A lamb wave approach for fatigue damage prognosis with piezoelectric transducers

Abstract

Fatigue is a progressive and localized damage that occurs when a material is subjected to cyclic loading. Historical cases have shown that undetected fatigue cracks often leads to catastrophic failure, including the loss of lives and assets. It is therefore important to have a robust non-destructive evaluation (NDE) technique to detect and monitor these cracks. The field Structural Health Monitoring (SHM) evolved in search for a robust system for damage detection and monitoring. As SHM develops, the need to be able to detect and monitor a damage no longer suffices. Being able to predict the system's remaining useful life arises as the next big challenge. This leads to the conceptualization of Damage Prognosis (DP). This research aims to expand the present capabilities of the Lamb wave propagation (LWP) technique for structural health monitoring (SHM) and damage prognosis (DP) and to develop a DP model to predict the growth of fatigue crack and subsequently the remaining useful life of material or structure in a narrow plate environment. A novel concept of an equivalent imaginary source (EIS) is proposed to model the displacement and strain field of reflected Lamb waves. This EIS incorporates information on the reflection coefficient and phase delay of reflected waves. The concept of EIS is then incorporated into an existing model for infinite plate to extend the model for voltage output of PZT transducers in a narrow plate. Using the extended model, a new damage index is proposed based on the change in peak amplitude of the sensor signals. Numerical simulations and experimental studies are conducted to verify the above model and damage index. Fatigue is stochastic in nature and there are numerous uncertainties in terms of the materials and damage detection techniques. Monte Carlo simulation is adopted to manage the uncertainties. An iterative framework is proposed to estimate the remaining life of a component when information on the fatigue properties of the material is not available. The error in prognosis reduces by adopting the proposed iterative framework. Damage prognosis is a multi-disciplinary and challenging problem. Thus, collective effort from the various field of research is required to solve this problem. The author hopes that the originality and contribution from this research will benefit the community and spur research in the field of damage prognosis.