Bayesian Reliability Assessment of Non-Destructive Evaluation (BRANDE)

ABSTRACT: Aircraft structural components are prone to damage typically assessed by various NDE tools. Damage to the aircraft structural components decreases operational readiness and compromises safety if improperly maintained; however, frequent inspections and maintenance is expensive. CBM+ equipped platforms, when supported by models and simulations typically experience an increase in operational readiness in the five percent range. CBM+ requires accurate quantitative damage descriptors with the addition of statistical metrics to feed risk assessments. MTC proposes to develop a software tool for Bayesian Reliability Assessment of NDE systems (BRANDE). BRANDE detects different aspects of damage characterization with an embedded statistical framework to decrease maintenance workloads, increase fleet readiness, and decrease safety risk by accurately representing confidence in quantitative NDE (QNDE). BRANDE will employ a generic statistical approach for quantitatively assessing the reliability of NDE systems considering critical controlling factors. Two basic problems in QNDE are considered: damage detection, where the performance (capacity) of NDE systems is represented by the Probability of Detection (POD) and False Call Rate (FCR), and damage sizing/localization, where the performance (accuracy) of NDE systems is represented by the Error of Estimation (EOE). Correspondingly, the proposed BRANDE, using Bayesian methods, solves the NDE assessment problem by providing a set of useful statistical metrics to quantify the variations/uncertainties in POD/FCR and EOE which depends on controlling factors. BENEFIT: Applications of BRANDE for both Military and Commercial applications are numerous. BRANDE can be readily applied to applications in structural health monitoring to make maintenance scheduling decisions. For example, Health and Usage Monitoring Systems would benefit directly from the proposed software developments as wireless systems to investigate cracks and loose bolts in critical structural components are embedded into the helicopter. BRANDE is applicable civil infrastructure where an embedded NDE tool could be compromised during a catastrophic event or where the size of global damage is to be predicted from local measurements on particular structures. The statistical framework can be used to evaluate multiple additional military applications and even the automotive industry.