论文部分内容阅读
SiC -CVD过程的本质复杂性制约了在碳 /碳复合材料表面有效地制备高性能抗氧化涂层 .本研究在对CVD工艺过程及机理试验研究的基础上 ,采用人工神经网络技术对SiC -CVD过程进行了辨识与模拟研究 ,建立了SiC -CVD过程的神经网络结构模型 ,并根据CVD过程复杂、工艺性强等特点 ,从样本选取、网络结构设计、学习参数调整等方面对神经网络学习算法进行了改进 .结果表明 :模型对工艺参数影响规律的预测结果与工艺实验结果相一致 ,所开发的神经网络模型 ,不仅可以对各种实验条件下的沉积结果进行准确的预测 ,还能更为全面地反映不同工艺因素对沉积规律的不同影响和判断各工艺因素间的交互作用的存在及作用大小 ,同时采用SiC -CVD模型可对沉积机理的时间效应进行预测和分析 .
The inherent complexity of SiC-CVD process restricts the efficient preparation of high-performance anti-oxidation coatings on the surface of carbon / carbon composites.Based on the experimental study of CVD process and mechanism, the artificial neural network The process of the CVD process is identified and simulated. The neural network structure model of the SiC-CVD process is established. According to the characteristics of the CVD process, such as complex process and strong technology, the neural network is studied from the aspects of sample selection, network structure design and learning parameter adjustment Algorithm is improved.The results show that the predicted results of the influence of the model on the process parameters are consistent with the experimental results, and the neural network model developed can not only accurately predict the deposition results under various experimental conditions, In order to comprehensively reflect the different influence of different process factors on the depositional law and determine the existence and role of the interaction between various process factors, the time effect of deposition mechanism can be predicted and analyzed by using SiC -CVD model.