The Use of Field Inspection Data in the Performance Measurement of Nondestructive Inspections

摘要

Inspection reliability is one of the corner stones of the safety by inspection approach for continuing airworthiness of aging aircraft and of the damage tolerance philosophy. Inspection reliability data, usually in the form of technique threshold data and Probability of Detection (POD) data, are essential for deriving inspection thresholds and inspection intervals. Most nondestructive inspection (NDI) reliability data available is from dedicated round-robin inspection programs. These data have been frequently challenged on the basis that the inspection conditions are not representative in terms of environment, access and human factors of the conditions seen in service. The North Atlantic Treaty Organisation (NATO) Research and Technology Organisation (RTO) Applied Vehicle Technology (AVT) Panel sponsored a Workshop Quantification Of Airframe Inspection Reliability Under Field Conditions in Brussels in May 1998 to address these issues. The AVT-051 Working Group was formed after this meeting to further investigate, and this paper presents a summary of the Working Group report. A fundamental flaw in using in-service inspection data alone to estimate the POD of an applied inspection technique was identified. The process only uses the 'missed' data associated with detected cracks. Essentially, in addition to the 'misses' identified through the process of 'back-growth' for detected cracks, there is another population of 'misses' for cracks that have not yet been detected. This non-conservatism leads to the conclusion that 'hit'/'miss' populations derived from detected cracks alone are insufficient to provide a useable POD for the technique.