Method and apparatus for drilling difficulty prediction and difficulty-solution generation, and device

Abstract

Embodiments of the present description relate to the technical field of artificial intelligence, and provide a method and apparatus for drilling difficulty prediction and difficulty-solution generation, and a device. The method comprises: on the basis of location information of a well point to be predicted, determining a list of neighboring wells of the well point to be predicted and a corresponding well-history knowledge graph; on the basis of the well-history knowledge graph, calculating a formation feature vector of each neighboring well in each formation, to obtain a formation feature vector of the well point to be predicted in each formation; and then on the basis of a pre-constructed difficulty-solution knowledge graph, determining a geological condition label, a drilling difficulty label and a difficulty-solution label of the well point to be predicted in each formation, and inputting same into a pre-trained drilling difficulty analysis large model, to obtain a drilling difficulty prediction text of the well point to be predicted in each formation and a corresponding difficulty-solution text.

Description

A drilling difficulty prediction and difficulty solution generation method, device and equipment

[0001] Related Applications

[0002] The present application claims priority to the Chinese patent application No. 202411388944.0 filed on September 30, 2024, and incorporates by reference the entire disclosure of the above patent application as part of the present application. TECHNICAL FIELD

[0003] Embodiments of the present specification relate to the field of artificial intelligence technology, in particular to a drilling difficulty prediction and difficulty solution generation method, device and equipment. BACKGROUND

[0004] Petroleum drilling engineering is an important means to achieve the goal of oil and gas exploration and development. Through drilling, one can go deep into the underground to explore and evaluate the existence and reserves of oil and gas resources. With the expansion of drilling operations to deep wells, ultra-deep wells and other areas, the geological conditions are more complex, and problems such as ultra-high temperature and ultra-high pressure, multiple pressure systems, hard formation and poor drillability, and rich acid fluids greatly increase the difficulty of drilling operations. At the same time, some exploration areas have less drilled well data and less drilling experience to learn from, and drilling accidents occur from time to time, which seriously affects the progress of exploration and development.

[0005] Currently, drilling difficulty analysis and difficulty solution formulation mainly rely on adjacent well data and expert experience, and lack intelligent analysis technology. The existing data-driven method for drilling risk identification and prediction lacks professional knowledge and experience embedding, and the model accuracy needs to be improved. Case-based reasoning and recommendation algorithms can recommend difficulty solutions, but the recommended results are only a copy of historical solutions, and cannot fuse and summarize the current well information and experience knowledge to generate targeted difficulty solutions. Therefore, there is an urgent need for a drilling difficulty prediction and difficulty solution generation method to accurately predict drilling difficulties and automatically generate corresponding difficulty solutions. SUMMARY

[0006] In view of the above problems of the prior art, the purpose of the embodiments of the present specification is to provide a drilling difficulty prediction and difficulty solution generation method, device and equipment to accurately predict drilling difficulties and automatically generate corresponding difficulty solutions.

[0007] To solve the above technical problems, the specific technical solutions of the embodiments of the present specification are as follows:

[0008] On the one hand, the embodiments of the present specification provide a drilling difficulty prediction and difficulty solution generation method, which comprises:

[0009] obtaining well point position information to be predicted;

[0010] determine a neighbor well list of the well point to be predicted and a corresponding well history knowledge graph according to the well point to be predicted position information, wherein the well history knowledge graph is pre-constructed using historical drilling data and includes formation characteristic data of different formations and relationships therebetween;

[0011] calculate a formation characteristic vector of each neighbor well in the neighbor well list in each formation according to the formation characteristic data in the well history knowledge graph;

[0012] calculate a formation characteristic vector of the well point to be predicted in each formation according to the formation characteristic vector of each neighbor well in the neighbor well list in each formation;

[0013] determine a geological condition label, a drilling difficulty label and a difficulty solution label of the well point to be predicted in each formation according to the formation characteristic vector of the well point to be predicted in each formation and a pre-constructed difficulty solution knowledge graph, wherein the difficulty solution knowledge graph includes geological condition class nodes, drilling difficulty class nodes, difficulty solution class nodes and relationships among the three types of nodes;

[0014] construct a prompt text according to the geological condition label, the drilling difficulty label and the difficulty solution label and input the prompt text into a pre-trained drilling difficulty analysis large model to obtain a drilling difficulty prediction text and a corresponding difficulty solution text of the well point to be predicted in each formation.

[0015] Further, the determination of the neighbor well list of the well point to be predicted according to the well point to be predicted position information includes:

[0016] determine a block to which the well point to be predicted belongs according to the well point to be predicted position information;

[0017] calculate a well spacing between the well point to be predicted and each well point in the block;

[0018] determine a well point with a well spacing less than a preset threshold as a neighbor well of the well point to be predicted to obtain the neighbor well list of the well point to be predicted.

[0019] Further, the calculation of the formation characteristic vector of each neighbor well in the neighbor well list in each formation according to the formation characteristic data in the well history knowledge graph includes:

[0020] perform representation learning on the well history knowledge graph according to the formation characteristic data to obtain formation characteristic vector data;

[0021] obtain a first formation characteristic vector and a second formation characteristic vector of a certain neighbor well in a certain formation according to the formation characteristic vector data;

[0022] fuse the first formation feature vector and the second formation feature vector to obtain a third formation feature vector of the certain adjacent well in a certain formation;

[0023] calculate formation feature vectors of an adjacent upper formation and an adjacent lower formation of the certain adjacent well in a certain formation;

[0024] fuse the third formation feature vector, the formation feature vectors of the adjacent upper formation and the adjacent lower formation to obtain a fourth formation feature vector of the certain adjacent well in a certain formation through nonlinear calculation;

[0025] obtain a well list of the certain adjacent well, and calculate a distance influence coefficient of each adjacent well in the well list of the certain adjacent well and a formation feature vector of each adjacent well in the certain formation;

[0026] calculate a final formation feature vector of the certain adjacent well in a certain formation according to the fourth formation feature vector, the distance influence coefficient and the formation feature vector of each adjacent well in the certain formation in the well list of the certain adjacent well;

[0027] repeat the above steps until a formation feature vector of each adjacent well in each formation in the well list is obtained.

[0028] Further, the calculation of the final formation feature vector of the certain adjacent well in a certain formation according to the fourth formation feature vector, the distance influence coefficient and the formation feature vector of each adjacent well in the certain formation in the well list of the certain adjacent well comprises:

[0029] the final formation feature vector of the certain adjacent well in a certain formation is calculated by the following formula:

[0030] wherein, E_W i S j represents a final formation feature vector of the i th adjacent well in the j th formation, E_W i S j _2 represents a fourth formation feature vector of the i th adjacent well in the j th formation, n1 represents the number of adjacent wells of the i th adjacent well, β k represents a distance influence coefficient of the i th adjacent well and the k th adjacent well in the well list of the i th adjacent well, E_W k S j _2 represents a formation feature vector of the k th adjacent well in the well list of the i th adjacent well in the j th formation.

[0031] Further, the calculation of the final formation feature vector of the certain adjacent well in a certain formation according to the fourth formation feature vector, the distance influence coefficient and the formation feature vector of each adjacent well in the certain formation in the well list of the certain adjacent well comprises:

[0032] The formation feature vector of the to-be-predicted well point in each formation is calculated by using the following formula:

[0033] wherein, E_W Loc S j denotes the formation feature vector of the to-be-predicted well point in the jth formation, n2 denotes the number of adjacent wells of the to-be-predicted well point, ρ i denotes the distance influence coefficient of the to-be-predicted well point and the ith adjacent well, E_W i S j denotes the formation feature vector of the ith adjacent well in the jth formation.

[0034] Further, the geological condition label, the drilling difficulty label and the difficulty solution label of the to-be-predicted well point in each formation are determined according to the formation feature vector of the to-be-predicted well point in each formation and the pre-constructed difficulty solution knowledge graph, comprising:

[0035] The difficulty solution knowledge graph is subjected to representation learning to obtain a geological condition feature vector corresponding to each geological condition class node in the difficulty solution knowledge graph;

[0036] The similarity between the formation feature vector of the to-be-predicted well point in each formation and each geological condition feature vector in the difficulty solution knowledge graph is calculated;

[0037] The labels corresponding to the first preset number of geological condition feature vectors with a similarity greater than a preset threshold are taken as the geological condition label of the to-be-predicted well point in each formation;

[0038] The drilling difficulty label and the difficulty solution label of the to-be-predicted well point in each formation are determined from the difficulty solution knowledge graph according to the geological condition label of the to-be-predicted well point in each formation.

[0039] Further, the drilling difficulty prediction text and the corresponding difficulty solution text of the to-be-predicted well point in each formation are obtained by inputting the geological condition label, the drilling difficulty label and the difficulty solution label into a pre-trained drilling difficulty analysis large model, comprising:

[0040] The geological condition label, the drilling difficulty label and the difficulty solution label are filled into the placeholders corresponding to the prompt text template to obtain a prompt text;

[0041] The prompt text is input into the pre-trained drilling difficulty analysis large model to generate the drilling difficulty prediction text and the corresponding difficulty solution text of the to-be-predicted well point in each formation.

[0042] In another aspect, the embodiments of the present specification provide a drilling difficulty prediction and difficulty solution generation device, the device comprising:

[0043] An acquisition module is configured to acquire well point position information to be predicted.

[0044] A neighboring well list determination module is configured to determine a neighboring well list of the well point to be predicted and a corresponding well history knowledge graph according to the well point position information to be predicted, wherein the well history knowledge graph is pre-constructed using historical drilling data and comprises formation feature data of different formations and relationships therebetween.

[0045] A neighboring well formation feature vector calculation module is configured to calculate a formation feature vector of each neighboring well in the neighboring well list at each formation according to the formation feature data in the well history knowledge graph.

[0046] A well point to be predicted formation feature vector calculation module is configured to calculate a formation feature vector of the well point to be predicted at each formation according to the formation feature vector of each neighboring well in the neighboring well list at each formation.

[0047] A label determination module is configured to determine a geological condition label, a drilling difficulty label and a difficulty solution label of the well point to be predicted at each formation according to the formation feature vector of the well point to be predicted at each formation and a pre-constructed difficulty solution knowledge graph, wherein the difficulty solution knowledge graph comprises geological condition class nodes, drilling difficulty class nodes, difficulty solution class nodes and relationships among the three types of nodes.

[0048] A drilling difficulty prediction and difficulty solution generation module is configured to construct a prompt text according to the geological condition label, the drilling difficulty label and the difficulty solution label, input the prompt text into a pre-trained drilling difficulty analysis large model, and obtain a drilling difficulty prediction text and a corresponding difficulty solution text of the well point to be predicted at each formation.

[0049] In yet another aspect, the embodiments of the present specification further provide a computer device comprising a memory, a processor and a computer program stored in the memory, wherein the computer program is executed by the processor to perform the instructions of any of the above methods.

[0050] In yet another aspect, the embodiments of the present specification further provide a computer readable storage medium having a computer program stored thereon, wherein the computer program is executed by a processor of a computer device to perform the instructions of any of the above methods.

[0051] In yet another aspect, the embodiments of the present specification further provide a computer program product, wherein the computer program product is executed by a processor of a computer device to perform the instructions of any of the above methods.

[0052] By adopting the drilling difficulty prediction and difficulty solution generation method provided in the embodiments of the present specification, the position information of the well point to be predicted can be used to determine the adjacent well list and the corresponding well history knowledge graph, and then the formation feature vector of each adjacent well in each formation in the adjacent well list can be calculated according to the formation feature data in the well history knowledge graph, so that the formation feature vector of the well point to be predicted in each formation is calculated according to the formation feature vector of each adjacent well in each formation. Thus, the formation feature vector of the well point to be predicted is obtained according to the formation feature vector of the adjacent well, so that the formation feature vector of the well point to be predicted is fused with the formation feature of the adjacent well and its own feature, thereby improving the accuracy and reliability of the formation feature of the well point to be predicted. Then, the geological condition label, drilling difficulty label and difficulty solution label of the well point to be predicted in each formation are determined according to the formation feature vector of the well point to be predicted in each formation and the pre-constructed difficulty solution knowledge graph, and input into the pre-trained drilling difficulty analysis large model to obtain the drilling difficulty prediction text and the corresponding difficulty solution text of the well point to be predicted in each formation. In this way, the features of the well point to be predicted and the historical drilling experience knowledge can be fused, and detailed drilling difficulty analysis text and corresponding difficulty solution text can be automatically generated to provide drilling guidance scheme for exploration personnel to perform actual drilling, thereby reducing the drilling operation risk, shortening the drilling cycle and improving the exploration and development efficiency.

[0053] The above description is only a summary of some technical solutions of the embodiments of the present specification. In order to more clearly understand the technical means of some embodiments of the present specification, the above and other purposes, characteristics and advantages of the embodiments of the present specification can be implemented according to the content of the specification, and the following preferred embodiments are described in detail below, and the accompanying drawings are described as follows. BRIEF DESCRIPTION OF DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present specification or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are only some embodiments of the present specification, and those skilled in the art can also obtain other drawings according to these drawings without creating any creative labor.

[0055] FIG. 1 shows a flowchart of a drilling difficulty prediction and difficulty solution generation method according to some embodiments of the present specification;

[0056] FIG. 2 shows a flowchart of determining the adjacent well list of the well point to be predicted according to some embodiments of the present specification;

[0057] FIG. 3 shows a flowchart of calculating the formation feature vector of each adjacent well in the adjacent well list of the well point to be predicted according to some embodiments of the present specification;

[0058] FIG. 4 shows a flowchart of determining the geological condition label, drilling difficulty label and difficulty solution label of each formation for the well point to be predicted in some embodiments of the present specification;

[0059] FIG. 5 shows a flowchart of generating drilling difficulty prediction text and corresponding difficulty solution text in some embodiments of the present specification;

[0060] FIG. 6 shows a schematic diagram of the module structure of a drilling difficulty prediction and difficulty solution generation device in some embodiments of the present specification;

[0061] FIG. 7 shows a schematic diagram of the structure of a computer device according to an embodiment of the present specification. DETAILED DESCRIPTION

[0062] The technical solutions in the embodiments of the present specification will be described clearly and completely below with reference to the drawings in the embodiments of the present specification. Obviously, the described embodiments are only some of the embodiments of the present specification, but not all the embodiments. Based on the embodiments in the present specification, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present specification.

[0063] To solve the above problems, the present specification provides a drilling difficulty prediction and difficulty solution generation method. FIG. 1 is a flowchart of a drilling difficulty prediction and difficulty solution generation method according to an embodiment of the present specification. The present specification provides the method operation steps as described in the embodiments or flowcharts, but more or fewer operation steps can be included based on conventional or non-creative labor. The order of the steps listed in the embodiments is only one of the many execution orders, and does not represent the only execution order. In actual system or device product execution, the method order shown in the embodiments or drawings can be executed in sequence or in parallel.

[0064] It should be noted that the terms "first", "second", and the like in the present specification and claims and the above drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present specification described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, device, product or equipment that includes a series of steps or units does not necessarily limit to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or equipment.

[0065] With reference to FIG. 1, the embodiment of the present specification provides a drilling difficulty prediction and difficulty solution generation method, the method comprising:

[0066] S101: acquiring well point position information to be predicted;

[0067] S102: determining a neighbor well list of the well point to be predicted and a corresponding well history knowledge graph according to the well point position information to be predicted, wherein the well history knowledge graph is pre-constructed using historical drilling data and comprises stratum characteristic data of different strata and relationships therebetween;

[0068] S103: calculating a stratum characteristic vector of each neighbor well in the neighbor well list in each stratum according to the stratum characteristic data in the well history knowledge graph;

[0069] S104: calculating a stratum characteristic vector of the well point to be predicted in each stratum according to the stratum characteristic vector of each neighbor well in the neighbor well list in each stratum;

[0070] S105: determining a geological condition label, a drilling difficulty label and a difficulty solution label of each stratum of the well point to be predicted according to the stratum characteristic vector of each stratum of the well point to be predicted and a pre-constructed difficulty solution knowledge graph, wherein the difficulty solution knowledge graph comprises a geological condition class node, a drilling difficulty class node, a difficulty solution class node and relationships among the three types of nodes;

[0071] S106: constructing a prompt text according to the geological condition label, the drilling difficulty label and the difficulty solution label and inputting the prompt text into a pre-trained drilling difficulty analysis large model to obtain a drilling difficulty prediction text and a corresponding difficulty solution text of each stratum of the well point to be predicted.

[0072] The method for predicting drilling difficulties and generating difficulty solution schemes provided in the embodiments of the present specification can determine the neighboring well list and the corresponding well history knowledge graph of the to-be-predicted well point according to the to-be-predicted well point position information, and then calculate the formation feature vector of each neighboring well in each formation in the neighboring well list according to the formation feature data in the well history knowledge graph, so as to calculate the formation feature vector of the to-be-predicted well point in each formation according to the formation feature vector of each neighboring well in each formation. Thus, the formation feature vector of the to-be-predicted well point is obtained according to the formation feature vector of the neighboring well, so that the formation feature vector of the to-be-predicted well point fuses the formation features of the neighboring well and its own features, thereby improving the accuracy and reliability of the formation features of the to-be-predicted well point. Then, the geological condition label, the drilling difficulty label and the difficulty solution scheme label of the to-be-predicted well point in each formation are determined according to the formation feature vector of the to-be-predicted well point in each formation and the pre-constructed difficulty solution scheme knowledge graph, and are input into the pre-trained drilling difficulty analysis large model to obtain the drilling difficulty prediction text and the corresponding difficulty solution scheme text of the to-be-predicted well point in each formation. In this way, the features of the to-be-predicted well point and the historical drilling experience knowledge can be fused, and detailed drilling difficulty analysis text and corresponding difficulty solution scheme text can be automatically generated to provide drilling guidance schemes for explorers to perform actual drilling, thereby reducing the drilling operation risk, shortening the drilling cycle and improving the exploration and development efficiency.

[0073] In the embodiments of the present specification, the to-be-predicted well point position information includes the position coordinates of the to-be-predicted well point. Referring to FIG. 2, the neighboring well list of the to-be-predicted well point is determined according to the to-be-predicted well point position information, including:

[0074] S201: determining the block to which the to-be-predicted well point belongs according to the to-be-predicted well point position information;

[0075] S202: calculating the well spacing between the to-be-predicted well point and each well point in the block to which the to-be-predicted well point belongs;

[0076] S203: determining the well points with well spacing less than a preset threshold as the neighboring wells of the to-be-predicted well point to obtain the neighboring well list of the to-be-predicted well point.

[0077] It can be understood that, in the process of oil exploration and development, the geological conditions of the adjacent well of the to-be-predicted well point are closely related to the geological conditions thereof. For example, the geological data of the adjacent well, such as lithology, structure, formation thickness, reservoir physical property and the like, can all provide important reference basis for the exploration and development of the to-be-predicted well point. By comparing the geological conditions of the adjacent well and the to-be-predicted well point, the similarity and difference of the geological features can be identified, and the law of geological evolution in the block to which the to-be-predicted well point belongs can be further analyzed, thereby providing basis for geological modeling and reservoir prediction. In addition, based on the drilling experience of the adjacent well, the risks that the to-be-predicted well point may encounter in the drilling process, such as abnormal formation pressure, blowout, well collapse and the like, can be evaluated, so that countermeasures can be formulated. Generally, since the geological conditions of the well points in the same block are relatively similar, in the embodiments of the present specification, the block to which the to-be-predicted well point belongs is first determined through the position coordinates of the to-be-predicted well point, and then the adjacent well is determined by calculating the well spacing between the to-be-predicted well point and the remaining well points in the block and the preset threshold, and finally the adjacent well list is formed. Exemplarily, in order to ensure the reliability of the formation features of the subsequent to-be-predicted well point, the preset threshold in the embodiments of the present specification is 5 km.

[0078] In step S102, the well history knowledge graph corresponding to the adjacent well list is constructed according to historical drilling data. Specifically, the well history knowledge graph construction process includes the following steps:

[0079] Step a: Obtain historical drilling data, which is derived from the engineering technical data of drilled wells, mainly including structured data and unstructured data. The structured data is mainly stored in a relational database, such as block, wellbore, formation, construction record, drill bit usage record, drilling fluid performance, drilling assembly data, logging data, coring data, time efficiency data, etc. The unstructured data is mainly stored in the file system in the form of pdf, word, etc., such as geological design report, drilling design report, construction summary report, case analysis report, and drilling-related policies, regulations, standards, implementation rules, technical manuals, etc.

[0080] Step b: Establish a single-well knowledge ontology model (i.e., the knowledge graph concept layer) according to the historical drilling data, mainly including blocks, wells, drilled formation features, different opening construction schemes, different well section drilling time efficiency, etc. entities, entity attributes and relationships between entities.

[0081] Step c: Extract the corresponding knowledge instances of the concept layer from the internal relational database using knowledge instance technology such as body mapping, R2RML (RDB (Relational Database) to RDF (Resource Description Framework) Mapping Language), D2RQ (Direct Database to RDF Querying), etc. to generate the well history knowledge graph. According to the single well knowledge ontology model, the missing entity examples, attribute examples and relationship examples in the well history knowledge graph are counted, and the knowledge instances are extracted from the unstructured documents such as single well geological design and single well engineering design by using natural language processing technology, and the well history knowledge graph is completed.

[0082] Step d: Characteristic learning of well history knowledge graph. Since most of the nodes and relationships in the well history knowledge graph are discrete and sparse data types, in order to facilitate calculation and reasoning, knowledge graph embedding technologies such as TransR, TransD and DistMult are used to map the nodes and relationships in the well history knowledge graph to a low-dimensional, dense numerical vector space, and finally generate feature vector data of different blocks, different wells, different formations and different construction schemes.

[0083] In some embodiments of the present specification, referring to FIG. 3, the formation feature vector of each adjacent well in each formation in the adjacent well list is calculated according to the formation feature data in the well history knowledge graph, including:

[0084] S301: Characteristic learning of well history knowledge graph according to formation feature data, to obtain formation feature vector data.

[0085] S302: Obtain the first formation feature vector and the second formation feature vector of a certain adjacent well in a certain formation according to the formation feature vector data.

[0086] Specifically, the first formation feature vector is the global formation feature vector of a certain adjacent well in a certain formation, denoted as E_W i S j _0, and the second formation feature vector is the local formation feature vector of a certain adjacent well in a certain formation, such as geological structure vector x1, deposition vector x2, lithology vector x3, temperature and pressure vector x4, etc. q vectors representing formation features, denoted as X={x1,x2,x3,x4,.....,x q}。

[0087] S303: Fuse the first formation feature vector and the second formation feature vector to obtain the third formation feature vector of a certain adjacent well in a certain formation.

[0088] It can be understood that by fusing the global stratum feature vector and the local stratum feature vector, the stratum features of a certain adjacent well in a certain stratum can be represented from multiple angles and levels, so that more comprehensive and rich information can be captured, thereby improving the accuracy and reliability of subsequent acquisition of corresponding geological labels according to the to-be-predicted well point. In some embodiments, the average method, weighted average method, KCNN (Knowledge-aware Convolutional Neural Network) and the like can be used for feature vector fusion. For example, the average method is used to obtain a fused stratum feature vector, i.e., a third stratum feature vector, as shown below:

[0089] wherein E_W i S j _1 represents the third stratum feature vector of the i th adjacent well in the j th stratum, E_W i S j _0 represents the first stratum feature vector of the i th adjacent well in the j th stratum, q represents the number of local stratum feature vectors, and x q represents the q th local stratum feature vector.

[0090] S304: Calculate the stratum feature vectors of the adjacent upper and lower strata of a certain adjacent well in a certain stratum.

[0091] Specifically, the stratum feature vectors of the adjacent upper and lower strata of a certain adjacent well in a certain stratum can be calculated by using the calculation methods of steps S302-S303.

[0092] S305: Perform nonlinear calculation on the third stratum feature vector and the stratum feature vectors of the adjacent upper and lower strata to obtain a fourth stratum feature vector of a certain adjacent well in a certain stratum.

[0093] It can be understood that the geological conditions between adjacent strata have high similarity, and by performing nonlinear calculation on the third stratum feature vector and the stratum feature vectors of the adjacent upper and lower strata, the complex correlation between the features of the current stratum and the adjacent strata can be captured, and a new feature vector containing more rich information can be generated, so that a more comprehensive stratum feature vector of a certain adjacent well in a certain stratum can be obtained. In some embodiments, the nonlinear calculation can be performed by the following formula: E_W i S j _2 = ActFunc(w1*E_W i S j _1 + w2*E_W i S j U_1 + w3*E_W i S jD_1) (2)

[0094] Among them, E_W i S j _2 represents the fourth formation feature vector of the i-th adjacent well in the j-th formation, ActFunc() represents the activation function, such as RelU or LeaklyReLU, w1, w2, and w3 are trainable weights, and E_W i S j _1 represents the third formation feature vector of the i-th adjacent well in the j-th formation, E_W i S j U_1 represents the feature vector of the adjacent upper stratum of the i-th well in the j-th stratum, and E_W i S j D_1 represents the feature vector of the adjacent lower strata of the i-th well in the j-th stratum.

[0095] S306: Obtain the list of neighboring wells of a certain neighboring well, calculate the distance influence coefficient between each neighboring well in the list of neighboring wells and the formation feature vector of each neighboring well in a certain formation.

[0096] Specifically, the list of neighboring wells of the i-th well is represented as W. i List = {W1, W2, ... W} k ,......}, for W i Each well in the list W k (k = 1, 2, ...), calculate W k With W i well spacing d k,i Then calculate W using the following formula. k With W i Distance influence coefficient β k : β k =(d k,max -d k,i ) / (d k,max -d k,min (3)

[0097] Where, β k d represents the distance influence coefficient between the i-th adjacent well and the k-th adjacent well in the list of adjacent wells. k,i W k With W i well spacing, d k,max d represents the maximum well spacing. k,min This represents the minimum well spacing. The distance influence coefficient characterizes the degree of influence of neighboring wells on the formation characteristics of this well; the larger the distance influence coefficient, the greater the influence of neighboring wells on the formation of this well.

[0098] The stratum feature vector of each neighbor well of the i th neighbor well of the well point to be predicted in each stratum can be calculated by using the calculation method of steps S302-S303.

[0099] S307: The final stratum feature vector of each neighbor well of the i th neighbor well of the well point to be predicted in each stratum is calculated according to the fourth stratum feature vector, the distance influence coefficient, and the stratum feature vector of each neighbor well of the neighbor well list of the neighbor well in each stratum.

[0100] Specifically, the final stratum feature vector of each neighbor well of the i th neighbor well of the well point to be predicted in each stratum is calculated by using the following formula in the embodiments of the present specification:

[0101] wherein, E_W i S j represents the final stratum feature vector of the i th neighbor well in the j th stratum, E_W i S j _2 represents the fourth stratum feature vector of the i th neighbor well in the j th stratum, n1 represents the number of neighbor wells of the i th neighbor well, β k represents the distance influence coefficient between the i th neighbor well and the k th neighbor well in the neighbor well list of the i th neighbor well, E_W k S j _2 represents the stratum feature vector of the k th neighbor well in the neighbor well list of the i th neighbor well in the j th stratum.

[0102] S308: The above steps are repeated until the stratum feature vector of each neighbor well in each stratum in the neighbor well list is obtained.

[0103] In this way, by fusing the feature vectors of the adjacent upper stratum, the adjacent lower stratum, and the neighbor well stratum of the i th neighbor well in the j th stratum, the feature vectors of multiple strata can be integrated, the omission of feature information caused by considering only a single stratum feature vector is avoided, the stratum feature of the i th neighbor well in the j th stratum can be more comprehensively described, and further accurate and reliable support can be provided for subsequent calculation of the stratum feature vector of each stratum of the well point to be predicted.

[0104] In some embodiments of the present specification, the stratum feature vector of the well point to be predicted in each stratum is calculated by using the following formula in step S104:

[0105] wherein, E_W Loc S j represents the stratum feature vector of the well point to be predicted in the j th stratum, n2 represents the number of neighbor wells of the well point to be predicted, ρ i represents the distance influence coefficient between the well point to be predicted and the i th neighbor well, E_W i S jdenotes the formation feature vector of the i-th adjacent well in the j-th formation. Thus, the formation feature vector of the predicted well point in each formation can be obtained by weighting and summing the formation feature vectors of the adjacent wells of the predicted well point according to the distance influence coefficient, thereby improving the accuracy of calculating the formation feature vector of each formation of the predicted well point and providing accurate and reliable data support for subsequent acquisition of the corresponding geological condition label according to the formation feature vector.

[0106] In step S105, the difficult point solution knowledge graph is formed from book documents, drilling logs and other data, and knowledge points such as geological conditions (such as piedmont structural belt, fault, salt and gypsum layer, etc.), drilling difficulties (such as coexistence of blowout and leakage, narrow safety density window, low mechanical drilling speed, etc.), difficult point solutions (such as high and steep structure broken hole stability technology, dense sandstone fractured gas layer protection technology, etc.) and the like. The main construction process includes the following steps:

[0107] Step a: Establish a drilling difficulty and difficult point solution label system. By establishing a label system, it is helpful for drilling engineers and other business experts to comprehensively master the complex geological conditions, drilling engineering difficulties and difficult point solutions of each block. The label system mainly includes geological labels, drilling difficulty labels and difficult point solution labels, as shown in Table 1.

[0108] Table 1

[0109] Step b: Establish a drilling difficulty and difficult point solution corpus. Specifically, first, taking “drilling difficulty and difficult point solution” as the theme, relevant data is obtained from the Internet, books, documents, reports and other sources, and the data is cleaned, de-duplicated, segmented and the like to form a text data set. For unstructured documents, image recognition, OCR, GROBID and other technical tools can be used to identify the document structure, extract chapter content, and further extract semantic relationships and cross relationships between texts. Then, by using keyword matching, text classification, large models and the like, the text content representing the drilling geological conditions, drilling difficulties or difficult point solutions is screened from the above text data set to form a drilling difficulty and difficult point solution corpus. Illustratively, by formulating prompt words, the relevant content can be extracted from the text data set by using a large model.

[0110] Step c: Establish an entity relationship extraction model of drilling difficulties and technical countermeasures. First, divide the corpus in step b into three parts according to a certain proportion (for example, 1:2:7, etc.): corpus1, corpus2, and corpus3. For corpus1, manually annotate the corresponding labels and the relationships between the labels for each text. The relationships between the labels include two types: the relationship between geological conditions and drilling difficulties, and the relationship between drilling difficulties and difficulty solutions. For corpus2, first use text classification technology to select matching instructions from a pre-set instruction set for each text. Then, use knowledge extraction tools to automatically extract entity relationships such as geological conditions, drilling difficulties, and difficulty solutions. After auditing and verifying the extracted json results, generate the corpus2 annotation dataset. Based on the annotated data of corpus1 and corpus2, build an entity relationship extraction dataset. Use relationship extraction models such as ETL-span, HBT, and SpERT to train on this dataset to generate an entity relationship extraction model. Divide corpus3 into n batches. For each batch, perform the following operations: use the entity relationship extraction model to identify all text content in the batch and extract entities such as geological conditions, drilling difficulties, and difficulty solutions, as well as their relationships. Randomly select a portion of the identified results and calculate the model accuracy. If the accuracy does not meet the requirements, add the misidentified data to corpus1 for manual annotation to further expand the dataset, optimize the entity relationship extraction, and apply the optimized entity relationship extraction model to the next batch. If the accuracy meets the requirements, perform the next batch.

[0111] Step d: Import the manual annotation results of corpus1, the large model extraction results of corpus2, and the entity relationship extraction model extraction results of corpus3 into a graph database to generate a difficulty solution knowledge graph. The node types in the difficulty solution knowledge graph are mainly divided into three categories: geological condition class, drilling difficulty class, and difficulty solution class. Use knowledge graph representation methods to map the nodes and relationships in the difficulty solution knowledge graph to a low-dimensional, dense numerical vector space to generate geological condition feature vectors, drilling difficulty feature vectors, and difficulty solution feature vectors.

[0112] In some embodiments of the present specification, referring to FIG. 4, the geological condition label, drilling difficulty label, and difficulty solution label of each formation for the to-be-predicted well point are determined according to the formation feature vector of each formation for the to-be-predicted well point and the pre-constructed difficulty solution knowledge graph, including:

[0113] S401: Perform representation learning on the difficulty solution knowledge graph to obtain a geological condition feature vector corresponding to each geological condition class node in the difficulty solution knowledge graph.

[0114] S402: Calculate the similarity between the formation feature vector of the to-be-predicted well point in each formation and each geological condition feature vector in the difficult point solution knowledge graph.

[0115] Specifically, the similarity between the formation feature vector of the to-be-predicted well point in each formation and each geological condition feature vector in the difficult point solution knowledge graph can be calculated by using a similarity calculation method such as cosine similarity, Euclidean distance, or Pearson correlation coefficient.

[0116] S403: Take the labels corresponding to the first preset number of geological condition feature vectors with a similarity greater than a preset threshold as the geological condition labels of the to-be-predicted well point in each formation.

[0117] S404: Determine the drilling difficulty label and the difficult point solution label of the to-be-predicted well point in each formation from the difficult point solution knowledge graph according to the geological condition label of the to-be-predicted well point in each formation.

[0118] Specifically, after determining the geological condition label of the to-be-predicted well point in each formation, a knowledge graph-based reasoning method, such as a recommendation algorithm based on a relationship path or a recommendation algorithm based on collaborative filtering, can be used to obtain the corresponding drilling difficulty label and difficult point solution label from the difficult point solution knowledge graph. Taking the recommendation algorithm based on the relationship path as an example, first search in the difficult point solution knowledge graph for all paths with the entity type of "drilling difficulty" as the starting point, the node of "difficult point solution" as the end point, and the hop count less than a specified value, and take the difficult point solution entities in these paths as a candidate solution set. Then, set the weights according to factors such as blocks, formations, and application effects, sort the candidate solutions, and take the first several candidate solutions as the recommendation results. In this way, since the difficult point solution knowledge graph contains the causal relationship between the geological condition, the drilling difficulty, and the difficult point solution, after determining the geological condition label from the difficult point solution knowledge graph according to the formation feature vector of the to-be-predicted well point in each formation, the corresponding drilling difficulty label and difficult point solution label can be obtained according to the causal relationship between them.

[0119] Since the geological condition label, the drilling difficulty label and the difficulty solution label are relatively general texts, in order to obtain more detailed drilling difficulty analysis texts and corresponding difficulty solution texts, the prompt text is constructed according to the geological condition label, the drilling difficulty label and the difficulty solution label and input into the pre-trained drilling difficulty analysis large model to obtain the drilling difficulty prediction text of each formation of the to-be-predicted well point and the corresponding difficulty solution text. In some embodiments of the present specification, referring to FIG. 5, the prompt text is constructed according to the geological condition label, the drilling difficulty label and the difficulty solution label and input into the pre-trained drilling difficulty analysis large model to obtain the drilling difficulty prediction text of each formation of the to-be-predicted well point and the corresponding difficulty solution text, including:

[0120] S501: Fill the geological condition label, the drilling difficulty label and the difficulty solution label into the placeholder corresponding to the pre-created prompt text template to obtain the prompt text.

[0121] Specifically, the prompt text template is designed based on the causal relationship among the three in the embodiments of the present specification. Exemplarily, the prompt text template can be as follows:

[0122] You are a senior {role} with {professional skill} and have worked in {oilfield or drilling company} for many years, and are very familiar with the geological background and drilling plan of this block. Please generate a technical solution report according to the following geological condition label, drilling difficulty label and difficulty solution label control conditions:

[0123] Geological condition label: {geological condition label};

[0124] Drilling difficulty label: {drilling difficulty label};

[0125] Difficulty solution label: {difficulty solution label}.

[0126] Among them, the content in {} is the template placeholder, which is filled according to the geological condition label, the drilling difficulty label and the difficulty solution label of the to-be-predicted well point.

[0127] S502: Input the prompt text into the pre-trained drilling difficulty analysis large model to generate the drilling difficulty prediction text of each formation of the to-be-predicted well point and the corresponding difficulty solution text.

[0128] The drilling difficulty analysis large model is obtained by training a general large model based on a drilling geology engineering corpus formed by collecting data such as professional books, technical manuals, standards and specifications of geology, drilling, logging, well logging, cementing and fracturing around the knowledge system. Specifically, the construction of the drilling difficulty analysis large model includes the stages of basic large model evaluation and optimization, pre-training, instruction fine-tuning and reinforcement learning. First, a drilling industry large model evaluation system is established to evaluate the understanding ability of different basic large models in the drilling field from different dimensions, and the optimal base large model is obtained. The basic large model can be any open source large model. In addition to general capabilities such as semantic understanding, generation, dialogue, encyclopedia, logical reasoning, computing ability, and code, the evaluation dimension mainly evaluates the understanding ability of the basic large model for professional knowledge, such as drilling professional terms, drilling theoretical knowledge, drilling technical operations, safety operation specifications, and emergency plans. Representative test questions can be selected from the test question bank of driller training, logging training, well control training, etc. to construct the evaluation dataset. Second, based on the optimized basic large model, self-supervised pre-training is performed on the corpus to learn geological engineering professional knowledge and language rules, and a drilling industry large model is obtained. In the pre-training stage, the drilling field dictionary and the full word mask strategy can be combined as a pre-training task to enhance the model's understanding of professional words. Then, according to the entity and relationship type distribution in the difficulty solution knowledge graph, the difficulty solution knowledge graph is sampled hierarchically to construct a drilling difficulty and technical countermeasure instruction fine-tuning dataset, including instructions, inputs and outputs. The specific example is shown in Table 2. The drilling industry large model is fine-tuned using this dataset to learn drilling difficulties and technical countermeasures under complex geological conditions, and a customized drilling difficulty analysis large model is obtained. Finally, based on the drilling difficulty and difficulty solution data of the drilled well, a difficulty solution evaluation dataset is constructed, and a reward model is trained based on the evaluation dataset. Through continuous iteration, the generation strategy of the drilling difficulty analysis large model is optimized to make the generated difficulty solution more instructive. In addition, to address potential data leakage, sensitive words, false information, and prompt word attacks, a drilling difficulty analysis large model safety evaluation index is developed to optimize the drilling difficulty analysis large model and ensure that the generated content is legal and compliant.

[0129] Table 2

[0130] For example, the user sends a request to the server through the user terminal: "Please help me write a drilling difficulty analysis and technical solution report based on the provided geological conditions, drilling difficulty and difficulty solution tags:"

[0131] For example, the user sends a request to the server through the user terminal: "Please help me write a drilling difficulty analysis and technical solution report based on the provided geological conditions, drilling difficulty and difficulty solution tags:"

[0132] Geological condition label: contains gypsum mudstone, strong water sensitivity;

[0133] Drilling difficulty label: the filling material of the formation micro-fracture is mainly montmorillonite, which is easy to hydrate and disperse, prone to shrinkage and collapse; poor stability;

[0134] Difficulty solution label: oil-based drilling fluid.

[0135] The server inputs the above prompt words into the drilling difficulty analysis large model, automatically generates the drilling difficulty analysis text and difficulty solution text of the formation, and sends them to the user terminal through communication with the user terminal for the user to consult, as follows:

[0136] “1, geological condition overview

[0137] The main geological conditions faced in this drilling operation are gypsum-containing mudstone formations, which are characterized by strong water sensitivity. The mudstone in this formation contains a high content of gypsum, which gives the formation special physical and chemical properties, making it a high requirement for drilling operations.

[0138] 2, drilling difficulty analysis

[0139] (1) Influence of filling material of formation micro-fracture

[0140] The micro-fractures present in the formation are a major difficulty in drilling. The filling material of these micro-fractures is mainly montmorillonite, which is a mineral with strong water absorption and easy to hydrate and disperse. Under the action of drilling fluid, montmorillonite is prone to hydration, causing the micro-fractures to expand, and thus causing wellbore instability, shrinkage and collapse, etc.

[0141] (2) Poor wellbore stability

[0142] The water sensitivity of gypsum-containing mudstone formations makes the wellbore rock prone to softening and peeling off after contact with water, resulting in extremely poor wellbore stability. This not only increases the safety risk during drilling, but also can lead to reduced drilling efficiency and increased drilling costs.

[0143] (3) Difficulty in selecting drilling fluid

[0144] Given the special properties of the filling material of the formation micro-fracture, selecting the appropriate drilling fluid becomes a major difficulty. Traditional water-based drilling fluids are prone to hydration and dispersion in gypsum-containing mudstone formations, causing wellbore instability, etc., so it is necessary to find a drilling fluid that can effectively inhibit the hydration of montmorillonite and improve wellbore stability.

[0145] 3, difficulty solution

[0146] In view of the above drilling difficulties, it is recommended to use oil-based drilling fluid as the drilling fluid solution for this drilling operation, and the specific technical solution is as follows:

[0147] (1) Drilling fluid formulation selection

[0148] Choose oil-based drilling fluids with excellent inhibition and stability, and optimize their formulation to effectively inhibit the hydration and dispersion of montmorillonite, thereby improving wellbore stability. At the same time, considering environmental requirements, environmentally friendly oil-based drilling fluids should be selected to reduce environmental pollution.

[0149] (2) Drilling fluid performance control

[0150] During drilling, the density, viscosity, and other performance parameters of the drilling fluid should be strictly controlled to ensure that the drilling fluid has good sand-carrying capacity and wall-protection performance. At the same time, the performance of the drilling fluid should be adjusted in a timely manner according to formation changes and drilling progress to meet the drilling needs at different stages.

[0151] 4. Summary

[0152] In summary, the high water sensitivity of gypsum-bearing mudstone formations and the unique properties of microfracture infill materials present numerous challenges to drilling operations. Using oil-based drilling fluids as a technical solution can effectively suppress the hydration and dispersion of montmorillonite, improve wellbore stability, and reduce drilling risks. Simultaneously, strengthening drilling fluid performance control and wellbore stability maintenance, along with developing detailed emergency response measures, can ensure the smooth progress of drilling operations.

[0153] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the acquisition, storage, use, and processing of data in the technical solutions described in the embodiments of this application all comply with relevant regulations.

[0154] Based on the aforementioned method for predicting drilling difficulties and generating solutions, this specification also provides a corresponding apparatus for predicting drilling difficulties and generating solutions. The apparatus may include a system (including a distributed system), software (application), module, component, server, client, etc., using the method described in this specification, combined with necessary hardware implementation. Based on the same innovative concept, the apparatuses in one or more embodiments provided in this specification are as described in the following embodiments. Since the implementation schemes and methods for solving problems by the apparatus are similar, the implementation of specific apparatuses in this specification can refer to the implementation of the aforementioned method, and repeated details will not be repeated. As used below, the terms "unit" or "module" can refer to a combination of software and / or hardware that performs a predetermined function. Although the apparatuses described in the following embodiments are preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0155] Specifically, FIG. 6 is a schematic diagram of a module structure of one embodiment of the drilling difficulty prediction and difficulty solution generation apparatus provided by the embodiments of the present specification. Referring to FIG. 6, the drilling difficulty prediction and difficulty solution generation apparatus provided by the embodiments of the present specification comprises:

[0156] The acquisition module 601 is configured to acquire well point position information to be predicted.

[0157] The adjacent well list determination module 602 is configured to determine, according to the well point position information to be predicted, an adjacent well list of the well point to be predicted and a corresponding well history knowledge graph, wherein the well history knowledge graph is pre-constructed by using historical drilling data and comprises stratum characteristic data of different strata and relationships therebetween.

[0158] The adjacent well stratum characteristic vector calculation module 603 is configured to calculate, according to the stratum characteristic data in the well history knowledge graph, a stratum characteristic vector of each adjacent well in the adjacent well list in each stratum.

[0159] The well point to be predicted stratum characteristic vector calculation module 604 is configured to calculate, according to the stratum characteristic vector of each adjacent well in the adjacent well list in each stratum, a stratum characteristic vector of the well point to be predicted in each stratum.

[0160] The label determination module 605 is configured to determine, according to the stratum characteristic vector of the well point to be predicted in each stratum and a pre-constructed difficulty solution knowledge graph, a geological condition label, a drilling difficulty label and a difficulty solution label of the well point to be predicted in each stratum, wherein the difficulty solution knowledge graph comprises a geological condition class node, a drilling difficulty class node, a difficulty solution class node and relationships among the three types of nodes.

[0161] The drilling difficulty prediction and difficulty solution generation module 606 is configured to construct a prompt text according to the geological condition label, the drilling difficulty label and the difficulty solution label, input the prompt text into a pre-trained drilling difficulty analysis large model, and obtain a drilling difficulty prediction text and a corresponding difficulty solution text of the well point to be predicted in each stratum.

[0162] The beneficial effects achieved by the apparatus provided by the embodiments of the present specification are consistent with the beneficial effects achieved by the above method, which will not be repeated here.

[0163] With reference to FIG. 7, based on the above-mentioned method for drilling difficulty prediction and difficulty solution generation, an embodiment of the present specification further provides a computer device 702, wherein the above-mentioned method runs on the computer device 702. The computer device 702 can include one or more processors 704, such as one or more central processing units (CPUs), each of which can implement one or more hardware threads. The computer device 702 can also include any memory 706 for storing any kind of information, such as code, settings, data, etc. Without limitation, for example, the memory 706 can include any one or a combination of the following: any type of RAM, any type of ROM, a flash memory device, a hard disk, an optical disk, etc. More generally, any memory can use any technology for storing information. Further, any memory can provide volatile or non-volatile retention of information. Further, any memory can represent a fixed or removable component of the computer device 702. In one case, the computer device 702 can perform any operation of the associated instructions when the processor 704 executes the associated instructions stored in any memory or combination of memories. The computer device 702 also includes one or more drive mechanisms 708, such as a hard disk drive mechanism, an optical disk drive mechanism, etc., for interacting with any memory.

[0164] The computer device 702 can also include an input / output module 710 (I / O) for receiving various inputs (via input devices 712) and for providing various outputs (via output devices 714). One particular output mechanism can include a presentation device 716 and an associated graphical user interface (GUI) 718. In other embodiments, the input / output module 710 (I / O), the input devices 712, and the output devices 714 can also not be included, just as a computer device in a network. The computer device 702 can also include one or more network interfaces 720 for exchanging data with other devices via one or more communication links 722. One or more communication buses 724 couple the above-described components together.

[0165] The communication links 722 can be implemented in any manner, for example, through a local area network, a wide area network (e.g., the Internet), a point-to-point connection, etc., or any combination thereof. The communication links 722 can include any combination of hardwired links, wireless links, routers, gateway functionality, name servers, etc., governed by any protocol or combination of protocols.

[0166] Corresponding to the method as shown in FIGS. 1 to 5, the embodiment of the present specification also provides a computer readable storage medium, and the computer readable storage medium stores a computer program, and the computer program is run by a processor to execute the steps of the above-mentioned method.

[0167] The embodiments of the present specification also provide a computer readable instruction, wherein when the processor executes the instruction, the program therein causes the processor to execute the method shown in FIGS. 1 to 5.

[0168] The embodiments of the present specification also provide a computer program product comprising at least one instruction or at least one program, which is loaded and executed by the processor to implement the method shown in FIGS. 1 to 5.

[0169] It should be understood that the size of the sequence number of each process described above in various embodiments of the present specification does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present specification.

[0170] It should also be understood that in the embodiments of the present specification, the term "and / or" is only to describe the association relationship of the associated objects, which means that there can be three relationships. For example, A and / or B can represent three cases of A alone, A and B together, and B alone. In addition, the character " / " in the present specification generally represents an "or" relationship between the front and rear associated objects.

[0171] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed in the present specification can be realized in electronic hardware, computer software or a combination of both. In order to clearly illustrate the interchangeability of hardware and software, the components and steps of each example have been described in general terms in the above description. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present specification.

[0172] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, which will not be repeated here.

[0173] In several embodiments provided in the present specification, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the device embodiments described above are only schematic, and the division of units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed each other can be indirect coupling or communication connection through some interfaces, devices or units, and can also be electrical, mechanical or other forms of connection.

[0174] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, i.e., may be located in one place, or may be distributed to multiple network units. Part or all of the units may be selected according to actual needs to achieve the purpose of the embodiment of the present specification.

[0175] In addition, each functional unit in each embodiment of the present specification can be integrated in one processing unit, or each unit can be physically present alone, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0176] If the integrated unit is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present specification essentially or the part of the prior art that contributes to the technical solutions, or all or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present specification. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various program code storage media.

[0177] The principles and implementation manners of the present specification are described in the specific embodiments in the present specification, and the above embodiment description is only used to help understand the method of the present specification and its core idea; meanwhile, for those skilled in the art, according to the idea of the present specification, the specific implementation manner and application range will have changes, and the above, the content of the present specification should not be understood as a limitation of the present specification.

Claims

1. A method for drilling difficulty prediction and difficulty solution generation, characterized by, The method comprises: obtaining well point position information to be predicted; determining a neighbor well list of the well point to be predicted and a corresponding well history knowledge graph according to the well point position information to be predicted, wherein the well history knowledge graph is pre-constructed using historical drilling data and comprises stratum characteristic data of different strata and relationships therebetween; calculating a stratum characteristic vector of each neighbor well in the neighbor well list in each stratum according to the stratum characteristic data in the well history knowledge graph; calculating a stratum characteristic vector of the well point to be predicted in each stratum according to the stratum characteristic vector of each neighbor well in the neighbor well list in each stratum; determining a geological condition label, a drilling difficulty label and a difficulty solution label of the well point to be predicted in each stratum according to the stratum characteristic vector of the well point to be predicted in each stratum and a pre-constructed difficulty solution knowledge graph, wherein the difficulty solution knowledge graph comprises a geological condition class node, a drilling difficulty class node, a difficulty solution class node and relationships among the three types of nodes; and constructing prompt text according to the geological condition label, the drilling difficulty label and the difficulty solution label and inputting the prompt text into a pre-trained drilling difficulty analysis large model to obtain drilling difficulty prediction text and corresponding difficulty solution text of the well point to be predicted in each stratum.

2. The method of claim 1, wherein, The determination of the neighbor well list of the well point to be predicted according to the well point position information to be predicted comprises: determining a block to which the well point to be predicted belongs according to the well point position information to be predicted; calculating a well spacing between the well point to be predicted and each well point in the block; and determining a well point with a well spacing less than a preset threshold as a neighbor well of the well point to be predicted to obtain the neighbor well list of the well point to be predicted.

3. The method of claim 1, wherein, The calculation of the stratum characteristic vector of each neighbor well in the neighbor well list in each stratum according to the stratum characteristic data in the well history knowledge graph comprises: performing representation learning on the well history knowledge graph according to the stratum characteristic data to obtain stratum characteristic vector data; obtaining a first stratum characteristic vector and a second stratum characteristic vector of a certain neighbor well in a certain stratum according to the stratum characteristic vector data; fusing the first stratum characteristic vector and the second stratum characteristic vector to obtain a third stratum characteristic vector of the certain neighbor well in the certain stratum; calculating stratum characteristic vectors of adjacent upper and lower strata of the certain neighbor well in the certain stratum; performing nonlinear calculation on the third stratum characteristic vector and the stratum characteristic vectors of the adjacent upper and lower strata to obtain a fourth stratum characteristic vector of the certain neighbor well in the certain stratum; obtaining a neighbor well list of the certain neighbor well, calculating a distance influence coefficient of each neighbor well in the neighbor well list of the certain neighbor well and a stratum characteristic vector of each neighbor well in the certain stratum; calculating a final stratum characteristic vector of the certain neighbor well in the certain stratum according to the fourth stratum characteristic vector, the distance influence coefficient and the stratum characteristic vector of each neighbor well in the neighbor well list of the certain neighbor well in the certain stratum; and repeating the above steps until the stratum characteristic vector of each neighbor well in each stratum in the neighbor well list is obtained.

4. The method of claim 3, wherein, The final formation feature vector of the certain adjacent well in the certain formation is calculated according to the fourth formation feature vector, the distance influence coefficient and the formation feature vector of each adjacent well in the adjacent well list of the certain adjacent well, and the method comprises the following steps: The final formation eigenvector of the certain formation of the certain adjacent well is calculated by using the following formula: wherein E_W i S j represents the final formation feature vector of the i-th adjacent well in the j-th formation, E_W i S j _2 represents the fourth formation feature vector of the i-th adjacent well in the j-th formation, n1 represents the number of adjacent wells of the i-th adjacent well, β k represents the distance influence coefficient of the i-th adjacent well and the k-th adjacent well in the adjacent well list thereof, E_W k S j _2 represents the formation feature vector of the k-th adjacent well in the adjacent well list of the i-th adjacent well in the j-th formation.

5. The method of claim 1, wherein, The formation feature vector of the to-be-predicted well point in each formation is calculated according to the formation feature vector of each adjacent well in each formation, and the method comprises the following steps: The formation feature vector of each formation for the well point to be predicted is calculated using the following formula: wherein E_W Loc S j denotes the formation feature vector of the jth formation of the well point to be predicted, n2 denotes the number of neighboring wells of the well point to be predicted, ρ i denotes the distance influence coefficient of the well point to be predicted and the ith neighboring well, E_W i S j denotes the formation feature vector of the jth formation of the ith neighboring well.

6. The method of claim 1, wherein, The geological condition label, the drilling difficulty label and the difficulty solution label of the to-be-predicted well point in each formation are determined according to the formation feature vector of the to-be-predicted well point in each formation and the pre-constructed difficulty solution knowledge graph, and the method comprises the following steps: The difficulty solution knowledge graph is subjected to representation learning to obtain a geological condition feature vector corresponding to each geological condition class node in the difficulty solution knowledge graph; The similarity between the formation feature vector of the to-be-predicted well point in each formation and each geological condition feature vector in the difficulty solution knowledge graph is calculated; The labels corresponding to the first preset number of geological condition feature vectors with a similarity greater than a preset threshold are taken as the geological condition label of the to-be-predicted well point in each formation; and The drilling difficulty label and the difficulty solution label of the to-be-predicted well point in each formation are determined from the difficulty solution knowledge graph according to the geological condition label of the to-be-predicted well point in each formation.

7. The method of claim 1, wherein, The geological condition label, the drilling difficulty label and the difficulty solution label are filled into the placeholders corresponding to the prompt text template to obtain a prompt text; and The prompt text is input into the pre-trained drilling difficulty analysis large model to generate the drilling difficulty prediction text and the corresponding difficulty solution text of the to-be-predicted well point in each formation. The device comprises:

8. A drilling difficulty prediction and difficulty solution generation apparatus, characterized by, An acquisition module is configured to acquire to-be-predicted well point position information; An adjacent well list determination module is configured to determine, according to the to-be-predicted well point position information, an adjacent well list of the to-be-predicted well point and a corresponding well history knowledge graph, wherein the well history knowledge graph is pre-constructed using historical drilling data and comprises formation feature data of different formations and relationships between the formations; An adjacent well formation feature vector calculation module is configured to calculate, according to the formation feature data in the well history knowledge graph, a formation feature vector of each adjacent well in each formation in the adjacent well list; A to-be-predicted well point formation feature vector calculation module is configured to calculate, according to the formation feature vector of each adjacent well in each formation in the adjacent well list, a formation feature vector of the to-be-predicted well point in each formation; ​ The label determination module is configured to determine, according to the stratum feature vector of each stratum of the well point to be predicted and a pre-constructed difficult point solution knowledge graph, a geological condition label, a drilling difficulty label and a difficult point solution label of each stratum of the well point to be predicted, wherein the difficult point solution knowledge graph comprises a geological condition class node, a drilling difficulty class node, a difficult point solution class node and a relationship among the three types of nodes. The drilling difficulty prediction and difficult point solution generation module is configured to input a prompt text constructed according to the geological condition label, the drilling difficulty label and the difficult point solution label into a pre-trained drilling difficulty analysis large model to obtain a drilling difficulty prediction text and a corresponding difficult point solution text of each stratum of the well point to be predicted.

9. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the method of any one of claims 1 to 7.

11. A computer program product, characterised in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the method of any one of claims 1 to 7. The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the method of any one of claims 1 to 7.

Images