CN-122021329-A - RBF-enhanced quadratic fit equation-PSO-based cable-stayed bridge power optimization method, electronic equipment and storage medium

CN122021329ACN 122021329 ACN122021329 ACN 122021329ACN-122021329-A

Abstract

The invention belongs to the technical field of bridge seismic resistance, and particularly relates to a cable-stayed bridge power optimization method based on RBF enhanced quadratic fit equation-PSO, which comprises the steps of establishing a finite element model of a cable-stayed bridge, generating samples based on a Box-Behnken design method, determining factor levels and test index values, modifying the finite element model according to factor levels of different test samples, loading earthquake vibration to obtain a response value, constructing an RBF neural network, randomly generating a supplementary sample, predicting the response value of the supplementary data by using the trained RBF neural network, and re-fitting a quadratic equation by using a weighted least square method. After constructing an objective function, carrying out iterative optimization by adopting a particle swarm algorithm to obtain a Pareto optimal solution set, modifying a finite element model by using the factor combination of the optimal solution, and then comparing the applied earthquake motion with the earthquake motion response of the original finite element model. The method can effectively solve the problems that the number of test sets required in bridge power optimization is too large, the number of fitting model data by adopting a quadratic equation is small, the weight required by multi-objective optimization is required, and the like.

Inventors

HU DONGYANG
LI WANXIONG
YAO ZHIGUO
LI HONGBO
LI CHONGJIE
YANG YEXIN
ZHOU XIAOZHUANG
LI CHUANLIN
LV LEI
LU SANPING
CHE WENQING
LIANG JUNXIONG

Assignees

中铁二院昆明勘察设计研究院有限责任公司

Dates

Publication Date: 20260512
Application Date: 20260205

Claims (8)

1. A cable-stayed bridge power optimization method based on RBF enhanced quadratic fitting equation-PSO is characterized by comprising the following steps: Modeling a cable-stayed bridge by finite element software, and selecting a matched earthquake motion input based on key parameters of site conditions and structural dynamic characteristics of the bridge; step 2, determining reasonable influence factors as design variables, and designing a Box-Behnken method by taking proper dynamic response as an index value test scheme of a test; step 3, constructing a BXF neural network, training the BXF neural network, and obtaining a trained bridge response prediction model; Step 4, randomly generating a supplementary sample in the parameter range of the original data, predicting the response value of the supplementary data by using the trained BXF neural network, combining the original data and the supplementary data generated by the BXF neural network, fitting a quadratic equation corresponding to each index value by using a least square method, and analyzing and checking whether the quadratic equation meets the requirement; Step 5, constructing two quadratic equations based on the precision requirement, and combining the two quadratic equations to form a cable-stayed bridge power optimization target equation set; solving the model by adopting a particle swarm algorithm to obtain a Pareto optimal solution set, and further extracting three design variable values corresponding to index combinations which enable the dynamic response of the cable-stayed bridge to be optimal; And 6, carrying out parameter optimization on the finite element model of the cable-stayed bridge according to the designed variable value corresponding to the optimal dynamic response index of the cable-stayed bridge obtained in the step 5, and then applying seismic wave input on the optimized model to finally obtain a seismic response time chart of the cable-stayed bridge.
2. The method for optimizing power of a cable-stayed bridge based on RBF reinforced quadratic fit equation-PSO according to claim 1, wherein the step 2 comprises the following specific steps: Step 201, the inclined angle of a guy cable of a cable-stayed bridge is influenced by the height h t of a cable-free area of a tower root of a main girder, the length a of the cable-free area of a main girder and the length a 1 of a span-neutral cable-free area of the main girder, and the three parameters are used as design variables; step 202, determining the value range and the horizontal dividing value of three factors according to the general design rule of the cable-stayed bridge, and executing step 203; Step 203. The box-Behnken test design method generates different numbers of test groups according to the number of factors and the level of the factors, and the different factor levels are marked as follows according to the factor value and the factor level of step 2.2 、、 Wherein h t,i is the value of the ith level h t , the other two parameters are the same, the ith level of h t , the ith level and the 1 ith level are obtained by adopting a Box-Behnken test design method to obtain different test groups ; Step 204, corresponding factor level of each test group at step 203 Removing the finite element model to obtain a finite element model corresponding to each test group, selecting a piece of earthquake motion obtained in the step 1, and applying the earthquake motion to different finite element models to obtain a bridge dynamic response value of the index value corresponding to each test group 、 W 1j represents the value of W 1 in the j-th combination, and W 2j represents the value of W 2 in the j-th combination.
3. The method for optimizing power of a cable-stayed bridge based on RBF enhanced quadratic fit equation-PSO according to claim 2, wherein in step 3, BXF neural network is constructed to different test groups in step 203 The value is used as the input of the BXF neural network model to 、 And the BXF neural network model is used as the output of the BXF neural network model, so that the BXF neural network is trained, and a trained bridge response prediction model is obtained.
4. The method for optimizing power of a cable-stayed bridge based on RBF reinforced quadratic fit equation-PSO according to claim 3, wherein the step 3 comprises the following specific steps: Step 301, respectively establishing 、 An independent RBF network, wherein the expansion coefficient is S, the training target error G and the maximum neuron number are M; Step 302, calculating 、、 The average value of each input is recorded as 、、 , , , Represents the mean value of h t 、a、a 1 under the ith combination to obtain the input value of the neural network For inputting training.
5. The method for optimizing power of a cable-stayed bridge based on RBF reinforced quadratic fit equation-PSO according to claim 4, wherein said step 4 comprises the specific steps of: Step 401, less samples are divided in step 203, and supplementary samples are randomly generated in the parameter range of the original data, a supplementary data point number set A is determined, and the supplementary samples A are repeated in step 3 to obtain a response value B of BXF neural network predicted supplementary data; Step 402, connecting Data set A and corresponding response value 、 Summarizing response value B, fitting the quadratic equation of W 1 and W 2 by using a least square method, and analyzing complex correlation coefficients of the quadratic equation If the requirements are met, the next step is carried out.
6. The method for optimizing power of a cable-stayed bridge based on RBF reinforced quadratic fit equation-PSO according to claim 5, wherein said step 5 comprises the specific steps of: Step 501, initializing maximum iteration number Defining the current iteration number as t, and initializing t=1; calculating the velocity of particle k in the t-th iteration using equation (1) And position : (1) Wherein, the The inertia coefficient is defined as S and, 、 For learning coefficients of T, U, 、 Is a random number between 0 and 1, For a historic optimal position of the particle k, Is a global optimal position; Step 502, combining to form a cable-stayed bridge power optimization objective equation set, Setting the iteration number as L, searching K particles, and carrying out iterative optimization by adopting a formula (1) to obtain a Pareto optimal solution set.
7. An electronic device comprising a memory and a processor, wherein the memory is configured to store a program that supports the processor to perform the cable-stayed bridge power optimization method of any of claims 1-6, the processor being configured to execute the program stored in the memory.
8. A computer readable storage medium having stored thereon a computer program, characterized in that the computer program when run by a processor performs the steps of the cable-stayed bridge power optimization method according to any one of claims 1-6.

Description

RBF-enhanced quadratic fit equation-PSO-based cable-stayed bridge power optimization method, electronic equipment and storage medium Technical Field The invention belongs to the technical field of bridge seismic resistance, and particularly relates to a cable-stayed bridge power optimization method, electronic equipment and a storage medium based on RBF reinforced quadratic fitting equation-PSO. Background Bridges are visually known as throats and hubs of transportation as key structures across obstacles. Once the bridge is damaged in an earthquake, the transportation of materials, the evacuation of people and emergency rescue are seriously hindered, and the social operation and disaster response are affected deeply. Among a plurality of large-span bridge forms, the cable-stayed bridge has emerged as a mainstream choice of the large-span bridge worldwide due to the characteristics of beautiful shape, reasonable structural stress, excellent economic performance and the like. In order to ensure the normal working state of the cable-stayed bridge and reduce the damage caused by earthquake, power optimization research on the cable-stayed bridge is necessary. The dynamic optimization of the bridge structure is a leading edge design concept and key technical method in the field of modern bridge engineering. The method breaks through the limitation of mainly focusing on static strength and rigidity in the traditional design, and actively and systematically performs collaborative adjustment and optimization search on mass distribution, rigidity configuration, damping characteristics and even overall morphology of the bridge in the scheme design stage from the essence of structural dynamic performance. The method is characterized in that by combining a mathematical optimization algorithm with refined structural power time course analysis, on the premise of meeting multiple constraints such as safety, economy, construction feasibility and the like, an optimal design scheme is intelligently found, so that the vibration response, fatigue damage and overall stability of the bridge under dynamic environments such as wind load, earthquake action, vehicle excitation, crowd activities and the like are in a comprehensive optimal state. Common optimization methods include orthogonal test design, latin hypercube sampling, response surface method and the like. The orthogonal design method can greatly reduce the test times, but can not explore the interaction of factors, the horizontal number of the factors is fixed, the flexibility is limited, the sample generation of Latin hypercube sampling is dependent on randomness, the sample number is fixed and is difficult to dynamically expand, the response surface method can analyze the interaction and the main utility, the horizontal number of the factors is a section instead of a fixed value, the gradient optimization and the multi-objective optimization are supported, but when the response surface method is adopted for carrying out the optimization design, the weight corresponding to each index value is needed to be known, and the weight is difficult to calculate for some unconventional index values. Disclosure of Invention The invention aims to provide a power optimization method, electronic equipment and a storage medium for a cable-stayed bridge based on RBF enhanced quadratic fit equation-PSO, so as to realize power optimization of the cable-stayed bridge, thereby effectively solving the problems of more test sets required in optimization design, less data of a quadratic fit model and weight required by multi-objective optimization. Compared with Latin hypercube sampling, the invention does not depend on randomness in sample generation, and is divided by small samples, compared with common quadratic polynomial, the invention adopts neural network prediction to enlarge the reasonable range of the samples, and compared with response surface method, the invention does not need the weight value of individual index. The technical scheme adopted by the invention is as follows: A cable-stayed bridge power optimization method based on RBF enhanced quadratic fit equation-PSO comprises the following steps: modeling a cable-stayed bridge by utilizing finite element software, comprehensively considering key parameters such as site conditions and structural dynamic characteristics of the bridge, and selecting a matched earthquake motion input; step 2, determining reasonable influence factors as design variables, and designing a Box-Behnken method by taking proper dynamic response as an index value test scheme of a test; The step 2 comprises the following specific steps: Step 201, the inclined angle of a guy cable of a cable-stayed bridge is influenced by the height h t of a cable-free area of a tower root of a main girder, the length a of the cable-free area of a main girder and the length a 1 of a span-neutral cable-free area of the main girder, and the three parameters are used as design v