Building seismic strength prediction and optimization method

Abstract

The invention belongs to the technical field of building quality detection and optimization, and relates to a building shock strength prediction and optimization method, which comprises the following steps: inputting the data of a current building into a trained shock strength prediction model to obtain the shock strength of the building, and comparing the predicted shock strength with the expected shock strength to obtain the shock strength of the building. Judging whether the seismic strength of the building reaches the standard or not, and if not, performing optimization; in the optimization process, a building structure scheme is generated in a random generation mode, and then a multi-objective optimization algorithm is adopted to optimize the generated initial structure scheme; predication and optimization design of the seismic strength of the building are achieved, and therefore the technical problem that an existing seismic design method cannot reach the optimal level is effectively solved; by means of the comprehensive data analysis and optimization algorithm, the anti-seismic performance of the building can be evaluated more comprehensively, and a targeted optimization scheme is provided to improve the safety and stability of the building in an earthquake.

Description

Technical Field

[0001] The present application belongs to the technical field of building quality detection and optimization, and more specifically, to a method for predicting and optimizing the seismic strength of a building. Background Art

[0002] In the field of construction engineering, improving the earthquake resistance of buildings is a vital task to ensure public safety and reduce losses caused by earthquake disasters. Buildings must have sufficient earthquake resistance to reduce the damage and losses caused by earthquakes.

[0003] However, traditional building seismic design methods may not fully utilize modern technology and data resources, resulting in suboptimal seismic performance of the design. Summary of the invention

[0004] The present invention provides a method for predicting and optimizing the seismic strength of a building, and is intended to solve the technical problem that the current seismic design method fails to reach the optimal level.

[0005] A method for predicting and optimizing the seismic strength of a building comprises the following steps:

[0006] Step 1: Obtain the building's structural data, historical earthquake data, and real-time vibration data of the building during an earthquake;

[0007] Step 2: Preprocess the data obtained in step 1 to obtain a data set;

[0008] Step 3: Divide the data set into a training set and a validation set, train the constructed seismic strength prediction model, and obtain a trained seismic strength prediction model;

[0009] Step 4: Input the data of the current building into the trained seismic strength prediction model to obtain the seismic strength of the building, and compare the predicted seismic strength with the expected seismic strength. If the predicted seismic strength is lower than the expected seismic strength, execute step 5, otherwise end;

[0010] Step 5: Based on the predicted earthquake resistance of the building and according to the existing building structure design parameters, an initial building structure scheme is generated by random generation;

[0011] Step 6: For the generated initial building structure plan, a multi-objective optimization algorithm is used to optimize the building structure.

[0012] The present invention inputs the data of the current building into a trained seismic strength prediction model to obtain the seismic strength of the building, compares the predicted seismic strength with the expected seismic strength, determines whether the seismic strength of the building meets the standard, and optimizes it if it does not meet the standard; during the optimization process, a building structure plan is generated by a random generation method, and then a multi-objective optimization algorithm is used to optimize the generated initial structure plan; the prediction and optimization design of the seismic strength of the building are realized, thereby effectively solving the technical problem that the current seismic design method fails to reach the optimal level; using comprehensive data analysis and optimization algorithms, the seismic performance of the building can be more comprehensively evaluated, and targeted optimization plans can be proposed to improve the safety and stability of the building in earthquakes.

[0013] Preferably, the seismic strength prediction model adopts a support vector machine model;

[0014] The Gaussian kernel is used as the kernel function of the support vector machine model; the loss function uses the hinge loss function.

[0015] Preferably, the training steps of the support vector machine model are as follows:

[0016] Model initialization: Initialize the parameters of the support vector model;

[0017] Kernel function parameter setting: For the Gaussian kernel function, set the bandwidth parameter of the Gaussian kernel;

[0018] Model training: Use the training set to train the support vector machine model and use the sequential minimum optimization algorithm for optimization:

[0019]

[0020] Where: w represents the weight vector of the hyperplane; b represents the bias term; C represents the regularization parameter; x i represents the feature vector of the i-th sample; y i Represents the true label of the i-th sample;

[0021] During the training process, the model parameters w and b are adjusted by optimizing the loss function:

[0022] max(0,1-y i (w·x i -b));

[0023] The validation set is used to evaluate the performance of the trained support vector machine model. If it does not meet expectations, the parameters of the support vector machine model are adjusted.

[0024] Preferably, step 5 comprises the following steps:

[0025] Determine the parameter space: confine the column position to a rectangular area whose boundaries are determined by the building's foundation dimensions;

[0026] Set the random sampling rule: use uniform distribution to generate the positions of the columns, and randomly generate N pairs of coordinates in the parameter space;

[0027] Structural evaluation: Conduct structural mechanics analysis on each generated building structure plan;

[0028] Screening qualified solutions: Compare the evaluation results with the set threshold. If the evaluation result is greater than the threshold, the solution is considered to meet the design requirements.

[0029] Preferably, step 6 comprises the following steps:

[0030] Define the optimization goal: Minimize the cost while achieving the seismic strength standard as the optimization goal;

[0031] Constructing the fitness function: taking seismic strength and cost as multiple optimization objectives and transforming them into a single optimization objective through weighted summation;

[0032] Initial population: The initial building structure scheme generated in step 5 is used as the initial population of the genetic algorithm;

[0033] Evolution: Generate new solutions through crossover and mutation operations, evaluate the quality of individuals according to the fitness function, and select individuals with high fitness for reproduction;

[0034] Form a Pareto optimal solution set: retain non-dominated solutions and obtain a set of Pareto optimal solutions.

[0035] Preferably, the fitness function is as follows:

[0036] F(x)=w s ×S(x)+w c ×C(x);

[0037] Where: F(x) represents the fitness of individual x; w s and w c represents the weights of seismic strength and cost; S(x) and C(x) represent seismic strength and cost respectively.

[0038] Preferably, the evolution process is as follows:

[0039] Crossover operation: select two individuals from the parent population, randomly select a crossover point, and then exchange the gene fragments of the two individuals after the crossover point:

[0040] Offspring=Parent1[:CrossPoint]+Parent2[CrossPoint:];

[0041] Where: Offspring represents the individual after crossover; Parent1 represents parent individual 1; CrossPoint represents the crossover point; Parent2 represents parent individual 2;

[0042] Mutation operation: Randomly change the value of the gene position:

[0043]

[0044] Where: P m represents the mutation rate; Mutated Offspring[i] represents the individual position after mutation; OriginalGene Value represents the original gene position value; Randomly Selected New Gene Value represents the randomly selected new gene position value;

[0045] Calculate fitness: Calculate the fitness value of the individual according to the fitness function;

[0046] Selection operation: Select a predetermined number of individuals from the parent and offspring populations, and then select the individuals with the best fitness as members of the next generation population.

[0047] The beneficial effects of the present invention include:

[0048] The present invention inputs the data of the current building into a trained seismic strength prediction model to obtain the seismic strength of the building, compares the predicted seismic strength with the expected seismic strength, determines whether the seismic strength of the building meets the standard, and optimizes it if it does not meet the standard; during the optimization process, a building structure plan is generated by a random generation method, and then a multi-objective optimization algorithm is used to optimize the generated initial structure plan; the prediction and optimization design of the seismic strength of the building are realized, thereby effectively solving the technical problem that the current seismic design method fails to reach the optimal level; using comprehensive data analysis and optimization algorithms, the seismic performance of the building can be more comprehensively evaluated, and targeted optimization plans can be proposed to improve the safety and stability of the building in earthquakes. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] 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 creative labor.

[0050] Figure 1 An overall step block diagram provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0052] See also Figure 1 As shown, the best embodiment of the present invention is further described;

[0053] A method for predicting and optimizing the seismic strength of a building comprises the following steps:

[0054] Step 1: Obtain the building's structural data, historical earthquake data, and real-time vibration data of the building during an earthquake;

[0055] Structural data acquisition:

[0056] Building structural characteristic data: including the building's floor plan, structural form (such as frame structure, reinforced concrete structure, etc.), structural component dimensions and materials, etc.

[0057] Structural parameter data: such as cross-sectional dimensions, material strength and other parameters of main load-bearing components such as columns, beams and walls.

[0058] Historical earthquake data acquisition:

[0059] Earthquake activity records: collect records of historical earthquake events, including information such as the magnitude, epicenter location, and time of occurrence of the earthquake.

[0060] Seismic parameter data: such as seismic acceleration time history, seismic wave propagation path and other parameters, which are used to analyze seismic effects under different seismic conditions.

[0061] Real-time vibration data acquisition:

[0062] Real-time monitoring equipment data: The real-time vibration response data of buildings during earthquakes is obtained through sensor networks and monitoring equipment, including vibration acceleration, vibration frequency, displacement and other information.

[0063] Vibration data processing: Perform preprocessing operations such as sampling, filtering, and data cleaning on real-time vibration data to extract effective vibration features.

[0064] Step 2: Preprocess the data obtained in step 1 to obtain a data set;

[0065] Data cleaning and screening: Clean and screen the acquired data to remove outliers and missing data to ensure data integrity and accuracy.

[0066] Data standardization and normalization: Standardize or normalize structural data and seismic data to make them have similar data distribution and magnitude to improve the convergence speed and effect of model training.

[0067] Dataset division: The preprocessed dataset is divided into a training set and a validation set. Cross-validation or holdout methods are usually used to ensure that the data of the training set and validation set are evenly distributed and independent of each other.

[0068] Feature engineering: Perform feature engineering on the data to extract effective feature variables, such as structural parameters, seismic parameters, and characteristic values ​​of real-time vibration data, for use in building a seismic strength prediction model.

[0069] The specific steps of the feature engineering are as follows:

[0070] Structural feature extraction:

[0071] Structural form coding: Encode the structural form of the building (such as frame structure, reinforced concrete structure, etc.) and convert it into numerical features.

[0072] Component size parameter extraction: Extract the size parameters of the main components of the building (such as columns, beams, walls, etc.), including cross-sectional area, length, width, etc.

[0073] Material property conversion: Convert material properties (such as elastic modulus, tensile strength, etc.) into numerical features to reflect the material quality and strength of the building.

[0074] Seismic data feature extraction:

[0075] Earthquake motion parameter extraction: Extracting earthquake motion parameters from historical earthquake data, including magnitude, source distance, spectrum characteristics, etc.

[0076] Earthquake time series characteristics: Spectral analysis and waveform feature extraction are performed on the earthquake time series to obtain richer earthquake motion characteristics.

[0077] Real-time vibration data feature extraction:

[0078] Frequency domain feature extraction: Extract frequency domain features of vibration data, such as main frequency, spectrum density, etc., through methods such as Fourier transform.

[0079] Time domain feature extraction: Extract the time domain statistical features of vibration data, such as mean, variance, peak value, kurtosis, etc.

[0080] Wavelet transform features: Wavelet transform is applied to extract the time-frequency domain features of vibration data to capture the local changes of vibration signals.

[0081] Data dimensionality reduction and feature selection:

[0082] Principal Component Analysis (PCA): Reduce the dimensionality of high-dimensional data, extract the main characteristic components, reduce the characteristic dimensions, and retain as much information as possible.

[0083] For example: Suppose there is a data set X containing m samples and n features, where each sample has n features. The goal of PCA is to find a transformation matrix W that projects the original data X into a new low-dimensional space Z so that the projected data has the largest variance. The dimension d of the new feature space is usually much smaller than the number of original features n;

[0084] Compute the sample mean vector:

[0085]

[0086] Where: represents the sample mean vector; x i represents the value of the i-th sample;

[0087] Compute the sample covariance matrix:

[0088]

[0089] Where: T represents the transpose operation of the matrix;

[0090] Compute the eigenvalues ​​and eigenvectors of the covariance matrix:

[0091] Eigenvalue: λ 1 ,λ 2 , ..., λ n , eigenvector: V = [v 1 , v 2 , ..., v n ];

[0092] Select the first d eigenvectors: Select the first d eigenvectors corresponding to the largest eigenvalue to form the transformation matrix W;

[0093] Project data: Project the original data X to a new low-dimensional space Z:

[0094] Z = X·W;

[0095] Feature selection algorithm: Use correlation analysis, information gain, variance analysis and other methods to select features that have a significant impact on earthquake intensity prediction.

[0096] Feature combination and crossover:

[0097] Feature combination: Combine multiple features to form new features to capture the interaction between features. For example, for two features x 1 and x 2 , we can construct a new feature x3 =x 1 × 2 ; As a combination of original features, this helps improve the model's ability to express complex relationships between features;

[0098] Feature crossover: Cross-combining different features to generate more feature combinations to improve the expressiveness of the model.

[0099] For feature x 1 and x 2 , we can generate new features x 3 =x 1 +x 2 ;x 4 =x 1 -x 2 Etc.; More feature combination information has been obtained; This helps to discover more complex correlations between features and improve the predictive ability of the model.

[0100] Dataset preparation: Finally, the preprocessed and feature-engineered dataset is prepared to be the input required for building the earthquake strength prediction model.

[0101] Step 3: Divide the data set into a training set and a validation set, train the constructed seismic strength prediction model, and obtain a trained seismic strength prediction model;

[0102] The seismic strength prediction model adopts a support vector machine model;

[0103] The Gaussian kernel is used as the kernel function of the support vector machine model; the loss function uses the hinge loss function.

[0104] The training steps of the support vector machine model are as follows:

[0105] Model initialization: Initialize the parameters of the support vector model;

[0106] Kernel function parameter setting: For the Gaussian kernel function, set the bandwidth parameter of the Gaussian kernel;

[0107] Model training: Use the training set to train the support vector machine model and use the sequential minimum optimization algorithm for optimization:

[0108]

[0109] Where: w represents the weight vector of the hyperplane; b represents the bias term; C represents the regularization parameter; x i represents the feature vector of the i-th sample; y i Represents the true label of the i-th sample;

[0110] During the training process, the model parameters w and b are adjusted by optimizing the loss function:

[0111] max(0, 1-y i (w·x i -b));

[0112] The validation set is used to evaluate the performance of the trained support vector machine model. If it does not meet expectations, the parameters of the support vector machine model are adjusted.

[0113] Step 4: Input the data of the current building into the trained seismic strength prediction model to obtain the seismic strength of the building, and compare the predicted seismic strength with the expected seismic strength. If the predicted seismic strength is lower than the expected seismic strength, execute step 5, otherwise end;

[0114] Step 5: Based on the predicted earthquake resistance of the building and according to the existing building structure design parameters, an initial building structure scheme is generated by random generation;

[0115] The step 5 comprises the following steps:

[0116] Determine the parameter space: confine the column position to a rectangular area whose boundaries are determined by the building's foundation dimensions;

[0117] Set the random sampling rule: use uniform distribution to generate the positions of the columns, and randomly generate N pairs of coordinates in the parameter space;

[0118] Structural evaluation: Conduct structural mechanics analysis on each generated building structure plan;

[0119] Screening qualified solutions: Compare the evaluation results with the set threshold. If the evaluation result is greater than the threshold, the solution is considered to meet the design requirements.

[0120] Step 6: For the generated initial building structure plan, a multi-objective optimization algorithm is used to optimize the building structure.

[0121] The step 6 comprises the following steps:

[0122] Define the optimization goal: Minimize the cost while achieving the seismic strength standard as the optimization goal;

[0123] Constructing the fitness function: taking seismic strength and cost as multiple optimization objectives and transforming them into a single optimization objective through weighted summation;

[0124] Initial population: The initial building structure scheme generated in step 5 is used as the initial population of the genetic algorithm;

[0125] Evolution: Generate new solutions through crossover and mutation operations, evaluate the quality of individuals according to the fitness function, and select individuals with high fitness for reproduction;

[0126] Form a Pareto optimal solution set: retain non-dominated solutions and obtain a set of Pareto optimal solutions.

[0127] Preferably, the fitness function is as follows:

[0128] F(x)=w s ×S(x)+w c ×C(x);

[0129] Where: F(x) represents the fitness of individual x; w s and w e represents the weights of seismic strength and cost; S(x) and C(x) represent seismic strength and cost respectively.

[0130] Preferably, the evolution process is as follows:

[0131] Crossover operation: select two individuals from the parent population, randomly select a crossover point, and then exchange the gene fragments of the two individuals after the crossover point:

[0132] Offspring=Parent1[:CrossPoint]+Parent2[CrossPoint:];

[0133] Where: Offspring represents the individual after crossover; Parent1 represents parent individual 1; CrossPoint represents the crossover point; Parent2 represents parent individual 2;

[0134] Mutation operation: Randomly change the value of the gene position:

[0135]

[0136] Where: P m represents the mutation rate; Mutated Offspring[i] represents the individual position after mutation; OriginalGene Value represents the original gene position value; Randomly Selected New Gene Value represents the randomly selected new gene position value;

[0137] Calculate fitness: Calculate the fitness value of the individual according to the fitness function;

[0138] Selection operation: Select a predetermined number of individuals from the parent and offspring populations, and then select the individuals with the best fitness as members of the next generation population.

[0139] The present invention inputs the data of the current building into a trained seismic strength prediction model to obtain the seismic strength of the building, compares the predicted seismic strength with the expected seismic strength, determines whether the seismic strength of the building meets the standard, and optimizes it if it does not meet the standard; during the optimization process, a building structure plan is generated by a random generation method, and then a multi-objective optimization algorithm is used to optimize the generated initial structure plan; the prediction and optimization design of the seismic strength of the building are realized, thereby effectively solving the technical problem that the current seismic design method fails to reach the optimal level; using comprehensive data analysis and optimization algorithms, the seismic performance of the building can be more comprehensively evaluated, and targeted optimization plans can be proposed to improve the safety and stability of the building in earthquakes.

[0140] The above are only preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for predicting and optimizing the seismic strength of a building, characterized in that: The following steps are involved: Step 1: Obtain the building's structural data, historical earthquake data, and real-time vibration data of the building during an earthquake; Step 2: Preprocess the data obtained in step 1 to obtain a data set; Step 3: Divide the data set into a training set and a validation set, train the constructed seismic strength prediction model, and obtain a trained seismic strength prediction model; Step 4: Input the real-time data into the trained seismic strength prediction model to obtain the seismic strength of the building, and compare the predicted seismic strength with the expected seismic strength. If the predicted seismic strength is lower than the expected seismic strength, execute step 5, otherwise end; Step 5: Based on the predicted earthquake resistance of the building and according to the existing building structure design parameters, an initial building structure scheme is generated by random generation; Step 6: For the generated initial building structure plan, a multi-objective optimization algorithm is used to optimize the building structure.

2. A method for predicting and optimizing seismic strength of a building according to claim 1, characterized in that: The seismic strength prediction model adopts a support vector machine model; The Gaussian kernel is used as the kernel function of the support vector machine model; the loss function uses the hinge loss function.

3. A method for predicting and optimizing the seismic strength of a building according to claim 2, characterized in that: The training steps of the support vector machine model are as follows: Model initialization: Initialize the parameters of the support vector model; Kernel function parameter setting: For the Gaussian kernel function, set the bandwidth parameter of the Gaussian kernel; Model training: Use the training set to train the support vector machine model and use the sequential minimum optimization algorithm for optimization: Where: w represents the weight vector of the hyperplane; b represents the bias term; C represents the regularization parameter; x i represents the feature vector of the i-th sample; y i Represents the true label of the i-th sample; During the training process, the model parameters w and b are adjusted by optimizing the loss function: max(0,1-y i (w·x i -b)); The validation set is used to evaluate the performance of the trained support vector machine model. If it does not meet expectations, the parameters of the support vector machine model are adjusted.

4. A method for predicting and optimizing seismic strength of a building according to claim 1, characterized in that: The step 5 comprises the following steps: Determine the parameter space: confine the column position to a rectangular area whose boundaries are determined by the building's foundation dimensions; Set the random sampling rule: use uniform distribution to generate the positions of the columns, and randomly generate N pairs of coordinates in the parameter space; Structural evaluation: Conduct structural mechanics analysis on each generated building structure plan; Screening qualified solutions: Compare the evaluation results with the set threshold. If the evaluation result is greater than the threshold, the solution is considered to meet the design requirements.

5. A method for predicting and optimizing seismic strength of a building according to claim 1, characterized in that: The step 6 comprises the following steps: Define the optimization goal: Minimize the cost while achieving the seismic strength standard as the optimization goal; Constructing a fitness function: Taking seismic strength and cost as multiple optimization objectives, and transforming them into a single optimization objective through weighted summation; Initial population: The initial building structure scheme generated in step 5 is used as the initial population of the genetic algorithm; Evolution: Generate new solutions through crossover and mutation operations, evaluate the quality of individuals according to the fitness function, and select individuals with high fitness for reproduction; Form a Pareto optimal solution set: retain non-dominated solutions and obtain a set of Pareto optimal solutions.

6. A method for predicting and optimizing seismic strength of a building according to claim 5, characterized in that: The fitness function is as follows: F(x)=w s ×S(x)+w c ×C(x); Where: F(x) represents the fitness of individual x; w s and w c represents the weights of seismic strength and cost; S(x) and C(x) represent seismic strength and cost respectively.

7. A method for predicting and optimizing the seismic strength of a building according to claim 6, characterized in that: The evolution process is as follows: Crossover operation: select two individuals from the parent population, randomly select a crossover point, and then exchange the gene fragments of the two individuals after the crossover point: Offspring=Parent1[:CrossPoint]+Parent2[CrossPoint:]; Where: Offspring represents the individual after crossover; Parent1 represents parent individual 1; CrossPoint represents the crossover point; Parent2 represents parent individual 2; Mutation operation: Randomly change the value of the gene position: Where: P m represents the mutation rate; Mutated Offspring[i] represents the individual position after mutation; Original GeneValue represents the original gene position value; Randomly Selected New Gene Value represents the randomly selected new gene position value; Calculate fitness: Calculate the fitness value of the individual according to the fitness function; Selection operation: Select a predetermined number of individuals from the parent and offspring populations, and then select the individuals with the best fitness as members of the next generation population.

Images