Semi-empirical life model of a cantilevered beam subject to random vibration

• Published in: International journal of fatigue Vol. 45; pp. 82 - 90
• Main Authors: Paulus, Mark, Dasgupta, Abhijit
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ltd 01.12.2012; Elsevier
• Subjects: Accelerated life; Accounting; Applied sciences; Cantilever beams; Constants; Exact sciences and technology; Fatigue; Fatigue (materials); Fatigue failure; Frequency shift; Life estimation; Mathematical models; Mechanical properties and methods of testing. Rheology. Fracture mechanics. Tribology; Metals. Metallurgy; Random vibration; Resonant frequency; Fatigue; Frequency shift; Accelerated life; Random vibration; Life estimation; Vibration; Mechanical properties; Modeling
• ISSN: 0142-1123; 1879-3452
• DOI: 10.1016/j.ijfatigue.2012.06.008

► Natural frequency change must be accounted for during random vibration. ► Natural frequency change is used to model damage accumulation. ► Life estimation can be performed using the damage accumulation rate. ► Modeling accuracy is improved by account for natural frequency change. Life estimation of structures subjected to non-uniform random vibration analysis has historically been done through the use of fatigue properties. In general, evaluation of the severity of the vibration environment is performed at the initial natural frequency. It is widely known, however, that many structures will undergo a frequency change during failure. If the random vibration is non-uniform with large peaks and valleys, the frequency change will result in a changing stress state that must be accounted for. Evaluation of the accumulated damage can be done through accounting for the natural frequency change. This model uses experimental data to determine two empirical constants to predict time to failure for various complex random vibration profiles. Although the new model will utilize an experimental component in lieu of FEA, the physics of the underlying failure will be retained. Additionally the model has shown good correlation to experimental data, and improvements over existing techniques.