A method development about fati̇gue damage estimation at metallic structures using PSD technique

The concept of fatigue is very important in industrial applications such as automotive, defense, avionics, aerospace, machine industry and it is become prominent in all fields of our lives. The products produced are expected to be durable and long-lasting against the loads they are exposed to under working conditions. Design criteria are also developed in this direction. There are three main factors that cause fatigue. The first factor that can cause fatigue is the geometry of the product. They are geometric details that directly affect fatigue such as the notch effect, including thin-form designs or sudden section contractions. The second factor is the mechanical and fatigue properties of the material used. The third factor is the type of loading that the product is subjected to, such as fixed, regular (stationary) loading or random (non-stationary) loading. By taking these three factors into consideration, the damage and corresponding life values of the products are obtained by using appropriate fatigue calculation methods. In this study, fatigue calculations of the sample were performed as a result of random vibration loading applied to the metallic test specimen which has notched geometry. Respectively, modal analysis, random vibration analysis and then fatigue calculations were performed numerically by using finite element software for the notched test specimen which is made of aluminum and steel material properties. Then, code was created in Python software by using contain formulations of Dirlik method which is one of the damage calculation methods for vibration fatigue. The damage and life values were calculated using the frequency-dependent PSD stress output obtained as a result of random vibration analysis with the finite element method or the frequency-dependent PSD stress output obtained from the test as well. For the verification of numerical and analytical calculations, it has been compared with the physical test result previously performed.