Multi-Scale Synergistic Topology Optimization Of Continuous Fiber Composite Panel Members

As a scientific and efficient structural design method,topology optimization can calculate the material distribution region that satisfies the design constraints and has the optimal target performance in a given design domain,and is widely used in many fields.However,the traditional structural topology optimization design based on homogeneous materials can no longer meet the needs of modern industrial products for ultra-lightweight structure,functional specialization and performance integration.How to give full play to the potential of material microstructure in improving product performance and realize multi-scale structural design considering both "material macroscopic distribution" and "enhanced fiber phase structure" has become a hot spot and difficult area in the field of structural topology optimization.For the structural optimization design problem of three-dimensional orthogonal woven composite panel components,a heuristic optimization algorithm structural multi-scale optimization strategy based on agent model response prediction is proposed in this paper.First,the homogenization method is used to obtain the equivalent properties of representative volume units at the fine scale and microscopic scale of the composite,and the macroscopic model of the composite is given homogenized material properties using a stepwise scaling method from the microscopic scale to the macroscopic scale;the variable density method applicable to orthotropic anisotropic materials is introduced to optimize the structural topology of the homogenized composite.The results show that using this method can systematically consider the effects on the macroscopic topology brought by the structural changes at multiple scales of continuous fiber composite sheets.Secondly,the topological flexibility response values under different fiber bundle weaving process parameters were fitted using Kriging model,and the root mean square error of the predicted values was obtained to judge the accuracy of the fitted model,and the minimum flexibility influence trend of the fiber weaving parameters on the optimal structure was obtained according to the predicted response surface.The parameter values of the sample points were obtained using the Latin hypercube sampling method,which ensures the randomness and comprehensiveness of sampling.Finally,a BP neural network improved by genetic algorithm was used as a proxy model to train the response values of the sample points;among them,the node threshold and interlayer node weight values of the BP neural network were obtained using the warp and weft yarn spacing and the warp and weft yarn bundle K bundle as input parameters and the minimum flexibility value under the optimal structure as response parameters,which were used as a prediction method for the flexibility response values;combined with the adaptive genetic algorithm to find the The optimal fiber weaving parameters under the constraint conditions are obtained,which can minimize the flexibility response of the composite material after the optimization of the macrostructure topology and reduce the flexibility of the structure by 67.2% compared with the average value of the response of the randomly sampled samples;and the simulation verification analysis is done to prove the effectiveness of the method.For the structural optimization design of continuous curvilinear laminated fiber composite sheet members,two coupling methods are proposed and analyzed in this paper:first,based on the improved particle swarm optimization method and parametric level set method synergistic optimization strategy;second,based on the unit maximum principal stress method and parametric level set method synergistic optimization strategy.First,the level set topology optimization method for planar anisotropic materials is constructed;the fiber unit orientation angle within the structure is dynamically adjusted using the particle swarm optimization algorithm and the maximum principal stress method,respectively,so that the discrete unit fiber angle is consistent with the direction of the maximum unit stress;the results show that both methods can obtain lower structural flexibility compared with unidirectional composites,in which the optimal The optimal structural flexibility obtained by the maximum principal stress method is 28.8% lower than that of the unidirectional composites,while the particle swarm method is only about 10% lower;and the optimal fiber placement paths parallel to the macrostructure boundary and clear are obtained by the maximum principal stress method.