Geotechnical engineering investigation report automatic generation method based on mapping knowledge domain

Abstract

The invention belongs to the technical field of geotechnical engineering, and discloses a geotechnical engineering investigation report automatic generation method based on a knowledge graph. Comprising the following steps: acquiring a multi-source geotechnical engineering investigation data set, and carrying out data cleaning and space correlation preprocessing; constructing a geotechnical engineering field knowledge graph, and forming a project specific knowledge sub-graph; performing multi-dimensional stratigraphic analysis based on the knowledge sub-graph, generating a three-dimensional model and obtaining a key profile map; geological condition reasoning is carried out in combination with historical investigation cases, and abnormal geological bodies and unfavorable geological phenomena are identified; engineering parameter intelligent recommendation is carried out based on the geological risk assessment result; carrying out adaptive selection and personalized configuration on the geotechnical engineering investigation report template library; and finally, carrying out intelligent assembly to generate a complete investigation report. According to the intelligent analysis method driven by the knowledge graph, automatic generation of the reconnaissance report is achieved, and the report compiling efficiency and quality are improved.

Description

Technical Field

[0001] The present invention relates to the field of geotechnical engineering technology, and more specifically, to a method for automatically generating geotechnical engineering investigation reports based on a knowledge graph. Background Art

[0002] Geotechnical engineering investigation is a crucial foundational task in engineering construction, and its results are directly related to the rationality of engineering design and the safety of construction. With the acceleration of urbanization and the increase in large-scale engineering projects, geotechnical engineering investigations face increasingly complex geological conditions and higher technical requirements. Currently, the preparation of geotechnical engineering investigation reports is primarily manual, relying on the personal experience and professional judgment of engineering technicians.

[0003] Due to the diverse sources of survey data, non-uniform data formats, and varying quality, data integration is difficult, hindering the full realization of data value. In actual engineering applications, geological condition analysis often relies too heavily on individual engineers' experience, lacking systematicity and comprehensiveness. This makes it difficult to accurately identify anomalous geological bodies (such as karst and weak interlayers) in complex strata and assess potential geological risks, which can easily lead to engineering safety hazards. The determination of engineering parameters often fails to fully consider the impact of geological risk factors, often relying on empirical formulas or simple statistical methods. This results in a lack of scientific basis for parameter recommendations, impacting the reliability of engineering designs. In report preparation, fixed templates are commonly used, making it difficult to flexibly adjust them to the geological characteristics and engineering requirements of different projects, affecting the report's relevance and practicality. Furthermore, a large amount of historical survey experience and case data is not effectively utilized, lacking knowledge transfer and experience reuse mechanisms, leading to repeated research on similar engineering problems and reducing survey efficiency. These problems are particularly prominent in large, complex engineering projects, such as subway projects, which span multiple geological units and face complex and variable geological conditions. Traditional manual compilation methods struggle to ensure the quality and efficiency of survey reports, potentially impacting engineering decision-making and construction safety.

[0004] In view of this, the present invention proposes a method for automatically generating geotechnical engineering survey reports based on knowledge graph to solve the above problems. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned objectives, the present invention provides the following technical solution: a method for automatically generating a geotechnical engineering investigation report based on a knowledge graph, comprising:

[0006] Step S1: Acquire a multi-source geotechnical engineering investigation dataset; perform data cleaning and quality assessment on the multi-source geotechnical engineering investigation dataset, and perform spatial correlation preprocessing to obtain standardized geotechnical engineering data;

[0007] Step S2: Construct a geotechnical engineering domain knowledge graph; perform entity extraction and relationship recognition on the standardized geotechnical engineering data, and map it to the geotechnical engineering domain knowledge graph to form a project-specific knowledge subgraph;

[0008] Step S3: Perform multi-dimensional stratigraphic analysis based on the project-specific knowledge subgraph to generate a three-dimensional stratigraphic spatial distribution model; dynamically slice the three-dimensional stratigraphic spatial distribution model to obtain key cross-sections;

[0009] Step S4: Obtain historical survey cases, perform geological condition reasoning based on the project-specific knowledge subgraph and historical survey cases, identify abnormal geological bodies and adverse geological phenomena, and generate geological risk assessment results;

[0010] Step S5: Based on the geological risk assessment results and the project-specific knowledge subgraph, intelligent engineering parameter recommendations are made to generate hierarchical and zoned foundation bearing capacity parameters and foundation treatment suggestions, i.e., engineering parameter recommendation results;

[0011] Step S6: Obtain a geotechnical engineering investigation report template library; adaptively select and personalize the geotechnical engineering investigation report template library based on the project-specific knowledge subgraph to generate a project-specific report framework;

[0012] Step S7: Based on the project-specific report framework, key profiles, geological risk assessment results, and engineering parameter recommendation results are intelligently assembled to generate a complete geotechnical engineering investigation report.

[0013] Preferably, the specific steps of step S1 are:

[0014] Acquire a multi-source geotechnical engineering investigation dataset consisting of drilling records, in-situ tests, laboratory tests, and historical investigation data;

[0015] Identify and classify data types of multi-source geotechnical engineering survey data sets to obtain structured survey data and unstructured description data;

[0016] Perform outlier detection and missing value identification on structured survey data, marking low-quality data points;

[0017] Perform interpolation repair or reliability assessment on low-quality data points based on geostatistical models to obtain corrected structured survey data;

[0018] Perform natural language processing on unstructured description data to extract key geological description features and obtain structured geological description features;

[0019] The modified structured survey data and structured geological description features are preprocessed by spatial coordinate association to obtain standardized geotechnical engineering data.

[0020] Preferably, the specific steps of step S2 are:

[0021] Construct an ontology model for geotechnical engineering that includes geotechnical classification system, geological structure relationship, and engineering property association;

[0022] Based on the geotechnical engineering ontology model, a knowledge graph is constructed in the geotechnical engineering field, which includes the relationship network between stratum types, lithologic characteristics, physical and mechanical parameters, and engineering properties.

[0023] Perform entity recognition on standardized geotechnical engineering data to extract entities of strata, lithology, parameters and engineering properties;

[0024] Reasoning about the relationships between entities based on spatial location and geological rules, and establishing associations between entities;

[0025] The extracted entities and relationships are mapped to the geotechnical engineering domain knowledge graph to form a project-specific knowledge subgraph that reflects the characteristics of the current project.

[0026] Preferably, the specific steps of step S3 are:

[0027] Extracting stratum spatial distribution information based on project-specific knowledge subgraph;

[0028] Perform three-dimensional interpolation on the stratum spatial distribution information to generate an initial stratum spatial distribution model;

[0029] Based on geological sedimentary laws and tectonic evolution knowledge, the initial stratigraphic spatial distribution model is constrained and optimized to generate a geological rationality correction model;

[0030] The geological rationality correction model is integrated with the engineering parameter distribution to generate a three-dimensional model of the stratum spatial distribution;

[0031] According to engineering design requirements and the importance of geological features, the three-dimensional model of the spatial distribution of the strata is dynamically sliced ​​to obtain key cross-sections.

[0032] Preferably, the specific steps of step S4 are:

[0033] Extract geological anomaly characteristic indicators from the project-specific knowledge subgraph;

[0034] Build a bad geological identification rule base, which includes characteristic patterns of various bad geological phenomena;

[0035] Based on the adverse geological identification rule library, pattern matching of geological anomaly characteristic indicators is performed to identify potential adverse geological phenomena;

[0036] Search the historical exploration case database to extract historical cases with similar geological conditions to the current project and their geological risk records, namely historical case risk records;

[0037] The currently identified potential adverse geological phenomena and historical case risk records are integrated to generate graded and zoned geological risk assessment results.

[0038] Preferably, the specific steps of step S5 are:

[0039] Extract physical and mechanical parameter data of each layer based on the project-specific knowledge subgraph;

[0040] Determine the key engineering parameter indicator set based on the design requirements of different engineering types;

[0041] Calculate preliminary engineering parameter recommendations based on physical and mechanical parameter data and key engineering parameter indicator sets;

[0042] Based on the geological risk assessment results, the preliminary engineering parameter recommendations are revised to generate risk-corrected engineering parameters;

[0043] Based on geotechnical specifications, the rationality of risk-corrected engineering parameters is verified to generate layered and zoned foundation bearing capacity parameters;

[0044] Generate targeted foundation treatment recommendations based on the layered and zoned foundation bearing capacity parameters and geological risk assessment results.

[0045] Preferably, step S6 specifically comprises the following steps:

[0046] Access a library of geotechnical investigation report templates categorized by project type, geological complexity, and report purpose;

[0047] Extract the engineering type, geological complexity and report purpose of the current project based on the project-specific knowledge subgraph, i.e., project characteristics;

[0048] Calculate the similarity of the geotechnical engineering investigation report template library based on the extracted project features and select the most matching basic report template;

[0049] Analyze the key geological features and engineering points in the project-specific knowledge sub-map to determine the key content of the report;

[0050] Optimize chapters and adjust the structure of the basic report template based on the key content of the report to generate a project-specific report framework.

[0051] Preferably, the specific steps of step S7 are:

[0052] Build report content filling rules based on the project-specific report framework;

[0053] Map key cross-sections to the stratigraphic description section of the project-specific reporting framework according to the report content filling rules;

[0054] Map the geological risk assessment results to the engineering geological issues section of the project-specific reporting framework according to the report content filling rules;

[0055] Map the engineering parameter recommendation results to the foundation and basic sections of the project-specific report framework according to the report content filling rules; then obtain the report text;

[0056] Generate supporting diagrams and appendix materials based on project-specific knowledge subgraphs;

[0057] Perform natural language optimization and professional terminology standardization on the generated report text;

[0058] Integrate report text, supporting charts and appendix materials to generate a complete geotechnical engineering investigation report.

[0059] Preferably, the steps for analyzing the key geological features and engineering points in the project-specific knowledge sub-graph and determining the key contents of the report are:

[0060] Identify key geological nodes and relationship strengths through topological analysis of project-specific knowledge subgraphs;

[0061] Calculate geological feature importance scores based on key geological nodes and relationship strengths;

[0062] Identify key geological features of the project based on geological feature importance scores;

[0063] Analyze the impact of key geological features on engineering construction in combination with the engineering type knowledge base;

[0064] Comprehensively analyze the impact on engineering construction and the importance score of geological features to determine the key contents of the report.

[0065] A device for automatically generating geotechnical engineering investigation reports based on a knowledge graph, the device comprising: a memory and at least one processor, wherein the memory stores instructions;

[0066] The at least one processor calls the instructions in the memory to enable the automatic generation of geotechnical engineering investigation reports based on knowledge graph to execute the automatic generation method of geotechnical engineering investigation reports based on knowledge graph as described.

[0067] The technical effects and advantages of the method for automatically generating geotechnical engineering investigation reports based on knowledge graphs of the present invention are as follows:

[0068] The present invention constructs a knowledge graph in the field of geotechnical engineering, organically combines professional knowledge with project data, forms a project-specific knowledge subgraph, and provides a knowledge basis for subsequent intelligent analysis; based on the project-specific knowledge subgraph, multi-dimensional stratigraphic analysis and three-dimensional model construction are carried out to make the description of geological conditions more accurate and intuitive; through geological condition reasoning and risk assessment, the system identifies potential geological risks and provides a basis for risk prevention and control for engineering design; intelligent recommendation of engineering parameters combines geological conditions and risk factors to make parameter recommendations more reasonable and reliable; the adaptive generation of project-specific report frameworks and intelligent assembly of content ensure the professionalism and pertinence of reports. Compared with traditional geotechnical engineering survey report preparation methods, the present invention has the following advantages: knowledge-driven intelligent analysis improves the accuracy and professionalism of geological condition interpretation; automated report generation greatly improves work efficiency and reduces the time and workload of manual writing; standardized processing procedures ensure the consistency and reliability of report quality; personalized report customization makes the report more in line with project characteristics and needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 Schematic diagram of the method for automatically generating geotechnical engineering investigation reports based on knowledge graph of the present invention;

[0070] Figure 2 Detailed implementation flow chart of step S1 of the present invention;

[0071] Figure 3 Schematic diagram of the automatic generation system of geotechnical engineering investigation reports based on knowledge graph of the present invention. DETAILED DESCRIPTION

[0072] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0073] Example 1

[0074] This application provides a method. The execution entities of the method include, but are not limited to, mechanical equipment, data processing platforms, cloud server nodes, network upload devices, etc. that are equipped with a system implementing the method, which can be considered as general computing nodes of this application. The data processing platform includes, but is not limited to, at least one of a geotechnical engineering management system, a geological information management system, and a cloud data management system.

[0075] See also Figure 1 The present invention provides a method for automatically generating a geotechnical engineering investigation report based on a knowledge graph, comprising the following steps:

[0076] Step S1: Acquire a multi-source geotechnical engineering investigation dataset; perform data cleaning and quality assessment on the multi-source geotechnical engineering investigation dataset, and perform spatial correlation preprocessing to obtain standardized geotechnical engineering data;

[0077] Step S2: Construct a geotechnical engineering domain knowledge graph; perform entity extraction and relationship recognition on the standardized geotechnical engineering data, and map it to the geotechnical engineering domain knowledge graph to form a project-specific knowledge subgraph;

[0078] Step S3: Perform multi-dimensional stratigraphic analysis based on the project-specific knowledge subgraph to generate a three-dimensional stratigraphic spatial distribution model; dynamically slice the three-dimensional stratigraphic spatial distribution model to obtain key cross-sections;

[0079] Step S4: Obtain historical survey cases, perform geological condition reasoning based on the project-specific knowledge subgraph and historical survey cases, identify abnormal geological bodies and adverse geological phenomena, and generate geological risk assessment results;

[0080] Step S5: Based on the geological risk assessment results and the project-specific knowledge subgraph, intelligent engineering parameter recommendations are made to generate hierarchical and zoned foundation bearing capacity parameters and foundation treatment suggestions, i.e., engineering parameter recommendation results;

[0081] Step S6: Obtain a geotechnical engineering investigation report template library; adaptively select and personalize the geotechnical engineering investigation report template library based on the project-specific knowledge subgraph to generate a project-specific report framework;

[0082] Step S7: Based on the project-specific report framework, key profiles, geological risk assessment results, and engineering parameter recommendation results are intelligently assembled to generate a complete geotechnical engineering investigation report.

[0083] The present invention improves data quality and consistency by cleaning and spatial correlation preprocessing multi-source geotechnical engineering survey data, providing a reliable data basis for subsequent analysis; constructs a knowledge graph in the field of geotechnical engineering and forms a project-specific knowledge subgraph, effectively capturing the complex relationships between geological entities and improving data semantic understanding capabilities; multi-dimensional stratigraphic analysis and three-dimensional model construction achieve accurate expression of stratigraphic spatial distribution, and dynamic slicing processing ensures the representativeness of key profiles; geological condition reasoning based on project-specific knowledge subgraphs and historical survey cases improves the recognition accuracy of abnormal geological bodies and adverse geological phenomena; intelligent recommendation of engineering parameters takes into account geological risk factors, and the generated foundation bearing capacity parameters and processing suggestions are more scientific and reliable; the adaptive selection and personalized configuration of the report template library ensure that the generated report framework conforms to the project characteristics; and finally realizes the intelligent assembly of survey reports, greatly improving the efficiency and quality of report generation.

[0084] In an embodiment of the present invention, the steps of the method for automatically generating a geotechnical engineering investigation report based on a knowledge graph include:

[0085] Step S1: Acquire a multi-source geotechnical engineering investigation dataset; perform data cleaning and quality assessment on the multi-source geotechnical engineering investigation dataset, and perform spatial correlation preprocessing to obtain standardized geotechnical engineering data;

[0086] In this embodiment, data from various sources are collected from geotechnical engineering investigation projects, including drilling records (such as borehole histograms, core photos, etc.), in-situ test data (such as standard penetration tests, static penetration tests, etc.), indoor test data (such as soil physical and mechanical properties, rock mechanics tests, etc.), geophysical investigation data (such as resistivity measurements, seismic wave tests, etc.) and historical investigation data. These heterogeneous multi-source data are integrated into a comprehensive geotechnical engineering investigation data set. The integrated data set is cleaned, including outlier detection, missing value processing, duplicate data removal, etc. Geostatistical methods are used to evaluate data quality, identify low-quality data points and mark them. Spatial correlation preprocessing is performed on the cleaned data based on the spatial coordinate system (latitude and longitude, elevation) to ensure spatial consistency and correlation of data from different sources. Through the above processing, standardized geotechnical engineering data is formed, providing a basis for subsequent knowledge graph construction and analysis.

[0087] Step S2: Construct a geotechnical engineering domain knowledge graph; perform entity extraction and relationship recognition on the standardized geotechnical engineering data, and map it to the geotechnical engineering domain knowledge graph to form a project-specific knowledge subgraph;

[0088] In this embodiment, an ontology model of the geotechnical engineering field is first constructed, which includes knowledge such as the geotechnical classification system, geological structure relationships, and engineering property associations. A knowledge graph of the geotechnical engineering field is constructed based on the ontology model to form a knowledge network that includes the relationships between stratum types, lithologic characteristics, physical and mechanical parameters, and engineering properties. Natural language processing and pattern recognition are performed on standardized geotechnical engineering data to extract entities such as strata, lithologies, and engineering parameters. Based on spatial position relationships and geological rules, the association relationships between entities are inferred, such as "coverage", "inclusion", "fracture", etc. The extracted entities and relationships are mapped to the domain knowledge graph to form a knowledge subgraph that reflects the characteristics of the current project, that is, a project-specific knowledge subgraph. This knowledge subgraph contains project-specific geological conditions, engineering characteristics, and risk factors, providing knowledge support for subsequent analysis.

[0089] Step S3: Perform multi-dimensional stratigraphic analysis based on the project-specific knowledge subgraph to generate a three-dimensional stratigraphic spatial distribution model; dynamically slice the three-dimensional stratigraphic spatial distribution model to obtain key cross-sections;

[0090] In this embodiment, the stratum spatial distribution information is extracted from the project-specific knowledge subgraph, including the stratigraphic sequence, thickness and spatial position of each borehole. The Kriging spatial interpolation algorithm or the improved triangular mesh method is used to perform three-dimensional interpolation on discrete stratigraphic data points to generate an initial stratigraphic spatial distribution model. Based on the geological sedimentation laws and tectonic evolution knowledge, the initial model is constrained and optimized to ensure that the model conforms to the geological laws. The distribution of physical and mechanical parameters is integrated with the optimized stratigraphic model to generate a three-dimensional stratigraphic spatial distribution model that integrates geological attributes and engineering characteristics. According to the engineering design requirements (such as building layout, underground structure direction, etc.) and the importance of geological features, the three-dimensional model is dynamically sliced ​​to obtain key profiles that can reflect the geological conditions and engineering risks to the greatest extent. These profiles provide core graphic data for the preparation of subsequent reports.

[0091] Step S4: Obtain historical survey cases, perform geological condition reasoning based on the project-specific knowledge subgraph and historical survey cases, identify abnormal geological bodies and adverse geological phenomena, and generate geological risk assessment results;

[0092] In this embodiment, characteristic indicators of geological anomalies, such as sudden changes in strata and areas with abnormal parameters, are extracted from the project-specific knowledge subgraph. A rule base for identifying adverse geology is constructed, which contains characteristic patterns of various types of adverse geological phenomena such as karst, weak interlayers, landslides, faults, etc. Pattern matching and reasoning are performed on the extracted geological anomaly characteristics based on the rule base to identify potential adverse geological phenomena. Cases similar to the geological conditions of the current project are retrieved from the historical survey case library, and their geological risk records and handling experience are extracted. By combining case reasoning and knowledge graph reasoning, the geological risks in the project area are comprehensively assessed to generate hierarchical and zoned geological risk assessment results. The assessment results include risk type, spatial distribution, risk level and potential impact, providing risk warnings for engineering design and construction.

[0093] Step S5: Based on the geological risk assessment results and the project-specific knowledge subgraph, intelligent engineering parameter recommendations are made to generate hierarchical and zoned foundation bearing capacity parameters and foundation treatment suggestions, i.e., engineering parameter recommendation results;

[0094] In this embodiment, the physical and mechanical parameter data of each stratum, such as compression modulus, internal friction angle, cohesion, etc., are extracted from the project-specific knowledge subgraph. According to the design requirements of the project type (such as high-rise buildings, bridges, underground projects, etc.), the key engineering parameter indicator set is determined. Based on the physical and mechanical parameters and empirical formulas, preliminary engineering parameter recommendation values ​​are calculated, such as foundation bearing capacity characteristic values, settlement calculation parameters, etc. In combination with the aforementioned geological risk assessment results, the preliminary parameter recommendation values ​​are revised to reduce the parameter values ​​in the risk area and generate risk-corrected engineering parameters. Based on geotechnical engineering specifications and standards, the rationality of the risk correction parameters is verified to ensure that the parameters meet the specification requirements and the safety of the project. Based on the stratum characteristics and risk assessment results, targeted foundation treatment suggestions are generated, such as replacement, pile foundation, foundation reinforcement and other measures. Combining the above content, a hierarchical and partitioned engineering parameter recommendation result is formed to provide reliable parameter support for engineering design.

[0095] Step S6: Obtain a geotechnical engineering investigation report template library; adaptively select and personalize the geotechnical engineering investigation report template library based on the project-specific knowledge subgraph to generate a project-specific report framework;

[0096] In this embodiment, a geotechnical engineering investigation report template library is constructed, which is classified by project type (such as buildings, roads, bridges, etc.), geological complexity (simple, medium, complex) and report purpose (such as feasibility study, preliminary design, construction drawing design, etc.). The characteristics of the current project, such as the project type, geological complexity and report purpose, are extracted from the project-specific knowledge subgraph. Based on the extracted project characteristics, the similarity of the templates in the template library is calculated, and the most matching basic report template is selected. The key geological features and engineering points in the project-specific knowledge subgraph are analyzed to determine the content that needs to be emphasized in the report. Based on the determined key content, the basic template is optimized and the structure is adjusted to generate an exclusive report framework suitable for the characteristics of the current project. The framework includes the overall structure of the report, the layout of the chapters and the key content prompts, providing a skeleton for subsequent report generation.

[0097] Step S7: Based on the project-specific report framework, key profiles, geological risk assessment results, and engineering parameter recommendation results are intelligently assembled to generate a complete geotechnical engineering investigation report.

[0098] In this embodiment, report content filling rules are constructed based on the project-specific report framework to guide how various analysis results are mapped into the report. The key profiles are mapped to the stratigraphic description section of the report framework according to the filling rules to intuitively display the geological conditions of the site. The geological risk assessment results are mapped to the engineering geological problem section to detail the adverse geological phenomena and potential risks existing on the site. The engineering parameter recommendation results are mapped to the foundation and foundation section to provide layered and zoned foundation bearing capacity parameters and foundation treatment suggestions. Based on the project-specific knowledge subgraph, supporting charts (such as geological profiles, physical and mechanical parameter statistical charts, etc.) and appendix materials (such as drilling bar charts, in-situ test results, etc.) are generated. The generated report text is optimized for natural language and professional terminology is standardized to ensure professional and accurate language expression. The report text, supporting charts and appendix materials are integrated to generate a complete geotechnical engineering survey report that meets industry standards and technical requirements.

[0099] See Figure 2 , which is a flowchart of detailed implementation steps of step S1. In a specific embodiment, the process of executing step S1 may specifically include the following steps:

[0100] Acquire a multi-source geotechnical engineering investigation dataset consisting of drilling records, in-situ tests, laboratory tests, and historical investigation data;

[0101] Identify and classify data types of multi-source geotechnical engineering survey data sets to obtain structured survey data and unstructured description data;

[0102] Perform outlier detection and missing value identification on structured survey data, marking low-quality data points;

[0103] Perform interpolation repair or reliability assessment on low-quality data points based on geostatistical models to obtain corrected structured survey data;

[0104] Perform natural language processing on unstructured description data to extract key geological description features and obtain structured geological description features;

[0105] The modified structured survey data and structured geological description features are preprocessed by spatial coordinate association to obtain standardized geotechnical engineering data.

[0106] Specifically, we first acquired a multi-source geotechnical engineering investigation dataset, including drilling records, in-situ tests, laboratory tests, and historical investigation data. Drilling records include borehole histograms, drilling logs, stratigraphic descriptions, and sampling records; in-situ test data include standard penetration tests (SPTs), cone penetration tests (CPTs), in-situ shear tests, and wave velocity tests; laboratory test data include geotechnical tests, rock mechanics tests, and soil chemical analysis data; and historical investigation data include surrounding engineering investigation reports, regional geological data, and historical geological disaster records.

[0107] The data types of the acquired multi-source geotechnical engineering investigation datasets are identified and classified into structured investigation data and unstructured descriptive data. Structured investigation data refers to data with a clear format and numerical values, such as borehole coordinates, formation depth, and test parameter values. Unstructured descriptive data refers to text-based descriptive data, such as formation lithology descriptions, field records, and descriptions of geological phenomena.

[0108] For structured survey data, outlier detection algorithms are applied to identify outliers and unreasonable values ​​in the data. Common outlier detection methods include the Z-Score method, the interquartile range (IQR) method, and the local outlier factor (LOF) method. Taking the Z-Score method as an example, the deviation of each data point from the overall mean is calculated and measured in units of standard deviation:

[0109] Where Z_i is the Z-score value of data point x_i, μ is the mean of the dataset, and σ is the standard deviation of the dataset. When |Z_i| exceeds a preset threshold (usually 3), it is identified as an outlier. Missing values ​​in the dataset are also identified, marking low-quality data points.

[0110] Interpolation repair or reliability assessment of low-quality data points is performed based on geostatistical models. Geostatistical models such as Kriging take into account spatial autocorrelation and are suitable for processing geological data. The basic formula of Kriging interpolation is:

[0111] Z(s_0)=∑ j λ_j·Z(s_j); where Z(s_0) is the attribute value at the location s_0 to be estimated, Z(s_j) is the attribute value at the known location s_j, λ_j is the weight coefficient corresponding to location s_j, and j is the index of the location. The weight coefficient λ_j is calculated using the variogram to minimize the variance of the interpolated estimate. Kriging interpolation is used to correct missing values ​​and outliers, resulting in corrected structured survey data.

[0112] For unstructured descriptive data, natural language processing techniques are applied to extract key geological descriptive features. Specific methods include text segmentation, keyword extraction, named entity recognition, and semantic analysis. Taking named entity recognition as an example, a geotechnical engineering dictionary is constructed to identify entities such as stratum names, lithologic characteristics, and tectonic phenomena in the text, converting unstructured text into structured geological descriptive features. The recognition process uses a conditional random field (CRF) model, whose probability formula is:

[0113] Where X is the observation sequence (text), Y is the tag sequence (entity category), y_I is the entity category corresponding to the Ith sequence position of the tag sequence; y_{I-1} is the entity category corresponding to the I-1th sequence position of the tag sequence; f_k is the characteristic function of the kth geological description feature, λ_k is the feature weight of the kth geological description feature, and Z(X) is the normalization factor. By training the CRF model, we can achieve automatic recognition and extraction of geotechnical engineering terminology and descriptions.

[0114] Finally, spatial coordinate association preprocessing is performed on the modified structured survey data and structured geological description features. All data points are associated with a unified spatial coordinate system to establish spatial correlation. Spatial indexing methods (such as R-trees or quadtrees) are used to organize the data to facilitate subsequent spatial queries and analysis. Through spatial association preprocessing, standardized geotechnical engineering data with geographic location information is obtained.

[0115] In a specific embodiment, the process of executing step S2 specifically includes the following steps:

[0116] Construct an ontology model for geotechnical engineering that includes geotechnical classification system, geological structure relationship, and engineering property association;

[0117] Based on the geotechnical engineering ontology model, a knowledge graph is constructed in the geotechnical engineering field, which includes the relationship network between stratum types, lithologic characteristics, physical and mechanical parameters, and engineering properties.

[0118] Perform entity recognition on standardized geotechnical engineering data to extract entities of strata, lithology, parameters and engineering properties;

[0119] Reasoning about the relationships between entities based on spatial location and geological rules, and establishing associations between entities;

[0120] The extracted entities and relationships are mapped to the geotechnical engineering domain knowledge graph to form a project-specific knowledge subgraph that reflects the characteristics of the current project.

[0121] Specifically, we first construct a geotechnical engineering domain ontology model as the framework for the knowledge graph. This domain ontology model consists of three main components: a geotechnical classification system, geological structural relationships, and engineering property associations. The geotechnical classification system follows international and national standards (such as the Unified Soil Classification System (USCS) or the China Geotechnical Classification Standard) and defines the classification hierarchy of soils and rocks. The geological structural relationships describe the genesis, spatial composition, and structural deformation characteristics of strata. The engineering property associations express the intrinsic connection between the physical and mechanical properties of geotechnical materials and engineering behavior.

[0122] Based on the geotechnical engineering domain ontology model, a knowledge graph for geotechnical engineering is constructed. A knowledge graph is a semantic network of entities and relationships. Entities include stratum types (e.g., clay layers, sand layers, bedrock), lithologic characteristics (e.g., particle composition, structural texture), physical and mechanical parameters (e.g., density, internal friction angle, cohesion), and engineering properties (e.g., bearing capacity, compressibility). Relationships include semantic associations such as "belongs to," "is located in," "covers," and "has." This is represented as a geotechnical engineering domain knowledge graph G = (V, E, T), where V is the entity set, E is the relationship set, and T is the triple set (v_i, e_j, v_k), indicating the existence of a relationship e_j between entity v_i and entity v_k.

[0123] Entity recognition was performed on standardized geotechnical engineering data to extract key entities. Named entity recognition technology and professional rules were used to extract stratigraphic entities (e.g., "medium-weathered granite," "saturated silty clay"), lithologic entities (e.g., "medium-grained structure," "layered structure"), parameter entities (e.g., "internal friction angle φ = 28°," "cohesion c = 35 kPa"), and engineering attribute entities (e.g., "bearing capacity characteristic value f_ak = 280 kPa," "compression modulus E_s = 12 MPa"). Entity recognition employed a hybrid approach based on rules and machine learning. The rule-based approach utilized a dictionary of professional terminology and regular expressions, while the machine learning approach employed a BiLSTM-CRF model for sequence labeling.

[0124] Reasoning on the relationships between entities based on spatial location and geological rules establishes associations between entities. Spatial location relationship reasoning is based on the entity's geographic coordinates and depth information to determine the upper and lower cover relationships and lateral change relationships of the strata. Geological rule reasoning is based on geological knowledge, such as sedimentation laws and tectonic evolution laws, to infer the causal relationships and attribute associations between entities. Relationship reasoning can be implemented using a rule-based reasoning engine or a probabilistic graphical model. The formula is:

[0125] where P(x) is the probability of a world state x, where x represents a set of relationship configurations between geological entities, e.g., one stratum is located above another; w_J is the weight of rule J, indicating the importance or credibility of the rule; n_J(x) is the number of times rule J is satisfied under x. For example, if rule J is “sandstone is usually located above shale,” n_J(x) is the number of times this rule is satisfied in x; and Z′ is a normalization constant.

[0126] The extracted entities and relationships are mapped into the geotechnical engineering domain knowledge graph to form a project-specific knowledge subgraph. This mapping process involves entity alignment, which matches the extracted entities with concepts in the knowledge graph, and relationship fusion, which integrates the inferred relationships into the knowledge graph framework. The project-specific knowledge subgraph is a subset of the geotechnical engineering domain knowledge graph, specifically reflecting the geological characteristics and engineering conditions of the current project, providing knowledge support for subsequent analysis.

[0127] In a specific embodiment, the process of executing step S3 specifically includes the following steps:

[0128] Extracting stratum spatial distribution information based on project-specific knowledge subgraph;

[0129] Perform three-dimensional interpolation on the stratum spatial distribution information to generate an initial stratum spatial distribution model;

[0130] Based on geological sedimentary laws and tectonic evolution knowledge, the initial stratigraphic spatial distribution model is constrained and optimized to generate a geological rationality correction model;

[0131] The geological rationality correction model is integrated with the engineering parameter distribution to generate a three-dimensional model of the stratum spatial distribution;

[0132] According to engineering design requirements and the importance of geological features, the three-dimensional model of the spatial distribution of the strata is dynamically sliced ​​to obtain key cross-sections.

[0133] Specifically, we first extract stratum spatial distribution information from the project-specific knowledge subgraph. This information includes data such as drillhole location coordinates, stratum interface depth, stratum type, and spatial extent. This information is stored in entities and relationships within the knowledge subgraph, and relevant data is extracted using a graph query language such as SPARQL or Cypher.

[0134] The extracted stratigraphic spatial distribution information is interpolated in three dimensions to generate an initial stratigraphic spatial distribution model. Three-dimensional interpolation methods include inverse distance weighting (IDW), radial basis functions (RBF), and three-dimensional kriging. Taking three-dimensional kriging as an example, its basic concept is to interpolate and estimate stratigraphic interfaces in three dimensions, determine spatial correlations through variogram analysis, calculate optimal weight coefficients, and achieve three-dimensional interpolation of stratigraphic interfaces to generate an initial stratigraphic spatial distribution model.

[0135] Based on geological sedimentary laws and tectonic evolution knowledge, the initial stratum spatial distribution model is constrained and optimized. Geological sedimentary laws include sequence stratigraphic principles, sedimentary facies models, and sediment distribution patterns; tectonic evolution knowledge includes fault development patterns, fold geometry, and tectonic deformation patterns. This geological knowledge is converted into mathematical constraints, and the initial 3D model is optimized. The objective function of the constrained optimization is expressed as:

[0136] minF(m)=||Gm-d|| 2 +α×||Lm|| 2 Where m is the model parameter vector, G is the forward operator, d is the observed data vector, L is the regularization operator, and α is the regularization parameter. The first term ensures the goodness of fit between the model and the observed data, while the second term introduces geological knowledge constraints. By solving this optimization problem, a correction model that satisfies geological plausibility is generated.

[0137] The geological rationality correction model is integrated with the engineering parameter distribution to generate a three-dimensional model of the stratum spatial distribution. The engineering parameter distribution is derived from standardized geotechnical engineering data, including soil strength parameters, deformation parameters, and hydrological parameters. The integration process uses parameter mapping and attribute interpolation techniques to map the engineering parameter values ​​to corresponding positions in three-dimensional space. It can be expressed as:

[0138] M = {G1, P1}; where M is the fused 3D model, G1 is the geological rationality correction model, and P1 is the engineering parameter distribution. The fused 3D model not only reflects the spatial distribution of the strata but also includes engineering parameter information at each location.

[0139] Dynamic slicing is performed on the 3D model of the stratum spatial distribution based on engineering design requirements and the importance of geological features to obtain key cross-sections. Engineering design requirements include building layout, foundation form, and construction plan; geological feature importance considers areas of intense geological change, areas with unfavorable geological distribution, and key engineering node areas. The dynamic slicing process is expressed as:

[0140] S = |M|_π; where S is the profile, M is the 3D model, and π is the slicing plane, determined by the slicing direction and position. By adjusting slicing parameters, key profiles from multiple perspectives and directions are generated, providing intuitive geological structure visualization for engineering design and report preparation.

[0141] In a specific embodiment, the process of executing step S4 specifically includes the following steps:

[0142] Extract geological anomaly characteristic indicators from the project-specific knowledge subgraph;

[0143] Build a bad geological identification rule base, which includes characteristic patterns of various bad geological phenomena;

[0144] Based on the adverse geological identification rule library, pattern matching of geological anomaly characteristic indicators is performed to identify potential adverse geological phenomena;

[0145] Search the historical exploration case database to extract historical cases with similar geological conditions to the current project and their geological risk records, namely historical case risk records;

[0146] The currently identified potential adverse geological phenomena and historical case risk records are integrated to generate graded and zoned geological risk assessment results.

[0147] Specifically, we first extract geological anomaly characteristic indicators from the project-specific knowledge subgraph. These indicators are quantitative parameters that reflect geological anomalies or potential risks, including abnormal stratum thickness, sudden changes in physical and mechanical parameters, abnormal groundwater levels, and changes in stratum inclination. These indicators are extracted through knowledge graph query and attribute analysis. For example, graph query can be used to identify areas with sudden parameter changes.

[0148] Build a rule library for identifying adverse geological conditions, including characteristic patterns of various adverse geological phenomena. These include karst, weak interlayers, expansive soil, liquefied soil, landfill areas, and landslides. For each adverse geological phenomenon, define its characteristic pattern, including parameter characteristics, spatial characteristics, and evolutionary characteristics. Rules can be expressed using production rules (IF-THEN rules), such as:

[0149] IF(soil type = "sand" AND density < 0.3 AND groundwater level is high) THEN liquefaction risk = "high";

[0150] Based on the adverse geology identification rule base, pattern matching is performed on geological anomaly characteristic indicators to identify potential adverse geological phenomena. Pattern matching utilizes a forward inference engine to match extracted geological anomaly characteristic indicators with the conditions in the rule base. If the conditions are met, the corresponding conclusion is triggered. The inference process is implemented using the Rete algorithm, which constructs a conditional network to efficiently match input facts with rule conditions. Through pattern matching, various types of adverse geological phenomena that may exist at the project site can be identified.

[0151] Search the historical exploration case database to extract historical cases with geological conditions similar to the current project and their geological risk records. Case retrieval is based on similarity calculations, comparing the current project with historical cases in terms of geological conditions, engineering characteristics, and environmental factors. Similarity calculations can be based on text-based distance calculation formulas or vector-level similarity calculations. Select historical cases with high similarity and extract the geological risk information recorded in them as historical case risk records.

[0152] The currently identified potential adverse geological phenomena are integrated with historical case risk records to generate hierarchical and regional geological risk assessment results. This fusion process utilizes evidence theory or fuzzy set theory to comprehensively consider risk information from different sources. Taking the Dempster-Shafer evidence theory as an example, assuming the basic probability distribution functions of the current identification results and historical case records are m_1 and m_2, respectively, the fused basic probability distribution function is m(A) = (m_1 ⊕ m_2)(A), where A is a subset (event) of the risk level set and ⊕ is the Dempster combination operator. Through evidence fusion, more reliable risk assessment results are obtained, which are then presented hierarchically and regionally according to risk level and spatial distribution.

[0153] In a specific embodiment, the process of executing step S5 may specifically include the following steps:

[0154] Extract physical and mechanical parameter data of each layer based on the project-specific knowledge subgraph;

[0155] Determine the key engineering parameter indicator set based on the design requirements of different engineering types;

[0156] Calculate preliminary engineering parameter recommendations based on physical and mechanical parameter data and key engineering parameter indicator sets;

[0157] Based on the geological risk assessment results, the preliminary engineering parameter recommendations are revised to generate risk-corrected engineering parameters;

[0158] Based on geotechnical specifications, the rationality of risk-corrected engineering parameters is verified to generate layered and zoned foundation bearing capacity parameters;

[0159] Based on the layered and zoned foundation bearing capacity parameters and geological risk assessment results, targeted foundation treatment suggestions are put forward to form a complete engineering parameter recommendation result.

[0160] Specifically, we first extract the physical and mechanical parameter data for each stratum based on the project-specific knowledge subgraph. These physical and mechanical parameters include soil's natural density, moisture content, liquid limit, plastic limit, compressibility, cohesion, and internal friction angle; and rock's uniaxial compressive strength, elastic modulus, and Poisson's ratio. These parameters are stored in the entity attributes of the knowledge subgraph and extracted using a graph query language.

[0161] Determine the key engineering parameter indicator set based on the design requirements of different project types. Different types of projects (such as high-rise buildings, bridges, tunnels, and underground projects) focus on different geotechnical parameters. For example, high-rise buildings focus on foundation bearing capacity and settlement characteristics, while tunnel projects focus more on rock mass strength and groundwater conditions. By querying the project type knowledge base and combining the specific project circumstances, determine the key engineering parameter indicator set for the project.

[0162] Preliminary engineering parameter recommendations are calculated based on physical and mechanical parameter data and a set of key engineering parameter indicators. The calculation process uses statistical methods and empirical formulas to convert foundation physical and mechanical parameters into the parameters required for engineering design. For example, the formula for estimating the foundation bearing capacity characteristic value f_ak using the static penetration test value q_c is:

[0163] f_ak = k1 × q_c; k1 is an empirical coefficient that depends on the soil type and condition. Similarly, foundation settlement can be estimated using the compression modulus E_s, while slope stability can be estimated using cohesion c and the internal friction angle φ. For each key engineering parameter, preliminary recommended values ​​are obtained using corresponding calculation models.

[0164] Based on the geological risk assessment results, the preliminary engineering parameter recommendations are revised to generate risk-corrected engineering parameters. Based on the geological risk assessment results, parameter values ​​are adjusted for areas located in risk zones or affected by adverse geology. The adjustment principle is: the higher the risk level, the more conservative the parameter values. The correction formula is expressed as:

[0165] P_adj = P_in × (1-α2 × Risk); where P_adj is the revised parameter value, P_in is the initial recommended value, α2 is the adjustment coefficient, and Risk is the risk level (a value between 0 and 1). Through parameter modification, engineering parameters that take geological risk factors into account are generated.

[0166] Based on geotechnical specifications, the rationality of risk-corrected engineering parameters is verified to generate layered and zoned foundation bearing capacity parameters. This rationality verification includes parameter range checks, parameter relationship checks, and specification compliance checks. Parameter range checks ensure that parameter values ​​are within reasonable physical ranges; parameter relationship checks examine the logical relationships between related parameters; and specification compliance checks ensure that parameter values ​​meet the requirements of relevant engineering specifications. Through rationality verification, foundation bearing capacity parameters that comply with specifications and physical laws are generated and presented in a layered and zoned manner by stratum and region.

[0167] Based on the layered and zoned foundation bearing capacity parameters and geological risk assessment results, targeted foundation treatment recommendations are proposed. Foundation treatment recommendations are based on decision rule reasoning, taking into account multiple factors such as foundation bearing capacity, engineering load, risk level, and economic factors. For example:

[0168] IF (foundation bearing capacity < building load requirement AND stratum = "soft soil" AND thickness > 5m) THEN recommendation = "use pile foundation";

[0169] IF (foundation bearing capacity < building load requirement AND stratum = "sand" AND density < 0.3) THEN recommendation = "use vibrocompaction reinforcement";

[0170] Through rule reasoning, foundation treatment suggestions for each area are generated, which together with the bearing capacity parameters constitute a complete engineering parameter recommendation result.

[0171] In a specific embodiment, the process of executing step S6 may specifically include the following steps:

[0172] Access a library of geotechnical investigation report templates categorized by project type, geological complexity, and report purpose;

[0173] Extract the engineering type, geological complexity and report purpose of the current project based on the project-specific knowledge subgraph, i.e., project characteristics;

[0174] Calculate the similarity of the geotechnical engineering investigation report template library based on the extracted project features and select the most matching basic report template;

[0175] Analyze the key geological features and engineering points in the project-specific knowledge sub-map to determine the key content of the report;

[0176] Optimize chapters and adjust the structure of the basic report template based on the key content of the report to generate a project-specific report framework.

[0177] Specifically, first obtain a library of geotechnical engineering investigation report templates categorized by project type, geological complexity, and report purpose. This template library is a pre-established collection of report templates for different application scenarios. By project type, templates are available for residential buildings, public buildings, industrial buildings, bridges, tunnels, and underground projects. By geological complexity, templates are available for simple, moderately complex, and complex geological conditions. By report purpose, templates are available for the feasibility study, preliminary design, and construction drawing design stages. Each template includes a standard chapter structure, required content elements, and specific formatting requirements.

[0178] Based on the project-specific knowledge subgraph, the project type, geological complexity, and report purpose of the current project are extracted to form project characteristics. The project type is directly extracted from the basic project information; the geological complexity is comprehensively assessed by analyzing factors such as the complexity of strata distribution, the distribution of adverse geological phenomena, and groundwater conditions; and the report purpose is determined based on the client's needs and the project stage. Project characteristics can be represented as a feature vector:

[0179] [Project type, geological complexity, report purpose];

[0180] For example, the feature vector may be ["high-rise residential", "medium complexity", "construction drawing design"], indicating that the current project is a high-rise residential project, the geological conditions are moderately complex, and the report is used in the construction drawing design stage.

[0181] Based on the extracted project features, the similarity calculation is performed on the geotechnical engineering investigation report template library, and the most matching basic report template is selected. The similarity calculation adopts the weighted cosine similarity method, taking into account the importance weight of each feature. The similarity calculation formula is:

[0182] Where F_pr is the project feature vector, F_te is the template feature vector, w_s is the weight of the sth feature, and sim_s is the similarity function for the sth feature. For categorical features (such as project type), similarity can be based on a predefined categorical similarity matrix; for hierarchical features (such as complexity), similarity can be calculated based on hierarchical differences. The similarity between the project and each template is calculated, and the template with the highest similarity is selected as the base report template.

[0183] Analyze the key geological features and engineering key points within the project-specific knowledge subgraph to determine the report's key content. Key geological features include major stratigraphic characteristics, special geological conditions, and potential geological risks; engineering key points include key design parameters, construction difficulties, and special treatment requirements. By querying and analyzing the knowledge subgraph, this key information is extracted and the key areas for emphasis in the report are determined. For example, if karst is present in the project area, the karst distribution characteristics and treatment recommendations will be the focus of the report; if the project is a high-rise building located in a soft soil area, foundation treatment and settlement control will be the focus of the report.

[0184] Based on the report's key content, the basic report template is optimized and restructured to create a project-specific report framework. This optimization includes adjusting the order of sections, expanding or simplifying specific content, and adding specialized analysis sections. For example, for projects with complex geological issues, a dedicated geological analysis section can be added; for projects with unique foundation types, the recommended foundation design parameters section can be enhanced. By restructuring the basic template, a report framework tailored to the specific project is generated, one that not only meets regulatory requirements but also highlights the project's key geological features and engineering highlights.

[0185] In a specific embodiment, the process of executing step S7 may specifically include the following steps:

[0186] Build report content filling rules based on the project-specific report framework;

[0187] Map key cross-sections to the stratigraphic description section of the project-specific reporting framework according to the report content filling rules;

[0188] Map the geological risk assessment results to the engineering geological issues section of the project-specific reporting framework according to the report content filling rules;

[0189] Map the engineering parameter recommendation results to the foundation and basic sections of the project-specific report framework according to the report content filling rules; then obtain the report text;

[0190] Generate supporting diagrams and appendix materials based on project-specific knowledge subgraphs;

[0191] Perform natural language optimization and professional terminology standardization on the generated report text;

[0192] Integrate report text, supporting charts and appendix materials to generate a complete geotechnical engineering investigation report.

[0193] Specifically, report content filling rules are constructed based on the project-specific report framework. Filling rules define how to map various analysis results to specific sections of the report framework, including content mapping relationships, display formats, and layout requirements. Filling rules are expressed in a structured template language, for example:

[0194] {"Chapter":"Stratigraphic Description","Data Source":"Key Profile","Mapping Rules":[{"Data Type":"Stratigraphic Profile","Display Position":"Chapter Body","Display Form":"Graphics + Text Description"},{"Data Type":"Drillhole Histogram","Display Position":"Attached Figures","Display Form":"Normalized Graphs"}],"Text Template":"The strata at the engineering site are from top to bottom: {Stratigraphic Description}. Among them, {Main Strata Name} Distribution {Distribution Characteristics}, Thickness {Thickness Range}"}; These rules guide how to reasonably place different types of content in the corresponding positions of the report.

[0195] Key cross-sections are mapped to the stratigraphic description section of the project-specific reporting framework according to the report content filling rules. Key cross-sections include engineering geology cross-sections, drillhole histograms, and 3D stratigraphic distribution visualizations. These graphical elements are inserted into the stratigraphic description section according to the filling rules, and accompanying textual descriptions are generated. For example, for engineering geology cross-sections, the stratigraphic sequence, thickness variations, and spatial distribution characteristics shown in the cross-section are extracted to generate standardized stratigraphic description text.

[0196] Map the geological risk assessment results to the Engineering Geology Issues section of the project-specific reporting framework according to the report content filling rules. The geological risk assessment results include the identification of adverse geological phenomena, risk level assessment, and spatial distribution information. Based on the filling rules, organize the risk assessment results into the Engineering Geology Issues section, including problem description, impact analysis, and risk level description. For each identified geological issue (such as weak layers, liquefied soils, expansive soils, etc.), generate a professional and standardized description paragraph detailing its distribution range, characteristics, and potential impacts.

[0197] The recommended engineering parameters are mapped to the Foundations and Substructures section of the project-specific report framework according to the report content filling rules, resulting in the report text. The recommended engineering parameters include layered and zoned foundation bearing capacity parameters and foundation treatment recommendations. According to the filling rules, these parameters and recommendations are organized into the Foundations and Substructures section, including parameter tables, zone descriptions, and treatment solution recommendations. For foundation bearing capacity parameters, standardized parameter tables are generated, including design parameter values ​​for different strata and regions. For foundation treatment recommendations, targeted engineering recommendations are generated, clearly stating the treatment method, scope of application, and expected results.

[0198] Supporting charts and appendices are generated based on the project-specific knowledge subgraph. These include site plans, survey point layouts, engineering geological profiles, and thematic analysis diagrams. Appendices include drilling records, test results tables, on-site photographs, and parameter statistical analysis tables. The system automatically generates these charts and appendices based on the data in the knowledge subgraph, ensuring standardized formatting and complete content. Chart generation utilizes a professional mapping engine that supports a variety of geological legends and engineering symbols, ensuring professionalism and readability.

[0199] The generated report text is optimized for natural language and standardized for professional terminology. Natural language optimization includes grammar checking, sentence structure optimization, and coherence improvement; professional terminology standardization ensures that the terminology used complies with industry standards and regulatory requirements. The system uses rule-based and statistical natural language processing technology to check for grammatical errors, redundant expressions, and logical problems in the text and optimizes them. At the same time, it utilizes a geotechnical engineering terminology library to ensure the consistency and standardization of terminology. For example, the non-standard expression "soft soil layer" is standardized as "soft plastic silty clay layer" to improve the professionalism of the report.

[0200] Integrate the report text, supporting charts, and appendix materials to generate a complete geotechnical engineering investigation report. The integration process includes document structure organization, page layout design, and format unification. The system organically combines the text content, charts, and appendices according to the report framework to generate a complete report document with a unified format and clear structure. Document generation supports multiple output formats (such as PDF, Word, etc.) and automatically processes page numbers, tables of contents, chapter references, etc. to ensure the professionalism and readability of the report. The generated report is not only complete in content and well-structured, but also meets industry standards and specifications, and can be directly used for engineering design reference and decision support.

[0201] In a specific embodiment, the process of analyzing key geological features and engineering key points in the project-specific knowledge subgraph to determine the key contents of the report may specifically include the following steps:

[0202] Identify key geological nodes and relationship strengths through topological analysis of project-specific knowledge subgraphs;

[0203] Calculate geological feature importance scores based on key geological nodes and relationship strengths;

[0204] Identify key geological features of the project based on geological feature importance scores;

[0205] Analyze the impact of key geological features on engineering construction in combination with the engineering type knowledge base;

[0206] Comprehensively analyze the impact on engineering construction and the importance score of geological features to determine the key contents of the report.

[0207] Specifically, we first identify key geological nodes and relationship strengths by performing topological analysis on the project-specific knowledge subgraph. Topological analysis is a graph structure analysis method used to identify important nodes and key relationships in the knowledge graph. The system uses centrality measurement algorithms, including degree centrality, betweenness centrality, and eigenvector centrality, to evaluate the importance of each node in the knowledge subgraph. Degree centrality calculates the number of connections of a node, and betweenness centrality calculates the "bridge" role of a node in the network. Through these centrality indicators, the system identifies key geological nodes (such as special strata, abnormal structures) and important relationships (such as stratum interface relationships, structural control relationships) in the knowledge subgraph.

[0208] The importance score of geological features is calculated based on key geological nodes and relationship strength. The importance score is a comprehensive indicator that takes into account the centrality index of the node, the strength of the relationship, and geological expertise. The calculation formula of the importance score is:

[0209] I(f) = w_1 × NC(f) + w_2 × RS(f) + w_3 × GK(f); where I(f) is the importance score of geological feature f, NC(f) is the node centrality index (normalized), RS(f) is the relationship strength index, GK(f) is the importance index based on geological knowledge, and w_1, w_2, and w_3 are weight coefficients. The relationship strength index RS(f) of a feature is based on the weights of the edges associated with the feature:

[0210] Where E_f is the set of edges related to geological feature f, weight(e) is the weight of edge e, and |E_f| is the size of the edge set. Through this calculation method, an importance score is assigned to each geological feature in the knowledge subgraph.

[0211] Identify key geological features of a project based on their importance scores. The system ranks all geological features by importance score and selects those with scores exceeding a threshold as key geological features. Key geological features typically include major stratigraphic types, special geological structures (such as faults and folds), unusual geological bodies (such as karst and paleochannels), and special soil layers (such as weak layers and expansive soils). These key features will be highlighted and analyzed in the report.

[0212] The impact of key geological features on project construction is analyzed in conjunction with the project type knowledge base. The project type knowledge base contains a sensitivity matrix of different project types to geological conditions, indicating the impact of various geological features on different projects. The impact assessment uses a matrix mapping method:

[0213] Imp(f,p) = Sen[f_type][p_type] × I(f); where Imp(f,p) is the impact of geological feature f on project p, Sen is the sensitivity matrix (a two-dimensional matrix used to quantify the impact of different geological features on various projects), f_type is the feature type, p_type is the project type, and I(f) is the importance score of geological feature f. This method assesses the specific impact of each key geological feature on the current project.

[0214] A comprehensive analysis of the impact on project construction and the importance of geological features determines the report's key content. Key geological features are ranked by their impact on the project, with those with the highest impact selected as the report's key content. The report's key content is represented as a priority queue, where each element contains a geological feature and its impact on the project. The queue is sorted in descending order of impact, ensuring that the most important content is prioritized. These key content will guide adjustments and optimizations to the report framework, ensuring that the report highlights the geological factors that have the greatest impact on the project, providing critical support for project design and decision-making.

[0215] It should be noted that the formulas of the present invention are all dimensionless and numerically calculated. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters and thresholds in the formulas are set by technicians in this field according to actual conditions.

[0216] The present invention constructs a knowledge graph in the field of geotechnical engineering, organically combines professional knowledge with project data, forms a project-specific knowledge subgraph, and provides a knowledge basis for subsequent intelligent analysis; based on the project-specific knowledge subgraph, multi-dimensional stratigraphic analysis and three-dimensional model construction are carried out to make the description of geological conditions more accurate and intuitive; through geological condition reasoning and risk assessment, the system identifies potential geological risks and provides a basis for risk prevention and control for engineering design; intelligent recommendation of engineering parameters combines geological conditions and risk factors to make parameter recommendations more reasonable and reliable; the adaptive generation of project-specific report frameworks and intelligent assembly of content ensure the professionalism and pertinence of reports. Compared with traditional geotechnical engineering survey report preparation methods, the present invention has the following advantages: knowledge-driven intelligent analysis improves the accuracy and professionalism of geological condition interpretation; automated report generation greatly improves work efficiency and reduces the time and workload of manual writing; standardized processing procedures ensure the consistency and reliability of report quality; personalized report customization makes the report more in line with project characteristics and needs.

[0217] The above describes the method for automatically generating a geotechnical engineering investigation report based on a knowledge graph in the embodiment of the present application. The following describes the system for automatically generating a geotechnical engineering investigation report based on a knowledge graph in the embodiment of the present application. Figure 3 In the embodiment of the present application, an embodiment of the automatic generation system of geotechnical engineering investigation report based on knowledge graph includes:

[0218] The data acquisition and evaluation module is used to obtain multi-source geotechnical engineering survey data sets; perform data cleaning and quality assessment on the multi-source geotechnical engineering survey data sets, and perform spatial correlation preprocessing to obtain standardized geotechnical engineering data;

[0219] Graph construction module, used to construct geotechnical engineering knowledge graph; extract entities and identify relationships from standardized geotechnical engineering data, and map them to the geotechnical engineering knowledge graph to form project-specific knowledge subgraphs;

[0220] The comprehensive segmentation module performs multi-dimensional stratigraphic analysis based on project-specific knowledge subgraphs to generate a three-dimensional stratigraphic spatial distribution model; it also dynamically slices the three-dimensional stratigraphic spatial distribution model to obtain key cross-sections;

[0221] The risk assessment module obtains historical exploration cases, performs geological condition reasoning based on project-specific knowledge subgraphs and historical exploration cases, identifies abnormal geological bodies and adverse geological phenomena, and generates geological risk assessment results;

[0222] The parameter recommendation module intelligently recommends engineering parameters based on geological risk assessment results and project-specific knowledge subgraphs, generating layered and partitioned foundation bearing capacity parameters and foundation treatment suggestions, i.e., engineering parameter recommendation results.

[0223] A report framework generation module is used to obtain a geotechnical engineering investigation report template library; based on the project-specific knowledge subgraph, the geotechnical engineering investigation report template library is adaptively selected and personalized to generate a project-specific report framework;

[0224] The survey report generation module, based on a project-specific report framework, intelligently assembles key profiles, geological risk assessment results, and engineering parameter recommendations to generate a complete geotechnical engineering survey report.

[0225] The present application also provides an automatic generation device for geotechnical engineering survey reports based on a knowledge graph. The automatic generation device for geotechnical engineering survey reports based on a knowledge graph includes a memory and a processor. The memory stores computer-readable instructions. When the computer-readable instructions are executed by the processor, the processor executes the steps of the automatic generation method for geotechnical engineering survey reports based on a knowledge graph in the above-mentioned embodiments.

[0226] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0227] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiment. All technical solutions based on the principles of the present invention are within the scope of protection of the present invention. It should be noted that for users of ordinary skill in the art, various improvements and modifications that do not depart from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. The automatic generation method of geotechnical engineering investigation report based on knowledge graph is characterized by: include: Step S1: Acquire multi-source geotechnical engineering investigation datasets; Perform data cleaning and quality assessment on multi-source geotechnical engineering survey datasets, and perform spatial correlation preprocessing to obtain standardized geotechnical engineering data; Step S2: Construct a geotechnical engineering domain knowledge graph; perform entity extraction and relationship recognition on the standardized geotechnical engineering data, and map it to the geotechnical engineering domain knowledge graph to form a project-specific knowledge subgraph; Step S3: Perform multi-dimensional stratigraphic analysis based on the project-specific knowledge subgraph to generate a three-dimensional stratigraphic spatial distribution model; dynamically slice the three-dimensional stratigraphic spatial distribution model to obtain key cross-sections; Step S4: Obtain historical survey cases, perform geological condition reasoning based on the project-specific knowledge subgraph and historical survey cases, identify abnormal geological bodies and adverse geological phenomena, and generate geological risk assessment results; Step S5: Based on the geological risk assessment results and the project-specific knowledge subgraph, intelligent engineering parameter recommendations are made to generate hierarchical and zoned foundation bearing capacity parameters and foundation treatment suggestions, i.e., engineering parameter recommendation results; Step S6: Obtaining a geotechnical engineering investigation report template library; Adaptively select and personalize the geotechnical engineering investigation report template library based on the project-specific knowledge subgraph to generate a project-specific report framework; Step S7: Based on the project-specific report framework, key profiles, geological risk assessment results, and engineering parameter recommendation results are intelligently assembled to generate a complete geotechnical engineering investigation report.

2. The method for automatically generating geotechnical engineering investigation reports based on knowledge graph according to claim 1 is characterized in that: The specific steps of step S1 are: Acquire a multi-source geotechnical engineering investigation dataset consisting of drilling records, in-situ tests, laboratory tests, and historical investigation data; Identify and classify data types of multi-source geotechnical engineering survey data sets to obtain structured survey data and unstructured description data; Perform outlier detection and missing value identification on structured survey data, marking low-quality data points; Perform interpolation repair or reliability assessment on low-quality data points based on geostatistical models to obtain corrected structured survey data; Perform natural language processing on unstructured description data to extract key geological description features and obtain structured geological description features; The modified structured survey data and structured geological description features are preprocessed by spatial coordinate association to obtain standardized geotechnical engineering data.

3. The method for automatically generating geotechnical engineering investigation reports based on knowledge graph according to claim 2 is characterized in that: The specific steps of step S2 are: Construct a geotechnical engineering ontology model that includes geotechnical classification system, geological structure relationship, and engineering property association; Construct a geotechnical engineering knowledge graph based on the geotechnical engineering ontology model, which includes the relationship network between stratum types, lithologic characteristics, physical and mechanical parameters, and engineering properties; Perform entity recognition on standardized geotechnical engineering data to extract entities of strata, lithology, parameters and engineering properties; Reasoning about the relationships between entities based on spatial locations and geological rules, and establishing associations between entities; The extracted entities and relationships are mapped to the geotechnical engineering domain knowledge graph to form a project-specific knowledge subgraph that reflects the characteristics of the current project.

4. The method for automatically generating geotechnical engineering investigation reports based on knowledge graph according to claim 3 is characterized in that: The specific steps of step S3 are: Extracting stratum spatial distribution information based on project-specific knowledge subgraph; Perform three-dimensional interpolation on the stratum spatial distribution information to generate an initial stratum spatial distribution model; Based on geological sedimentary laws and tectonic evolution knowledge, the initial stratigraphic spatial distribution model is constrained and optimized to generate a geological rationality correction model; The geological rationality correction model is integrated with the engineering parameter distribution to generate a three-dimensional model of the stratum spatial distribution; According to engineering design requirements and the importance of geological features, the three-dimensional model of the spatial distribution of the strata is dynamically sliced ​​to obtain key cross-sections.

5. The method for automatically generating geotechnical engineering investigation reports based on knowledge graph according to claim 4 is characterized in that: The specific steps of step S4 are: Extract geological anomaly characteristic indicators from the project-specific knowledge subgraph; Build a bad geological identification rule base, which includes characteristic patterns of various bad geological phenomena; Based on the adverse geological identification rule library, pattern matching of geological anomaly characteristic indicators is performed to identify potential adverse geological phenomena; Search the historical exploration case database to extract historical cases with similar geological conditions to the current project and their geological risk records, namely historical case risk records; The currently identified potential adverse geological phenomena and historical case risk records are integrated to generate graded and zoned geological risk assessment results.

6. The method for automatically generating geotechnical engineering investigation reports based on knowledge graph according to claim 5 is characterized in that: The specific steps of step S5 are: Extract physical and mechanical parameter data of each layer based on the project-specific knowledge subgraph; Determine the key engineering parameter indicator set based on the design requirements of different engineering types; Calculate preliminary engineering parameter recommendations based on physical and mechanical parameter data and key engineering parameter indicator sets; Based on the geological risk assessment results, the preliminary engineering parameter recommendations are revised to generate risk-corrected engineering parameters; Based on geotechnical specifications, the rationality of risk-corrected engineering parameters is verified to generate layered and zoned foundation bearing capacity parameters; Generate targeted foundation treatment recommendations based on the layered and zoned foundation bearing capacity parameters and geological risk assessment results.

7. The method for automatically generating geotechnical engineering investigation reports based on knowledge graph according to claim 6 is characterized in that: The specific steps of step S6 are: Access a library of geotechnical investigation report templates categorized by project type, geological complexity, and report purpose; Extract the engineering type, geological complexity and report purpose of the current project based on the project-specific knowledge subgraph, i.e., project characteristics; Calculate the similarity of the geotechnical engineering investigation report template library based on the extracted project features and select the most matching basic report template; Analyze the key geological features and engineering points in the project-specific knowledge sub-map to determine the key content of the report; Optimize chapters and adjust the structure of the basic report template based on the key content of the report to generate a project-specific report framework.

8. The method for automatically generating geotechnical engineering investigation reports based on knowledge graph according to claim 7 is characterized in that: The specific steps of step S7 are: Build report content filling rules based on the project-specific report framework; Map key cross-sections to the stratigraphic description section of the project-specific reporting framework according to the report content filling rules; Map the geological risk assessment results to the engineering geological issues section of the project-specific reporting framework according to the report content filling rules; Map the engineering parameter recommendation results to the foundation and basic sections of the project-specific report framework according to the report content filling rules; then obtain the report text; Generate supporting diagrams and appendix materials based on project-specific knowledge subgraphs; Perform natural language optimization and professional terminology standardization on the generated report text; Integrate report text, supporting charts and appendix materials to generate a complete geotechnical engineering investigation report.

9. The method for automatically generating geotechnical engineering investigation reports based on knowledge graph according to claim 7 is characterized in that: The steps for analyzing the key geological features and engineering points in the project-specific knowledge sub-map and determining the key contents of the report are as follows: Identify key geological nodes and relationship strengths through topological analysis of project-specific knowledge subgraphs; Calculate geological feature importance scores based on key geological nodes and relationship strengths; Identify key geological features of the project based on geological feature importance scores; Analyze the impact of key geological features on engineering construction in combination with the engineering type knowledge base; Comprehensively analyze the impact on engineering construction and the importance score of geological features to determine the key contents of the report.

10. Automatic generation equipment for geotechnical engineering investigation reports based on knowledge graph, characterized by: The device for automatically generating geotechnical engineering investigation reports based on knowledge graphs includes: a memory and at least one processor, wherein the memory stores instructions; The at least one processor calls the instructions in the memory to enable the device for automatically generating geotechnical engineering investigation reports based on knowledge graphs to execute the method for automatically generating geotechnical engineering investigation reports based on knowledge graphs as described in any one of claims 1 to 9.

Images