Abstract
The problem of identifying intergranular stress corrosion cracking (IGSCC) in 26-inch diameter, 1. 3 inch thick austenitic stainless steel type 304 pipe is discussed in this paper. A computer based pattern recognition approach is utilized for classifying IGSCC's versus fabrication anomolies (geometric reflectors). The simplified pattern recognition approach utilized for classification purposes yields an 89. 3% overall index of performance, with 94% correct classification of IGSCC's. The approach is practical enough to be implemented in the field in hardware as an add-on device to existing flaw detectors.
Original language | English (US) |
---|---|
Specialist publication | Solid State Technology |
State | Published - Jan 1 1979 |
All Science Journal Classification (ASJC) codes
- Electronic, Optical and Magnetic Materials
- Condensed Matter Physics
- Electrical and Electronic Engineering
- Materials Chemistry