A D V A N C E D M A T E R I A L S & P R O C E S S E S | S E P T E M B E R 2 0 1 8 2 6 A NONCOMPREHENSIVE GUIDE TO NONDESTRUCTIVE EVALUATION This article is an excerpt from the completely revised ASM Handbook, Volume 17, Nondestructive Evaluation of Materials, just published this month. Samuel W. Glass III Pacific Northwest National Laboratory, Richland, Wash. T he most common application of nondestructive evaluation (NDE) is detection of flaws caused by man- ufacturing anomalies, service or envi- ronmental stresses, or natural material aging. More generally, applications may include estimation of mechanical and material properties, stress/strain, and dy- namic response behavior. NDE comprises a large family of specific test disciplines including visual inspection, dimension- al metrology, ultrasound, radiography, penetrant tests, magnetic particle tests, leak tests, eddy current tests, potential drop tests, flash and vibrothermography, shearography, acoustic emissions, and many other methods. This article offers an overview of NDE science and discuss- es some key considerations for choosing specific techniques. FLAW DETECTION AND EVALUATION Before an NDE method is cho- sen, the inspection objective should be clearly defined. Considerations include: • Reasons for performing the NDE (failure prevention, performance en- hancement, end-of-life prediction, quality control) • Types of flaws or material charac- teristics of interest (fatigue cracks, stress corrosion cracks, creep, pit- ting, erosion, embrittlement, wear, planar cracks/voids, transverse/ axial cracks, color variation, density, resonant frequency) • Size and orientation of rejectable/ reportable flaws • Anticipated locations of flaws of interest (surface, volumetric, welds, heat affected zones, high stress points, areas subject to wear) • Material characteristics (hardness, toughness, density, strength) • Size and environmental, temporal, and spatial accessibility of test component Typical inspection techniques examine a particular metric or sig- nal against a threshold level. Wheth- er the signal is a simple indicator level or a complex multidimensional image, the basic concept of flaw detection is that the signal exceeds a detection threshold. Quantification of this aspect of NDE is known as the probability of de- tection (POD)  . The detection threshold must be set sufficiently low to ensure detection of significant flaws, yet not so low that noise or normal variations in the signal frequently exceed the thresh- old and increase the probability of false alarms (PFA) (Fig. 1). Realistic expecta- tions for any NDE technique appreciate that there is some threshold of flaw size belowwhich it is unlikely for the inspec- tion to detect degradation. Converse- ly, there is a signal level at which the chance of flaw detection is very high. NDE methods are designed to allow de- tection thresholds to be set tomaximize POD while minimizing PFA. Factors affecting POD include signal sensitivity to degradations of Fig. 1 — Probability of detection (POD) and probability of false alarm (PFA) are determined by fractions of signal and noise distributions above a threshold. Signal distribution generally shifts to higher levels as flaw size increases, leading to the sigmoidal POD curve.