Premium
An improved back computation procedure for the parting‐out step of a destructive method for measuring residual stresses in pipes
Author(s) -
SHEALY W. S.,
SHADLEY J. R.,
RYBICKI E. F.
Publication year - 1984
Publication title -
strain
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.477
H-Index - 47
eISSN - 1475-1305
pISSN - 0039-2103
DOI - 10.1111/j.1475-1305.1984.tb00541.x
Subject(s) - residual stress , computation , strain gauge , structural engineering , residual , curvature , gauge (firearms) , line (geometry) , stress (linguistics) , welding , dimension (graph theory) , coupon , materials science , mechanics , engineering , mathematics , geometry , composite material , algorithm , physics , metallurgy , linguistics , philosophy , pure mathematics , finance , economics
One method for measuring through the thickness residual stress distributions in welded pipes involves a destructive, three step laboratory procedure consisting of parting out, splitting, and layer removal operations. In the parting out step, a coupon of material is removed from the wall of the pipe while strain gauges on the inner and outer surfaces are monitored. In back computing the through the thickness residual stresses from the parting‐out strain gauge data, it is generally assumed that stress changes through the thickness of the coupon vary as a straight line. If the circumferential dimension of the coupon is large compared to the thickness, this assumption is valid. However, in the case of a pipe, the circumferential dimension is usually limited to reduce curvature effects during the subsequent splitting and layer removal steps. If this dimension is too small, the assumption of straight line stress changes through the thickness can lead to serious inaccuracies. In this paper, the validity of the straight line assumption is examined over a range of conditions. To handle cases where the straight line assumption is not valid, a modified back computation procedure for the analysis of parting‐out strain gauge data is developed and presented with numerical results.