论文部分内容阅读
建立了基于遗传算法和误差反传(GA-BP)神经网络的化学气相渗透(CVI)工艺参数优化模型。以新型等温CVI工艺制备C/C复合材料时采集的实验数据作为模型评价样本,分析了主要可控影响因素(沉积温度、前驱气体分压与滞留时间等)对C/C复合材料制件密度及其密度均匀性的作用规律。在该模型指导下,样本的期望密度和实测密度最大误差不超过6.2%,密度差最大误差不超过8.2%。实验结果也证明了该模型具有较高的精度和良好的泛化能力,可以用于CVI工艺参数的优化。
A chemical vapor infiltration (CVI) process parameters optimization model based on genetic algorithm and error back propagation (GA-BP) neural network was established. The experimental data collected from the CVI composites prepared by the isothermal CVI process were used as the model evaluation samples. The effects of the main controllable factors (deposition temperature, partial pressure and residence time of precursor gas) on the density of C / C composites And the law of its density uniformity. Under the guidance of the model, the maximum error of the expected density and the measured density of the sample does not exceed 6.2%, and the maximum error of the density difference does not exceed 8.2%. The experimental results also prove that the model has high precision and good generalization ability, which can be used to optimize the CVI process parameters.