Springback analysis in bending through finite element method based artificial neural networks

---

Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Makina Mühendisliği Bölümü, Türkiye

Tezin Onay Tarihi: 2013

Öğrenci: ÖZGÜ ŞENOL

Danışman: HALUK DARENDELİLER

Özet:

Springback prediction is vital in order to obtain the desired part shape in metal forming processes. In most of the applications, springback amount is determined by trial and error procedures, and recently by using numerical methods or through handbook tables. Artificial Neural Network (ANN) is a helpful tool for the engineers and applied in this study to determine the springback amounts in air, V-die and wipe bending processes. For this purpose, bending processes are analyzed by commercial finite element (FE) software and springback amounts are collected for different parameters such as thickness, die radius, bending angle, etc. Then, by developing a feedforward neural network with backpropagation learning algorithm, the springback amounts for bending applications are determined. ANN results of three bending operations are combined to analyze an industrial workpiece. In addition to this, an experimental bending operation is analyzed for air bending process. It is shown that ANN can be effectively applied to determine springback amount in air, V-die and wipe bending.