the iteration time is ite < GEN, continuing to the step (3); otherwise, outputting the particles in the external set to obtain an optimal solution set optimized by the algorithm, namely the actual measurement length of the well track and the actual control torque reach relative optimization.
According to the method, an external set is constructed by adopting a dominance relation, non-dominance solutions found from the beginning to the present of the population are saved, the guiding algorithm approaches to a Pareto front end more quickly, and the distributivity of feasible solutions is kept. And a mutation operator is introduced to disturb the particles, so that the algorithm is prevented from falling into local optimization, the selection mode of individual optimization and global optimization positions is improved, a multi-target particle swarm optimization MOPSO is designed, the MOPSO is used for optimizing the actual measurement depth TMD and the actual control torque TCT of the complex three-dimensional well track, the optimization of the optimal parameters of the multiple well tracks, such as the depth of a deviation point, the well inclination angle and the azimuth angle, is completed, the multi-target well track parameter optimization under the actual drilling condition is realized, the drilling success rate is improved, and the drilling cost is reduced.
The experimental result of the MOPSO for realizing the complex borehole trajectory optimization problem solution shows that the optimized TMD and TCT results overcome the defect that the traditional borehole trajectory single-target optimization has larger deviation with the actual drilling requirement in real drilling, the borehole trajectory parameters under multiple target conditions under the actual drilling condition are obtained, the method is applied to the borehole trajectory optimization of the intelligent drilling process, the requirement of the borehole trajectory actual parameter optimization is basically met, the decision of drilling workers in the borehole trajectory process is facilitated, and therefore the oil layer can be reached more accurately. Through the research on the well track MOPSO, a theoretical decision basis is laid for further realizing interactive multi-target dynamic optimization of the well track while drilling, accurate control, deviation correction and collision prevention of the well track, improvement of the target hitting rate, multi-target well track parameter optimization under the actual drilling condition, improvement of the drilling success rate and reduction of the drilling cost.
Drawings
FIG. 1 is a vertical cross-section of a complex wellbore trajectory.
FIG. 2 is a three-dimensional schematic of the D1, D5 ramp segment.
Fig. 3 is a schematic diagram of torques T3 and T5 corresponding to steady-slope stages of D2 and D4.
FIG. 4 is a schematic illustration of the torques T2, T6 for the D1 and D5 ramp-up sections.
FIG. 5 is a schematic diagram of torque T4 corresponding to the D3 ramp down segment.
FIG. 6 is a graph of the optimization results of using MOPSO to achieve a three-dimensional borehole trajectory.
Description of the attached tables
Table 4 compares the preferred wellbore trajectory parameters with the single-objective optimization results of other intelligent algorithms after the practice of the present invention.
Detailed Description
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
The complex borehole trajectory optimization method based on the improved multi-objective particle swarm optimization algorithm comprises the following steps of:
(1) and setting parameters of the multi-target particle swarm algorithm MOPSO, including the maximum value and the maximum value of the dynamic inertia weight, an acceleration factor, the population scale and the maximum iteration GEN.
The constraint conditions and the argument constraint boundary conditions are shown in table 2.
TABLE 1 MOPSO parameter settings
Figure BDA0002062744670000101
In table 1, POP is a population scale, and ex POP is an external file scale, and POP is generally taken as ex POP; GEN is the maximum iteration number; in MOPSO, psoc1 and psoc2 are accelerators; psowmax、psowminThe maximum value and the minimum value of the inertia weight are obtained.
TABLE 2 wellbore trajectory variable constraint boundaries and constraints
Figure BDA0002062744670000102
Figure BDA0002062744670000111
(2) Initializing a population, wherein the population comprises the speed and initial position of particles, the position according with constraint conditions, an external archive, an individual optimum and a global optimum; azimuth angle, inclination angle, dog-leg angle, curvature radius, actual measured length of each well section, actual vertical depth, actual control torque and casing length; wherein the initial positions of the particles in the population are randomly generated, the initial positions of the population are assigned to the initial positions which meet the constraint condition, and the individual optimal position and the global optimal position are set as the particles.
(3) When the initial position of the population meets constraint conditions, calculating an objective function value, wherein the constraint conditions comprise a value range of an independent variable, a casing length range, a target vertical well Depth non-negative constraint range, a dog leg angle range in a stratum and non-negative constraints in the optimization problem of the actual well track, the non-negative constraints mean that the actual Measurement Depth and the vertical Depth cannot be negative, selecting 12 geometric parameters of the well track for optimization, and realizing that the actual Measurement Depth TMD (TMD) of a binocular target to be optimized and the actual Control Torque TCT (TCT) reach relative optimality, and the geometric parameters comprise a deflecting point Depth, a well inclination angle and an azimuth angle;
wherein a vertical cross-section of a complex wellbore trajectory to be optimized is shown in figure 1.
Wherein the objective function can be defined as:
Figure BDA0002062744670000121
in the formula (1)
Figure BDA0002062744670000122
I.e. the solution space R12The system consists of 12-dimensional decision vectors X, namely parameters to be optimized; TMD and TCT are optimization objective functions, the unit is ft and N x ft respectively, and i is the number of sections designed by the sleeve. Wherein, D1: a First burst-up section First section deflecting section; d2: tangent section positive cutting segment D3: a Drop-off ramp-down section; d4: a Hold section steady-slope section; d5: a Second slope increasing section of the Second structured-up section; HD: horizontal section wellbore length, ft; TVDmax，TVDminRespectively the upper and lower limits of the actual vertical depth of the well track.
In fig. 1, the calculation formula for each section of the borehole trajectory is defined as:
Figure BDA0002062744670000123
D2＝(Dd-Dkop-D1×(sinφ1-sinφ0)/(φ1-φ0))/cos(φ1) (43)
Figure BDA0002062744670000124
D4＝(DB-Dd-D3×(sinφ2-sinφ1)/(φ2-φ1))/cos(φ2) (45)
Figure BDA0002062744670000131
the total actual measured length is:
TMD＝Dk+D1+D2+D3+D4+D5+HD (47)
wherein D1 and D5 are the first and the second oblique increasing sections, and the three-dimensional schematic diagram is shown in figure 2. Theta1～θ2The inclination angles of the two measuring points are set;
Figure BDA0002062744670000132
the azimuth is shown for two measuring points.
In fig. 2, the actual measured curve length increment of the increased slope section is:
Figure BDA0002062744670000133
in the formula (48), r is a radius of curvature,
Figure BDA0002062744670000134
the incremental calculation of the curve segment of equation (48) in three-dimensional coordinates may be defined as:
Figure BDA0002062744670000135
Figure BDA0002062744670000136
Figure BDA0002062744670000137
in fig. 3 to 5, P1 and P2 are two observation points; e1, e2 is a unit vector in the direction of the wellbore; f1 is the axial force at the bottom of the directional well, N; f2 is the axial force at the top of the directional well, N; beta is the total angular change; b is buoyancy coefficient, B is 0.7; mu is friction coefficient, and mu is 0.3; w is the unit drilling tool weight w is 0.3 kN/ft; and r is the radius r of the drill rod which is 10 mm.
The stress F1-F7 of each well section is calculated as follows:
F7＝0 (53)
F6＝F7+Bw·HDcosφ3＝Bw·HDcosφ3 (54)
F5＝F6+Bw·D5(sinφ3-sinφ2)/(φ3-φ2) (55)
F4＝F5+Bw·D4cosφ2 (56)
F3＝F4+Bw·D3(sinφ2-sinφ1)/(φ2-φ1) (57)
F2＝F3+Bw·D2cosφ1 (58)
F1＝F2+Bw·D1(sinφ1-sinφ0)/(φ1-φ0) (59)
calculating torques T1-T7 corresponding to F1-F7 according to the following calculation formula:
T1＝μrw·Dksinφ0 (60)
Figure BDA0002062744670000141
T3＝μrw·D2sinφ1 (62)
Figure BDA0002062744670000142
T5＝μrw·D4sinφ2 (64)
Figure BDA0002062744670000143
T7＝μrw·HD·sinφ3 (66)
the total actual control torque is:
TCT＝T1+T2+T3+T4+T5+T6+T7 (67)
the length calculation formula of each section of the casing is as follows:
C1＝Dk+D1*sin(θ1)/θ1 (68)
Figure BDA0002062744670000144
Figure BDA0002062744670000145
the actual sag depth of each segment is calculated as follows:
Figure BDA0002062744670000151
Figure BDA0002062744670000152
TVDdrop＝D3*(sin(rad*θ2)-sin(rad*θ1))/((θ2-θ1)*rad) (73)
Figure BDA0002062744670000153
Figure BDA0002062744670000154
the total actual vertical depth is:
Figure BDA0002062744670000155
the meanings and the value ranges of the parameters in the formulae (42) to (76) are shown in table 2. And calculating actual measurement lengths D1-D5, actual stresses F1-F7 and corresponding actual control torques T1-T7 of the trajectories of the well sections by the formulas (42) - (75). Calculating target functions TMD and TCT and storing optimal solutions meeting constraint conditions by the formulas (1) to (36), temporarily storing the first generation of optimal feasible solutions and optimal target function values meeting the constraint conditions as global optimal solutions and global target functions, calculating target fitness values TMD and TCT according to the formula (1) along with the increase of iteration times, and recording the individual optimal solution xp, the global optimal xg, the individual optimal target function yp and the global optimal target function yg; if the constraint condition is met, saving the global optimal solution xg and the global optimal objective function value yg according to the domination relation; otherwise, it remains unchanged. Let ite be ite + 1;
(4) updating the position and speed of each generation of particles according to the formula (77) and the formula (78), and if the particles exceed the boundary in the process, adopting boundary constraint; according to the application research of the improved MOPSO, the speed influences the local optimization capability of the algorithm, a larger speed corresponds to a stronger global optimization capability, and a smaller speed is beneficial to the local optimization of particles. In the invention, MOPSO adopts dynamically changed inertial weight psow to facilitate population to keep diversity and searching capability, the later development capability of the algorithm is increased, and the speed updating mode of particles is shown as a formula (77).
Figure BDA0002062744670000156
Figure BDA0002062744670000161
In the formula (77), vpj(t) is the speed of the particle p in the jth dimension at time t; psowc1 and psowc2 are positive acceleration constants; r is1j(t),r2j(t) is the interval [0,1]]The random number generated in (1); yp represents the current optimal position of the particle p;
Figure BDA0002062744670000162
finding optimal positions for all particles in the current population; psow is the introduced inertial weight;
a matrix approach is used to update the position of the particle, and the velocity of the particle is updated as shown in equation (78).
xp(t+1)＝xp(t)+vp(t+1) (78)
In the formula (78), xp(t) is the position of the particle p at time t, xpj(t) is the position of the particle p in dimension j at time t, and xi～U(xmin，xmax)。
(5) Performing mutation operation on the particles. Both the particle swarm algorithm and the genetic algorithm are population algorithms which are widely applied, but GA generates new individuals based on intersection and variation, and the particle swarm algorithm is generated by learning from the optimal positions of the individuals and the global optimal positions. For the advantages and disadvantages of the two algorithms, for example, although the GA has strong global search capability, the GA has a slow convergence rate and a high computational complexity; the particle swarm algorithm has a faster convergence speed, but when the problem of multiple peaks is encountered, the algorithm is easy to fall into a local optimal solution, so that a hybrid algorithm combining the two algorithms appears to improve respective disadvantages.
In view of the successful application of mutation operation in the single-target particle swarm algorithm, the invention adopts a mutation operator to act on the MOPSO to guide the flight of the population, improve the ability of the population to jump out of the local optimum and strengthen the local and global search strength. In addition, after the algorithm is searched for a period of time, the number of individuals participating in mutation needs to be reduced, local development is carried out, and a mutation operator is adopted to apply disturbance to the particles, so that the particles are prevented from falling into local optimum; for particle p, the positional variation of the particle is:
xp＝xp+16*va.rsig(-randr)*vp (79)
in the formula (79), varsig ═ 1 indicates that the particle has changed and moved in the original directionWhether they are the same; 16 is a coefficient that enables the particles to jump out of the local optimum position; v. ofpIs the mutation probability.
Figure BDA0002062744670000171
In the formula (80), the ite is the current algebra of the algorithm, and the GEN is the maximum iteration number of the algorithm.
Generating a random number randr in an interval [0,1] for each particle in the population, and carrying out mutation operation on the particles when the randr is less than randp, otherwise, not carrying out mutation; in addition, when the particles are varied, it is necessary to define the particles that exceed the boundary of the domain on the boundary.
(6) For particles meeting the constraint conditions, calculating an objective function value of each particle in the population according to each target of each particle; if the constraint condition is not satisfied for the particles that do not satisfy the constraint condition four times in succession, the velocity and position of the particles need to be reassigned according to the steps (4) and (5), and the variation and boundary of the particles need to be constrained.
(7) Updating the individual optimum is the process of the algorithm from the beginning iteration to the current optimum position, if the current position x dominates the individual extreme position xpThen update to current position x.
(8) The non-dominating set nd is ordered, and after each particle in the population is locally and optimally updated, non-inferior solutions in algorithm iteration need to be stored. In order to improve the efficiency of the MOPSO and reduce the computational complexity of the MOPSO, the algorithm searches for the optimal individual by adopting a non-dominating set algorithm, and the non-dominating set is selected and quickly ordered by adopting a multi-target domination relationship.
The specific method comprises the following steps:
selecting an individual apb in a population ap, usually selecting a first individual, and deleting the apb from the ap;
② comparing other individuals in the population with the individual apb. The population is divided into two groups, one is composed of individuals dominated by individual apb, and the other is composed of individuals dominated by individual apb or unrelated to individual apb;
if the individual apb is not dominated by other individuals in the population, namely the individual apb is a non-dominated solution, putting the individual apb into a non-dominated set nd; otherwise, the materials are not put in;
fourthly, deleting the first type of individuals from the ap, and if the ap is not empty, turning to the fourth step;
when ap is empty, nd is the required non-dominating set.
The method constructs a non-dominating set, the population at the beginning of each cycle is composed of the dominating individual apb obtained last time or individuals irrelevant to the individual apb, the whole population is not, the range is reduced when individual comparison is carried out, and the running speed is improved.
(9) The external set ex is updated. With the iteration, each group of non-inferior solutions is compared with the solution of the current non-dominating set one by one, and then the non-inferior solutions in the external files in the MOPSO are arranged in a descending order according to the objective function values, so that the calculation complexity of the algorithm is reduced.
The updating method comprises the following steps:
firstly, when an external set is empty, namely when an algorithm just starts to run, putting individuals in a non-dominating set nd into the external set ex;
selecting a certain individual p in the non-dominating set nd when the external set is not empty, comparing the individual p in the external set ex with the individual p, and if the individual p is dominated by the individual in the external set ex, not putting the individual p into the external set ex; otherwise, the individual p is placed in the external set ex and those dominated by the individual p are deleted.
And thirdly, circulating until the comparison is finished.
(10) And dynamically controlling the scale of the external files by adopting a dynamic congestion distance method. In order to maintain the distribution of the external files, improve the distribution of non-inferior solutions and the global search capability of the algorithm, improve the efficiency of the algorithm, need to keep the particles at the sparse position of the external files ex, reduce the particles at the dense position, need to recalculate the crowding distance, and control the scale of the external set. And when the external set ex exceeds the set range, adopting a truncation method to delete the subsequent redundant individuals.
The congestion distance calculation method is as follows:
firstly, initializing and setting an expPOP individual congestion distance dis as 0;
sorting ex in ascending order according to the value of the yd-th target function;
thirdly, after sorting, setting the crowding distance of the first and last individuals to be infinite;
and fourthly, calculating the crowding distance of the other individuals p according to the formula (80).
Figure BDA0002062744670000191
In equation (80), ex (p, j) is the jth objective function value of the pth particle in ex.
And fifthly, changing the value of yd, and turning to the step two until all target dimensions are traversed.
(11) The global optimum selects a particle from the forefront part in the external archive after descending sorting as the global optimum, guides the particle group to continuously search for a better solution, and ensures the distributivity of the algorithm. When the MOPSO solves the multi-objective problem, each iteration generates a group of non-inferior solutions. The global optimum setting is chosen from the individuals in the top 10% of the outer set. The selection mode of the global optimal position has different global optimal positions for each particle in the population, and ensures the distribution of the algorithm.
(12) if the iteration time is ite < GEN, continuing to the step (3); otherwise, outputting the particles in the external set to obtain an optimal solution set optimized by the algorithm, namely the actual measurement length of the well track and the actual control torque reach relative optimization.
The external file ex is the best result obtained by the algorithm and outputs the result. Outputting a global optimal position xgAnd an optimal objective function yg. Namely the output
Figure BDA0002062744670000192
ygTMD (TCT). The results of the MOPSO-based complex wellbore trajectory optimization were compared to the results of the single-objective optimization of several other intelligent optimization algorithms, as shown in Table 3. The preferred results of the simulation results using MOPSO to optimize TMD and TCT are shown in FIG. 6。
Three sets of optimal solutions selected in Table 3
Figure BDA0002062744670000201
Table 3 is the three sets of globally optimal objective function solutions selected. Through comparison of three groups of global optimal solutions, it can be found that the actual measurement length in the solution of the group 2 is not changed much, but the actual control torque is much higher than that of the solution of the group 1, and the actual measurement length in the solution of the group 3 is reduced, but the actual control torque is much higher than that of the solution of the group 2, so that the parameters in the optimization process of the well drilling well track are influenced and restricted mutually. Therefore, in the actual drilling process, when the actual measurement length meets the actual requirement in a certain range, the 1 st group of solutions are selected as the optimal solutions, namely the requirements of the actual measurement length are met, and the actual control torque can be effectively reduced.
To further illustrate the optimized performance of the improved MOPSO algorithm, the simulation results of the improved MOPSO algorithm of the present invention are compared with the classical and GA (Shokir et al.2004), NPSO (Amin Atashnezhad2014) and PSO (Shokir et al.2004), and the specific data are shown in table 4.
Table 4 comparison of MOPSO optimized borehole trajectory optimization results with other algorithms for single-objective optimization results
Figure BDA0002062744670000211
As can be seen from Table 4, in the optimization of the complex three-dimensional borehole trajectory, the MOPSO is adopted to realize the optimization of the multi-objective TMD and the TCT, the optimization result is better, the result of the single-objective optimization is basically reserved, and the requirement of the multi-objective optimization in practice is met.
The MOPSO is adopted to realize the optimization of the three-dimensional borehole trajectory, and the optimization results of TMD and TCT are shown in FIG. 6. FIG. 6 adopts MOPSO to realize the optimization of complex three-dimensional borehole trajectory TMD and TCT, the abscissa is the actual measurement length TMD of the optimal borehole trajectory, and the ordinate is the actual control torque of the optimal borehole trajectory. As can be seen in FIG. 6, the Pareto front of the two objective optimization problems for the wellbore trajectory parameters shows that the value of one objective function decreases as the value of the other objective function increases. The Pareto boundary does not mean that the lowest torque corresponds to the longest wellbore trajectory, or that the shortest wellbore trajectory corresponds to the greatest torque. The two objective functions are nonlinear, and the optimal solution is in a descending trend.
According to the method, an external set is constructed by adopting a dominance relation, non-dominance solutions found from the beginning to the present of the population are saved, the guiding algorithm approaches to a Pareto front end more quickly, and the distributivity of feasible solutions is kept. And a mutation operator is introduced to disturb the particles, so that the algorithm is prevented from falling into local optimization, the selection mode of individual optimization and global optimization positions is improved, a multi-target particle swarm optimization MOPSO is designed, the MOPSO is used for optimizing the actual measurement depth TMD and the actual control torque TCT of the complex three-dimensional well track, the optimization of the optimal parameters of the multiple well tracks, such as the depth of a deviation point, the well inclination angle and the azimuth angle, is completed, the multi-target well track parameter optimization under the actual drilling condition is realized, the drilling success rate is improved, and the drilling cost is reduced.
The experimental result of the MOPSO for realizing the complex borehole trajectory optimization problem solution shows that the optimized TMD and TCT results overcome the defect that the traditional borehole trajectory single-target optimization has larger deviation with the actual drilling requirement in real drilling, the borehole trajectory parameters under multiple target conditions under the actual drilling condition are obtained, the method is applied to the borehole trajectory optimization of the intelligent drilling process, the requirement of the borehole trajectory actual parameter optimization is basically met, the decision of drilling workers in the borehole trajectory process is facilitated, and therefore the oil layer can be reached more accurately. Through the research on the well track MOPSO, a theoretical decision basis is laid for further realizing interactive multi-target dynamic optimization of the well track while drilling, accurate control, deviation correction and collision prevention of the well track, improvement of the target hitting rate, multi-target well track parameter optimization under the actual drilling condition, improvement of the drilling success rate and reduction of the drilling cost.

Claims (2)

1. The complex borehole trajectory optimization method based on the improved multi-objective particle swarm optimization algorithm is characterized by comprising the following steps of:
(1) setting parameters of a multi-target particle swarm algorithm MOPSO, including the maximum value and the maximum value of the dynamic inertia weight, an acceleration factor, a population scale and a maximum iteration GEN;
(2) initializing a population, wherein the population comprises the speed and initial position of particles, the position according with constraint conditions, an external archive, an individual optimum and a global optimum; azimuth angle, inclination angle, dog-leg angle, curvature radius, actual measured length of each well section, actual vertical depth, actual control torque and casing length; wherein the initial positions of the particles in the population are randomly generated, the initial positions of the population are assigned to the initial positions which accord with the constraint conditions, and the individual optimal position and the global optimal position are set as the particles per se;
(3) when the initial position of the population meets constraint conditions, calculating an objective function value, wherein the constraint conditions comprise the value range of independent variables, the casing length range, the non-negative constraint range of the target vertical well depth, the dog leg angle range in the stratum and non-negative constraints in the optimization problem of the actual well track, the non-negative constraints mean that the actual measurement depth and the vertical depth cannot be negative, selecting 12 geometric parameters of the well track for optimization, and realizing that the actual measurement depth TMD and the actual control torque TCT of the binocular target to be optimized reach relative optimization, wherein the 12 geometric parameters comprise the deflecting point depth Dd、DB、DkWell angle theta1～θ3And azimuth angle
Figure FDA0003561168070000011
(4) Updating the position and speed of each generation of particles according to the formula (37) and the formula (38), and if the particles exceed the boundary in the process, adopting boundary constraint; the velocity of the particles is updated as shown in equation (37),
Figure FDA0003561168070000012
in the formula (37), vpj(t) is the speed of the particle p in the jth dimension at time t; psowc1And psowc2Is a positive acceleration constant; r is1j(t)，r2j(t) is the interval [0,1]]The random number generated in (1); y ispRepresenting the current optimal position of the particle p;
Figure FDA0003561168070000021
finding optimal positions for all particles in the current population; psow is the introduced inertial weight;
updating the positions of the particles in a matrix mode, wherein the speed updating mode of the particles is shown as a formula (38);
xp(t+1)＝xp(t)+vp(t+1) (38)
in the formula (38), xp(t) is the position of the particle p at time t, xpj(t) is the position of the particle p in dimension j at time t, and xi～U(xmin，xmax)；
(5) Carrying out mutation operation on the particles, and adopting a mutation operator to act on MOPSO (metal oxide semiconductor) so as to guide the flight of the particles, improve the ability of the population to jump out of the local optimum and strengthen the local and global search strength; in addition, after the algorithm is searched for a period of time, the number of individuals participating in mutation is reduced, local development is carried out, and mutation operators are adopted to apply disturbance to the particles, so that the particles are prevented from falling into local optimum; for particle p, the positional variation of the particle is:
xp＝xp+16*varsig*(1-randr)*vp (39)
in the formula (39), varsig ═ 1 indicates whether the direction of movement of the particles after variation is the same as the original direction of movement; 16 is a coefficient that enables the particles to jump out of the local optimum position; v. ofpIs the variation probability;
Figure FDA0003561168070000022
in the formula (40), the ite is the current algebra of the algorithm, and the GEN is the maximum iteration number of the algorithm;
generating a random number randr in an interval [0,1] for each particle in the population, and carrying out mutation operation on the particles when the randr is less than randp, otherwise, not carrying out mutation; in addition, when the particles are changed, the particles beyond the boundary of the definition domain are defined on the boundary;
(6) aiming at the particles meeting the constraint conditions, calculating an objective function value of each particle in the population according to each target of each particle; if the constraint condition is not satisfied for the particles which do not satisfy the constraint condition for four times continuously, the speed and the position of the particles are redistributed according to the step (4) and the step (5), and the variation and the boundary constraint of the particles are carried out;
(7) updating the process of the individual optimal algorithm from the beginning iteration to the current optimal position, if the current position x dominates the position x of the individual extreme value thereofpIf yes, updating the current position x;
(8) ordering a non-dominating set nd, after local optimal updating is carried out on each particle in the population, storing non-inferior solutions in algorithm iteration, searching optimal individuals by adopting a non-dominating set algorithm, and quickly ordering the non-dominating set by adopting a multi-target domination relationship;
(9) with the iteration, each group of non-inferior solutions is compared with the solution of the current non-dominating set one by one, and then the non-inferior solutions in the external files in the MOPSO are arranged in a descending order according to the objective function value, so that the calculation complexity of the algorithm is reduced;
(10) dynamically controlling the scale of the external files by adopting a dynamic congestion distance method, and when the external set ex exceeds a set range, deleting the following redundant individuals by adopting a truncation method;
(11) selecting a particle from the foremost part of the external files after descending sorting as a global optimum, guiding the particle swarm to continuously search a better solution, and ensuring the distributivity of the algorithm;
(12) if the iteration time is ite < GEN, continuing to the step (3); otherwise, outputting the particles in the external set to obtain an optimal solution set optimized by the algorithm, namely the actual measurement length of the well track and the actual control torque reach relative optimization.
2. The complex wellbore trajectory optimization method based on the improved multi-objective particle swarm algorithm according to claim 1,
in step (3), the objective function can be defined as:
Figure FDA0003561168070000041
in the formula (1)
Figure FDA0003561168070000042
I.e. the feasible solution space R12The system consists of 12-dimensional decision vectors X, namely parameters to be optimized; TMD and TCT are optimization objective functions, and the units are ft and N.ft respectively; i number of stages designed for casing, TVDmax，TVDminRespectively the upper and lower limits of the actual vertical depth of the well track;
the calculation formula of each section of the borehole trajectory is defined as:
Figure FDA0003561168070000043
D2＝(Dd-Dkop-D1×(sinφ1-sinφ0)/(φ1-φ0))/cos(φ1) (3)
Figure FDA0003561168070000044
D4＝(DB-Dd-D3×(sinφ2-sinφ1)/(φ2-φ1))/cos(φ2) (5)
Figure FDA0003561168070000045
the total actual measured length is:
TMD＝Dk+D1+D2+D3+D4+D5+HD (7)
wherein D1 and D5 are the first and second ramp sections, theta1～θ2: the inclination angles of the two measuring points are set;
Figure FDA0003561168070000051
prescription position angles of two measuring points;
the actual measurement curve length increment of the deflection increasing section is as follows:
Figure FDA0003561168070000052
in the formula (8), r is a curvature radius,
Figure FDA0003561168070000053
the incremental calculation of the curve segment of equation (8) in three-dimensional coordinates can be defined as:
Figure FDA0003561168070000054
Figure FDA0003561168070000055
Figure FDA0003561168070000056
the stress F1-F7 of each well section is calculated as follows:
F7＝0 (13)
F6＝F7+Bw·HDcosφ3＝Bw·HDcosφ3 (14)
F5＝F6+Bw·D5(sinφ3-sinφ2)/(φ3-φ2) (15)
F4＝F5+Bw·D4cosφ2 (16)
F3＝F4+Bw·D3(sinφ2-sinφ1)/(φ2-φ1) (17)
F2＝F3+Bw·D2cosφ1 (18)
F1＝F2+Bw·D1(sinφ1-sinφ0)/(φ1-φ0) (19)
calculating torques T1-T7 corresponding to F1-F7 according to the following calculation formula:
T1＝μrw·Dksinφ0 (20)
Figure FDA0003561168070000061
T3＝μrw·D2sinφ1 (22)
Figure FDA0003561168070000062
T5＝μrw·D4sinφ2 (24)
Figure FDA0003561168070000063
T7＝μrw·HD·sinφ3 (26)
the total actual control torque is:
TCT＝T1+T2+T3+T4+T5+T6+T7 (27)
the length calculation formula of each section of the casing is as follows:
C1＝Dk+D1*sin(θ1)/θ1 (28)
Figure FDA0003561168070000064
Figure FDA0003561168070000065
the actual sag depth of each segment is calculated as follows:
Figure FDA0003561168070000066
Figure FDA0003561168070000067
TVDdrop＝D3*(sin(rad*θ2)-sin(rad*θ1))/((θ2-θ1)*rad) (33)
Figure FDA0003561168070000068
Figure FDA0003561168070000069
the total actual vertical depth is:
Figure FDA00035611680700000610
in the formulae (2) to (36), the meanings of the parameters are as follows:
θ1～θ3-for the well angle, degree, to be optimized;
Figure FDA0003561168070000071
-is the azimuth angle, degree, to be optimized;
Dk-the depth of the kick-off point, m;
dd is the depth of a two-section target point in the depth of the deflecting point, m;
DB-three target point depths in the deflecting point depth, m;
TVD-Length of horizontal wellbore section, m
Calculating actual measurement lengths D1-D5, actual stresses F1-F7 and corresponding actual control torques T1-T7 of the trajectories of the well sections according to the formulas (2) - (25); calculating the target functions TMD and TCT by the formulas (1) to (36), storing the optimal solution meeting the constraint condition, temporarily storing the first generation of optimal feasible solution and optimal target function value meeting the constraint condition as the global optimal solution and global target function, calculating the target adaptability values TMD and TCT according to the formula (1) along with the increase of the iteration times, and recording the current optimal solution.
CN201910410779.7A 2019-05-16 2019-05-16 Complex borehole trajectory optimization method based on improved multi-target particle swarm optimization Expired - Fee Related CN110134006B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910410779.7A CN110134006B (en) 2019-05-16 2019-05-16 Complex borehole trajectory optimization method based on improved multi-target particle swarm optimization

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910410779.7A CN110134006B (en) 2019-05-16 2019-05-16 Complex borehole trajectory optimization method based on improved multi-target particle swarm optimization

Publications (2)

Publication Number Publication Date
CN110134006A CN110134006A (en) 2019-08-16
CN110134006B true CN110134006B (en) 2022-04-29

Family

ID=67574737

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910410779.7A Expired - Fee Related CN110134006B (en) 2019-05-16 2019-05-16 Complex borehole trajectory optimization method based on improved multi-target particle swarm optimization

Country Status (1)

Country Link
CN (1) CN110134006B (en)

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111382543A (en) * 2020-03-13 2020-07-07 中国海洋石油集团有限公司 Offshore cluster well drilling sequence optimization method based on improved ant colony algorithm
CN112101672B (en) * 2020-09-22 2023-05-26 郑州铁路职业技术学院 Gas field group collaborative development optimization method based on particle swarm optimization algorithm
CN112163763B (en) * 2020-09-25 2023-09-01 大连大学 Weapon target distribution solving method based on improved multi-target HQPSOGA algorithm
CN113051775B (en) * 2021-04-13 2023-03-24 中南大学 Horizontal directional drilling track optimization method based on improved radial movement algorithm
CN113205171B (en) * 2021-05-07 2024-05-14 柳州华世通汽车部件股份有限公司 Multi-objective optimization method for discrete binary particle swarm motor
CN113688486B (en) * 2021-07-29 2024-05-28 国网吉林省电力有限公司四平供电公司 Ultrasonic detection method for lead in transformer bushing based on echo model parameter estimation
CN113610314A (en) * 2021-08-17 2021-11-05 中海石油(中国)有限公司 Drilling parameter optimization method, system, equipment and medium based on particle swarm optimization
CN113761730B (en) * 2021-08-27 2023-09-29 浙江理工大学 Tea garden four-bar intertillage mechanism multi-target point track comprehensive method based on improved difference
CN113935132A (en) * 2021-10-15 2022-01-14 福州大学 Parameter optimization method for viscous damper based on improved multi-target particle swarm algorithm
CN116954239B (en) * 2023-06-12 2024-03-19 成都丰千达科技有限公司 Unmanned aerial vehicle track planning method and system based on improved particle swarm optimization

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106408111A (en) * 2016-08-29 2017-02-15 方洋旺 Novel multi-target particle swarm optimization method
CN106940742A (en) * 2017-03-07 2017-07-11 西安石油大学 Bad hole track optimizing method based on quick self-adapted quantum genetic algorithm
CN107992051A (en) * 2017-12-26 2018-05-04 江南大学 Unmanned vehicle paths planning method based on improved multi-objective particle swarm algorithm

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106408111A (en) * 2016-08-29 2017-02-15 方洋旺 Novel multi-target particle swarm optimization method
CN106940742A (en) * 2017-03-07 2017-07-11 西安石油大学 Bad hole track optimizing method based on quick self-adapted quantum genetic algorithm
CN107992051A (en) * 2017-12-26 2018-05-04 江南大学 Unmanned vehicle paths planning method based on improved multi-objective particle swarm algorithm

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
Designing and optimizing deviated wellbore trajectories using novel particle swarm algorithms;AminAtashnezhad等;《Journal of Natural Gas Science and Engineering》;ELSEVIER;20141130;第21卷;第1184-1204页 *
Multi-objective cellular particle swarm optimization for wellbore trajectory design;Jun Zheng等;《Applied Soft Computing》;ELSEVIER;20190430;第77卷;第106-117页 *
基于NPSO的三维复杂井眼轨迹控制转矩的优选;沙林秀等;《石油机械》;20171010;第45卷(第10期);第5-10页 *
基于多目标粒子群算法与TOPSIS的射孔方案优选研究;爨莹等;《科学技术与工程》;20130908;第13卷(第25期);第7302-7306页 *
基于粒子群算法的钻进参数多目标优化研究;张栋栋;《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅰ辑》;中国学术期刊（光盘版）电子杂志社;20150525(第05期);第1-57页 *

Also Published As

Publication number Publication date
CN110134006A (en) 2019-08-16

Similar Documents

Publication Publication Date Title
CN110134006B (en) Complex borehole trajectory optimization method based on improved multi-target particle swarm optimization
WO2022179266A1 (en) Deviation rectification control method and apparatus for shield tunneling attitude
CN113051775B (en) Horizontal directional drilling track optimization method based on improved radial movement algorithm
CN107122861A (en) A kind of Forecast of Gas Emission method based on PCA PSO ELM
CN107843920B (en) A kind of clustering optimization population reservoir attribute prediction technique
CN110348046B (en) Drilling track multi-target optimization method based on fuzzy comprehensive evaluation
CN108388702A (en) Engineering ceramics electrical discharge machining effect prediction method based on PSO neural networks
CN106875050A (en) A kind of Engineering constraint parameter optimization method based on improvement chaos ant colony algorithm
CN108894768B (en) Drilling track design method and system based on bat algorithm and well wall stability
CN110836670B (en) Mixed firework particle swarm cooperation method for solving unmanned aerial vehicle constrained route planning
CN109885061B (en) Improved NSGA-II-based dynamic positioning multi-objective optimization method
CN110738365A (en) flexible job shop production scheduling method based on particle swarm optimization
CN112817312A (en) Path planning method based on double search optimization algorithm
CN106940742A (en) Bad hole track optimizing method based on quick self-adapted quantum genetic algorithm
CN114253215B (en) Civil cabin door automatic drilling and riveting path planning method based on improved ant colony algorithm
CN104390534A (en) Tunnel smooth surface blasting quality control method
Hou et al. Improved particle swarm optimization for selection of shield tunneling parameter values
CN105868858A (en) Method for optimizing track of engraving machine
Xu et al. Research on well trajectory optimization design based on ant colony algorithm
CN111997583A (en) Rotating guide motor speed regulation method based on fuzzy neural network
CN115903816A (en) Low-energy-consumption mobile robot path planning method
Huang et al. Multi-objective drilling trajectory optimization with a modified complexity index
Liu et al. A three-dimensional path planning method of autonomous burrowing robot for lunar subsurface exploration
CN111637882B (en) Differential evolution geomagnetic navigation method based on grid features
CN110439567B (en) Deep punching lamination freezing scheme design method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20220429