AK-GWO: a novel hybrid optimization method for accurate optimum hierarchical stiffened shells

Engineering with Computers - An attempt has been made to propose a novel prediction model based on the Gaussian process regression (GPR) approach. The proposed GPR was used to predict blast-induced...     

Gradient tree boosting machine learning on predicting the failure modes of the RC panels under impact loads

This paper proposed a new approach in predicting the local damage of reinforced concrete (RC) panels under impact loading using gradient boosting machine learning (GBML), one of the most powerful techniques in machine learning. A number of experimental data on the impact test of RC panels were collected for training and testing of the proposed model. With the lack of test data due to the high cost and complexity of the structural behavior of the panel under impact loading, it was a challenge to predict the failure mode accurately. To overcome this challenge, this study proposed a machine-learning model that uses a robust technique to solve the problem with a minimal amount of resources. Although the accuracy of the prediction result was not as high as expected due to the lack of data and the unbalance experimental output features, this paper provided a new approach that may alternatively replace the conventional method in predicting the failure mode of RC panel under impact loading. This approach is also expected to be widely used for predicting the structural behavior of component and structures under complex and extreme loads.

     

Multi-extremum-modified response basis model for nonlinear response prediction of dynamic turbine blisk

For the nonlinear dynamic analyses of complex mechanical components, it is necessary to apply efficient modeling framework to reduce computational burden. The accurate surrogate model for approximating the nonlinear responses of several failures is a vital issue to provide robust and safe design conditions in complex engineering applications. In this paper, two different Modified multi-extremum Response Surface basis Models (MRSM) are proposed for dynamic nonlinear responses of failure capacities for turbine blisk responses. The proposed MRSM is established using two regression processes including regressed the input variables by linear or exponential basis functions in first calibrating phase and regressed the second-order polynomial basis function using inputs data provided by first stage in second calibrating procedure. A sensitivity analysis using MRSM is proposed to consider the variation of input variables on the nonlinear responses. In the sensitivity analysis procedure, the effects of input variables are evaluated using the calibrating results given from the first regressed process. To evaluate the performance of the proposed MRSM, three multi-extremum failure modes including radial deformation of compressor blisk, maximum strain, and stress of compressor blade and disk are considered. the prediction of MRSM of nonlinear responses for Thermal-fluid–structure system with dynamical nonlinear finite-element analyses is compared with response surface method (RSM) and artificial neural network (ANN). The predicted results of modeling approaches showed that the sensitivity analysis based on MRSM accurately provided the effective degree for input variables. The gas temperature has the highest effects on nonlinear responses of turbine blisk which is followed by angular speed and material density. The MRSM combined with basic exponential function performs better than other models, while the MRSM coupled with linear function is more accurate than ANN and RSM. The proposed MRSM models have illustrated the accurate and efficient framework for approximating dynamic structural analysis of complex components.