Uncertainty-Based Warpage Prediction And Optimization Technology For Fiber-Reinforced Composite Components

Composite materials are widely used in the aerospace field due to their excellent performance,but the warpage of the components due to the curing process is difficult to predict,which limits their further promotion.The reason is that uncertainty factors such as material properties,geometric measurements,and boundary loads during the curing process will cause the warpage of the component to have certain parameter randomness.Therefore,it is necessary to characterize the uncertainty that affects the final deformation during the curing manufacturing process,and to predict and control the optimization of the curing warpage deformation.For the general idea of uncertainty analysis based on "characterization-analysisoptimization",the article proposes a general framework of curing deformation prediction control based on parameter uncertainty factors.Starting from the geometric randomness of the components to the randomness of performance,the two uncertainties are characterized,quantified and propagated under the probability framework.This article first describes the design and manufacturing process of the composite material,modeling from the perspective of macro design,meso-layout and micro fiber,respectively,and completes the uncertainty characterization of geometric parameters based on the probability model.Prediction of the warpage of composite materials is accomplished through Bayes' theorem to quantify the propagation of parameter uncertainty.Based on the constitutive relationship of the curing and deformation of the composite material,the prediction model determines the influence relationship between its internal parameters and establishes a standardized hierarchical Bayesian network.In this paper,MCMC hybrid sampling is used to complete the prior correction of geometric parameters,and a hierarchical Bayes model is constructed to complete the prior determination of physical parameters.After Bayesian inference is completed,the physical posterior and geometric parameter posterior are used together to realize the uncertainty propagation calculation with the help of the parameterized finite element model,which verifies the feasibility of the warpage prediction under the influence of uncertainty.The optimization of warpage deformation of composite materials is based on the response surface model,and the adaptive genetic algorithm is used to solve the model.Metamodel theory is used to describe the curing process,and the whole curing process is decomposed into two meta-models of temperature field and strain field,which together form the overall response surface.The response surface is used to replace the original finite element calculation model,and combined with MC sampling to complete the uncertainty propagation calculation.After the global optimization model is constructed based on the response surface model,considering the high-dimensional characteristics of the response surface-based objective function,the cross-mutation probability of the genetic algorithm and the population number are adaptively reduced.This method effectively improves the convergence of the algorithm and verifies the effectiveness of the optimization model.An optimized software system for curing deformation prediction of composite materials was developed.The software system uses MATLAB as the algorithm core and realizes finite element calculation with the help of Abaqus solver.The software realizes the integration of probabilistic characterization of uncertain parameters,warpage prediction based on Bayesian theory and deformation control optimization of response surface proxy model.Finally,the wing skin laminates are used as examples to verify the correctness of the warpage prediction control analysis.