Concrete interface shear strength prediction method, device and equipment and storage medium

Abstract

The invention is suitable for the technical field of structural design, and provides a concrete interface shear strength prediction method, device and equipment and a storage medium, and the method comprises the following steps: obtaining the shear strength and concrete parameters of a concrete interface; establishing an initial LightGBM model, recording the initial LightGBM model as a first model, and preprocessing shear strength and concrete parameters to obtain a training set and a test set; based on a collaborative group optimization algorithm and a genetic algorithm, optimizing hyper-parameters of the first model to obtain a target hyper-parameter set; applying the target hyper-parameter set to the first model to obtain a second model, and training the second model based on the training set and the test set to obtain a shear strength prediction model; and inputting the concrete parameters of the target concrete interface into the shear strength prediction model to obtain a shear strength prediction value of the target concrete interface. According to the method, the shear strength of the concrete interface can be accurately predicted under the condition of complex working conditions.

Description

Technical Field

[0001] The present application belongs to the field of structural design technology, and in particular relates to a method, device, equipment and storage medium for predicting the shear strength of a concrete interface. Background Art

[0002] With the widespread application of prefabricated buildings, bridges and other structures, as well as the reinforcement and renovation of existing buildings, the problem of predicting the shear strength of the interface between new and old concrete has become an important topic in the field of civil engineering. Existing prediction methods usually rely on empirical formulas or simplified theoretical models. However, due to the complexity of the working conditions and the nonlinear relationship between multidimensional parameters, the accuracy of existing methods is often limited. In recent years, the development of machine learning technology has provided new possibilities for solving this problem. Machine learning can handle multiple parameters and complex working conditions more effectively by automatically learning complex nonlinear relationships from large amounts of data, and has been widely used in the field of civil engineering.

[0003] Among various machine learning models, LightGBM (Light Gradient Boosting Machine) has become a powerful tool for predicting the shear strength of the interface between new and old concrete due to its efficient learning mechanism and excellent generalization ability. LightGBM can capture the nonlinear characteristics in the data through the combination of multiple weak decision trees, and shows high prediction accuracy in regression problems. However, the hyperparameters of the LightGBM model have a significant impact on the prediction performance. If the hyperparameters are not selected properly, the model performance may be degraded. Although common grid search and random search methods can optimize hyperparameters to a certain extent, they have problems such as low efficiency and high computational overhead, and cannot fully adapt to the optimization requirements of multidimensional parameters in complex engineering problems. Therefore, traditional methods have many limitations under complex working conditions and nonlinear relationships between multidimensional parameters. Summary of the invention

[0004] The embodiments of the present application provide a method, device, equipment and storage medium for predicting the shear strength of a concrete interface to improve the prediction performance of the LightGBM model in predicting shear strength, and can accurately predict the shear strength of a concrete interface under complex working conditions.

[0005] This application is implemented through the following technical solutions:

[0006] In a first aspect, an embodiment of the present application provides a method for predicting the shear strength of a concrete interface, comprising:

[0007] Obtain the shear strength of the concrete interface and concrete parameters.

[0008] An initial LightGBM model is established, recorded as the first model, and the shear strength and concrete parameters are preprocessed to obtain the training set and test set.

[0009] Based on the collaborative group optimization algorithm and the genetic algorithm, the hyperparameters of the first model are optimized to obtain the target hyperparameter set; the target hyperparameter set is applied to the first model to obtain the second model, and the second model is trained based on the training set and the test set to obtain the shear strength prediction model.

[0010] The concrete parameters of the target concrete interface are input into the shear strength prediction model to obtain the shear strength prediction value of the target concrete interface.

[0011] In combination with the first aspect, in some possible implementations, based on a collaborative group optimization algorithm and a genetic algorithm, the hyperparameters of the first model are optimized to obtain a target hyperparameter set, including:

[0012] Initialize the parameters of the collaborative swarm optimization algorithm, the parameters of the genetic algorithm and the population parameters, and obtain the number and upper and lower bounds of the hyperparameters of the first model; wherein the fitness function of the collaborative swarm optimization algorithm is the negative mean square error of the first model on the training set.

[0013] Based on the number and upper and lower bounds of the hyperparameters of the first model, the initial particle positions in the collaborative swarm optimization algorithm are determined to obtain the initial population.

[0014] For each iterative population, the particles in the population are iterated based on the collaborative swarm optimization algorithm to obtain updated particles.

[0015] Based on the fitness value of the best particle among the updated particles, the selection probability of the genetic algorithm is determined.

[0016] Based on the selection probability, the parent particles of the genetic algorithm are determined.

[0017] The parent particles are iterated based on the genetic algorithm to obtain the child particles.

[0018] The offspring particles and the updated particles are combined to form a new population, which is recorded as the first population.

[0019] The particles in the first population are sorted according to the fitness values, and a preset number of particles are selected as the second population.

[0020] The global optimal position is updated according to the particle with the smallest fitness value in the second population, and when the current number of iterations does not reach the maximum number of iterations, the second population is used as the population for the next iteration.

[0021] After the current number of iterations reaches the maximum number of iterations, the iteration is stopped and the global optimal position at this time is used as the target hyperparameter set.

[0022] In conjunction with the first aspect, in some possible implementations, determining the selection probability of the genetic algorithm based on the fitness value of the best particle among the updated particles includes:

[0023] Based on the fitness value of the best particle among the updated particles, combined with the first formula, the selection probability of the genetic algorithm is determined.

[0024] The first formula is:

[0025]

[0026] Among them, Pi represents the selection probability, N represents the number of particles, and f j represents the fitness value of particle j, and f represents the fitness value of the optimal particle.

[0027] In conjunction with the first aspect, in some possible implementations, initializing parameters of the collaborative group optimization algorithm, parameters of the genetic algorithm, and population parameters includes:

[0028] Initialize particle velocity, personal best position, personal best fitness value, global best position, global best fitness value, population size, maximum number of iterations, initial inertia weight, inertia weight reduction rate, acceleration coefficient, crossover probability, and mutation probability.

[0029] In combination with the first aspect, in some possible implementations, training a second model based on a training set and a test set to obtain a shear strength prediction model includes:

[0030] The second model is trained based on the training set to obtain a trained model.

[0031] The accuracy of the trained model was checked based on the test set, mean squared error, and mean absolute error.

[0032] If the accuracy of the trained model meets the preset conditions, the trained model will be used as the shear strength prediction model.

[0033] In combination with the first aspect, in some possible implementations, shear strength and concrete parameters are preprocessed to obtain a training set and a test set, including:

[0034] Determine missing data in shear strength and concrete parameters.

[0035] The missing data are completed based on the K nearest neighbor algorithm.

[0036] Based on the first model, the importance score of each type of concrete parameter in the completed concrete parameters is calculated.

[0037] For each type of concrete parameter, if the importance score is greater than a preset threshold, the data of the concrete parameter of this type is retained; if the importance score is less than or equal to the preset threshold, the data of the concrete parameter of this type is deleted to obtain a processed data set.

[0038] The processed data set is divided into training set and test set according to the preset ratio.

[0039] In combination with the first aspect, in some possible implementations, based on the first model, calculating the importance score of each type of concrete parameter in the completed concrete parameters includes:

[0040] Based on the first model and in combination with the second formula, the importance score of each type of concrete parameter in the completed concrete parameters is calculated.

[0041] The second formula is:

[0042]

[0043] Among them, I j represents the importance score of concrete parameters of type j, T represents the total number of decision trees in the first model, S t,j represents the set of all split nodes in decision tree t using concrete parameters of type j, Gain S represents the information gain of node s.

[0044] In a second aspect, an embodiment of the present application provides a device for predicting the shear strength of a concrete interface, comprising:

[0045] The data acquisition module is used to obtain the shear strength of the concrete interface and concrete parameters.

[0046] The data processing module is used to establish the initial LightGBM model, recorded as the first model, and pre-process the shear strength and concrete parameters to obtain the training set and the test set.

[0047] The first modeling module is used to optimize the hyperparameters of the first model based on a collaborative group optimization algorithm and a genetic algorithm to obtain a target hyperparameter set; apply the target hyperparameter set to the first model to obtain a second model, and train the second model based on a training set and a test set to obtain a shear strength prediction model.

[0048] The result prediction module is used to input the concrete parameters of the target concrete interface into the shear strength prediction model to obtain the shear strength prediction value of the target concrete interface.

[0049] In a third aspect, an embodiment of the present application provides a terminal device, comprising: a processor and a memory, the memory being used to store a computer program, and when the processor executes the computer program, the method for predicting the shear strength of a concrete interface as described in any one of the first aspects is implemented.

[0050] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method for predicting the shear strength of a concrete interface as described in any one of the first aspects is implemented.

[0051] It can be understood that the beneficial effects of the second to fourth aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here.

[0052] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0053] This application uses the Social Spider Optimization Algorithm (SSOA) for global search. The Social Spider Optimization Algorithm SSOA can efficiently explore in a wide range of parameter spaces and find preliminary high-quality hyperparameter combinations by simulating group collaboration and information exchange between individuals. Subsequently, the Genetic Algorithm (GA) takes over the solution obtained by the Social Spider Optimization Algorithm SSOA, further optimizes the hyperparameters through crossover and mutation operations, enhances the diversity of the population, and promotes a more comprehensive exploration of the solution space. The hybrid optimization strategy combining the Social Spider Optimization Algorithm SSOA and the Genetic Algorithm GA not only speeds up the convergence speed of hyperparameter optimization and significantly shortens the model training time, but also effectively avoids the problem that a single optimization algorithm is prone to falling into local optimality, improves the prediction performance of the LightGBM model, and ensures accurate prediction of the shear strength of the concrete interface under complex working conditions.

[0054] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present specification. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0056] Figure 1 It is a flow chart of a method for predicting the shear strength of a concrete interface provided in one embodiment of the present application;

[0057] Figure 2 It is a structural schematic diagram of a concrete interface shear strength prediction device provided by an embodiment of the present application;

[0058] Figure 3 It is a structural diagram of a terminal device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0059] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0060] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0061] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0062] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0063] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0064] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0065] The present application provides a method for predicting the shear strength of concrete interfaces. Figure 1 is a schematic diagram of the process of predicting the shear strength of concrete interface provided by an embodiment of the present application, with reference to Figure 1 , the concrete interface shear strength prediction method is described as follows:

[0066] Step 101, obtaining the shear strength of the concrete interface and concrete parameters.

[0067] Specifically, the concrete parameters may include: compressive strength of the first poured concrete, compressive strength of the second poured concrete, material type of the second poured concrete, length along the shear surface, shear surface area, interface treatment method, interface roughness, interface stress, anchoring ratio, anchoring depth, anchoring yield strength and anchoring diameter, etc.

[0068] Step 102, establish an initial LightGBM model, recorded as the first model, and preprocess the shear strength and concrete parameters to obtain a training set and a test set.

[0069] Exemplarily, preprocessing the shear strength and concrete parameters to obtain a training set and a test set may include:

[0070] Determine missing data in shear strength and concrete parameters.

[0071] The missing data are completed based on the K nearest neighbor algorithm.

[0072] Based on the first model, the importance score of each type of concrete parameter in the completed concrete parameters is calculated.

[0073] For each type of concrete parameter, if the importance score is greater than a preset threshold, the data of the concrete parameter of this type is retained; if the importance score is less than or equal to the preset threshold, the data of the concrete parameter of this type is deleted to obtain a processed data set.

[0074] The processed data set is divided into training set and test set according to the preset ratio.

[0075] Specifically, the preset threshold value may be 0.1, and the preset ratio of the training set to the test set may be 8:2. In some specific embodiments, the importance scores corresponding to the two types of concrete parameters, namely, the rebar diameter and the rebar depth, are both less than 0.1. In order to reduce the data dimension and the model complexity, this feature item is deleted in subsequent training.

[0076] Specifically, the K-nearest neighbor (KNN) algorithm was introduced to fill missing values, especially for missing concrete parameters, to ensure the integrity and quality of the data set and avoid prediction bias caused by missing data. By preprocessing the data, the relevance and reliability of the data were further improved, providing a solid foundation for model training.

[0077] Exemplarily, based on the first model, calculating the importance score of each type of concrete parameter in the completed concrete parameters may include:

[0078] Based on the first model and in combination with the second formula, the importance score of each type of concrete parameter in the completed concrete parameters is calculated.

[0079] The second formula can be:

[0080]

[0081] Among them, I j represents the importance score of concrete parameters of type j, T represents the total number of decision trees in the first model, S t,j represents the set of all split nodes in decision tree t using concrete parameters of type j, Gain S represents the information gain of node s.

[0082] For example, in some embodiments, the process of determining the missing data in the shear strength and concrete parameters and completing the missing data based on the K nearest neighbor algorithm may specifically include the following process:

[0083] Step S1, the obtained shear strength of the concrete interface and the concrete parameters are defined as a data set Among them, x i represents the feature vector of the i-th sample, y i represents the shear strength of the i-th sample, n represents the number of samples, and for a sample’s characteristic vector x i Let's say x i ={x i1 , x i2 , x i3 ,......,x im}, m represents the number of concrete parameter types.

[0084] Step S2, find the data set with missing data: D 缺失 The missing data in the data set of missing data are generally concrete parameters. If the missing data is shear strength, the data set corresponding to the shear strength is directly eliminated.

[0085] Step S3, after finding the data set with missing data, calculate the Euclidean distance from the data set with complete data to the data set with missing data, select the 10 data sets closest to the Euclidean distance as the neighbors of the data set with missing data, determine the type of missing data, and use the mean of the data of this type of the 10 neighbors to complete the missing data.

[0086] Specifically, in order to more clearly illustrate the calculation method of Euclidean distance, define x1 = {x 11 , x 12 , x 13 ,......,x 1m} and x2={x 21 , x 22 , x 23 ,......,x 2m}, assuming x 11 If there is missing data, the calculation formula of Euclidean distance is: In short, it calculates the distance between two vectors based on data other than the missing data type.

[0087] Specifically, in some specific embodiments, the first model may be a LightGBM regression model M:

[0088] M = LightGBMRegressor(Θ)

[0089] Among them, Regressor represents the regression model, and Θ represents the initial hyperparameter set of the first model.

[0090] Step 103, based on the collaborative group optimization algorithm and the genetic algorithm, the hyperparameters of the first model are optimized to obtain a target hyperparameter set; the target hyperparameter set is applied to the first model to obtain a second model, and the second model is trained based on the training set and the test set to obtain a shear strength prediction model.

[0091] Exemplarily, based on the collaborative group optimization algorithm and the genetic algorithm, optimizing the hyperparameters of the first model to obtain the target hyperparameter set may include:

[0092] Initialize the parameters of the collaborative swarm optimization algorithm, the parameters of the genetic algorithm and the population parameters, and obtain the number and upper and lower bounds of the hyperparameters of the first model; wherein the fitness function of the collaborative swarm optimization algorithm is the negative mean square error of the first model on the training set.

[0093] Based on the number and upper and lower bounds of the hyperparameters of the first model, the initial particle positions in the collaborative swarm optimization algorithm are determined to obtain the initial population.

[0094] For each iterative population, the particles in the population are iterated based on the collaborative swarm optimization algorithm to obtain updated particles.

[0095] Based on the fitness value of the best particle among the updated particles, the selection probability of the genetic algorithm is determined.

[0096] Based on the selection probability, the parent particles of the genetic algorithm are determined.

[0097] The parent particles are iterated based on the genetic algorithm to obtain the child particles.

[0098] The offspring particles and the updated particles are combined to form a new population, which is recorded as the first population.

[0099] The particles in the first population are sorted according to the fitness values, and a preset number of particles are selected as the second population.

[0100] The global optimal position is updated according to the particle with the smallest fitness value in the second population, and when the current number of iterations does not reach the maximum number of iterations, the second population is used as the population for the next iteration.

[0101] After the current number of iterations reaches the maximum number of iterations, the iteration is stopped and the global optimal position at this time is used as the target hyperparameter set.

[0102] Exemplarily, determining the selection probability of the genetic algorithm based on the fitness value of the best particle among the updated particles may include:

[0103] Based on the fitness value of the best particle among the updated particles, combined with the first formula, the selection probability of the genetic algorithm is determined.

[0104] The first formula can be:

[0105]

[0106] Among them, Pi represents the selection probability, N represents the number of particles, and f j represents the fitness value of particle j, and f represents the fitness value of the optimal particle.

[0107] Exemplarily, initializing the parameters of the collaborative swarm optimization algorithm, the parameters of the genetic algorithm, and the population parameters may include:

[0108] Initialize particle velocity, personal best position, personal best fitness value, global best position, global best fitness value, population size, maximum number of iterations, initial inertia weight, inertia weight reduction rate, acceleration coefficient, crossover probability, and mutation probability.

[0109] Exemplarily, before using the collaborative swarm optimization algorithm and the genetic algorithm, the parameters need to be initialized, specifically including:

[0110] 1.1 Parameter settings

[0111] Population size: Set the number of particles N = 50.

[0112] Maximum number of iterations: Set the number of iterations T = 100.

[0113] Problem Dimensionality: Number of hyperparameters Dim=11.

[0114] Upper and lower bounds of hyperparameters: Set upper and lower bounds LB and UB for each hyperparameter.

[0115] Inertia weight parameters: initial inertia weight w(0)=0.9, inertia weight reduction rate k=0.5.

[0116] Acceleration coefficient: c1=c2=2.

[0117] Genetic algorithm parameters: crossover probability Pc=0.8, mutation probability Pm=0.1.

[0118] 1.2. Initialize particle position

[0119] Randomly generate an initial position X for each particle between the upper and lower bounds of each dimension i .

[0120] X i =LB+rand(Dim)*(UB-LB)i=1,2,,,,N

[0121] Among them, rand(Dim) generates a random vector of dimension Dim, with elements between [0,1].

[0122] 1.3. Initialize particle speed

[0123] Set the initial velocity of all particles to the zero vector v i =0.

[0124] 1.4. Calculate the initial fitness value

[0125] For each particle X i , calculate its fitness value f(X i ).

[0126] 30% of the training set is used as the validation set, and the fitness function uses the negative mean square error (MSE) of the model on the validation set as the fitness value:

[0127]

[0128] Among them, l is the number of samples in the validation set, y i To verify the actual value of concentrated shear strength, is the predicted value of the second model.

[0129] 1.5. Initialize personal best position and fitness value

[0130] Take each particle's current position as its personal best position:

[0131] p besti =X i

[0132] Take each particle's fitness value as its personal best fitness value:

[0133] f pbesti =f(X i )

[0134] 1.6. Initialize the global best position and fitness value

[0135] Find the particle with the largest fitness value in the current population and set it as the global optimal position:

[0136] g best =X g ,

[0137] f gbest =f(X g )

[0138] In the algorithm iterative optimization phase, for each generation (from t=1 to T), the following steps are performed:

[0139] 2.1. Update inertia weight

[0140] Adaptive update formula using inertia weight:

[0141] w(t+1)=w(t)×(1-e -k×t )

[0142] Among them, w(0) is the initial inertia weight.

[0143] 2.2. For each particle i (from i=1 to N), perform the following operations:

[0144] Generate random numbers: Generate a random vector between [0,1]

[0145] r1,r2,r3,r4,r5∈[0,1]

[0146] Inertia Weight Term (IWV):

[0147] IWV i =w(t)×v i

[0148] Personal Best (PBC):

[0149] PBC i =r1×(ε×rand(pbest i )-X i )

[0150] Among them, ε represents a small constant used to introduce small disturbances, rand(pbest i ) represents a random perturbation of the personal best position.

[0151] Global Best Term (GBC):

[0152] GBC i =r2×(g best -X i )

[0153] Dynamic Attractor (DAC):

[0154]

[0155] Calculate attract i The specific process includes: determining the neighborhood, setting the neighborhood size H = 5, for particle i, according to the Euclidean distance between particles, selecting the H particles closest to form a neighborhood Ni, and finding the particle with the largest fitness value in Ni, whose position is attract i .

[0156] Adaptive Neighborhood Interaction Term (ANIC):

[0157] Set neighborbest i =attract i ,but

[0158] ANIC i =r4×(neighborbest i -X i )

[0159] Diversity Maintenance Item (DMC):

[0160] DMC i =c2×r5×(diversity i -Xi )

[0161] diversity i = X i + b × (X center - X i )

[0162]

[0163] where b is a random number in the range of [0, 1].

[0164] 2.3. Update speed

[0165] Sum up the above items to update the speed:

[0166] v i = IWV i + PBC i + GBC i + DAC i + ANIC i + DMC i

[0167] Update the position: Update the position of the particle

[0168] X i = X i + v i

[0169] Boundary handling: If X i exceeds the upper and lower bounds LB and UB, then perform boundary handling:

[0170] If X i < LB, then X i = LB.

[0171] If X i > UB, then X i = UB.

[0172] 2.4. Calculate the fitness value

[0173] Calculate the new fitness value of the particle: f(X i )

[0174] 2.5. Update the personal best position and fitness value

[0175] If the current fitness value of the particle is better than its personal best fitness value, then update:

[0176] p besti = X i

[0177] f pbesti = f(Xi )

[0178] Specifically, the specific process of genetic algorithm is as follows:

[0179] 3.1. Select Operation

[0180] Calculate the selection probability Pi based on the fitness value:

[0181]

[0182] Using the roulette method, parent particles are selected according to Pi to form a mating pool.

[0183] 3.2 Crossover Operation

[0184] With the crossover probability Pc, the particles in the mating pool are crossovered.

[0185] Using single point crossover:

[0186] For each pair of parent particles, randomly select an intersection point c (between [1, Dim-1]). Generate two child particles:

[0187] offspring1=[parent1(1:c),parent2(1+c:Dim)]

[0188] offspring2=[parent2(1:c), parent1(1+c:Dim)]

[0189] Among them, parent represents the parent generation and offspring represents the offspring generation.

[0190] 3.3 Mutation Operation

[0191] With the mutation probability Pm, each dimension of the descendant particles is mutated.

[0192] Variation method:

[0193] For each dimension j of the particle, mutation is performed with probability Pm:

[0194] Calculate the magnitude of variation:

[0195] O j =0.1(UB j -LB j ),j=1,2,3,,,,,Dim

[0196] Generate random perturbations:

[0197] T=randn()×O j

[0198] Among them, randn() is a random number generation function of standard normal distribution.

[0199] Update offspring i，j =offspring i，j +T.

[0200] 3.4 Boundary Processing

[0201] For the mutated offspring particles, check whether they exceed the boundary and perform boundary processing.

[0202] 3.5 Fitness Evaluation

[0203] For the mutated offspring particles, calculate their fitness value f(offspring i ).

[0204] The above iterations are repeated multiple times. When the number of iterations reaches the maximum number of iterations, the global optimal position is used as the target hyperparameter set. After that, the target hyperparameter set is applied to the first model to obtain the second model.

[0205] Exemplarily, training the second model based on the training set and the test set to obtain the shear strength prediction model may include:

[0206] The second model is trained based on the training set to obtain a trained model.

[0207] The accuracy of the trained model was checked based on the test set, mean squared error, and mean absolute error.

[0208] If the accuracy of the trained model meets the preset conditions, the trained model will be used as the shear strength prediction model.

[0209] For example, after obtaining the shear strength prediction model, the key features affecting the shear strength of the concrete interface can be output according to the shear strength prediction model. The specific process is as follows:

[0210] Step S1, combining the third formula, calculating the SHAP value of each concrete parameter; wherein the SHAP value of each type of concrete parameter represents the contribution of the feature to the prediction result.

[0211] The third formula can be:

[0212]

[0213] Among them, Φ irepresents the SHAP value of feature i, N represents the set of all features, S represents any feature subset that does not contain feature i, |S| represents the number of features in set S, v(S) represents the contribution of feature set S to the model prediction output, and v(S∪{i}) represents the contribution of feature set S∪{i} containing feature i to the model prediction output.

[0214] Step S2, determining the key features that affect the shear strength of the concrete interface according to the magnitude of each SHAP value.

[0215] In this embodiment, the SHAP analysis method is combined to perform an explanatory analysis on the shear strength prediction model, and the contribution of each input feature to the shear strength prediction is deeply understood, which enhances the interpretability of the shear strength prediction model and the transparency of engineering applications, and solves many limitations of traditional methods under complex working conditions and multi-dimensional parameter nonlinear relationships. It has important theoretical significance and broad engineering application prospects.

[0216] Step 104: input the concrete parameters of the target concrete interface into the shear strength prediction model to obtain the shear strength prediction value of the target concrete interface.

[0217] The above concrete interface shear strength prediction method uses the Social Spider Optimization Algorithm (SSOA) for global search. The Social Spider Optimization Algorithm SSOA can efficiently explore in a wide range of parameter spaces and find preliminary high-quality hyperparameter combinations by simulating group collaboration and information exchange between individuals. Subsequently, the Genetic Algorithm (GA) takes over the solution obtained by the Social Spider Optimization Algorithm SSOA, further optimizes the hyperparameters through crossover and mutation operations, enhances the diversity of the population, and promotes a more comprehensive exploration of the solution space. The hybrid optimization strategy combining the Social Spider Optimization Algorithm SSOA and the Genetic Algorithm GA not only speeds up the convergence speed of hyperparameter optimization and significantly shortens the model training time, but also effectively avoids the problem that a single optimization algorithm is prone to fall into the local optimum, improves the prediction performance of the LightGBM model, and ensures accurate prediction of the shear strength of the concrete interface under complex working conditions.

[0218] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean 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 the present application.

[0219] Corresponding to the concrete interface shear strength prediction method described in the above embodiment, Figure 2 A structural block diagram of a concrete interface shear strength prediction device provided in an embodiment of the present application is shown. For ease of explanation, only the parts related to the embodiment of the present application are shown.

[0220] See also Figure 2 , the concrete interface shear strength prediction device in the embodiment of the present application may include:

[0221] The data acquisition module 201 is used to obtain the shear strength of the concrete interface and concrete parameters.

[0222] The data processing module 202 is used to establish an initial LightGBM model, recorded as the first model, and pre-process the shear strength and concrete parameters to obtain a training set and a test set.

[0223] The first modeling module 203 is used to optimize the hyperparameters of the first model based on a collaborative group optimization algorithm and a genetic algorithm to obtain a target hyperparameter set; apply the target hyperparameter set to the first model to obtain a second model, and train the second model based on a training set and a test set to obtain a shear strength prediction model.

[0224] The result prediction module 204 is used to input the concrete parameters of the target concrete interface into the shear strength prediction model to obtain the shear strength prediction value of the target concrete interface.

[0225] Exemplarily, the data processing module 202 may be used to:

[0226] Initialize the parameters of the collaborative swarm optimization algorithm, the parameters of the genetic algorithm and the population parameters, and obtain the number and upper and lower bounds of the hyperparameters of the first model; wherein the fitness function of the collaborative swarm optimization algorithm is the negative mean square error of the first model on the training set.

[0227] Based on the number and upper and lower bounds of the hyperparameters of the first model, the initial particle positions in the collaborative swarm optimization algorithm are determined to obtain the initial population.

[0228] For each iterative population, the particles in the population are iterated based on the collaborative swarm optimization algorithm to obtain updated particles.

[0229] Based on the fitness value of the best particle among the updated particles, the selection probability of the genetic algorithm is determined.

[0230] Based on the selection probability, the parent particles of the genetic algorithm are determined.

[0231] The parent particles are iterated based on the genetic algorithm to obtain the child particles.

[0232] The offspring particles and the updated particles are combined to form a new population, which is recorded as the first population.

[0233] The particles in the first population are sorted according to the fitness values, and a preset number of particles are selected as the second population.

[0234] The global optimal position is updated according to the particle with the smallest fitness value in the second population, and when the current number of iterations does not reach the maximum number of iterations, the second population is used as the population for the next iteration.

[0235] After the current number of iterations reaches the maximum number of iterations, the iteration is stopped and the global optimal position at this time is used as the target hyperparameter set.

[0236] Exemplarily, the data processing module 202 may also be used for:

[0237] Based on the fitness value of the best particle among the updated particles, combined with the first formula, the selection probability of the genetic algorithm is determined.

[0238] The first formula can be:

[0239]

[0240] Among them, Pi represents the selection probability, N represents the number of particles, and f j represents the fitness value of particle j, and f represents the fitness value of the optimal particle.

[0241] Exemplarily, initializing the parameters of the collaborative swarm optimization algorithm, the parameters of the genetic algorithm and the population parameters includes:

[0242] Initialize particle velocity, personal best position, personal best fitness value, global best position, global best fitness value, population size, maximum number of iterations, initial inertia weight, inertia weight reduction rate, acceleration coefficient, crossover probability, and mutation probability.

[0243] Exemplarily, the first modeling module 203 may be used to:

[0244] The second model is trained based on the training set to obtain a trained model.

[0245] The accuracy of the trained model was checked based on the test set, mean squared error, and mean absolute error.

[0246] If the accuracy of the trained model meets the preset conditions, the trained model will be used as the shear strength prediction model.

[0247] Exemplarily, the data processing module 202 may be used to:

[0248] Determine missing data in shear strength and concrete parameters.

[0249] The missing data are completed based on the K nearest neighbor algorithm.

[0250] Based on the first model, the importance score of each type of concrete parameter in the completed concrete parameters is calculated.

[0251] For each type of concrete parameter, if the importance score is greater than a preset threshold, the data of the concrete parameter of this type is retained; if the importance score is less than or equal to the preset threshold, the data of the concrete parameter of this type is deleted to obtain a processed data set.

[0252] The processed data set is divided into training set and test set according to the preset ratio.

[0253] Exemplarily, the data processing module 202 may also be used for:

[0254] Based on the first model and in combination with the second formula, the importance score of each type of concrete parameter in the completed concrete parameters is calculated.

[0255] The second formula can be:

[0256]

[0257] Among them, I j represents the importance score of concrete parameters of type j, T represents the total number of decision trees in the first model, S t,j represents the set of all split nodes in decision tree t using concrete parameters of type j, Gain S represents the information gain of node s.

[0258] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0259] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by 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 embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0260] The present application also provides a terminal device, see Figure 3The terminal device 300 may include: at least one processor 310 and a memory 320, wherein the memory 320 is used to store a computer program 321, and the processor 310 is used to call and run the computer program 321 stored in the memory 320 to implement the steps in any of the above-mentioned method embodiments, for example Figure 1 Steps 101 to 104 in the illustrated embodiment. Alternatively, when the processor 310 executes the computer program, the functions of each module / unit in the above-mentioned device embodiments are implemented, for example Figure 2 The functions of each module are shown.

[0261] Exemplarily, the computer program 321 may be divided into one or more modules / units, one or more modules / units are stored in the memory 320, and are executed by the processor 310 to complete the present application. The one or more modules / units may be a series of computer program segments that can complete specific functions, and the program segments are used to describe the execution process of the computer program in the terminal device 300.

[0262] Those skilled in the art will understand that Figure 3 It is only an example of a terminal device and does not constitute a limitation on the terminal device. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components, such as input and output devices, network access devices, buses, etc.

[0263] The processor 310 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.

[0264] The memory 320 may be an internal storage unit of the terminal device, or an external storage device of the terminal device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. The memory 320 is used to store the computer program and other programs and data required by the terminal device. The memory 320 may also be used to temporarily store data that has been output or is to be output.

[0265] The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the bus in the drawings of the present application is not limited to only one bus or one type of bus.

[0266] The concrete interface shear strength prediction method provided in the embodiment of the present application can be applied to terminal devices such as computers, wearable devices, vehicle-mounted devices, tablet computers, laptops, netbooks, etc. The embodiment of the present application does not impose any restrictions on the specific type of terminal devices.

[0267] The embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in each embodiment of the above-mentioned concrete interface shear strength prediction method can be implemented.

[0268] An embodiment of the present application provides a computer program product. When the computer program product is run on a mobile terminal, the mobile terminal can implement the steps in each embodiment of the above-mentioned concrete interface shear strength prediction method when executing the computer program product.

[0269] If the integrated unit is implemented in the form of 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, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device that can carry the computer program code to the camera device / terminal device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electrical carrier signal, telecommunication signal and software distribution medium. For example, a USB flash drive, a mobile hard disk, a disk or an optical disk.

[0270] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0271] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example 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 performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0272] In the embodiments provided in the present application, it should be understood that the disclosed devices / network equipment and methods can be implemented in other ways. For example, the device / network equipment embodiments described above are merely schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0273] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0274] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for predicting the shear strength of concrete interface, characterized in that: include: Obtain the shear strength of the concrete interface and concrete parameters; An initial LightGBM model is established, which is recorded as the first model, and the shear strength and the concrete parameters are preprocessed to obtain a training set and a test set; Based on the collaborative group optimization algorithm and the genetic algorithm, the hyperparameters of the first model are optimized to obtain a target hyperparameter set; the target hyperparameter set is applied to the first model to obtain a second model, and the second model is trained based on the training set and the test set to obtain a shear strength prediction model; The concrete parameters of the target concrete interface are input into the shear strength prediction model to obtain the shear strength prediction value of the target concrete interface.

2. The method for predicting the shear strength of concrete interface according to claim 1, characterized in that: The method of optimizing the hyperparameters of the first model based on the collaborative group optimization algorithm and the genetic algorithm to obtain a target hyperparameter set includes: Initializing parameters of the collaborative swarm optimization algorithm, parameters of the genetic algorithm and population parameters, and obtaining the number and upper and lower bounds of the hyperparameters of the first model; wherein the fitness function of the collaborative swarm optimization algorithm is the negative mean square error of the first model on the training set; Based on the number and upper and lower bounds of the hyperparameters of the first model, determining initial particle positions in the collaborative swarm optimization algorithm to obtain an initial population; For each iterative population, the particles in the population are iterated based on the collaborative swarm optimization algorithm to obtain updated particles; Determining the selection probability of the genetic algorithm based on the fitness value of the best particle among the updated particles; Based on the selection probability, determining the parent particle of the genetic algorithm; Iterate the parent particles based on the genetic algorithm to obtain child particles; Combining the offspring particles and the updated particles to form a new population, recorded as the first population; Sort the particles in the first population according to fitness values, and select a preset number of particles as the second population; updating the global optimal position according to the particle with the smallest fitness value in the second population, and using the second population as the population for the next iteration when the current number of iterations does not reach the maximum number of iterations; After the current number of iterations reaches the maximum number of iterations, the iteration is stopped, and the global optimal position at this time is used as the target hyperparameter set.

3. The method for predicting the shear strength of concrete interface according to claim 2, characterized in that: The step of determining the selection probability of the genetic algorithm based on the fitness value of the best particle among the updated particles comprises: Based on the fitness value of the best particle among the updated particles, combined with the first formula, determining the selection probability of the genetic algorithm; The first formula is: Where Pi represents the selection probability, N represents the number of particles, and f j represents the fitness value of particle j, and f represents the fitness value of the optimal particle.

4. The method for predicting the concrete interface shear strength according to claim 2, characterized in that: The parameters of the initialization collaborative group optimization algorithm, the parameters of the genetic algorithm and the population parameters include: Initialize particle velocity, personal best position, personal best fitness value, global best position, global best fitness value, population size, maximum number of iterations, initial inertia weight, inertia weight reduction rate, acceleration coefficient, crossover probability, and mutation probability.

5. The method for predicting the concrete interface shear strength according to claim 1, characterized in that: The step of training the second model based on the training set and the test set to obtain a shear strength prediction model comprises: Training the second model based on the training set to obtain a trained model; Testing the accuracy of the trained model based on the test set, mean square error and mean absolute error; If the accuracy of the trained model meets the preset conditions, the trained model is used as the shear strength prediction model.

6. The method for predicting the concrete interface shear strength according to claim 1, characterized in that: The preprocessing of the shear strength and the concrete parameters to obtain a training set and a test set includes: determining missing data in the shear strength and the concrete parameters; Completing the missing data based on a K-nearest neighbor algorithm; Calculating, based on the first model, the importance score of each type of concrete parameter in the completed concrete parameters; For each type of concrete parameter, if the importance score is greater than a preset threshold, the data of the concrete parameter of this type is retained; if the importance score is less than or equal to the preset threshold, the data of the concrete parameter of this type is deleted, thereby obtaining a processed data set; The processed data set is divided into the training set and the test set according to a preset ratio.

7. The method for predicting the shear strength of concrete interface according to claim 6, characterized in that: The calculating, based on the first model, the importance score of each type of concrete parameter in the completed concrete parameters includes: Based on the first model and in combination with the second formula, calculating the importance score of each type of concrete parameter in the completed concrete parameters; The second formula is: Among them, I j represents the importance score of concrete parameters of type j, T represents the total number of decision trees in the first model, S t,j represents the set of all split nodes in decision tree t using concrete parameters of type j, Gain S represents the information gain of node s.

8. A device for predicting the shear strength of concrete interface, characterized in that: include: A data acquisition module is used to obtain the shear strength of the concrete interface and concrete parameters; A data processing module is used to establish an initial LightGBM model, recorded as a first model, and pre-process the shear strength and the concrete parameters to obtain a training set and a test set; A first modeling module is used to optimize the hyperparameters of the first model based on a collaborative group optimization algorithm and a genetic algorithm to obtain a target hyperparameter set; apply the target hyperparameter set to the first model to obtain a second model, and train the second model based on a training set and a test set to obtain a shear strength prediction model; The result prediction module is used to input the concrete parameters of the target concrete interface into the shear strength prediction model to obtain the shear strength prediction value of the target concrete interface.

9. A terminal device, comprising: A processor and a memory, wherein the memory stores a computer program that can be run on the processor, wherein when the processor executes the computer program, the method for predicting the shear strength of a concrete interface according to any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for predicting the concrete interface shear strength according to any one of claims 1 to 7 is implemented.

Images