A method for designing a pile plank bridge deck pavement

Abstract

This invention discloses a design method for pile-slab bridge deck pavement, belonging to the interdisciplinary field of engineering design and artificial intelligence. Specifically, it includes: data acquisition and preprocessing to establish a standardized design database; constructing a classification prediction model, completing model training and optimization; selecting core feature variables through intelligent feature importance analysis; calculating the pile-slab collaborative stress coefficient and predicting the equivalent elastic modulus of the pavement layer; calculating key parameters and iteratively optimizing the model; predicting and verifying the pavement layer thickness and material ratio parameters; calculating bond strength and predicting service life to ensure that design requirements are met; and completing final parameter optimization, outputting a final design scheme including core parameters, material ratios, and construction suggestions. This invention solves the problems of traditional design relying on experience, low accuracy, and difficulty in quantifying multi-parameter coupling, improving design efficiency and accuracy, extending the service life of the pavement layer, and is applicable to the pavement pavement design of various highway and municipal pile-slab bridges.

Description

Technical Field

[0001] This invention belongs to the field of intelligent technology application, involving the intersection of engineering design and artificial intelligence. Specifically, it relates to a design method for pile-slab bridge deck pavement, which integrates digital data processing technology to achieve intelligent prediction, accurate calculation and iterative optimization of pavement parameters. It is applicable to the pavement structure design of various highway and municipal pile-slab bridges, solving the technical problems of traditional design relying on experience, low accuracy, poor efficiency and difficulty in quantifying multi-parameter coupling, and promoting the development of bridge pavement design towards intelligence and precision. Background Technology

[0002] Pile-slab bridges, with their significant advantages such as lightweight structure, strong span adaptability, convenient construction, and good adaptability to complex terrain, are widely used in highway engineering, especially suitable for sections with complex terrain conditions such as mountainous areas and soft soil areas. As a key component of pile-slab bridges, the bridge deck pavement directly bears the load impact and shearing action of high-speed vehicles on highways, while also resisting the effects of natural environment (such as temperature changes, rain and snow erosion, and ultraviolet radiation). Its design rationality and structural stability directly determine the service life, driving safety, and comfort of the bridge.

[0003] Currently, the design of pile-slab bridge deck pavement still faces many unresolved issues. Existing design methods largely rely on engineers' experience and lack systematic data analysis and intelligent prediction support. They also fail to effectively integrate the core advantages of intelligent technologies, resulting in design efficiency and accuracy that cannot meet engineering requirements. Specifically, existing technologies have the following shortcomings: First, they lack classification prediction models, making it impossible to build prediction models using historical design data. This hinders accurate prediction of design parameters, leading to a disconnect between design parameters and actual engineering conditions, and increasing the risk of early problems such as pavement layer cracking, voids, and shoving. Second, they lack intelligent technology support, making it difficult to efficiently mine and analyze the characteristics of massive amounts of engineering, load, and environmental data, and to quantify the impact weight of multi-parameter coupling on pavement layer performance. Third, the design process has low automation; parameter calculation and optimization rely on manual operation, which is not only inefficient but also prone to human error, failing to meet the pavement layer design requirements under the complex conditions of highways. Fourth, the calculation formulas used in existing designs are relatively traditional and lack innovation, failing to accurately couple the influence of multiple factors, resulting in significant deviations between calculation results and actual engineering conditions.

[0004] With the rapid development of intelligent technology, its application in engineering design is becoming increasingly widespread. Intelligent technology enables efficient data processing and feature mining, and classification prediction models can be used to construct accurate prediction models, achieving intelligent prediction and optimization of design parameters. Therefore, there is an urgent need for a piling-slab bridge deck pavement design method that deeply integrates intelligent technology and possesses innovative calculation formulas. This method should fully leverage the advantages of intelligent technology, address the shortcomings of existing technologies, improve the accuracy and efficiency of pavement layer design, extend the service life of bridges, and reduce operation and maintenance costs. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies in pile-slab bridge deck pavement design, such as lack of integration of intelligent technology, low design accuracy, poor efficiency, difficulty in quantifying multi-parameter coupling, and lack of innovation in calculation formulas. This invention provides a pile-slab bridge deck pavement design method that integrates digital data processing technology to construct a full-process intelligent design system of "data acquisition - model training - parameter prediction - iterative optimization". This system enables accurate prediction and efficient optimization of pavement layer design parameters, improves the load-bearing capacity and service life of the pavement layer, reduces operation and maintenance costs, and adapts to the design needs of pile-slab bridge deck pavement under various complex working conditions.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A design method for pile-slab bridge deck pavement utilizes a computer system to complete data acquisition, model building, model training, parameter calculation, iterative optimization, and design output, achieving intelligent and precise design of pile-slab bridge deck pavement. The method specifically includes the following steps:

[0008] Step 1: Data acquisition and preprocessing. The computer data acquisition module is used to acquire the foundation parameters, load parameters, environmental parameters and historical design data of similar bridge pavement. Based on digital data processing technology, the acquired data is processed to establish a standardized design database, providing high-quality data support for model training.

[0009] Step 2: Construct a classification prediction model. Use the historical design data preprocessed in Step 1 as training samples, pavement layer thickness, material ratio, service life, and shear strength as target variables, and pile parameters, slab parameters, load parameters, and environmental parameters as input feature variables. Determine the node splitting rules and thresholds of the classification prediction model to complete the initial construction of the classification prediction model.

[0010] Step 3: Train and optimize the classification prediction model constructed in Step 2. Divide the data in the standardized design database into a training set and a test set in a 7:3 ratio. Iteratively optimize the depth, threshold, and weights of the classification prediction model using the training set, and verify the model accuracy using the test set to obtain the trained classification prediction model.

[0011] Step 4: Rank the input feature variables extracted in Step 1 by importance, and select the core feature variables with the top 80% weights that have an impact on the pavement layer design performance as the key input parameters for subsequent parameter calculations and model iterations;

[0012] Step 5: Based on the core feature variables selected in Step 4, and combined with digital data processing technology, calculate the pile-slab co-stress coefficient, input the calculation results into the trained classification prediction model, and obtain the preliminary predicted value of the equivalent elastic modulus of the pavement layer.

[0013] Step 6: Based on the preliminary predicted value of the equivalent elastic modulus output by the classification prediction model, calculate the critical shear strength, internal maximum stress and crack control coefficient of the pavement layer, and feed the calculation results back to the classification prediction model to perform the first iteration optimization of the model;

[0014] Step 7: Based on the iteratively optimized classification prediction model, predict the minimum design thickness and material ratio parameters of the pavement layer, and verify the rationality of the parameters by combining digital data processing technology. If the parameters do not meet the design requirements, adjust the feature weights of the classification prediction model and recalculate the parameters and iterate the model.

[0015] Step 8: Calculate the bond strength between the pavement layer and the pile plate structure, input the bond strength data into the classification prediction model, predict the service life of the pavement layer, and verify whether the service life meets the requirement of ≥20 years. If it does not meet the requirement, return to step 7 to readjust the parameters and iterate.

[0016] Step 9: Integrate the output results of the classification prediction model and the regression method, and perform final optimization on all design parameters of the pavement layer to ensure that all parameters meet the design standards and actual engineering needs;

[0017] Step 10: Output the final pavement layer design scheme, including core design parameters, material ratio scheme and construction reference suggestions, to complete the design of the pile-slab bridge deck pavement.

[0018] Further technology of the present invention:

[0019] Preferably, in step 2, the classification prediction model adopts a classification regression tree structure, and the node splitting rule is determined based on the information gain ratio. By traversing all input feature variables and splitting thresholds, the feature with the largest information gain ratio and the threshold are selected as the basis for node splitting.

[0020] Preferably, step 3 of the training process includes: initializing the classification prediction model parameters, calculating the prediction error of the training set, constructing a new classification prediction model branch based on the error, adjusting the model weights, iteratively training until the model prediction error reaches a preset threshold, and optimizing the classification prediction model structure.

[0021] Preferably, step 4 quantifies the influence of each input feature variable on the target variable by calculating the Gini coefficient of each input feature variable, and selects core feature variables with a Gini coefficient of less than 0.2 to ensure that the cumulative influence weight of the core feature variables on the design performance reaches more than 80%.

[0022] Preferably, the formula for calculating the pile-slab co-stress coefficient λ in step 5 is:

[0023] ;

[0024] in: The elastic modulus of the pile is expressed in MPa. Let m be the moment of inertia of the pile section. 4 ; (where the length of the pile is in meters). The elastic modulus of the plate is expressed in MPa. Let m be the moment of inertia of the plate section. 4 ; The length of the plate is in meters (m). The average density of the pile-slab structure is kg / m³. The acceleration due to gravity is expressed in m / s². The cross-sectional area of ​​the pile is in m². The cross-sectional area of ​​the plate is in m². The Poisson's ratio of the pile body; The Poisson's ratio of the plate; The maximum allowable stress of the pile body is expressed in MPa. The maximum allowable stress of the plate is expressed in MPa. This is the pile-slab connection coefficient, with a value ranging from 0.85 to 0.98. The value of the pile-slab coordinated displacement deviation ranges from 0.1mm to 0.5mm. The calculation process of this formula is realized through digital data processing technology, and the calculation result serves as the core input parameter of the classification prediction model.

[0025] Preferably, the critical shear strength of the pavement layer in step 6 Maximum internal stress and crack control coefficient The calculation formulas are as follows:

[0026] ;

[0027] ;

[0028] ;

[0029] in: The equivalent elastic modulus of the pavement layer is expressed in MPa. The density of the paving layer material is kN / m³. The standard value of the cubic compressive strength of the pavement concrete is given in MPa. The reinforcement ratio of the pavement layer is %; The coefficient of stress for pile-slab co-contraction. The maximum shear strength of the pavement material is given in MPa. Preliminary thickness estimate for the pavement layer, in meters; This is the shear correction factor, with a value ranging from 1 to 5; The maximum axle load of the vehicle is kN; This refers to the vehicle load impact coefficient. The width of the wheel touching the ground, in meters; The design thickness of the paving layer is in meters (m). The elastic modulus of the plate is expressed in MPa. The coefficient of thermal expansion of the pavement layer is given in °C. , These are the local annual highest and lowest temperatures, in °C; For the Poisson's ratio of the pavement layer, The shear modulus of the pavement layer, in MPa; This is the environmental stress correction value, ranging from 0.1 MPa to 0.3 MPa; The allowable stress of the pavement layer is expressed in MPa. The tensile strength of the paving material is expressed in MPa. This is the crack suppression coefficient, with a value ranging from 5 to 10.

[0030] Preferably, the minimum design thickness of the paving layer in step 7 is... and material proportioning parameters The calculation formulas are as follows:

[0031] ;

[0032] ;

[0033] in: The maximum axle load of the vehicle is kN; The wheelbase of the vehicle is in meters (m). The equivalent elastic modulus of the pavement layer is expressed in MPa. The coefficient of stress for pile-slab co-contraction. The elastic modulus of the plate is expressed in MPa. The width of the wheel touching the ground, in meters; The critical shear strength of the pavement layer is given in MPa. The tensile strength of the paving material is expressed in MPa. This refers to the vehicle load impact coefficient. The coefficient of friction between the vehicle tires and the pavement layer. This is a thickness correction amount, with a value ranging from 5mm to 15mm; The bond strength between the pavement layer and the pile plate is expressed in MPa. This is the crack control coefficient. The density of the paving layer material is kN / m³. The design thickness of the paving layer is in meters (m). , These are the local annual highest and lowest temperatures, in °C; The reinforcement ratio of the pavement layer is %; This is the material proportioning correction factor, with a value ranging from 1 to 3.

[0034] Preferably, the bond strength between the pavement layer and the pile-slab structure in step 8 is... and service life The calculation formulas are as follows:

[0035] ;

[0036] ;

[0037] in: This is the crack control coefficient. The equivalent elastic modulus of the pavement layer is expressed in MPa. The reinforcement ratio of the pavement layer is %; The design thickness of the paving layer is in meters (m). The coefficient of stress for pile-slab co-contraction. The standard value of the cubic compressive strength of the pavement concrete is given in MPa. The critical shear strength of the pavement layer is given in MPa. The coefficient of adhesion friction, The elastic modulus of the plate is expressed in MPa. This is a correction value for bond reinforcement, ranging from 0.1 MPa to 0.3 MPa. The maximum stress inside the pavement layer, in MPa; The local annual precipitation is in mm. This is the highest annual temperature in the area, in °C. The lifespan correction period ranges from 0.5 years to 2 years.

[0038] Preferably, in step 9, the intelligent model fuses weights. The calculation formula is:

[0039]

[0040] Where: model prediction accuracy is the test set accuracy of the classification prediction model, %; prediction error is the test set error of the classification prediction model, %; and the weights of the classification prediction model are: The weights of the regression method are .

[0041] Preferably, the standardized design database can be updated in real time, with the weight of newly added data. The calculation formula is:

[0042]

[0043] Wherein: the similarity of newly added data is the feature similarity between newly added design data and historical data, %; the average similarity of historical data is the average feature similarity of historical design data, %.

[0044] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0045] On the one hand, it is highly intelligent and calculates parameters accurately, deeply integrating digital data processing technology and intelligent models to replace traditional manual experience design, greatly improving design efficiency and shortening the design cycle of a single bridge pavement, thus increasing efficiency; on the other hand, it accurately couples multiple factors such as pile-slab synergistic stress, load impact, and environmental influence to achieve accurate calculation of each core parameter, improving accuracy compared to traditional design methods and effectively reducing human error.

[0046] On the other hand, it has strong adaptability and practicality. By using intelligent feature importance analysis to select core feature variables, and combining intelligent model fusion and multi-round iterative optimization, it ensures that all parameters such as pavement layer thickness and material ratio meet the design standards and actual engineering needs. This can effectively avoid early defects such as pavement layer cracking and delamination, and extend the service life to more than 20 years. The standardized design database can be updated in real time through intelligent incremental learning, and can adapt to different working conditions such as mountainous areas and soft soil areas. It is suitable for the design of various highway and municipal pile-slab bridges. The output design scheme includes core parameters, material ratios and construction reference suggestions. It is easy to operate and can be directly applied to actual engineering, promoting the development of bridge pavement design towards intelligence and precision. 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 specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0048] A pile-slab bridge on a mountainous expressway has a span of 30m. The piles are reinforced concrete cast-in-place piles, and the slab is prestressed concrete. The design speed is 100km / h, with an average daily traffic volume of 5000 vehicles, of which 35% are heavy vehicles. The local annual maximum temperature is 38℃, the annual minimum temperature is -5℃, and the annual precipitation is 1200mm. The roadbed is soft soil, and the pavement layer is required to have a service life of ≥20 years. The bridge deck pavement design is carried out using the method of this invention, and the specific steps are as follows:

[0049] 1. Data Acquisition and Preprocessing: Through sensors, engineering survey reports, and historical design archives of similar bridges, 30 basic parameters in four categories—piles, slabs, loads, and environment—were collected. Some core parameters are as follows: Pile elastic modulus. Moment of inertia of pile section Pile length Pile cross-sectional area Plate elastic modulus Moment of inertia of plate section Plate length plate cross-sectional area Maximum axle load of the vehicle Wheelbase Wheel ground contact width Vehicle load impact coefficient The highest annual temperature in the area The lowest temperature of the year Annual precipitation The collected data was cleaned, and two groups of samples with missing values ​​exceeding 10% were deleted. Three groups of samples with missing values ​​less than 10% were supplemented using intelligent interpolation. After normalization, three groups of abnormal data were removed. Twenty-five feature variables related to the pavement layer design performance were selected through feature extraction. Combined with 120 groups of historical design data of similar bridge pavements, a standardized design database was established.

[0050] 2. Classification Prediction Model Construction: Using 120 sets of preprocessed historical data as training samples, pavement layer thickness, material ratio, service life, and shear strength as target variables, and 25 feature variables as input feature variables, a classification prediction model was constructed using a classification regression tree structure. The node splitting rule was determined based on the information gain ratio. After traversing all feature variables and candidate thresholds, the pile-slab coordinated displacement deviation was selected as the root node splitting feature, with a splitting threshold of 0.3 mm. The model branches were constructed step by step by splitting. When the model depth reached 8 layers, the number of child node samples was ≥5 sets, and the information gain ratio was ≥0.01, the splitting stopped, and the initial construction was completed. The model contained 32 nodes and 63 branches.

[0051] 3. Model Training and Optimization: The 120 sets of data were divided into a training set (84 sets) and a test set (36 sets) in a 7:3 ratio. The model parameters were initialized as follows: learning rate 0.1, maximum number of iterations 100, and maximum model depth 10. The model was trained iteratively using the training set, and the sample prediction error was calculated and new branches were constructed. The model weights were adjusted using gradient descent. After 86 iterations, the average prediction error of the training set was reduced to 4.1%, which met the requirement of ≤5%. Six redundant branches were removed using pruning, resulting in a model with 26 nodes and 57 branches. The test set data was then input into the trained model. The prediction accuracy of the test set was 95.9%, and the prediction error was 4.1%, thus obtaining the trained classification prediction model.

[0052] 4. Core Feature Screening: The Gini coefficient method was used to calculate the Gini coefficient of 25 input feature variables. The variables were sorted by Gini coefficient from smallest to largest and their cumulative influence weight was calculated. Sixteen core feature variables with a Gini coefficient < 0.2 were selected, with a cumulative influence weight of 83.9%. The core feature variables include: pile elastic modulus, slab elastic modulus, pile-slab coordinated displacement deviation, maximum vehicle axle load, annual maximum temperature, annual minimum temperature, annual precipitation, standard value of cubic compressive strength of pavement concrete, reinforcement ratio of pavement steel, average density of pile-slab structure, pile Poisson's ratio, slab Poisson's ratio, vehicle load impact coefficient, wheel ground contact width, pavement material unit weight, and pavement temperature expansion coefficient.

[0053] 5. Calculation of pile-slab synergistic stress coefficient and prediction of equivalent elastic modulus: Parameters required for calculation are extracted from 16 core characteristic variables, including the average density of the pile-slab structure. Poisson's ratio of the pile Poisson's ratio of the plate Pile-slab coordinated displacement deviation Maximum allowable stress of pile body Maximum allowable stress of plate Pile-slab connection coefficient gravitational acceleration Substitute into the formula for calculating the pile-slab co-stress coefficient:

[0054] ;

[0055] Calculated using digital data processing technology ,Will The classification and prediction model, trained by inputting other core feature variables, outputs a preliminary predicted value of the equivalent elastic modulus of the pavement layer. .

[0056] 6. Calculation of key parameters of the pavement layer and first iteration of the model: based on Based on the core parameters: density of the paving layer material Standard value of cubic compressive strength of concrete in pavement layer reinforcement ratio of pavement layer steel bars Maximum shear strength of pavement material Preliminary estimated thickness of the paving layer Shear correction factor Pavement layer allowable stress Tensile strength of paving layer material Crack suppression coefficient The coefficient of thermal expansion of the pavement layer Poisson's ratio of the paving layer shear modulus of pavement layer Environmental stress correction value Substituting the three key parameters into the calculation formula, the result is obtained through intelligent regression:

[0057] ;

[0058] ;

[0059] ;

[0060] After comparison, all three key parameters met the preset standard values ​​( , , The calculation results are fed back to the classification prediction model, and the weights of the relevant features of the equivalent elastic modulus and the pile-slab co-stress coefficient are adjusted to complete the first model iteration. After the iteration, the prediction accuracy of the model test set is improved to 96.3%, and the prediction error is reduced to 3.7%.

[0061] 7. Prediction and Verification of Pavement Layer Thickness and Material Mix Ratio Parameters: The parameters calculated in step 6... , , Input the iterative classification prediction model, output the preliminary predicted value of the minimum design thickness of the pavement layer. Preliminary predicted values ​​of material proportioning parameters Combining the core parameter: the coefficient of friction between the vehicle tires and the pavement layer. Thickness correction amount The bond strength is tentatively based on the preliminary estimate. Material proportion correction factor Substitute the minimum design thickness of the pavement layer and the material ratio parameters into the calculation formula:

[0062] ;

[0063] ;

[0064] Parameter verification: Material proportioning parameters The values ​​are within a reasonable range (250-270), the parameters are acceptable, and no further iteration is needed.

[0065] 8. Bond strength calculation and service life prediction verification: Based on the results obtained in step 7 , And the relevant parameters obtained in step 6, combined with the core parameter: the coefficient of adhesion friction. Bond reinforcement correction value Life correction period Substitute into the formula for calculating bond strength and service life:

[0066] ;

[0067] ;

[0068] Lifetime verification: The design requirement of a predicted lifetime of 22.8 years ≥ 20 years is met. At the same time, dual verification is performed through digital data processing technology. The lifetime prediction error is 2.7% ≤ 3%, which meets the design requirements. There is no need to return to step 7 to adjust the parameters.

[0069] 9. Model Fusion and Final Parameter Optimization: Based on the accuracy (96.3%) and prediction error (3.7%) of the trained classification prediction model on the test set, substitute them into the fusion weight calculation formula:

[0070] ;

[0071] The classification prediction model has a weight of 0.75745, and the regression model has a weight of 1 - 0.75745 = 0.24255. By weighted fusion, the predicted parameter values ​​output by the classification prediction model and the actual values ​​calculated by the regression model are integrated to obtain the final design parameter: the equivalent elastic modulus of the pavement layer. Critical shear strength Maximum internal stress Crack control coefficient Minimum design thickness Material proportioning parameters Bond strength The service life is 22.8 years. Verification has shown that all parameters meet design standards and actual engineering conditions, and parameter optimization is complete.

[0072] 10. Design Scheme Output: Output the final design scheme for the pile-slab bridge deck pavement of this mountainous expressway. The core design parameters are as follows: equivalent elastic modulus of the pavement layer. Critical shear strength: 3.22 MPa; maximum internal stress: 11.48 MPa; crack control coefficient: 0.818; minimum design thickness: 0.13 m; material mix ratio: 262 (rounded to the nearest integer); steel reinforcement ratio: 0.8%; bond strength: 1.92 MPa; service life: 22.8 years. Material mix design: C40 concrete, cement content: 420 kg / m³; sand ratio: 38%; aggregate content: 1280 kg / m³; water-cement ratio: 0.45; admixture dosage: 1.2% of cement content. Construction recommendations: The pavement layer should be poured in layers, with a bottom layer thickness of 60 mm and a top layer thickness of 70 mm. After pouring, a membrane should be used for curing for at least 14 days. During construction, the flatness error of the pavement layer should be controlled to ≤3 mm / m. After construction, shear strength and bond strength tests should be conducted to ensure that the test results meet the design parameter requirements.

[0073] This embodiment completes the design of pile-slab bridge deck pavement using the method of the present invention. The design process is efficient and the parameters are accurate, which solves the problems of traditional design relying on experience and difficulty in quantifying multi-parameter coupling. The service life of the pavement layer meets the design requirements and is suitable for the working conditions of soft soil subgrade and high proportion of heavy vehicles on mountain highways, thus verifying the feasibility and practicality of the method of the present invention.

[0074] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. A method of designing a deck pavement for a pile plank bridge, characterized in that The intelligent and precise design of pile-slab bridge deck paving is achieved by using a computer system to complete data acquisition, model building, model training, parameter calculation, iterative optimization, and design output. Specifically, this includes the following steps: Step 1: Data acquisition and preprocessing. The computer data acquisition module is used to acquire the foundation parameters, load parameters, environmental parameters and historical design data of similar bridge pavement. Based on digital data processing technology, the acquired data is processed to establish a standardized design database, providing high-quality data support for model training. Step 2: Construct a classification prediction model. Use the historical design data preprocessed in Step 1 as training samples, pavement layer thickness, material ratio, service life, and shear strength as target variables, and pile parameters, slab parameters, load parameters, and environmental parameters as input feature variables. Determine the node splitting rules and thresholds of the classification prediction model to complete the initial construction of the classification prediction model. Step 3: Train and optimize the classification prediction model constructed in Step 2. Divide the data in the standardized design database into a training set and a test set in a 7:3 ratio. Iteratively optimize the depth, threshold, and weights of the classification prediction model using the training set, and verify the model accuracy using the test set to obtain the trained classification prediction model. Step 4: Rank the input feature variables extracted in Step 1 by importance, and select the core feature variables with the top 80% weights that have an impact on the pavement layer design performance as the key input parameters for subsequent parameter calculations and model iterations; Step 5: Based on the core feature variables selected in Step 4, and combined with digital data processing technology, calculate the pile-slab co-stress coefficient, input the calculation results into the trained classification prediction model, and obtain the preliminary predicted value of the equivalent elastic modulus of the pavement layer. Step 6: Based on the preliminary predicted value of the equivalent elastic modulus output by the classification prediction model, calculate the critical shear strength, internal maximum stress and crack control coefficient of the pavement layer, and feed the calculation results back to the classification prediction model to perform the first iteration optimization of the model; Step 7: Based on the iteratively optimized classification prediction model, predict the minimum design thickness and material ratio parameters of the pavement layer, and verify the rationality of the parameters by combining digital data processing technology. If the parameters do not meet the design requirements, adjust the feature weights of the classification prediction model and recalculate the parameters and iterate the model. Step 8: Calculate the bond strength between the pavement layer and the pile plate structure, input the bond strength data into the classification prediction model, predict the service life of the pavement layer, and verify whether the service life meets the requirement of ≥20 years. If it does not meet the requirement, return to step 7 to readjust the parameters and iterate. Step 9: Integrate the output results of the classification prediction model and the regression method, and perform final optimization on all design parameters of the pavement layer to ensure that all parameters meet the design standards and actual engineering needs; Step 10: Output the final pavement layer design scheme, including core design parameters, material ratio scheme and construction reference suggestions, to complete the design of the pile-slab bridge deck pavement.

2. A design method for deck pavement of a pile plank bridge according to claim 1, wherein In step 2, the classification prediction model adopts a classification regression tree structure. The node splitting rule is determined based on the information gain ratio. By traversing all input feature variables and splitting thresholds, the feature with the largest information gain ratio and the threshold are selected as the basis for node splitting.

3. The design method of a pile plank bridge pavement according to claim 1, characterized in that, Step 3, the training process, includes: initializing the classification prediction model parameters, calculating the prediction error of the training set, constructing a new classification prediction model branch based on the error, adjusting the model weights, iteratively training until the model prediction error reaches a preset threshold, and optimizing the classification prediction model structure.

4. The design method of a pile plank bridge pavement according to claim 1, characterized in that, Step 4 quantifies the influence of each input feature variable on the target variable by calculating the Gini coefficient of each input feature variable, and selects core feature variables with a Gini coefficient of less than 0.2 to ensure that the cumulative weight of the core feature variables on the design performance reaches more than 80%.

5. The design method of a pile plank bridge pavement according to claim 1, characterized in that, The formula for calculating the pile-slab co-stress coefficient λ in step 5 is: ； in: The elastic modulus of the pile, in MPa; Let m be the moment of inertia of the pile section. 4 ; (where the length of the pile is in meters). The elastic modulus of the plate is expressed in MPa. Let m be the moment of inertia of the plate section. 4 ; The length of the plate is in meters (m). The average density of the pile-slab structure is kg / m³. The acceleration due to gravity is expressed in m / s². The cross-sectional area of ​​the pile is in m². The cross-sectional area of ​​the plate is in m². The Poisson's ratio of the pile body; The Poisson's ratio of the plate; The maximum allowable stress of the pile body is expressed in MPa. The maximum allowable stress of the plate is expressed in MPa. This is the pile-slab connection coefficient, with a value ranging from 0.85 to 0.

98. The value of the pile-slab coordinated displacement deviation ranges from 0.1mm to 0.5mm. The calculation process of this formula is realized through digital data processing technology, and the calculation result serves as the core input parameter of the classification prediction model.

6. The design method for pile-slab bridge deck pavement according to claim 1, characterized in that, Critical shear strength of the pavement layer in step 6 Maximum internal stress and crack control coefficient The calculation formulas are as follows: ； ； ； in: The equivalent elastic modulus of the pavement layer is expressed in MPa. The density of the paving layer material is kN / m³. The standard value of the cubic compressive strength of the pavement concrete is given in MPa. The reinforcement ratio of the pavement layer is %; The coefficient of stress for pile-slab co-contraction. The maximum shear strength of the pavement material is given in MPa. Preliminary thickness estimate for the pavement layer, in meters; This is the shear correction factor, with a value ranging from 1 to 5; The maximum axle load of the vehicle is kN; This refers to the vehicle load impact coefficient. The width of the wheel touching the ground, in meters; The design thickness of the paving layer is in meters (m). The elastic modulus of the plate is expressed in MPa. The coefficient of thermal expansion of the pavement layer is given in °C. , These are the local annual highest and lowest temperatures, in °C; For the Poisson's ratio of the pavement layer, The shear modulus of the pavement layer, in MPa; This is the environmental stress correction value, ranging from 0.1 MPa to 0.3 MPa; The allowable stress of the pavement layer is expressed in MPa. The tensile strength of the paving material is expressed in MPa. This is the crack suppression coefficient, with a value ranging from 5 to 10.

7. The design method for pile-slab bridge deck pavement according to claim 1, characterized in that, Minimum design thickness of paving layer in step 7 and material proportioning parameters The calculation formulas are as follows: ； ； in: The maximum axle load of the vehicle is kN; The wheelbase of the vehicle is in meters (m). The equivalent elastic modulus of the pavement layer is expressed in MPa. The coefficient of stress for pile-slab co-contraction. The elastic modulus of the plate is expressed in MPa. The width of the wheel touching the ground, in meters; The critical shear strength of the pavement layer is given in MPa. The tensile strength of the paving material is expressed in MPa. This refers to the vehicle load impact coefficient. The coefficient of friction between the vehicle tires and the pavement layer. This is a thickness correction amount, with a value ranging from 5mm to 15mm; The bond strength between the pavement layer and the pile sheet is expressed in MPa. This is the crack control coefficient. The density of the paving layer material is kN / m³. The design thickness of the paving layer is in meters (m). , These are the local annual highest and lowest temperatures, in °C; The reinforcement ratio of the pavement layer is %; This is the material proportioning correction factor, with a value ranging from 1 to 3.

8. The design method for pile-slab bridge deck pavement according to claim 1, characterized in that, The bond strength between the pavement layer and the pile-slab structure in step 8 and service life The calculation formulas are as follows: ； ； in: This is the crack control coefficient. The equivalent elastic modulus of the pavement layer is expressed in MPa. The reinforcement ratio of the pavement layer is %; The design thickness of the paving layer is in meters (m). The coefficient of stress for pile-slab co-contraction. The standard value of the cubic compressive strength of the pavement concrete is given in MPa. The critical shear strength of the pavement layer is given in MPa. The coefficient of adhesion friction, The elastic modulus of the plate is expressed in MPa. This is a correction value for bond reinforcement, ranging from 0.1 MPa to 0.3 MPa. The maximum stress inside the pavement layer, in MPa; The local annual precipitation is in mm. This is the highest annual temperature in the area, in °C. The lifespan correction period ranges from 0.5 years to 2 years.

9. The design method for pile-slab bridge deck pavement according to claim 1, characterized in that, Step 9: Intelligent model fusion weights The calculation formula is: ； Where: model prediction accuracy is the test set accuracy of the classification prediction model, %; prediction error is the test set error of the classification prediction model, %; and the weights of the classification prediction model are: The weights of the regression method are .

10. The design method for pile-slab bridge deck pavement according to claim 1, characterized in that, The standardized design database can be updated in real time, with new data having corresponding weights. The calculation formula is: ； Wherein: the similarity of newly added data is the feature similarity between newly added design data and historical data, %; the average similarity of historical data is the average feature similarity of historical design data, %.

Images