CN115640759A - Drill jamming early warning method and system based on machine learning - Google Patents
Drill jamming early warning method and system based on machine learning Download PDFInfo
- Publication number
- CN115640759A CN115640759A CN202211670817.0A CN202211670817A CN115640759A CN 115640759 A CN115640759 A CN 115640759A CN 202211670817 A CN202211670817 A CN 202211670817A CN 115640759 A CN115640759 A CN 115640759A
- Authority
- CN
- China
- Prior art keywords
- rock debris
- friction
- drill
- time
- torque
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000010801 machine learning Methods 0.000 title claims abstract description 28
- 239000011435 rock Substances 0.000 claims abstract description 141
- 238000005553 drilling Methods 0.000 claims abstract description 48
- 238000013508 migration Methods 0.000 claims abstract description 42
- 230000005012 migration Effects 0.000 claims abstract description 42
- 238000005457 optimization Methods 0.000 claims abstract description 29
- 238000009826 distribution Methods 0.000 claims abstract description 27
- 238000007405 data analysis Methods 0.000 claims abstract description 13
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 12
- 239000000126 substance Substances 0.000 claims description 18
- 238000005520 cutting process Methods 0.000 claims description 17
- 238000005259 measurement Methods 0.000 claims description 17
- 230000009286 beneficial effect Effects 0.000 claims description 10
- 238000004140 cleaning Methods 0.000 claims description 10
- 239000012530 fluid Substances 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 7
- 239000012634 fragment Substances 0.000 claims description 6
- 230000005484 gravity Effects 0.000 claims description 6
- 230000001133 acceleration Effects 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims description 3
- 238000009987 spinning Methods 0.000 claims 1
- 239000003814 drug Substances 0.000 description 6
- 229940079593 drug Drugs 0.000 description 6
- 238000012544 monitoring process Methods 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000007619 statistical method Methods 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 3
- 229910003460 diamond Inorganic materials 0.000 description 2
- 239000010432 diamond Substances 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012567 pattern recognition method Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Abstract
The invention discloses a drill sticking early warning method and system based on machine learning, belonging to the technical field of drill sticking early warning, wherein the method comprises the following steps: obtaining rock debris data distributed along a well track, and constructing a real-time rock debris migration model; obtaining real-time dynamic distribution of the rock debris bed in the well hole based on the real-time rock debris migration model; constructing a friction and torque balance model when the drill rod rotates in the borehole; obtaining a friction torque value on a drill rod based on real-time dynamic distribution of a rock debris bed in a well hole and a friction torque balance model; optimizing a friction torque balance model by using a Bayesian optimization algorithm and a Nash efficiency coefficient based on the friction torque value on the drill rod; carrying out real-time drill sticking early warning by utilizing a time sequence data analysis method based on the optimized friction resistance torque balance model; the method solves the problem that the drilling sticking risk is difficult to accurately and quickly predict in real time under different working conditions.
Description
Technical Field
The invention belongs to the technical field of drill sticking early warning, and particularly relates to a drill sticking early warning method and system based on machine learning.
Background
The time loss of the drilling process is mostly caused by the blockage of the drilling tool; if the drill jamming is not found in time in the drilling process, the drill stopping treatment is often needed, a large amount of time and cost are needed for treatment, and more serious secondary hazards are possibly caused.
The method mainly comprises big data statistical analysis and well drilling model construction analysis, wherein the big data statistical analysis and well drilling model construction analysis are mainly used as the existing drill sticking early warning method, the big data statistical analysis and well drilling model construction analysis are mainly divided into a discriminant analysis method and a pattern recognition method, and on the premise that a large amount of effective data are collected, the model is established to judge whether the drill sticking occurs or not by analyzing the occurrence reasons of the drill sticking, the operation conditions during the drill sticking and the like and utilizing methods such as neural network multivariate statistical analysis, fuzzy logic, analytic hierarchy process and the like; the drilling parameters can be abnormally changed when the sticking occurs, but the traditional physical model has large calculation amount and is difficult to completely reflect the change rule when the sticking occurs, the false alarm rate is high, the accuracy rate is low, the real-time performance is poor in practical application, and the occurring sticking is difficult to classify.
Disclosure of Invention
Aiming at the defects in the prior art, the drill sticking early warning method and the drill sticking early warning system based on machine learning provided by the invention can continuously and automatically adjust and adapt to real-time changing working conditions, and solve the problem that the drill sticking risk is difficult to accurately and quickly predict in real time under different working conditions.
In order to achieve the purpose of the invention, the invention adopts the technical scheme that:
the invention provides a drill sticking early warning method based on machine learning, which comprises the following steps:
s1, obtaining rock debris data distributed along a well track, and constructing a real-time rock debris migration model;
s2, obtaining real-time dynamic distribution of the rock debris bed in the borehole based on the real-time rock debris migration model;
s3, constructing a friction resistance torque balance model when the drill rod rotates in the borehole based on the space rectangular coordinate system, the Ferner coordinate system, the borehole track, the drill rod stress and the rock debris migration model;
s4, obtaining a friction torque value on the drill rod based on real-time dynamic distribution of the rock debris bed in the well and a friction torque balance model;
s5, optimizing a friction-resistance torque balance model by using a Bayes optimization algorithm and a Nash efficiency coefficient based on the friction-resistance torque value on the drill rod;
and S6, carrying out real-time drill sticking early warning by utilizing a time sequence data analysis method based on the optimized friction resistance torque balance model.
The invention has the beneficial effects that: the invention provides a stuck drill early warning method based on machine learning, which combines a real-time rock debris migration model, a friction resistance torque balance model and a machine model to realize stuck drill risk monitoring and early warning based on real-time drilling data, carries out real-time monitoring on drilling rock debris bed distribution and well cleaning through the real-time rock debris migration model, constructs a friction resistance torque balance model when a drill rod rotates in a well hole based on the real-time rock debris migration model, utilizes Bayesian optimization to train the friction resistance torque balance model to adapt to different working conditions, can continuously improve model precision even if the well condition or the well type is changed, has high model training speed, can complete training only through real-time logging data of the current well, and obtains a stuck drill risk index by analyzing and comparing real-time prediction results and actual measurement data of a mixed model and carrying out time sequence data analysis to realize the real-time early warning of the drilling stuck drill.
Further, the calculation expression of the real-time rock debris migration model in the step S1 is as follows:
wherein, the first and the second end of the pipe are connected with each other,tthe time is represented by the time of day,which means that the time is to be subjected to a partial derivation,to representThe cross-sectional area of the rock debris layer,representThe concentration of the cuttings in the cuttings layer,to representThe flow velocity of the rock debris in the rock debris layer,the partial derivative of the thickness of the rock debris layer is obtained,representing the rate of volume exchange of rock cuttings between different layers of rock cuttings,to representThe average density of the layer of rock debris,the pressure of the flow is indicated by the expression,gwhich represents the acceleration of the force of gravity,which represents an angle with the direction of gravity,to representThe shear stress of the rock debris layer is,representThe perimeter of the layer of rock debris,is represented byDirection of rock debris layerThe shear stress of the rock debris layer is,to representThe rock debris is layered onThe distance of the rock debris layer is greater than the distance of the rock debris layer,nthe total number of layers of the rock debris migration model is represented,represents the average density of the rock debris migration model,the average flow velocity of the rock debris migration model is represented.
The beneficial effect of adopting the further scheme is as follows: the real-time rock debris migration model is provided for monitoring the distribution of the drilling rock debris bed and the cleaning of the well hole in real time, and a foundation is provided for constructing a friction torque balance model when the drill rod rotates in the well hole.
Further, the calculation expression of the friction torque balance model in step S3 is as follows:
wherein the content of the first and second substances,the axial force of the drill rod is shown,sindicating the length of the drill pipe immersed in the drilling fluid,representing the weight of a unit of drill pipe in the drilling fluid,the angle of the borehole is shown,representing the unit tangential vector in the z-direction of the drill rod,kthe axial force coefficient of the drill rod is shown,representing a unit normal vector in the z-direction of the drill rod,which is indicative of the contact force,representing the angle between the normal of the contact plane and the contact force,the coefficient of friction resistance is expressed as,indicating the additional force on the drill pipe,representing the weight of a unit of drill pipe in the drilling fluid,represents a unit pair normal vector in the z direction of the drill rod,the axial torque required to rotate the drill rod is indicated,the radius of the drill rod is shown,representing the curvature.
The beneficial effect of adopting the above further scheme is that: and providing a calculation method of the friction torque balance model, and providing a basis for obtaining the friction torque value on the corresponding drill rod and realizing real-time drill sticking early warning.
Further, the step S4 includes the steps of:
s41, based on the real-time dynamic distribution and moment balance model of the detritus bed in the borehole, standardizing the height of the detritus bed and the change rate of the height of the detritus bed to the torque, and obtaining the torque measured values of the height of the standardized detritus bed and the height of the standardized detritus bed:
wherein, the first and the second end of the pipe are connected with each other,the normalized height of the rock debris bed is shown,the height of the rock debris bed is shown,Dthe diameter of the borehole is shown as,representing the rate of change of torque caused by the debris,representing the torque measurement at the normalized cuttings bed height,representing calculated torque values for the same wellbore configuration;
s42, constructing a model of the height of the rock fragment bed and a torque index based on the standard height of the rock fragment bed and a torque measured value under the standard height of the rock fragment bed:
wherein the content of the first and second substances,representing a parameter of the first exponential model,representing a second index model parameter;
s43, calculating to obtain the change rate of the friction coefficient based on the height of the rock debris bed and the torque index model:
wherein the content of the first and second substances,the change rate of the friction coefficient is shown,the coefficient of friction resistance at the standardized cuttings bed height is expressed,represents the coefficient of friction resistance without debris;
and S44, obtaining a friction torque value based on the friction coefficient change rate.
The beneficial effect of adopting the above further scheme is that: the method for obtaining the friction torque value on the drill rod based on the real-time dynamic distribution of the rock debris bed in the borehole and the friction torque balance model is provided, and a basis is provided for the optimization of the friction torque balance model.
Further, the step S5 includes the steps of:
s51, acquiring actually measured friction torque data in historical drilling data;
s52, inputting real-time dynamic distribution data of the detritus bed in the borehole in the historical drilling data into a friction torque balance model for simulation to obtain a corresponding friction torque value on the drill rod;
s53, constructing a Nash efficiency coefficient optimization function based on the friction torque measured data and the friction resistance torque value on the drill rod:
wherein the content of the first and second substances,the coefficient of the Nash efficiency is expressed,represents a time step ofiThe measured value of the frictional torque at the time of operation,represents a time step ofiThe friction torque value on the drill rod is measured,represents the average value of the friction torque value on the drill pipe,nrepresenting a total number of time steps;
s54, optimizing a friction-resistance torque balance model by using a Bayes optimization algorithm based on a Nash efficiency optimization function;
s55, judging whether the Nash efficiency coefficient is larger than a preset threshold value or not based on the optimization result, if so, entering a step S57, and otherwise, entering a step S56;
s56, updating historical drilling data, and returning to the step S51;
and S57, completing the optimization of the friction torque balance model, and entering the step S6.
The beneficial effect of adopting the above further scheme is that: and providing a friction torque value based on a drill rod, optimizing a friction torque balance model by using a Bayesian optimization algorithm and a Nash efficiency coefficient, and providing a basis for real-time drilling sticking early warning by using the optimized friction torque balance model.
Further, the step S6 includes the steps of:
s61, acquiring a real-time actual measurement friction resistance torque value;
s62, obtaining a predicted value of the friction resistance torque on the drill rod based on the optimized friction resistance torque balance model according to actually measured drilling data;
s63, calculating to obtain the relative deviation of the friction resistance torque by utilizing a time sequence data analysis method according to the real-time actual measurement friction resistance torque value and the predicted value of the friction resistance torque on the drill rod:
wherein, the first and the second end of the pipe are connected with each other,representing the relative deviation of the predicted value of the friction torque on the drill rod,representing a predicted value of the friction torque on the drill pipe considering the influence of the rock debris,representing a predicted value of friction torque on the drill pipe without considering the influence of rock debris,representing the relative deviation of the real-time actual measurement friction resistance torque value and the predicted value of the friction resistance torque on the drill rod considering the influence of rock debris,a measurement representing off-bottom free-wheeling torque;
s64, calculating a real-time stuck drill risk index based on the friction resistance torque relative deviation, and carrying out real-time stuck drill early warning based on the stuck drill risk index:
wherein the content of the first and second substances,indicating that insufficient cleaning of the cuttings results in a stuck drill risk index,indicating that inadequate cleaning of non-cuttings results in a stuck drill risk index,a weighting factor that represents the value of the relative deviation,a weighting factor representing the moving average deviation value,a first non-dimensional parameter is represented,a second non-dimensional parameter is represented,a third non-dimensional parameter is represented,representThe average deviation value of the moving average is calculated,representThe moving average deviation value of (2).
The beneficial effect of adopting the above further scheme is that: the method for carrying out real-time drilling sticking early warning by utilizing a time sequence data analysis method based on the optimized friction resistance torque balance model is provided, drilling sticking risk indexes are provided, whether drilling sticking is carried out due to insufficient rock debris cleaning or not is respectively predicated, the drilling sticking risk indexes are between 0 and 1, wherein 0 represents no drilling sticking risk, and 1 represents high-probability sticking of a drill rod.
The invention also provides a system of the drill sticking early warning method based on machine learning, which comprises the following steps:
the drilling rock debris migration module is used for acquiring rock debris data distributed along a borehole trajectory and constructing a real-time rock debris migration model;
the dynamic distribution data module is used for obtaining real-time dynamic distribution of the rock debris bed in the well hole based on the real-time rock debris migration model;
the force and moment balance module is used for constructing a friction and torque balance model when the drill rod rotates in the borehole based on a space rectangular coordinate system, a Ferner coordinate system, a borehole track, a drill rod stress and a rock debris migration model;
the friction torque data module is used for obtaining a friction torque value on the drill rod based on real-time dynamic distribution of the detritus bed in the well and a friction torque balance model;
the machine learning optimization module is used for optimizing a friction torque balance model by utilizing a Bayesian optimization algorithm and a Nash efficiency coefficient based on the friction torque value on the drill rod;
and the friction drill sticking early warning module is used for carrying out real-time drill sticking early warning by utilizing a time sequence data analysis method based on the optimized friction torque balance model.
The beneficial effects of the invention are as follows: the invention provides a system of a drilling sticking early warning method based on machine learning, which is a system correspondingly arranged on the drilling sticking early warning method based on machine learning and is used for realizing the drilling sticking early warning method based on machine learning.
Drawings
Fig. 1 is a flowchart illustrating steps of a method for warning of drill sticking based on machine learning according to an embodiment of the present invention.
FIG. 2 is a schematic diagram of a friction torque balance model of a drill pipe rotating in a borehole according to an embodiment of the present invention.
Fig. 3 is a block diagram of a system of a drill sticking early warning method based on machine learning according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention is provided to facilitate the understanding of the present invention by those skilled in the art, but it should be understood that the present invention is not limited to the scope of the embodiments, and it will be apparent to those skilled in the art that various changes may be made without departing from the spirit and scope of the invention as defined and defined by the appended claims, and all changes that can be made by the invention using the inventive concept are intended to be protected.
As shown in fig. 1, in an embodiment of the present invention, the present invention provides a stuck drill early warning method based on machine learning, including the following steps:
s1, obtaining rock debris data distributed along a well track, and constructing a real-time rock debris migration model;
the calculation expression of the real-time rock debris migration model in the step S1 is as follows:
wherein the content of the first and second substances,tthe time is represented by the time of day,it is shown that the partial derivative is calculated over time,to representThe cross-sectional area of the rock debris layer,to representThe concentration of the rock debris in the rock debris layer,representThe flow velocity of the rock debris in the rock debris layer,the partial derivative of the thickness of the rock debris layer is obtained,representing the volume exchange rate of rock debris between different rock debris layers,to representThe average density of the layer of rock debris,the pressure of the flow is indicated by the expression,gwhich represents the acceleration of the force of gravity,which represents an angle with the direction of gravity,to representThe shear stress of the rock debris layer is,representThe perimeter of the layer of rock debris,is represented byDirection of rock debris layerThe shear stress of the rock debris layer is,representThe rock debris is layered onThe distance between the layers of rock debris,nthe total number of layers of the rock debris migration model is represented,represents the average density of the rock debris migration model,representing the average flow velocity of the rock debris migration model;
s2, obtaining real-time dynamic distribution of the rock debris bed in the borehole based on the real-time rock debris migration model;
s3, constructing a friction resistance torque balance model when the drill rod rotates in the borehole based on the space rectangular coordinate system, the Ferner coordinate system, the borehole track, the drill rod stress and the rock debris migration model, as shown in FIG. 2;
the calculation expression of the friction torque balance model in the step S3 is as follows:
wherein the content of the first and second substances,the axial force of the drill rod is shown,sindicating the length of the drill pipe immersed in the drilling fluid,representing the weight of a unit of drill pipe in the drilling fluid,the angle of the well bore is shown,representing the unit tangential vector in the z-direction of the drill rod,kthe axial force coefficient of the drill rod is shown,representing a unit normal vector in the z-direction of the drill rod,which is indicative of the contact force,representing the angle between the normal of the contact plane and the contact force,the coefficient of friction resistance is expressed as,indicating the additional force on the drill pipe,representing the weight of a unit of drill pipe in the drilling fluid,represents a unit pair normal vector in the z direction of the drill rod,the axial torque required to rotate the drill rod is indicated,the radius of the drill rod is shown,represents a curvature;
the flener coordinate system, in which the borehole centerline is the reference line, the target itself is the origin, and the coordinate axes simplify the problem of trajectory prediction compared to the cartesian coordinate systemsThe direction being along the centre line, called the longitudinal direction, the axes of the coordinatelThe direction is normal to the center line through the origin, called lateral,sdirection andlthe directions are perpendicular to each other, so that the flener coordinate system uses the tangent vector of the central line and the normal vector to establish a coordinate system;
s4, obtaining a friction torque value on the drill rod based on real-time dynamic distribution of the debris bed in the well and a friction torque balance model;
the step S4 includes the steps of:
s41, based on the real-time dynamic distribution and moment balance model of the rock debris bed in the well, standardizing the height of the rock debris bed and the torque change rate of the height of the rock debris bed to obtain the torque measured values of the height of the standardized rock debris bed and the height of the standardized rock debris bed:
wherein, the first and the second end of the pipe are connected with each other,the normalized height of the rock debris bed is shown,the height of the rock debris bed is shown,Dthe diameter of the borehole is shown as,representing the rate of change of torque caused by the debris,representing the torque measurement at the normalized cuttings bed height,representing calculated torque values for the same wellbore configuration;
s42, constructing a model of the height of the rock debris bed and a torque index based on the height of the standardized rock debris bed and a torque measured value under the height of the standardized rock debris bed:
wherein the content of the first and second substances,representing the parameters of the first exponential model,representing a second index model parameter;
s43, calculating to obtain the change rate of the friction coefficient based on the height of the rock debris bed and the torque index model:
wherein the content of the first and second substances,the change rate of the friction coefficient is shown,representing the friction coefficient under the height of the standardized detritus bed,represents the coefficient of friction resistance without debris;
s44, obtaining a friction torque value based on the friction coefficient change rate;
s5, optimizing a friction-resistance torque balance model by using a Bayes optimization algorithm and a Nash efficiency coefficient based on the friction-resistance torque value on the drill rod;
the step S5 includes the steps of:
s51, acquiring actually measured friction torque data in historical drilling data;
s52, inputting real-time dynamic distribution data of the detritus bed in the borehole in the historical drilling data into a friction torque balance model for simulation to obtain a corresponding friction torque value on the drill rod;
s53, constructing a Nash efficiency coefficient optimization function based on the friction torque measured data and the friction resistance torque value on the drill rod:
wherein the content of the first and second substances,the coefficient of the nash efficiency is expressed,represents a time step ofiThe measured value of the frictional torque at the time of operation,represents a time step ofiThe friction torque value on the drill rod is measured,represents the average value of the friction torque value on the drill pipe,nrepresenting a total number of time steps;
s54, optimizing a friction-resistance torque balance model by using a Bayes optimization algorithm based on a Nash efficiency optimization function;
s55, judging whether the Nash efficiency coefficient is larger than a preset threshold value or not based on the optimization result, if so, entering a step S57, and if not, entering a step S56;
s56, updating historical drilling data, and returning to the step S51;
s57, completing the optimization of the friction resistance torque balance model, and entering the step S6;
s6, carrying out real-time drill sticking early warning by utilizing a time sequence data analysis method based on the optimized friction resistance torque balance model;
the step S6 includes the steps of:
s61, acquiring a real-time actual measurement friction resistance torque value;
s62, obtaining a predicted value of the friction resistance torque on the drill rod based on the optimized friction resistance torque balance model according to actually measured drilling data;
s63, calculating to obtain the relative deviation of the friction torque by utilizing a time sequence data analysis method according to the real-time actual measurement friction torque value and the predicted value of the friction torque on the drill rod:
wherein the content of the first and second substances,representing the relative deviation of the predicted value of the friction torque on the drill rod,representing a predicted value of the friction torque on the drill pipe considering the influence of the rock debris,representing a predicted value of friction torque on the drill pipe without considering the influence of rock debris,representing the relative deviation of the real-time actual measurement friction resistance torque value and the predicted value of the friction resistance torque on the drill rod considering the influence of rock debris,a measurement representing off-bottom free-wheeling torque;
s64, calculating a real-time stuck drill risk index based on the friction resistance torque relative deviation, and carrying out real-time stuck drill early warning based on the stuck drill risk index:
wherein the content of the first and second substances,indicating that insufficient cleaning of the cuttings results in a stuck drill risk index,indicating that inadequate cleaning of non-cuttings results in a stuck drill risk index,a weighting factor that represents the value of the relative deviation,indicating movementThe weighting factor for the average deviation value is,a first non-dimensional parameter is represented,a second non-dimensional parameter is represented,a third non-dimensional parameter is represented,to representThe average deviation value is moved,to representThe moving average deviation value of (a);
the stuck drill risk index is between 0 and 1, wherein 0 represents no stuck drill risk and 1 represents high probability of stuck drill pipe.
The invention provides a stuck drill early warning method based on machine learning, which combines a real-time rock debris migration model, a friction resistance torque balance model and a machine model to realize stuck drill risk monitoring and early warning based on real-time drilling data, carries out real-time monitoring on drilling rock debris bed distribution and well cleaning through the real-time rock debris migration model, constructs a friction resistance torque balance model when a drill rod rotates in a well hole based on the real-time rock debris migration model, utilizes Bayesian optimization to train the friction resistance torque balance model to adapt to different working conditions, can continuously improve model precision even if the well condition or the well type is changed, has high model training speed, can complete training only through real-time logging data of the current well, and obtains a stuck drill risk index by analyzing and comparing real-time prediction results and actual measurement data of a mixed model and carrying out time sequence data analysis to realize the real-time early warning of the drilling stuck drill.
Example 2
As shown in fig. 3, in this embodiment, the present invention further provides a system of a drill sticking early warning method based on machine learning, including:
the drilling rock debris migration module is used for acquiring rock debris data distributed along a borehole trajectory and constructing a real-time rock debris migration model;
the dynamic distribution data module is used for obtaining real-time dynamic distribution of the rock debris bed in the borehole based on the real-time rock debris migration model;
the force and moment balance module is used for constructing a friction resistance and torque balance model when the drill rod rotates in the well hole based on the space rectangular coordinate system, the Ferner coordinate system, the well track, the drill rod stress and the rock debris migration model;
the friction torque data module is used for obtaining a friction torque value on the drill rod based on real-time dynamic distribution of the detritus bed in the well and a friction torque balance model;
the machine learning optimization module is used for optimizing a friction torque balance model by utilizing a Bayes optimization algorithm and a Nash efficiency coefficient based on the friction torque value on the drill rod;
and the friction drag drill sticking early warning module is used for carrying out real-time drill sticking early warning by utilizing a time sequence data analysis method based on the optimized friction drag torque balance model.
The system of the drug relocation method based on the drug classification diagram neural network provided by the embodiment can execute the technical scheme shown in the drug relocation method based on the drug classification diagram neural network in the above method embodiment, and the implementation principle and the beneficial effect are similar, and the details are not repeated here.
In the embodiment of the present invention, the functional units may be divided according to a drug relocation method based on a drug classification map neural network, for example, each function may be divided into each functional unit, or two or more functions may be integrated into one processing unit. The integrated unit may be implemented in the form of hardware, or may be implemented in the form of a software functional unit. It should be noted that the division of the cells in the present invention is schematic, and is only a logical division, and there may be another division manner in actual implementation.
In the embodiment of the invention, in order to realize the principle and the beneficial effect of the machine learning-based diamond sticking early warning method, the system of the machine learning-based diamond sticking early warning method comprises a hardware structure and/or a software module which are corresponding to the execution of each function. It should be readily appreciated by those of ordinary skill in the art that while the exemplary elements and algorithm steps described in connection with the embodiments disclosed herein may be embodied in hardware and/or in a combination of hardware and computer software, whether such functionality is implemented as hardware or computer software, the functionality described may be implemented using different approaches for each particular application depending upon the particular application and design constraints imposed on the technology, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
Claims (7)
1. A stuck drill early warning method based on machine learning is characterized by comprising the following steps:
s1, obtaining rock debris data distributed along a well track, and constructing a real-time rock debris migration model;
s2, obtaining real-time dynamic distribution of the rock debris bed in the well hole based on the real-time rock debris migration model;
s3, constructing a friction resistance torque balance model when the drill rod rotates in the borehole based on the space rectangular coordinate system, the Ferner coordinate system, the borehole track, the drill rod stress and the rock debris migration model;
s4, obtaining a friction torque value on the drill rod based on real-time dynamic distribution of the debris bed in the well and a friction torque balance model;
s5, optimizing a friction-resistance torque balance model by using a Bayes optimization algorithm and a Nash efficiency coefficient based on the friction-resistance torque value on the drill rod;
and S6, carrying out real-time drill sticking early warning by utilizing a time sequence data analysis method based on the optimized friction resistance torque balance model.
2. The machine learning-based stuck drill early warning method according to claim 1, wherein the computational expression of the real-time rock debris migration model in the step S1 is as follows:
wherein the content of the first and second substances,tthe time is represented by a time-of-day,it is shown that the partial derivative is calculated over time,representThe cross-sectional area of the rock debris layer,to representThe concentration of the rock debris in the rock debris layer,to representThe flow velocity of the rock debris in the rock debris layer,the partial derivative of the thickness of the rock debris layer is obtained,representing the rate of volume exchange of rock cuttings between different layers of rock cuttings,to representThe average density of the layer of rock debris,the pressure of the flow is indicated by the expression,gwhich represents the acceleration of the force of gravity,which represents an angle with respect to the direction of gravity,to representThe shear stress of the rock debris layer is,representThe perimeter of the layer of rock debris,is represented byDirection of rock debris layerThe shear stress of the rock debris layer is,to representFormation of rock debrisThe distance of the rock debris layer is greater than the distance of the rock debris layer,nthe total number of layers of the rock debris migration model is represented,represents the average density of the rock debris migration model,the average flow velocity of the rock debris migration model is represented.
3. The machine learning-based stuck drill early warning method according to claim 1, wherein the calculation expression of the friction torque balance model in the step S3 is as follows:
wherein, the first and the second end of the pipe are connected with each other,the axial force of the drill rod is shown,sindicating the length of the drill pipe immersed in the drilling fluid,representing the weight of a unit of drill pipe in the drilling fluid,the angle of the well bore is shown,representing a unit tangential vector in the z-direction of the drill rod,kthe axial force coefficient of the drill rod is shown,representing a unit normal vector in the z-direction of the drill rod,which is indicative of the contact force,representing the angle between the normal of the contact plane and the contact force,the coefficient of friction resistance is expressed as,indicating the additional force on the drill pipe,representing the weight of a unit of drill pipe in the drilling fluid,represents a unit secondary normal vector in the z direction of the drill rod,the axial torque required to rotate the drill rod is indicated,the radius of the drill rod is shown,representing the curvature.
4. The machine learning-based stuck drill early warning method according to claim 1, wherein the step S4 comprises the steps of:
s41, based on the real-time dynamic distribution and moment balance model of the detritus bed in the borehole, standardizing the height of the detritus bed and the change rate of the height of the detritus bed to the torque, and obtaining the torque measured values of the height of the standardized detritus bed and the height of the standardized detritus bed:
wherein the content of the first and second substances,the standardized height of the rock debris bed is shown,the height of the rock debris bed is shown,Dthe diameter of the borehole is shown as,representing the rate of change of torque caused by the debris,representing the torque measurement at the normalized cuttings bed height,representing calculated torque values for the same wellbore configuration;
s42, constructing a model of the height of the rock fragment bed and a torque index based on the standard height of the rock fragment bed and a torque measured value under the standard height of the rock fragment bed:
wherein the content of the first and second substances,representing a parameter of the first exponential model,representing a second index model parameter;
s43, calculating to obtain the change rate of the friction coefficient based on the height of the rock debris bed and the torque index model:
wherein the content of the first and second substances,the change rate of the friction coefficient is expressed,the coefficient of friction resistance at the standardized cuttings bed height is expressed,represents the friction coefficient without rock debris;
and S44, obtaining a friction torque value based on the friction coefficient change rate.
5. The machine learning-based stuck drill early warning method according to claim 4, wherein the step S5 comprises the steps of:
s51, acquiring actually measured friction torque data in historical drilling data;
s52, inputting real-time dynamic distribution data of the detritus bed in the borehole in the historical drilling data into a friction torque balance model for simulation to obtain a corresponding friction torque value on the drill rod;
s53, constructing a Nash efficiency coefficient optimization function based on the friction torque measured data and the friction resistance torque value on the drill rod:
wherein the content of the first and second substances,the coefficient of the nash efficiency is expressed,represents a time step ofiThe measured value of the friction torque at the time of starting,represents a time step ofiThe friction torque value on the drill rod is measured,represents the average value of the friction torque value on the drill pipe,nrepresenting a total number of time steps;
s54, optimizing a friction torque balance model based on a Nash efficiency optimization function and beneficial to a Bayesian optimization algorithm;
s55, judging whether the Nash efficiency coefficient is larger than a preset threshold value or not based on the optimization result, if so, entering a step S57, and otherwise, entering a step S56;
s56, updating historical drilling data, and returning to the step S51;
and S57, completing the optimization of the friction resistance torque balance model, and entering the step S6.
6. The machine learning-based stuck drill early warning method according to claim 5, wherein the step S6 comprises the steps of:
s61, acquiring a real-time actual measurement friction resistance torque value;
s62, obtaining a predicted value of friction resistance torque on the drill rod based on the optimized friction resistance torque balance model according to the actually measured drilling data;
s63, calculating to obtain the relative deviation of the friction resistance torque by utilizing a time sequence data analysis method according to the real-time actual measurement friction resistance torque value and the predicted value of the friction resistance torque on the drill rod:
wherein, the first and the second end of the pipe are connected with each other,representing the relative deviation of the predicted value of the friction torque on the drill rod,representing a predicted value of the friction torque on the drill pipe considering the influence of the rock debris,representing a predicted value of friction torque on the drill pipe without considering the influence of rock debris,representing the relative deviation of the real-time actual measurement friction resistance torque value and the predicted value of the friction resistance torque on the drill rod considering the influence of rock debris,a measurement representing an off-bottom free-spinning torque;
s64, calculating a real-time stuck drill risk index based on the friction resistance torque relative deviation, and carrying out real-time stuck drill early warning based on the stuck drill risk index:
wherein the content of the first and second substances,indicating that insufficient cuttings cleaning results in a stuck drill risk index,indicating that inadequate cleaning of non-cuttings results in a stuck drill risk index,a weighting factor that represents the value of the relative deviation,a weighting factor representing the moving average deviation value,a first non-dimensional parameter is represented,a second non-dimensional parameter is represented,a third non-dimensional parameter is represented,representThe average deviation value of the moving average is calculated,representThe moving average deviation value of (2).
7. A system of the machine learning based stuck drill early warning method as claimed in any one of claims 1 to 6, comprising:
the drilling rock debris migration module is used for acquiring rock debris data distributed along a borehole track and constructing a real-time rock debris migration model;
the dynamic distribution data module is used for obtaining real-time dynamic distribution of the rock debris bed in the well hole based on the real-time rock debris migration model;
the force and moment balance module is used for constructing a friction resistance and torque balance model when the drill rod rotates in the well hole based on the space rectangular coordinate system, the Ferner coordinate system, the well track, the drill rod stress and the rock debris migration model;
the friction torque data module is used for obtaining a friction torque value on the drill rod based on real-time dynamic distribution of the detritus bed in the well and a friction torque balance model;
the machine learning optimization module is used for optimizing a friction torque balance model by utilizing a Bayes optimization algorithm and a Nash efficiency coefficient based on the friction torque value on the drill rod;
and the friction drag drill sticking early warning module is used for carrying out real-time drill sticking early warning by utilizing a time sequence data analysis method based on the optimized friction drag torque balance model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211670817.0A CN115640759A (en) | 2022-12-26 | 2022-12-26 | Drill jamming early warning method and system based on machine learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211670817.0A CN115640759A (en) | 2022-12-26 | 2022-12-26 | Drill jamming early warning method and system based on machine learning |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115640759A true CN115640759A (en) | 2023-01-24 |
Family
ID=84949918
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211670817.0A Pending CN115640759A (en) | 2022-12-26 | 2022-12-26 | Drill jamming early warning method and system based on machine learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115640759A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116362143A (en) * | 2023-06-02 | 2023-06-30 | 中国石油天然气集团有限公司 | Drill string friction analysis method and device |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104153730A (en) * | 2014-07-23 | 2014-11-19 | 中国石油大学(华东) | Drilling tool for clearing borehole |
CN105952437A (en) * | 2016-06-20 | 2016-09-21 | 中国石油大学(华东) | Indoor experimental research apparatus for friction and torque of strings in three-dimensional curved boreholes |
CN115203877A (en) * | 2021-04-08 | 2022-10-18 | 中国石油化工股份有限公司 | Closed-loop drilling optimization system and method for simulating drilling state in real time |
-
2022
- 2022-12-26 CN CN202211670817.0A patent/CN115640759A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104153730A (en) * | 2014-07-23 | 2014-11-19 | 中国石油大学(华东) | Drilling tool for clearing borehole |
CN105952437A (en) * | 2016-06-20 | 2016-09-21 | 中国石油大学(华东) | Indoor experimental research apparatus for friction and torque of strings in three-dimensional curved boreholes |
CN115203877A (en) * | 2021-04-08 | 2022-10-18 | 中国石油化工股份有限公司 | Closed-loop drilling optimization system and method for simulating drilling state in real time |
Non-Patent Citations (1)
Title |
---|
李紫璇 等: "钻井模型与机器学习耦合的实时卡钻预警技术" * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116362143A (en) * | 2023-06-02 | 2023-06-30 | 中国石油天然气集团有限公司 | Drill string friction analysis method and device |
CN116362143B (en) * | 2023-06-02 | 2023-08-22 | 中国石油天然气集团有限公司 | Drill string friction analysis method and device |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109508827B (en) | Drilling accident early warning method based on time recursion neural network | |
US20210364482A1 (en) | Prediction method of part surface roughness and tool wear based on multi-task learning | |
CN105089620B (en) | Monitoring system, the method and device of bit freezing | |
CN110110740B (en) | Drilling process working condition identification method based on multi-time scale features and neural network | |
Leng et al. | A hybrid data mining method for tunnel engineering based on real-time monitoring data from tunnel boring machines | |
CN112529341A (en) | Drilling well leakage probability prediction method based on naive Bayesian algorithm | |
CN115640759A (en) | Drill jamming early warning method and system based on machine learning | |
CN107292467A (en) | A kind of drilling risk Forecasting Methodology | |
Murillo et al. | Pipe sticking prediction and avoidance using adaptive fuzzy logic and neural network modeling | |
Li et al. | Incipient fault detection for geological drilling processes using multivariate generalized Gaussian distributions and Kullback–Leibler divergence | |
CN113431496A (en) | Drilling and production ship cooperative operation fault diagnosis and decision-making assisting method | |
CN113062731A (en) | Intelligent identification method for complex working conditions under drilling well | |
CN107292754A (en) | A kind of drilling risk forecasting system | |
CN115329657A (en) | Drilling parameter optimization method and device | |
CN115234220A (en) | Method and device for identifying underground stick-slip vibration in real time by using intelligent drill bit | |
CN115841247A (en) | Digital drilling risk monitoring method and device | |
CN111434886B (en) | Mechanical drilling speed calculation method and device for drilling process | |
CN116066062A (en) | Drilling sticking real-time early warning method based on parameter change trend abnormity diagnosis | |
Qiu et al. | TBM tunnel surrounding rock classification method and real-time identification model based on tunneling performance | |
CN115688956A (en) | While-drilling blocking early warning method and system based on intelligent learning of mobile sliding window | |
CN113685166B (en) | Drilling accident early warning method and system | |
US20230111036A1 (en) | Integrated drilling dysfunction prediction | |
CN111625916A (en) | Method and system for calculating stability value of well wall | |
CN111598366A (en) | Real-time drilling auxiliary decision-making method and system | |
CN115374827A (en) | PCA-LSTM-based stick-slip vibration level evaluation method and device |
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 | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20230124 |
|
RJ01 | Rejection of invention patent application after publication |