Estimate of Probability of Crack Detection from Service Difficulty Report Data.
-
1995-09-01
-
By Brewer, J.
Details:
-
Creators:
-
Corporate Creators:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Corporate Publisher:
-
Abstract:The initiation and growth of cracks in a fuselage lap joint were simulated. Stochastic distribution of crack initiation and rivet interference were included. The simulation also contained a simplified crack growth. Nominal crack growth behavior of large cracks was derived from the simulation results. These calculations implied that large cracks spend most of their lives as multiple small cracks and that the final growth and coalescence occur rapidly. The nominal crack growth histories were applied to 'C-check' crack detection data from the Service Difficulty Report (SDR) database to estimate the sizes of cracks when they were missed during previous inspections. These nondetection events were also estimated by assuming that each detected crack had always been a single crack. The results of each method were appropriately filtered and used to estimate probability of crack detection (POD) using the maximum likelihood technique. The POD estimates obtained from the single crack model are probably conservative, but the underlying assumption of slow crack growth (and therefore more inspection opportunities) might result in unconservative damage tolerance analyses.
-
Content Notes:Introduction, Crack growth simulation, Nominal crack growth histories, Analysis of data from service difficulty reports, Conclusions and recommendations, References, Appendices
-
Format:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type: