Estimation of Mechanical Properties of PAN Nanofibers Based on Polymeric Structural Characteristics by Artificial Intelligence Modeling

Introduction: Today, nanofibers are commonly used in filtration, composites, tissue engineering, drug delivery systems and many scientific and industrial applications. Here, investigating of nanofiber mechanical properties is important. Measuring mechanical properties of thin nanofiber is very difficult, time consuming and expensive. In this research, mechanical properties of nanofibers have been studied based on their structural characteristics. Method: From the presented experiments, polymeric structural parameters and mechanical properties of parallel PAN nanofibers were measured for 150 samples in five categories of electrospinning conditions. After that, adaptive back propagation neural network was designed and optimized by genetic algorithm for experimental data. Result: The results presented 0.89% and 0.006% for test and train errors which were acceptable for mechanical properties estimation. Conclusion: The presented intelligent modeling method can be an accurate choice for mechanical properties estimation of nanofibers especially, where the experimental measuring is difficult or unavailable. Also, sensitivity test presented that distance between crystal in L1020 and polymeric crystal size had more effect on the strength of the nanofibers.