【摘要】 为了降低固化过程中复合材料结构内部温度场和固化度场的非均匀性,提出一种基于径向基(RBF)神经网络的固化均匀性优化方法。通过拉丁超立方试验设计方法选取设计变量(升温速度、保温温度及保温时间)的样本点,采用固化反应的参数化有限元分析模型对样本点进行计算以获取响应(温度梯度、固化度梯度及总的固化时间),以此建立固化反应的RBF神经网络,最后利用主要目标法和多岛遗传算法优化该网络以求得最优解。AS4/3501-6复合材料算例的优化结果表明:对比原设计,优化后温度梯度和固化度梯度的最大值分别降低了71.34%和51.47%,总的固化时间仅增加4.71%,优化效果显著。更多还原

【Abstract】 In order to reduce the inhomogeneity of temperature field and curing degree field in the composite structure during curing, this paper proposes a curing uniformity optimization method based on radial basis function(RBF) neural network. It chooses the design variables(temperature, heat preservation temperature and holding time) of the sample points by Latin hypercube experimental design method, obtains sample points response(temperature gradient, the curing degree of the gradient and the total cure time) by calculating the curing reaction of parametric finite element analysis model, and establishes curing reaction of RBF neural network based on this. Finally, this paper uses the method of main goal and more genetic algorithm to optimize the network to ensure the optimal solution. The optimization results of AS4/3501-6 composite materials show that, compared with the original design, the maximum values of the temperature gradient and the curing degree gradient after optimization are reduced by 71.34% and 51.47% respectively, and the total curing time is only increased by 4.71%, indicats a significant optimization effect.更多还原