Seismic exploration equipment intelligent manufacturing method and system based on large language model, terminal and storage medium

Abstract

The invention relates to the technical field of data processing, and discloses a seismic exploration equipment intelligent manufacturing method and system based on a large language model, a terminal and a storage medium, and the method comprises the steps: obtaining multi-source seismic data corresponding to initial seismic exploration parameters of seismic exploration equipment, and carrying out the data processing, and obtaining target multi-source seismic data; performing knowledge graph construction to obtain a marine geological knowledge graph; determining a target large language model, and performing autonomous learning processing according to the marine geological knowledge map by adopting a retrieval enhancement technology to obtain a seismic exploration parameter adjustment model; obtaining current seismic profile data, performing autonomous decision processing, outputting seismic exploration parameter configuration suggestions, and adjusting initial seismic exploration parameters of seismic exploration equipment; and optimizing the seismic exploration parameter adjustment model according to the actual exploration data. According to the invention, the imaging resolution, the penetrating power and the data acquisition efficiency of the seismic exploration equipment are obviously improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method, system, terminal and computer-readable storage medium for intelligent manufacturing of seismic exploration equipment based on a large language model. Background Art

[0002] Marine geological exploration is a crucial foundation for engineering activities such as offshore oil and gas resource development, submarine wind farm layout, submarine pipeline laying, and geological hazard monitoring. As the scale of marine engineering continues to expand, higher requirements are placed on the accuracy, efficiency, and intelligence of geological exploration. Marine seismic exploration, as a mainstream method for geological structure imaging, has been widely used in many fields.

[0003] However, existing marine seismic exploration systems generally use source and receiver arrays with fixed parameter configurations. Key parameter ranges such as source frequency, power, and receiver spacing are preset before the equipment leaves the factory. Basic exploration goals can be achieved in sea areas with relatively simple geological conditions. However, in complex geological environments such as wind power exploration, submarine landslides, and shallow gas-rich areas, conventional parameter combinations make it difficult to achieve high-quality imaging. Imaging resolution and penetration depth often cannot be taken into account at the same time, affecting the accuracy of geological target identification and exploration efficiency.

[0004] Therefore, there is an urgent need for an intelligent manufacturing method and system with self-learning and parameter self-adjustment capabilities to meet the urgent needs of high-resolution seismic imaging in a changing marine environment and realize the automatic configuration and precise operation of seismic exploration equipment in complex geological scenarios. Summary of the Invention

[0005] The main purpose of the present invention is to provide an intelligent manufacturing method, system, terminal and computer-readable storage medium for seismic exploration equipment based on a large language model, aiming to solve the problem that marine seismic exploration systems in the prior art generally use source and receiver arrays with fixed parameter configurations, lack an adaptive adjustment mechanism, and are difficult to cope with the requirements of refined imaging in complex geological environments.

[0006] To achieve the above objectives, the present invention provides a method for intelligent manufacturing of seismic exploration equipment based on a large language model, the method comprising the following steps: Acquiring multi-source seismic data corresponding to initial seismic exploration parameters of the seismic exploration equipment, and performing data processing on the multi-source seismic data to obtain target multi-source seismic data; Constructing a knowledge graph based on the target multi-source seismic data to obtain a marine geological knowledge graph; Determining a target large language model, and using retrieval enhancement technology to perform autonomous learning processing on the target large language model based on the marine geological knowledge graph to obtain a seismic exploration parameter adjustment model; Acquiring current seismic profile data, performing autonomous decision-making based on the current seismic profile data using the seismic exploration parameter adjustment model, outputting a seismic exploration parameter configuration suggestion, and adjusting the initial seismic exploration parameters of the seismic exploration equipment based on the seismic exploration parameter configuration suggestion to obtain a target seismic exploration equipment; Actual exploration data is acquired, and a reinforcement learning strategy is adopted to optimize the seismic exploration parameter adjustment model carried in the target seismic exploration equipment according to the actual exploration data.

[0007] Optionally, in the intelligent manufacturing method for seismic exploration equipment based on a large language model, the data processing includes preprocessing and reprocessing; The acquiring of multi-source seismic data corresponding to the initial seismic exploration parameters of the seismic exploration equipment and performing data processing on the multi-source seismic data to obtain target multi-source seismic data specifically includes: Determining initial seismic exploration parameters of the seismic exploration equipment, collecting reflected wave signals according to the initial seismic exploration parameters, and obtaining an exploration environment corresponding to the reflected wave signals, wherein the initial seismic exploration parameters include an initial source frequency and an initial receiving array layout; Acquiring existing seismic exploration data, and constructing multi-source seismic data based on the existing seismic exploration data, the initial seismic exploration parameters, the reflected wave signal, and the exploration environment; Preprocessing and reprocessing the multi-source seismic data to obtain target multi-source seismic data; The pre-processing includes cleaning, standardization, anomaly detection and error correction. The reprocessing includes standardization processing, format conversion processing, noise removal and completion processing, and accuracy verification processing.

[0008] Optionally, the intelligent manufacturing method for seismic exploration equipment based on a large language model, wherein the knowledge graph is constructed based on the target multi-source seismic data to obtain a marine geological knowledge graph, specifically comprising: Determining a map node, wherein the map node includes stratigraphic units, reflection characteristics, sedimentary types, equipment parameters, and imaging effects; A triplet model is used to construct a knowledge graph for the target multi-source seismic data according to the graph nodes to obtain a marine geological knowledge graph, wherein the triples in the triplet model include source parameters, geological characteristics and exploration effects.

[0009] Optionally, in the intelligent manufacturing method for seismic exploration equipment based on a large language model, determining a target large language model specifically includes: Acquiring a preset geological environment and preset earthquake source parameters, and encoding the preset geological environment and the preset earthquake source parameters to obtain a coding sequence; Determining a preset large language model, and performing model training on the preset large language model according to the encoding sequence to obtain an initial large language model; The initial large language model is fine-tuned using LoRA technology to obtain a target large language model.

[0010] Optionally, the intelligent manufacturing method for seismic exploration equipment based on a large language model, wherein the retrieval enhancement technology is used to perform model training on the target large language model according to the marine geological knowledge graph to obtain a seismic exploration parameter adjustment model, further comprising: Obtaining vector data in the marine geological knowledge map, and performing similarity calculation on the vector data using a hierarchical clustering algorithm to obtain a similarity result; Classify the vector data according to the similarity result to obtain a visual data cluster; Tag metadata is added to each data sample in the visualization data cluster, wherein the tag metadata includes the region to which it belongs, the data type, the exploration target, the stratigraphic age, the acquisition method, and the source batch.

[0011] Optionally, the intelligent manufacturing method for seismic exploration equipment based on a large language model, wherein the use of retrieval enhancement technology to train the target large language model according to the marine geological knowledge graph to obtain a seismic exploration parameter adjustment model specifically includes: Obtaining a user query statement, and inputting the user query statement into the target large language model; Perform keyword analysis and weight calculation on the user query statement using the target large language model to obtain an initial candidate text; Performing high-dimensional vectorization encoding processing, semantic similarity calculation processing, and reordering processing on the initial candidate text to obtain semantic matching data; The semantic matching data is subjected to a targeted search process according to the tag metadata in the marine geological knowledge graph, and a search result is output. The training of the target large language model is completed, and a seismic exploration parameter adjustment model is obtained.

[0012] Optionally, the intelligent manufacturing method for seismic exploration equipment based on a large language model, wherein the acquiring of current seismic profile data, performing autonomous decision-making processing based on the current seismic profile data by the seismic exploration parameter adjustment model, outputting seismic exploration parameter configuration suggestions, and adjusting the initial seismic exploration parameters of the seismic exploration equipment based on the seismic exploration parameter configuration suggestions to obtain target seismic exploration equipment, specifically includes: Acquiring current seismic profile data, and inputting the current seismic profile data into the seismic exploration parameter adjustment model; Analyzing and processing the current seismic profile data using the seismic exploration parameter adjustment model to obtain the geological environment and seismic exploration targets; Keyword extraction and knowledge graph matching are performed on the geological environment and the seismic exploration target to obtain seismic exploration parameter configuration suggestions, and the initial seismic exploration parameters are adjusted according to the seismic exploration parameter configuration suggestions, wherein the seismic exploration parameter configuration suggestions include a target source frequency and a target receiving array layout.

[0013] In addition, to achieve the above-mentioned purpose, the present invention further provides a large language model-based intelligent manufacturing system for seismic exploration equipment, wherein the large language model-based intelligent manufacturing system for seismic exploration equipment includes: A multi-source seismic data acquisition module is used to acquire multi-source seismic data corresponding to initial seismic exploration parameters of the seismic exploration equipment, and perform data processing on the multi-source seismic data to obtain target multi-source seismic data; A knowledge graph construction module is used to construct a knowledge graph based on the target multi-source seismic data to obtain a marine geological knowledge graph; An autonomous learning module is used to determine a target large language model and use retrieval enhancement technology to perform autonomous learning processing on the target large language model based on the marine geological knowledge map to obtain a seismic exploration parameter adjustment model; an autonomous decision-making module, configured to obtain current seismic profile data, perform autonomous decision-making processing based on the current seismic profile data using the seismic exploration parameter adjustment model, output a seismic exploration parameter configuration suggestion, and adjust the initial seismic exploration parameters of the seismic exploration equipment based on the seismic exploration parameter configuration suggestion to obtain a target seismic exploration equipment; The optimization feedback module is used to obtain actual exploration data and optimize the seismic exploration parameter adjustment model carried by the target seismic exploration equipment based on the actual exploration data using a reinforcement learning strategy.

[0014] In addition, to achieve the above-mentioned purpose, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a large language model-based intelligent manufacturing program for seismic exploration equipment stored in the memory and runnable on the processor. When the large language model-based intelligent manufacturing program for seismic exploration equipment is executed by the processor, the steps of the large language model-based intelligent manufacturing method for seismic exploration equipment are implemented as described above.

[0015] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a large language model-based intelligent manufacturing program for seismic exploration equipment, and when the large language model-based intelligent manufacturing program for seismic exploration equipment is executed by a processor, the steps of the large language model-based intelligent manufacturing method for seismic exploration equipment are implemented as described above.

[0016] In the present invention, multi-source seismic data corresponding to the initial seismic exploration parameters of the seismic exploration equipment are obtained, and the multi-source seismic data are processed to obtain target multi-source seismic data; a knowledge graph is constructed based on the target multi-source seismic data to obtain a marine geological knowledge graph; a target large language model is determined, and the target large language model is autonomously learned based on the marine geological knowledge graph using retrieval enhancement technology to obtain a seismic exploration parameter adjustment model; current seismic profile data are obtained, and autonomous decision-making processing is performed based on the current seismic profile data using the seismic exploration parameter adjustment model to output seismic exploration parameter configuration suggestions, and the initial seismic exploration parameters of the seismic exploration equipment are adjusted based on the seismic exploration parameter configuration suggestions to obtain target seismic exploration equipment; actual exploration data are obtained, and a reinforcement learning strategy is used to optimize the seismic exploration parameter adjustment model carried in the target seismic exploration equipment based on the actual exploration data. The present invention constructs a knowledge graph and uses retrieval enhancement technology to combine the knowledge graph with a large language model for autonomous learning, thereby constructing a seismic exploration parameter adjustment model; the seismic exploration parameter adjustment model can make autonomous decisions based on actual geological conditions, thereby automatically adjusting the exploration parameters of marine earthquakes, significantly improving seismic imaging resolution, data acquisition efficiency and automation level. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a comparison chart of traditional penetration depth and resolution; Figure 2 This is a first flow chart of the intelligent manufacturing method of seismic exploration equipment based on a large language model of the present invention; Figure 3 This is a second flow chart of the intelligent manufacturing method for seismic exploration equipment based on a large language model of the present invention; Figure 4 This is a schematic diagram of the architecture of the marine geological intelligent manufacturing system of the intelligent manufacturing method of seismic exploration equipment based on the large language model of the present invention; Figure 5 This is a third flow chart of a preferred embodiment of the method for intelligent manufacturing of seismic exploration equipment based on a large language model of the present invention; Figure 6 This is a fourth flow chart of a preferred embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model of the present invention; Figure 7 This is a fifth flow chart of a preferred embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model of the present invention; Figure 8 This is a sixth flow chart of a preferred embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model of the present invention; Figure 9 This is a seventh flow chart of a preferred embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model of the present invention; Figure 10 This is a schematic diagram of the intelligent manufacturing system hardware and software architecture and data acquisition module composition of a preferred embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model of the present invention; Figure 11 This is a schematic diagram of a knowledge graph relationship network for seismic exploration parameter optimization according to a preferred embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model of the present invention; Figure 12 This is a schematic diagram of the intelligent seismic data processing and decision optimization process of a preferred embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model of the present invention; Figure 13 This is a schematic diagram of the retrieval and recommendation process of the RAG (retrieval enhanced generation) system of a preferred embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model of the present invention; Figure 14 It is a schematic diagram of a multi-level semantic retrieval system of a preferred embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model of the present invention; Figure 15 This is a schematic diagram of a model fine-tuning process of a preferred embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model of the present invention; Figure 16 It is a schematic diagram of the strategy iteration and fine-tuning process of the optimization feedback module of a preferred embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model of the present invention; Figure 17 This is a structural diagram of a preferred embodiment of the intelligent manufacturing system for seismic exploration equipment based on a large language model of the present invention; Figure 18 It is a structural diagram of a preferred embodiment of the terminal of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solutions and advantages of the present invention more clear and distinct, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0019] like Figure 1As shown in the figure, while high-frequency sources have good vertical resolution and are suitable for imaging shallow details, their rapid signal attenuation and limited penetration depth make it difficult to detect deep structures. Low-frequency sources, on the other hand, have strong penetration and are suitable for medium- to deep-layer detection, but their low resolution can easily miss shallow geological information. In actual exploration, it is often necessary to image both shallow and deep structures. However, existing marine seismic exploration systems lack the ability to automatically adjust source and receiver parameters according to different exploration objectives, making it difficult to strike a balance between imaging quality and efficiency.

[0020] Furthermore, the geological environment of the seabed varies significantly, with sediment types, stratigraphic structures, and pore properties varying across regions. Currently, seismic exploration operations are typically categorized into four types: conventional, high-resolution, ultra-high-resolution, and shallow-stratum profiling. Each operating mode employs a fixed configuration of source power, frequency, and receiver spacing, lacking the ability to adaptively optimize parameters for specific geological conditions. This results in the inability to effectively identify some key geological anomalies.

[0021] In summary, existing marine seismic exploration technologies suffer from fixed source and receiver parameters, limited imaging capabilities, poor environmental adaptability, and a lack of intelligent adjustment mechanisms. These issues make them difficult to meet the technical requirements of the next generation of marine resource development and exploration in complex deepwater environments. Therefore, an intelligent seismic exploration system is urgently needed that can adaptively adjust and optimize source parameters based on geological environmental characteristics to improve imaging resolution, expand detection depth, and enhance overall operational efficiency.

[0022] To address these issues, the present invention proposes a method and system for intelligent manufacturing of seismic exploration equipment based on a large language model. This system is particularly well-suited for seismic exploration missions under complex geological conditions (capable of acquiring and analyzing high-precision seismic data). It improves the resolution and data acquisition efficiency of marine seismic exploration, enabling precise detection and imaging of seafloor geological structures. By integrating autonomous learning, decision-making, and continuous optimization capabilities, the system intelligently adjusts source frequency, energy, and receiver placement parameters, significantly improving seismic imaging resolution, data acquisition efficiency, and automation. It is suitable for a wide range of operations, including high-resolution shallow-layer stratigraphic exploration and mid-layer structural imaging.

[0023] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 2 The first embodiment of the intelligent manufacturing method of seismic exploration equipment based on a large language model in the embodiment of the present invention includes: Basic scheme of the present invention: Step S10: Acquire multi-source seismic data corresponding to initial seismic exploration parameters of the seismic exploration equipment, and perform data processing on the multi-source seismic data to obtain target multi-source seismic data.

[0024] The present invention obtains, on the one hand, reflected wave signal data collected in real time by seismic exploration equipment during operation, and on the other hand, obtains data resources from different survey lines, historical operation areas, and public scientific research institutions, integrates them into multi-source seismic data, and performs data preprocessing on them. Afterwards, the data is reprocessed when stored in a database, thereby obtaining the target multi-source seismic data.

[0025] Step S20: construct a knowledge graph based on the target multi-source seismic data to obtain a marine geological knowledge graph.

[0026] In order to improve the semantic understanding and associative reasoning capabilities of data, this invention constructs a marine geological knowledge graph based on target multi-source seismic data, revealing the multidimensional logical relationship between data through a graph structure. The graph integrates historical exploration records, scientific research data and expert annotation results, and can achieve deep coupling of exploration parameters and geological background, and provide knowledge support for parameter recommendation and intelligent reasoning.

[0027] Step S30: determine a target large language model, and use retrieval enhancement technology to perform autonomous learning processing on the target large language model according to the marine geological knowledge graph to obtain a seismic exploration parameter adjustment model.

[0028] The present invention combines a large language model, a marine geological knowledge graph, and retrieval enhancement technology for autonomous learning, thereby constructing a seismic exploration parameter adjustment model. The seismic exploration parameter adjustment model has the ability to identify geological scenarios, generate parameter suggestions, and dynamically correct them, and can efficiently and accurately analyze user query statements and seismic profile data.

[0029] Step S40: Acquire current seismic profile data, perform autonomous decision-making based on the current seismic profile data through the seismic exploration parameter adjustment model, output seismic exploration parameter configuration suggestions, and adjust the initial seismic exploration parameters of the seismic exploration equipment based on the seismic exploration parameter configuration suggestions to obtain target seismic exploration equipment.

[0030] When the present invention receives real-time seismic profile data, it automatically analyzes the waveform characteristics, reflection structure and stratum interface information through the seismic exploration parameter adjustment model, and combines it with the existing knowledge graph for semantic matching to determine the geological environment type of the current operating area, such as soft sediment layer, hard bedrock, landslide or fault zone, so as to achieve accurate output of seismic exploration parameter configuration recommendations (automatically updating the source frequency, power and receiving array configuration to generate the final seismic exploration equipment parameters).

[0031] Step S50: Acquire actual exploration data, and use a reinforcement learning strategy to optimize the seismic exploration parameter adjustment model carried in the target seismic exploration equipment according to the actual exploration data.

[0032] Based on an optimized feedback mechanism, the present invention inputs the imaging effects, equipment feedback data and user evaluation results obtained during the actual exploration process into the system, and uses a reinforcement learning strategy to continuously adjust and optimize the seismic exploration parameter adjustment model, thereby improving the accuracy and stability of parameter recommendations and realizing closed-loop adaptive optimization.

[0033] In summary, this invention can intelligently optimize and recommend source frequency, power, and receiver array configuration based on existing seismic profile data, significantly improving the device's imaging resolution, penetration capability, and data acquisition efficiency. This invention also possesses continuous optimization capabilities, dynamically fine-tuning parameter strategies based on feedback, making it suitable for the intelligent manufacturing of high-precision seismic exploration equipment for complex submarine geological structures.

[0034] Further technical solutions of the present invention: See also Figure 3 The second embodiment of the intelligent manufacturing method of seismic exploration equipment based on a large language model in the embodiment of the present invention includes: S101, determining initial seismic exploration parameters of a seismic exploration device, collecting reflected wave signals according to the initial seismic exploration parameters, and obtaining an exploration environment corresponding to the reflected wave signals, wherein the initial seismic exploration parameters include an initial source frequency and an initial receiving array layout; like Figure 4 As shown, the present invention includes four core modules, namely data acquisition module, autonomous learning module, autonomous decision-making module and optimization feedback module. Each module operates in coordination to form an exploration system for adaptive optimization of dynamic geological environment.

[0035] The data acquisition module utilizes a multi-frequency source and a variable-spacing receiver array to dynamically adjust the source frequency between 300Hz and 3000Hz. It also supports flexible configuration of receiving hydrophone arrays with spacing between 0.5m and 2m to meet the needs of seismic exploration at varying depths and resolutions. This module acquires reflected wave signals from seafloor formations in real time and generates high-precision raw seismic profiles, providing a data foundation for subsequent geological interpretation and parameter optimization.

[0036] S102, acquiring existing seismic exploration data, and constructing multi-source seismic data based on the existing seismic exploration data, the initial seismic exploration parameters, the reflected wave signal, and the exploration environment; In addition to the function of physically collecting measured data, the present invention also integrates a structured storage mechanism and semantic parsing framework for instrument parameters, geological structures, and unstructured data. It supports the integration and storage of seismic instrument configuration data (such as source type, receiver layout method, and sampling accuracy) and geological exploration data (such as seismic profiles, well logging data, and fault distribution), thereby building a unified data management system.

[0037] S103, preprocessing and reprocessing the multi-source seismic data to obtain target multi-source seismic data; S104, wherein the pre-processing includes cleaning, standardization, anomaly detection and error correction; S105, the reprocessing includes standardization processing, format conversion processing, noise removal and completion processing, and accuracy verification processing.

[0038] See also Figure 5 The third embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model in the embodiment of the present invention includes: S201, determining a map node, wherein the map node includes stratigraphic units, reflection characteristics, sedimentary types, equipment parameters, and imaging effects; S202. Use a triplet model to construct a knowledge graph for the target multi-source seismic data according to the graph nodes to obtain a marine geological knowledge graph, wherein the triples in the triplet model include source parameters, geological characteristics, and exploration effects.

[0039] The present invention adopts a knowledge graph construction method to perform semantic modeling of core elements such as sedimentation, structure, lithologic combination, and oil and gas reservoir indications in the submarine geological environment, and constructs a geological relationship network driven by seismic data, thereby improving the modeling and reasoning capabilities of the complex geological background of the target area.

[0040] The present invention provides an autonomous learning module to undertake the intelligent learning and understanding of historical data, measured exploration data, expert experience, and literature. This module integrates retrieval enhancement generation technology and knowledge graph construction technology to form a unified data integration and semantic enhancement framework. The present invention first standardizes key information such as source parameters, receiver configuration, and exploration area characteristics, and then vectorizes and stores them to achieve efficient retrieval based on semantics and similarity. The knowledge graph uses "exploration target-geological unit-instrument parameters-imaging effect" as the core logical chain to construct an association network between geological elements, thereby improving the system's ability to understand complex marine geological conditions.

[0041] In practical applications, the autonomous learning module can automatically analyze various types of geological exploration data, including but not limited to seismic profiles, sedimentary environment descriptions, fault annotations, well location measurements, and multi-source geophysical survey results. Geological exploration data is primarily divided into two categories: seismic instrument parameters and geological exploration results data. Seismic instrument parameters include technical indicators such as source type, operating frequency, pulse energy, waveform characteristics, receiving array configuration, receiving spacing, layout method, dynamic range, and sensitivity. These parameters are standardized and stored as high-dimensional vectors. These vectors are then quickly matched with the geological characteristics of a specific area to intelligently recommend the optimal source plan and receiving array layout. Geological exploration results data, on the other hand, includes seismic profiles, well logging data, sedimentary environment characteristics, laboratory analysis, exploration reports, well location information, stratigraphic divisions, sediment physical parameters, and historical exploration lines. Before storage, all data undergoes noise removal, format standardization, and accuracy verification to ensure information integrity and consistency. Subsequently, the present invention uses vector retrieval technology to quickly compare and reuse similar cases in the current exploration area, providing highly relevant data support for parameter recommendation and strategy formulation.

[0042] See also Figure 6 The fourth embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model in the embodiments of the present invention includes: S301, obtaining a preset geological environment and preset earthquake source parameters, and encoding the preset geological environment and the preset earthquake source parameters to obtain a coding sequence; S302: Determine a preset large language model, and perform model training on the preset large language model according to the encoding sequence to obtain an initial large language model.

[0043] S303 , using LoRA technology (Low-Rank Adaptation, a parameter adjustment method for deep learning models) to fine-tune the initial large language model to obtain a target large language model.

[0044] The autonomous decision-making module, incorporated into this system, is its core intelligence. Leveraging a large language model and deep neural network model, it possesses multiple functions, including seismic data analysis, exploration strategy inference, and parameter generation. Through a dialogue-based model, exploration requirements are input and instrument parameters are obtained. This module receives real-time input and background data provided by the data acquisition and autonomous learning modules, automatically analyzes the geological characteristics of the exploration target area, and, combining historical best practices with model predictions, generates optimal recommendations for parameters such as source type, power, frequency combination, receiver spacing, and array configuration.

[0045] Furthermore, the present invention possesses adaptive control capabilities. During actual exploration, if the input measured data exhibits poor imaging quality, a low signal-to-noise ratio, or unclear structures, the autonomous decision-making module can automatically adjust parameter settings and provide new recommendations based on a feedback mechanism. For example, in shallow water with soft sediments, the system tends to recommend high-frequency, low-energy sources to improve near-surface resolution; whereas in deep water with hard formations, it automatically prioritizes low-frequency, high-energy sources to image deeper targets.

[0046] See also Figure 7 The fifth embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model in the embodiment of the present invention includes: S304, obtaining vector data in the marine geological knowledge map, and performing similarity calculation on the vector data using a hierarchical clustering algorithm to obtain a similarity result; S305, classifying the vector data according to the similarity result to obtain a visual data cluster; S306: Adding label metadata to each data sample in the visualization data cluster, wherein the label metadata includes the region to which it belongs, the data type, the exploration target, the stratigraphic age, the acquisition method, and the source batch.

[0047] See also Figure 8 The sixth embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model in the embodiments of the present invention includes: S307: Obtain a user query statement, and input the user query statement into the target large language model; S308: Perform keyword analysis and weight calculation on the user query using the target large language model to obtain initial candidate texts; S309, performing high-dimensional vectorization encoding processing, semantic similarity calculation processing, and reordering processing on the initial candidate text to obtain semantic matching data; S310: Performing a targeted search process on the semantic matching data according to the tag metadata in the marine geological knowledge graph, outputting the search results, completing the training of the target large language model, and obtaining a seismic exploration parameter adjustment model.

[0048] The optimization feedback module in this invention, based on reinforcement learning and closed-loop control strategies, continuously receives exploration results from actual operations, user evaluations, and measured data transmitted by equipment, automatically iteratively updating the model. This module uses a reward-penalty mechanism to evaluate the contribution of different parameter configurations to imaging results and mission objectives, and dynamically optimizes model parameters accordingly.

[0049] After each exploration mission, the system automatically compares the deviations between predicted parameters and measured data, analyzes the sources of imaging errors, and records these deviations in the model training feedback set. Through multiple rounds of optimization, the system has gradually developed a library of parameter configuration strategies suitable for different geological scenarios. The reinforcement learning framework ensures that the model can rapidly adapt and make judgments when faced with new geological scenarios or unusual areas.

[0050] In addition, the optimization feedback module supports a human-machine collaborative feedback mechanism. Geological experts can use the visual interface to mark key anomalies, correct parameter recommendation results, and then update the knowledge graph and parameter bias, gradually realizing human-machine joint training. This mechanism not only improves the model's explanatory power and expert acceptance, but also enhances the model's reliability and transparency in actual production scenarios. The present invention is equipped with a three-dimensional seismic imaging visualization module, which can display the impact of source layout, profile response, stratigraphic distribution, and parameter changes on imaging results in real time, supports users to interactively adjust plans, and significantly improves the efficiency of exploration strategy decision-making.

[0051] See also Figure 9 The seventh embodiment of the intelligent manufacturing method for seismic exploration equipment based on a large language model in the embodiments of the present invention includes: S401, obtaining current seismic profile data, and inputting the current seismic profile data into the seismic exploration parameter adjustment model; S402: Analyze and process the current seismic profile data using the seismic exploration parameter adjustment model to obtain a geological environment and a seismic exploration target.

[0052] Based on the large language model, the present invention analyzes the waveform characteristics and reflection structure information in the seismic profile data to identify the geological environment type and the corresponding seismic exploration target.

[0053] S403. Perform keyword extraction and knowledge graph matching on the geological environment and the seismic exploration target to obtain a seismic exploration parameter configuration suggestion, and adjust the initial seismic exploration parameters of the seismic exploration equipment according to the seismic exploration parameter configuration suggestion to obtain a target seismic exploration equipment, wherein the seismic exploration parameter configuration suggestion includes a target source frequency and a target receiving array layout.

[0054] The present invention uses a large language model to extract keywords from geological environments and exploration targets, and combines this with a marine geological knowledge graph to perform semantic matching and reasoning to generate seismic exploration parameter adjustment results. Based on this result, the initial parameters of the seismic exploration equipment are intelligently adjusted to obtain the target seismic exploration equipment configuration.

[0055] Example application: The large language model-based intelligent manufacturing method and system for seismic exploration equipment described in the present invention are aimed at complex marine geological environments. The system includes a data acquisition module, an autonomous learning module, an autonomous decision-making module and an optimization feedback module, forming a closed-loop intelligent optimization system for exploration strategies.

[0056] like Figure 10 As shown, the hardware platform of the present invention includes a multi-frequency seismic source, a variable-spacing receiving array, high-speed data communication, and a shipboard power supply system. The source system (corresponding to the multi-frequency seismic source) includes chirps (linear frequency modulation signals, whose frequency varies linearly with time), boomers (a type of marine seismic exploration source device, known as boomer sources), and sparkers (electric spark sources), enabling multi-band selection based on the depth and structural complexity of the geological target. The receiving system (corresponding to the variable-spacing receiving array) adopts an array layout, optimizing the array spacing and arrangement to achieve wide-area coverage and fine sampling of seismic signals. The data transmission module (corresponding to the high-speed data communication) supports high-speed, low-latency transmission, ensuring data integrity and real-time performance in harsh marine environments. The shipboard power supply system adapts to the needs of long-term operations and provides continuous and stable power support.

[0057] During the data acquisition phase, the system is deployed aboard an oceanographic survey vessel, acquiring seismic data through a combination of a selected source (e.g., Sparker, Boomer, or Chirp) and a variable-spacing underwater receiving array (e.g., a towed hydrophone array). Frequency adjustment between 300Hz and 3000Hz enables the selection of source devices, automatically adjusting transmission frequency and energy output based on target depth and geological complexity. The receiver array supports flexible spacing between 0.5m and 2.0m to accommodate varying resolution and penetration depth requirements.

[0058] like Figure 10 As shown in the figure, on the software platform, the data acquisition control unit generates an initial acquisition parameter combination scheme for the source and receiving array based on the geological model of the task area, historical exploration data, and current operation feedback. This is then applied to the control program. During operation, the system collects reflected wave signals in real time and performs preliminary imaging processing. The acquisition parameters, environmental information, and raw waveform data are simultaneously stored, providing a basis for subsequent model evaluation and parameter optimization.

[0059] The data management unit establishes a unified, multi-source data management system, integrating standardized data resources from various survey lines, historical operational areas, and publicly available research institutions. This data includes seismic profiles, well logs, sediment experiment results, and source configuration parameters. The software platform connects to the subsequent autonomous learning and optimization feedback modules through database expansion, creating a structured index system that improves retrieval efficiency and enhances subsequent autonomous learning capabilities.

[0060] The data preprocessing unit cleans, standardizes, detects anomalies, and corrects errors in acquired data. For example, it automatically identifies and removes abnormal waveforms and low-quality data. A format converter standardizes different data formats (such as SEG-Y, CSV, and LAS), improving data versatility and model compatibility. Historical data, after comparison and correction, is used as input for model pretraining, significantly reducing sample bias and improving prediction accuracy.

[0061] To enhance semantic understanding and associative reasoning capabilities of data, this paper introduces a mechanism for constructing a seismic exploration knowledge graph. This knowledge graph, with instrument parameters, seismic waveform characteristics, geological environment parameters, and imaging effects as core nodes, reveals the multidimensional logical relationships between data through a graph structure. This graph (the marine geology knowledge graph in this paper) integrates historical exploration records, scientific research data, and expert annotations. It enables a deep coupling of exploration parameters with geological context, providing knowledge support for parameter recommendations and intelligent reasoning.

[0062] For example, in a certain exploration area, the present invention identified an area dominated by interbedded sand and mud layers based on the knowledge graph. The source signal exhibited significant low-frequency attenuation in the high-porosity sandstone region. Through a semantic matching mechanism, the present invention linked the reflection characteristics with the coarse sand annotation information and adjusted the source frequency to 300Hz to enhance the imaging effect of the target layer's reflection signal, thereby improving reservoir identification accuracy. For another example, in areas with dense fault zones, the present invention automatically triggered a frequency up-regulation mechanism based on the fault-fluid channel coupling model in the knowledge graph and optimized the receiver layout density to enhance the resolution of fault details and the ability to identify interlayer relationships.

[0063] like Figure 11 As shown, to enhance knowledge reasoning capabilities, the present invention constructs a knowledge index (referred to as the marine geological knowledge graph in this invention) and establishes a network of associations between "source parameters, geological characteristics, and exploration results" based on a triplet model. Nodes in the graph include stratigraphic units, reflection characteristics, sedimentary types, equipment parameters, and imaging results. This graph is refined through expert annotation and system learning to form a knowledge structure that supports reasoning. This graph not only supports seismic anomaly identification, stratigraphic comparative analysis, and parameter optimization recommendations, but also enables model migration across different geological scenarios, improving system adaptability.

[0064] The knowledge graph, the semantic backbone of this invention, connects data, models, and application scenarios, supporting transfer learning between different geological environments and enabling cross-regional and cross-project knowledge reuse and strategy transfer. For example, the source placement strategy for shallow delta mud deposits in offshore wind power surveys can be applied to deepwater marginal basins with similar sediment properties and reflection structure characteristics in oil and gas development projects through graph similarity reasoning, thereby improving the system's adaptability and intelligent parameter recommendation capabilities in new work areas.

[0065] like Figure 12 As shown in the figure, in the autonomous learning stage, the present invention builds a knowledge-driven structure based on localized data, and uniformly encodes and deeply processes the multi-source heterogeneous data related to marine seismic exploration by integrating a vector database and a semantic indexing engine. The data covers historical seismic profile images, well logging curves, sediment physical parameters, exploration reports, source equipment parameters, etc. All data must be standardized, format converted, denoised and supplemented, and precision checked before being stored (i.e. Figure 12 Data preprocessing in , to ensure that the data quality meets the requirements of model training and semantic reasoning.

[0066] The present invention adopts Retrieval-Augmented Generation (RAG) technology to achieve deep semantic modeling and intelligent retrieval of unstructured data (such as Figure 12 and Figure 13 First, at the base layer, text data is vectorized using embedded semantic encoders (such as BERT (Bidirectional Encoder Representations from Transformers), a pre-trained language representation model, SBERT, or Sentence-BERT, a type of BERT model). This is combined with similarity calculations in a high-dimensional feature space to achieve semantic matching between complex geological text and instrument parameter descriptions. This invention integrates a local vector index memory (such as FAISS or Milvus) to enable intelligent knowledge retrieval completely independent of the external network environment, ensuring information security and engineering controllability.

[0067] Furthermore, at the logical level, data is categorized by type into two main groups: seismic instrument parameters and geological exploration results. Seismic instrument parameters include source type, excitation frequency, pulse duration, energy level, typical penetration depth, signal resolution, and receiver configuration (see Table 1). This data is structured and converted into high-dimensional vectors. Combined with the geological background characteristics of the exploration area, the system automatically recommends optimal source configurations and receiver placements to meet different exploration depth and resolution requirements (see Table 2).

[0068] Table 1: Comparison of parameters of commonly used source equipment for marine seismic exploration

[0069] Regional geological information data, classified as exploration results, includes profile images, sedimentary environment classifications, structural features, anomalous reflection characteristics, well location and depth information, logging response curves, experimental analysis results, structural divisions, sediment physical parameters, and historical exploration lines, as shown in Table 2. This paper uses regional-level partition indexing for this type of data and establishes a multidimensional index system based on well location or work area numbers. This supports semantic matching and inversion logic reasoning based on vectorized coding.

[0070] Table 2: Comparison of typical marine geological environment characteristics and applicable scenarios for seismic exploration

[0071] The present invention uses a hierarchical clustering algorithm to automatically classify vector data with high similarity, forming visual data clusters and building a multi-dimensional labeling system. Each data sample is attached with label metadata, including the region, data type, exploration target, stratigraphic age, acquisition method, source batch, etc., to facilitate user-defined filtering and improve the accuracy and interpretation ability of semantic retrieval. The vector retrieval engine outputs candidate knowledge items based on label filtering and similarity sorting for subsequent autonomous decision-making modules to call (such as Figure 13 shown).

[0072] In order to improve the retrieval efficiency and semantic matching accuracy of geological exploration data, based on the realization of data vector storage and the construction of a multi-dimensional label system, the present invention further introduces a multi-level semantic retrieval strategy at the strategy layer (such as Figure 14 This strategy combines the traditional keyword matching mechanism (the BM25 algorithm is an information retrieval algorithm that calculates the matching score between documents and query terms based on the similarity between the two terms) with semantic vector ranking methods (such as SBERT / BERT). It implements the full retrieval process from keyword triggering to semantic understanding, from coarse screening to fine tuning, and can effectively cope with multi-source heterogeneous and complex data environments.

[0073] The specific implementation steps of the multi-level semantic retrieval strategy are as follows: The first-level search layer uses the BM25 inverted index mechanism. The system analyzes and weights the user's query, quickly identifying semantically relevant initial candidate texts. This layer focuses on coverage speed and keyword accuracy, and is suitable for queries with clear structure and concentrated terminology, such as engineering search tasks like "fault zone + earthquake source." It can complete text screening in milliseconds.

[0074] The second-level search focuses on improving semantic accuracy. The system uses semantic embedding models (such as Sentence-BERT) to perform high-dimensional vector encoding on candidate data, and uses vector databases (such as FAISS, Milvus, or Qdrant) to calculate semantic similarity and re-rank. At this stage, the system can effectively identify sentence variants and contextual logical relationships, and is particularly suitable for analysis scenarios with inconsistent expressions, overlapping terminology, or non-standardized input. For example, when a user asks a question about "landfall source configuration strategy," the system can link to data related to "unstable sediment imaging optimization" to achieve precise matching at the semantic level.

[0075] The third-level search enables refined label-assisted screening. The system adds multi-dimensional label metadata to each data sample, covering factors such as region, data type, exploration target, stratigraphic age, acquisition method, and data source. Users can combine and restrict tags to further narrow the query scope and enhance professional matching. For example, when searching for data such as "deepwater area + landslide body + high-frequency earthquake source + grid layout", the system can achieve targeted search for specific engineering needs while retaining semantic matching, improving the controllability and interpretability of the results.

[0076] All data processing is performed in the local operating environment, and the system has the ability to maintain and dynamically update local data. New data can be entered into the system manually or through batch import. After standardization, vector encoding and map attachment are automatically completed. The system supports version control and update tracking mechanisms to ensure data traceability and version consistency. It also provides archiving suggestions for long-term unused data, achieving structural cleanup of data resources and space optimization.

[0077] In the autonomous decision-making stage, the large language model (LLM) integrated by the present invention, as well as the marine geological knowledge graph and retrieval enhancement generation system (such as Figure 12(as shown in the figure), the system has the ability to identify geological scenarios and generate parameter recommendations. Upon receiving real-time seismic profile data, the system automatically analyzes waveform characteristics, reflection structures, and stratigraphic interface information. Using existing knowledge graphs for semantic matching, it determines the geological environment type of the current operating area, such as soft sedimentary layers, hard bedrock, landslides, or fault zones. After identifying the geological environment, the system automatically generates optimal parameter recommendations, including source type, operating frequency, and pulse interval, and proposes receiver array layout. During parameter generation, the system also references successful cases in similar geological scenarios and uses RAG to select the optimal parameter combination. These recommendations not only take into account historical performance but also incorporate real-time analysis of current indicators such as signal-to-noise ratio and waveform continuity to ensure that the recommended solution is highly consistent with current geological conditions. The system supports a closed-loop decision-making model of "operation, analysis, and optimization," dynamically adjusting exploration strategies during data acquisition to improve operational efficiency and imaging accuracy.

[0078] The autonomous decision-making capability of the present invention relies on the following key mechanisms: First, LLM can simultaneously receive structured and unstructured data inputs, including historical seismic profiles, well logging data, sedimentary environment descriptions, equipment parameter logs, etc.; second, the present invention has a built-in geological knowledge graph module, which can analyze the correlation between geological body characteristics and equipment configuration, identify similar historical scenarios through triple matching, and retrieve parameter configuration solutions from past successful cases; then, the model generates source configuration recommendations based on similarity sorting and semantic reasoning, such as recommending the use of source equipment such as electric spark, boomer, sparker or air gun, as well as the corresponding frequency range, pulse interval and energy level; finally, the system supports user feedback mechanism and reinforcement learning method, which can self-correct and optimize the parameter recommendation strategy during continuous use.

[0079] For example, in a shallow soft sediment environment, the present invention identifies seismic wave characteristics characterized by a continuous but rapidly decaying reflective interface. Combined with historical knowledge graphs, the system finds that a Boomer source provides better resolution in similar environments. The system then recommends this source, with a frequency range of 1500Hz to 3000Hz and a pulse interval of 0.1 seconds. It also recommends a linear array layout with a 1-meter spacing between receivers to obtain clear images of shallow structures.

[0080] In another case, faced with a deepwater work area characterized by frequent fault activity and complex landslide structures, the system automatically read unconformities, inclined bedding, and reflection interference signals from the seismic profile. Comparing this with historical fault zone exploration data, it identified insufficient penetration of medium- and high-frequency signals. The system then recommended switching to an electric spark source with an operating frequency of 600-1000 Hz and a pulse interval of 0.5 seconds. It also recommended a fan-shaped arrangement of the receiver array, with spacing compressed to 0.5 meters, to enhance detection of fault boundaries.

[0081] This decision-making capability is implemented using a locally fine-tuned LLM (or LLaMA) model as an example. This invention utilizes LoRA technology to achieve model adaptation by training only a small number of learnable weights, without modifying the underlying model parameters. First, the system loads the underlying LLaMA model and its tokenizer (a core component that converts raw text into structured data that the model can process) and initializes the model structure. During the loading phase, undefined pad_tokens (padding tokens) in the tokenizer are set to eos_tokens (end-of-sequence tokens) to prevent alignment errors when batching inputs. Subsequently, the system injects a trainable LoRA-based subsystem, fine-tuning only key submodules to effectively control computing resource consumption.

[0082] like Figure 15 As shown, model training data is generated using a structured template. The content consists of prompt (request) and completion (response) fields, focusing on typical geological environments (such as shallow-water soft sediments, fault zones, and landslide areas) and recommended source parameters (such as type, frequency, and receiver layout). This data is then encoded into a sequence input using a Tokenizer. During training, the system constructs a DataLoader, batches the training data (i.e., batching), and initializes an optimizer (such as AdamW). Training proceeds through a standard cycle of forward propagation, backpropagation, and parameter updates. After each round of training, the system evaluates the loss value to optimize model performance. After training is complete, the model and its Tokenizer are saved and deployed in the actual inference system.

[0083] During the deployment inference process, the system interprets the user's input question as a geological environment and exploration target. Using keyword extraction and knowledge graph matching, it identifies the geological scenario category to which the query belongs. It then retrieves a table of recommended parameters and returns corresponding source and receiver configuration suggestions. If the system cannot identify a valid scenario, it prompts the user to provide additional information to improve matching accuracy.

[0084] The present invention also supports real-time feedback from the receiving system regarding source response and received waveform quality. Simultaneously, it calculates metrics such as signal-to-noise ratio, propagation delay, and reflection integrity, and compares them with model expectations. If anomalies are detected, such as signal weakening, poor coupling, or blurred imaging, the system instantly adjusts the source frequency, receiver spacing, or array configuration, and updates recommended parameters, achieving data-driven automated tuning.

[0085] In the optimization feedback phase, the system introduces the reinforcement learning (RL) mechanism to build a parameter tuning strategy. According to the seismic imaging effect and signal-to-noise ratio performance of each exploration mission, the source and receiver configuration (such as Figure 12 and Figure 16 The system automatically compares the imaging results with the expected model and updates its internal strategy value function based on evaluation metrics such as seismic profile clarity, key horizon identification accuracy, and signal continuity, thereby optimizing the next round of parameter recommendations. If insufficient image resolution is detected in a soft sedimentary environment, the system prioritizes increasing the proportion of high-frequency wave components. If penetration depth is insufficient, the system appropriately increases the low-frequency source power and widens the receiver array spacing to improve imaging depth.

[0086] User behavior feedback is also a core component of this invention. The system supports recording geological interpreters' modifications to prediction results, especially in areas such as determining key horizons or seismic reflection characteristics. Every user modification is automatically recorded as a feedback sample and entered into the system's fine-tuning process. For example, if a user manually adjusts the system's recommended source frequency or receiver spacing, the system will analyze and modify the recommendation logic based on the geological characteristics of the exploration area. This mechanism can gradually reduce prediction errors, enabling the system to continuously optimize with use.

[0087] The system supports continuous iterative optimization based on newly added samples or user operations. If a user is dissatisfied with a particular exploration suggestion, their modification behavior can be automatically interpreted as a new training sample. After review, it is added to the training set, triggering the fine-tuning process and forming a closed loop of human-machine collaborative enhancement. This feedback-driven mechanism continuously strengthens the model's understanding and adaptability to complex geological scenarios in actual operation.

[0088] like Figure 16 Figure 1 illustrates the user feedback and self-optimization process. The system first receives actual exploration data input and compares and analyzes it with the geological model. Based on the comparison results, the system automatically adjusts the source parameters and receiver layout to form a preliminary parameter recommendation scheme. If the user does not provide feedback, the system retains the original recommendation strategy. If the user adjusts the recommendation results, the system enters the feedback optimization process. During this process, the user modifies the source and receiver configuration through human-computer interaction, and the system fine-tunes the parameters accordingly and revises the original recommendation strategy. All feedback data is organized into training samples and added to the sample library for incremental fine-tuning of the large language model. The system retrains the model based on the updated samples, generating a more accurate recommendation strategy and incorporating the new strategy into the next round of recommendation processes, achieving continuous learning and capability evolution. This closed-loop process combines user behavior feedback, reinforcement learning, and fine-tuning mechanisms, effectively improving the model's adaptability and the accuracy of parameter recommendations in complex geological environments.

[0089] To further enhance its intelligence, the system integrates expert annotations and historical corrections into its parameter optimization strategy, adaptively adjusting the recommended range through a parameter offset attribution mechanism. For example, if the system identifies that a user has repeatedly lowered the recommended source frequency from 800Hz to 600Hz in a shallow gas reservoir, its strategy will automatically prioritize frequencies near 600Hz in similar scenarios and make fine adjustments based on the actual seismic signal acquisition results. If the noise level of the device's return signal is too high or the layer identification is unclear, the system can also automatically adjust the receiver deployment density, array structure, or source spacing to ensure continuous optimization of data acquisition quality.

[0090] Beneficial effects of the present invention: 1. The intelligent manufacturing system provided by this invention possesses significant technical advantages and application value in complex marine geological exploration scenarios. First, by incorporating an autonomous decision-making mechanism and adaptive parameter optimization technology, the system intelligently matches the optimal source frequency and receiving array layout based on the actual geological environment. Compared to traditional fixed-parameter acquisition schemes, this invention can achieve high-resolution imaging even in thick sand layers, significantly improving the accuracy of identifying shallow to mid-layer geological targets.

[0091] 2. Leveraging a large language model and a reinforcement learning framework, this system continuously learns from field survey data and automatically generates parameter recommendation strategies. During operation, the system dynamically adjusts key parameters such as source power, frequency range, and receiver spacing using real-time feedback, achieving closed-loop "acquisition-while-optimization" control, significantly reducing manual intervention and trial-and-error costs.

[0092] 3. The present invention integrates structured database and knowledge graph construction technology, making the implicit relationship between exploration parameters, geological structure characteristics, survey line layout and imaging effects explicit and structured in a semantic network manner, thereby enhancing the semantic understanding and logical reasoning ability of seismic data.

[0093] 4. This invention possesses a high degree of geological adaptability and scenario-shifting capabilities, making it suitable for a variety of exploration tasks, including deepwater oil and gas resource surveys, submarine landslide risk assessments, submarine pipeline site selection, offshore wind farm foundation surveys, and other engineering scenarios. The system maintains high-quality data acquisition capabilities in diverse environments, effectively reducing ineffective survey lines and duplication of work, while significantly improving operational efficiency and economic benefits.

[0094] Furthermore, this system possesses continuous learning and autonomous evolution capabilities, with built-in model fine-tuning and strategy optimization mechanisms. It automatically absorbs feedback data after each exploration mission and continuously refines its parameter recommendation logic. Through a closed-loop process of "self-learning-self-adaptation-self-optimization," the system's intelligence continuously increases with usage, offering excellent lifecycle scalability and system evolution potential.

[0095] Furthermore, if Figure 17 As shown, based on the above-mentioned intelligent manufacturing method of seismic exploration equipment based on a large language model, the present invention also provides an intelligent manufacturing system of seismic exploration equipment based on a large language model, wherein the intelligent manufacturing system of seismic exploration equipment based on a large language model includes: The multi-source seismic data acquisition module 51 is used to acquire multi-source seismic data corresponding to the initial seismic exploration parameters of the seismic exploration equipment, and perform data processing on the multi-source seismic data to obtain target multi-source seismic data; A knowledge graph construction module 52 is configured to construct a knowledge graph based on the target multi-source seismic data to obtain a marine geological knowledge graph; An autonomous learning module 53 is configured to determine a target large language model and perform autonomous learning processing on the target large language model based on the marine geological knowledge graph using a search enhancement technique to obtain a seismic exploration parameter adjustment model; An autonomous decision-making module 54 is configured to obtain current seismic profile data, perform autonomous decision-making processing based on the current seismic profile data using the seismic exploration parameter adjustment model, output a seismic exploration parameter configuration suggestion, and adjust the initial seismic exploration parameters of the seismic exploration equipment based on the seismic exploration parameter configuration suggestion to obtain a target seismic exploration equipment; The optimization feedback module 55 is used to obtain actual exploration data and optimize the seismic exploration parameter adjustment model carried by the target seismic exploration equipment according to the actual exploration data using a reinforcement learning strategy.

[0096] Furthermore, if Figure 18 As shown, based on the above-mentioned intelligent manufacturing method of seismic exploration equipment based on a large language model, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 18 Only some of the components of the terminal are shown, but it should be understood that implementation of all of the shown components is not required, and more or fewer components may be implemented instead.

[0097] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard drive or memory of the terminal. In other embodiments, the memory 20 may also be an external storage device of the terminal, such as a plug-in hard drive, a Smart Media Card (SMC), a Secure Digital (SD) card, a flash memory card, etc. equipped on the terminal. Furthermore, the memory 20 may include both the internal storage unit of the terminal and an external storage device. The memory 20 is used to store application software installed on the terminal and various types of data, such as program code of the installed terminal. The memory 20 may also be used to temporarily store data that has been output or is about to be output. In one embodiment, the memory 20 stores a large language model-based intelligent manufacturing program 40 for seismic exploration equipment. This large language model-based intelligent manufacturing program 40 for seismic exploration equipment can be executed by the processor 10, thereby implementing the large language model-based intelligent manufacturing method for seismic exploration equipment described in this application.

[0098] In some embodiments, the processor 10 can be a central processing unit (CPU), a microprocessor or other data processing chip, used to run the program code or process data stored in the memory 20, such as executing the intelligent manufacturing method of seismic exploration equipment based on a large language model.

[0099] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch screen, etc. The display 30 is used to display information on the terminal and to display a visual user interface.

[0100] In one embodiment, when the processor 10 executes the large language model-based intelligent manufacturing program 40 for seismic exploration equipment in the memory 20 , the steps of the large language model-based intelligent manufacturing method for seismic exploration equipment are implemented.

[0101] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a large language model-based intelligent manufacturing program for seismic exploration equipment, and when the large language model-based intelligent manufacturing program for seismic exploration equipment is executed by a processor, the steps of the large language model-based intelligent manufacturing method for seismic exploration equipment are implemented as described above.

[0102] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or terminal comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or terminal. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or terminal comprising the element.

[0103] Of course, those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware (such as a processor, controller, etc.) through a computer program. The program can be stored in a computer-readable storage medium that can be read by a computer. When executed, the program can include the processes in the above-described method embodiments. The computer-readable storage medium can be a memory, a magnetic disk, an optical disk, etc.

[0104] It should be understood that the application of the present invention is not limited to the above examples. For those skilled in the art, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A method for intelligent manufacturing of seismic exploration equipment based on a large language model, characterized in that: The intelligent manufacturing method for seismic exploration equipment based on a large language model includes: Acquiring multi-source seismic data corresponding to initial seismic exploration parameters of the seismic exploration equipment, and performing data processing on the multi-source seismic data to obtain target multi-source seismic data; Constructing a knowledge graph based on the target multi-source seismic data to obtain a marine geological knowledge graph; Determining a target large language model, and using retrieval enhancement technology to perform autonomous learning processing on the target large language model based on the marine geological knowledge graph to obtain a seismic exploration parameter adjustment model; Acquiring current seismic profile data, performing autonomous decision-making based on the current seismic profile data using the seismic exploration parameter adjustment model, outputting a seismic exploration parameter configuration suggestion, and adjusting the initial seismic exploration parameters of the seismic exploration equipment based on the seismic exploration parameter configuration suggestion to obtain a target seismic exploration equipment; Actual exploration data is acquired, and a reinforcement learning strategy is adopted to optimize the seismic exploration parameter adjustment model carried in the target seismic exploration equipment according to the actual exploration data.

2. The intelligent manufacturing method for seismic exploration equipment based on a large language model according to claim 1, characterized in that: The data processing includes preprocessing and reprocessing; The acquiring of multi-source seismic data corresponding to the initial seismic exploration parameters of the seismic exploration equipment and performing data processing on the multi-source seismic data to obtain target multi-source seismic data specifically includes: Determining initial seismic exploration parameters of the seismic exploration equipment, collecting reflected wave signals according to the initial seismic exploration parameters, and obtaining an exploration environment corresponding to the reflected wave signals, wherein the initial seismic exploration parameters include an initial source frequency and an initial receiving array layout; Acquiring existing seismic exploration data, and constructing multi-source seismic data based on the existing seismic exploration data, the initial seismic exploration parameters, the reflected wave signal, and the exploration environment; Preprocessing and reprocessing the multi-source seismic data to obtain target multi-source seismic data; The pre-processing includes cleaning, standardization, anomaly detection and error correction. The reprocessing includes standardization processing, format conversion processing, noise removal and completion processing, and accuracy verification processing.

3. The intelligent manufacturing method for seismic exploration equipment based on a large language model according to claim 1, characterized in that: The knowledge graph is constructed based on the target multi-source seismic data to obtain a marine geological knowledge graph, specifically including: Determining a map node, wherein the map node includes stratigraphic units, reflection characteristics, sedimentary types, equipment parameters, and imaging effects; A triplet model is used to construct a knowledge graph for the target multi-source seismic data according to the graph nodes to obtain a marine geological knowledge graph, wherein the triples in the triplet model include source parameters, geological characteristics and exploration effects.

4. The intelligent manufacturing method for seismic exploration equipment based on a large language model according to claim 1, characterized in that: Determining the target large language model specifically includes: Acquiring a preset geological environment and preset earthquake source parameters, and encoding the preset geological environment and the preset earthquake source parameters to obtain a coding sequence; Determining a preset large language model, and performing model training on the preset large language model according to the encoding sequence to obtain an initial large language model; The initial large language model is fine-tuned using LoRA technology to obtain a target large language model.

5. The intelligent manufacturing method for seismic exploration equipment based on a large language model according to claim 1, characterized in that: The method adopts the retrieval enhancement technology to perform autonomous learning processing on the target large language model according to the marine geological knowledge map to obtain a seismic exploration parameter adjustment model, and further includes: Obtaining vector data in the marine geological knowledge map, and performing similarity calculation on the vector data using a hierarchical clustering algorithm to obtain a similarity result; Classify the vector data according to the similarity result to obtain a visual data cluster; Tag metadata is added to each data sample in the visualization data cluster, wherein the tag metadata includes the region to which it belongs, the data type, the exploration target, the stratigraphic age, the acquisition method, and the source batch.

6. The intelligent manufacturing method for seismic exploration equipment based on a large language model according to claim 5, characterized in that: The use of the retrieval enhancement technology to autonomously learn the target large language model according to the marine geological knowledge graph to obtain a seismic exploration parameter adjustment model specifically includes: Obtaining a user query statement, and inputting the user query statement into the target large language model; Perform keyword analysis and weight calculation on the user query statement using the target large language model to obtain an initial candidate text; Performing high-dimensional vectorization encoding processing, semantic similarity calculation processing, and reordering processing on the initial candidate text to obtain semantic matching data; The semantic matching data is subjected to a directed search process according to the tag metadata in the marine geological knowledge graph, and the search results are output to obtain a seismic exploration parameter adjustment model.

7. The intelligent manufacturing method for seismic exploration equipment based on a large language model according to claim 1, characterized in that: The acquiring of current seismic profile data, performing autonomous decision-making processing based on the current seismic profile data by the seismic exploration parameter adjustment model, outputting a seismic exploration parameter configuration suggestion, and adjusting the initial seismic exploration parameters of the seismic exploration equipment based on the seismic exploration parameter configuration suggestion to obtain a target seismic exploration equipment specifically includes: Acquiring current seismic profile data, and inputting the current seismic profile data into the seismic exploration parameter adjustment model; Analyzing and processing the current seismic profile data using the seismic exploration parameter adjustment model to obtain the geological environment and seismic exploration targets; Keyword extraction and knowledge graph matching are performed on the geological environment and the seismic exploration target to obtain a seismic exploration parameter configuration suggestion, and the initial seismic exploration parameters of the seismic exploration equipment are adjusted according to the seismic exploration parameter configuration suggestion to obtain a target seismic exploration equipment, wherein the seismic exploration parameter configuration suggestion includes a target source frequency and a target receiving array layout.

8. An intelligent manufacturing system for seismic exploration equipment based on a large language model, characterized in that: The large language model-based intelligent manufacturing system for seismic exploration equipment includes: A multi-source seismic data acquisition module is used to acquire multi-source seismic data corresponding to initial seismic exploration parameters of the seismic exploration equipment, and perform data processing on the multi-source seismic data to obtain target multi-source seismic data; A knowledge graph construction module is used to construct a knowledge graph based on the target multi-source seismic data to obtain a marine geological knowledge graph; An autonomous learning module is used to determine a target large language model and use retrieval enhancement technology to perform autonomous learning processing on the target large language model based on the marine geological knowledge map to obtain a seismic exploration parameter adjustment model; an autonomous decision-making module, configured to obtain current seismic profile data, perform autonomous decision-making processing based on the current seismic profile data using the seismic exploration parameter adjustment model, output a seismic exploration parameter configuration suggestion, and adjust the initial seismic exploration parameters of the seismic exploration equipment based on the seismic exploration parameter configuration suggestion to obtain a target seismic exploration equipment; The optimization feedback module is used to obtain actual exploration data and optimize the seismic exploration parameter adjustment model carried by the target seismic exploration equipment based on the actual exploration data using a reinforcement learning strategy.

9. A terminal, characterized in that: The terminal includes: a memory, a processor, and a large language model-based intelligent manufacturing program for seismic exploration equipment stored in the memory and runnable on the processor. When the large language model-based intelligent manufacturing program for seismic exploration equipment is executed by the processor, the steps of the large language model-based intelligent manufacturing method for seismic exploration equipment as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a large language model-based intelligent manufacturing program for seismic exploration equipment. When the large language model-based intelligent manufacturing program for seismic exploration equipment is executed by a processor, the steps of the large language model-based intelligent manufacturing method for seismic exploration equipment as described in any one of claims 1 to 7 are implemented.

Images