Development of artificial neural network based design tool for aircraft engine bolted flange connection subject to combined axial and moment load

---

Tezin Türü: Yüksek Lisans

Tezin Yürütüldüğü Kurum: Orta Doğu Teknik Üniversitesi, Mühendislik Fakültesi, Havacılık ve Uzay Mühendisliği Bölümü, Türkiye

Tezin Onay Tarihi: 2018

Öğrenci: TAHİR VOLKAN SANLI

Danışman: ALTAN KAYRAN

Özet:

In this thesis, a design tool using artificial neural network (ANN) is developed for the bolted flange connections, which enables the user to analyze typical aircraft engine connections subjected to combined axial and bending moment in a fast yet very accurate way. The neural network trained for the design tool uses the database generated by numerous finite element analyses for different combinations of parametric design variables of the bolted flange connection. The defined parameters are the number of bolts, the bolt size, the shaft thickness, the flange thickness, the pretension load acting on the bolt and the combined external axial force and bending moment. The outputs gathered from total 12000 FE analyses are the bolt reaction force and the average flange stress, which are collected to be used as database for the ANN training process together with the input design parameters. The results of the trained ANN are then compared with the FEA results. The comparison proves that the neural network shows great compliance with the non-linear FEA within the range of design parameters. As the last step, a graphical user interface is developed to turn the neural network into a user-friendly bolted flange design tool. It is believed that the developed design tool can replace the non-linear finite element analysis and be used very effectively in an optimization framework for the weight minimization of cylindrical bolted flange connections of aircraft engines.