CN116955540B

CN116955540B - Recommendation method and device for drilling accident handling scheme

Info

Publication number: CN116955540B
Application number: CN202311204126.6A
Authority: CN
Inventors: 赵艳红; 丁建新; 温欣; 李雪松; 王海涛; 王建华; 王玉芬; 陈星燃; 徐澎; 胡宏涛
Original assignee: Kunlun Digital Technology Co ltd; China National Petroleum Corp
Current assignee: Kunlun Digital Technology Co ltd; China National Petroleum Corp
Priority date: 2023-09-19
Filing date: 2023-09-19
Publication date: 2024-02-09
Anticipated expiration: 2043-09-19
Also published as: CN116955540A

Abstract

A method and apparatus for recommending a drilling event handling scenario, the method comprising: acquiring attributes of one or more drilling incidents to be treated; obtaining a treatment method recommendation result of the drilling accident to be treated according to the attribute of one or more drilling accidents to be treated and the recommendation model of the treatment scheme; wherein, the recommended model of the treatment scheme is established according to the drilling accident ontology; the drilling accident ontology is established according to the related theoretical knowledge of the drilling accident and the historical treatment cases of the drilling accident.

Description

Recommendation method and device for drilling accident handling scheme

Technical Field

The present disclosure relates to the field of information technology and the field of petroleum industry, and more particularly, to a method and apparatus for recommending a drilling accident treatment scheme.

Background

Drilling is a concealed underground project, and has a great deal of ambiguity, randomness and uncertainty, thus being a real high-risk operation. In the drilling operation, the drilling complexity and accidents caused by the factors of geology, engineering, management, personnel and the like account for about 3-8% of the total drilling time, and huge economic loss is caused. At present, accident handling at the drilling site is still based on expert experience. An experienced drilling engineer can timely identify accident signs, accurately judge the accident types and take effective treatment measures. On the contrary, the accident is not accurately researched and judged, the accident treatment is not timely, the treatment method is not reasonable, and the like, so that the best moment for the accident treatment is missed, the accident treatment time is prolonged, a large amount of manpower and material resources are consumed if the accident treatment time is light, the drilling construction period is prolonged, and the whole well is abandoned if the accident treatment time is heavy. Therefore, a recommendation system for a complex knowledge body management and disposal scheme of the drilling accident is established, and the recommendation system has important functions of guaranteeing safe drilling, improving drilling efficiency and saving drilling cost.

Disclosure of Invention

The application provides a recommendation method and a recommendation device for a drilling accident treatment scheme, wherein the method comprises the steps of establishing a drilling accident knowledge body; and a recommended model of a disposal scheme is established according to the drilling accident knowledge body, the emergency disposal capability of a site engineer can be improved by utilizing the recommended model of the disposal scheme, the accident disposal time is shortened, the accident disposal efficiency is improved, and the disposal cost after the accident occurs is saved.

In a first aspect, the present application provides a method of recommending a drilling accident handling scenario, the method comprising:

acquiring attributes of one or more drilling incidents to be treated;

obtaining a recommended result of the treatment scheme of the drilling accident to be treated according to the attribute of one or more drilling accidents to be treated and by using a recommended model of the treatment scheme;

the recommendation model of the treatment scheme is established according to the drilling accident knowledge body; the drilling accident ontology is established according to the related theoretical knowledge of the drilling accident and the historical treatment cases of the drilling accident.

In an exemplary embodiment, the drilling accident ontology includes: a theoretical ontology and a treatment case ontology;

The theoretical knowledge ontology comprises one or more of the following: an accident complex type knowledge body, an accident phenomenon knowledge body, an influence factor knowledge body and a disposal method knowledge body;

the treatment case-like ontology includes one or more of: block ontology, well ontology, formation ontology, and historical disposal case ontology.

In an exemplary embodiment, the incident complex type ontology includes one or more of the following attributes: accident complexity cause attribute, accident complexity phenomenon attribute, precaution measure attribute and disposal method attribute;

the accident phenomenon knowledge body comprises one or more of the following attributes: phenomenon description attributes, possible cause attributes, inspection item attributes, and measure attributes;

the influence factor ontology comprises one or more of the following properties: factor description attributes, reason description attributes, possible complex type attributes, possible incident type attributes;

the treatment method ontology includes one or more of the following attributes: method definition attributes, principle description attributes, implementation step attributes, advantage attributes, disadvantage attributes, application scenario attributes, and notes attributes;

The block ontology includes one or more of the following attributes: block name attributes, geologic structure attributes, and geographic environment attributes;

the well ontology includes one or more of the following attributes: well number attribute, well type attribute and coordinate attribute;

the stratum ontology comprises one or more of the following attributes: stratum name attribute, stratum top depth attribute, stratum bottom depth attribute and lithology attribute;

the historical treatment case ontology includes one or more of the following attributes: date of occurrence attribute, name of incident attribute, base case attribute, elapsed time of incident attribute, process of incident attribute, analysis of cause of incident attribute, summary and experience attribute.

In an exemplary embodiment, the recommended model building process of the treatment scheme includes:

labeling each attribute in the drilling accident knowledge body, and establishing a drilling accident label system;

generating a theoretical knowledge feature matrix based on theoretical knowledge labels in the drilling accident label system;

generating a treatment case feature matrix based on historical treatment case tags in the drilling incident tag system;

respectively setting corresponding weights for the theoretical knowledge feature matrix and the treatment case feature matrix, and combining to obtain a feature matrix;

And constructing a recommendation model of the treatment scheme according to the feature matrix.

In an exemplary embodiment, the theoretical knowledge tag in the drilling event tag system comprises: accident complex type tags, accident phenomenon tags, influencing factor tags and disposal method tags;

wherein the accident-complex-type tag includes one or more of: drilling complex type tags and drilling accident type tags;

the complex drilling label comprises one or more of the following: lost circulation, well invasion, overflow, balling, water hole drop, water hole blockage and well wall collapse;

the drilling accident type tag comprises one or more of the following: drilling sticking accidents, logging accidents, well cementation accidents, well control accidents, drilling tool breaking and down-hole junk;

the incident label includes one or more of the following: stuck point position, drilling tool activity condition, engineering parameter change, wellhead display condition, gas measurement abnormality and drilling fluid property change;

the influence factor tags include one or more of the following: geological factors, engineering factors, management and human factors;

the treatment method tag includes one or more of the following: a well killing method, a lost circulation treatment method, a card measuring method, a card releasing method and a down-hole junk treatment method;

In an exemplary embodiment, the generating a theoretical knowledge feature matrix based on theoretical knowledge tags in the drilling accident tag system includes:

taking the number of the accident complex types as the number of matrix rows, taking the total number of labels of the theoretical knowledge labels as the number of matrix columns, and establishing a theoretical knowledge feature matrix M _theory ；

Wherein the theoretical knowledge feature matrix M _theory Element M of (3) _i-theory，j I-they represent accident complex type numbers, j represents tag numbers; if the j-th tag exists for the i-th incident complex type, M _i-theory，j =1; if the j-th tag does not exist in the i-th accident complex type, M _i-theory，j =0。

In an exemplary embodiment, the generating a treatment case feature matrix based on historical treatment case tags in the drilling incident tag system includes:

taking the number of the historical treatment cases as the number of matrix rows, taking the total number of labels of the theoretical knowledge labels as the number of matrix columns, and establishing a characteristic matrix M of the historical treatment cases _case ；

Wherein M is _i-case，j Representing the historic treatment case feature matrix M _case I-case represents a treatment case number, j represents a label number; if the j-th label exists for the i-case treatment case, M _i-case，j =1; if the j-th tag does not exist in the i-case treatment case, M _i-case，j =0。

In an exemplary embodiment, the obtaining, according to the attribute of the one or more drilling incidents to be treated, the treatment recommendation model, using the treatment recommendation model, obtains a treatment recommendation result for the drilling incident to be treated, including:

converting attributes of one or more drilling incidents to be treated into a query vector;

converting a feature matrix corresponding to the recommended model of the treatment scheme into a row vector;

calculating the similarity of each element in the query vector and the row vector;

and (3) arranging the similarity values from large to small, and determining row vectors corresponding to the similarity values of topN preceding the similarity rows to generate recommended results.

In an exemplary embodiment, the determining the row vector corresponding to the similarity value of TopN preceding the similarity row to generate the recommendation result includes:

determining feature vectors corresponding to similarity values of the topN preceding similarity rows to form a recommendation result matrix M _TopN ；

Setting an accident complex type input mark TypeFlag, an accident phenomenon input mark SignFlag and an accident influencing factor input mark FactorFlag according to the query vector respectively;

According to the accident complex type input mark, the accident phenomenon input mark, the accident influencing factor input mark and the recommended result matrix M _TopN And generating a recommendation result.

In an exemplary embodiment, generating possible accident types MayTypes according to the accident complex type input flag TypeFlag includes:

from the recommendation result matrix M _TopN Acquiring a feature vector corresponding to a column of the accident complex type;

calculating corresponding scores according to each accident complex type by adopting a score formula;

sorting the obtained scores from big to small, and determining the complex types of the first a accidents as the MayTypes value;

wherein, the score formula is:

in the above-mentioned formula(s),representing accident-complex type scores, i' representing the rows in the recommendation result matrix,column representing recommendation result matrix, 0<K1 is less than or equal to; topN represents the number of rows of the recommendation result matrix, a being a positive integer.

In an exemplary embodiment, generating possible incident phenomena maysign from the incident phenomenon input signature SignFlag includes:

from the recommendation result matrix M _TopN Acquiring a characteristic vector corresponding to a column of the accident phenomenon type;

calculating corresponding scores by adopting a score formula aiming at each accident phenomenon type;

Sequencing the obtained scores from big to small, and determining the types of the accident phenomena ranked in the first a as a MaySigns value;

wherein, the score formula is:

in the above-mentioned formula(s),a score representing the accident phenomenon, i' representing a row in the recommendation result matrix,TopN represents the number of rows, K1, of the recommendation result matrix<≤(K1+K2)。

In an exemplary embodiment, generating a possible accident phenomenon MayFactors according to the influence factor input flag FactorFlag includes:

from the recommendation result matrix M _TopN Acquiring a characteristic vector corresponding to a column of the influence factor class;

calculating corresponding scores according to each accident reason by adopting a calculation formula;

sorting the obtained scores from big to small, and determining the types of the reasons of the first a accidents as a MayFactors value;

wherein, the score formula is:

in the above-mentioned formula(s),a score representing the cause of the incident, i' representing a row in the recommendation result matrix,TopN represents the number of rows of the recommendation result matrix, (K1+K2)<≤(K1+K2+K3)。

In an exemplary embodiment, the input labels are based on the accident-complex type, the accident-event input labels, the accident-influencing-factor input labels, and the recommendation-result matrix M _TopN Generating a recommendation result, comprising:

from M _TopN Acquiring a feature vector corresponding to a disposal method label column;

calculating corresponding scores by adopting a score formula for each treatment method;

sorting the obtained scores from big to small, and determining the top a treatment methods as MayActions values;

generating a recommendation result according to the determined MayTypes value, maysign value, mayFactors value and MayActions value;

wherein,

in the above-mentioned calculation formula, the calculation formula,representing the score of the treatment method, i' representing the row in the recommendation result matrix,TopN represents the number of rows of the recommendation result matrix, (K1+K2+K3)<≤(K1+K2+K3+K4)。

In a second aspect, an embodiment of the present invention provides a recommendation device for a drilling accident treatment plan, the device comprising: a memory and a processor; the memory is for storing a program for performing a recommended method of drilling accident treatment plan, and the processor is for reading the program for performing the recommended method for performing drilling accident treatment plan, performing the method of any of the above embodiments.

In a third aspect, embodiments of the present invention provide a computer-readable storage medium having stored thereon a data processing program for execution by a processor of the recommended method of drilling event management scheme of any of the above embodiments.

Compared with the related art, the recommendation method and device of the drilling accident handling scheme comprise the following steps: acquiring attributes of one or more drilling incidents to be treated; obtaining a recommended result of the treatment scheme of the drilling accident to be treated according to the attribute of one or more drilling accidents to be treated and by using a recommended model of the treatment scheme; the recommendation model of the treatment scheme is established according to the drilling accident knowledge body; the drilling accident ontology is established according to the related theoretical knowledge of the drilling accident and the historical treatment cases of the drilling accident. The application establishes a drilling accident knowledge body; and a recommended model of a disposal scheme is established according to the drilling accident knowledge body, the emergency disposal capability of a site engineer can be improved by utilizing the recommended model of the disposal scheme, the accident disposal time is shortened, the accident disposal efficiency is improved, and the disposal cost after the accident occurs is saved.

Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the application. Other advantages of the present application may be realized and attained by the structure particularly pointed out in the written description and drawings.

Drawings

The accompanying drawings are included to provide an understanding of the technical aspects of the present application, and are incorporated in and constitute a part of this specification, illustrate the technical aspects of the present application and together with the examples of the present application, and not constitute a limitation of the technical aspects of the present application.

FIG. 1 is a flow chart of a recommended method of drilling event handling scenario according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a drilling event ontology in some exemplary embodiments;

FIG. 3 is a schematic diagram of generating a feature matrix based on theoretical knowledge in some example embodiments;

fig. 4 is a schematic diagram of generating a feature matrix based on treatment cases in some example embodiments;

FIG. 5 is a feature matrix merge schematic in some example embodiments;

FIG. 6 is M in some exemplary embodiments _action A feature matrix row vector diagram;

FIG. 7 is a schematic diagram of obtaining a feature vector corresponding to TopN in some exemplary embodiments;

fig. 8 is a schematic diagram of a recommendation device for a drilling accident handling scheme according to an embodiment of the present application.

Detailed Description

The present application describes a number of embodiments, but the description is illustrative and not limiting and it will be apparent to those of ordinary skill in the art that many more embodiments and implementations are possible within the scope of the embodiments described herein. Although many possible combinations of features are shown in the drawings and discussed in the detailed description, many other combinations of the disclosed features are possible. Any feature or element of any embodiment may be used in combination with or in place of any other feature or element of any other embodiment unless specifically limited.

The present application includes and contemplates combinations of features and elements known to those of ordinary skill in the art. The embodiments, features and elements of the present disclosure may also be combined with any conventional features or elements to form a unique inventive arrangement as defined in the claims. Any feature or element of any embodiment may also be combined with features or elements from other inventive arrangements to form another unique inventive arrangement as defined in the claims. Thus, it should be understood that any of the features shown and/or discussed in this application may be implemented alone or in any suitable combination. Accordingly, the embodiments are not to be restricted except in light of the attached claims and their equivalents. Further, various modifications and changes may be made within the scope of the appended claims.

Furthermore, in describing representative embodiments, the specification may have presented the method and/or process as a particular sequence of steps. However, to the extent that the method or process does not rely on the particular order of steps set forth herein, the method or process should not be limited to the particular sequence of steps described. Other sequences of steps are possible as will be appreciated by those of ordinary skill in the art. Accordingly, the particular order of the steps set forth in the specification should not be construed as limitations on the claims. Furthermore, the claims directed to the method and/or process should not be limited to the performance of their steps in the order written, and one skilled in the art can readily appreciate that the sequences may be varied and still remain within the spirit and scope of the embodiments of the present application.

The key technologies for constructing the intelligent recommendation system comprise a knowledge base construction technology, a text feature extraction technology and a sparse vector processing technology. In knowledge base construction, knowledge extraction techniques are typically employed to extract domain knowledge from structured, semi-structured, and unstructured data sources. The knowledge extraction for unstructured data mainly adopts an information extraction technology to automatically extract the fact information such as entities, attributes, relations, events and the like from natural language texts. Information extraction is an important research direction of natural language processing, and comprises three subtasks of named entity recognition, relation extraction and event extraction. Named entity recognition (Named Entity Recognition, NER) has been a research hotspot in the field of NLP, mainly aimed at identifying entities and entity types from unstructured text, belonging to the category of sequence labeling. With the development of artificial intelligence technology, NER has evolved from early dictionary and rule-based methods to traditional machine learning-based methods, and in recent years to deep learning-based methods and Attention-based methods. In the aspect of text feature extraction, a text vector is generally generated by adopting methods such as One-Hot, TF-IDF, word2vec and the like. In the aspect of processing sparse feature vectors, the sparse matrix is generally converted into a dense matrix by adopting methods such as simple filling, labeling, clustering, singular value decomposition, feature value decomposition, dimension reduction, embedding and the like.

Although the intelligent recommendation technology is widely applied in the fields of medical treatment, catering, traveling, shopping and the like, the following technical problems still exist in the field of drilling engineering:

(1) The construction of a complex knowledge base of drilling accidents. Numerous scholars and research institutions at home and abroad summarize theoretical knowledge of a large number of drilling accidents and complex problems in theory and practice, including reasons, symbolization, prevention and treatment methods of different accidents; meanwhile, various drilling companies have accumulated a great number of technical measures, successful experience and failed training in the drilling construction process for decades. The theoretical knowledge, the disposal cases and the experience training are mainly stored in the enterprise in the form of unstructured documents, the data is difficult to obtain, a unified knowledge classification system is not established, and the manual arrangement is time-consuming and labor-consuming.

(2) The drilling treatment cases are mainly stored in unstructured documents, and information such as reasons, phenomena and treatment methods in the drilling cases cannot be effectively obtained based on the traditional text feature extraction technology.

(3) The characteristic dimension recommended by the drilling accident handling scheme comprises influencing factors, accident characterization and the like, the characteristic dimension is large, handling cases are few, and the data sparseness problem is serious.

In order to solve the problems, a recommendation method and a device for a drilling accident disposal scheme are provided, wherein theoretical knowledge and historical disposal cases of drilling accidents are fused, a drilling accident knowledge body is built, and a recommendation model of the disposal scheme is further built to solve the problems.

The embodiment of the invention provides a recommendation method of a drilling accident handling scheme, as shown in fig. 1, the method comprises steps S100-S110, and the method specifically comprises the following steps:

s100: acquiring attributes of one or more drilling incidents to be treated;

s110: and obtaining a recommended result of the treatment scheme of the drilling accident to be treated according to the attribute of one or more drilling accidents to be treated and the recommended model of the treatment scheme.

In this embodiment, the recommended model of the treatment plan is built according to the drilling accident ontology; the drilling accident ontology is established according to the related theoretical knowledge of the drilling accident and the historical treatment cases of the drilling accident.

The ontology defines core concepts, concept attributes, relationships between concepts, and concept instances within a particular domain. The drilling accident knowledge body mainly comprises core concepts such as blocks, wells, stratum, disposal cases, accident complex types, accident phenomena, influence factors, disposal methods and the like. Drilling incident ontology as shown in fig. 2, about 8 core concepts, 11 class relationships, and 50 attributes.

The drilling accident ontology includes: a theoretical ontology and a treatment case ontology;

In an exemplary embodiment, the attributes included in the theoretical knowledge base and the treatment case-base are specifically as follows:

1. accident complex type knowledge body

There is no clear boundary between drilling complications and drilling accidents, which may occur if the complications are not handled effectively in time. For each type of drilling complexity and drilling accidents, theoretical knowledge of the drilling complexity and the drilling accidents comprises the attributes of accident complexity reasons, accident complexity phenomena, preventive measures, treatment schemes and the like.

(1) Complex types of drilling

The well drilling complexity comprises lost circulation, well invasion, overflow, balling, water hole dropping, water hole blocking, well wall collapse and the like.

(2) Type of drilling accident

Drilling accidents comprise stuck drilling accidents, logging accidents, well cementation accidents, well control accidents, drilling tool breakage, underground falling objects and the like, wherein the stuck drilling accidents comprise adsorption stuck drilling, collapse stuck drilling, sand bridge stuck drilling, diameter reduction stuck drilling, key slot stuck drilling, cement stuck drilling and the like; well control accidents include kick, blowout runaway and the like.

2. Accident phenomenon knowledge body

The accident phenomenon is mainly divided into stuck point position, drilling tool condition, logging instrument condition, engineering parameter change condition, wellhead display condition, gas measurement abnormality, drilling fluid change and the like. For each type of accident, the theoretical knowledge includes the properties of explanation, possible reasons, inspection items, measures, etc.

(1) Clamping point position

The stuck point position is mainly divided into the vicinity of the drill bit, on the drill collar or drill rod, on the top of the drill collar, unmeasured and the like.

(2) Drilling tool condition

(1) Drilling tool activity conditions: when the drilling tool is lifted, the drilling tool is overtraked, lifted with resistance, lifted without resistance, lowered without resistance, blocked when moving up and down, abnormal rotation when encountering blocking and no resistance, and the like.

(2) Drilling tool failure: the pin is disconnected, etc.

(3) Logging instrument conditions

The condition of the logging instrument is mainly divided into damage of the logging instrument, blockage of the logging instrument, cable jump, lowering of the logging instrument but lifting of the logging instrument and the like.

(4) Engineering parameter variation

The engineering parameter changes are mainly divided into turntable rotation conditions, torque, pump pressure, mechanical drilling speed, suspended weight, vertical pressure, casing pressure and the like.

(1) The rotation condition of the turntable: and suddenly gets bigger when drilling quickly or when drilling is empty.

(2) Torque: abrupt torque increase, gradual torque increase, abrupt torque decrease, normal torque, etc.

(3) Pumping pressure: pumping pressure fluctuation, pumping pressure slow descending, pumping pressure suddenly ascending, pumping pressure slowly ascending, pump holding and the like.

(4) Mechanical drilling rate: rapid mechanical drilling speed, slow mechanical drilling speed, etc.

(5) Hanging weight: fluctuation of the suspended weight, decrease of the suspended weight, increase of the suspended weight, and the like.

(6) Vertical pressing: rise in vertical pressure, fluctuation in vertical pressure, fall in vertical pressure, etc.

(7) And (3) sleeve pressing: the sleeve pressure fluctuates, the sleeve pressure decreases, the sleeve pressure increases, etc.

(5) Wellhead display condition

The wellhead display condition is mainly divided into slurry pool liquid level variable quantity, drill cuttings display and the like.

(1) The change amount of the liquid level of the slurry pool: overflow, leakage, slurry injection height, upward flushing of drilling fluid to the rotating disc surface, etc.

(2) Drilling fluid circulation conditions: the drilling fluid can normally circulate, can circulate with small displacement, lose circulation, the liquid level of an annular space on a drill string does not drop, the drilling fluid returns to a wellhead along with the drill string, the drilling fluid is reversely sprayed into an inner hole of the drill string, the normal outlet of an inlet is reduced, and the like.

(3) Cuttings display: the amount of the returning is increased by a large amount of collapsed matter, the amount of returning is reduced, and the sheet-like falling blocks, the block-like falling blocks and the like are returned.

(6) Abnormal measurement of qi

(1) Full hydrocarbon and composition variation: background gas, negative value of heat conduction full hydrocarbon, single gas connection, gas tripping, single gas connection failure, alarm of H2S sensor and the like.

(2) Oil gas channeling upwards: change in upward movement altitude, change in upward movement speed, etc.

(7) Drilling fluid performance variation

The property change of the drilling fluid is mainly divided into the change of drilling fluid density, the change of drilling fluid conductivity, the change of drilling fluid temperature, the change of drilling fluid viscosity, the change of solid phase content, the change of PH value and the like.

(1) Drilling fluid density variation: unstable inlet density, unstable outlet density, sudden decrease in outlet density, decrease in inlet density, increase in inlet density, etc.

(2) Drilling fluid conductivity changes, namely inlet conductivity increases, inlet conductivity decreases, outlet conductivity increases, drilling fluid conductivity does not change (zero or positive value), drilling fluid conductivity suddenly changes and the like.

(3) The temperature of the drilling fluid is changed, the inlet temperature or the outlet temperature is not changed, the inlet temperature is rapidly decreased, the inlet temperature gradient is decreased, and the like.

(4) The viscosity of the drilling fluid is changed, namely, the viscosity of the drilling fluid is increased, the viscosity of the drilling fluid is decreased, and the like.

(5) The solid phase content changes, the solid phase content increases, and the like.

(6) PH value change, PH value rise, PH value fall, etc.

3. Accident influencing factor ontology

Accident influencing factors are mainly divided into geological factors, engineering factors, management and human factors. For each type of influencing factors, the theoretical knowledge comprises attributes such as factor description, reason description, possible complex types, possible accident types and the like.

(1) Geological factors

Geological factors may be further classified into unstable formations, abnormal formation pressures, specific geological formations, specific formation fluids, other specific formations, and the like.

(1) Unstable rock formations: carbonate formations, salt formations, cream salt formations, coal seams, bitumen formations, shale, siltstone, sandstone, sand, etc.

(2) Abnormal formation pressure: high pore pressure (abnormally high pressure), low fracture pressure (abnormally low pressure), stress relief, collapse pressure.

(3) Special geological structure: wrinkles, faults, cracks, karst cave, non-integrated contact surfaces, etc.

(4) Special formation fluids: high content of hydrogen sulfide, shallow gas and carbon dioxide.

(5) Other special formations: high permeability layer, easy-to-leak stratum, low stratum pressure bearing capacity, etc.

(2) Engineering factors

Geological factors can be further divided into equipment factors, design factors, disposal measures, well control factors, well cementation factors, process defects and the like.

(1) Equipment factor: well control equipment, well drilling equipment, well logging equipment and the like.

(2) Design factors: drilling fluid design problems, improper drill bit selection, no formation pressure prediction or large prediction error, unreasonable drilling tool combination design and the like.

(3) Treatment measures: emergency or handling measures are not reasonable.

(4) Well control factors: well control fluid, and the like.

(5) Cementing factors: poor cementing quality, cementing slurry, etc.

(6) And (5) process defects.

(3) Management and human factor

(1) Management factors: insufficient field management, weak combination of geological engineering integrated construction management and the like.

(2) Human factors: problems of operation specifications, experience inadequacies, poor awareness of post responsibility and safety, design execution problems, lack of technical instruction, inadequate well bore condition mastery, and the like.

4. Knowledge body of treatment method

The disposal method is mainly divided into an overflow disposal method, a lost circulation disposal method, a card measuring method, a card releasing method, a downhole junk disposal method and the like. For each class of treatment methods, its theoretical knowledge includes method definitions, principle descriptions, implementation steps, advantages, disadvantages, application scenarios, notes, etc.

(1) Overflow disposal method

The overflow disposal method is mainly classified into a well closing method, a well killing tool, and the like.

(1) The well closing method comprises the following steps: soft shut-in, hard shut-in, etc.

(2) Well killing method

Conventional well killing methods: driller's method of well killing (secondary circulation method), engineering method of well killing (primary circulation method or waiting for weighting method), and simultaneous circulation and weighting method of well killing.

Unconventional well killing method: balance point well killing, displacement well killing, pressure return well killing, low-throttling well killing, reverse circulation well killing, positive circulation well killing, throttling circulation well killing, etc.

The well control method under special conditions comprises the following steps: shallow well section overflow, same-layer blowout and leakage well killing, well killing without drilling tools in the well, well killing with upper blowout and lower leakage, well killing with lower blowout, well killing after overflow in tripping (a method of tripping after temporary well killing), well killing after overflow in tripping (a waiting circulation overflow method), well killing after overflow in tripping, and the like.

(3) Well killing tool

Blowout prevention tool in drilling tool: kelly up-cock, down-cock, etc.

Drilling tool back pressure valve: dished back pressure valves, arrow-shaped back pressure valves, drop-in check valves, ball back pressure valves, drop-in ball check valves, blowout preventers within a hous drill string, and the like.

(2) Lost circulation treatment method

(1) The large leakage treatment method comprises the following steps: static plugging during large leakage, plugging by fine particles and fibrous substances, plugging by unidirectional pressure sealing agents, plugging by bridging agents, plugging by high-water-loss slurry, plugging by acid-soluble plugging agents, plugging by cement slurry, and the like.

(2) The method for plugging the large karst cave with large cracks comprises the following steps: filling and plugging agent composite plugging, plugging mixture prepared by drilling fluid-colloid cement slurry and cement slurry, nylon bag plugging, net sleeve plugging tool and the like.

(3) The method for drilling by reducing the bottom hole pressure comprises the following steps: this refers to drilling without focusing on lost circulation, but with the drilling fluid column pressure balanced with formation pressure, for example: foam drilling fluid, low density water-in-oil drilling fluid, oxidized polyacrylamide aqueous solution, and the like.

(3) Card measuring method

The method is mainly divided into a calculation method, a card measuring method of a card measuring instrument and the like.

(4) Method for releasing card

The method is mainly divided into a card unlocking tool and a card unlocking method.

(1) Card releasing tool

The unlocking tool comprises a jarring unlocking tool, a back-off tool and the like.

Jarring unclamping tool: YSJ hydraulic jars, while drilling jars, super hydraulic jars, hydraulic compensated jars, double acting hydraulic jars, closed jars, and the like.

And (3) back-off tool: a reverser, a back-off joint, a back-off fishing spear, a back-off fishing barrel and the like.

(2) The unlocking method comprises the following steps: the movable drilling tool is released, the jar is released, the depressurization method is released, the pressure-holding circulation method is adopted, the reverse buckling sleeve is milled and released, the side drilling tool and the explosion are released, and the like.

(5) Method for treating objects falling in well

The method is mainly divided into a fishing tool and a fishing method.

(1) Fishing tool: slip overshot, fishing spear, male and female cone, cutting fishing tool, auxiliary fishing tool, etc.

(2) The salvaging method comprises the following steps: milling salvage, cutting salvage, slip salvage, piercing salvage, male cone salvage, female cone salvage and the like.

(6) New well hole drawing treatment method

The new well treatment method mainly comprises the steps of finding an old well scheme and redesigning a well track.

5. Handling case ontology

The system mainly comprises attributes such as a block, a well number, a stratum, an occurrence date, an accident complex name, a basic condition, an accident occurrence process, an accident handling process, accident cause analysis, summarization, experience and the like.

The treatment case knowledge instance is automatically extracted from unstructured documents such as accident drilling engineering design, accident summary report and the like by mainly adopting a natural language processing technology.

6. Block knowledge body

The method mainly comprises the attributes of block names, superior blocks and the like.

The block knowledge instance mainly originates from an enterprise relational database and can be directly obtained through an ontology mapping technology. The ontology mapping technique may employ a technique commonly used in the art, which is not particularly limited.

7. Well knowledge body

The method mainly comprises the attributes of well numbers, belonging blocks, well types, well position coordinates and the like.

The well knowledge instance mainly originates from an enterprise relational database and can be obtained directly through an ontology mapping technology. The ontology mapping technique may employ a technique commonly used in the art, which is not particularly limited.

8. Stratum knowledge body

The method mainly comprises the attributes of well number, stratum code, stratum name, top depth, ground depth, lithology and the like.

The stratum knowledge instance mainly originates from an enterprise relational database and can be directly obtained through an ontology mapping technology. The ontology mapping technique may employ a technique commonly used in the art, which is not particularly limited.

In an exemplary embodiment, establishing a recommendation model of a treatment plan according to the drilling accident ontology includes:

the first step, labeling each attribute in the drilling accident knowledge body, and establishing a drilling accident label system;

generating a disposal case feature matrix based on historical disposal case labels in the drilling accident label system;

step four, respectively setting corresponding weights for the theoretical knowledge feature matrix and the treatment case feature matrix, and combining to obtain a feature matrix M;

and fifthly, constructing a recommended model of the treatment scheme according to the feature matrix M.

In the first step, labeling each attribute in the drilling accident knowledge body, and establishing a specific implementation process of the drilling accident label system comprises the following steps:

splitting and marking the accident complex theoretical knowledge and the related attribute values of the treatment cases to form a label system of the theoretical knowledge and the case key information; the labeling can be performed by a manual labeling method or an automatic text classification method.

1. Manual labeling method

The manual labeling method mainly adopts a manual mode to analyze the complex theoretical knowledge of the accident and the relevant attribute values of the treatment cases, and selects classification nodes from the corresponding knowledge body as labels.

(1) Complex theoretical knowledge of labeled accident

Aiming at the theoretical knowledge of each type in the accident complex type knowledge ontology, analyzing attribute values such as accident complex reasons, accident complex phenomena, disposal schemes and the like, and selecting matched nodes from influence factor knowledge ontologies, accident phenomenon ontologies and disposal method knowledge ontologies as labels thereof.

(2) Labeling disposal cases

And analyzing attribute values such as accident complex names, accident occurrence passes, accident handling processes, accident reason analysis and the like of each treatment case, and selecting classification nodes from the corresponding ontology as labels of the cases.

(1) Labeling accident complex names: extracting attribute values of 'accident complex names' of the treatment cases, and selecting corresponding classification nodes from 'accident complex type knowledge ontology' as accident complex type labels of the cases. One treatment case may have multiple accident complex type tags.

(2) The labeling accident happens through: the 'accident passing' attribute value of the disposal case is extracted, the content related to the accident phenomenon is analyzed, and the corresponding classification node is selected from the 'accident phenomenon knowledge ontology' to be used as the accident phenomenon label of the case. One treatment case may have multiple incident labels.

(3) Labeling accident causes: the attribute value of 'accident cause analysis' of the treatment case is extracted, the content related to the accident influence factors is analyzed, and the corresponding classification nodes are selected from 'influence factor knowledge ontology' to be used as influence factor labels of the case. A treatment case may have multiple influencing factor tags.

(4) The labeling accident handling process comprises the following steps: the method comprises the steps of extracting attribute values of an accident handling process of a handling case, analyzing content related to an accident handling method, and selecting corresponding classification nodes from an accident handling method ontology as an accident handling method label of the case. One treatment case may have multiple incident treatment method tags.

2. Automatic text classification method

An automatic text classification method refers to a process of mapping a piece of text to a predetermined tag or tags.

(1) Data set construction

And extracting accident occurrence process, accident handling process and accident reason attribute values of each handling case, and manually marking to generate a training data set. Each sample in the training dataset includes both text and labels.

(2) Text preprocessing

And performing word segmentation, word deactivation and other operations on each text in the data set.

(3) Text feature extraction and vectorization

After the text is preprocessed, a text vector can be generated by adopting One-Hot, TF-IDF, word2vec and other methods, and a text feature vector is generated by adopting the word2vec method.

(4) Construction of classification models

After the text feature vector is generated, the text feature vector can be trained by adopting a support vector machine, a random forest, a neural network and other methods to generate a text classification model.

(5) Model application

And applying the trained classification model to the text which is not marked yet, obtaining a classification result, and taking the classification result as a label of the text.

(6) Classification result auditing

In order to ensure the correctness of the label, the classification result can be manually checked and verified.

By adopting any method, each attribute in the drilling accident knowledge body is subjected to labeling treatment, and a drilling accident label system is established.

Wherein, the theoretical knowledge label in the drilling accident label system includes: an accident complex type tag (K1), an accident phenomenon tag (K2), an influence factor tag (K3) and a disposal method tag (K4);

the accident-complex-type tag includes one or more of the following: drilling complex type tags and drilling accident type tags;

the treatment method tag includes one or more of the following: a well killing method, a lost circulation treatment method, a card measuring method, a card releasing method and a down-hole junk treatment method.

In an exemplary embodiment, the generating a theoretical knowledge feature matrix based on theoretical knowledge tags in the drilling accident tag system in the second step includes:

the number (K1) of the accident complex types is used as a matrix line number, the total number (K) of the labels of the theoretical knowledge labels is used as a matrix column number, and the construction is carried outTheoretical knowledge feature matrix M _theory ；

Wherein K represents the total number of labels of the theoretical knowledge labels, K1 represents the number of labels of accident complex types, K2 represents the number of labels of accident phenomena, K3 represents the number of labels of influencing factors and K4 represents the number of labels of disposal methods; k=k1+k2+k3+k4.

The theoretical knowledge characteristic matrix M _theory Element M of (3) _i-theory，j I-they represent accident complex type numbers, j represents tag numbers; if the j-th tag exists for the i-th incident complex type, M _i-theory，j =1; if the j-th tag does not exist in the i-th accident complex type, M _i-theory，j =0. In this embodiment, the theoretical knowledge feature matrix construction process is as shown in fig. 3:

after labeling the accident complex theoretical knowledge, a theoretical knowledge feature matrix can be generated, and the theoretical knowledge feature matrix M is adopted _theory1 Number of theoretical knowledge of accident complex type N _theory ×K _theory Representation, where N _theory Represents the theoretical knowledge number, K of the accident complex type _theory The number of features representing each theoretical knowledge. Since the theoretical knowledge is mainly composed of different accident complex types of theoretical knowledge, N _theory =K1，K _theory The theoretical knowledge feature matrix can be expressed as k2+k3+k4

M _theory 1=N _theory ×K _theory =K1×(K2+K3+K4)。

As shown in fig. 3, each accident complex type theoretical knowledge is numbered as theoretical knowledge 1 and knowledge 2.

M _theory2 =N _theory ×K=N _theory X (k1+k2+k3+k4). For each element M of the theoretical knowledge feature matrix _ij I represents a theoretical knowledge number, j represents a tag number, and if the ith theoretical knowledge has the jth tag, M _ij =1; if the ith theoretical knowledge does not have the jth tag, M _ij =0。

In an exemplary embodiment, generating a treatment case feature matrix based on historical treatment case tags in the drilling incident tag system in a third step comprises:

taking the number of historical treatment cases as the number of rows, taking the total number of labels (K, K=K1+K2+K3+K4) of the theoretical knowledge labels as the number of matrix columns, and establishing a historical treatment case feature matrix M _case ；

Wherein the historic treatment case feature matrix M _case Element M of (3) _i-case，j I-case represents a treatment case number, j represents a label number; if the j-th label exists for the i-case treatment case, M _i-case，j =1; if the j-th tag does not exist in the i-case treatment case, M _i-case，j =0. In this embodiment, the feature matrix is generated based on the treatment case, as shown in fig. 4:

after labeling the treatment cases, a case knowledge feature matrix can be generated, wherein the case knowledge feature matrix adopts M _case =N _case ×K _case Representation, where N _case Representing the number of cases (1 well several incidents), each row representing one case, each case being K _case Dimension vector, K _case =k=k1+k2+k3+k4, i.e. K features. The treatment case feature matrix can be represented as M _case =N _case ×K _case =N _case ×K=N _case ×（K1+K2+K3+K4）。

All cases are numbered 1 2 the term "n", for each element M of the feature matrix _ij I represents a case number, j represents a label number, and if the ith case has the jth label, M _ij =1; if the jth tag does not exist in the ith case, M _ij =0。

In an exemplary embodiment, in the fourth step, the theoretical knowledge feature matrix and the treatment case feature matrix are respectively set with corresponding weights, and are combined to obtain a feature matrix M, which is implemented as follows:

1. setting theoretical knowledge weight and treatment case weight

Theoretical knowledge is the basis and case knowledge is derived from production practice. The importance of different types of knowledge in the recommendation process is embodied by setting theoretical knowledge weight and case knowledge weight. In this embodiment, the weight of the theoretical knowledge feature matrix is represented by w1, the weight of the treatment case feature matrix is represented by w2, and the weights of w1 and w2 can be set according to specific situations; the sum of the w1 weight and the w2 weight is 1.

2. Combining theoretical knowledge feature matrices and treatment case feature matrices

As shown in FIG. 5, the theoretical knowledge feature matrix M _theory Multiplying by weight w1, case knowledge feature matrix M _case Multiplied by a weight w2 and then combined in rows, denoted m=n×k, where N represents the total number of knowledge and n=k1+n _case K represents the total number of features, k=k1+k2+k3+k4. Each element M in the feature matrix after combination _ij I represents a knowledge number (theoretical knowledge+case knowledge), and j represents a tag number.

In an exemplary embodiment, obtaining a treatment method recommendation for a drilling incident to be treated according to attributes of one or more drilling incidents to be treated and using the treatment method recommendation model, comprising:

the method comprises the steps of firstly, converting the attribute of one or more drilling accidents to be treated into a query vector Q;

secondly, converting a feature matrix M corresponding to the treatment method recommendation model into a row vector;

thirdly, calculating the similarity of each element in the row vectors converted by the query vector and the feature matrix;

and fourthly, sequencing the obtained similarity values, and determining a preset TopN row to generate a recommendation result.

In this embodiment, the process of generating the feature matrix is as follows:

(1) Selecting feature columns according to recommended targets

The drilling accident complex feature matrix M consists of accident complex types, accident phenomena, influence factors and treatment methods, and the feature number K=K1+K2+K3+K4. If the treatment scheme is selected as the recommended target, the accident complex type, the accident phenomenon and the shadow are selectedGenerating and disposing method characteristic matrix by using M _action =N×K _action Representation, where K _action Representing the number of features, K _action =K1+K2+K3。

(2) Generating query vectors from user input

The user can select one or more attributes from the "accident complex type ontology", "accident phenomenon ontology" and "influence factor ontology" as input of the method. The system automatically converts user input into K _action Dimension query vector q= (Q) ₁ ,q ₂ ,q ₃ ,...，q _kaction ) If the user selects the j-th tab, q _j =1, otherwise q _j =0。

(3) Calculating similarity

1. Representing a treatment method feature matrix M by using row vectors _action As shown in FIG. 6, where m _i Is a K _action Dimension column vector, m _i =（m _i1 ,m _i2 ,...,.m _ij ,...,m _ik ）。

2. Calculating query vector Q and processing method feature matrix M _action Each element m in the row vector _i Similarity ρ of (1) _i The following is shown:

3. through the steps, N similarity values are obtained in total, wherein the similarity values are ρ ₁ 、ρ ₂ …ρ _i …ρ _n Ordering the values according to the similarity, and obtaining the previous TopN values.

(4) Generating recommendation results

1. For the row corresponding to the front TopN similarity values in the similarity sequence, and obtain the feature vector of the corresponding row from the combined feature matrix M, as shown in fig. 7, obtain the feature vector corresponding to TopN, and adopt M _TopN Representing a matrix of TopN rows of eigenvectors.

2. Setting TypeFlag, signFlag and FactorFlag according to user query vector Q

TypeFlag: the accident complex type inputs the sign. Typeflag=false if the user does not select any accident complex class label, otherwise typeflag=true.

SignFlag: the accident phenomenon is input with a sign. Sign flag=false if the user does not select any incident class label, otherwise sign flag=true.

FactorFlag: the accident influencing factors are input with marks. If the user does not select any influence factor class label, then factorflag=false, otherwise factorflag=true.

(1) Generating a possible accident type from TypeFlag

Using MayTypes to represent the possible accident types, if typeflag=true, mayTypes are empty; if typeflag=false, then an inference of the likely accident type will be given in the recommendation, as follows:

a: from M _TopN Acquiring a feature vector corresponding to the accident complex type tag column;

b: the complex type for each accident is noted asCalculated by the following formulaThe score, wherein, the score formula is:

in the above-mentioned formula(s),representing accident-complex type scores, i' representing the rows in the recommendation result matrix,column representing recommendation result matrix, 0<K1 is less than or equal to; topN represents the number of rows of the recommendation result matrix.

c: sorting the obtained scores from big to small, and determining the types of the accident phenomena ranked in the first a as a MayTypes value; such as: after ranking the K1 scores, the 3 types with the higher scores are selected as MayTypes values.

(2) Generating possible accident phenomena according to SignFlag

Using maysign to represent a possible accident phenomenon, if signflag=true, then maysign is empty; if sign=false, then the recommendation will give an inference of a possible accident as follows:

a: from M _TopN Acquiring a feature vector corresponding to the accident phenomenon type tag column;

b: calculating corresponding scores by adopting a score formula aiming at each accident phenomenon type;

wherein, the score formula is:

c: sequencing the obtained scores from big to small, and determining the types of the accident phenomena ranked in the first a as a MaySigns value; such as: after the K2 scores are ranked, 3 accident phenomena with higher scores are selected as the maysign values.

(3) Generating possible accident causes according to FactorFlag

The possible accident cause is represented by using a mayfactor, and if the factor flag=true, the mayfactor is empty; if the factor flag=false, then the recommendation will give an inference of the likely cause of the accident as follows:

a: from M _TopN Acquiring a feature vector corresponding to the influence factor type tag column;

b: calculating corresponding scores according to each accident reason by adopting a calculation formula;

wherein, the score formula is:

c: sorting the obtained scores from big to small, and determining the types of the reasons of the first a accidents as a MayFactors value; for example: after the K3 scores are ranked, 3 accident reasons with higher scores are selected as the MayFactors value.

(4) Recommendation generation method

The recommended treatment method is represented by MayActions, and the calculation flow is as follows:

a: from M _TopN Acquiring a feature vector corresponding to a disposal method label column;

b: calculating corresponding scores by adopting a score formula for each treatment method;

wherein, the score formula is:

c: sorting the obtained scores from big to small, and determining the top a treatment methods as MayActions values; for example: after the K4 scores are ranked, the 3 treatment methods with the higher scores are selected as MayActions values.

(5) Generating recommendation results

Through the above steps, the final recommendation RecResult, recResult =maytypes value+maysign value+mayfactors value+mayactions value is generated.

In the embodiment of the application, a complex knowledge body of the drilling accident is established first; then splitting and labeling the historical cases to form a label system of case key information, and constructing a drilling accident handling scheme feature matrix by combining theoretical knowledge; and finally, generating a feature vector according to the case label, and establishing a treatment scheme recommendation model. Based on the established drilling accident handling scheme recommendation model, possible types and possible reasons of the accident are intelligently analyzed according to user input attributes, an accident handling method and related cases are recommended, timely and effective data support is provided for analysis and judgment of an accident scene, and accident handling efficiency is improved.

The embodiment of the invention also provides a device for handling drilling accidents, as shown in fig. 8, the device comprises: a memory 800 and a processor 810; the memory is for storing a program for performing a drilling incident treatment, and the processor is for reading and executing the program for performing a drilling incident treatment, and executing the method of any one of the above embodiments.

Embodiments of the present invention also provide a computer-readable storage medium having stored thereon a data processing program that is executed by a processor to perform the method of any of the above embodiments.

Example one

Aiming at the accident phenomenon input by a user, the method for handling the drilling accident determines a recommended result, and the implementation process is as follows:

inputting an up-mentioned resistance attribute, and obtaining a treatment method recommendation result of the drilling accident to be treated by using a treatment method recommendation model, wherein the treatment method recommendation result comprises a possible complex type, a possible reason and a recommendation method.

(1) Possibly of complex type

Sequence number	Complex name	Case of association of this type
			1	Diamond clip	【1】 XX1 stuck drill [ 2 ] XX2 stuck drill [ 3 ] XX3 stuck drill [ 4 ] XX4 stuck drill [ 5 ] XX5 stuck drill [ 6 ] XX6 stuck drill [ 7 ] XX7 stuck drill
2	Reducing stuck drill	【1】 B oil field YY1 reducing stuck drill [ 2 ] B oil field YY2 reducing stuck drill [ 3 ] C oil field YY3 reducing stuck drill
			3	Collapse stuck drill	【1】 Afield ZZ1 collapse stuck drill [ 2 ] A field ZZ2 collapse stuck drill

(2) Possible reasons for

Sequence number	Factor name	Cause of generation	Of the main complex type	Can generate accidents	Association case
						1	Emergency or department No reason measures Reasonable and reasonable	Underground condition judgment The cutting-off is not accurate, and the cutting-off time is short, treatment measure system The determination is made as to whether the error is correct, emergency means Less treating prescription Case analysis correction Untimely to the ground	Finding downhole complications The condition is not processed in time, is not complex in nature Complexity of problem solving To make it unnecessary to use The accident occurred Artificially cause	To recover the underground condition Hybridization	(1) A oilfield XXX11 key slot is stuck on The blocking phenomenon is existed when the thirty third column is started, the third column is not to be started 400kN is mentioned to 1000kN. Visible worker The thought of the person is that there is no key slot for clamping A concept. (2) C oil field XXX12 reducing diameter No plugging measures are taken for the stuck well (3) D oil field XXX13 mud-clad stuck drill Unreasonable measures of management(1) Found to have mud bag pulling The phenomenon of the piston should not be raised further Lowering to the well section without blocking, starting the pump for circulation, try to eliminate balling. (4) A oilfield XXX14 collapse stuck drill has no technical sleeve Side drill (5C) for sealing and separating well Oil field XXX15 reducing stuck drill treatment measure The application is unreasonable. The drill is not blocked at all Should be hard to handle. If a circulating drilling fluid is adopted, The method of slightly lifting and slowly turning is not to treat Accidents can occur. (6) Afield XXX16 Dry drill stuck check drill without immediate drill-up With means for continuing drilling without decreasing pump capacity In the case of a pump pressure of 18MPa to 14MPa, and further down to 9MPa, is Ming And the short circuit circulation phenomenon is obvious. (7) XXX17 Backing-up is not done in the process of cutting off drilling tool to reaming Well controlling the process of drilling, blocking and back reaming In the middle, the backing is not well controlled, resulting in a drilling tool Tripping and falling into the well. (8) XXX18 stuck drill Improper measure is carried out when the underground is blocked, and the measure is not taken In complex well sections, however, risk awareness is unclear, three-valve and offset displacement are adopted when the eyes are scratched Large, resulting in collapse and accumulation of pump out. (9) XXX19 stuck drill is not timely taken countermeasures Discontinuity of the grouting leakage stopping slurry during the application And the problem is that countermeasures are not taken in time.
2	Drilling fluid device Problem counting	Drilling fluid system And rheology and formation property is not good Is adapted to	Easily-eroded well wall rock Stone, causing collapse; easy to use Causing shrinkage of the soft formation Diameter of the pipe	Some of the open hole sections Some formations shrink in diameter or Collapse, resulting in a well Blowout, lost circulation or well Collapse of	(1) XXX21 stuck drilling fluid performance is not available The method can meet the requirement that the drilling fluid is not available Meet the requirement of stabilizing the well wall of the strip lake group. (2) XXX22 stuck mud performance is not satisfactory The underground collapse prevention requirement asphalt amount is slightly followed Hysteresis. Slurry propertiesCannot meet the underground anti-theft requirement Collapse requirements. (3) XXX23 stuck drilling mud carrying device Poor mud carrying capacity with poor capacity and cuttings Can not move in time in the horizontal section to form rock The chip bed also causes the upward lifting drilling tool to pump, The main reason for sticking.
						3	Operating Specification Problem(s)	Technical quality Low operation skill Improper operation or loss of function Error, violation, department Irregularities in drilling operations Is not strict, Inexperienced delivery of Drill non-uniformity	Operator pair of drilling machine Critical part and device Lack of stringent requirements And a checking means for making Complications in downhole situations Or cause a larger accident	To recover the underground condition Hybridization, presence of stringency Heavy well control hidden trouble	(1) XXX31 broken bit (2) XXX32 clip The operator of the drilling brake lever is operated improperly Is too much lifted and is not reached in the pumping up process The time picking pump causes pump jamming and jamming. (3) The XXX33 drilling sticking construction plan is imperfect, and the injection is carried out The construction of the plugging slurry is discontinuous.

(3) Recommended method bath well stuck freeing (bubble stuck freeing agent)

Case 1 AAA11

The treatment method comprises the following steps:

【1】 The injection of the detergent, the diesel oil and the clean water [ fail ] has no effect

The injection of the stuck releasing agent, the diesel oil soaking and the clean water injecting have no effect, and the pumping pressure is increased from 16MPa to 22MPa, so that lost circulation is realized.

【2】 Cement plug start sidetrack drilling [ successful ]

The rotary disc is used for back-off, and the rotary disc is used for back-off from 4763.80m, and the fish 128.48m (comprising 83/8in tricone bit 0.24m+430×4A10 joint 0.63m+61/4in drill collar 9.11m+81/8in centralizing joint 0.38m+61/4in drill collar 118.12 m) is used. The cement plug was sidetracked starting from 4289.20 m.

Case 2 AAA12

The treatment method comprises the following steps:

【1】 The foam-releasing agent (failure) is ineffective

The amount of the bubble releasing card is 3 times, and the fresh water drilling fluid is 1 time ineffective.

【2】 Milling and unclamping [ successfully ]

The density of the drilling fluid is increased to 1.92g/cm ³ And (5) bursting the back-off button, and sleeving, milling and releasing the clamp. The loss was 28.5 days.

Case 3 AAA13

The treatment method comprises the following steps:

【1】 The jar is not effective in the down stroke (failure)

The ground plane jar is down hit and is not effective.

【2】 Ineffective infusion of diesel fuel [ failure ]

The diesel oil is injected for soaking, and the diesel oil is hit from the ground, so that the diesel oil is ineffective.

【3】 Inverted sleeve milling [ failure ]

And (5) reversely buckling by using an original drilling tool, and reversely opening from the well depth 952.02 m. The fish length was 135.25m.

The lower left-handed threaded drill rod is reversely buckled for the second time, and the male cone is buckled in a sliding way.

And a milling barrel with the diameter of lower phi 193.67mm is sleeved and milled for 10m.

The lower left-handed thread male cone is reversely buckled and still is slipped.

【4】 Side drilling of filled wells [ successful ]

Case 4 AAA14

The treatment method comprises the following steps:

【1】 Lifting the lower movable drilling tool (failure)

The method is characterized in that the method is adopted to gradually increase the discharge capacity and keep the pumping pressure not to exceed 15MPa, the circulation is carried out until about 1:00 of 27 days of 6 months, the discharge capacity is increased to 36L/s, the pumping pressure is 20MPa, and the discharge capacity and the pumping pressure are recovered to those in normal drilling. During the process, the movable drilling tool is continuously lifted and lowered, the maximum lifting is 1800KN, the pressing is carried out until 1000KN, and the slip-sitting intermittent strong rotating drilling tool is 20 circles and returns to 20 circles.

【2】 Soaking mixed crude oil; soaking of Mixed diesel oil [ failure ]

The mixed crude oil is soaked for 6 months and 27 days, and the circulation movement drilling tool is 9:00 to 14:30, the maximum is lifted to 1800KN, the mixed crude oil is pressed down to 800KN, the drilling tool is rotated for 20 circles, and the drilling tool is rotated for 20 circles.

【3】 Explosion trip [ failure ]

7 months and 8 days, and 7:20 testing truck testing card point to 3690m, and explosion releasing the buckle at 3711.25 m. 5 single pump-on top passes, tiecard, were started from 14:40 to 15:00. The single valve displacement is 11L/s, the pumping pressure is 3MPa, the pumping pressure is circulated to 16:00, the pumping pressure is suddenly increased to 9MPa, and the wellhead is lost. The movable drilling tool is lifted up and down, the maximum lifting is 2600KN, the maximum lifting is 400KN, the movable drilling tool is 18:00, and 5 single drilling tools are lifted up. 18:00 removal of the kelly begins tripping out and tripping out 19 columns. And (3) connecting a kelly bar at a ratio of 20:30, opening the pump, pumping up to 8MP, taking 5 single units out at a ratio of 20:30, opening the pump, and pumping up to 2.5MPa. Starting tiecard, circulating to 7 months and 9 days 5:40, and starting normal drill-out.

【4】 Sidetracking [ success ]

Case 5 AAA15

The treatment method comprises the following steps:

【1】 Acid soaking [ failure ]

Acid soaking twice: after soaking for 2 times, the acid movement drilling tool is not unclamped, and the acid can not be effectively soaked to the stuck point.

【2】 Jarring [ failure ]

Testing card and setting hole: the stuck point 2350m was measured and perforation was successful at well depth 2385.72 m. Shock and retest stuck point: the jarring is not unclamped, the stuck point is repeatedly measured, the stuck point is not measured, the rotary table rotates for four circles during the stuck point measurement, no reverse torque exists, the suspended weight of the lifting drilling tool is reduced, the original suspended weight is 54 tons, the current suspended weight is 46 tons, the drilling tool can move up and down, the broken drilling tool is analyzed and judged to be started up for examination, and the fish top position is: 2154.70m and 337.74m.

【3】 Explosion and loosening of lower instrument (failure)

Buckling and explosion loosening: the lower back-buckling buckle successfully rises to 520KN in suspended weight, the lower instrument explodes and loosens the buckle, and the plugging sidetracking is determined through research and discussion.

Example two

Aiming at the type of the accident input by the user, the method for handling the drilling accident determines a recommended result, and the implementation process is as follows:

the user inputs the attribute of 'drilling tool breaking', and obtains the recommended result of the treatment method of the drilling accident to be treated by using the recommended model of the treatment method, wherein the recommended result comprises possible reasons, possible phenomena and recommended methods.

(1) Possible reasons for

Sequence number	Factor name	Cause of generation	Of the main complex type	May generate things Therefore, it is	Association case
						1	High sulfur content Hydrogen gas	During drilling, sulfur Hydrogen sulfide working with a wellbore Fluid or produced fluid Into the ground, which has And (5) strong corrosiveness. Hydrogen sulfide Will cause the viscosity of the drilling fluid Greatly promote and cause Fluidity of drilling fluid Difference; will also be strongly corroded Damage the drilling tool; in drilling well Liquid outlet groove surface detection Less than hydrogen sulfide, with Greater "concealment", easy bring to well control Risk, poor control Easily causes casualties.	Causing the drilling tool to generate hydrogen The bubbling caused a crack to occur, embrittling the steel material to obtain Will not be controlled in time Causing serious problems Therefore, the method is simple and convenient.	Expansion of hydrogen sulfide Powder for human bodyHas the following components Important injuries, severe Heavy causes of people Casualties of the person.	(1) XXX1 overflow and breaking drilling tool pair The block has high concentration hydrogen sulfide Is unknown to the present block The hazard of high concentration hydrogen sulfide Not known and not adopting pressing in time The additive back process presses the hydrogen sulfide back into the formation, monitoring the concentration of Hydrogen sulfide at the surface 46ppm, while the actual concentration downhole is greater High conditions cause the drilling tool to appear Severe corrosion phenomena, downhole drilling tools Damage in advance, causing severe drilling tools An accident.
2	Drilling fluid device Problem counting	Drilling fluid system and flow Denaturation and formation Properties Is not adapted to	Easily-eroded well wall rock Stone, causing collapse; easy to use Causing shrinkage of the soft formation Diameter of the pipe	In open hole section Some formations shrink Diameter or collapse, build Blowout and lost circulation Or collapse of well	(1) XXX2 overflow and breaking drilling tool The well fluid density cannot meet the downhole pressure Well needs to find overflow after closing Minimum measured from Guan Jingli pressure The density of the well-killing drilling fluid is as follows 1.22g/cm3, under specific pressure In well construction, due to the lower-in-tower drill The stratum is limestone of Oregano system Stratum, density according to adjacent well experience Too high, possibly out during well killing Large lost circulation, so in the first two times The engineering method adopts density 1.17 g/cm3 solids free drilling A hydraulic well, the density of which cannot meet the underground condition The need for well killing.
						3	Emergency or department No reason measures Reasonable and reasonable	Judging whether the underground condition is not Accurate, disposition ofMeasure system Incorrect positioning emergency hand Few sections, treatment scheme is divided into Failure to correct the place	Finding downhole complications The condition is not processed in time, is not complex in nature Complexity of problem solving To make it unnecessary to use The accident occurred Artificially cause	To enable downhole conditions Complicating the process	(1) XXX3 cutting tool to scratch In the process, the reversing is not controlled well Drilling in the process of drilling in the resistance and back reaming The backing is not well controlled, and the drilling tool is disengaged And (5) buckling and falling the well.

(2) Possible phenomena

Sequence number	Phenomenon name	Cause of generation	Examination item	Treatment measures	The phenomenon is related to cases
						1	Male buckle disconnection	Post-onset of 12:20 presentation at 21 months The male buckle of the non-magnetic drill collar is disconnected XXX well breaking tool pin breaking tool Drill inspection, 3:00 drill-out completed, found Bending 30 drill rods, the last The male buckle end is broken at 1.86 m.
2	Overflow volume	Formation pressure is not known The accuracy is high; drill-out Shi Jing Inward short of filling with banjo Also; excessive suction Pressure; drilling fluid-tight The degree is low; drilling fluid leakage Loss of function; formation pressure difference Often times	Checking drilling fluid return In a measuring and drilling fluid pool Drilling fluid amount; inspection of Well drilling after stopping pump Whether or not the liquid overflows	After finding overflow Immediately closing the well; at the position of The wellhead is led to a certain degree Is to (1) local resistance Increasing the annular space Pressure of (i.e. well) Back pressure is applied to the mouth.	XXX well overflow and broken drilling tool overflow Quantity overflow 0.6m ³
						3	Upward lifting resistance	XXX well breaking drilling tool tripping and blocking Starting the drill starting at 22:00 and starting at 22:30 Drilling to 1030m to meet resistance, and connecting kelly Pump-on circulation back-scribing
4	Drop of suspended weight	Indicating that the weight table is bad; drilling tool Breaking; pump pressure drops; lost circulation and blowout	Check weight table transmission Good presser and rubber tube Bad, whether there is a table Liquid and zero in table Whether or not the member is damaged Post-recalibration suspension Weight, e.g. indicating that the weight is complete Good immediate drill-up detection Checking drilling tools according to the drill Determination of breaking condition The treatment method is as follows.	If the drilling tool is disconnected The drill should be started up immediately and the drill should be started, breaking according to drilling tools Condition determination salvaging Method of	Suspension weight drop of XXX well breaking drilling tool The well depth is 1027m by back-scribing 23:05, the drilling tool is tripped, and the suspended weight is reduced from 550KN To 60KN

(3) Recommended method for unlocking movable drilling tool

Case 1 XXXX drilling tool

The treatment method comprises the following steps:

【1】 Stuck-removing agent soaking well section, open hole section soaking oil-based mud [ failure ]

【2】 Jarring of jarring device [ failure ]

【3】 Movable drilling tool [ failure ]

2021, 8, 19, 00:00-8:00 active drilling tool, releasing the clamp (suspended weight 130T, torque 20 KN.m, pressing down to suspended weight 28T, observing 30 minutes and 10 minutes of activity once

【4】 Explosion trip [ failure ]

The 0:20 weighting rod meets resistance at 4420 m position in 8 month of 2021, the cable is lifted out from 1:20, the original suspension weight of the top drive is connected with torque 25 KN.m and is lowered to 28T for rest 30min, the original suspension weight is lifted to 200T for rest 1 h to 9:50, the operation is repeated, the explosive is released from the cable lifting (the release position is 4400 m) button, the powder is assembled from 14:30, the instrument is lowered, the explosive is released from the 18:00, the cable lifting (the release position is 4381 m) button is not released, the pumping pressure is pumped to 25MPa, the lifting is carried out for 200T-210T, the lifting is carried out for 28T for 30min, the lifting is carried out to 160T of the original suspension weight, the torque is increased for 30 KN.m, the lifting is carried out for 28T, and the operation is repeated for 30 min.

Example three

The user inputs the attributes of the cracks and the karst cave and obtains the recommended results of the treatment method of the drilling accidents to be treated by using the recommended model of the treatment method, wherein the recommended results comprise the types of the possible accidents, the possible phenomena and the recommended methods.

(1) Type of possible accident

Sequence number	Complex name	Case of association of this type
			1	Lost circulation	【1】 XXX well leakage [ 2 ] XXX well leakage [ 3 ] XXX well leakage [ 4 ] XXX well leakage [ 5 ] XXX well leakage [ 6 ] XXX well leakage [ 7 ] XXX well leakage [ 8 ] XXX well leakage [ 9 ] XXX well leakage [ 10 ] XXX well leakage 【11】 XXX well lost circulation
2	Overflow flow	【1】 XXX well overflow, drill hydrogen embrittlement fracture [ 2 ] XXX well overflow [ 3 ] XXX well overflow [ 4 ] XXX well overflow Flow of
			3	Blowout	【1】 XXX well blowout [ 2 ] No. I gas field XXX well blowout

(2) Possible phenomena

Sequence number	Phenomenon name	Cause of generation	Examination item	Treatment measures	The phenomenon is related to cases
						1	Leakage amount	Drilling fluid leakage Loss of function	Without any means for	Lowering drilling fluid-tightness A degree; altering a well Viscosity of liquid and cutting Force; adjusting a well Engineering measures	(1) XXX well lost circulation working fluid (comprising mud, water Mud, completion fluids, and other fluids, etc.) under differential pressure Drilling into stratum by using leakage into well depth 4636m to generate leakage Loss of 1m ³ 1. Reducing the density of the drilling fluid. 2. Altering a well Liquid viscosity and shear force. 3. Adjusting drilling engineering measures (2) XXX well lost circulation working fluid (comprising mud, water Mud, completion fluids, and other fluids, etc.) under differential pressure With a total leak-off of 7m into the stratum ³ 1. Lowering drilling Liquid density. 2. The viscosity and shear force of the drilling fluid are changed. 3. Adjusting drilling engineering measures (3) XXX well lost circulation work Working fluids (including mud, cement slurry, completion fluids, and others) Liquid, etc.) leaking into the formation under differential pressure Total loss of impuritiesDrilling fluid 30m ³ 1. Reducing drilling fluid Density. 2. The viscosity and shear force of the drilling fluid are changed. 3. Adjustment of Whole drilling engineering measures
2	Sleeve pressure rise	Ground equipment Abnormality; downhole Tool anomalies; formation conditions A change; casing pipe Channeling pressure between them; completion string Leakage of	Inspection floor Apparatus, downhole Tool with no Abnormal conditions; the sleeves are arranged between No channeling pressure Image generation	Replacement of fresh oil pipe Downhole tool, lifting tool High construction quality; add on top of the top seal Top packer	Pressure rise of XXX well overflow and hydrogen embrittlement fracture well closing sleeve of drilling tool High casing pressure 2.3 ↗ 8.6.8.6 MPaXXX well overflow shut-in After closing the well for 2 minutes, the casing pressure is increased to 11MPa garage XXX well overflow shut-in casing pressure rise Closing the well and casing pressure to 7.4MPa, and expanding to 7.7MPa after 5 minutes Generating fluctuation according to the abnormality of ground equipment and underground workers Abnormal and stratum condition change taking treatment measures
						3	Overflow volume	Formation pressure Inaccurate mastering Determining; when the drill is started No filling is carried out in the well Manchurian medicine's strength Also; too large Suction pressure; drilling fluid-tight The degree is low; drilling well Liquid leakage; ground (floor) Different lamination force Often times	Inspection of well bores A liquid return amount, Drilling fluid pool Well drilling fluid An amount of; inspection stop In the back well of the pump The drilling fluid being Whether or not to overflow	After finding overflow Immediately closing the well; at the position of The wellhead is led to a certain degree Is to (1) local resistance Increasing the annular space Pressure of (i.e. well) Back pressure is applied to the mouth.	(1) XXX well overflow volume 0.4m ³ (2) XXX Well overflow quantity well closing process for 6 minutes Stream 24m ³

(3) Static plugging method

Case 1 XXXX well lost circulation

The treatment method comprises the following steps:

matching leakage-stopping mud (leakage-stopping agent while drilling) [ failure ]

Plugging slurry 70m with concentration of 9% in 15:00-20:00 ³ (6T, KH-n 3t, 0.5T, 0.3T SP-8) of plugging agent while drilling, and 29 th of slurry is lost. Complex release of stationary lost circulation [ successful ].

Case 2 XXXX well lost circulation

The treatment method comprises the following steps:

Complex release of static leak stopping [ successful ]

Co-directional annular hanging and grouting 25m during static plugging period of 12:00-19:00 ³ 15% plugging slurry 60m ³ (walnut shell powder: 3t, comprehensive plugging agent: 3t, superfine calcium carbonate: 2t, plugging agent while drilling 3t, natural asphalt powder 2 t), 19:00 down to 3470m, and pumping into the plugging slurry with a 9L/s displacement pump.

Case 3 XXXX well lost circulation

The treatment method comprises the following steps:

【1】 Bridge plugging with plugging slurry (failure)

15% and 20% Density 1.55g/m are used ³ Complicated release of plugging slurry bridge failure [ 2 ] stopping and plugging [ success ]

Static plugging from the start of drilling to 1900 m, and matching 20 percent density of 1.55g/m ³ 60 sides of the slurry are blocked, and the slurry is respectively pressurized at 2200m and 2000m without fruits; after the well is drilled down to 2393 meters of static plugging in a sectional circulation mode, 20% plugging slurry is used for being segmented to the bottom, and the large-displacement well is washed without leakage and is released in a complex mode.

Case 4 XXXX well lost circulation

【1】 Ground supplement plugging agent while drilling [ failure ] ground supplement density40m of plugging agent with the concentration of 1.31g/cm3 and 6% while drilling ³ The method comprises the steps of carrying out a first treatment on the surface of the The plugging agent while drilling with the concentration of 10% is supplemented on the ground.

【2】 Static plugging [ failure ] 2:00 drill is lifted to 2170m static plugging, and the ground supplement density is 1.30g/cm ³ 40m of plugging slurry ³ 。

【3】 And (3) injecting ash (failure) and lifting the drill rod under the drill by a test drill 13:40. Drilling to the depth of 2440 m in the course of 6 months and 4 days and 8:00. And (3) ending the 11:30 ash injection, and losing 4.5m ³ (Ash 18t, front liquid 11.3m, cement paste 15.1m, density 1.73 g/cm) ³ 2.7m of post-treatment liquid, 15.6m of pulp, and 1.30g/cm of density ³ ) 13:30 lifting the drill until 1700m is used for bearing pressure, and setting mud into the drill for 2m ³ The highest pressure is 1.6mpa, the pressure is stabilized at 0.6mpa, and the drilling and the curing are carried out.

Example four

Aiming at user input influencing factors and accident phenomena, a well drilling accident treatment method determines a recommended result, and the implementation process is as follows:

inputting an influence factor fault attribute and an accident phenomenon pump pressure fluctuation attribute, and obtaining a treatment method recommendation result of the drilling accident to be treated by using a treatment method recommendation model, wherein the treatment method recommendation result comprises a possible accident type and a recommendation method.

(1) Type of possible accident

Sequence number	Complex name	Case of association of this type
			1	Lost circulation	【1】 XXX well lost circulation [ 2 ] XXX oil zone XX1 well lost circulation, overflow
2	Overflow flow	【1】 XXX oil zone XX1 well lost circulation and overflow

(2) Static plugging method

Case 1 XXX well lost circulation

The treatment method comprises the following steps:

static plugging in technical sleeve [ failure ] is complicated, 12 columns of the drill are not lifted into the technical sleeve, static plugging is achieved, meanwhile, 40m of plugging slurry with the concentration of 8% (1 t walnut shell 1t vermiculite 2 t) is matched on the ground, and 10m of slurry is hung and irrigated during the process.

And (3) starting to lift the drill to stop leakage in the technical sleeve at 20:30, supplementing 40m of slurry (1.50 g/cm < 3 > PMHA-2.3 t JT8880.4t caustic soda 0.2t HY-2.4 t) and hoisting the slurry for 6m during the process, wherein the pressure is 21:30 Guan Jingbie.

Case 2 XXX well lost circulation, overflow

The treatment method comprises the following steps:

pumping well control fluid from the annulus to press the overflow into the wellbore back into the formation [ successful ]

The well is pumped by a 1:30 pressure return method after 11 months and 9 days, and 2.25g/cm is reversely extruded by a slurry pump after 23:10-0:30 ³ ×64m ³ 3 ↗ 6 ↗ 9 ↗ 20l/s displacement, 7.5 ↗ MPa casing pressure 14 ↘ MPa pumping pressure 9 ↗ 12 ↘ 4MPa casing pressure reduced to 0 after stopping pumping; forward extrusion 2.23g/cm 3X 24m at 1:20 for 0:45-10 days ³ The discharge capacity is 9l/s, the vertical pressure is 20 ↘ 18.5.5 MPa, and the vertical pressure is 0 after the pump is stopped. Annulus liquid level 682-642m, observing well opening, and hanging and filling 1.97g/cm in annulus every 30min ³ ×0.5m ³ Hanging irrigation 1.97g/cm for each 1h water hole ³ ×0.5m ³ The annular liquid level is 507-309m, and the water eye liquid level is 199-355m. When the drill is started, the water hole is blocked, and the pressure is not blocked.

Case 3 XXX well lost circulation, overflow

The treatment method comprises the following steps:

drilling fluid circulation (successful) of well closing casing pressure within a required range is controlled by adjusting a choke manifold valve

19:15 throttle cycle, displacement 8-9l/s, pumpThe pressure is 19-20MPa, and the inlet density is 2.10g/cm ³ The outlet density was 2.03 ↘, 1.96 ↗ 2.07.07 g/cm ³ Igniting the ignition cylinder in the ratio of 14:40 to 17:22, igniting the ignition cylinder with flame of 4 to 5m, and pumping the ignition cylinder into the ignition cylinder with the flame of 2.10g/cm ³ ×238m ³ Return to 203m ³ Stopping the pump, and opening the well under the pressure of 0.

According to the example, based on the established drilling accident handling scheme recommendation model, possible types and possible reasons of the accident are intelligently analyzed according to the user input attribute, the accident handling method and related cases are recommended, timely and effective data support is provided for analysis and research of the accident scene, and the accident handling efficiency is improved.

Those of ordinary skill in the art will appreciate that all or some of the steps, systems, functional modules/units in the apparatus, and methods disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof. In a hardware implementation, the division between the functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may be performed cooperatively by several physical components. Some or all of the components may be implemented as software executed by a processor, such as a digital signal processor or microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on computer readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). The term computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, as known to those skilled in the art. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can be accessed by a computer. Furthermore, as is well known to those of ordinary skill in the art, communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media.

Claims

1. A method of recommending a drilling event disposition, the method comprising:

acquiring attributes of one or more drilling incidents to be treated;

obtaining a recommended result of the treatment scheme of the drilling accident to be treated according to the attribute of one or more drilling accidents to be treated and by utilizing the recommended model of the treatment scheme;

the recommendation model of the treatment scheme is established according to the drilling accident knowledge body; the drilling accident knowledge body is established according to related theoretical knowledge of the drilling accident and historical treatment cases of the drilling accident;

the recommended model building process of the treatment scheme comprises the following steps:

2. The method of recommending a drilling event disposal facility according to claim 1, wherein,

3. The method of recommending a drilling event disposal facility according to claim 2, wherein,

the accident complex type knowledge body comprises one or more of the following attributes: accident complexity cause attribute, accident complexity phenomenon attribute, precaution measure attribute and disposal method attribute;

4. A method of recommending a drilling event disposal facility according to claim 3, wherein,

the theoretical knowledge label in the drilling accident label system comprises: accident complex type tags, accident phenomenon tags, influencing factor tags and disposal method tags;

5. The method of recommending a drilling incident treatment plan of claim 4, wherein the generating a theoretical knowledge feature matrix based on theoretical knowledge tags in the drilling incident tag system comprises:

6. A recommended method of drilling incident treatment plan as claimed in claim 3, wherein the generating a treatment case feature matrix based on historical treatment case tags in the drilling incident tag system comprises:

7. The method of recommending a drilling event disposal plan according to claim 1, wherein the obtaining a disposal method recommendation for a drilling event to be disposed according to the attribute of one or more drilling events to be disposed and using the recommended model of the disposal plan comprises:

calculating the similarity of each element in the query vector and the feature matrix row vector;

8. The method of claim 7, wherein determining a row vector corresponding to a similarity value of TopN preceding the similarity row generates a recommendation, comprising:

9. The method of recommending drilling event handling scenarios according to claim 8, wherein generating a likely event type MayTypes from the event complex type input tag TypeFlag comprises:

wherein, the score formula is:

；

in the above-mentioned formula(s),representing the accident complex type score,i'representing rows in the recommendation result matrix, +.>Column representing recommendation result matrix, 0</>K1 is less than or equal to; topN represents the number of rows of the recommendation result matrix.

10. A method of recommending drilling event dispositions according to claim 9, wherein generating possible event events maysign from the event input signature SignFlag comprises:

wherein, the score formula is:

；

in the above-mentioned formula(s),a score representing the event of an accident,i'representing rows in the recommendation result matrix, +. >TopN represents the number of rows, K1, of the recommendation result matrix</>And a is a positive integer which is less than or equal to (K1+K2).

11. The method of claim 10, wherein generating a probable accident phenomenon mayfactor based on the influence factor input flag FactorFlag comprises:

wherein, the score formula is:

；

in the above-mentioned formula(s),a score representing the cause of the incident,i'representing rows in the recommendation result matrix, +.>TopN represents the number of rows of the recommendation result matrix, (K1+K2)</>≤ (K1+K2+K3)。

12. A method of recommending a well drilling accident management program according to claim 11, wherein the accident-complex-type input signature, the accident-event input signature, the accident-influencing-factor input signature, and the recommendation-result matrix M are based on the accident-complex-type input signature _TopN Generating a recommendation result, comprising:

wherein,

；

in the above-mentioned calculation formula, the calculation formula,is a score of (a) and (b),i'representing rows in the recommendation result matrix, +.>TopN represents the number of rows of the recommendation result matrix, (K1+K2+K3)</>≤ (K1+K2+K3+K4)。

13. A recommendation device for a drilling accident management program, the device comprising: a memory and a processor; the memory is for storing a program for performing a recommended method of a drilling accident treatment plan, the processor is for reading the program for performing the recommended method for performing a drilling accident treatment plan, and performing the method of any one of claims 1-12.