• Version
• 259 Downloads
• 2.18 MB File Size
• 1 File Count
• March 7, 2017 Create Date
• March 7, 2017 Last Updated

Correlation of Fatigue Data for Aluminum Aircraft Wing and Tail Structures

S-N curves are derived for aluminum wing and tail structures by 
fitting various regression models to. 246 full-scale constant-amplitude fatigue 
test results from twelve types of aircraft structures. The derived curves were 
tested by comparing the predicted lives with actual. test results of various 
aircraft structures fatigue tested to variable-amplitude loads spectra. More 
reliable predictions resulted from these derived S-N curves than from existing 
S-N curves.

1.0 INTRODUCTION 
This document establishes more consistent S-N curves for the fatigue life prediction of alu- 
minum aircraft wing structures than the presently available RAeS-ESDU S-N curves for "Typical Wings 
and Tailplanes" (Ref. 1). 
Constant-amplitude fatigue test results from 246 aircraft wings and tails from twelve aircraft 
types were pooled and various regression models were fitted in order to obtain mathematical expressions 
for sets of S.N curves. 
These curves and the existing RAeS.ESDU curves were then used to predict the lives Of 
variable-amplitude tests and the calculated lives were compared with the actual test results. 
In the life calculations the method Of linear cumulative dam$.ge (Palmgen-Miner Rule) was 
utilized. 
2.0 s-N CURVES DERIVED FROM FULL-SCALE STRUCTURES 
When estimating the life of a wing structure when structural details and an extensive stress 
analysis of the wing are not available it is normally better to use S-N curves derived from built-up struc• 
tures than those derived from notched material coupons. Reasons for this include the fact that there 
is no fretting in a simple specimen and a simple specimen has only one load path whereas many struc- 
tures are redundant. 
RAeS•ESDU published a set Of S-N curves derived from 137 full-scale fatigue tests on wings 
and tailplanes of various aluminum alloy aircraft structures (Ref. 1). These data and curves are re- 
produced in Figures 1 and 2. 
Upon inspection of Figure 2 it is noticed that the curves do not follow the same general 
trend. It is easily seen that the line for five ksi mean stress does not follow the pattern of the lines for 
the other mean stresses. It is understood that the curves were drawn by "eye" and to illustrate this 
inconsistency the slopes and intercepts Of the lines were plotted.