Fuzzy system model for isothermal chemical vapor infiltration process of carbon/carbon composites

Isothermal chemical vapor infiltration (CVI) is an important conventional technique for the fabrication of ceramic and carbon matrix composites. The major disadvantage of isothermal CVI process is that the extremely long processing time is required. Therefore the key point of research on the isothermal CVI technique is to optimize process parameters and to raise the deposition efficiency. Based on the experimental results, the fuzzy rules were automatically obtained and optimized by genetic algorithm, thus the fuzzy system for isothermal CVI of carbon/carbon composites was established. Through the analysis of the outputs of training and testing samples, we could draw a conclusion that this system is characterized by higher precision and good generalization ability. With the help of this system, the regulations affecting infiltration temperature, fiber content and chamber pressure in isothermal CVI were analyzed. The results show that this system has a good instructive function for the formulating of CVI processes in practical application.