Fastener Damage Estimation in a Square Aluminum Plate

Abstract

To reduce costs and meet turn-around goals of future space vehicles, structural health monitoring systems are needed to assess the health of the entire vehicle structure within hours of the completed mission. In particular, the thermal protection system must be in good condition before launch due to its critical role in protecting the vehicle’s primary structure and subsystems. The capability to detect fastener failure is critical, as the overall thermal protection system for a vehicle will likely include a number of adjacent panels that are mechanically fastened to the substructure. This article discusses experimental and analytical efforts focused on estimating fastener damage in a square aluminum plate test article. Fastener condition is estimated using classifiers based on statistical pattern recognition methods. The classifiers utilize features from measured vibration data with the ability to discriminate between damage conditions. Finite element analyses are used to provide a physical understanding of the structural dynamics of the aluminum plate and an interpretation of the selected features used for the classifier. Techniques to detect, locate, and assess damage resulting due to fastener failure have been investigated. The classification systems presented perform reasonably well in detecting and localizing damage, particularly if the damage is fairly severe. However, the classification system performance is much worse at assessing the severity of damage, particularly for lower levels of damage.