Research On The Closure Of Fiber Orientation Tensor And Prediction Of Mechanical Performance For The Fiber Reinforced Polymer Matrix Composites

Short-carbon-fiber reinforced injection-moulded composites are very attractive in aviation,aerospace and other high-tech fields because of their superior mechanical properties.As multiphase materials,short-carbon-fiber reinforced injection-moulded polyetheretherketone composites have obvious microstructure characteristics;the mechanical properties change significantly with selected different components and parameters of microstructure and show significant anisotropy.Injection molding is one of the important processes of short-carbon-fiber reinforced composites,it has the advantages of high efficiency,excellent stability and excellent molding flexibility as opposed to other processes and can be applied to produce those plastics with complex shapes.It is of great significance for the design and application of such high-performance injection-moulded parts to predict the distribution of fiber orientation and the mechanical properties.This paper focuses on the fiber orientation tensor closure during injection molding processing of short-fiber reinforced composites and prediction of mechanical performance for the short-fiber reinforced polymer matrix composites,and a serial of researches on solving fiber orientation distribution function and developing an efficient and accurate closure have been carried out.And the main works are as follows.1.To construct fitting fiber orientation tensor closure,it is necessary to solve the Fokker-Planck equation to obtain the fiber orientation probability distribution function and then calculate the fiber orientation tensor according to the definition.In this paper,a finite volume numerical method based on flux conservation is used to solve the Fokker-Planck equation.By solving the simple flow field,the combined flow field and the center gate disk flow field,it is shown that the numerical solution method has better numerical stability.2.To overcome the premature convergence during Kriging model adaptive iterative process in efficient global optimization(EGO)algorithm,the improved EGO and more efficient convergence criteria are proposed.According to the engineering optimization problem paying more attention to the optimal solution,in each iteration of reconstruction of Kriging model,the optimal design of the recently constructed Kriging model is added to sample points and update Kriging model if it is better than the previous optimum,otherwise,the design which maximizes the expected improvement function is added to sample points.Baed on the dividing and conquering strategy,a parallel efficient global optimization,named subregional efficient global optimization(SEGO)is proposed.the SEGO algorithm attempts to divide the domain of every continuous variable into two equal subregions and calculates the expected improvement(El)function in each subregion.To tackle the Pareto optimum of injection process parameters for multi-objective optimization of the quality of plastic part,Non-dominated sorting-based genetic algorithm ?(NSGA-?),the improved EGO and SEGO are used to find a much better spread of design solutions and better convergence near the true Pareto optimal front.A cover of liquid crystal display part is optimized to show the method.The results show that the Pareto fronts obtained by NSGA-? are distributed uniformly,and this algorithm has good convergence and robustness.3.Based on the existing fiber orientation tensor closure,we firstly numerically solved the probability distribution function of simple flow fields by the finite volume method to obtain the second order orientation tensor and the fourth order orientation tensor in the principal axis system.Secondly,based on uniform distribution of two principal values of the second order orientation tensor,we selected six flow fields to approximate the fourth order orientation tensor in terms of the second order orientation tensor,and firstly give the surface plot of contracted fourth-order tensor components versus second-order tensor components in the principal axes.The maximization of expected improvement adds flow fields to improve the accuracy of the closure.Results demonstrate that the EGO-based closure outperforms others closure in simple shear flow,and the EGO-based closure and the natural(NAT)closure give good performance in the combined flow field.However,EGO-based closure shows nonphysical oscillations in the center-gated disk flow.One more flow state is added to the training data set,thereby eliminating nonphysical oscillation and increasing the accuracy of the closure.4.The effective properties of Poly(lactic acid)reinforced by the randomly distributed halloysite nanotubes(HNTs)were studied using a novel numerical implementation of asymptotic homogenization method.The effects of different morphologies and the two different boundary conditions were evaluated and were found to be not of significance for the results of numerical simulation.Compared with other numerical methods,good agreement was found between the novel numerical implementation of asymptotic homogenization method and the experimental data.5.Based on the experimental analysis,the multi-scale analysis and calculation model of the particle-reinforced polymer nanocomposites was constructed by using the theory of asymptotic homogenization from the microstructure characteristics of the material.Numerical simulation results show that the novel numerical implementation of asymptotic homogenization(NIAH)method is more reasonable and accurate for simulating the micro-mechanical properties of periodic composite materials.With increase of the number of the unit cell particles,the results of NIAH method converge to a stable value.Based on the results obtained,it can be concluded that spatial distribution for the two kinds of particles in the unit cell has a significant impact on the material macroscopic elastic properties.