首页|A deep learning unified model for predicting the residual stiffness of plain weave composites under combined high and low cycle fatigue loading
A deep learning unified model for predicting the residual stiffness of plain weave composites under combined high and low cycle fatigue loading
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
Elsevier
The study of the fatigue stiffness degradation behavior of carbon fiber composites under combined high and low cycle fatigue (CCF) loading is essential for the design of aviation structures subjected to cyclic loads. In this paper, an experimental study of plain weave composites (PWCs) is performed under designed CCF loading spectra, indicating that the superimposed high-cycle fatigue load significantly accelerates the damage accumulation process and leads to a more pronounced fatigue stiffness degradation phenomenon. An improved latent variable dynamic evolution learning strategy combined with the attention mechanism is proposed to capture the nonlinear dynamic characteristics of fatigue stiffness degradation and a deep learning unified model (DLU model) for residual stiffness prediction is developed. The experimental data from uniaxial and biaxial CCF loading are utilized to demonstrate the superior performance of the DLU model. The results show that residual stiffness can be accurately predicted using limited observational data. Moreover, by disentangling the latent variable, complete extrapolation of residual stiffness is achieved through latent variable analysis. This paper provides a novel method for overcoming the challenge of residual stiffness monitoring and prediction under complex fatigue loading conditions.
Plain weave compositesCombined high and low cycle fatigue loadingDeep learningResidual stiffness prediction
NETL
NSTL
Elsevier
