Decomposed-Coordinated Framework With Enhanced Extremum Kriging for Multicomponent Dynamic Probabilistic Failure Analyses

Abstract

For multicomponent structures enduring dynamic workloads coming from multi-physical fields, safety assessment is significant to guarantee the normal operation of entire structure system. In this paper, an enhanced extremum Kriging-based decomposed coordinated framework (E2K-DCF) is proposed to improve the dynamic probabilistic failure analyses of multicomponent structures. In this method, extremum Kriging model (EKM) is developed by introducing Kriging model into extremum response surface method (ERSM) to process the transient response problem and shorten computational burden in dynamic probabilistic failure analyses. Multiple population genetic algorithm (MPGA) is employed to solve maximum likelihood equation (MLE) and find the optimal hyperparameter θ in the EKM, which is promising to enhance approximate accuracy; decomposed-coordinated (DC) strategy is used to handle the coordinated relationship of multiple analytical objectives. To validate the proposed E2K-DCF, the probabilistic failure analysis of turbine blisk radial deformation is conducted by comparing with different methods within time domain [0 s, 215 s], considering fluid-thermal-structural interaction. It is revealed that the failure probability of blisk radial deformation is only 0.0022 when the allowable value is 2.5702 × 10 −3 m acquired from real world practice. Compared to the other approaches, this E2K-DCF has obvious advantages in fitting time and accuracy as well as simulation efficiency and accuracy. The results illustrate that the E2K-DCF is effective and applicable in dynamic probabilistic failure analysis. The efforts of this paper provide a novel viewpoint for the transient reliability evaluation of multicomponent structures, which is likely to enrich mechanical reliability theory. INDEX TERMS E2K-DCF, multicomponent structure, multiple population genetic algorithm, probabilistic failure, turbine blisk.

Reliability Optimization of Structural Deformation with Improved Support Vector Regression Model

“…e basis of an SVM model is statistical learning algorithm so that the SVM is suitable for small samples of structural design analysis, which are gained from a few FE simulations. e SR model is good at the solution of high nonlinear problems between input variables and output response, by introducing a maximum classification margin subject to inequality constraints [30]. Hence, the SR can improve the computational efficiency and accuracy of structural reliability optimization [9].…”

“…Comparing with the GA, the MPGA holds more flexible and adaptive design space exploration, which has the potential to avoid the effect of plateau-like function [34]. Besides, the MPGA breaks the limitation of a single population evolution of GA and uses many populations with different control parameters for optimization iterations [29,30]. Besides, the MPGA derived from GA inherits natural selection and genetic properties, and the optimal solution of the objective function can be gained via enough iterations with selection, crossover, and mutation.…”

Adaptive Local Maximum-Entropy Surrogate Model and Its Application to Turbine Disk Reliability Analysis

“…The performance of an aero-engine is determined by numerous factors, such as its geometric dimensions, material properties, and environments. In the process of the equipment design and manufacturing, these factors will inevitably experience certain random changes, which lead to security risks in the deterministic design [1][2][3][4]. The traditional method is to set a safety factor to consider these random uncertainties, which often results in an increased structural weight.…”

Multi-Surrogate Collaboration Approach for Creep-Fatigue Reliability Assessment of Turbine Rotor

“…Under such circumstances, as valuable alternatives to direct Monte Carlo simulation, surrogate model methods were developed to lessen excess simulation tasks and widely employed into reliability assessment, sensitivity analysis and probabilistic design [27]- [30]. Typical surrogate models involve polynomial response surface [31], [32], support vector regression [33], [34], Kriging surrogate (KS) [35], [36] and neural network surrogate (NNS) [37]- [39]. By integrating flexible network topology and strong nonlinear fitting ability, NNS holds the potentials to improve the computational efficiency in performing complex reliability assessment issues.…”