A large model-based urban underground space geological suitability quantitative evaluation method

Abstract

This invention discloses a quantitative evaluation method for the geological suitability of urban underground space based on a large model, belonging to the field of intelligent evaluation technology for geological information. The method includes: performing prior calculations on geological observation data based on voxel-level spatial weighted layers and suitability calculation templates, and assimilating and updating the data with newly added geological observation data to generate voxel-level posterior suitability volumes and posterior uncertainty volumes; superimposing the engineering constraint set with the voxel-level posterior suitability volumes and posterior uncertainty volumes, and calculating the geological suitability score according to the measurement rules of the geological suitability rule set; performing voxel confidence-weighted verification on the geological suitability score, re-evaluating areas that fail the verification through micro-surveys and backflow, and generating a geological suitability evaluation report. This invention achieves linked optimization of geological suitability scores and uncertainties, ultimately achieving stable, interpretable, and iterative results.

Description

Technical Field

[0001] This invention relates to the field of geological information intelligent evaluation technology, and in particular to a quantitative evaluation method for the geological suitability of urban underground space based on a large model. Background Technology

[0002] Geological suitability assessment of urban underground space is widely used in decision-making for rail transit, integrated utility tunnels, and underground engineering. The conventional approach is based on borehole and in-situ / indoor tests, integrating geophysical exploration, remote sensing, and monitoring data to construct an indicator system. This system is then used in conjunction with hierarchical analysis, entropy weighting, or weighted overlay on a GIS / 3D model to generate graded results and recommendations. At the same time, thresholds and criteria are set with reference to regulatory provisions and typical engineering experience. The industry is evolving towards 3D gridding, data fusion, and traceable management.

[0003] In complex urban environments, traditional processes often rely on batch updates to absorb time-varying observations, making it difficult to form a closed loop for quantifying and assimilating uncertainties that are time-series and source-specific. Meanwhile, constraints such as engineering red lines, neighbor disturbance control, and construction organization are often applied retroactively, and the coupling of scoring and constraints and the confidence verification mechanism are not uniform enough, affecting cross-project comparability and dynamic verification. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a quantitative evaluation method for the geological suitability of urban underground space based on a large model, which solves the problems of difficulty in forming an uncertainty assimilation loop in time-varying observations and the lack of consistency between engineering constraints and scoring coupling and confidence verification.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides a method for quantitative evaluation of the geological suitability of urban underground space based on a large model, which includes constructing a three-dimensional voxel grid and aggregating multi-source data and standardized case corpus, and extracting underground space suitability thresholds and measurement rules through Transformer to generate a traceable geological suitability rule set;

[0008] The geological suitability rule set is mapped to a three-dimensional voxel grid and boundary constraints are applied. Combined with geological observation data, a voxel-level spatial weighted layer is generated and solidified as a suitability calculation template.

[0009] Based on voxel-level spatial weighted layers and suitability calculation templates, prior calculations are performed on geological observation data, and assimilation and updates are performed with newly added geological observation data to generate voxel-level posterior suitability volumes and posterior uncertainty volumes.

[0010] The engineering constraint set is superimposed with the voxel-level posterior suitability volume and posterior uncertainty volume, and the geological suitability score is calculated according to the measurement rules of the geological suitability rule set.

[0011] The geological suitability score is verified by voxel confidence weighting. Areas that fail the verification are re-evaluated by micro-exploration and backflow, and a geological suitability evaluation report is generated.

[0012] As a preferred embodiment of the quantitative evaluation method for the geological suitability of urban underground space based on a large model as described in this invention, the construction of a three-dimensional voxel grid refers to setting the voxel resolution and layer thickness, as well as the starting point and grid range, according to the boundary and unified coordinates of the urban underground space, generating a regular three-dimensional grid by equal XYZ interval subdivision, and assigning a unique index to each voxel.

[0013] As a preferred embodiment of the quantitative evaluation method for the geological suitability of urban underground space based on a large model as described in this invention, the multi-source data includes borehole columnar sections, geophysical profiles, remote sensing and InSAR deformation time series, groundwater level and settlement monitoring time series, pipeline and underground structure vectors, and the IFC model.

[0014] The standardized case literature includes compilation standards, guidelines, local standards, historical engineering reports, and case texts.

[0015] As a preferred embodiment of the quantitative evaluation method for the geological suitability of urban underground space based on a large model as described in this invention, the multi-source data and standardized case corpus are subjected to standardization processing and cleaning to remove duplicates.

[0016] Input the standardized case corpus into Transformer to perform term alignment and clause extraction, and output the underground space suitability threshold, measurement rules and source records;

[0017] By utilizing the scale of three-dimensional voxel grids and the data caliber of multi-source data, the consistency of underground space suitability thresholds and measurement rules is checked, and a traceable set of geological suitability rules is generated after format solidification and number verification.

[0018] As a preferred embodiment of the large-model-based quantitative evaluation method for the geological suitability of urban underground space described in this invention, the steps of mapping the geological suitability rule set to a three-dimensional voxel mesh and applying boundary constraints are as follows:

[0019] By combining GIS spatial overlay and 3D rasterization, the geological suitability rule set is mapped to a 3D voxel grid according to the unique voxel index and a list of applicable rule voxels is generated.

[0020] Based on geological boundary data, a boundary constraint mask layer is generated using a spatial rasterization method, and the list of applicable regular voxels is masked and filtered to generate a list of applicable regular voxels after boundary constraints.

[0021] As a preferred embodiment of the urban underground space geological suitability quantitative evaluation method based on a large model as described in this invention, the steps for generating a voxel-level spatial weighted layer and solidifying it as a suitability calculation template are as follows:

[0022] Based on geological observation data, voxel observation feature sets are extracted, and combined with a constrained list of applicable rule voxels, an entropy weighting method is used to generate a voxel-level spatial weighted layer.

[0023] The measurement caliber and adjudication order of the voxel-level spatial weighted layer and the geological suitability rule set are solidified into a suitability calculation template.

[0024] As a preferred embodiment of the large-model-based quantitative evaluation method for the geological suitability of urban underground space described in this invention, the steps of performing prior calculations on geological observation data and assimilating and updating it with newly added geological observation data are as follows.

[0025] Based on voxel-level spatial weighting layers and suitability calculation templates, geological observation data are aligned according to voxel unique indexes and timestamps, and aggregated to generate voxel observation sequences as prior input packages.

[0026] The prior input package is subjected to threshold determination, weight aggregation and level mapping according to the suitability calculation template, and uncertainty assessment is performed using the variance propagation method to generate voxel-level prior suitability volume and prior uncertainty volume.

[0027] As a preferred embodiment of the large-model-based quantitative evaluation method for the geological suitability of urban underground space described in this invention, the steps for generating voxel-level posterior suitability volumes and posterior uncertainty volumes are as follows:

[0028] The newly added geological observation data are aligned with the voxel unique index and timestamp, and combined with the voxel-level prior suitability volume and prior uncertainty volume to generate the assimilation input set;

[0029] The assimilation of the input set is performed using the Bayesian update method, and outlier removal and adaptive weight allocation are carried out to generate voxel-level posterior suitability and posterior uncertainty volumes.

[0030] As a preferred embodiment of the large-model-based quantitative evaluation method for urban underground space geological suitability described in this invention, the steps for calculating the geological suitability score according to the measurement rules of the geological suitability rule set are as follows:

[0031] Acquire engineering constraint data and resolve conflicts through standardization and voxelization encoding to generate an engineering constraint set;

[0032] The set of engineering constraints is superimposed with the voxel-level posterior suitability volume and posterior uncertainty volume to output the constrained posterior suitability volume and posterior uncertainty volume.

[0033] Based on the constrained posterior suitability body and posterior uncertainty body, the geological suitability score is calculated according to the measurement rules of the geological suitability rule set.

[0034] As a preferred embodiment of the large-model-based quantitative evaluation method for the geological suitability of urban underground space described in this invention, the steps of performing voxel confidence-weighted verification on the geological suitability score, conducting micro-surveys and re-evaluation on areas that fail the verification, and generating a geological suitability evaluation report are as follows.

[0035] Based on voxel-level posterior uncertainty and geological suitability scores, the inverse variance weighted normalization method is used for weighted fusion, and data coverage is used as the availability factor to calculate voxel confidence.

[0036] If the voxel confidence level is not lower than the confidence threshold and the geological suitability score is not lower than the underground space suitability threshold at the same time, the verification is judged to be unsuccessful, and a mask of unsuccessful regions is formed by connecting domain aggregation.

[0037] Using the mask of areas that failed the verification as the scope, micro-surveys were carried out, and the newly added engineering geological observation data were aligned and recalculated to obtain the updated voxel-level posterior suitability volume and posterior uncertainty volume, as well as the optimized geological suitability score, and a geological suitability evaluation report was generated.

[0038] The beneficial effects of this invention are as follows: by mapping the geological suitability rule set to a three-dimensional voxel grid and applying boundary constraints, a voxel-level spatial weighted layer is generated and solidified as a suitability calculation template, thereby achieving spatial adaptive fusion of rules and observations; by performing voxel confidence weighted verification on the geological suitability score and performing micro-re-exploration and re-evaluation on areas that fail the verification, the geological suitability score and uncertainty are linked for optimization, ultimately achieving the effect of stable, interpretable and iterative results. Attached Figure Description

[0039] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1This is a flowchart of a method for quantitatively evaluating the geological suitability of urban underground space based on a large model.

[0041] Figure 2 A flowchart for generating a voxel-level spatial weighted layer and solidifying it as a suitability calculation template.

[0042] Figure 3 Flowchart for generating voxel-level posterior suitability and posterior uncertainty volumes. Detailed Implementation

[0043] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0044] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0045] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0046] Reference Figures 1-3 This is one embodiment of the present invention, which provides a method for quantitative evaluation of the geological suitability of urban underground space based on a large model, including the following steps:

[0047] S1. Construct a three-dimensional voxel mesh and collect multi-source data and standardized case corpus. Extract underground space suitability thresholds and measurement rules through Transformer to generate a traceable set of geological suitability rules.

[0048] Based on the urban underground space boundary and unified coordinates, the voxel resolution, layer thickness, starting point and grid range are set, and a regular three-dimensional grid is generated by equal XYZ interval subdivision and a unique index is assigned to each voxel to form a three-dimensional voxel grid.

[0049] Furthermore, based on the urban underground space boundary and unified coordinates, the voxel resolution and layer thickness, the starting point and grid range are first determined, and the number of elements in the X, Y, and Z directions are calculated to establish the XYZ coordinate axes. Then, a regular 3D grid is generated by equidistant XYZ subdivision. The regular 3D grid is clipped using a spatial mask generated by the urban underground space boundary to retain voxels within the urban underground space boundary. For each voxel, the coordinates of the voxel center point and the coordinates of the voxel's outer boundary are calculated and assigned a unique index using both row and column layer encoding and globally unique encoding. At the same time, the starting point, grid range, voxel resolution, and layer thickness are recorded to ensure that the coordinates are consistent with the elevation datum. After clipping and numbering, the voxel list and spatial index relationship are output to form a 3D voxel grid.

[0050] Multi-source data include borehole columnar sections, geophysical profiles (GPR / resistivity / shallow seismic activity), remote sensing and InSAR deformation time series, groundwater level and settlement monitoring time series, pipeline and underground structure vectors, and IFC models.

[0051] The standardized case literature includes compilation standards, guidelines, local standards, historical engineering reports, and case texts;

[0052] It should be noted that multi-source data is obtained through interface downloads or batch imports from survey unit databases, geophysical exploration and monitoring platforms, remote sensing and InSAR services, urban pipeline archives, and IFC model databases; standard case data is obtained through online downloads or scanning and identification from national and local standards platforms, industry association and construction unit archives. Standard case data is obtained by searching standard release platforms, industry association websites, and construction unit and survey and design unit archives.

[0053] Standardize and clean the multi-source data and standardized case corpus to remove duplicates;

[0054] Furthermore, the data formats of the collected borehole columnar sections, geophysical profiles, remote sensing and InSAR deformation time series, groundwater level and settlement monitoring time series, pipeline and underground structure vectors, and IFC models were standardized. By removing duplicate records and irrelevant information, the dataset was ensured to be clean and concise. For the standardized case corpus, including compilation specifications, guidelines, local standards, historical engineering reports, and case texts, similar cleanup work was performed to eliminate redundant and unnecessary parts, and their content was adjusted to a consistent structured format.

[0055] Input the standardized case corpus into Transformer to perform term alignment and clause extraction, and output the underground space suitability threshold, measurement rules and source records;

[0056] Furthermore, the specific process of inputting the standardized case corpus into Transformer for terminology alignment and clause extraction includes: segmenting and tokenizing the compiled specifications, guidelines, local standards, historical engineering reports, and case texts to adapt each text unit to Transformer's input format; using Transformer's self-attention mechanism to identify professional terms with different expressions but consistent semantics in different standardized texts, completing terminology alignment; based on this, through sequence labeling and structural parsing, accurately locating and extracting the judgmental statements related to underground space development from the aligned text, thereby generating underground space suitability thresholds, measurement rules, and traceability records. Underground space suitability thresholds refer to the quantitative limits or limiting values ​​extracted from the standardized case corpus used to determine the feasibility of urban underground space development; measurement rules refer to the evaluation logic, combination methods, weight allocation principles, and calculation processes extracted from the standardized case corpus for integrating multi-source data for suitability scoring; traceability records refer to the traceable information including the source text location, document name, clause number, and publication time, generated simultaneously during the extraction of underground space suitability thresholds and measurement rules.

[0057] By utilizing the scale of three-dimensional voxel grids and the data caliber of multi-source data, the consistency of underground space suitability thresholds and measurement rules is checked, and a traceable set of geological suitability rules is generated after format solidification and number verification.

[0058] Furthermore, based on the voxel resolution and layer thickness of the three-dimensional voxel grid, it is determined whether the spatial granularity of the underground space suitability threshold matches the spatial expression precision of the multi-source data. For example, if the underground space suitability threshold comes from a standard provision applicable to the evaluation of areas at the hundred-meter level, and the three-dimensional voxel grid resolution is 10 meters and the multi-source data such as borehole columnar sections and InSAR deformation time series have a spatial expression precision of 10 meters, then it is determined to be mismatched; if the threshold itself is based on the engineering scale and is consistent with the voxel scale and the spatial granularity of the multi-source data, then it is determined to be a match. By comparing the data calibers of borehole columnar sections, geophysical profiles, remote sensing and InSAR deformation time series, groundwater level and settlement monitoring time series, pipeline and underground structure vectors, and IFC models, the compatibility of parameter types, units of measurement, and observation frequencies involved in the measurement rules is verified. For example, if the measurement rules require the use of "annual average settlement rate (unit: mm / year)" as an input parameter, while the settlement monitoring time series in the multi-source data provides "monthly cumulative settlement (unit: mm)," it is necessary to determine whether consistent parameters can be obtained through time aggregation and unit conversion. If the parameter types, units of measurement, or observation frequencies cannot be aligned, they are considered incompatible. Underground space suitability thresholds and measurement rules with scale or caliber conflicts are adjusted or labeled. The verified underground space suitability thresholds and measurement rules are formatted according to a unified structure and assigned a unique number. The continuity, uniqueness, and standardization of the numbering are then reviewed to finally generate a traceable set of geological suitability rules.

[0059] It should be noted that the traceable set of geological suitability rules provides a normative basis for the assessment results of underground space suitability, facilitating review and verification, accountability, and dynamic updates.

[0060] S2. Map the geological suitability rule set to a three-dimensional voxel grid and apply boundary constraints. Combine this with geological observation data to generate a voxel-level spatial weighted layer and solidify it as a suitability calculation template.

[0061] By combining GIS spatial overlay and 3D rasterization, the geological suitability rule set is mapped to a 3D voxel grid according to the unique voxel index and a list of applicable rule voxels is generated.

[0062] Furthermore, each rule in the geological suitability rule set is converted into a two-dimensional or three-dimensional geographic element according to its spatial applicability range, and then subjected to GIS spatial overlay analysis with a three-dimensional voxel grid under a unified coordinate system. Through three-dimensional rasterization, the overlay results are aligned to the three-dimensional voxel grid according to the XYZ equidistant subdivision structure, so that the spatial scope of each rule accurately corresponds to a voxel unit with a unique index. All voxel unique indexes covered by the rules and their corresponding rule contents are summarized to form a rule voxel applicability list.

[0063] The list of applicable voxels enables precise binding between geological rules and spatial voxels, providing a rule-based basis for subsequent voxel-level suitability calculations.

[0064] Based on geological boundary data, a boundary constraint mask layer is generated using a spatial rasterization method, and the list of applicable regular voxels is masked and filtered to generate a list of applicable regular voxels after boundary constraints.

[0065] Furthermore, the geological boundary data (such as bedrock surfaces, fault zones, aquifer boundaries, or planned control lines) are rasterized in three-dimensional space under a unified coordinate system and converted into a binary mask layer aligned with the three-dimensional voxel grid. In this layer, the effective area voxels are assigned a value of 1, and the invalid area voxels are assigned a value of 0. Then, the mask layer and the list of applicable rule voxels are logically ANDed with each voxel according to the unique voxel index. Voxel rule items located in invalid areas are removed, and rule voxel records that are only within the effective geological boundary range are retained. This generates a list of applicable rule voxels with boundary constraints, ensuring that the geological suitability rules are only effective within the legal or feasible geological space and avoiding distortion of evaluation results due to exceeding the boundaries.

[0066] Based on geological observation data, voxel observation feature sets are extracted, and combined with a constrained list of applicable rule voxels, an entropy weighting method is used to generate a voxel-level spatial weighted layer.

[0067] Furthermore, geological parameters, physical properties, and temporal variation characteristics associated with the unique index of each voxel are extracted from geological observation data to form a voxel observation feature set. The voxel observation feature set is matched with the rule voxel applicable list after boundary constraints to select the observation feature subset applicable to each voxel. On this basis, the information entropy of each applicable feature is calculated, and the weight of each feature at the voxel is determined according to the inverse ratio of the entropy value. The entropy weight method is used to generate a voxel-level spatial variable weight layer that reflects the dynamic importance of different geological factors in space. This enables the weight to be adaptively adjusted according to geological conditions and rule constraints, thereby improving the spatial refinement level of suitability evaluation.

[0068] It should be noted that the geological observation data includes borehole stratification parameters and test indices that can be aligned by voxels, geophysical inversion attributes, InSAR deformation time series, groundwater level time series, and monitoring time series of surface subsidence and structural deformation.

[0069] The measurement caliber and adjudication order of the voxel-level spatial weighted layer and the geological suitability rule set are solidified into a suitability calculation template.

[0070] Furthermore, using GIS attribute join with voxel unique indexes as keys, a mapping list of weight fields and voxel unique indexes is established for voxel-level spatial weighted layers. The measurement scope (value range, statistical window, normalization method) and adjudication order (triggering conditions, priority, missing replacement strategy) of each rule in the geological suitability rule set are extracted. Subsequently, based on a unified parameter table, the weight fields of the voxel-level spatial weighted layers, the threshold table of the geological suitability rule set, and the adjudication order are written into the input field specifications, calculation flow table, and output field specifications of the suitability calculation template. This clarifies the execution order and stopping conditions of threshold determination, weight aggregation, and level mapping, and completes the one-to-one binding of the weight calling method of the voxel-level spatial weighted layers with the threshold table and adjudication order of the geological suitability rule set. This solidifies the measurement scope and adjudication order of the voxel-level spatial weighted layers and the geological suitability rule set into a suitability calculation template.

[0071] It should be noted that the measurement caliber is a unified definition of the calculation and value specification of the geological suitability rule set, including the value range, statistical or sampling window, normalization and directional caliber, summary and grade mapping caliber; the adjudication order is the order of execution and priority arrangement when multiple rules or thresholds are applicable at the same time, including triggering conditions, conflict handling strategies, missing substitution strategies and stopping conditions.

[0072] S3. Based on the voxel-level spatial weighted layer and suitability calculation template, perform prior calculations on the geological observation data and assimilate and update it with the newly added geological observation data to generate voxel-level posterior suitability volume and posterior uncertainty volume.

[0073] Based on voxel-level spatial weighting layers and suitability calculation templates, geological observation data are aligned according to voxel unique indexes and timestamps, and aggregated to generate voxel observation sequences as prior input packages.

[0074] Furthermore, the drilling geological observation data are associated with the corresponding unique voxel index in the three-dimensional voxel grid according to their spatial location and timestamp. The observations of multiple phases and types under the same index are time-aligned and time-series integrated to form a set of observation records organized in chronological order with unique voxel indexes as identifiers. Then, combined with the weight configuration of each voxel observation feature in the voxel-level spatial weighted layer and the measurement caliber and adjudication order specified in the suitability calculation template, the observation records are structured and encapsulated, and the voxel observation feature sequence is aggregated as a priori input package.

[0075] The prior input package is subjected to threshold determination, weight aggregation and level mapping according to the suitability calculation template, and uncertainty assessment is performed using the variance propagation method to generate voxel-level prior suitability and prior uncertainty.

[0076] Furthermore, the voxel observation sequence in the prior input package is compared with the underground space suitability threshold item by item according to the measurement caliber and adjudication order specified in the suitability calculation template to complete the threshold determination. The determination results of each voxel observation feature are weighted and summarized according to the weights provided by the voxel-level spatial weighting layer, and the summarized value is mapped to the suitability level according to the preset grading standard. At the same time, based on the measurement error or temporal dispersion of each voxel observation feature, the variance propagation method is used to calculate the error accumulation in the weighted summarization process layer by layer, quantify the uncertainty of the voxel-level evaluation results, and finally generate the voxel-level prior suitability body and prior uncertainty body, providing an uncertain prior basis for subsequent assimilation and update.

[0077] It should be noted that the grading standard mapping is based on the underground space suitability threshold extracted from the geological suitability rules and is set according to the evaluation logic of suitability scoring specified in the measurement rules.

[0078] The newly added geological observation data are aligned with the voxel unique index and timestamp, and combined with the voxel-level prior suitability volume and prior uncertainty volume to generate the assimilation input set;

[0079] Furthermore, based on the spatial location and acquisition time of the newly added geological observation data, they are matched to the corresponding unique voxel index and timestamp in the three-dimensional voxel grid to complete spatiotemporal alignment. Subsequently, the aligned newly added geological observation data is associated with the existing voxel-level prior suitability volume and prior uncertainty volume according to the same unique voxel index, integrating the observation values, prior suitability scores and their uncertainty information to form a structured data set containing prior states and newly added geological observation data, generating an assimilation input set to provide the input basis for Bayesian updates.

[0080] The assimilation of the input set is performed using the Bayesian update method, and outlier removal and adaptive weight allocation are carried out to generate voxel-level posterior suitability and posterior uncertainty volumes.

[0081] Furthermore, the Bayesian update method is used to perform assimilation on the input set. This involves using the voxel-level prior suitability volume as the prior distribution and the newly added geological observation data as the likelihood information. The posterior probability distribution is calculated based on the Bayesian formula to obtain the updated voxel suitability estimate. During the assimilation process, residual analysis is used to identify observation points that deviate from the prior state, and outlier observations are removed. The contribution weights of each observation in the fusion are dynamically adjusted according to the observation accuracy and prior uncertainty to achieve adaptive weight allocation. Finally, voxel-level posterior suitability and posterior uncertainty volumes are generated, reflecting the latest geological suitability state and its reliability after the fusion of new observations.

[0082] S4. Superimpose the engineering constraint set with the voxel-level posterior suitability volume and posterior uncertainty volume, and calculate the geological suitability score according to the measurement rules of the geological suitability rule set.

[0083] Acquire engineering constraint data and resolve conflicts through standardization and voxelization encoding to generate an engineering constraint set;

[0084] Furthermore, engineering constraint data is obtained by collecting spatial control elements from urban planning approval documents, underground engineering design drawings, existing facility protection specifications, and construction organization plans, including engineering red lines, NIMBY control zones, safety protection zones for existing underground structures, and construction organization restriction areas. The engineering constraint data is uniformly converted to a coordinate system consistent with the 3D voxel mesh, standardized to eliminate format and semantic differences, and voxelized and encoded based on unique voxel indexes, marking spatial conflict areas as infeasible or partially feasible. On this basis, for cases where multiple engineering constraints overlap within the same voxel, conflict adjudication is performed according to a pre-defined spatial occupancy logic, ultimately generating a set of engineering constraints with voxel units and values ​​ranging from 0 to 1, used for constraint overlay in subsequent suitability scoring.

[0085] It should be noted that the spatial conflict area refers to the voxel area in which multiple control elements in the engineering constraint data overlap or interfere in three-dimensional space. It is identified by voxelizing and encoding the engineering constraint data and comparing it with the unique voxel index. The spatial occupancy logic refers to the rules for determining the constraint priority and superposition method when multiple engineering constraint conditions coexist in the same voxel. It is defined based on the safety control level, control intensity and construction stage sequence information in the engineering constraint data.

[0086] The set of engineering constraints is superimposed with the voxel-level posterior suitability volume and posterior uncertainty volume to output the constrained posterior suitability volume and posterior uncertainty volume.

[0087] Furthermore, the engineering constraint set and the voxel-level posterior suitability volume and posterior uncertainty volume are superimposed one voxel at a time according to the unique voxel index. The posterior suitability value of each voxel is multiplied by the engineering constraint mask value corresponding to that voxel in the engineering constraint set, so that the suitability of the infeasible region is set to zero or reduced proportionally. At the same time, the value of the posterior uncertainty volume at the same voxel position remains unchanged or is synchronously masked, and the constrained posterior suitability volume and posterior uncertainty volume are output.

[0088] Based on the constrained posterior suitability body and posterior uncertainty body, the geological suitability score is calculated according to the measurement rules of the geological suitability rule set, and the expression is:

[0089] ;

[0090] in, It is a voxel index, corresponding to a single voxel in a 3D voxel grid; It is a voxel The geological suitability score ranges from 0 to 1. It is the total number of geological observation data; It is a voxel The engineering constraint mask (derived from the engineering constraint set, with values ​​ranging from 0 to 1, where 1 indicates complete feasibility, values ​​between 0 and 1 indicate proportional reduction, and 0 is used for infeasible regions). It is the first Spatial weighting of geological observation data, representing voxels Weight at each location; It is the first Posterior scores of geological observation data; It is the uncertainty penalty coefficient (the range of values ​​is: 0≤ ≤1, used for pressing Discounted geological suitability score); It is a voxel Posterior uncertainty; Geological observation data Threshold constraint strength coefficient ( ≥0, used to control the severity of punishment when the threshold is not reached); Geological observation data The corresponding underground space suitability threshold.

[0091] It should be noted that in the expression , , as well as All were normalized to eliminate dimensional differences.

[0092] In the expression, First pass voxels were obtained The weighted aggregation results integrate the multi-source posterior scores according to importance into a comprehensive suitability score; subsequently, this is combined with the threshold compliance factor. The comprehensive suitability obtained after multiplication and rule constraints is used to assess the suitability of underground spaces that do not meet the suitability threshold. According to the strength coefficient Imposing penalties ensures the outcome reflects the binding force of the clauses; then multiplying by an uncertainty reduction term. The obtained "confidence-corrected overall suitability" reflects the posterior uncertainty based on voxel values. By coefficient The reduction is used to reflect the impact of confidence on the results; finally, the results are multiplied by an engineering constraint mask. Applying engineering constraints as feasibility scaling to the aforementioned values ​​to reflect engineering boundaries and limitations, thereby obtaining voxels. Geological suitability score .

[0093] S5. Perform voxel confidence weighted verification on the geological suitability score, conduct micro-re-exploration and re-evaluation on areas that fail the verification, and generate a geological suitability evaluation report.

[0094] Based on voxel-level posterior uncertainty and geological suitability scores, an inverse variance weighted normalization method is used for weighted fusion, and voxel confidence is calculated using data coverage as an availability factor. The expression is as follows:

[0095] ;

[0096] in, It is the voxel confidence level, with a value ranging from 0 to 1; This is the data coverage, with a value ranging from 0 to 1, representing the current voxel availability factor;

[0097] It should be noted that the data coverage is based on the suitability calculation template, which lists the expected observations of the voxel and assigns weights. After the collected observations are spatially associated with the voxels, quality checks and timeliness reductions are performed on each item. The data is then obtained by taking a weighted average based on effectiveness and representativeness and normalizing the result.

[0098] If the voxel confidence level is not lower than the confidence threshold and the geological suitability score is not lower than the underground space suitability threshold at the same time, the verification is judged to be unsuccessful, and a mask of unsuccessful regions is formed by connecting domain aggregation.

[0099] Furthermore, for each voxel, it is determined whether its voxel confidence score is lower than the confidence threshold or whether its geological suitability score is lower than the underground space suitability threshold. If either condition is not met, the voxel is marked as failing the verification. All voxels that fail the verification are subjected to adjacency analysis in three-dimensional space, and are aggregated into a continuous spatial region according to the six-neighbor or twenty-six-neighbor connectivity criterion. The corresponding binarized three-dimensional raster data is generated, which is the verification failure area mask, used to guide the delineation of the subsequent micro-survey scope.

[0100] Using the mask of areas that failed the verification as the scope, micro-surveys were carried out, and the newly added engineering geological observation data were aligned and recalculated to obtain the updated voxel-level posterior suitability volume and posterior uncertainty volume, as well as the optimized geological suitability score, and a geological suitability evaluation report was generated.

[0101] Furthermore, taking the area mask that failed the verification as the scope, additional drilling, supplementary geophysical profiles, or groundwater level and settlement monitoring points are deployed in the corresponding spatial area to carry out micro-supplementary exploration. The newly acquired engineering geological observation data are aligned to the unique voxel index in the three-dimensional voxel grid according to the collection location and timestamp, and used as new observation input to re-execute the prior calculation, Bayesian assimilation update, and engineering constraint superposition process to obtain the updated voxel-level posterior suitability volume and posterior uncertainty volume, as well as the optimized geological suitability score. Finally, all evaluation results and traceability information are integrated to generate a geological suitability evaluation report, realizing closed-loop optimization and reliable delivery of evaluation results.

[0102] This embodiment also provides a computer device applicable to the quantitative evaluation method of urban underground space geological suitability based on a large model, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the quantitative evaluation method of urban underground space geological suitability based on a large model as proposed in the above embodiment.

[0103] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0104] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the method for quantitative evaluation of the geological suitability of urban underground space based on a large model, as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0105] In summary, this invention achieves spatial adaptive fusion of rules and observations by mapping the geological suitability rule set to a three-dimensional voxel grid and applying boundary constraints to generate a voxel-level spatial weighted layer and solidifying it as a suitability calculation template; and by performing voxel confidence weighted verification on the geological suitability score and micro-re-excavation re-evaluation on areas that fail the verification, it achieves linkage optimization between geological suitability score and uncertainty, ultimately achieving stable, interpretable, and iterative results.

[0106] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A quantitative evaluation method for the geological suitability of urban underground space based on a large model, characterized in that: include, A three-dimensional voxel grid was constructed and multi-source data and standardized case corpus were aggregated. Then, the suitability threshold and measurement rules of underground space were extracted through Transformer to generate a traceable set of geological suitability rules. The geological suitability rule set is mapped to a 3D voxel mesh and boundary constraints are applied. Combined with geological observation data, a voxel-level spatial weighted layer is generated and solidified as a suitability calculation template. The steps are as follows. By combining GIS spatial overlay and 3D rasterization, the geological suitability rule set is mapped to a 3D voxel grid according to the unique voxel index and a list of applicable rule voxels is generated. Based on geological boundary data, a boundary constraint mask layer is generated using a spatial rasterization method, and the list of applicable regular voxels is masked and filtered to generate a list of applicable regular voxels after boundary constraints. Based on geological observation data, voxel observation feature sets are extracted, and combined with a constrained list of applicable rule voxels, an entropy weighting method is used to generate a voxel-level spatial weighted layer. The measurement scope and adjudication order of the voxel-level spatial weighted layer and the geological suitability rule set are solidified into a suitability calculation template; Based on voxel-level spatial weighted layers and suitability calculation templates, prior calculations are performed on geological observation data, and assimilation and updates are performed with newly added geological observation data to generate voxel-level posterior suitability volumes and posterior uncertainty volumes. The engineering constraint set is superimposed with the voxel-level posterior suitability volume and posterior uncertainty volume, and the geological suitability score is calculated according to the measurement rules of the geological suitability rule set. The geological suitability score is validated using voxel confidence weighting. Areas that fail the validation are reassessed through micro-surveys and backflow, and a geological suitability evaluation report is generated. The steps are as follows: Based on voxel-level posterior uncertainty and geological suitability scores, the inverse variance weighted normalization method is used for weighted fusion, and data coverage is used as the availability factor to calculate voxel confidence. If the voxel confidence level is not lower than the confidence threshold and the geological suitability score is not lower than the underground space suitability threshold at the same time, the verification is judged to be unsuccessful, and a mask of unsuccessful regions is formed by connecting domain aggregation. Using the mask of areas that failed the verification as the scope, micro-surveys were carried out, and the newly added engineering geological observation data were aligned and recalculated to obtain the updated voxel-level posterior suitability volume and posterior uncertainty volume, as well as the optimized geological suitability score, and a geological suitability evaluation report was generated.

2. The method for quantitative evaluation of the geological suitability of urban underground space based on a large model as described in claim 1, characterized in that: The construction of the three-dimensional voxel grid refers to setting the voxel resolution and layer thickness, as well as the starting point and grid range, based on the urban underground space boundary and unified coordinates, generating a regular three-dimensional grid by equal XYZ interval subdivision, and assigning a unique index to each voxel.

3. The method for quantitative evaluation of the geological suitability of urban underground space based on a large model as described in claim 2, characterized in that: The multi-source data includes borehole columnar sections, geophysical profiles, remote sensing and InSAR deformation time series, groundwater level and settlement monitoring time series, pipeline and underground structure vectors, and IFC models. The standardized case literature includes compilation standards, guidelines, local standards, historical engineering reports, and case texts.

4. The method for quantitative evaluation of the geological suitability of urban underground space based on a large model as described in claim 3, characterized in that: Standardize and clean the multi-source data and standardized case corpus to remove duplicates; Input the standardized case corpus into Transformer to perform term alignment and clause extraction, and output the underground space suitability threshold, measurement rules and source records; By utilizing the scale of three-dimensional voxel grids and the data caliber of multi-source data, the suitability thresholds and measurement rules for underground space are checked for consistency. After format solidification and number verification, a traceable set of geological suitability rules is generated.

5. The method for quantitative evaluation of the geological suitability of urban underground space based on a large model as described in claim 1, characterized in that: The steps for performing prior calculations on geological observation data and assimilating and updating it with newly added geological observation data are as follows: Based on voxel-level spatial weighting layers and suitability calculation templates, geological observation data are aligned according to voxel unique indexes and timestamps, and aggregated to generate voxel observation sequences as prior input packages. The prior input package is subjected to threshold determination, weight aggregation and level mapping according to the suitability calculation template, and uncertainty assessment is performed using the variance propagation method to generate voxel-level prior suitability volume and prior uncertainty volume.

6. The method for quantitative evaluation of the geological suitability of urban underground space based on a large model as described in claim 5, characterized in that: The steps for generating the voxel-level posterior suitability volume and posterior uncertainty volume are as follows: The newly added geological observation data are aligned with the voxel unique index and timestamp, and combined with the voxel-level prior suitability volume and prior uncertainty volume to generate the assimilation input set; The assimilation of the input set is performed using the Bayesian update method, and outlier removal and adaptive weight allocation are carried out to generate voxel-level posterior suitability and posterior uncertainty volumes.

7. The method for quantitative evaluation of the geological suitability of urban underground space based on a large model as described in claim 1, characterized in that: The steps for calculating the geological suitability score according to the measurement rules of the geological suitability rule set are as follows: Acquire engineering constraint data and resolve conflicts through standardization and voxelization encoding to generate an engineering constraint set; The set of engineering constraints is superimposed with the voxel-level posterior suitability volume and posterior uncertainty volume to output the constrained posterior suitability volume and posterior uncertainty volume. Based on the constrained posterior suitability body and posterior uncertainty body, the geological suitability score is calculated according to the measurement rules of the geological suitability rule set.

Images