Open Access Open Access  Restricted Access Subscription or Fee Access

Entropy-Based Damage Evaluation of Composite Structure and Its Application to Structural Health Management



Assessment of material state damage in a composite structure using fundamental thermodynamic concept of entropy has been proposed in this article. The health management system of composite structure is mainly based on installation of piezoelectric material on the structure itself. To supplement this existing system, in this article, it has been demonstrated that the change in entropy due to material degradation can be used as a measure of damage. The advantage of this entropy-based damage index measurement is that it offers a unified and more comprehensive indication of damage irrespective of the path of the damage mode. Since the dissipation of energy due to structural loading initiates damage from an ideal material-state, an estimation of material state disorganization from an unloaded-state can be considered as a change in entropy. Over the years, many attempts had been made to implement this concept in metallic structure, however, the application of this approach to composite structure is far from trivial. To circumvent this gap, an entropic characterization of fatigue degradation mechanism under various loading rates has been proposed and demonstrated for carbon fiber based composite structure using material state property information. Essentially, an experimental setup was designed to assess the change in structural integrity of the composite specimens at multiple fatigue cycles using the stateof- the-art scanning acoustic microscope. The statistical mechanics-based probability distribution function for the change in structural properties are then expressed in terms of fundamental entropy distribution and a damage index of the structure is formulated. Next, the specimens went through a time dependent stress relief cycle, and the entropy as a thermodynamic state function for the damage was assessed by using the change in probability distributions. Based on initial results, the proposed methodology provides enough proof to effectively identify the damage initiation irrespective of its type and the localization of the damage


Full Text: