Page 281 - Aircraft Stuctures for Engineering Student
P. 281

262  Airworthiness and airframe loads

                 would cause failure in Ni  cycles the number of cycles nj required to cause total fatigue
                 failure is given by

                                                                                    (8.38)


                 Although S-N  curves may be  readily obtained for different materials by testing a
                 large number of  small specimens (coupon  tests), it is not practicable to adopt the
                 same approach for  aircraft components since these  are expensive to manufacture
                 and the test programme too expensive to run for long periods of time. However,
                 such a programme was initiated in the early 1950s to test the wings and tailplanes
                 of Meteor and Mustang fighters. These were subjected to constant amplitude loading
                 until failure with different specimens being tested at different load levels. Stresses were
                 measured at points where fatigue was expected (and actually occurred) and S-N
                 curves plotted for the complete structure. The curves had the usual appearance and
                 at low stress levels had such large endurances that fatigue did not occur; thus a fatigue
                 limit existed. It was found that the average S-N  curve could be approximated to by
                 the equation

                                          Sa = 10.3( 1 + lOOO/dZ)                   (8.39)
                 in which the mean stress was 90 N/mm2. In general terms, Eq. (8.39) may be written as
                                            s, = S,(l+  C/dZ)                       (8.4)

                 in which S,  is the fatigue limit and  C is  a constant.  Thus Sa + S,  as N + cy).
                 Equation (8.40) may be rearranged to give the endurance directly, i.e.

                                                                                    (8.41)


                 which shows clearly that as Sa + S,,  N --+ cy).
                   It has been found experimentally that N is inversely proportional to the mean stress
                 as the latter varies in the region of  90N/mm2 while C  is  virtually constant. This
                 suggests a method of determining a ‘standard’ endurance curve (corresponding to a
                 mean stress level of  90N/mm2) from tests carried out on a few specimens at other
                 mean stress levels. Suppose S,  is the mean stress level, not 90 N/mm2, in tests carried
                 out on a few specimens at an alternating stress level Sa:,, where failure occurs at a
                 mean number of cycles N,.  Then assuming that the S-N  curve has the same form
                 as Eq. (8.40)
                                                                                    (8.42)

                 in which C = 1000 and Smqm is the fatigue limit stress corresponding to the mean
                 stress S,.  Rearranging Eq. (8.42) we have
                                                                                    (8.43)
                 The number of cycles to failure at a mean stress of 90 N/mm2 would have been, from
                 the above
                                                     Sm
                                                N  =-N,
                                                     90
   276   277   278   279   280   281   282   283   284   285   286