Optimization Of Curing Cycle Of CFWRP With Combination Of Artificial Neural Network And Genetic Algorithm

The carbon fiber wrapped reinforced plastics (CFWRP) have been widely applied in aerospace, marine transportation, building, medicine and so on, because of its outstanding characteristic including high specific tenacity and high specific modulus, anti-corrosive and so on.Whether the wrapped products to obtain the ideal design strength, It not only depended on material itself performance and the proportion of fiber and resin, but also had an important relations with the resin cure process. Traditionally the cure cycle was gained by trial-and-error, the model optimization or the on-line monitoring method, however they were either not adapt to varied environment or high cost and operation difficultly. Thus the thesis brought forward a new method that established model using neural network and optimized it by genetic algorithm, the result indicated that it was feasible.The researches object of this thesis was wet wrapped NOL-rings with carbon fiber and epoxy resin. The curing temperatures was gotten by the DSC experiments, the elevation of temperature rate was kept from 0.5℃/min to 1℃/min, the curing preservative time was arranged using the orthogonal tables of three factors and three levels ,and then we wrapped and cured NOL-rings under 1kg wrapping strain and 23% glue content process conditions. Following we measured its tensile strength and degree of cure after forming. Then we dealt the experimental results using orthogonal analysis. Meanwhile we divided the results into two parts: training samples and testing samples.We respectively established two BP neural network models between the tensile strength, degree of cure and curing time using the Ann toolbox of Matlab 6.5, then trained and tested models with training samples and testing samples. On the base of models, we got an optimization result through optimizing a double objective function using genetic algorithm with floating point encoded. Compare with the orthogonal design result: we cut down the curing time 1h and enhanced the tensile strength at the same time. Finally we analyzed each preservative time level influencing tensile strength through the contoured block diagram; the result...