Shale oil and gas intelligent drilling speed increasing method and system based on benchmark library

Abstract

The invention relates to the technical field of shale oil and gas engineering construction, in particular to a shale oil and gas intelligent drilling speed increasing method and system based on a benchmark library. According to the method, the block drilling speed increase benchmark library is established, the difference between the current mechanical drilling speed and the optimal mechanical drilling speed of the benchmark library is dynamically compared in the drilling process, the early warning that the mechanical drilling speed is too low is automatically sent out, and meanwhile the knowledge library is utilized to intelligently output the optimal drilling construction scheme; the method is delivered to engineering technicians for implementation in a text and voice broadcast mode, all drilling construction in a block can be promoted to automatically benchmark optimal indexes in real time, the drilling speed and efficiency of the whole block are improved, the drilling cost is reduced, and the economic benefits of shale oil and gas development are improved; according to the method, it can be ensured that each level of each well in the block can adopt the current optimal and latest construction parameters, acceleration and cost reduction of large-batch well drilling blocks are facilitated, the commercial development value of shale oil and gas is improved, and the method is wide in application prospect and easy to popularize.

Description

Technical Field

[0001] This invention relates to the field of shale oil and gas engineering construction technology, specifically to an intelligent drilling acceleration method and system for shale oil and gas based on a benchmark database. Background Art

[0002] Accelerating drilling is a core theme for improving efficiency and reducing costs in oil and gas exploration and development. For shale oil and gas development areas, which utilize large-scale drilling to achieve large-scale production, accelerating drilling to reduce costs is particularly crucial to enhancing efficient development capabilities. Currently, accelerating drilling involves two main approaches: First, comprehensively collecting geological data and deeply analyzing formation characteristics to develop precise drilling plans, establishing efficient communication mechanisms, and promoting collaborative work among geology, engineering, and logistics departments to reduce non-productive time. Second, real-time acquisition of drilling parameters and application of big data analysis to promptly detect anomalies, provide early warnings of potential risks, and introduce artificial intelligence algorithms to automatically optimize drilling parameters based on real-time data and historical experience, achieving intelligent drilling decision-making and ensuring drilling operations are always at their optimal state, continuously improving drilling speed.

[0003] However, for shale oil and gas development blocks with a large number of wells, the formation characteristics penetrated by each well have many subtle differences, and each drilling team adopts differentiated construction plans based on the actual drilling conditions. Current technologies mainly rely on regular personnel exchanges and technical comparisons to promote faster and cheaper large-scale drilling. However, this communication and information sharing efficiency is low and cannot meet the needs of speeding up and reducing costs in blocks with a large number of wells. Current technologies lack a more quantitative, scientific, and efficient means to ensure that each well can share the latest and most valuable construction experience within the work area, thereby maximizing the speed of drilling in the work area. Summary of the Invention

[0004] The technical problem to be solved by this invention is to provide an intelligent drilling speed-up method and system for shale oil and gas based on a benchmark database. This system can dynamically compare the difference between the current mechanical drilling rate and the optimal mechanical drilling rate in the benchmark database and automatically issue low-speed warnings. At the same time, it can intelligently output the optimal drilling construction plan using a knowledge base, thereby achieving overall drilling speed-up and efficiency improvement in the block, reducing drilling costs, and improving the economic benefits of shale oil and gas development.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A method for accelerating intelligent drilling in shale oil and gas based on a benchmark database includes the following steps: Step S1: Establish a dynamically updated benchmark database for block drilling acceleration, filter data from completed wells or drilled sections, construct a benchmark sample database for drilling plans, and dynamically update the benchmark sample database in real time. The benchmark sample database includes benchmark formation characteristic data. and benchmark project parameter data ; Step S2: Establish a drilling acceleration knowledge base, slice the benchmark sample database based on data type, and use vectorization tools to convert the sliced ​​data into high-dimensional semantic vectors and store them in the vector knowledge base; Step S3: Match the current drilling formation characteristics. During drilling, acquire the current formation characteristic data in real time and quantize it. The weighted Euclidean distance method is used to calculate the current stratigraphic characteristic data. Compared with benchmark stratigraphic characteristics data The vector distance is used to filter out the top N formation characteristic data and their corresponding engineering parameter data, and the maximum value of the average mechanical drilling rate ROP of the well section is taken as the optimal mechanical drilling rate ROP0. Step S4, Average Drilling Rate Calculation and Low-Speed ​​Warning: Dynamically calculate the average mechanical drilling rate ROP of the current well and current formation, and compare it with the optimal mechanical drilling rate ROP0 in the benchmark library. When ROP-ROP0<0, issue a warning of excessively low mechanical drilling rate in a timely manner. Step S5: Intelligently recommend the optimal drilling plan. The constructed vector knowledge base is integrated into the Large Language Deep Learning Model (LLM). When a warning of low mechanical drilling speed is issued, based on the Large Language Deep Learning Model, combined with the current formation characteristics and engineering parameters, and using the Retrieval Enhancement Generation Algorithm (RAG), the optimal drilling construction plan and parameter optimization measures that can be implemented at present are output. Step S6: Human-computer interaction is implemented. Based on the optimal drilling construction plan, the corresponding original construction plan text V0 is extracted, and the original construction plan text V0 is converted into professional terms to derive the corresponding speech reading text V1. Then, based on the speech reading text V1, the speech synthesis model is called to generate the corresponding speech broadcast.

[0006] Preferably, in step S1, the benchmark formation characteristic data ={strata position, lithology, apparent dip angle, vertical depth}, the benchmark engineering parameter data ={Rate of drilling (ROP) of the well section, type and parameters of the drill bit used, depth and footage of the drill bit, speed-up tools, pressure on the drill bit, rotational speed, displacement, drilling fluid system and performance}.

[0007] Preferably, in step S1, when constructing the benchmark sample database of the drilling scheme, sample data of well sections that have experienced complex failures due to improper construction measures during the drilling process are removed; The average mechanical drilling rate ROP of the well section is calculated as (H0-H') / (t0-t'), where H0 and t0 are the total depth of the well section and the total drilling time of the well section, respectively, and H' and t' are the depth of the well section where complex failures occurred during drilling and the drilling time of the well section, respectively.

[0008] Preferably, in step S2, the step of converting the sliced ​​data into high-dimensional semantic vectors using vectorization tools includes: S21, Input preprocessing: Cleaning (removing redundant data) and standardizing the data slices to avoid noise affecting vector quality; S22, Create a set: Determine the dimension and distance metric of the high-dimensional semantic vector based on the types of engineering parameters; S23, Build an index: Build a knowledge base index for data keywords, including: well number, layer, and mechanical drilling rate of OP; S24, Write Vectors and Metadata: Associate each high-dimensional semantic vector with the metadata in the benchmark sample database and store it in the vector knowledge base.

[0009] Preferably, in step S3, the weighted Euclidean distance calculation method uses weighted values ​​for stratigraphic position, stratigraphic lithology, apparent dip angle, and vertical depth in the ranges of 0.3~0.5, 0.3~0.4, 0.1~0.2, and 0.1~0.2, respectively.

[0010] Preferably, in step S4, Δ = ROP - ROP0 is calculated in real time. When Δ < 0, it means that the average mechanical drilling rate of the current well and the current formation is lower than the optimal mechanical drilling rate in the benchmark database, and a warning of low mechanical drilling rate is issued.

[0011] Preferably, in step S5, the large language deep learning model includes: RAG retrieval layer: From the vector knowledge base, match the historical data most similar to the current stratigraphic characteristics and engineering parameters as the "factual basis" for the large language deep learning model LLM generation scheme; LLM generation layer: Based on the retrieved vector knowledge base data and combined with the principles of formation mechanics and fluid mechanics, it infers the optimal drilling construction plan that is suitable for the current scenario.

[0012] Preferably, in step S6, the large speech synthesis model is specifically a large speech synthesis model based on TTS technology.

[0013] A benchmark database-based intelligent drilling acceleration system for shale oil and gas operates using the drilling acceleration methods described above, mainly including: 1) Low Drilling Rate Early Warning Module: Used to dynamically calculate the average mechanical drilling rate (ROP) of the current well and current formation, and compare it with the optimal mechanical drilling rate (ROP0) in the benchmark library, so as to issue an early warning of low mechanical drilling rate in a timely manner. 2) Speed-up and optimization module: This module integrates the constructed vector knowledge base with the Large Language Deep Learning Model (LLM), combines the current formation characteristics and engineering parameters, and outputs the optimal drilling construction plan and parameter optimization measures that can be implemented at present based on the Retrieval Enhancement Generation Algorithm (RAG). 3) Human-computer interaction module: It is used to derive the corresponding voice reading text V1 based on the original construction plan text V0, and display the text V0 through the software interface. Then, it calls the corresponding large voice synthesis model to output the text V1 in the form of voice broadcast.

[0014] Compared with the prior art, the present invention has the following main advantages: 1. This invention establishes a benchmark library for improving drilling speed in a block and dynamically compares the current mechanical drilling speed with the optimal mechanical drilling speed in the benchmark library during drilling. It automatically issues a warning that the mechanical drilling speed is too low. At the same time, it uses a knowledge base to intelligently output the optimal drilling construction plan and delivers it to engineering technicians for implementation through text and voice broadcast. This can promote all drilling construction in the block to automatically benchmark against the optimal indicators in real time, thereby improving the overall drilling speed and efficiency of the block, reducing drilling costs, and improving the economic benefits of shale oil and gas development. 2. This invention can automatically issue low-speed warnings during construction and intelligently recommend the optimal drilling construction plan. It can ensure that each well in each layer of the block can use the current optimal and latest construction parameters, which helps to accelerate and reduce costs in large-scale drilling blocks and enhance the commercial development value of shale oil and gas. Its application prospects are broad and it is easy to promote. Attached Figure Description

[0015] Figure 1 This is an overall flowchart of the intelligent drilling acceleration method for shale oil and gas based on a benchmark library in an embodiment of the present invention; Figure 2 This is a schematic diagram of an intelligent drilling acceleration system for shale oil and gas based on a benchmark database, as described in an embodiment of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0017] It should be noted that, depending on the implementation needs, the various steps / components described in this application can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.

[0018] In this invention, unless otherwise expressly specified and limited, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise expressly and specifically limited.

[0019] To further explore the potential of drilling operations in terms of speed and efficiency, this application proposes a new "single-well speed efficiency" index to scientifically and accurately quantify the drilling speed and efficiency of shale oil and gas single wells, and proposes a shale oil and gas drilling management method based on the single-well speed efficiency index to achieve faster and more efficient shale oil and gas drilling.

[0020] Example 1: This example provides a method for accelerating intelligent drilling in shale oil and gas based on a benchmark database, such as... Figure 1 As shown, the main steps include the following: Step S1: Establish a dynamically updated benchmark database for block drilling acceleration, filter data from completed wells or drilled sections, construct a benchmark sample database for drilling plans, and dynamically update the benchmark sample database in real time. The benchmark sample database includes benchmark formation characteristic data. and benchmark project parameter data ; Step S2: Establish a drilling acceleration knowledge base, slice the benchmark sample database based on data type, and use vectorization tools to convert the sliced ​​data into high-dimensional semantic vectors and store them in the vector knowledge base; Step S3: Match the current drilling formation characteristics. During drilling, acquire the current formation characteristic data in real time and quantize it. The weighted Euclidean distance method is used to calculate the current stratigraphic characteristic data. Compared with benchmark stratigraphic characteristics data The vector distance is used to filter out the top N formation characteristic data and their corresponding engineering parameter data, and the maximum value of the average mechanical drilling rate ROP of the well section is taken as the optimal mechanical drilling rate ROP0. Step S4, Average Drilling Rate Calculation and Low-Speed ​​Warning: Dynamically calculate the average mechanical drilling rate ROP of the current well and current formation, and compare it with the optimal mechanical drilling rate ROP0 in the benchmark library. When ROP-ROP0<0, issue a warning of excessively low mechanical drilling rate in a timely manner. Step S5: Intelligently recommend the optimal drilling plan. The constructed vector knowledge base is integrated into the Large Language Deep Learning Model (LLM). When a warning of low mechanical drilling speed is issued, based on the Large Language Deep Learning Model, combined with the current formation characteristics and engineering parameters, and using the Retrieval Enhancement Generation Algorithm (RAG), the optimal drilling construction plan and parameter optimization measures that can be implemented at present are output. Step S6: Human-computer interaction is implemented. Based on the optimal drilling construction plan, the corresponding original construction plan text V0 is extracted, and the original construction plan text V0 is converted into professional terms to derive the corresponding speech reading text V1. Then, based on the speech reading text V1, the speech synthesis model is called to generate the corresponding speech broadcast.

[0021] Furthermore, in step S1, the benchmark formation characteristic data ={strata position, lithology, apparent dip angle, vertical depth}, the benchmark engineering parameter data ={Rate of drilling (ROP) of the well section, type and parameters of the drill bit used, depth and footage of the drill bit, speed-up tools, pressure on the drill bit, rotational speed, displacement, drilling fluid system and performance}.

[0022] Furthermore, in step S1, when constructing the benchmark sample database of the drilling scheme, sample data of well sections that have experienced complex failures due to improper construction measures during the drilling process are removed; The average mechanical drilling rate ROP of the well section is calculated as (H0-H') / (t0-t'), where H0 and t0 are the total depth of the well section and the total drilling time of the well section, respectively, and H' and t' are the depth of the well section where complex failures occurred during drilling and the drilling time of the well section, respectively.

[0023] Furthermore, in step S2, the process of converting the sliced ​​data into high-dimensional semantic vectors using vectorization tools includes: S21, Input preprocessing: Cleaning (removing redundant data) and standardizing the data slices to avoid noise affecting vector quality; S22, Create a set: Determine the dimension and distance metric of the high-dimensional semantic vector based on the types of engineering parameters; S23, Build an index: Build a knowledge base index for data keywords, including: well number, layer, and mechanical drilling rate of OP; S24, Write Vectors and Metadata: Associate each high-dimensional semantic vector with the metadata in the benchmark sample database and store it in the vector knowledge base.

[0024] Furthermore, in step S3, in the weighted Euclidean distance calculation method, the weighted values ​​of the stratigraphic position, stratigraphic lithology, apparent dip angle, and vertical depth are respectively 0.3~0.5, 0.3~0.4, 0.1~0.2, and 0.1~0.2.

[0025] Furthermore, in step S4, Δ = ROP - ROP0 is calculated in real time. When Δ < 0, it indicates that the average mechanical drilling rate of the current well and current formation is lower than the optimal mechanical drilling rate in the benchmark database, and a warning of excessively low mechanical drilling rate is issued.

[0026] Furthermore, in step S5, the large language deep learning model includes: RAG retrieval layer: From the vector knowledge base, match the historical data most similar to the current stratigraphic characteristics and engineering parameters as the "factual basis" for the large language deep learning model LLM generation scheme; LLM generation layer: Based on the retrieved vector knowledge base data and combined with the principles of formation mechanics and fluid mechanics, it infers the optimal drilling construction plan that is suitable for the current scenario.

[0027] Furthermore, in step S6, the large speech synthesis model is specifically a large speech synthesis model based on TTS technology.

[0028] Example 2: This example provides an intelligent drilling acceleration method for shale oil and gas based on a benchmark database. It can automatically issue low-speed warnings during the construction process and intelligently recommend the optimal and feasible drilling construction plan.

[0029] First, a benchmark library for improving drilling speed in the block is established, storing the corresponding mechanical rate of drilling (ROP) and drilling measures according to formation level, drilling trip, and drilling session. Based on professional characteristics, the content of the benchmark library is sliced ​​and vectorized, and indexes are created for key parameters based on the vector library. During drilling, the current ROP is automatically compared with the optimal ROP0 for the same formation. If the ROP is lower than ROP0, an early warning is issued. At the same time, the optimal drilling plan is intelligently recommended based on the vector knowledge base using a large model (LLM), and delivered to the engineering and technical personnel for implementation through text and voice broadcast.

[0030] Furthermore, during the drilling process, the current rate of drilling (RPD) is dynamically compared with the optimal RPD in the benchmark database. If the current RPD is lower than the optimal RPD in the benchmark database, an automatic warning is issued. Simultaneously, the optimal drilling acceleration plan is intelligently output using a knowledge base, and implemented through text and voice broadcasts to interact with technical personnel. This ensures that the current optimal parameters are used at every level of every well, achieving an overall acceleration of batch drilling in shale oil and gas blocks and enhancing the commercial development value of shale oil and gas.

[0031] Furthermore, when constructing the vector knowledge base, in order to improve the retrieval efficiency of the knowledge base, indexes can be established for keywords such as well number, formation level, and ROP. Indexing methods can include HNSW index, nnoy index, Faiss index, etc.

[0032] Furthermore, during the drilling process, obtain the current formation characteristics and vectorize them into , and use the weighted Euclidean distance calculation method (the value ranges of the weighted values of formation level, formation lithology, apparent dip angle, and vertical depth are 0.3 - 0.5, 0.3 - 0.4, 0.1 - 0.2, 0.1 - 0.2 in sequence), and compare it with the in the vector library to calculate the vector distance, and obtain the with the vector distance ranking N and the corresponding , and obtain the average ROP of each well section in each vector x , find the maximum value of ROP among them, and set it as ROP0 Furthermore, integrate the constructed vector knowledge base with a large model (such as DeepSeek). When a low-speed warning is issued, use the RAG technology to obtain the knowledge fragments related to the current formation speed increase, and then input these knowledge fragments and the problem into the large model to quickly generate accurate and professional drilling optimization plans.

[0033] Furthermore, when a low-speed warning is issued and a speed increase optimization plan (set as V0) is output, call the speech synthesis large model to perform voice broadcast on it to improve the efficiency of human-machine collaboration. Since most current speech synthesis base models cannot comprehensively and accurately synthesize drilling professional vocabulary or units (such as the polyphonic character "drill", unit kN, etc.), in this example, the special vocabulary in V0 is converted into a professional pronunciation (such as converting "kN" into "kilo-newton"), and V1 is derived. When implemented in software, V0 is provided to the software interface for display output to ensure its professionalism, and V1 is provided to the speech synthesis large model for voice broadcast to ensure the accuracy of its pronunciation.

[0034] Embodiment 3, based on the same inventive concept, this embodiment also provides a shale oil and gas intelligent drilling speed increase system based on a benchmark library, which adopts the drilling speed increase method as described above, as Figure 2 shown, mainly including: 1) Low drilling speed warning module: Dynamically calculate the difference between the current ROP and the benchmark ROP0, and issue a warning according to the rules.

[0035] 2) Speed increase optimization module: Use the large language deep learning model (LLM) combined with the vector knowledge base to think, and output the optimal drilling plan and parameter optimization measures.

[0036] 3) Human-computer interaction module: Based on the original drilling speed-up scheme text V0, a ​​suitable text V1 for voice reading is derived. The software interface displays V0, and at the same time, a suitable large speech synthesis model is called to output V1 in speech mode.

[0037] Example 4: Based on the same inventive concept, this example also provides a non-transitory readable storage medium on which a program is stored. When the program is executed by the intelligent drilling speed-up system, it implements the drilling speed-up method as described above.

[0038] Furthermore, all parts of this application that are not described in detail are the same as or implemented using existing technology.

[0039] In summary: 1. This invention establishes a benchmark library for improving drilling speed in a block and dynamically compares the current mechanical drilling speed with the optimal mechanical drilling speed in the benchmark library during drilling. It automatically issues a warning that the mechanical drilling speed is too low. At the same time, it uses a knowledge base to intelligently output the optimal drilling construction plan and delivers it to engineering technicians for implementation through text and voice broadcast. This can promote all drilling construction in the block to automatically benchmark against the optimal indicators in real time, thereby improving the overall drilling speed and efficiency of the block, reducing drilling costs, and improving the economic benefits of shale oil and gas development. 2. This invention can automatically issue low-speed warnings during construction and intelligently recommend the optimal drilling construction plan. It can ensure that each well in each layer of the block can use the current optimal and latest construction parameters, which helps to accelerate and reduce costs in large-scale drilling blocks and enhance the commercial development value of shale oil and gas. Its application prospects are broad and it is easy to promote.

[0040] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0041] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0042] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for accelerating intelligent drilling in shale oil and gas based on a benchmark database, characterized in that, The steps include: S1. Establish a dynamically updated benchmark database for accelerating block drilling, filter data of completed wells or drilled sections, construct a benchmark sample database for drilling schemes, and dynamically update the benchmark sample database in real time. The benchmark sample database includes benchmark formation characteristic data and benchmark engineering parameter data. S2, establish a drilling acceleration knowledge base, slice the benchmark sample database based on data type, and use vectorization tools to convert the sliced ​​data into high-dimensional semantic vectors and store them in the vector knowledge base; S3, Match the current drilling formation characteristics. During the drilling process, the current formation characteristic data is acquired in real time and vectorized. The weighted Euclidean distance calculation method is used to calculate the vector distance between the current formation characteristic data and the benchmark formation characteristic data. The formation characteristic data and their corresponding engineering parameter data with the top N vector distances are selected. The maximum value of the average mechanical drilling speed of the well section is taken as the optimal mechanical drilling speed. S4, Average Drilling Rate Calculation and Low-Speed ​​Warning: Dynamically calculate the average mechanical drilling rate of the current well and current formation, and compare it with the optimal mechanical drilling rate in the benchmark database. When the average mechanical drilling rate is lower than the optimal mechanical drilling rate in the benchmark database, a warning of excessively low mechanical drilling rate is issued in a timely manner. S5 intelligently recommends the optimal drilling plan. It integrates the constructed vector knowledge base into the large language deep learning model. When a warning of low mechanical drilling speed is issued, based on the large language deep learning model, combined with the current formation characteristics and engineering parameters, and using the retrieval enhancement generation algorithm, it outputs the currently feasible optimal drilling construction plan and parameter optimization measures. S6, Human-computer interaction is implemented. Based on the optimal drilling construction plan, the corresponding original construction plan text V0 is extracted, and the original construction plan text V0 is converted into professional terms to derive the corresponding speech reading text V1. Then, based on the speech reading text V1, the speech synthesis model is called to generate the corresponding speech broadcast.

2. The intelligent drilling acceleration method for shale oil and gas based on a benchmark database as described in claim 1, characterized in that... In step S1, the benchmark formation characteristic data ={strata position, lithology, apparent dip angle, vertical depth}, the benchmark engineering parameter data ={Rate of drilling (ROP) of the well section, type and parameters of the drill bit used, depth and footage of the drill bit, speed-up tools, pressure on the drill bit, rotational speed, displacement, drilling fluid system and performance}.

3. The intelligent drilling acceleration method for shale oil and gas based on a benchmark database as described in claim 1, characterized in that... In step S1, when constructing the benchmark sample database of the drilling scheme, sample data of well sections that have experienced complex failures due to improper construction measures during the drilling process are removed; The average mechanical drilling rate ROP of the well section is calculated as (H0-H') / (t0-t'), where H0 and t0 are the total depth of the well section and the total drilling time of the well section, respectively, and H' and t' are the depth of the well section where complex failures occurred during drilling and the drilling time of the well section, respectively.

4. The intelligent drilling acceleration method for shale oil and gas based on a benchmark database as described in claim 1, characterized in that... In step S2, the step of converting the sliced ​​data into high-dimensional semantic vectors using vectorization tools includes: S21, Input preprocessing: Cleaning and standardizing the data slices; S22, Create a set: Determine the dimension and distance metric of the high-dimensional semantic vector based on the types of engineering parameters; S23, Build an index: Build a knowledge base index for data keywords, including: well number, layer, and mechanical drilling rate of OP; S24, Write Vectors and Metadata: Associate each high-dimensional semantic vector with the metadata in the benchmark sample database and store it in the vector knowledge base.

5. The intelligent drilling acceleration method for shale oil and gas based on a benchmark database as described in claim 1, characterized in that... In step S3, the weighted Euclidean distance calculation method uses weighted values ​​for stratigraphic position, stratigraphic lithology, apparent dip angle, and vertical depth in the ranges of 0.3~0.5, 0.3~0.4, 0.1~0.2, and 0.1~0.2, respectively.

6. The intelligent drilling acceleration method for shale oil and gas based on a benchmark database according to claim 1, characterized in that... In step S4, Δ = ROP - ROP0 is calculated in real time. When Δ < 0, it means that the average mechanical drilling rate of the current well and current formation is lower than the optimal mechanical drilling rate in the benchmark database, and a warning of low mechanical drilling rate is issued.

7. The intelligent drilling acceleration method for shale oil and gas based on a benchmark database according to claim 1, characterized in that... In step S5, the large language deep learning model includes: RAG retrieval layer: From the vector knowledge base, match the historical data most similar to the current stratigraphic characteristics and engineering parameters as the "factual basis" for the large language deep learning model LLM generation scheme; LLM generation layer: Based on the retrieved vector knowledge base data and combined with the principles of formation mechanics and fluid mechanics, it infers the optimal drilling construction plan that is suitable for the current scenario.

8. The intelligent drilling acceleration method for shale oil and gas based on a benchmark database as described in claim 1, characterized in that... In step S6, the large speech synthesis model is specifically a large speech synthesis model based on TTS technology.

9. A shale oil and gas intelligent drilling acceleration system based on a benchmark database, which operates the drilling acceleration method as described in any one of claims 1 to 8, characterized in that, include: 1) Low Drilling Rate Early Warning Module: Used to dynamically calculate the average mechanical drilling rate of the current well and current formation, and compare it with the optimal mechanical drilling rate in the benchmark library, so as to issue an early warning of excessively low mechanical drilling rate in a timely manner. 2) Speed-up and optimization module: This module integrates the constructed vector knowledge base with the large language deep learning model, combines the current formation characteristics and engineering parameters, and outputs the optimal drilling construction plan and parameter optimization measures that can be implemented at present based on the retrieval enhancement generation algorithm. 3) Human-computer interaction module: It is used to derive the corresponding voice reading text based on the original construction plan text, display the original construction plan text through the software interface, and then call the corresponding large voice synthesis model to output the voice reading text in the form of voice broadcast.

10. A non-transitory readable storage medium having a program stored thereon, characterized in that, When executed by the intelligent drilling acceleration system, the program implements the drilling acceleration method as described in any one of claims 1 to 8.

Images