A T-beam bridge splice joint stress prediction method and device, a terminal and a storage medium

Abstract

The application provides a T-beam bridge splicing joint stress prediction method and device, a terminal and a storage medium, and relates to the technical field of bridge stress prediction. The method comprises the following steps: acquiring the structural parameters of the upper part of a T-beam bridge, the T-beam bridge comprising a new T-beam bridge and an old T-beam bridge, the structural parameters comprising section parameters, splicing joint parameters, support axial stiffness parameters and the number of T-beam pieces; sampling the structural parameters by using a Latin hypercube sampling technology, and taking the sampling results as splicing joint bridge samples; obtaining the maximum transverse tensile stress at the splicing joint position according to the splicing joint bridge samples; constructing a splicing joint stress prediction model by using the maximum transverse tensile stress at the splicing joint position, and determining the splicing joint stress of a target T-beam bridge based on the splicing joint stress prediction model. The application can improve the prediction efficiency of the bridge splicing joint stress.

Description

Technical Field

[0001] This application relates to the field of bridge stress prediction technology, and in particular to a method, device, terminal and storage medium for predicting stress at splice joints of T-beam bridges. Background Technology

[0002] Uneven settlement between the old and new bridges is an unavoidable problem in bridge widening projects. This uneven settlement can significantly impact the overall structure of the widened bridge, especially at the joints. Therefore, predicting the stress state of these joints is crucial. However, current theoretical research in this area is insufficient. Due to the numerous boundary conditions and complex geometry of bridges, it is impossible to derive a reasonable explicit formula for quickly predicting the stress state of the joints. Summary of the Invention

[0003] This application provides a method, device, terminal, and storage medium for predicting the stress at the splice joint of a T-beam bridge, in order to solve the problem that predicting the stress at the splice joint of a bridge is difficult due to the numerous boundary conditions and complex geometry of the bridge.

[0004] Firstly, this application provides a method for predicting the stress at splice joints of T-beam bridges, including:

[0005] Obtain the structural parameters of the superstructure of the T-beam bridge, which includes new T-beam bridges and old T-beam bridges. The structural parameters include cross-sectional parameters, splice parameters, support axial stiffness parameters, and the number of T-beam segments.

[0006] The structural parameters were sampled using Latin hypercube sampling technique, and the sampling results were used as samples for widening bridges.

[0007] Based on the bridge widening sample, the maximum transverse tensile stress at the splice joint location is obtained;

[0008] Using the maximum transverse tensile stress at the splice joint location, a splice joint stress prediction model is constructed, and based on the splice joint stress prediction model, the splice joint stress of the target T-beam bridge is determined.

[0009] Secondly, this application provides a stress prediction device for splice joints of T-beam bridges, comprising:

[0010] The parameter acquisition module is used to acquire the structural parameters of the superstructure of the T-beam bridge, which includes new T-beam bridges and old T-beam bridges. The structural parameters include cross-sectional parameters, splice parameters, support axial stiffness parameters, and the number of T-beam segments.

[0011] The sampling module is used to sample the structural parameters using Latin hypercube sampling technology and use the sampling results as a sample of the bridge to be widened.

[0012] The stress calculation module is used to obtain the maximum transverse tensile stress at the splice joint location based on the spliced ​​bridge sample.

[0013] The stress prediction module is used to construct a splice joint stress prediction model using the maximum transverse tensile stress at the splice joint location, and to determine the splice joint stress of the target T-beam bridge based on the splice joint stress prediction model.

[0014] Thirdly, this application provides a terminal including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the method as described in the first aspect or any possible implementation of the first aspect above.

[0015] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method as described in the first aspect or any possible implementation of the first aspect.

[0016] This application provides a method, apparatus, terminal, and storage medium for predicting the stress at splice joints of T-beam bridges. The method involves acquiring structural parameters of the superstructure of the T-beam bridge (including both new and existing bridges), such as cross-sectional parameters, splice joint parameters, support axial stiffness parameters, and the number of T-beam segments. Latin hypercube sampling is used to sample these structural parameters, and the sampling results are used as samples of the widened bridges. Based on these widened bridge samples, the maximum lateral tensile stress at the splice joint location is obtained. Using this maximum lateral tensile stress, a splice joint stress prediction model is constructed, and based on this model, the splice joint stress of the target T-beam bridge is determined. This application utilizes Latin hypercube sampling to uniformly sample structural parameters, generating representative widened bridge samples and enriching the diversity of the dataset. By constructing a splice joint stress prediction model, the prediction efficiency of bridge splice joint stress can be improved. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating the implementation of the stress prediction method for T-beam bridge splice joints provided in the embodiments of this application.

[0019] Figure 2 This is a schematic diagram of the structural parameters of the superstructure of a T-beam bridge provided in an embodiment of this application;

[0020] Figure 3 This is a schematic diagram of the structure of the stress prediction device for T-beam bridge splice joints provided in the embodiments of this application;

[0021] Figure 4 This is a schematic diagram of the terminal provided in the embodiments of this application. Detailed Implementation

[0022] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0023] To make the objectives, technical solutions, and advantages of this application clearer, the following description will be provided in conjunction with the accompanying drawings and specific embodiments.

[0024] Due to the numerous boundary conditions and complex geometry of bridges, it is impossible to derive a reasonable explicit formula to quickly predict the stress state of the splice joints at the supports of a T-beam bridge under the settlement load of a new bridge. With the development of machine learning, its powerful nonlinear fitting capability provides a reasonable method to solve this problem. Based on this, this application provides a machine learning-based method for predicting the stress state of splice joints in T-beam bridges. The technical concept is as follows: determine the structural parameters of the superstructure of the T-beam bridge; generate spliced ​​bridge samples through Latin hypercube sampling; construct a finite element model based on the generated spliced ​​bridge samples and apply settlement load to the new bridge; extract the maximum tensile stress at the splice joint location and set a tensile stress threshold to generate a machine learning database; process the generated machine learning database using Borderline-SMOTE data balancing technology to generate a uniform and fast evaluation database; train the generated database using the XGBoost algorithm and optimize the hyperparameters using the GA-RIME algorithm to obtain the XGBoost model with optimal hyperparameters; and predict the stress state of the bridge splice joints based on the XGBoost model with optimal hyperparameters.

[0025] Figure 1 The implementation flowchart of the stress prediction method for T-beam bridge splice joints provided in this application embodiment is described in detail below:

[0026] In step 101, the structural parameters of the superstructure of the T-beam bridge are obtained. The T-beam bridge includes new T-beam bridges and old T-beam bridges. The structural parameters include cross-sectional parameters, splice parameters, support axial stiffness parameters, and the number of T-beam segments.

[0027] In this embodiment of the application, when building the splice joint stress prediction model, it is necessary to obtain the structural parameters of the superstructure of the T-beam bridge as training samples. The T-beam bridge includes both new and old T-beam bridges, and the structural parameters include cross-sectional parameters, splice joint parameters, support axial stiffness parameters, and the number of T-beam segments.

[0028] The cross-sectional parameters include the cross-sectional parameters of the old T-beam bridge at the support and the cross-sectional parameters of the new T-beam bridge. In bridge widening, the same cross-sectional form is generally used. Considering that increasing the cross-sectional form will increase the data dimension and require higher computational cost, the embodiments of this application only consider the general case. The cross-sectional parameters of the T-beam bridge specifically include six cross-sectional parameters: web height, web width, flange width, flange thickness, chamfer width, and chamfer height.

[0029] The splice joint parameters include the splice joint thickness and the splice joint length. For the widened structure of the T-beam bridge with only wet joint connections on the superstructure, the splice joint thickness is equal to the flange thickness. Since a portion of the flange plate needs to be removed in the splice joint area, the flange width parameter for the joint portion also needs to be included. Both the joint concrete material and the bridge concrete material are C50 concrete.

[0030] Regarding the bearing axial stiffness parameter, under the action of settlement load on the new bridge bearing, the bearing axial stiffness has the greatest impact on the joint stress, while the stiffness in other directions has an extremely small impact. For example, when a new bridge is spliced ​​with an old bridge, if the bearing axial stiffness of the old bridge is small, applying a settlement load to the new bridge may cause unevenness at the splice point between the new and old bridges. Therefore, the structural parameters in the embodiments of this application should include the bearing axial stiffness parameter, and the bearing axial stiffness parameter is the bearing axial stiffness parameter of the old bridge.

[0031] In this embodiment of the application, the structural parameters also include the number of T-beam segments, namely the number of T-beam segments of the new bridge and the number of T-beam segments of the old bridge.

[0032] For the specific structural parameters and their corresponding locations, please refer to... Figure 2 As shown, the chamfer is represented by its width and height, and both the width and height of the chamfer are [value missing]. Figure 2 The chamfer should point to either the front or back; one refers to the width and the other to the height. Additionally, the flange width of the seam is the same as the seam thickness.

[0033] This application embodiment can also define numerical ranges for the above-mentioned structural parameters. These ranges can be divided according to actual engineering parameters, and the size of the range determines the applicability range of the subsequently constructed splice joint stress prediction model. Preferably, the parameter ranges are divided according to actual engineering projects, including collecting a large number of engineering drawings to obtain the range of each structural parameter.

[0034] For example, by collecting a large number of engineering drawings, the following parameters were determined: web height, web width, flange width, flange thickness, chamfer width, chamfer height, joint length, joint thickness (i.e., flange width at the joint), support axial stiffness parameters, and the range of the number of T-beams for new and old bridges. The specific ranges are shown in the table below:

[0035] parameter interval parameter interval Web height [a1,b1] Web width [a2,b2] wing width [a3,b3] Edge thickness [a4,b4] chamfer width [a5,b5] Chamfer height [a6,b6] seam length [a7,b7] seam thickness [a8,b8] Axial stiffness parameters of support [a9,b9] Number of T-beams on the new bridge [a10,b10] Number of T-beams in the old bridge [a11,b11]

[0036] In step 102, the structural parameters are sampled using Latin hypercube sampling technique, and the sampling results are used as samples for widening the bridge.

[0037] Latin hypercube sampling (LHS) is a method for approximate random sampling from a multivariate parameter distribution. It belongs to stratified sampling techniques and is often used in computer experiments or Monte Carlo integration.

[0038] In this embodiment of the application, the structural parameters of the superstructure of the T-beam bridge obtained in step 101 are uniformly sampled using Latin hypercube sampling technology, and the sampling results after uniform sampling are used as samples for widening the bridge.

[0039] In one possible implementation, the structural parameters are sampled using Latin hypercube sampling, and the sampling results are used as samples for widening the bridge. This can include:

[0040] Using the Latin hypercube sampling technique, the range of values ​​for each structural parameter is divided into a preset number of equal-width intervals, and the width of the corresponding structural parameter in each equal-width interval is obtained.

[0041] By using the range of values ​​for each structural parameter and the width of the corresponding structural parameter in each equal-width interval, the random value of the corresponding structural parameter in each equal-width interval is calculated.

[0042] For each structural parameter, a random value is randomly selected from all random values ​​as the target random value, and the target random values ​​corresponding to each structural parameter are combined into a bridge width-splitting sample.

[0043] Optionally, for each structural parameter, obtain the range of values ​​for each structural parameter [a i b i The number of samples N to be generated for each structural parameter is determined, i.e., a preset number of equal-width intervals of N. Then, using Latin hypercube sampling, the parameter value range of each structural parameter is divided into equal-width intervals of the preset number N, and the width of each structural parameter in each equal-width interval is calculated, i.e.:

[0044] For each structural parameter, find the maximum value b within its range.i and the minimum value of parameter a i And by inputting the preset quantity N into the first formula, the width ΔX of each structural parameter in each equal-width interval is obtained. i The first formula is:

[0045]

[0046] Where, ΔX i Let b be the width of the i-th structural parameter in each equal-width interval. i Let a be the maximum value of the i-th structural parameter. i Let N be the minimum value of the i-th structural parameter, and N be the preset number.

[0047] The width ΔX of each structural parameter in each equal-width interval is obtained. i Then, using the parameter value range of each structural parameter and the width ΔX of that structural parameter in each equal-width interval... i The corresponding structural parameters are calculated to take random values ​​within each equal-width interval, i.e.:

[0048] Find the minimum value 'a' of each structural parameter. i And the corresponding structural parameters, including the width ΔX in each equal-width interval. i The input into the second formula calculates the random values ​​X of the corresponding structural parameters within each equal-width interval. i,j The second formula is:

[0049] X i,j =a i +(j-1)·ΔX i +u i ·ΔX i

[0050] Among them, X i,j Let a be a random value of the i-th structural parameter within the j-th equal-width interval. o Let ΔX be the minimum value of the i-th structural parameter, j be the number of the equal-width interval (i.e., 1 to N), and ΔX be the minimum value of the i-th structural parameter. i Let u be the width of the i-th structural parameter in each equal-width interval. i A random number that is not less than 0 and less than 1.

[0051] Then, for each structural parameter, a random value is randomly selected from all random values ​​of that structural parameter as the target random value of that structural parameter, and the target random values ​​of each structural parameter are combined into a bridge width-splitting sample.

[0052] For example, given N1 structural parameters, and N2 samples for each structural parameter, each structural parameter generates N2 random values. For each structural parameter, a random value x1 is randomly selected from these N2 random values ​​for the current sampling count as the target random value for that current sampling count. Then, the target random values ​​corresponding to the N1 structural parameters for the current sampling count are combined to form the bridge width-spanning sample for that current sampling count.

[0053] It should be noted that the number of draws is N2, and the target random value drawn each time does not enter the subsequent draw process.

[0054] This application introduces Latin hypercube sampling technology to uniformly sample bridge structural parameters, generating representative widened bridge samples and enriching the diversity of the dataset.

[0055] In step 103, the maximum transverse tensile stress at the splice joint is obtained based on the sample of the widened bridge.

[0056] In this embodiment of the application, the maximum lateral stress at the splice joint of each spliced ​​bridge sample is obtained.

[0057] In one possible implementation, obtaining the maximum transverse tensile stress at the splice joint location based on a sample of bridges with widened sections can include:

[0058] Based on each bridge width-splitting sample, a corresponding finite element model is built;

[0059] For each finite element model, a preset settlement load is applied to each new T-beam bridge in the finite element model to obtain the maximum transverse tensile stress at the splice joint position corresponding to the finite element model.

[0060] Optionally, each bridge widening sample calculated in step 102 is used to perform corresponding finite element modeling, resulting in a finite element model for each bridge widening sample. Then, for each finite element model, a preset settlement load is applied to each new T-beam bridge in the finite element model, and the preset settlement load applied to each new T-beam bridge is uniform. For example, for each finite element model, solid elements are used for modeling, and a 5mm displacement load is applied to the new bridge in the model.

[0061] After applying a preset settlement load to the new bridge in each finite element model, the maximum transverse tensile stress at the splice joint in the corresponding finite element model is extracted.

[0062] In step 104, the maximum transverse tensile stress at the splice joint location is used to construct a splice joint stress prediction model, and based on the splice joint stress prediction model, the splice joint stress of the target T-beam bridge is determined.

[0063] In this embodiment of the application, the maximum transverse tensile stress at the splice joint location obtained in step 103 is used to construct a splice joint stress prediction model, and then the splice joint stress condition of the target T-beam bridge is predicted based on the splice joint stress prediction model.

[0064] In one possible implementation, a stress prediction model for the splice joint is constructed using the maximum transverse tensile stress at the splice joint location, which may include:

[0065] For each maximum transverse tensile stress, determine whether the maximum transverse tensile stress exceeds the tensile stress threshold. If it does, the maximum transverse tensile stress exceeding the tensile stress threshold is determined as the first sample. If it does not exceed the threshold, the maximum transverse tensile stress not exceeding the tensile stress threshold is determined as the second sample.

[0066] Construct a classification database from all first samples and all second samples;

[0067] The Borderline-SMOTE data balancing technique is used to process the categorical database and generate the target database.

[0068] The XGBoost algorithm was used to construct a splice joint stress prediction model for the target database.

[0069] Optionally, after obtaining the maximum transverse tensile stress at the splice joint location corresponding to each bridge width sample, it is necessary to determine the label of each maximum transverse tensile stress, that is, to set a tensile stress threshold. In this embodiment, the tensile stress threshold is set as the standard value of the tensile strength of C50 concrete. Then, it is determined whether each maximum transverse tensile stress exceeds the tensile stress threshold. All maximum transverse tensile stresses that exceed the tensile stress threshold are recorded as the first sample, and all maximum transverse tensile stresses that do not exceed the tensile stress threshold are recorded as the second sample. The first sample and the second sample form a new classification database, denoted as M.

[0070] Then, since the newly generated classification database M may have a small number of classes with a low proportion, that is, the number of maximum transverse tensile stresses that do not exceed or exceed the tensile stress threshold is small, this embodiment introduces the Borderline-SMOTE data balancing technology to generate a balanced dataset. That is, for the newly generated classification database M, the Borderline-SMOTE data balancing technology is used to process the classification database M to generate a uniform and fast evaluation database, which is denoted as the target database.

[0071] The specific steps of the Borderline-SMOTE data balancing technique are as follows:

[0072] Calculate the K nearest neighbors for each minority class sample. Let the training set be S = (x... p ,y p ), where xp It is a feature sample, y p ∈{0,1} is the class label, where 1 represents the minority class and 0 represents the majority class. For each minority class sample x q Calculate its 10 nearest neighbor samples, denoted as x. r r = 1, 2, ..., 10.

[0073] The nearest neighbor determination formula uses Euclidean distance, and the data is normalized to (0, 1) before calculating the nearest neighbors, i.e.:

[0074]

[0075] Where t is a characteristic, t = 1, 2, ..., 11.

[0076] Determine sample features x q Calculate the number of majority class samples n0 and minority class samples n1 in its k nearest neighbors.

[0077] Determine sample type:

[0078] Based on the ratio of majority class samples n0 to minority class samples n1, the sample features x q They are classified into three types:

[0079] Safe sample: n0 = 0 or n0 / 10 < 0.25.

[0080] Boundary sample: 0.25≤n0 / 10≤0.75.

[0081] Noise sample: n0 / 10≥0.75.

[0082] Randomly select sample features x from all boundary samples. q and its minority class nearest neighbors x r Generate new samples, that is:

[0083] x n =x q +λ·(x r -x q )

[0084] Where λ is a random number that follows a uniform distribution (0, 1).

[0085] Repeat the above steps until the generated minority class samples equal the majority class samples, and denote the generated new sample set as R.

[0086] Finally, the XGBoost algorithm is used to train the target database to obtain the splice joint stress prediction model. For example, all data in the target database can be divided into a training set and a test set, with a ratio of 8:2, or the division can be adjusted according to the actual situation.

[0087] The embodiments of this application employ Borderline-SMOTE data balancing technology, which solves the problem of insufficient minority class samples in the dataset and improves the predictive performance of the machine learning model.

[0088] In one possible implementation, the XGBoost algorithm is used to construct a splice joint stress prediction model for the target database, which may include:

[0089] Construct an XGBoost model using the XGBoost algorithm;

[0090] The GA-RIME algorithm is used to optimize the hyperparameters of the XGBoost model, and the XGBoost model with the optimal hyperparameters is obtained.

[0091] A splice seam stress prediction model was obtained by training the target database using an XGBoost model with optimal hyperparameters.

[0092] Optionally, after training the splice joint stress prediction model using the XGBoost algorithm, the XGBoost model is first constructed using the XGBoost algorithm, then the hyperparameters of the XGBoost model are optimized using the GA-RIME algorithm to obtain the XGBoost model with the optimal hyperparameters, and finally, the XGBoost model with the optimal hyperparameters is used to train the target database to obtain the splice joint stress prediction model.

[0093] The steps for obtaining the optimal hyperparameter combination using the GA-RIME algorithm are as follows:

[0094] Step 1: Parameter settings.

[0095] Define a fitness function; for class-balanced samples, use the error rate as the fitness function:

[0096]

[0097] Where F is the fitness function.

[0098] TP (True Positive): The number of samples that are both true and predicted to be positive.

[0099] TN (True Negative): The number of samples where the actual value is negative and the prediction is also negative.

[0100] FP (False Positive): The number of samples where the actual value is negative but the prediction is positive.

[0101] FN (False Negative): The number of samples that are actually positive but predicted to be negative.

[0102] Set upper and lower bounds Ub and Lb for each hyperparameter. The specific hyperparameters and ranges are as follows:

[0103] n_estimators[100, 600];

[0104] max_depth[2, 6];

[0105] learning_rate[0.01, 1];

[0106] subsample[0.8, 1];

[0107] colsample_bytree[0.5, 1];

[0108] min_child_weight[1, 10];

[0109] reg_lambda(L2)[0,1].

[0110] Set the number of iterations T = 300, where t is the current iteration number.

[0111] Hyperparameter dimension: The number of hyperparameters, Dim = 13.

[0112] The population size is n.

[0113] Step 2: Initialize population parameters.

[0114]

[0115] Where R is the population, i is the individual code of the population, i = 1, 2, ..., n, and j is the hyperparameter dimension.

[0116] The population individuals are randomly generated within the hyperparameter space, and the individual R with the lowest fitness value is found. best and minimum fitness value F(R) best ).

[0117] Enter the loop process.

[0118] Step 3: Ion position update, which is performed for each ion.

[0119] Step 3.1: If r2 is less than E, update the position of each particle using the soft frost search strategy, i.e.:

[0120]

[0121] in, For the updated particle position, R best,jLet r1 be the j-th particle of the best individual in population R, and r1 be a random number between (-1, 1) and h be a random number between (0, 1). w=5, Ub ij and Lb ij These represent the range of values ​​for the hyperparameter ion. r2 is a random number between (0, 1).

[0122] Step 3.2: If r3 is less than F normr (S i ), to execute the hard frost perforation mechanism, that is:

[0123] r3 <F normr (S i )

[0124] Where r3 is a random number between (-1, 1), F normr (S i ) represents the normalized value of the fitness value of the current individual.

[0125] Step 3.3: Implement an active greedy selection mechanism:

[0126] if F is the fitness function, then

[0127] if but

[0128] This eventually leads to the formation of a new population R.

[0129] Step 3.4: Execute the genetic algorithm mechanism.

[0130] For individuals in the population after the greedy selection mechanism, perform selection, crossover, and mutation operations.

[0131] Step 3.4.1 Parent Selection: Based on the normalized probability P of the fitness function F, select a parent R from the current population. P .

[0132] Choose probability calculation:

[0133] For each individual R i Calculate its selection probability P i ,in, To give individuals with lower fitness a higher selection probability, f is used. i =1-F is used as the standard.

[0134] Calculate the cumulative probability:

[0135]

[0136] Parent selection: For i = (1, n), generate a random number r i ~U(0,1).

[0137] If Q z-1 <γ i z If the z-th individual is selected as the i-th parent, then the total number of parents is n, and they are also numbered using the denoting 'i'.

[0138] Step 3.4.2 Cross operation:

[0139] Parents are randomly paired up in pairs, resulting in a total of n / 2 pairs, where n is an even number. There are p parent pairs. 2m-1 and p 2m , where m = (1, n / 2).

[0140] For each parent generation, two child generations C are generated using simulated binary crossover. 2m-1 and C 2m , where m = (1, n / 2).

[0141] Take the crossover probability P c =0.8, first generate a random number r c ~(0,1), if r c >P c If the child generation equals the parent generation, then the child generation equals the parent generation; otherwise, perform the following operations:

[0142] For parameter dimension j, generate random numbers u ~ U(0,1) and calculate:

[0143]

[0144] Where, η c =2.

[0145] Generate offspring parameter values:

[0146] C 2m-1,j =0.5[(1+β q )P 2m-1,j +(1-β q )P 2m,j ]

[0147] C 2m,j =0.5[(1-β q )P 2m-1,j +(1+β q )P 2m,j ]

[0148] After pairing, a total of n offspring are generated.

[0149] Step 3.4.3 Mutation operation:​

[0150] Take the mutation probability P m =0.1, generate random number r m ~U(0,1).

[0151] If r m ≤P m If the mutation operation is successful, then the mutation operation is performed; otherwise, the variable remains unchanged.

[0152] The specific steps are as follows:

[0153] Generate random numbers r ~ U(0,1), and calculate the variance:

[0154]

[0155] Update parameter C i,j =C i,j +δ j (U j -L j ), where U j and L j These are the upper and lower bounds of the hyperparameters.

[0156] Step 3.4.4 Merging Selection Strategy:

[0157] The resulting offspring population C and population R are merged. The fitness function of each individual in the population is calculated. The top n individuals with the lowest fitness values ​​are selected to form a new population R. The individual with the lowest fitness value in the current population is then updated. best and minimum fitness value F(R) best ).

[0158] Repeat the above steps until the iteration count t = T, at which point the loop ends.

[0159] The individual with the lowest fitness value is ultimately identified as the optimal hyperparameter combination. The XGBoost model with this optimal hyperparameter combination is then used to train the target database to obtain a splice seam stress prediction model.

[0160] The embodiments of this application use the GA-RIME algorithm to optimize the hyperparameters of the XGBoost model, avoiding the inefficiency of manual optimization and improving the prediction accuracy of bridge splice joint stress.

[0161] This application provides a method for predicting the stress at splice joints of T-beam bridges. The method involves acquiring the structural parameters of the superstructure of the T-beam bridge, including both new and existing bridges. These parameters include cross-sectional parameters, splice joint parameters, support axial stiffness parameters, and the number of T-beam segments. Latin hypercube sampling is used to sample these structural parameters, and the sampling results are used as samples of the widened bridges. Based on these widened bridge samples, the maximum lateral tensile stress at the splice joint location is obtained. Using this maximum lateral tensile stress, a splice joint stress prediction model is constructed, and based on this model, the splice joint stress of the target T-beam bridge is determined. This application utilizes Latin hypercube sampling to uniformly sample structural parameters, generating representative widened bridge samples and enriching the diversity of the dataset. By constructing a splice joint stress prediction model, the prediction efficiency of bridge splice joint stress can be improved.

[0162] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0163] The following are device embodiments of this application. For details not described in detail, please refer to the corresponding method embodiments described above.

[0164] Figure 3 A schematic diagram of the stress prediction device for T-beam bridge splice joints provided in this application embodiment is shown. For ease of explanation, only the parts related to this application embodiment are shown, and are described in detail below:

[0165] like Figure 3 As shown, the stress prediction device 3 for T-beam bridge splice joints includes:

[0166] The parameter acquisition module 31 is used to acquire the structural parameters of the superstructure of the T-beam bridge. The T-beam bridge includes new T-beam bridges and old T-beam bridges. The structural parameters include section parameters, splice parameters, support axial stiffness parameters and number of T-beam segments.

[0167] Sampling module 32 is used to sample structural parameters using Latin hypercube sampling technology and use the sampling results as a sample of the bridge to be widened.

[0168] The stress calculation module 33 is used to obtain the maximum transverse tensile stress at the splice joint location based on the spliced ​​bridge sample;

[0169] The stress prediction module 34 is used to construct a splice joint stress prediction model by utilizing the maximum transverse tensile stress at the splice joint location, and to determine the splice joint stress of the target T-beam bridge based on the splice joint stress prediction model.

[0170] This application provides a device for predicting the stress at the splice joints of T-beam bridges. The device acquires structural parameters of the superstructure of the T-beam bridge, including both new and old bridges. These parameters include cross-sectional parameters, splice joint parameters, support axial stiffness parameters, and the number of T-beam segments. Latin hypercube sampling is used to sample these structural parameters, and the sampling results are used as samples of the widened bridges. Based on these widened bridge samples, the maximum lateral tensile stress at the splice joint location is obtained. Using this maximum lateral tensile stress, a splice joint stress prediction model is constructed, and based on this model, the splice joint stress of the target T-beam bridge is determined. This application utilizes Latin hypercube sampling to uniformly sample structural parameters, generating representative widened bridge samples and enriching the diversity of the dataset. By constructing a splice joint stress prediction model, the prediction efficiency of bridge splice joint stress can be improved.

[0171] In one possible implementation, the sampling module can be used for:

[0172] Using the Latin hypercube sampling technique, the range of values ​​for each structural parameter is divided into a preset number of equal-width intervals, and the width of the corresponding structural parameter in each equal-width interval is obtained.

[0173] By using the range of values ​​for each structural parameter and the width of the corresponding structural parameter in each equal-width interval, the random value of the corresponding structural parameter in each equal-width interval is calculated.

[0174] For each structural parameter, a random value is randomly selected from all random values ​​as the target random value, and the target random values ​​corresponding to each structural parameter are combined into a bridge width-splitting sample.

[0175] In one possible implementation, the sampling module can also be used for:

[0176] Input the maximum and minimum values ​​of each structural parameter within its range, along with a preset quantity, into the first formula to obtain the width of each structural parameter in each equal-width interval. The first formula is:

[0177]

[0178] Where, ΔX i Let b be the width of the i-th structural parameter in each equal-width interval. i Let a be the maximum value of the i-th structural parameter. i Let N be the minimum value of the i-th structural parameter, and N be the preset number.

[0179] In one possible implementation, the sampling module can also be used for:

[0180] For each structural parameter, the minimum value of that parameter and the width of that parameter in each equal-width interval are input into the second formula to calculate a random value for that structural parameter within each equal-width interval. The second formula is:

[0181] X i,j =a i +(j-1)·ΔX i +u i ·ΔX i

[0182] Among them, X i,j Let a be a random value of the i-th structural parameter within the j-th equal-width interval. i Let ΔX be the minimum value of the i-th structural parameter, j be the number of the equal-width interval, and ΔX be the minimum value of the i-th structural parameter. i Let u be the width of the i-th structural parameter in each equal-width interval. i A random number that is not less than 0 and less than 1.

[0183] In one possible implementation, the stress calculation module can be used for:

[0184] Based on each bridge width-splitting sample, a corresponding finite element model is built;

[0185] For each finite element model, a preset settlement load is applied to each new T-beam bridge in the finite element model to obtain the maximum transverse tensile stress at the splice joint position corresponding to the finite element model.

[0186] In one possible implementation, the stress prediction module can be used for:

[0187] For each maximum transverse tensile stress, determine whether the maximum transverse tensile stress exceeds the tensile stress threshold. If it does, the maximum transverse tensile stress exceeding the tensile stress threshold is determined as the first sample. If it does not exceed the threshold, the maximum transverse tensile stress not exceeding the tensile stress threshold is determined as the second sample.

[0188] Construct a classification database from all first samples and all second samples;

[0189] The Borderline-SMOTE data balancing technique is used to process the categorical database and generate the target database.

[0190] The XGBoost algorithm was used to construct a splice joint stress prediction model for the target database.

[0191] In one possible implementation, the stress prediction module can be used for:

[0192] Construct an XGBoost model using the XGBoost algorithm;

[0193] The GA-RIME algorithm is used to optimize the hyperparameters of the XGBoost model, and the XGBoost model with the optimal hyperparameters is obtained.

[0194] The XGBoost model with optimal hyperparameters was trained on the target database to obtain a splice seam stress prediction model.

[0195] Figure 4 This is a schematic diagram of the terminal provided in an embodiment of this application. For example... Figure 4 As shown, the terminal 4 in this embodiment includes: a processor 40, a memory 41, and a computer program 42 stored in the memory 41 and executable on the processor 40. When the processor 40 executes the computer program 42, it implements the steps in the above embodiments of the T-beam bridge splice joint stress prediction method, for example... Figure 1 Steps 101 to 104 are shown. Alternatively, when the processor 40 executes the computer program 42, it implements the functions of each module / unit in the above-described device embodiments, for example... Figure 3 The functions of each module are shown.

[0196] For example, the computer program 42 can be divided into one or more modules / units, which are stored in the memory 41 and executed by the processor 40 to complete this application. The one or more modules / units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program 42 in the terminal 4. For example, the computer program 42 can be divided into... Figure 3 The modules shown.

[0197] The terminal 4 can be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. The terminal 4 may include, but is not limited to, a processor 40 and a memory 41. Those skilled in the art will understand that... Figure 4 This is merely an example of terminal 4 and does not constitute a limitation on terminal 4. It may include more or fewer components than shown, or combine certain components, or different components. For example, the terminal may also include input / output devices, network access devices, buses, etc.

[0198] The processor 40 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0199] The memory 41 can be an internal storage unit of the terminal 4, such as a hard disk or memory of the terminal 4. The memory 41 can also be an external storage device of the terminal 4, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card equipped on the terminal 4. Furthermore, the memory 41 can include both internal storage units and external storage devices of the terminal 4. The memory 41 is used to store the computer program and other programs and data required by the terminal. The memory 41 can also be used to temporarily store data that has been output or will be output.

[0200] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0201] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0202] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0203] In the embodiments provided in this application, it should be understood that the disclosed devices / terminals and methods can be implemented in other ways. For example, the device / terminal embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0204] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0205] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0206] If the integrated module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above-described embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various T-beam bridge splice joint stress prediction method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately added to or subtracted from the content as required by the legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium may not include electrical carrier signals and telecommunication signals.

[0207] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for predicting stress at splice joints in T-beam bridges, characterized in that, include: Obtain the structural parameters of the superstructure of the T-beam bridge, which includes new T-beam bridges and old T-beam bridges. The structural parameters include cross-sectional parameters, splice parameters, support axial stiffness parameters, and the number of T-beam segments. The structural parameters were sampled using Latin hypercube sampling technique, and the sampling results were used as samples for widening bridges. Based on the bridge widening sample, the maximum transverse tensile stress at the splice joint location is obtained; Using the maximum transverse tensile stress at the splice joint location, a splice joint stress prediction model is constructed, and based on the splice joint stress prediction model, the splice joint stress of the target T-beam bridge is determined. The method of sampling the structural parameters using Latin hypercube sampling and using the sampling results as a sample for widening the bridge includes: Using the Latin hypercube sampling technique, the range of values ​​for each structural parameter is divided into a preset number of equal-width intervals, and the width of the corresponding structural parameter in each equal-width interval is obtained. By using the range of values ​​for each structural parameter and the width of the corresponding structural parameter in each equal-width interval, the random value of the corresponding structural parameter in each equal-width interval is calculated. For each structural parameter, a random value is randomly selected from all random values ​​as the target random value, and the target random values ​​corresponding to each structural parameter are combined into a bridge width-splitting sample. The step of obtaining the maximum transverse tensile stress at the splice joint location based on the bridge width sample includes: Based on each bridge width-splitting sample, a corresponding finite element model is built; For each finite element model, a preset settlement load is applied to each new T-beam bridge in the finite element model to obtain the maximum transverse tensile stress at the splice joint position corresponding to the finite element model. The step of constructing a splice joint stress prediction model using the maximum transverse tensile stress at the splice joint location includes: For each maximum transverse tensile stress, determine whether the maximum transverse tensile stress exceeds the tensile stress threshold. If it does, the maximum transverse tensile stress exceeding the tensile stress threshold is determined as the first sample. If it does not exceed the tensile stress threshold, the maximum transverse tensile stress not exceeding the tensile stress threshold is determined as the second sample. Construct a classification database from all first samples and all second samples; The classification database is processed using Borderline-SMOTE data balancing technology to generate the target database; The XGBoost algorithm is used to construct the splice joint stress prediction model for the target database.

2. The method for predicting stress at splice joints of T-beam bridges according to claim 1, characterized in that, The step of obtaining the width of the corresponding structural parameters in each equal-width interval includes: The maximum and minimum values ​​of each structural parameter within its range, along with the preset quantity, are input into the first formula to obtain the width of each structural parameter in each equal-width interval. The first formula is: in, For the first Each structural parameter represents the width of each equal-width interval. For the first The maximum value of each structural parameter. For the first The minimum value of each structural parameter. The preset quantity.

3. The method for predicting stress at splice joints of T-beam bridges according to claim 1, characterized in that, The step of calculating the random value of each structural parameter within each equal-width interval by utilizing the parameter value range of each structural parameter and the width of the corresponding structural parameter in each equal-width interval includes: For each structural parameter, the minimum value of that parameter and the width of that parameter in each equal-width interval are input into the second formula to calculate a random value for that structural parameter within each equal-width interval. The second formula is: in, For the first The structural parameter is at the first Random values ​​within equal-width intervals For the first The minimum value of each structural parameter. Numbering of equal-width intervals, For the first Each structural parameter represents the width of each equal-width interval. A random number that is not less than 0 and less than 1.

4. The method for predicting stress at splice joints of T-beam bridges according to claim 1, characterized in that, The step of constructing the splice joint stress prediction model using the XGBoost algorithm on the target database includes: Construct an XGBoost model using the XGBoost algorithm; The hyperparameters of the XGBoost model are optimized using the GA-RIME algorithm to obtain an XGBoost model with optimal hyperparameters. The target database is trained using an XGBoost model with optimal hyperparameters to obtain the splice seam stress prediction model.

5. A device for predicting stress at splice joints of T-beam bridges, characterized in that, include: The parameter acquisition module is used to acquire the structural parameters of the superstructure of the T-beam bridge, which includes new T-beam bridges and old T-beam bridges. The structural parameters include cross-sectional parameters, splice parameters, support axial stiffness parameters, and the number of T-beam segments. The sampling module is used to sample the structural parameters using Latin hypercube sampling technology and use the sampling results as a sample of the bridge to be widened. The stress calculation module is used to obtain the maximum transverse tensile stress at the splice joint location based on the spliced ​​bridge sample. The stress prediction module is used to construct a splice joint stress prediction model using the maximum transverse tensile stress at the splice joint location, and to determine the splice joint stress of the target T-beam bridge based on the splice joint stress prediction model. The sampling module is used for: Using the Latin hypercube sampling technique, the range of values ​​for each structural parameter is divided into a preset number of equal-width intervals, and the width of the corresponding structural parameter in each equal-width interval is obtained. By using the range of values ​​for each structural parameter and the width of the corresponding structural parameter in each equal-width interval, the random value of the corresponding structural parameter in each equal-width interval is calculated. For each structural parameter, a random value is randomly selected from all random values ​​as the target random value, and the target random values ​​corresponding to each structural parameter are combined into a bridge width-splitting sample. The stress calculation module is used for: Based on each bridge width-splitting sample, a corresponding finite element model is built; For each finite element model, a preset settlement load is applied to each new T-beam bridge in the finite element model to obtain the maximum transverse tensile stress at the splice joint position corresponding to the finite element model. The stress prediction module is used for: For each maximum transverse tensile stress, determine whether the maximum transverse tensile stress exceeds the tensile stress threshold. If it does, the maximum transverse tensile stress exceeding the tensile stress threshold is determined as the first sample. If it does not exceed the tensile stress threshold, the maximum transverse tensile stress not exceeding the tensile stress threshold is determined as the second sample. Construct a classification database from all first samples and all second samples; The classification database is processed using Borderline-SMOTE data balancing technology to generate the target database; The XGBoost algorithm is used to construct the splice joint stress prediction model for the target database.

6. A terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the stress prediction method for splice joints of T-beam bridges as described in any one of claims 1 to 4.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the stress prediction method for splice joints of T-beam bridges as described in any one of claims 1 to 4.

Images