CN-122020791-A - BIM prestressed pipe pile matching optimization system based on multisource geological data

CN122020791ACN 122020791 ACN122020791 ACN 122020791ACN-122020791-A

Abstract

The invention relates to the technical field of geotechnical engineering intelligent design, and discloses a BIM prestressed pipe pile matching optimization system based on multisource geological data. The system comprises a data preprocessing module, a feature extraction module, a feature perception module, a feature fusion module and a BIM pile matching generation module. The system performs standardized processing on original geological data, performs cluster analysis on the original geological data from a historical engineering case library, extracts geological mode features with space influence radius, encodes current site data through a multi-level feature perception architecture, fuses features of different levels by utilizing a bidirectional attention mechanism, maps the fused comprehensive geological features to a BIM environment, and automatically drives and generates an optimized pile allocation scheme. The invention realizes the intelligent multiplexing of engineering geological knowledge and the deep fusion of multi-scale geological features, can automatically and accurately generate the design scheme of the prestressed pipe pile highly adapted to the conditions of the complex field, and improves the design efficiency and the scientificity of the scheme.

Inventors

ZHU DAOFENG
LIU MIN
CHEN YUXIN
ZHAO YAN
LI HELI
LIU YINGTAO
ZHANG ZIJIAN
ZHANG HUAJIAN

Assignees

广东地山基础工程有限公司

Dates

Publication Date: 20260512
Application Date: 20260127

Claims (10)

1. A BIM prestressed pipe pile-mating optimization system based on multi-source geological data, the system comprising: The data preprocessing module receives an original investigation information set of a target site from a geological data source, performs data integrity and consistency verification, and generates a standardized multi-source geological data set; The feature extraction module is used for calling a stored engineering case library, performing clustering division on the historical prestressed pipe pile schemes based on stratum lithology and mechanical characteristic indexes to form a plurality of pipe pile scheme clusters with different geological backgrounds, extracting a core geological mode of each pipe pile scheme cluster, calculating the influence radius in space, and outputting geological mode feature vectors and feature influence domains; The feature perception module is used for constructing a multi-level feature perception architecture according to the dimensionality of the feature vectors of the geological modes, feeding the standardized multi-source geological data set into the multi-level feature perception architecture, carrying out space region division and feature coding on the data set at each level according to the corresponding feature influence domain, and outputting the level geological feature coding; the feature fusion module is used for carrying out bidirectional attention weighting and splicing on the hierarchical geological feature codes from different hierarchies through a built-in feature cross fusion unit to generate a fused site comprehensive geological feature expression; And the BIM pile matching generation module is used for mapping the site comprehensive geological feature expression to a BIM model environment, driving a pile matching assembly in the model and generating a pile matching optimization scheme according to the specification, the length and the arrangement parameters of the prestressed pipe piles.
2. A BIM prestressed pipe pile-fitting optimization system based on multi-source geological data according to claim 1, wherein extracting the core geological pattern of each pile plan cluster and calculating its spatially affected radius comprises: constructing a geological sequence sample set taking pile position coordinates and stratum parameter sequences as input for each tubular pile scheme cluster; Initializing a deep embedding network, wherein the deep embedding network comprises a stacking module for sequence modeling and a pooling module for feature concentration; Inputting the geological sequence sample set into the deep embedding network, capturing long-range dependency relationship among stratum parameters in the sample through the stacking module, and outputting high-dimensional sequence characteristics; inputting the high-dimensional sequence characteristics into the pooling module, and extracting a dense vector representing the common geological law of the tubular pile scheme cluster as the core geological mode through self-adaptive weighting aggregation operation; after extracting the core geologic pattern, calculating a feature space distance between original sequence features and the core geologic pattern of each geologic sequence sample in the tubular pile scheme cluster; And counting the distribution of the feature space distances of all samples, and defining the numerical value corresponding to the preset dividing point of the distance distribution as the feature influence domain of the core geological mode.
3. A BIM prestressed pipe pile-mating optimization system based on multi-source geological data according to claim 2, wherein said standardized multi-source geological data sets are fed into said multi-level feature-aware architecture, and wherein spatial region partitioning and feature encoding of the data sets is performed at each level according to corresponding feature impact domains, comprising: the perception range of each level of the multi-level feature perception architecture is matched with one feature influence domain; at a specific level of the multi-level feature perception architecture, acquiring a numerical value of the feature influence domain associated with the specific level, and taking the numerical value as a space neighborhood radius; defining a cube space area in the standardized multi-source geological data set by taking each exploration point or grid point in a target site as a center and the radius of the space neighborhood as a side length, and taking all geological data points in the cube space area as neighborhood data blocks of the current center point; Organizing all data points in the neighborhood data block into a three-dimensional tensor with a channel structure according to the spatial position and the geological attribute type of the data points; performing convolution operation on the three-dimensional tensor by using a three-dimensional convolution kernel group to extract local geological structure characteristics in the neighborhood of the neighborhood data block space; After processing all exploration points or grid points, rearranging all extracted local geological structure features according to the spatial positions of the central points of the local geological structure features to form the hierarchical geological feature code of the specific hierarchy, wherein the hierarchical geological feature code is a feature map corresponding to the space coordinates of a field.
4. A BIM prestressed pipe pile-mating optimization system based on multi-source geological data according to claim 3, wherein said hierarchical geological feature codes from different hierarchical levels are bi-directionally weighted and spliced by means of a built-in feature cross-fusion unit, comprising: The feature cross fusion unit receives hierarchical geologic feature codes from at least two different levels of the multi-level feature-aware architecture; encoding the hierarchical geologic features from a coarser sensing hierarchy, namely the hierarchy with a larger feature influence domain, and enabling the spatial resolution of the hierarchical geologic features to be consistent with the hierarchical geologic feature encoding from a finer sensing hierarchy through an up-sampling operation; splicing the up-sampled coarser level feature codes and finer level feature codes along the feature channel dimension to form a combined feature tensor; inputting the combined feature tensor into two independent convolution layers respectively to generate query feature mapping and key feature mapping for calculating attention weight; Calculating the dot product similarity of the query feature map and the key feature map at each spatial position, and generating a spatial attention weight map through normalization function processing; carrying out channel type weighting on the combined feature tensor by using the space attention weight graph, and highlighting feature response of important space positions; And inputting the re-weighted characteristic tensor into a fusion convolution layer to perform dimension reduction and fusion, and finally outputting the fused site comprehensive geological characteristic expression.
5. The BIM prestressed pipe pile-matching optimization system based on multi-source geological data of claim 4, wherein said combined feature tensors are respectively input into two independent convolution layers to generate a query feature map and a key feature map for computing attention weights, comprising: The two independent convolution layers have the same convolution kernel size but each possess an independent trainable weight parameter; After one convolution layer processes the combined feature tensor, the output feature map is defined as the query feature map, and the query feature map is used for representing a feature mode of information required to be acquired at each position; After the combined feature tensor is processed by another convolution layer, the output feature map is defined as the key feature map, and the key feature map is used for representing a feature mode of each position available for providing information; the query feature map and the key feature map have identical spatial dimensions and feature channel numbers to facilitate a point-by-point similarity matching calculation.
6. The BIM prestressed pipe pile-mating optimization system based on multi-source geological data of claim 1, wherein invoking the stored engineering case library performs cluster partitioning of historical prestressed pipe pile solutions based on formation lithology and mechanical property indexes, comprising: Reading all historical prestressed pipe pile scheme records from the engineering case library, wherein each record comprises a stratum lithology description sequence and a corresponding soil body mechanical parameter sequence of a construction site; digitally encoding the stratum lithology description sequence, converting lithology categories into multidimensional independent heat vectors, and simultaneously carrying out standardized treatment on the soil body mechanical parameter sequence to eliminate dimension differences; Connecting the processed lithologic vector sequence and the mechanical parameter sequence of each historical prestressed pipe pile scheme in depth to form a high-dimensional geological feature representation of the historical prestressed pipe pile scheme; searching a region with dense sample distribution in a space formed by the high-dimensional geological feature representation by adopting a clustering algorithm based on density; dividing all historical prestress tubular pile schemes in the same dense area into a set, wherein each set forms one tubular pile scheme cluster, and distributing a unique cluster identifier for each tubular pile scheme cluster.
7. The BIM prestressed pipe pile-matching optimization system based on multi-source geological data of claim 6, wherein searching for areas with dense sample distribution in the space formed by the high-dimensional geological feature representation by adopting a clustering algorithm based on density comprises: Setting a neighborhood searching radius parameter and a minimum neighborhood sample number parameter for the density-based clustering algorithm; Randomly selecting an unviewed historical prestressed pipe pile scheme sample as a starting point in the high-dimensional geological feature representation space; searching all other sample points in the hypersphere defined by the neighborhood searching radius parameter by taking the starting point as the center; if the number of the found sample points is greater than or equal to the minimum neighborhood sample number parameter, judging that the starting point is a core point, continuously recursively expanding the neighborhood of the starting point based on the sample points, and marking all the sample points which pass through the density as the same cluster; if the number of the found sample points is smaller than the minimum neighborhood sample number parameter, temporarily marking the starting point as a noise point; And repeating the processes of selecting the starting point, searching the neighborhood and marking until all historical prestressed pipe pile scheme samples are accessed, and finally finishing the division of a plurality of pipe pile scheme clusters.
8. The BIM prestressed pipe pile-mating optimization system based on multi-source geological data of claim 1, wherein receiving the original survey information set of the target site from the geological data source, performing data integrity and consistency checks, comprises: Creating a list of standard geological data fields, said list defining the data items and their formats that must be contained; Each data record in the received original investigation information set is compared with the standard geological data field list item by item, missing data fields, data fields with wrong formats and data fields with values obviously exceeding a reasonable range are identified, and a data exception report is generated; for the missing data field, according to the spatial position of the missing data field, adopting a Kriging interpolation algorithm, and carrying out spatial interpolation filling by using the similar data of the surrounding effective exploration points; For the data field with the format error, converting the data field into a standard format according to a predefined data conversion rule base; For the data field with the value obviously exceeding the reasonable range, marking the data field as data to be checked, triggering a manual checking flow, and after the manual checking, replacing the data field with a correction value or rejecting according to the technical specification; And outputting the standardized multi-source geological data set after all the verification and processing steps are completed.
9. The BIM prestressed pipe pile-matching optimization system based on multi-source geological data of claim 1, wherein mapping the site integrated geological feature expression to a BIM model environment drives pile-matching components in the model to generate a set of pile-matching optimization schemes according to specifications, lengths and arrangement parameters of prestressed pipe piles, comprising: In a BIM model environment, predefining a member family library of the prestressed pipe piles, wherein the member family library comprises pipe pile family types with various specification diameters, wall thicknesses, concrete strength grades and prestressed reinforcement configurations; inputting the site comprehensive geological feature expression into a parameterized rule reasoning engine, wherein a mapping rule set from geological features to tubular pile bearing capacity and sedimentation features is preset in the reasoning engine; The parameterization rule reasoning engine is used for reasoning a single pile bearing capacity characteristic value and a predicted settlement amount required by each pile-planning position according to the comprehensive geological feature expression of the field and the structural design load requirement; according to the single pile bearing capacity characteristic value and the estimated settlement, tubular pile group types meeting the conditions are automatically matched in the member group library, and based on preset economy and construction convenience weight, a plurality of feasible tubular pile specifications and length combinations are ordered; And generating a plurality of different tubular pile arrangement plan views and section views according to the sequencing result, wherein each arrangement scheme is used for listing the specifications, the lengths, the numbers, the pile top elevations and the arrangement intervals of the tubular piles in detail to form the group of pile allocation optimization schemes.
10. The BIM prestressed pipe pile matching optimization system based on multi-source geological data as claimed in claim 9, wherein said parameterization rule reasoning engine, according to said site comprehensive geological feature expression, combined with structural design load requirements, reasoning out single pile bearing capacity feature values and estimated settlement required by each quasi-pile position, comprises: The parameterized rule reasoning engine loads the site comprehensive geological feature expression, wherein the expression comprises characteristic values of soil layer compression modulus, internal friction angle, cohesive force and standard penetration number at each pile-simulating position; simultaneously, the parameterized rule reasoning engine reads the structural design load requirements from an external interface, wherein the requirements comprise vertical axial force, horizontal force and bending moment; The parameterization rule reasoning engine calls a built-in pile foundation bearing capacity calculation formula, takes soil layer parameters and structural load at a pile-simulating position as input variables, iteratively calculates pile end resistance and pile side resistance required for meeting bearing capacity requirements, and further reversely calculates to obtain the required single pile bearing capacity characteristic value; The parameterization rule reasoning engine calls a built-in settlement calculation formula by a layered sum method, calculates settlement of the pile foundation under different load levels by utilizing the characteristic value of the soil layer compression modulus and the estimated additional stress of the pile end to obtain the estimated settlement, and compares and verifies the calculated characteristic value of the single pile bearing capacity and the estimated settlement with a design specification allowable value to ensure that the single pile bearing capacity meets the safety requirement.

Description

BIM prestressed pipe pile matching optimization system based on multisource geological data Technical Field The invention relates to the technical field of geotechnical engineering intelligent design, in particular to a BIM prestressed pipe pile matching optimization system based on multisource geological data. Background Currently, application of building information model technology in pile foundation engineering is mostly limited to three-dimensional visual display and collision inspection of a given design scheme. The design process itself still relies heavily on the personal experience of the engineer, making decisions based on limited current site survey reports. The existing BIM platform lacks depth mining and structuring multiplexing capability for internal association of massive and multi-source historical engineering geological data and a successful pile allocation scheme, and the knowledge source of design decision is single and discrete. In the aspect of intelligent auxiliary decision making by using geological data, the existing method mostly adopts a mode of directly carrying out statistical analysis on the stratum parameters of the current site or carrying out simple comparison with a few similar cases. The methods are difficult to systematically refine and quantify the commonality rules of the pile matching modes which are proved by practice and are economical and reasonable under different geological conditions. Meanwhile, the conventional data processing mode generally regards the geological features as homogeneity or independent attributes to be overlapped, and the differential and nonlinear comprehensive influence of the geological features with different scales and different influence ranges on design parameters such as pile length, pile distance and the like cannot be effectively expressed. This results in the resulting pile deployment solution often being either conservative or poorly suited to the actual geology of the site. Disclosure of Invention The invention aims to provide a BIM prestressed pipe pile matching optimization system based on multi-source geological data, so as to solve the problems in the background art. To achieve the above object, the present invention provides a BIM prestressed pipe pile-matching optimization system based on multi-source geological data, the system comprising: The data preprocessing module receives an original investigation information set of a target site from a geological data source, performs data integrity and consistency verification, and generates a standardized multi-source geological data set; The feature extraction module is used for calling a stored engineering case library, performing clustering division on the historical prestressed pipe pile schemes based on stratum lithology and mechanical characteristic indexes to form a plurality of pipe pile scheme clusters with different geological backgrounds, extracting a core geological mode of each pipe pile scheme cluster, calculating the influence radius in space, and outputting geological mode feature vectors and feature influence domains; The feature perception module is used for constructing a multi-level feature perception architecture according to the dimensionality of the feature vectors of the geological modes, feeding the standardized multi-source geological data set into the multi-level feature perception architecture, carrying out space region division and feature coding on the data set at each level according to the corresponding feature influence domain, and outputting the level geological feature coding; the feature fusion module is used for carrying out bidirectional attention weighting and splicing on the hierarchical geological feature codes from different hierarchies through a built-in feature cross fusion unit to generate a fused site comprehensive geological feature expression; And the BIM pile matching generation module is used for mapping the site comprehensive geological feature expression to a BIM model environment, driving a pile matching assembly in the model and generating a pile matching optimization scheme according to the specification, the length and the arrangement parameters of the prestressed pipe piles. Preferably, extracting the core geologic pattern of each pile plan cluster and calculating the spatially affected radius thereof comprises: constructing a geological sequence sample set taking pile position coordinates and stratum parameter sequences as input for each tubular pile scheme cluster; Initializing a deep embedding network, wherein the deep embedding network comprises a stacking module for sequence modeling and a pooling module for feature concentration; Inputting the geological sequence sample set into the deep embedding network, capturing long-range dependency relationship among stratum parameters in the sample through the stacking module, and outputting high-dimensional sequence characteristics; inputting the high-dimensional sequence characte