CN117473848A - Drilling learning curve construction method, system, equipment and storage medium - Google Patents

Abstract

The invention provides a method, a system, equipment and a storage medium for constructing a drilling learning curve, wherein the method comprises the following steps: acquiring historical drilling data of a reference well; performing similarity calculation on historical drilling data of the reference drilling well and drilling related design data of the target drilling well, and determining a fastest drilling scheme; carrying out fusion processing based on the related fault information and the fastest drilling scheme to construct a drilling learning curve; the corresponding fastest drilling scheme is determined through historical drilling data, and then the fastest drilling scheme and relevant fault information are fused and considered to construct an optimal learning curve, so that the scientificity and the accuracy of learning curve construction are improved, and intelligent optimization of the drilling process scheme is realized.

Description

Drilling learning curve construction method, system, equipment and storage medium

Technical Field

The invention belongs to the field of petroleum drilling engineering, and particularly relates to a drilling learning curve construction method, a system, equipment and a storage medium.

Background

The learning curve is also called experience curve, reflects the production rule that the technical cost and the production efficiency are reduced along with the expansion of the technical scale by the measures such as experience accumulation, adjustment, optimization and the like, and is applied to a plurality of industries. In the drilling construction process, technological parameter selection is carried out by different methods for different geological conditions and actual conditions, a construction scheme suitable for the technological parameter selection is designed, a learning curve is established and used for management, the efficient and optimal drilling process can be realized, the drilling period is optimized, and the aims of reducing cost and enhancing efficiency are achieved. At present, the well drilling design or construction mainly depends on the history of adjacent wells or the construction experience of areas to be summarized, and the scientificity and the accuracy are poor.

Disclosure of Invention

The invention aims to provide a method, a system, equipment and a storage medium for constructing a well drilling learning curve, which solve the technical problems, and the corresponding fastest well drilling scheme is determined through historical well drilling data, and then the fastest well drilling scheme and relevant fault information are fused and considered to construct an optimal learning curve, so that the scientificity and the accuracy of the construction of the learning curve are improved, and the intelligent optimization of a well drilling process scheme is realized.

In a first aspect of the present invention, a method for constructing a learning curve of a well is provided, the method comprising:

acquiring historical drilling data of a reference well, wherein the historical drilling data comprises: recording relevant fault information of the reference drilling well in the exploitation stage;

performing similarity calculation on the historical drilling data of the reference drilling and drilling related design data of the target drilling, and determining a fastest drilling scheme;

and carrying out fusion processing based on the related fault information and the fastest drilling scheme to construct a drilling learning curve, wherein the drilling learning curve is used for forming a drilling scheme corresponding to the target drilling.

Optionally, obtaining historical drilling data for the reference well includes:

Selecting a reference well drilling which meets the well condition of the target well drilling and the position relation of the reference well drilling and the target well drilling meets the preset requirement;

historical drilling data for each of the reference wells at each production stage is extracted.

Optionally, before the similarity calculation between the historical drilling data of the reference well and the drilling related design data of the target well, and after the obtaining of the historical drilling data of the reference well, the method further comprises:

and carrying out noise reduction and impurity removal treatment on the historical drilling data of each reference drilling in each exploitation stage.

Optionally, the drilling-related design data of the target well includes: drilling data of the target drilling at each production stage during the design process;

the method for determining the fastest drilling scheme includes the steps of performing similarity calculation on historical drilling data of reference drilling and drilling related design data of target drilling, and determining the fastest drilling scheme, wherein the method comprises the following steps:

respectively carrying out multi-dimensional similarity calculation on the drilling data of the target drilling in each exploitation stage and the historical drilling data of each reference drilling in the corresponding exploitation stage to obtain the fastest drilling scheme in each exploitation stage; wherein the dimensions include one or more of: wellbore trajectory similarity, lithology similarity, formation pressure, and fracture pressure similarity.

Optionally, the historical drilling data further includes: well condition information, rock stratum information, formation pressure and vertical depth information of each reference well;

and respectively carrying out multidimensional similarity calculation on the drilling data of the target drilling in each production stage and the historical drilling data of the reference drilling in the corresponding production stage, wherein the multidimensional similarity calculation comprises the following steps:

respectively inputting the well condition information of each reference well to an LSTM (least squares) so as to finish the calculation of the similarity of the well track of each reference well in each production stage;

and/or based on the rock stratum information of each reference well, completing lithology similarity calculation of each reference well in each exploitation stage by editing a distance and sequence matching algorithm;

and/or completing the similarity of the formation pressure and the fracture pressure of each reference well in each production stage through a Euclidean distance algorithm based on the formation pressure and the sag information of each reference well.

Optionally, the related fault information includes: production fault information for each of said reference wells at each production stage;

the fusion processing is performed based on the related fault information and the fastest drilling scheme to construct a drilling learning curve, which comprises the following steps:

Carrying out fusion calculation on the exploitation fault information of each reference well in each exploitation stage and the fastest well drilling scheme in the corresponding exploitation stage to obtain an optimal learning curve in each exploitation stage;

and according to the time sequence of each exploitation stage, combining the optimal learning curve corresponding to each exploitation stage, and constructing a drilling learning curve corresponding to the target drilling.

Optionally, the fusion calculation is performed on the production fault information of each reference drilling at each production stage and the fastest drilling scheme at the corresponding production stage, so as to obtain an optimal learning curve in each production stage, including:

setting a first-level weight for the exploitation fault information of each reference well in each exploitation stage, and setting a second-level weight for the fastest well in each exploitation stage based on the ageing of each fastest well scheme;

and summing the primary weight and the secondary weight corresponding to each mining stage to select an optimal learning curve in each mining stage.

In a second aspect of the present invention, there is provided a drilling learning curve construction system, the system comprising:

a historical data acquisition module for acquiring historical drilling data of a reference drilling, wherein the historical drilling data comprises: recording relevant fault information of the reference drilling well in the exploitation stage;

The similarity calculation module is used for calculating the similarity between the historical drilling data of the reference drilling and the drilling related design data of the target drilling, and determining the fastest drilling scheme;

and the curve construction module is used for carrying out fusion processing based on the related fault information and the fastest drilling scheme to construct a drilling learning curve, wherein the drilling learning curve is used for forming a drilling scheme corresponding to the target drilling.

In a third aspect of the present invention, there is provided a computer apparatus comprising: one or more processors; a memory; one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more applications configured to: the well drilling learning curve construction method according to the above is performed.

In a fourth aspect of the present invention, a computer readable storage medium is provided, wherein the storage medium stores at least one instruction, at least one program, a set of codes, or a set of instructions, and the at least one instruction, the at least one program, the set of codes, or the set of instructions are loaded and executed by a processor to implement a well learning curve construction method as described above.

Compared with the prior art, the invention has the advantages that: the corresponding fastest drilling scheme is determined through historical drilling data, and then the fastest drilling scheme and relevant fault information are fused and considered to construct an optimal learning curve, so that the scientificity and the accuracy of learning curve construction are improved, and intelligent optimization of the drilling process scheme is realized.

The foregoing description is only an overview of the present invention, and is intended to be implemented in accordance with the teachings of the present invention in order that the same may be more clearly understood and to make the same and other objects, features and advantages of the present invention more readily apparent.

Drawings

Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to designate like parts throughout the figures.

In the drawings:

the invention will be described in more detail hereinafter on the basis of embodiments and with reference to the accompanying drawings.

FIG. 1 is a schematic flow chart of a method for constructing a learning curve for drilling according to a first embodiment of the present invention;

FIG. 2 is a schematic flow chart of a method for constructing a learning curve for drilling according to a second embodiment of the present invention;

FIG. 3 is a schematic illustration of a well learning curve according to a first embodiment of the present invention;

FIG. 4 is a schematic illustration of a preferred drilling scenario for three-way drilling in accordance with a second embodiment of the present invention;

FIG. 5 is a schematic illustration of wellbore trajectory similarity calculation in accordance with the present invention;

FIG. 6 is a schematic diagram of lithology similarity calculation according to the present invention;

FIG. 7 is a schematic representation of formation pressure and fracture pressure similarity calculations in accordance with the present invention;

fig. 8 is a schematic structural diagram of a drilling learning curve construction system according to a third embodiment of the present invention.

Detailed Description

Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

FIG. 1 is a schematic flow chart of a method for constructing a learning curve for drilling according to a first embodiment of the present invention; FIG. 2 is a schematic flow chart of a method for constructing a learning curve for drilling according to a second embodiment of the present invention; FIG. 3 is a schematic illustration of a well learning curve according to a first embodiment of the present invention; FIG. 4 is a schematic illustration of a preferred drilling scenario for three-way drilling in accordance with a second embodiment of the present invention; FIG. 5 is a schematic illustration of wellbore trajectory similarity calculation in accordance with the present invention; FIG. 6 is a schematic view of a lithology similarity calculation according to a second embodiment of the present invention; FIG. 7 is a schematic diagram of formation pressure and fracture pressure similarity calculation according to a second embodiment of the present invention; fig. 8 is a schematic structural diagram of a drilling learning curve construction system according to a third embodiment of the present invention.

According to a first embodiment of the present invention, there is provided a well drilling learning curve construction method, the method including: acquiring historical drilling data of a reference well, wherein the historical drilling data comprises: recording relevant fault information of the reference drilling well in the exploitation stage; performing similarity calculation on the historical drilling data of the reference drilling and drilling related design data of the target drilling, and determining a fastest drilling scheme; and carrying out fusion processing based on the related fault information and the fastest drilling scheme to construct a drilling learning curve, wherein the drilling learning curve is used for forming a drilling scheme corresponding to the target drilling.

In this way, the corresponding fastest drilling scheme is determined through historical drilling data, and then the fastest drilling scheme and relevant fault information are fused and considered to construct an optimal learning curve, so that the scientificity and the accuracy of learning curve construction are improved, and intelligent optimization of the drilling process scheme is realized.

Specifically, according to fig. 1, a first embodiment of the present invention provides a method for constructing a learning curve of drilling, which can be applied to land drilling or sea drilling, and can be drilling of metal ores, coal mines, petroleum, natural gas, or flammable ice, etc., the method comprising:

S110: acquiring historical drilling data of a reference well;

wherein the historical drilling data comprises: recording relevant fault information of the reference drilling well in the exploitation stage;

for example: relevant fault information, including but not limited to, information about the sag, type, grade, and detail, is extracted from daily reports and wells Shi Shuju (included in historical drilling data) that characterizes each reference well as experiencing a fault during the drilling operation.

S120: performing similarity calculation on the historical drilling data of the reference drilling and drilling related design data of the target drilling, and determining a fastest drilling scheme;

s130: and carrying out fusion processing based on the related fault information and the fastest drilling scheme to construct a drilling learning curve, wherein the drilling learning curve is used for forming a drilling scheme corresponding to the target drilling.

Specifically, in the above steps S110-S130, step S110 is performed first to obtain historical drilling data of a reference well, where the reference well may provide references for construction, engineering management, instrument application, or geological exploration related aspects for the target well. In another embodiment, the selection needs to be made in advance for the reference well.

In general, multiple wells (i.e., reference wells) may be deployed in the same area, and the wells, well types, destination layers, operation projects, etc. of these wells may be the same or similar, and the development periods of these wells should be similar according to the normal algorithm, but the development periods of these wells may be quite different in the actual drilling process.

Therefore, when step S110 is specifically performed, a set of adjacent wells may be formed by selecting wells with well conditions and positional relationships satisfying preset requirements as reference wells from the wells according to the relationships between the well conditions of the wells and the well conditions of the target well and the positional relationships between the wells and the target well, and then acquiring historical well data of each well (reference well) in the set of adjacent wells. The number of reference wells may be one or a plurality of reference wells, and in this embodiment, the number of reference wells is not limited. Wherein the historical drilling includes, but is not limited to, one or more of the following: logging real-time data, drill bit usage records, formation stratification, well logging data. For example, logging is a service process of observing, collecting, recording and analyzing information of returned materials of a shaft such as solid, liquid and gas in the process of drilling by using methods such as rock and mineral analysis, geophysics, geochemistry and the like, and can judge underground geology and oil-gas conditions, analyze and judge underground drilling engineering profiles such as underground stratum, structure and oil-gas conditions of a local well, oil field development scheme design and the like.

And after the historical drilling data of the reference drilling is obtained, carrying out similarity calculation on the historical drilling data of the reference drilling and drilling related design data of the target drilling, determining a fastest drilling scheme, and carrying out fusion processing on the basis of the related fault information and the fastest drilling scheme to construct a drilling learning curve, wherein the drilling learning curve is used for forming a drilling scheme corresponding to the target drilling.

In addition, in another embodiment, in order to ensure accuracy and save calculation resources, after the historical drilling data are obtained, noise reduction and impurity removal processing can be performed on the historical drilling data, specifically, cleaning processing is performed on null values, abnormal values, repeated values and the like on the historical drilling data, and of course, the cleaning processing includes performing Gaussian filtering noise reduction processing on real-time data (including logging real-time data) to obtain available clean data.

In addition, to further improve the accuracy of constructing the well learning curve, in another embodiment, the following settings are made:

in obtaining historical drilling data for each reference well, historical drilling data for each reference well at various production stages (including, but not limited to, open time and formation) is acquired. Wherein each reference well is sequentially drilled according to the plan for different times and/or sections during the production process. Such as: three reference wells are provided: reference well 1, reference well 2 and reference well 3, each of which is 4 times. Therefore, in this embodiment, it is necessary to obtain the historical drilling data corresponding to 1 to 4 times of the production reference well 1, obtain the historical drilling data corresponding to 1 to 4 times of the production reference well 2, and obtain the historical drilling data corresponding to 1 to 4 times of the production reference well 3.

And then, carrying out noise reduction and impurity removal treatment on the historical drilling data of each reference drilling at each production stage. Of course, the noise reduction and impurity removal process is not necessary, and the noise reduction and impurity removal method in this embodiment is not limited, and may be determined according to a specific scheme.

In addition, the drilling-related design data of the target drilling includes: during the design process, the target wells are drilled with well data at various production stages including, but not limited to, start-up and formation. Such as: the drilling-related design data of the target drilling is characterized by: and carrying out pre-estimated design on the related drilling data of the exploitation stage of the target drilling according to experimental data, drilling simulation or other means. Such as: the target well is not yet started to perform the well drilling work, so the following data are estimated in advance according to experimental data, well drilling simulation or other means, including but not limited to: parameters such as MQ516J drill bit, HJT537JK drill bit, weight on bit, rate of penetration, lithology, three pressures, weight on bit, rotational speed, rate of penetration, wellbore trajectory, drilling tool assembly, or drilling fluid performance are used; alternatively, the target well has completed a first time of the associated work, but has not started a second time and thereafter each time of the associated work, so the second time and thereafter each time of the associated work are estimated in advance based on experimental data, well simulation or other means, including but not limited to the following data: parameters such as MQ516J bit and HJT537JK bit, weight on bit, rate of penetration, lithology, three pressures, weight on bit, rotational speed, rate of penetration, wellbore trajectory, drilling tool assembly, or drilling fluid performance are used.

In the process of calculating the similarity between the historical drilling data of the reference drilling well and the drilling related design data of the target drilling well to determine the fastest drilling scheme, the multi-dimensional similarity calculation is required to be respectively performed on the drilling data of the target drilling well in each production stage and the historical drilling data of the reference drilling well in the corresponding production stage, namely: respectively carrying out multi-dimensional similarity calculation on historical drilling data of each exploitation stage of each reference drilling well, and also respectively carrying out multi-dimensional similarity calculation on drilling related design data of the target drilling well in each exploitation stage; such as: there are provided a target well and three reference wells: reference well 1, reference well 2 and reference well 3, each well needs 4 times. In this embodiment, the similarity calculation is performed on 1 to 4 times of the reference well 1 to obtain the respective similarity calculation results of the reference well 1 to 4 times, and the target well and the reference well 2 to 3 times are calculated in the same way to obtain the respective similarity calculation results of the target well 1 to 4 times, the respective similarity calculation results of the reference well 2 to 4 times, and the respective similarity calculation results of the reference well 3 to 1 to 4 times. Wherein the multi-dimensions include, but are not limited to, one or more of the following: wellbore trajectory similarity, lithology similarity, formation pressure, and fracture pressure similarity. Specifically, the well condition information of each reference well is input to a Long Short-Term Memory (LSTM) to complete the calculation of the wellbore trajectory similarity of each reference well at each production stage (as shown in fig. 5), where the well condition information includes, but is not limited to: well depth, well inclination, azimuth, and dogleg (rate of change of full angle); and/or based on the formation information of each reference well, completing lithology similarity calculation (shown in fig. 6) of each reference well at each production stage by editing a distance and sequence matching algorithm; and/or based on the formation pressure and the vertical depth information of each reference well, completing the formation pressure and fracture pressure similarity of each reference well in each production stage through a Euclidean distance algorithm (shown in figure 7). Wherein the historical drilling data further comprises: well condition information, formation pressure, and sag information for each of the reference wells at each production stage.

Such as: there are provided a target well and three reference wells: target well, reference well No. 1, reference well No. 2, reference well No. 3. According to the method, 4 times of opening are needed for the target well drilling and each reference well drilling, the 1-time well drilling data and the 1-3 times of reference well drilling history well drilling data are combined, similarity calculation is conducted on three dimensions of wellbore track similarity, lithology similarity, formation pressure, fracture pressure similarity and the like, so that respective similarity calculation results of the target well drilling and the 1-3 times of reference well drilling are obtained, each similarity calculation result corresponds to a 1-time well drilling scheme of the target well drilling and the 1-3 times of reference well drilling, if the 1-time similarity calculation result of the 1-time reference well drilling is the largest, the 1-time well drilling scheme of the 1-time reference well drilling is selected as the fastest well drilling scheme according to the size of the similarity calculation result; combining the 2-time drilling data of the target drilling well and the 2-time historical drilling data of the 1-3 reference drilling well, performing similarity calculation on three dimensions such as wellbore track similarity, lithology similarity, formation pressure and fracture pressure similarity to obtain respective similarity calculation results of the 2-time drilling well of the target drilling well and the 1-3 reference drilling well, wherein each similarity calculation result corresponds to a 2-time drilling scheme of the target drilling well and the 1-3 reference drilling well, if the 2-time similarity calculation result of the 2-time reference drilling well is the largest, selecting the 2-time drilling scheme of the 2-time reference drilling well as the fastest drilling scheme according to the size of the similarity calculation result; combining 3-time drilling data of the target drilling well and 3-time historical drilling data of the 1-3 reference drilling well, performing similarity calculation on three dimensions of wellbore track similarity, lithology similarity, formation pressure, fracture pressure similarity and the like to obtain respective similarity calculation results of the target drilling well and the 1-3 reference drilling well respectively at 3 times, wherein each similarity calculation result corresponds to a 3-time drilling scheme of the target drilling well and the 1-3 reference drilling well, if the 3-time similarity calculation result of the 3-time reference drilling well is the largest, selecting the 3-time drilling scheme of the 3-time reference drilling well as the fastest drilling scheme according to the size of the similarity calculation result; combining 4-time drilling data of the target drilling and 1-3 reference drilling at 4 times of historical drilling data, performing similarity calculation on three dimensions such as wellbore track similarity, lithology similarity, formation pressure and fracture pressure similarity to obtain respective similarity calculation results of the target drilling and 1-3 reference drilling at 4 times of 4 times, wherein each similarity calculation result corresponds to a drilling scheme of the target drilling and 1-3 reference drilling at 4 times, and if the similarity calculation result of the target drilling at 4 times is the largest, the drilling scheme of the target drilling at 4 times is selected as the fastest drilling scheme according to the size of the similarity calculation result; thus, each well (target well, reference well 1, reference well 2, reference well 3) has a corresponding fastest well plan at each production stage (e.g., 1, 2, 3, or 4).

Wherein, in combination with fig. 6, the calculating of the lithology similarity of each reference well based on the editing distance and the sequence matching algorithm comprises: for two character strings a, b, the lengths are |a|, |b|, respectively, and the editing distances are as follows:

wherein: i and j represent the subscripts of string a and string b, respectively, beginning with 1. For example, the sitten word is converted to sitting: ('sitten' and 'sitting' edit distance is 3).

sitten(k→s)

sittin(e→i)

sitting(→g)

Similarity=1-edit distance/mate. Max (str1. Length, str2. Length) =1-3/7=0.571.

Further, the related failure information includes: production fault information of each reference well at each production stage, the types of the production fault information being set to be plural; therefore, in the process of carrying out fusion processing based on the related fault information and the fastest drilling scheme in each exploitation stage to construct a drilling learning curve, fusion calculation is required to be carried out on the exploitation fault information of each reference drilling in each exploitation stage and the fastest drilling scheme in the corresponding exploitation stage, so as to obtain an optimal learning curve in each exploitation stage; and then according to the time sequence of each exploitation stage, combining the optimal learning curve corresponding to each exploitation stage to construct the well drilling learning curve corresponding to the target well drilling.

Such as: three reference wells, namely a reference well 1, a reference well 2 and a reference well 3 are arranged, and each reference well needs 4 times;

and collecting fault information of each 1-4 times of faults of the 1-3 reference drilling, and sorting the collected fault information to obtain mining fault information of each 1-4 times of faults of the 1-3 reference drilling. Moreover, the production fault information includes, but is not limited to, one or more of the following: stuck, broken, lost circulation, or overflow.

Under the condition that each reference well (reference well 1, reference well 2, reference well 3 and reference well 4) has a corresponding fastest well drilling scheme in each exploitation stage (such as 1, 2, 3 or 4), carrying out fusion calculation on exploitation fault information of the reference well 1 to 4 and the fastest well drilling scheme to obtain an 11 initial learning curve of the reference well 1 at 1, a 12 initial learning curve of the reference well 1 at 2, a 13 initial learning curve of the reference well 1 at 3 and a 14 initial learning curve of the reference well 1 at 4; similarly, a 21 initial learning curve of the reference well drilling number 2 at 1 time, a 22 initial learning curve of the reference well drilling number 2 at 2 time, a 23 initial learning curve of the reference well drilling number 2 at 3 time and a 24 initial learning curve of the reference well drilling number 2 at 4 time can be obtained; similarly, a 31 initial learning curve of the reference well 3 at 1 time, a 32 initial learning curve of the reference well 3 at 2 times, a 33 initial learning curve of the reference well 3 at 3 times, and a 34 initial learning curve of the reference well 3 at 4 times can be obtained. Then, selecting a 1 st optimal learning curve (51 learning curve) from the 1 st 11 th initial learning curve, the 21 st initial learning curve and the 31 st initial learning curve of the 1 st-3 th reference well drilling; 2 times of optimal learning curves (52 learning curves) are selected from 12 initial learning curves, 22 initial learning curves and 32 initial learning curves of 2 times of reference drilling in numbers 1-3; 3 times of optimal learning curves (52 learning curves) are selected from the 13 initial learning curves, the 23 initial learning curves and the 33 initial learning curves of the 3 times of reference drilling in the 1-3 times; selecting 4 optimal learning curves (54 learning curves) from the 14 initial learning curves, the 24 initial learning curves and the 34 initial learning curves of the 4 initial learning curves of the reference well drilling 1-3; namely: the 51 learning curve, the 52 learning curve, the 53 learning curve, i.e. the 54 learning curve correspond to the optimal learning curves of reference well 1 to 3 in 1, 2, 3 and 4 times respectively. And then, sequentially connecting the 51 learning curve, the 52 learning curve and the 53 learning curve, namely, the 54 learning curve into a complete learning curve according to the time sequence, wherein the complete learning curve is the drilling learning curve corresponding to the target drilling. The initial learning curve is as shown in dashed lines in fig. 3; the above-described 51 learning curve, 52 learning curve, 53 learning curve, i.e., 54 learning curve is similar to the solid line shown in fig. 3.

In addition, in the process of carrying out fusion calculation on the exploitation fault information of each reference well in each exploitation stage and the fastest well drilling scheme in the corresponding exploitation stage to obtain the optimal learning curve in each exploitation stage, the exploitation fault information of each reference well in each exploitation stage needs to be respectively provided with a first-level weight, and the fastest well drilling scheme in each exploitation stage is respectively provided with a second-level weight based on the ageing of each fastest well drilling scheme; and summing the primary weight and the secondary weight corresponding to each mining stage to select an optimal learning curve in each mining stage.

Examples are as follows: the first-order weight is used for characterization as follows: and (3) calibrating the weight of fault information of each mining stage according to the types and the severity level, such as: the first-level weight of the faults of drilling sticking, drilling breaking and the like caused by engineering reasons is changed to 0.2, the first-level weight of lost circulation is changed to 0.15, the first-level weight of overflow is changed to 0.1, and the first-level weights are accumulated successively; and according to the ageing of the fastest drilling scheme (namely, the completion time of each fastest drilling scheme), the secondary weight of the fastest drilling scheme of 1 time is changed into 10, and finally the primary weight and the secondary weight are integrated to select the optimal learning curve of each mining stage, so that the highest division of the optimal drilling construction scheme is realized according to the total of two factors of drilling speed and faults.

In addition, in another embodiment, in the process of respectively performing multidimensional similarity calculation on the drilling data of the target drilling in each production stage and the historical drilling data of each reference drilling in the corresponding production stage, and obtaining the fastest drilling scheme in each production stage according to the multidimensional similarity calculation, if the similarity calculation result of the target drilling and each reference drilling in each opening is obtained, selecting at least two drilling schemes corresponding to the similarity calculation result as the fastest drilling scheme corresponding to the production stage according to the size of the similarity calculation result corresponding to each production stage; therefore, in the process of carrying out fusion calculation on the exploitation fault information of each reference well in each exploitation stage and the fastest well drilling scheme in the corresponding exploitation stage to obtain the optimal learning curve in each exploitation stage, the exploitation fault information of each reference well in each exploitation stage needs to be respectively provided with a first-level weight, and the fastest well drilling scheme in each exploitation stage is respectively provided with a second-level weight based on the ageing of each fastest well drilling scheme; and summing the primary weight and the secondary weight corresponding to each mining stage to select an optimal learning curve in each mining stage.

Examples are as follows: the first-order weight is used for characterization as follows: calibrating the fault information of each exploitation stage according to the type and severity level, taking the reference well drilling of the No. 1 as an example: the first-level weight of the faults of drilling sticking, drilling breaking and the like caused by engineering reasons is changed to 0.2, the first-level weight of lost circulation is changed to 0.15, the first-level weight of overflow is changed to 0.1, and the first-level weights are accumulated successively; according to the time effect of the fastest drilling scheme (namely, the completion time of each fastest drilling scheme), carrying out hierarchical arrangement on the fastest drilling scheme of 1-3 reference drilling at 1 time, changing the second-level weight of the shortest failure (namely, the completion speed of the fastest drilling scheme) into 10, changing the second-level weight of the longest failure into 6, finally selecting the optimal learning curve of each mining stage by combining the first-level weight and the second-level weight, and then constructing the drilling learning curve corresponding to the target drilling according to the time sequence of each mining stage and combining the optimal learning curve corresponding to each mining stage; therefore, the method realizes the highest division into the optimal drilling construction schemes according to two factors of drilling speed and faults.

In this regard, in combination with the intelligent method, the fastest drilling scheme of the reference well at each production stage is formed by extracting historical well drilling data of the reference well at each production stage (including each time and stratum), and in combination with well drilling data of the target well at each production stage; then, an optimal learning curve of each exploitation stage is constructed by fusing the fastest well drilling scheme and exploitation fault information of each exploitation stage, and the optimal learning curve of the target well drilling can be obtained based on the optimal learning curve of each exploitation stage according to the time sequence of each exploitation stage, so that a well drilling scheme corresponding to the target well drilling can be formed; therefore, not only is manual intervention reduced, but also the fastest drilling scheme and fault factors are fused and considered, so that the scientificity and the accuracy of the construction of the optimal learning curve are improved, and the intelligent optimization of the drilling process scheme is realized.

According to a second embodiment of the present invention, as shown in fig. 2, there is provided a method for constructing a learning curve of a well, specifically, the method includes:

(1) Constructing an area adjacent well set according to well type, distance and well types in historical well drilling data of each reference well, wherein the historical well drilling data comprise logging real-time data, drill bit use records, stratum layering, well logging data and the like; of course, after a new well is drilled (target well production is complete), the new well is taken as a subset into the set of adjacent wells

(2) Performing problems such as null value, abnormal value and repeated value on historical drilling data, and performing Gaussian filtering noise reduction treatment on real-time data (including logging real-time data) to obtain available clean data;

(3) Extracting logging real-time data to extract the average value of engineering parameters such as drilling pressure, rotating speed, mechanical drilling speed and the like at each meter depth interval according to a well depth scale and a pure drilling working condition to form well drilling aging whole meter data;

moreover, the well structure information of each reference well is extracted to obtain the top depth and the bottom depth of each reference well in each exploitation stage (beginning and stratum), and parameters such as lithology, three pressures, weight on bit, rotating speed, mechanical drilling speed, drill bit use record, well track, drilling tool combination, drilling fluid performance and the like are extracted from historical well drilling data according to the depth interval of the beginning and stratum;

(4) And (3) carrying out similarity calculation with the historical well according to the well design data, namely: based on the drilling-related design data of the target drilling (the present well drilling design data shown in fig. 2), using the wellbore trajectory similarity, the lithology similarity, the formation pressure and the fracture pressure similarity as criteria, targeting the fastest rate of penetration, preferably determining the fastest drilling plan (the fastest drilling well set plan shown in fig. 2) of each reference drilling at each production stage (open time and formation); wherein the wellbore trajectory is similarly calculated using LSTM (as shown in fig. 5), and the input parameters include well depth, well inclination angle, azimuth angle, dog leg angle (full angle change rate), etc.; the lithology similarity is calculated by adopting an editing distance and sequence matching algorithm (shown in fig. 6); the similarity of the formation pressure and the fracture pressure is calculated by Euclidean distance (shown in figure 7);

wherein, in connection with fig. 6, according to the edit distance and the sequence matching algorithm, the step of completing the lithology similarity calculation includes: for two character strings a, b, the lengths are |a|, |b|, respectively, and the editing distances are as follows:

wherein: i and j represent the subscripts of string a and string b, respectively, beginning with 1. For example, the sitten word is converted to sitting (edit distance of 3 for sitten ' and ' sitting ').

sitten(k→s)

sittin(e→i)

sitting(→g)

Similarity=1-edit distance/mate. Max (str1. Length, str2. Length) =1-3/7=0.571.

(5) Extracting relevant information such as the sagging depth, the type, the grade, the detail and the like of the complex fault occurrence from daily reports and wells Shi Shuju (contained in historical drilling data), namely: extraction of production fault data for each reference well at each production stage from historical well data, including but not limited to: related information such as sagging depth, type, grade, detail and the like of the complex faults;

(6) Extracting a pure drilling period according to the well construction progress of the fastest drilling speed of each section, and constructing and drawing an optimal drilling learning curve of a virtual area, wherein the shortest time is selected from the area of each period of completion in each beginning;

(7) Fusing the fastest drilling speed scheme of each well section with complex faults (mining fault data), calibrating weights for the complex faults according to the occurrence types and severity levels, setting the faults such as drilling sticking, drilling breaking and the like to be 0.2, setting the lost circulation to be 0.15, setting the overflow to be 0.1, and accumulating gradually; and (3) grading the fastest drilling speed scheme, wherein the fastest score is 10, the slowest score is 6, and finally, integrating the two factors of drilling speed and complexity, and dividing the fastest drilling speed scheme into the optimal drilling construction scheme.

According to fig. 4, the technique is applied to northwest oilfield division, the drilling process is optimized through multi-well analysis and regional drilling learning curve construction, and the first, second, third and fourth openings of the technique are optimized in a segmented manner according to the geological condition and design data of northwest Nissan drilling forward direction of 53-2H, and according to a comprehensive evaluation model integrating drilling speed and complex faults, the first opening of the technique preferably adopts a SHB1-5H drilling scheme comprising an MQ519J drill bit; the two-way and three-way drilling schemes are preferably SHB1-1H, wherein the three-way drilling head comprises MQ516J, HJT537JK drilling heads and the like, and is matched with a 127mm screw drilling tool; the quarter prefers the SHB1-5H drilling scheme. And finally, recording the shortest drilling period of the three-open-hole section creating zone blocks for 21.48 days. The technology makes drilling experience tangible, improves engineering analysis decision and process optimization level, improves rationality of drilling design and construction scheme, optimizes technological measures such as drill bit, drilling fluid, power drilling tool and the like according to the characteristics of the starting time and stratum, and achieves targeted speed and efficiency improvement of regional drilling.

Combining an intelligent method, namely, extracting historical drilling data of the reference drilling in each production stage (comprising each time and stratum), and combining the drilling data of the target drilling in each production stage to form a fastest drilling scheme of the reference drilling in each production stage; then, constructing an optimal learning curve of each exploitation stage by fusing the fastest well drilling scheme and exploitation fault information of each exploitation stage, and obtaining the optimal learning curve of the target well drilling based on the optimal learning curve of each exploitation stage according to the time sequence of each exploitation stage, so that a well drilling scheme corresponding to the target well drilling can be formed; therefore, not only is manual intervention reduced, but also the fastest drilling scheme and fault factors are fused and considered, so that the scientificity and the accuracy of the construction of the optimal learning curve are improved, and the intelligent optimization of the drilling process scheme is realized.

According to a third embodiment of the present invention, shown in fig. 8, there is provided a well learning curve construction system, the system comprising:

a historical data acquisition module 210 for acquiring historical drilling data of the reference well;

wherein the historical drilling data comprises: recording relevant fault information of the reference drilling well in the exploitation stage;

a similarity calculation module 220, configured to perform similarity calculation on the historical drilling data of the reference drilling and drilling related design data of the target drilling, and determine a fastest drilling scheme of the reference drilling;

and a curve construction module 230, configured to perform fusion processing based on the related fault information and the fastest drilling scheme of the reference drilling, so as to construct a drilling learning curve, where the drilling learning curve is used to form a drilling scheme corresponding to the target drilling.

Optionally, the historical data obtaining module 210 is specifically configured to select a reference well drilling that meets the preset requirements with respect to the well condition of the target well drilling and the positional relationship with the target well drilling; historical drilling data for each of the reference wells at each production stage is extracted.

Optionally, the system further comprises: the noise reduction and impurity removal module is used for carrying out noise reduction and impurity removal treatment on the historical drilling data of each reference drilling at each exploitation stage before the similarity calculation is carried out on the historical drilling data of the reference drilling and the drilling related design data of the target drilling and after the historical drilling data of the reference drilling is acquired.

Optionally, the similarity calculation module 220 includes: the fusion calculation unit is used for combining the drilling data of the target drilling at each production stage with the historical drilling data of each reference drilling at the corresponding production stage, and respectively carrying out multidimensional similarity calculation on each reference drilling at each production stage to obtain the fastest drilling scheme of each drilling at each production stage;

wherein the dimensions include one or more of: wellbore trajectory similarity, lithology similarity, formation pressure, and fracture pressure similarity; the drilling-related design data for the target well includes: drilling data of the target drilling at each production stage during the design process;

optionally, the fusion calculation unit is specifically configured to input the well condition information of each reference well to the LSTM, so as to complete calculation of the wellbore trajectory similarity of each reference well in each production stage; and/or based on the rock stratum information of each reference well, completing lithology similarity calculation of each reference well in each exploitation stage by editing a distance and sequence matching algorithm; and/or based on the formation pressure and the sag information of each reference well in each production stage, completing the similarity of the formation pressure and the fracture pressure of each reference well through a Euclidean distance algorithm;

Wherein the historical drilling data further comprises: well condition information, rock stratum information, formation pressure and vertical depth information of each reference well;

optionally, the curve construction module 230 includes: the fusion calculation unit is used for carrying out fusion calculation on the exploitation fault information of each reference well in each exploitation stage and the fastest well drilling scheme of each reference well in the corresponding exploitation stage to obtain an optimal learning curve in each exploitation stage; and the curve construction unit is used for constructing a drilling learning curve corresponding to the target drilling according to the time sequence of each exploitation stage and combining the optimal learning curve corresponding to each exploitation stage.

Wherein the relevant fault information includes: production fault information for each of said reference wells at each production stage;

optionally, the fusion calculation unit is configured to set a first-level weight for the production fault information of each reference well in each production stage, and set a second-level weight for the fastest well plan of each reference well in each production stage based on aging of each fastest well plan; and summing the primary weight and the secondary weight corresponding to each mining stage to select an optimal learning curve in each mining stage.

In this regard, historical drilling data of the reference drilling at each production stage (including each time and stratum) is extracted through the historical data acquisition module 210, and then the fastest drilling scheme of the reference drilling at each production stage is formed by combining the drilling data of the target drilling at each production stage through the similarity calculation module 220; then, the curve construction module 230 is used for constructing an optimal learning curve of each exploitation stage by fusing the fastest drilling scheme and exploitation fault information of each exploitation stage, and the optimal learning curve of the target drilling can be obtained based on the optimal learning curve of each exploitation stage according to the time sequence of each exploitation stage, so that a drilling scheme corresponding to the target drilling can be formed; therefore, not only is manual intervention reduced, but also the fastest drilling scheme and fault factors are fused and considered, so that the scientificity and the accuracy of the construction of the optimal learning curve are improved, and the intelligent optimization of the drilling process scheme is realized.

A fourth embodiment of the present invention provides a computer apparatus including: one or more processors; a memory; one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more applications configured to: the well drilling learning curve construction method according to the above is performed.

The specific embodiment of the steps of the method can be referred to the first and second embodiments, and the description of this embodiment is not repeated here.

A fifth embodiment of the present invention provides a computer readable storage medium storing at least one instruction, at least one program, a set of codes, or a set of instructions, the at least one instruction, the at least one program, the set of codes, or the set of instructions being loaded and executed by a processor to implement a well learning curve construction method as described above.

The specific embodiment of the steps of the method can be referred to the first and second embodiments, and the description of this embodiment is not repeated here.

From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment. Based on such understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art in the form of a software product stored in a storage medium (e.g. ROM/RAM, magnetic disk, optical disk) comprising a number of instructions for causing a terminal (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the method according to the embodiments of the present invention.

The embodiments of the present invention have been described above with reference to the accompanying drawings, but the present invention is not limited to the above-described embodiments, which are merely illustrative and not restrictive, and many forms may be made by those having ordinary skill in the art without departing from the spirit of the present invention and the scope of the claims, which are to be protected by the present invention.

Claims (10)

1. A method of drilling a well learning curve, the method comprising:

acquiring historical drilling data of a reference well, wherein the historical drilling data comprises: recording relevant fault information of the reference drilling well in the exploitation stage;

performing similarity calculation on the historical drilling data of the reference drilling and drilling related design data of the target drilling, and determining a fastest drilling scheme;

and carrying out fusion processing based on the related fault information and the fastest drilling scheme to construct a drilling learning curve, wherein the drilling learning curve is used for forming a drilling scheme corresponding to the target drilling.

2. The method of claim 1, wherein obtaining historical drilling data for the reference well comprises:

Selecting a reference well drilling which meets the well condition of the target well drilling and the position relation of the reference well drilling and the target well drilling meets the preset requirement;

historical drilling data for each of the reference wells at each production stage is extracted.

3. The method of claim 2, wherein prior to the similarity calculation of the historical drilling data of the reference well with the drilling-related design data of the target well, and after the obtaining of the historical drilling data of the reference well, the method further comprises:

and carrying out noise reduction and impurity removal treatment on the historical drilling data of each reference drilling in each exploitation stage.

4. A method according to claim 2 or 3, wherein the drilling-related design data of the target well comprises: drilling data of the target drilling at each production stage during the design process;

the method for determining the fastest drilling scheme includes the steps of performing similarity calculation on historical drilling data of reference drilling and drilling related design data of target drilling, and determining the fastest drilling scheme, wherein the method comprises the following steps:

respectively carrying out multi-dimensional similarity calculation on the drilling data of the target drilling in each exploitation stage and the historical drilling data of each reference drilling in the corresponding exploitation stage to obtain the fastest drilling scheme in each exploitation stage; wherein the dimensions include one or more of: wellbore trajectory similarity, lithology similarity, formation pressure, and fracture pressure similarity.

5. The method of claim 4, wherein the historical drilling data further comprises: well condition information, rock stratum information, stratum pressure and vertical depth information of each reference well drilling at each exploitation stage;

and respectively carrying out multidimensional similarity calculation on the drilling data of the target drilling in each production stage and the historical drilling data of the reference drilling in the corresponding production stage, wherein the multidimensional similarity calculation comprises the following steps:

respectively inputting the well condition information of each reference well to an LSTM (least squares) so as to finish the calculation of the similarity of the well track of each reference well in each production stage;

and/or based on the rock stratum information of each reference well, completing lithology similarity calculation of each reference well in each exploitation stage by editing a distance and sequence matching algorithm;

and/or completing the similarity of the formation pressure and the fracture pressure of each reference well in each production stage through a Euclidean distance algorithm based on the formation pressure and the sag information of each reference well.

6. The method of claim 4, wherein the related fault information comprises: production fault information for each of said reference wells at each production stage;

The fusion process is performed on the basis of the related fault information and the fastest drilling scheme in each exploitation stage, so as to construct a drilling learning curve, which comprises the following steps:

carrying out fusion calculation on the exploitation fault information of each reference well in each exploitation stage and the fastest well drilling scheme in the corresponding exploitation stage to obtain an optimal learning curve in each exploitation stage;

and according to the time sequence of each exploitation stage, combining the optimal learning curve corresponding to each exploitation stage, and constructing a drilling learning curve corresponding to the target drilling.

7. The method of claim 6, wherein the merging of the production fault information of each reference well at each production stage with the fastest well plan at the corresponding production stage to obtain the optimal learning curve at each production stage comprises:

setting a first-level weight for the exploitation fault information of each reference well in each exploitation stage, and setting a second-level weight for the fastest well in each exploitation stage based on the ageing of the fastest well scheme;

and summing the primary weight and the secondary weight corresponding to each mining stage to select an optimal learning curve in each mining stage.

8. A well drilling learning curve construction system, the system comprising:

a historical data acquisition module for acquiring historical drilling data of a reference drilling, wherein the historical drilling data comprises: recording relevant fault information of the reference drilling well in the exploitation stage;

the similarity calculation module is used for calculating the similarity between the historical drilling data of the reference drilling and the drilling related design data of the target drilling, and determining the fastest drilling scheme;

and the curve construction module is used for carrying out fusion processing based on the related fault information and the fastest drilling scheme to construct a drilling learning curve, wherein the drilling learning curve is used for forming a drilling scheme corresponding to the target drilling.

9. A computer device, comprising: one or more processors; a memory; one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the one or more processors, the one or more applications configured to: a well drilling learning curve construction method according to any one of claims 1 to 7 is performed.

10. A computer readable storage medium having stored thereon at least one instruction, at least one program, code set, or instruction set, the at least one instruction, the at least one program, the code set, or instruction set being loaded and executed by a processor to implement the well learning curve construction method of any of claims 1 to 7.