CN-121997774-A - Intelligent design system of slide-resistant pile structure suitable for complicated stratum condition

CN121997774ACN 121997774 ACN121997774 ACN 121997774ACN-121997774-A

Abstract

The application belongs to the field of geotechnical engineering and structural design, in particular to an intelligent design system of an anti-slide pile structure suitable for complex stratum conditions, which comprises a multi-source data standardization integration module, a space topological graph construction module based on constraint Delong triangulation, a layer sequence evolution reasoning engine based on a graph attention network, a three-dimensional geological probability field and confidence degree mapping module, a pile soil coupling stress analysis module and a safety reserve self-adaptive design module.

Inventors

ZHANG FENGSHU
QIN YI
LI GEN
LIU XIN
ZHAO JIE
ZHANG SHUNTAO
YANG YUJING
WU E
ZENG WENZHAO
LV WEI

Assignees

四川省地质环境调查研究中心

Dates

Publication Date: 20260508
Application Date: 20260407

Claims (10)

1. An intelligent design system of a slide-resistant pile structure suitable for complex stratum conditions, which is characterized by comprising: The multi-source geological survey data standardization integration module is used for carrying out digital extraction, standardization processing and logic deviation correction on original survey information to construct a high-dimensional node data set containing space three-dimensional coordinates and attribute feature vectors; The space topological graph construction module is used for converting the discrete high-dimensional nodes output by the multi-source geological survey data standardization integration module into a global space topological graph containing a horizontal space adjacent relation and a vertical layer sequence constraint; The layer sequence evolution reasoning engine based on the graph attention network is used for aggregating node characteristics of the global space topological graph in a non-Euclidean space through a multi-head attention mechanism, reasoning a layer evolution rule of a non-sampling area according to a geological layer sequence rule and outputting layer type probability distribution; The three-dimensional geological probability field and confidence level mapping module is used for constructing a dense three-dimensional grid element model in a region to be designed based on the output result of the layer sequence evolution reasoning engine, and generating a three-dimensional confidence level distribution map reflecting the reliability of a geological structure in a quantization mode by utilizing an information entropy algorithm; The pile-soil coupling stress analysis and dynamic pressure field calculation module is used for extracting probabilistic mechanical parameters of the pile-laying position according to the three-dimensional geological probability field and the three-dimensional confidence distribution map, wherein the probabilistic mechanical parameters comprise cohesive force, internal friction angle and deformation modulus obtained by stratum type probability weighting statistics, and a pile circumference dynamic pressure field distribution interval under the action of landslide thrust is simulated through a nonlinear finite element analysis unit; the anti-slide pile structure safety reserve self-adaptive design module based on the confidence coefficient is used for establishing a risk-driven parameter optimization mechanism, and the safety coefficient adjusting unit dynamically compensates the structural design parameters of the anti-slide pile according to the confidence coefficient level of the area where the pile body is located, so that the intelligent self-adaptive design of the anti-slide pile structure is realized.
2. The intelligent design system of the slide pile structure suitable for complex stratum conditions according to claim 1, wherein the multi-source geological survey data standardization integration module is provided with a geological drilling data analyzer, the geological drilling data analyzer extracts characteristic fields containing drilling coordinates, orifice elevations, stratum demarcation point depths, groundwater levels and physical and mechanical parameters of a rock and soil body from an unstructured text survey report through a regular expression algorithm, the physical and mechanical parameters of the rock and soil body comprise cohesive force, internal friction angles, weight, compression modulus and permeability coefficients, the attribute characteristic vectors comprise rock and soil layer numbers, lithology descriptors, standard penetration numbers, shear wave speeds, cohesive force, internal friction angles and weight, the multi-source geological survey data standardization integration module further comprises a normalization processing unit, the multi-source geological survey data standardization integration module is further provided with a geological information correction sub-module, and the geological information correction sub-module compares predicted values with actual values through an embedded knowledge base, and generates a new-level conflict map data when the same spatial contradiction between the original dynamic map data is updated and the network map data is updated.
3. The intelligent design system of the slide pile structure suitable for complex stratum conditions, according to claim 1, is characterized in that the space topological graph construction module based on constraint Delong triangulation executes the following topological modeling flow: Constructing a constraint Delong triangle network based on two-dimensional coordinates of the drilling holes on a horizontal projection plane, and constructing a logic connection path of adjacent drilling holes and abstracting the logic connection path as edges of the graph; Constructing a longitudinal sequence chain in each drilling node according to the depth sequence revealed by the stratum, arranging stratum contact points with different depths according to the depth sequence, and endowing the stratum contact points with sequence topology labels to form a transverse and longitudinal interweaved global graph structure; The space topological graph construction module based on constraint Delould internal triangulation supports the explicit expression of a three-dimensional fracture network and faults, and the information transfer direction in the structure of the constraint graph is forced by inserting a discontinuous surface operator between stratum nodes so as to simulate the cutoff effect of fracture on stratum continuity.
4. The intelligent design system of the slide pile structure suitable for complex stratum conditions is characterized in that a layer sequence evolution reasoning engine based on a graph attention network adopts a deep residual error connection structure, comprises a plurality of parallel graph attention layers, each layer is provided with a plurality of independent attention heads, each drilling node is used as a central node in the reasoning process, weight coefficients are automatically learned and distributed by calculating cosine similarity of the central node and a neighborhood node in a feature space, the weight coefficients are constrained by geological layer sequence rules, stratum type probability distribution of spatial points of a non-sampling area is predicted by feature aggregation of continuous extension, pinch and non-integration contact evolution rules of a stratum, and layer normalization processing and residual error addition operation are arranged after each layer of graph attention operation.
5. The intelligent design system of the slide pile structure for complex formation conditions according to claim 4, wherein the layer sequence evolution inference engine based on the graph attention network supports non-isotropic attention weight distribution, and the attention mechanism introduces azimuth operators to enable the weight distribution to have directional relevance, namely the engine can distribute higher attention weight along the formation trend and lower weight along the layer sequence direction perpendicular to the formation trend so as to adapt to the non-isotropic characteristics of the inclined formation or the fold structure.
6. The intelligent design system of the anti-slide pile structure suitable for complex stratum conditions is characterized in that the three-dimensional geological probability field and confidence mapping module is used for endowing each voxel in a three-dimensional grid model with a probability vector, the probability vector represents probability values of the position belonging to different stratum categories, the three-dimensional confidence distribution map is obtained by conducting Shannon entropy calculation on the probability vectors, the three-dimensional geological probability field and confidence mapping module is provided with a space-time dynamic updating function, drilling parameter monitoring data are accessed in real time in the construction stage of the anti-slide pile, and stratum information revealed in the field is fed back into the probability field model through a Bayesian updating algorithm for dynamic correction.
7. The intelligent design system of the anti-slide pile structure suitable for complex stratum conditions, which is characterized in that the three-dimensional geological probability field and confidence mapping module adopts a probability modeling method based on variation reasoning to subdivide a confidence map into two sub-dimensions of model confidence and observation confidence, wherein the model confidence reflects the inference capability of the system on geological evolution rules, and the observation confidence reflects the coverage density of investigation holes.
8. The intelligent design system of the anti-slide pile structure suitable for complex stratum conditions is characterized in that pile-soil coupling stress analysis and dynamic pressure field calculation modules utilize nonlinear finite element analysis units to simplify pile-periphery soil bodies into resistance media formed by multiple layers of nonlinear springs, stiffness coefficients of the springs are determined by mechanical parameters weighted by stratum type probability, when landslide thrust is simulated, the system searches potential sliding surfaces by adopting an intensity reduction method, calculates residual sliding force on the sliding surfaces based on three-dimensional stratum topology logic, outputs pile-periphery lateral pressure fields with standard deviation distribution intervals, and further adopts an improved p-y curve model, skeleton curve parameters of the pile-soil coupling stress analysis and dynamic pressure field calculation modules are obtained by nonlinear integration of stratum probability distribution functions to cover extreme stress working conditions caused by complex stratum heterogeneity, the pile-soil coupling stress analysis and dynamic pressure field calculation modules integrate fluid-solid coupling analysis units, and simulate the automatic calculation of the contribution of the pile-periphery lateral pressure fields to the dynamic pressure load by converting the stratum probability fields into permeability coefficient fields in the static pore and fracture zones.
9. The intelligent design system of the anti-slide pile structure suitable for complex stratum conditions is characterized in that a safety coefficient adjusting unit in an anti-slide pile structure safety reserve self-adaptive design module based on confidence coefficient is linked with a confidence coefficient mapping module in real time, when the pile diameter, the reinforcement ratio and the anchoring depth of an anti-slide pile are designed, the system performs dynamic safety compensation according to confidence coefficient indexes of areas where sections of a pile body are located, the anti-slide pile structure safety reserve self-adaptive design module based on the confidence coefficient has a multi-objective optimization function, and pareto optimal solutions are found among safety confidence coefficient, engineering cost and construction period through a genetic algorithm.
10. The intelligent design system of the anti-slide pile structure suitable for complex stratum conditions is characterized by further comprising a virtual reality monitoring interface, a parallel computing acceleration layer and an expert rule inspection engine, wherein the virtual reality monitoring interface is used for conducting three-dimensional visual rendering and slicing observation on an anti-slide pile stress cloud picture, a three-dimensional geological probability field and a confidence cloud picture, the parallel computing acceleration layer is used for conducting hardware-level acceleration on graph injection force operation and finite element analysis by means of parallel processing capacity of a graphic processor, and the expert rule inspection engine comprises geological disaster management standard terms used for conducting automatic compliance on minimum spacing, minimum burial depth and structural requirements of the anti-slide pile.

Description

Intelligent design system of slide-resistant pile structure suitable for complicated stratum condition Technical Field The invention belongs to the field of geotechnical engineering and structural design, and particularly relates to an intelligent design system of an anti-slide pile structure suitable for complex stratum conditions. Background The anti-slide pile is a core supporting structure for slope stability control and landslide control, the design accuracy is related to infrastructure safety and engineering cost optimization, civil engineering digitally promotes anti-slide pile parametric modeling, stress analysis and pile distribution optimization to become the main stream, traditional intelligent design is based on geological drilling discrete data, a three-dimensional geological model is built through an interpolation algorithm, sliding body evolution is simulated, pile body design parameters are deduced, applicability is strong in simple stratum, and design efficiency can be improved. Along with the progress of engineering to complex geological areas, the existing intelligent design system of the anti-slide pile has obvious limitation, a main stream interpolation algorithm regards stratum as a continuous isotropic scalar field, geological sequence topological logic is stripped, stratum dislocation and horizon distortion are easy to generate when complex stratum is processed, and the pile periphery soil body resistance distribution is inconsistent with the actual stress environment and the pressure field calculation distortion is caused. The prior art lacks a geological uncertainty processing mechanism, the geological risk of a non-sampling area in a complex stratum is high, the system can only provide a single deterministic three-dimensional model, and reliability evaluation and risk early warning cannot be given, so that a designer cannot judge the design reliability, and hidden danger of insufficient design strength or excessive conservation of engineering easily occurs. Therefore, the invention provides an intelligent design system of the slide-resistant pile structure, which is suitable for complex stratum conditions. Disclosure of Invention In order to overcome the defects in the prior art and solve at least one technical problem proposed in the background art, The technical proposal adopted for solving the technical problems is that the intelligent design system of the anti-slide pile structure suitable for complex stratum conditions comprises a multisource geological survey data standardization integration module, a space topological graph construction module based on constraint Delong triangulation, a layer sequence evolution reasoning engine based on a graph attention network, a three-dimensional geological probability field and confidence degree mapping module, a pile soil coupling stress analysis and dynamic pressure field calculation module and an anti-slide pile structure safety reserve self-adaptive design module based on confidence degree, The multi-source geological survey data standardization integration module is technically characterized in that a geological drilling data analyzer is arranged and is used for digitally extracting original survey information comprising drilling coordinates, orifice elevations, stratum demarcation point depths and physical and mechanical parameters of a rock and soil body, the geological drilling data analyzer converts an unstructured text survey report into a structured space vector data set through a built-in regular expression algorithm, each sampling point is defined as a high-dimensional node comprising a space three-dimensional coordinate and an attribute characteristic vector, the attribute characteristic vector comprises a rock and soil layer number, a lithology descriptor, a standard penetration number, a shear wave speed, a cohesive force, an internal friction angle and a gravity, in order to eliminate the influence of different dimensions on a follow-up reasoning engine, the module is further provided with a normalization processing unit which is used for carrying out linear scaling processing on the physical and mechanical parameters, The module is used for converting discrete drilling nodes into a graph structure with topological connection relation, and executing the following engineering flow, wherein firstly, a constrained deluxe triangle network is constructed based on two-dimensional coordinates of drilling holes on a horizontal projection plane to ensure that a logic connection path is established between adjacent drilling holes, secondly, each drilling hole is abstracted into nodes in the graph structure, edges in the triangle network are abstracted into edges of the graph, thus constructing a global graph reflecting the space distribution characteristics of a site, further, in order to capture the sequence characteristics of stratum in the vertical direction, the module is used for constructing a longitudinal layer sequence ch