| Summary: | 碩士 === 國立成功大學 === 航空太空工程學系碩博士班 === 96 === This research uses finite element method’s simulation and nano-indentation’s experiment parameters of loading and depth. It takes neural network’s method to induce inverse computational conclusion of Young’s modulus. In finite element method’s simulation, the problem is non-linear analysis. By computer analytic software of ANSYS, it simulates three different conditions, including plane stress model, plane strain model, and axial symmetric model. They are used to analyze the depth in different loadings. The contact effect must be considered in analytic process. Therefore, the indenter moves single axial as rigid target surface and the specimen is linear-elastic contact surface at simulation setting. In nano-indentation experiment, Center for Micro/Nano Science and Technology of National Cheng Kung University (N.C.K.U.) provides nano-indentation test. We must choose the standard specimen, fused silica, in test material and use the indenter which has Berkovich and conical shapes. During the process, we set boundary conditions and environmental factors. Then, the results get loading-depth curve at different shapes of indenter.
Finally, the literatures, the finite element simulation, and the experiment results demonstrate that they have analogical values. By neural network training, the results and the Young’s modulus compare each other and may improve nano-indentation’s research method.
Keywords：finite element method, nano-indentation, neural network, plane stress, plane strain, axial symmetric
|
|---|
