When characterizing the NDI capability for detecting cracks, the natural metric for measuring crack damage was the linear crack dimension used in damage tolerance analyses.  The selection of the appropriate metric for corrosion damage, however, is not immediately apparent.  There are different types of corrosion damage and different metrics can be used to quantify the damages.  For example, in hidden corrosion in lap joints and doublers on fuselage structures there are several possible metrics: thickness loss, pit depth and/or frequency, surface roughness, and joint pillowing.  When inspecting for intergranular and exfoliation corrosion, around fasteners, useful metrics might be the maximum radial distance that the corrosion extends from the fastener hole or the corrosion area about each fastener.  In some sense, each metric plays a role in the effect that the corrosion defect has on the structure.  Consequently, it is important to consider all of the metrics for a given application.  Each corrosion type must be considered separately, but the important aspect of the metric is that it measures corrosion severity.  Ideally, the metric should be based on an “effect of defects” study; however, in practice the important metrics are generally known, and, in order to keep the assessment focused, it becomes necessary to select only one metric at a time for detection assessment.  If it is absolutely essential to include an evaluation of more than one metric, then multiple evaluations must be performed (one evaluation per metric).

There is a necessary relation between the corrosion metric and the NDI technique.  Obviously, the NDI technique must be responsive to changes in the corrosion damage metric.  For example, in inspecting for hidden corrosion in lap joints, eddy current is responsive to thickness loss but may not be sufficiently responsive to pit depth.  If pit depth is a critical parameter, a different NDI technique would be needed.