Statistical Flaw Detection: Application to Flaws Below Curved Surfaces

摘要: The detection of the presence flaws in structural materials is most important function which Non-Destructive Evaluation (NDE) performs. As structures are designed to higher performance criteria and as safety life cycle cost factors become more important, it becomes necessary detect smaller difficult find flaws. This paper presents a practical approach optimum noise signals. A decision theoretic (described detail companion by Fertig, et al.1) used derive algorithm adapted environment particular measurement being made. An automatic procedure for characterizing noises developing presented. Two implementations this have been tested on experimental data show substantial improvement over conventional techniques. One flexible research purposes, other real-time suitable field implementation.