Structural Reliability Analysis And Optimization Design Based On The Hybrid Intelligent Algorithm

As the rapid development of modern science, technology and economy, theproblems of structural reliability and safety, in engineering structure design, more andmore occupies an important position. The structural reliability analysis and optimizationdesign is a method to obtain the optimal design variable, which define the structuralreliability index as the objective function or constraint condition.Compared with the traditional structure design method, this method can get bettereconomic benefit, so its study has a very important significance.However, in the real engineering structures, there are many uncertainty factors, andthe nonlinear degree is higher between the input and response, so the method of thetraditional reliability analysis and optimization design can’t meet the need of theengineering structures, and sometimes make the result not ideal and even not correct.In this paper, the study for the structural reliability analysis, optimization designmethod and the intelligent optimization algorithm was carried out. The probabilisticreliability optimization model of the structural system based on the PSO-DE hybridalgorithm, the non-probabilistic reliability model based on the PSO-DE hybridalgorithm, the non-probabilistic reliability optimization model based on the PSO-DEhybrid algorithm, the non-probabilistic reliability model based on the SAPSO-DEhybrid algorithm and the non-probabilistic reliability optimization model based on theSAPSO-DE hybrid algorithm were proposed.The main contents and some conclusions are briefed as below:1. The intelligent optimization algorithm is studied. Considering the convergenceproperty of the particle swarm optimization (PSO) and the differential evolution (DE), aPSO-DE hybrid algorithm based on the evolution of cognitive experience was proposed.Numerical examples show that the proposed hybrid algorithm overcomes the premature phenomenon of the PSO algorithm, the convergence speed is fast and the calculationaccuracy is higher, the correctness and feasibility of the hybrid algorithm in engineeringapplication are verified.2. The probabilistic reliability optimization design of the structural system wasstudied. The probabilistic reliability optimization design problem is a nested loopproblem. The outer loop is to optimize the objective function and the inner loop is tooptimize the reliability analysis. In the process of reliability analysis, the probabilisticreliability index was calculated by the improved first order second moment method; inthe process of reliability optimization, applying the PSO-DE hybrid algorithm, theprobabilistic reliability optimization design of the structural system was carried out. Forthe minimization problems of the structure mass under the constraints of the structuralsystem probabilistic reliability index, the probabilistic reliability optimization designmethod of the structural system based on the PSO-DE hybrid algorithm is proposed.Numerical examples show that this method has a good stability, is higher in precisionand easy to achieve, and overcomes the premature phenomenon of the PSO algorithm.3. For the structural non-probabilistic reliability analysis and optimization designproblems, the structural non-probabilistic reliability analysis and optimization designusing the non-probabilistic reliability model was carried out. In this paper, acomputational method of non-probabilistic reliability index using the convex modelsmethod and the PSO-DE hybrid algorithm was proposed. And using the amendediterative algorithm, combined with the PSO-DE hybrid algorithm and thenon-probabilistic reliability optimization model, the method of structuralnon-probabilistic reliability optimization design was proposed. Through the numericalexamples, the correctness and feasibility of suggested structural non-probabilisticreliability analysis and optimization design method were verified4. In order to get higher calculation accuracy and convergence properties, aSAPSO-DE hybrid algorithm based on combination of the simulated annealing-particleswarm optimization algorithm and differential evolution is proposed. Applying thishybrid algorithm to the design of the structural reliability analysis and optimization, thenon-probabilistic reliability model was established, the methods of structural non-probabilistic reliability analysis and optimization design based on the SAPSO-DEhybrid algorithm were proposed. Through the numerical examples, the validity andfeasibility of suggested method of the structural non-probabilistic reliability analysisand optimization design were verified.