Damage detection of composite overwrapped pressure vessels using ARX models

Abstract

After the Columbia accident, aged composite overwrapped pressure vessels (COPVs) employed for space missions are being subjected to various examination and monitoring techniques to identify different failure modes. As an innovative approach, this paper explores the capability of Auto-Regressive model with eXogenous input (ARX) model analysis method to detect damages/flaws in COPVs utilizing the pressure as the input and the corresponding strains as the response data. Different levels of pressure were applied to the instrumented COPVs; both pressure and strains were simultaneously recorded, and later were analyzed to detect anomalies. Also investigated is the possibility of obtaining indices/features related to the performance of COPVs. It was discovered that varying strain measurements were recorded of damage induced COPVs while consistent strain measurements were recorded of intact COPVs. Furthermore, a Damage Feature (DF) was identified to compare the existing conditions of COPVs. A deviation of the DF values from an established threshold level was apparent for damage-induced COPVs while other COPVs were below the threshold level.