Pros and Cons of Proof Testing Carbon Composite Overwrapped Pressure Vessels: A Comparison of Two Mathematical Models

Abstract

Proof testing of carbon/epoxy composite overwrapped pressure vessels (COPVs) is often viewed as a process of weeding out weak vessels, thus improving future reliability. Unlike in metal vessels, however, considerable damage is done in terms of fiber breakage. This is clear from acoustic emission and is expected based on data on Weibull fiber strength, which allows estimation of the number of fiber breaks to expect at a given proof pressure. Thus if a proof pressure level is too high, long term reliability may degrade rather than improve. Classical reliability models are not based on fiber breakage processes that capture the above drawbacks. In fact, when applied to carbon/epoxy COPVs, the classic power law-Weibull version predicts that the higher the proof test level, the greater the reliability benefit. Here we develop a modelcalled the stochastic fiber breakage model, that explicitly accounts for micromechanical and statistical fiber failure processes in the overwrap. We find that any long term benefits of a proof test depend critically on the proof pressure level. Examples are given that are motivated by certain pressure vessels that are currently in service. These examples are based on data for types of carbon fiber (T1000G and AS4) in an epoxy matrix, and for comparison purposes results are also given for Kevlar 49 fiber. Under a proof pressure levels above 0.70 the proof test is predicted to degrade long-term reliability in the case of carbon fibers, but not in the case of Kevlar 49 fibers.