An optimization method and system for bridge abutment backfilling scheme

Abstract

This invention discloses a method and system for optimizing bridge abutment backfill schemes, relating to the field of digital engineering scheme optimization. The bridge abutment backfill scheme optimization system includes: a data acquisition module, a modeling and partitioning module, a frost heave prediction module, a mechanical analysis module, a scheme generation module, a simulation analysis module, an optimization module, a scheduling generation module, and a BIM management module. This invention constructs a meteorological-geological coupled prediction model, integrating meteorological and geological information, to achieve quantitative prediction of freezing depth and identification of frost heave risk levels, significantly improving the adaptability and foresight of backfill design in climate-sensitive areas. Through simulation analysis of structural stress and deformation, combined with the load conditions and foundation response characteristics of the backfill area, a mechanical fit level is generated to ensure that candidate backfill schemes meet structural safety requirements.

Description

Technical Field

[0001] This invention relates to the field of digital engineering scheme optimization, and in particular to a method and system for optimizing bridge abutment backfill schemes. Background Technology

[0002] Backfill behind bridge abutments is a crucial transitional area between the bridge substructure and the roadbed. Its construction quality directly affects the long-term stability of the abutment, the efficiency of structural stress transfer, and the development trend of joint defects (such as vehicle slumping and differential settlement). Especially in high-altitude or cold-temperate regions with frequent seasonal freeze-thaw cycles, the backfill area behind the abutment is highly susceptible to frost heave stress and hydrothermal migration effects, inducing structural deformation, loosening of fill material, and local voids. In severe cases, this can affect the load-bearing capacity of the abutment and traffic safety.

[0003] Traditional optimization of bridge abutment backfill schemes often relies on experience-based material selection and standardized construction paths. The abutment area lacks fine-grained delineation, making it difficult to identify risk levels and differentiate materials based on local conditions. Frost heave response assessments lack multi-source data coupling and prediction, resulting in poor sensitivity of material configuration and construction schemes to climate change. The lack of digital simulation analysis of backfill schemes makes it difficult to quantitatively compare and reconstruct boundaries among multiple candidate schemes. Furthermore, there is a lack of optimization methods for backfill repair schemes addressing backfill anomalies in the abutment areas of existing bridges.

[0004] Therefore, there is an urgent need for a digital optimization method for bridge abutment backfilling schemes that integrates multi-source data-driven approaches, supports regional risk identification, scheme generation and simulation evaluation, and is suitable for permafrost regions, so as to achieve intelligent and quantitative support for the construction or repair of abutment backfilling for new and existing bridges. Summary of the Invention

[0005] This invention proposes an optimization method and system for bridge abutment backfill schemes, which integrates modules such as 3D modeling, frost heave prediction, structural analysis, material selection and simulation optimization. It breaks through the technical limitations of existing construction experience-oriented and holistic processing, and realizes the refinement, data-driven approach and intelligent decision support of bridge abutment backfill schemes.

[0006] An optimization method for bridge abutment backfilling schemes includes:

[0007] Collect topographic and geological information of the bridge abutment backfill area;

[0008] The design information of the bridge abutment is obtained, and a three-dimensional model of the abutment and abutment back area is constructed by combining the terrain information. The foundation backfill area is automatically divided, and a BIM model of the bridge abutment backfill is generated.

[0009] For each basic backfill area in the BIM model, obtain meteorological information related to frost heave and input it into the meteorological-geological coupled prediction model. Combine the geological information to calculate the predicted value of freezing depth and output the corresponding frost heave risk level.

[0010] Based on the frost heave risk level of each foundation backfill area, and combined with the load conditions and foundation response characteristics in the design information, a simulation analysis of structural stress and deformation is conducted to generate the mechanical fit level of each foundation backfill area.

[0011] Based on the frost heave risk level and mechanical compatibility level, the material performance database is called to generate multiple candidate backfill schemes for each basic backfill area. Each scheme includes backfill depth, backfill material combination and ratio, and outputs the corresponding material type, layer thickness and compaction parameters.

[0012] Simulation is performed on each candidate backfill scheme for each basic backfill area to evaluate performance indicators. Based on the preset optimization objective function, the final backfill scheme for each basic backfill area is selected. The results of the simulation are combined to adjust the boundaries of the basic backfill areas and merge and reorganize them to generate an optimized construction area division scheme.

[0013] An optimized construction plan, including work sequence, work path, and resource allocation, is generated based on the construction area division scheme and the final backfilling scheme.

[0014] As a preferred embodiment of the present invention, the automatic division of the basic backfill area includes:

[0015] Based on the 3D model of the bridge abutment and its back area, structural boundary features, slope ratio, elevation distribution, and relative positional relationship with the bridge abutment are extracted. The 3D space is divided into multiple foundation backfill areas according to preset partitioning rules, including slope threshold, layer elevation, and structural support and non-support areas. The automatic partitioning results are written into the attribute fields of the corresponding components in the BIM model.

[0016] As a preferred embodiment of the present invention, the meteorological-geological coupled prediction model includes:

[0017] The input module is used to receive meteorological and geological information corresponding to each basic backfill area;

[0018] The feature construction module is used to normalize, expand time series data, and encode features from the input data.

[0019] The prediction module, built on a random forest regression model, is used to establish a nonlinear mapping relationship between input data and predicted frozen depth values, and to calculate the predicted frozen depth values.

[0020] The training module is used to perform supervised training of the prediction module using historical measured frozen depth data and output the optimal model weight parameters.

[0021] As a preferred embodiment of the present invention, the calculation of the predicted freezing depth includes:

[0022] For each basic backfill area, a multi-dimensional input vector containing meteorological and geological information is constructed. The meteorological information includes the daily average surface temperature sequence, the number of days of freezing, the snowfall thickness, and the temperature fluctuation range. The geological information includes the underground temperature profile, soil type, soil moisture content, and regional permafrost distribution type. The multi-dimensional input vector is mapped to the trained random forest regression model through the prediction module. The predicted value of local freezing depth is output according to the partition path of each decision tree. The predicted value of freezing depth of the area is generated by the weighted average of all trees.

[0023] As a preferred embodiment of the present invention, the simulation analysis of structural stress and deformation includes:

[0024] A coupled model of a multi-layer backfill structure foundation was constructed using the finite element method. Based on the frost heave risk level, load conditions, and foundation response characteristics of each foundation backfill area, corresponding frost heave boundary conditions and load boundary conditions were set. The vertical deformation, equivalent stress distribution, and contact pressure response of the foundation backfill area under frost heave displacement were calculated through simulation. The response values ​​were compared with the preset safety threshold to determine the structural safety level range and generate the corresponding mechanical adaptation level.

[0025] As a preferred embodiment of the present invention, the simulation includes:

[0026] Candidate backfill schemes corresponding to each foundation backfill area are written into the temporary attribute fields of the corresponding components in the BIM model, and the spatial location, soil conditions, boundary constraints and material parameters of the components are extracted to construct a structured simulation input dataset. The simulation input dataset is imported into the finite element simulation platform, and settlement response analysis, lateral stability analysis and frozen boundary transmission analysis are performed on each candidate backfill scheme based on frost heave loading conditions and structural load conditions to obtain the response values ​​of the candidate backfill schemes under three types of performance indicators.

[0027] As a preferred embodiment of the present invention, the optimized construction area division scheme includes:

[0028] Based on performance indicators, a multi-objective weighted optimization function with performance balance and material coordination as weights is constructed to select the final backfill scheme from candidate backfill schemes. According to the final backfill scheme results of each foundation backfill area, its spatial continuity is analyzed. Based on the set area reorganization rules, the boundaries of adjacent areas are adjusted and merged and reorganized to output an optimized construction area division scheme that meets the requirements of construction continuity, material uniformity and risk controllability. This scheme is then written into the attribute fields of the corresponding components in the BIM model.

[0029] As a preferred embodiment of the present invention, an optimization method for bridge abutment backfill scheme further includes:

[0030] For existing bridges with abutment anomalies, structural measurement information is obtained through 3D laser scanning and ground-penetrating radar, replacing design information as the input for building the BIM model. Based on the integrity of the backfill structure, the distribution of voids, and the compaction state, the initial state information of the foundation backfill area is defined. The process of obtaining the corresponding frost heave risk level, mechanical compatibility level, and candidate backfill schemes takes into account the physical performance damage of the original backfill material and the uncertainty of the structural boundary. The simulation is compared between two types of schemes: replacement reconstruction or superimposed reinforcement, and corresponding repair backfill schemes and construction area adjustment schemes are generated.

[0031] A bridge abutment backfill optimization system includes:

[0032] The data acquisition module is used to collect topographic, geological and meteorological information of the bridge abutment backfill area, as well as to obtain design information or measured structural information of the bridge abutment.

[0033] The modeling and partitioning module is used to construct a 3D model of the bridge abutment and its backfill area, automatically partition the foundation backfill area, and generate a BIM model of the bridge abutment backfill.

[0034] The frost heave prediction module, based on a meteorological and geological coupled prediction model, calculates the predicted freezing depth for each basic backfill area and outputs the frost heave risk level.

[0035] The mechanical analysis module is used to simulate and analyze the stress and deformation of the structure, and generate the mechanical fit level for each foundation backfill area.

[0036] The scheme generation module is used to generate candidate backfill schemes by calling the material performance database based on the frost heave risk level and mechanical compatibility level.

[0037] The simulation analysis module is used to simulate candidate backfill schemes and output performance indicators.

[0038] The optimization module is used to select the final backfill scheme based on a preset optimization objective function, reorganize the boundaries of the foundation backfill area, and output the construction area division scheme.

[0039] The scheduling generation module is used to generate optimized construction plans based on the construction area division plan and the final backfill plan;

[0040] The BIM management module is used to manage the optimization process of backfill schemes based on BIM models.

[0041] The present invention has the following advantages:

[0042] This invention achieves refined management and component-level data representation of the backfill area of ​​the bridge abutment and its backfill region by automatically dividing the foundation backfill area and writing its structural boundaries, elevation information and spatial characteristics into the BIM model, providing basic support for subsequent differential analysis and optimization. By being compatible with the existing bridge abutment repair process, it introduces three-dimensional laser scanning and ground-penetrating radar data to construct the BIM model, and generates repair-type backfill schemes under the conditions of material damage and structural uncertainty, thus expanding the applicability and engineering scenarios of the method.

[0043] This invention constructs a meteorological-geological coupled prediction model, integrating meteorological and geological information, to achieve quantitative prediction of freezing depth and identification of frost heave risk levels, significantly improving the adaptability and foresight of backfill design in climate-sensitive areas. Through simulation analysis of structural stress and deformation, combined with the load conditions and foundation response characteristics of the backfill area, a mechanical fit level is generated to ensure that candidate backfill schemes meet structural safety requirements.

[0044] This invention constructs a database of candidate backfill schemes and combines it with simulation to evaluate performance, enabling quantitative comparison and optimization across multiple dimensions such as settlement control, structural stability, and freezing response. This enhances the engineering adaptability and cost rationality of material configuration. By reorganizing and merging the boundaries of the foundation backfill area based on an optimization objective function, an optimized construction area division scheme that satisfies construction continuity, material uniformity, and construction feasibility is generated, thereby improving the overall backfill quality and efficiency. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are only schematic diagrams of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0046] Figure 1 This is a schematic diagram of a bridge abutment backfill optimization system used in an embodiment of the present invention. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0048] Example 1: An optimization method for bridge abutment backfilling scheme, comprising the following steps:

[0049] Step S1: Collect topographic and geological information of the bridge abutment backfill area; obtain the design information of the bridge abutment backfill, and construct a three-dimensional model of the abutment and abutment backfill area in combination with the topographic information; automatically divide the foundation backfill area and generate a BIM model of the bridge abutment backfill.

[0050] Terrain information: Obtain terrain data around the bridge through UAV oblique photogrammetry, LiDAR scanning or 3D laser point cloud to obtain a high-precision digital terrain model (DTM), which provides geometric attribute information including at least the regional slope, aspect and elevation distribution.

[0051] Geological information: Extracted using drilling, in-situ testing, ground-penetrating radar, and other methods in conjunction with existing geological survey reports, including at least the soil type, stratification, groundwater level depth, water content, permafrost distribution type, and underground temperature profile of the platform backwater area. This information will serve as an important input variable for subsequent frost heave risk prediction and backfill material selection.

[0052] For example, for a highway bridge located in a cold region, the soil layers in the abutment area, from top to bottom, are fill, silty clay, and strongly weathered rock. Ground temperature monitoring data shows that the average freezing depth in winter is 1.3m, and the groundwater level is typically 2.1m deep.

[0053] Design information: Extracted from design documents, including at least the abutment structure type (gravity type, rib type, or enlarged foundation type), abutment back structure dimensions, design elevation, support location, elevation difference, and load distribution characteristics, to guide the parametric modeling of BIM components and the setting of area division rules.

[0054] By integrating topographic and design information, a three-dimensional geometric model of the abutment and backfill area is constructed based on a modeling platform (such as Revit), and geological information is attached to the corresponding stratigraphic components in the form of attributes to generate a BIM basic component model with attribute information.

[0055] The foundation backfill area refers to multiple independent analysis units between the bridge abutment structure and the surrounding soil, which are divided according to factors such as geometric location, structural function and geological characteristics, and are used to realize regional risk assessment and differentiated solution configuration.

[0056] The automatic division of the basic backfill area includes:

[0057] Based on the 3D model of the abutment and its back slope area, structural boundary features (slope aspect), slope ratio (source: DTM model calculation), elevation distribution (from laser scanning or photogrammetry), and relative positional information with respect to the abutment (such as the intersection of the back slope of the abutment and the slope) are extracted. The 3D space is divided into multiple basic backfill areas according to preset zoning rules. The zoning rules include slope threshold (areas with a slope greater than 1:1.5 need to be separately divided into high-risk areas), layer elevation (dividing backfill layers every 0.5m or 1m), and the logic for dividing the structural support area and non-support area (distinguishing between the direct bearing area and the free backfill area by the abutment bottom plate, wing wall, or support pile). The results of the automatic division are written into the attribute fields of the corresponding components in the BIM model.

[0058] The results of the automatic partitioning are written into the attribute fields of the corresponding components in the BIM model through script tools (such as Dynamo), including at least the area number, the layer in which it is located, the identifier of adjacent structural components, and whether it is in the critical influence zone of the structure.

[0059] Example: In the abutment back area of ​​a continuous rigid frame bridge, the model is divided into three areas according to the zoning rules: structural support area A1, structural free backfill area B1, and slope high-risk area C1. Each area has a unique number and attribute identifier.

[0060] The data generated in this step will flow into subsequent steps through the following paths: the basic backfill area and its attribute fields extracted from the 3D model will be used as input to the meteorological and geological coupled prediction model for freezing depth prediction; the spatial location and layering information in the same model will be input to the finite element simulation platform for structural response analysis; all partitioning results will be uniformly visualized through the BIM platform, supporting attribute-based querying, filtering, and dynamic updates.

[0061] Step S2: For each basic backfill area in the BIM model, obtain meteorological information related to frost heave and input it into the meteorological-geological coupled prediction model. Combine the geological information to calculate the predicted value of freezing depth and output the corresponding frost heave risk level.

[0062] Meteorological information: obtained through historical meteorological station data or remote sensing climate models, including at least the daily average surface temperature sequence, used to construct the heat conduction boundary conditions during the freezing process; the number of consecutive days of freezing, the number of days with continuous temperatures below 0°C; snowfall thickness, the snow layer as an insulating layer affecting the freezing depth; and the temperature fluctuation range, measuring the degree of disturbance to the soil caused by repeated freeze-thaw cycles.

[0063] For example, in a bridge abutment area in Northeast China, the average number of days of freezing in the winter of 2021 was 84 days, the maximum daily snowfall thickness was 26 cm, and the temperature fluctuation range was as high as ±12℃.

[0064] The geological information used in the model is analyzed as follows: underground temperature profile, reflecting the initial thermal field distribution at different depths; soil type, such as clay, silt, and sand, with significant differences in thermal conductivity and frost heave coefficient; soil moisture content, with higher moisture content indicating greater freezing potential; and regional permafrost distribution type, such as seasonally frozen soil, perennially frozen soil, or permafrost-free areas.

[0065] The meteorological-geological coupled prediction model includes the following modules:

[0066] The input module is used to receive combined meteorological and geological information data corresponding to each basic backfill area, and encode them uniformly to form a structured input vector;

[0067] The feature construction module is used to process the input data as follows:

[0068] Normalization: Used to eliminate dimensional differences (such as different units for temperature and thickness).

[0069] Time series expansion: Converting daily temperature series into statistical features (maximum value, mean, trend term);

[0070] Feature encoding: Convert categorical variables (such as soil type) into one-hot encodings or embedding vectors.

[0071] The prediction module establishes a nonlinear mapping relationship between frost heave-related variables and predicted freezing depth values ​​based on a random forest regression model. The random forest model is an ensemble learning method composed of multiple regression trees, exhibiting good anti-overfitting and feature selection capabilities. The model output is the predicted freezing depth (in meters) for each region.

[0072] The training module utilizes historical frozen depth data of existing regions for supervised learning. During training: the measured frozen depth is used as the supervision variable; an optimization problem is constructed with mean squared error (MSE) as the objective function; feature importance is ranked and splitting strategies are optimized for all tree nodes during training; the model performance control objective is to keep the average relative error below 10%.

[0073] The calculation of the predicted freezing depth includes:

[0074] For each basic backfill area, a multi-dimensional input vector containing meteorological and geological information is constructed, with the following dimensional structure:

[0075] Meteorological information: {Daily average temperature series average: -6.4℃, number of days of freezing: 72 days, maximum snowfall thickness: 24cm, maximum daily temperature fluctuation: ±10℃};

[0076] Geological information: {Soil type (unique thermal code): clayey soil, underground temperature gradient: -0.4℃ / m, moisture content: 18%, permafrost type: short seasonal permafrost (annual freezing time greater than 2 months and less than 4 months)};

[0077] The input vector is fed into a pre-trained random forest regression model, and each regression tree outputs a frozen depth sub-prediction value according to its partition path. The final frozen depth prediction value is the weighted average of the outputs of all trees.

[0078] The frost heave risk level refers to the possibility and degree of impact of frost heave deformation in the bridge abutment back area under existing geological and meteorological conditions. It is divided into three levels: high risk, medium risk, and low risk.

[0079] Judgment method: The predicted freezing depth is analyzed in conjunction with factors such as the design backfill thickness, groundwater level, and soil type; the classification rules are defined by empirical weights or expert systems. For example, freezing depth > 75% of backfill thickness and moisture content > 15% is defined as high risk; freezing depth between 50% and 75% of backfill thickness is medium risk; and freezing depth below 50% is low risk.

[0080] The calculation results of the frost heave risk level are written into the attribute fields of the corresponding foundation backfill area in the BIM model for subsequent structural simulation and material matching.

[0081] The data flow and processing path in this step is as follows: Geological and meteorological information is imported into the coupled prediction model through an interface; after feature construction, a standardized input vector is formed; the input vector is calculated by a random forest model to output the predicted freezing depth; compared with the design backfill layer thickness, the risk level is calculated; the risk level value is written into the attribute field of the corresponding area in the BIM model (e.g., Risk_Frost_Level=High); the output result will enter the next stage of structural response simulation to set boundary conditions and structural adaptation parameters.

[0082] Step S3: Based on the frost heave risk level of each foundation backfill area, and combined with the load conditions and foundation response characteristics in the design information, perform a simulation analysis of structural stress and deformation to generate the mechanical fit level of each foundation backfill area.

[0083] Frost heave risk level (input parameter): Calculated from step S2, used to set the frost heave boundary conditions in the simulation. The risk level affects boundary parameters such as freezing depth, frost heave displacement loading, and frost heave duration in the simulation.

[0084] Load conditions are structural load input parameters extracted from the design information, including: dead load (self-weight, fill weight); live load (vehicle load, traffic load); special loads (such as snow load, impact load); and load case combinations, which are the most unfavorable combinations input for simulation according to the design specifications.

[0085] Foundation response characteristics, constitutive relations and parameters of foundation soil, including: modulus (compression modulus, deformation modulus); Poisson's ratio; shear strength parameters (internal friction angle, cohesion); frost heave deformation modulus (used for material response simulation under freezing conditions); characteristic values ​​of foundation bearing capacity.

[0086] Mechanical compatibility level refers to the degree to which the backfill area of ​​the foundation adapts to the effects of frost heave under structural simulation, and is divided into three levels: good compatibility, moderate compatibility, and poor compatibility. The level assessment is based on the comparison results between structural response indicators and safety standards.

[0087] The simulation analysis of structural stress and deformation includes:

[0088] A coupled model of a multi-layer backfill structure foundation was constructed using the finite element method. Each foundation backfill area was modeled as an independent subdomain. The model included a rigid or semi-rigid superstructure (bridge abutment); multi-layer backfill (divided according to material type and compaction degree); and foundation soil (layered soil, saturation distribution). Equivalent parameters were used to convert the frost heave effect into initial strain or displacement loading.

[0089] Set boundary conditions, including frost heave boundary conditions and load boundary conditions;

[0090] Frost heave boundary conditions: freezing depth determined by the frost heave risk level; vertical frost heave displacement load (calculated based on frost heave coefficient × freezing depth); horizontal constraint boundary or slip boundary set according to site conditions.

[0091] Example: When a certain area is classified as high-risk, with a freezing depth of 1.2m and a frost heave coefficient of 2%, an initial upward frost heave displacement of 24mm is applied to the corresponding node in the simulation.

[0092] Load boundary conditions: Dead load is applied to the top surface of the abutment; live load is applied according to the vehicle load standard (such as Highway-I load) distributed at nodes; support conditions are set as fixed or elastic supports according to the foundation type (rigid foundation, pile foundation).

[0093] Through simulation calculations, the response of each foundation backfill area under frost heave displacement is as follows: vertical deformation, simulating the vertical displacement of the node under freezing loading conditions; equivalent stress distribution, the equivalent stress field in the soil, reflecting the structural adaptability; contact pressure response, the reaction force distribution at the bottom of the abutment or between backfill layers; safety margin factor, comparing the calculated value with the design allowable value to output whether the structure is in a safe state.

[0094] All metrics were compared with the safety thresholds defined by design specifications or the project.

[0095] For example, the following settings should be set: vertical differential settlement <15mm; contact stress not exceeding 60% of the soil compressive strength; maximum displacement difference between backfill layers <10mm.

[0096] Mechanical adaptability level definition: measures the structural and foundation response capability of a region under frost heave deformation, indicating whether it is suitable to maintain the existing structural function and stability under current design parameters.

[0097] The determination method is as follows: when all structural response indicators are less than the threshold and have a margin of >20%, it is determined to be well adapted; if some indicators are close to the limit (margin <10%), it is determined to be moderately adapted; if there are problems such as exceeding the threshold or significant differential settlement, it is considered poorly adapted.

[0098] The adaptation level is written as a label into the attribute field of the basic backfill area components in the BIM model to guide the matching of backfill materials and schemes in the next stage.

[0099] The data input and output path for this step is as follows: Input data includes the frost heave risk level from step S2; geometric and load parameters from step S1 and the design documents; and foundation response parameters from the geological survey. The processing involves constructing a finite element model and setting boundary conditions; performing simulations for each region and outputting multiple structural response indices; comparing with preset standards and outputting safety judgment results; output data includes the mechanical fit level (used for matching candidate schemes in S4); all structural response values ​​can be optionally written to BIM attribute fields or exported to a simulation report; the model results are visualized to present structural weak areas and areas with optimization potential.

[0100] Step S4: Based on the frost heave risk level and mechanical compatibility level, call the material performance database to generate multiple candidate backfill schemes for each basic backfill area. Each scheme includes backfill depth, backfill material combination and ratio, and outputs the corresponding material type, layer thickness and compaction parameters.

[0101] Candidate backfill schemes refer to a set of feasible backfill designs selected from a material performance database for a specific backfill area, based on its frost heave risk level and structural adaptability. Each scheme includes the following parameters: number of backfill layers and total depth; type of backfill material for each layer; material mix proportions (e.g., particle size distribution, cement content); layer thickness; and target compaction degree.

[0102] The materials performance database (data structure) stores the engineering physical properties and construction parameters of various backfill materials in a structured format. Key fields include (as shown in the table below):

[0103]

[0104] Data sources include laboratory material testing, accumulated engineering experience, standardized material manuals, and preliminary engineering measurement data.

[0105] The candidate solution generation logic and process include:

[0106] Obtain input variables, including the frost heave risk level from S2; the mechanical fit level from S3; and the regional spatial location, layered structure, and design load level from the BIM model.

[0107] Set screening and matching rules. For areas with high risk of frost heave and poor compatibility, prioritize lightweight materials with strong frost heave resistance (low frost heave coefficient) (such as foamed soil and lime-soil-gravel composite layers). For medium-risk areas with moderate compatibility, choose a combination scheme, such as a coarse-grained bottom layer and a fine-grained top layer. For low-risk areas with good compatibility, use conventional fillers, such as compacted loess or natural graded sand.

[0108] Several alternative solutions are generated using a rule-driven matching engine or rule base (such as an expert system); at least 2–3 technically feasible and economically differentiated combination solutions are generated for each region; the generated solution data is bound and written into the temporary attribute fields of the BIM components.

[0109] Example: The following candidate solutions are generated for a certain medium-risk area of ​​frost heave:

[0110] Option A: Top layer of lime-soil (15cm) + middle layer of lime-stabilized sand (25cm) + bottom layer of crushed stone (40cm); compaction degree 92%; Option B: Foamed lightweight soil integral backfill (1.0m); compaction degree 85%, higher unit cost but less deformation; Option C: Sand + geotextile layered combination, compaction degree 90%, good material synergy.

[0111] Parameter output description: The following technical parameters are output for each candidate scheme and written into the BIM model as data fields: material type, represented by the component attribute field for each layer; layer thickness (Layer_Thickness), in millimeters or centimeters; compaction degree (Compaction_Ratio), the target compaction ratio, such as 95%; backfill depth (Backfill_Depth), the total depth of the area; material combination number and scheme ID, used for simulation calls and version control; material performance score (optional), freeze sensitivity score, construction convenience score, etc.

[0112] The generation of candidate backfill schemes forms the following data flow: Input: geometric attributes such as area location and layer thickness in the BIM model; frost heave level and compatibility level output by S2 / S3; Processing: performing condition matching in the material property database; automatically combining and generating schemes based on the rule base; Output: generating 2-3 sets of candidate schemes for each area; writing all schemes into the BIM model attributes; data is synchronously output to the simulation platform as input for the next stage of performance evaluation.

[0113] Step S5: Simulate each candidate backfill scheme for each foundation backfill area, evaluate performance indicators, select the final backfill scheme for each foundation backfill area based on the preset optimization objective function, and adjust and merge the boundaries of the foundation backfill area in combination with the simulation results to generate an optimized construction area division scheme.

[0114] Simulation of candidate backfill schemes: For each scheme in each basic backfill area, a finite element model is constructed under actual frost heave and load, and its performance indicators such as settlement, stability and heat conduction are evaluated, and its engineering response is quantified.

[0115] Optimization objective function: Used to comprehensively evaluate the multi-dimensional performance indicators of multiple schemes, and to rank and select the results in a weighted manner. The objective function includes: settlement performance indicators (maximum settlement); stability indicators (slope safety factor); degree of influence of frozen boundary (heat conduction path); material compatibility and balance; and economic parameters (optional).

[0116] Boundary adjustment and regional merging and reorganization: This refers to redefining the backfill construction boundary based on the spatial continuity and material consistency of the performance of each region in the simulation results, so as to achieve construction continuity, process simplification and centralized material allocation.

[0117] The simulation includes:

[0118] Candidate backfill schemes corresponding to each basic backfill area are written into the temporary attribute fields of the corresponding components in the BIM model. The spatial location (regional boundary, coordinate location, layer thickness, from the BIM model), stratum conditions (soil layer distribution, foundation parameters, groundwater level, from geological survey data), boundary constraints (live load, dead load combination, frost heave loading boundary, from design information, S2 results) and material parameters (material type, compaction degree, physical parameters of each layer, from the material performance database) of the components are extracted to construct a structured simulation input dataset.

[0119] For example, candidate scheme B includes a combination of foamed lightweight soil upper layer and crushed stone base layer. Its material parameters (such as modulus of 30MPa and thermal conductivity of 0.42W / m·K) are provided by the material database. The corresponding backfill thickness is 80cm. The frost heave displacement is calculated by multiplying the freezing depth in step S2 by the frost heave coefficient and is applied to the top surface of the simulation model.

[0120] Using a finite element simulation tool (such as ABAQUS), perform simulations for the following three types of performance evaluation conditions:

[0121] Settlement response analysis: Simulate the vertical deformation behavior of the foundation backfill area after frost heave loading, and output the maximum settlement and uneven settlement difference;

[0122] Lateral stability analysis: used for slopes or open areas to calculate safety factors and slippage trends;

[0123] Freezing boundary conduction analysis: Evaluate the impact of different material combinations on temperature field propagation and identify whether frost heave concentration zones are likely to form. Each backfill scheme outputs corresponding simulation results, which are written to the model or exported in a structured format (such as JSON).

[0124] The optimized construction area division scheme includes:

[0125] Based on performance indicators, a multi-objective weighted optimization function is constructed with performance balance and material compatibility as weights to select the final backfill scheme from the candidate backfill schemes;

[0126] Performance balance refers to the specific requirements of the three types of performance evaluation results (such as the settlement priority project, which sets a larger settlement performance index).

[0127] Material coordination refers to the degree of consistency in material types, compaction standards, and processes among different backfill areas. Higher coordination leads to simpler construction organization and more efficient material allocation. Example: If both areas A and B use a double-layer structure of lime-soil + sand cushion, and both have a compaction degree of 92%, then they are considered to have high coordination; if the material types differ significantly, then the coordination is low.

[0128] For each candidate scheme in a region, the objective function value is calculated and optimized. The optimal scheme is selected as the final backfill scheme, written into the main attribute field of the BIM model, and used for the next construction scheme.

[0129] The optimized construction area division scheme includes:

[0130] The following factors are considered when adjusting regional boundaries: spatial continuity (adjacent areas have no structural boundaries and minimal geological differences); material uniformity (the final scheme uses similar material types and compaction parameters); and construction convenience (areas that do not affect the work sequence and workflow are prioritized for merging). Region merging methods include automatic merging (based on BIM attribute analysis of adjacent area similarity); semi-automatic adjustment (requiring engineer confirmation); and manual splitting (if simulation results reveal local anomalies, large areas can be reverse-divided). The final output is an optimized construction area division scheme, written into the component attribute fields of the BIM model, and used for subsequent construction simulation, scheduling, and resource allocation.

[0131] Step S5 data flow path description: Input: Candidate backfill scheme data from S4; BIM geometry and attribute model; Geological, structural and load information extracted from S1–S3; Processing: Batch generation of simulation models; Extraction of result indicators; Construction and calculation of optimization objective function; Output: Final backfill scheme for each area; Simulation scoring report; Optimized area division boundary (spatial data + attribute fields); Visualized map or IFC partitioning results.

[0132] Step S6: Generate an optimized construction plan that includes work sequence, work path and resource allocation based on the construction area division plan and the final backfill plan.

[0133] Optimized construction plan: This refers to a feasible and operable construction organization plan formulated based on the regional division results and final backfill design, comprehensively considering the coordination of construction sequence, equipment routes, material transportation, compaction technology, and resource organization. This plan is not the same as traditional construction drawings or processes, but rather a data-driven construction scheduling and planning result that highly relies on model-driven approaches. It has structured expression capabilities and supports construction simulation, progress simulation, and dynamic updates.

[0134] The data sources for optimizing the construction plan are mainly as follows: area division boundaries, from the area reorganization results (BIM) after simulation in step S5, used for determining work units; final backfill scheme parameters, from the final backfill scheme output in step S5, used for material and compaction control; work surface geometry, from the spatial location and hierarchical relationship of BIM components, used for path generation and equipment layout; material properties and loading / unloading requirements, from the material performance database, used for transportation and storage organization; and on-site resource list, from the engineering resource scheduling platform (interface with external systems), used for matching available equipment, manpower, and materials.

[0135] Work sequence refers to the arrangement of backfilling construction in different areas or layers during the construction process, taking into account structural safety, construction technology, material transportation, and the continuity of equipment paths. Work sequence planning principles: prioritize high-risk areas and areas adjacent to critical structures; proceed layer by layer from bottom to top based on elevation differences; if there is a compaction dependency (e.g., upper layer compaction requires a stable lower layer surface), enforce the established work sequence relationship. For example, work in area A (near the abutment backfill) must precede work in area B (outer edge free backfill area) to prevent construction vehicles from disturbing unstable areas. Output format: structured task schedule table (Gantt format or node dependency diagram); each area is assigned a construction priority field, written into the BIM model component attributes.

[0136] The work path is the dynamic path planning result of construction equipment (such as road rollers, transport vehicles, and vibratory compactors) on the construction site, emphasizing the accessibility, continuity, and minimum construction interference principles of the path. Path generation logic: A feasible path network is constructed based on the spatial coordinate information of the BIM model and the location of obstacles; turning radii and passage width constraints for construction equipment are introduced; the optimal work path is planned using A* search, Dijkstra's algorithm, or genetic algorithm (automatic / semi-automatic methods can be selected depending on site conditions); the round-trip paths of the work equipment, material transport vehicle routes, and the location of temporary roads are considered. Output format: Path data is written to the associated path field of a GIS layer or BIM component; the output is a path coordinate sequence, a 2D graphic, or an animated simulation trajectory.

[0137] Resource allocation refers to the organizational method of rationally distributing various construction resources (personnel, machinery, and materials) to different work areas and time periods to ensure efficiency and quality. Resource types include: human resources (construction teams, technical personnel); machinery (road rollers, loaders, water trucks, compaction equipment); material resources (various types of backfill materials, additives, and auxiliary materials); and auxiliary resources (water supply, power supply, lighting, access facilities, etc.). Scheduling principles include: matching with the work sequence to avoid resource conflicts or idleness; prioritizing merging supply batches when using the same materials in different areas; prioritizing highly qualified and experienced personnel in high-risk areas; and linking route and resource allocation: the longer the route, the greater the transportation capacity needs to be. Resource allocation output formats include: work unit-resource allocation table; daily construction plan (daily input of personnel, machinery, and materials); and marking of construction stage status and resource consumption attributes in the BIM model.

[0138] This step establishes the following data processing chain: Input phase: Import construction area boundary and final scheme data from step S5; obtain resource library and construction site layout map from project management system; Generation phase: Generate work sequence and path through scheduling algorithm and rule engine; call resource matching module to generate daily resource allocation plan; Output phase: Write construction scheme results into BIM model in attribute form; synchronously export task table and resource plan to construction management system.

[0139] Step S7: For bridges that have been built and have abutment anomalies, structural measurement information is obtained through 3D laser scanning and ground-penetrating radar. This information is used as the input for building the BIM model instead of the design information. Based on the integrity of the backfill structure, the distribution of voids, and the compaction state, the initial state information of the foundation backfill area is defined. The process of obtaining the corresponding frost heave risk level, mechanical compatibility level, and candidate backfill schemes takes into account the physical performance damage of the original backfill material and the uncertainty of the structural boundary. The simulation is compared between two types of schemes: replacement reconstruction or superimposed reinforcement. Corresponding repair backfill schemes and construction area adjustment schemes are generated.

[0140] Bridges with abutment anomalies: This refers to bridges that exhibit one of the following problems in the abutment area during the operational period: settlement difference exceeding the allowable value specified in the code; backfill voids or collapses behind the abutment wall; frost heave or approach slab settlement; or detection of localized structural loosening or insufficient compaction. Such structures lack effective original construction records or drawings; therefore, subsequent models must be constructed primarily based on on-site measured data.

[0141] Structural measurement information is obtained through on-site non-destructive testing and high-precision scanning methods, including: three-dimensional laser scanning (TLS) to obtain the surface morphology and settlement deformation of the abutment and backfill; ground-penetrating radar (GPR) to detect voids, thickness distribution and uneven compaction within the backfill layer; dynamic load response testing (optional) to evaluate the structure's response characteristics under load (such as dynamic deformation modulus); and sampling testing to obtain residual backfill material for mechanical / thermal property testing.

[0142] BIM model construction logic: Based on 3D laser point cloud as the geometric reconstruction foundation, the structural outline is automatically generated; the backfill layer interface is analyzed through ground-penetrating radar scanning maps and AI image recognition models; the detection results are attached to BIM components in the form of attributes to generate a digital abutment model with historical deformation and current status. Since the original design drawings may be missing or unreliable, measured data is used to replace the following fields: dimensions and boundary locations of the abutment back area; thickness and material type of each backfill layer (some require estimation); compaction status and integrity label of the backfill structure (e.g., presence of voids, insufficient compaction); actual deformation field distribution (settlement, heave, cracks, etc.).

[0143] Example: Radar detection revealed a 60cm deep and 1.2m long cavity 2m behind the abutment of an operating bridge, with the surrounding compaction degree below 85%. This area was marked as a local instability zone and prioritized for inclusion in subsequent repair simulations.

[0144] In the constructed BIM model, each foundation backfill area will have the following initial state information attribute fields attached:

[0145] Layer_Integrity_Status: Structural integrity status (intact / void / delamination);

[0146] Compaction_Estimate: Compaction estimate (derived from radar signal analysis or compaction testing);

[0147] Material_Damage_Index: Material damage index (estimated by laboratory testing or empirical model).

[0148] Freeze_History_Tag: Whether the site has experienced significant frost heave;

[0149] Uncertainty_Level: The level of uncertainty at the structural boundary (high / medium / low).

[0150] This information will serve as input correction factors for the S2 and S3 models, participating in the generation of the repair strategy.

[0151] In such existing abnormal bridges, since the original material condition is uncontrollable, the models in S2 and S3 need to be adjusted as follows:

[0152] Correction of damaged material parameters: correction of frost heave coefficient, adjusted according to the model of moisture content increase and microcrack evolution; modulus reduction factor, the elastic modulus of the material is reduced according to the empirical residual curve (e.g., 30~60%); thermal conductivity change, taking into account the difference between the void area and the dense area (heat conduction is blocked at the void).

[0153] Interval parameter modeling or Monte Carlo simulation methods are used to introduce boundary condition uncertainty zones: fuzzification of the formation interface (±10cm tolerance); instability of the load diffusion path; and transformation of the contact stress boundary into a flexible contact model.

[0154] Two repair strategies were designed for each unstable region, and simulations were conducted to compare their effects.

[0155] The simulated alternative reconstruction scheme (Scheme Type I) involves completely removing all existing materials in the abnormal areas; refilling with new high-frost-resistant materials (such as lightweight foamed soil and lime-soil stabilized gravel); compaction according to new construction standards; and setting the simulation to ideal reconstruction conditions to re-simulate the overall structural deformation and stress field.

[0156] The superimposed reinforcement scheme (Scheme Type II) does not completely remove the original material; an antifreeze and heat insulation layer (such as extruded polystyrene board + coarse-grained pad) is added above the abnormal area; local grouting or spraying is used to increase the contact density; the simulation process is set to a double-layer combination of raw material retention + new layer material overlay.

[0157] Simulation outputs and comparison indicators: frost heave deformation control capability; project cost estimation (which can be estimated using a unit cost model); construction period and operational feasibility score. The output results will be written into the repaired BIM model for subsequent construction decisions and on-site scheduling.

[0158] Example 2, an optimization system for bridge abutment backfill scheme, see [link / reference] Figure 1 As shown, it includes the following modules:

[0159] The data acquisition module is used to collect topographic, geological and meteorological information of the bridge abutment backfill area, as well as to obtain design information or measured structural information of the bridge abutment.

[0160] The modeling and partitioning module is used to construct a 3D model of the bridge abutment and its backfill area, automatically partition the foundation backfill area, and generate a BIM model of the bridge abutment backfill.

[0161] The frost heave prediction module, based on a meteorological and geological coupled prediction model, calculates the predicted freezing depth for each basic backfill area and outputs the frost heave risk level.

[0162] The mechanical analysis module is used to simulate and analyze the stress and deformation of the structure, and generate the mechanical fit level for each foundation backfill area.

[0163] The scheme generation module is used to generate candidate backfill schemes by calling the material performance database based on the frost heave risk level and mechanical compatibility level.

[0164] The simulation analysis module is used to simulate candidate backfill schemes and output performance indicators.

[0165] The optimization module is used to select the final backfill scheme based on a preset optimization objective function, reorganize the boundaries of the foundation backfill area, and output the construction area division scheme.

[0166] The scheduling generation module is used to generate optimized construction plans based on the construction area division plan and the final backfill plan;

[0167] The BIM management module is used to manage the optimization process of backfill schemes based on BIM models.

[0168] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An optimization method for bridge abutment backfilling scheme, characterized in that, include: Collect topographic and geological information of the bridge abutment backfill area; The design information of the bridge abutment is obtained, and a three-dimensional model of the abutment and abutment back area is constructed by combining the terrain information. The foundation backfill area is automatically divided, and a BIM model of the bridge abutment backfill is generated. The automatic division of the basic backfill area includes: Based on the 3D model of the bridge abutment and its back area, structural boundary features, slope ratio, elevation distribution and relative positional relationship with the bridge abutment are extracted. The 3D space is divided into multiple foundation backfill areas according to preset partitioning rules, including slope threshold, layer elevation, and structural support and non-support areas. The automatic partitioning results are written into the attribute fields of the corresponding components in the BIM model. For each basic backfill area in the BIM model, obtain meteorological information related to frost heave and input it into the meteorological-geological coupled prediction model. Combine the geological information to calculate the predicted value of freezing depth and output the corresponding frost heave risk level. The meteorological-geological coupled prediction model includes: The input module is used to receive meteorological and geological information corresponding to each basic backfill area; The feature construction module is used to normalize, expand time series data, and encode features from the input data. The prediction module, built on a random forest regression model, is used to establish a nonlinear mapping relationship between input data and predicted frozen depth values, and to calculate the predicted frozen depth values. The training module is used to perform supervised training of the prediction module using historical measured frozen depth data and output the optimal model weight parameters. The calculation of the predicted freezing depth includes: For each basic backfill area, a multi-dimensional input vector containing meteorological and geological information is constructed. The meteorological information includes the daily average surface temperature sequence, the number of days of freezing, the snowfall thickness, and the temperature fluctuation range. The geological information includes the underground temperature profile, soil type, soil moisture content, and regional permafrost distribution type. The multi-dimensional input vector is mapped to the trained random forest regression model through the prediction module. The predicted value of local freezing depth is output according to the partition path of each decision tree. The predicted value of freezing depth of the area is generated by the weighted average of all trees. Based on the frost heave risk level of each foundation backfill area, and combined with the load conditions and foundation response characteristics in the design information, a simulation analysis of structural stress and deformation is conducted to generate the mechanical fit level of each foundation backfill area. Based on the frost heave risk level and mechanical compatibility level, the material performance database is called to generate multiple candidate backfill schemes for each basic backfill area. Each scheme includes backfill depth, backfill material combination and ratio, and outputs the corresponding material type, layer thickness and compaction parameters. Simulation is performed on each candidate backfill scheme for each basic backfill area to evaluate performance indicators. Based on the preset optimization objective function, the final backfill scheme for each basic backfill area is selected. The results of the simulation are combined to adjust the boundaries of the basic backfill areas and merge and reorganize them to generate an optimized construction area division scheme. An optimized construction plan, including work sequence, work path, and resource allocation, is generated based on the construction area division scheme and the final backfilling scheme.

2. The method for optimizing bridge abutment backfill scheme according to claim 1, characterized in that, The simulation analysis of structural stress and deformation includes: A coupled model of a multi-layer backfill structure foundation was constructed using the finite element method. Based on the frost heave risk level, load conditions, and foundation response characteristics of each foundation backfill area, corresponding frost heave boundary conditions and load boundary conditions were set. The vertical deformation, equivalent stress distribution, and contact pressure response of the foundation backfill area under frost heave displacement were calculated through simulation. The response values ​​were compared with the preset safety threshold to determine the structural safety level range and generate the corresponding mechanical adaptation level.

3. The method for optimizing bridge abutment backfill scheme according to claim 1, characterized in that, The simulation includes: Candidate backfill schemes corresponding to each foundation backfill area are written into the temporary attribute fields of the corresponding components in the BIM model, and the spatial location, soil conditions, boundary constraints and material parameters of the components are extracted to construct a structured simulation input dataset. The simulation input dataset is imported into the finite element simulation platform, and settlement response analysis, lateral stability analysis and frozen boundary transmission analysis are performed on each candidate backfill scheme based on frost heave loading conditions and structural load conditions to obtain the response values ​​of the candidate backfill schemes under three types of performance indicators.

4. The method for optimizing bridge abutment backfill scheme according to claim 1, characterized in that, The optimized construction area division scheme includes: Based on performance indicators, a multi-objective weighted optimization function with performance balance and material coordination as weights is constructed to select the final backfill scheme from candidate backfill schemes. According to the final backfill scheme results of each foundation backfill area, its spatial continuity is analyzed. Based on the set area reorganization rules, the boundaries of adjacent areas are adjusted and merged and reorganized to output an optimized construction area division scheme that meets the requirements of construction continuity, material uniformity and risk controllability. This scheme is then written into the attribute fields of the corresponding components in the BIM model.

5. The method for optimizing bridge abutment backfill scheme according to claim 1, characterized in that, Also includes: For bridges that have been built and have abutment anomalies, the measured structural information is obtained by 3D laser scanning and ground-penetrating radar, which replaces the design information as the input for building the BIM model. The initial state information of the foundation backfill area is defined based on the integrity of the backfill structure, the distribution of voids and the compaction state. The process of obtaining the corresponding frost heave risk level, mechanical compatibility level and candidate backfill scheme takes into account the physical performance damage of the original backfill material and the uncertainty of the structural boundary. It compares and simulates two types of schemes: simulation replacement and reconstruction or superimposed reinforcement, and generates corresponding repair backfill schemes and construction area adjustment schemes.

6. A system for optimizing bridge abutment backfill schemes, characterized in that, The system applies any one of the bridge abutment backfill optimization methods described in claims 1 to 5, including: The data acquisition module is used to collect topographic, geological and meteorological information of the bridge abutment backfill area, as well as to obtain design information or measured structural information of the bridge abutment. The modeling and partitioning module is used to construct a 3D model of the bridge abutment and its backfill area, automatically partition the foundation backfill area, and generate a BIM model of the bridge abutment backfill. The frost heave prediction module, based on a meteorological and geological coupled prediction model, calculates the predicted freezing depth for each basic backfill area and outputs the frost heave risk level. The mechanical analysis module is used to simulate and analyze the stress and deformation of the structure, and generate the mechanical fit level for each foundation backfill area. The scheme generation module is used to generate candidate backfill schemes by calling the material performance database based on the frost heave risk level and mechanical compatibility level. The simulation analysis module is used to simulate candidate backfill schemes and output performance indicators. The optimization module is used to select the final backfill scheme based on a preset optimization objective function, reorganize the boundaries of the foundation backfill area, and output the construction area division scheme. The scheduling generation module is used to generate optimized construction plans based on the construction area division plan and the final backfill plan; The BIM management module is used to manage the optimization process of backfill schemes based on BIM models.

Images