Wei Bai-Jian, Chuang Yao-Chen, Wang Kai-Hong, Yao Yuan
Department of Chemical Engineering, National Tsing Hua University, Hsinchu 30013, Taiwan.
Polymers (Basel). 2016 Sep 8;8(9):337. doi: 10.3390/polym8090337.
Resin transfer molding (RTM) is a popular manufacturing technique that produces fiber reinforced polymer (FRP) composites. In this paper, a model-assisted flow front control system is developed based on real-time estimation of permeability/porosity ratio using the information acquired by a visualization system. In the proposed control system, a radial basis function (RBF) network meta-model is utilized to predict the position of the future flow front by inputting the injection pressure, the current position of flow front, and the estimated ratio. By conducting optimization based on the meta-model, the value of injection pressure to be implemented at each step is obtained. Moreover, a cascade control structure is established to further improve the control performance. Experiments show that the developed system successfully enhances the performance of flow front control in RTM. Especially, the cascade structure makes the control system robust to model mismatch.
树脂传递模塑成型（RTM）是一种用于生产纤维增强聚合物（FRP）复合材料的常用制造技术。本文基于利用可视化系统获取的信息对渗透率/孔隙率比进行实时估计，开发了一种模型辅助的流动前沿控制系统。在所提出的控制系统中，通过输入注射压力、流动前沿的当前位置以及估计的比率，利用径向基函数（RBF）网络元模型来预测未来流动前沿的位置。通过基于元模型进行优化，可获得在每个步骤要施加的注射压力值。此外，建立了一种级联控制结构以进一步提高控制性能。实验表明，所开发的系统成功提高了RTM中流动前沿控制的性能。特别是，级联结构使控制系统对模型失配具有鲁棒性。
