: One of the commonly encountered fracture problems in pressure vessels and piping systems is concerned with longitudinal inner surface-cracks in cylinders subjected to pressure. Such flaws could be the resultant due to manufacturing difficulties (e.g. tool marks) or material defects. The presence of such cracks might lead to reduce the safe operating lifetime of pressure vessels considerably, and in this connection, incidents of many pressure vessels have been traced.
    Thus, it has recently been recognized that the fatigue crack growth behavior of such cracks is one of the important subjects for the design of a pressure vessel and the reliable prediction of the operating lifetime. However, few data have been obtained on this behavior in thick-walled cylinders.
    In this study, fatigue tests exposing the inner surfaces of cylinders to cyclic pressure were carried out to clarify the fatigue crack growth behavior. The material tested is a nickel chromium molybdenum steel, whose mechanical properties are listed in the Table I. Cylindrical specimens are with a small surface notch, prepared by electric-discharging method and the ratio of the outer diameter to the inner is 2.4. Pressure ratio is varied in the range of 0.04 - 0.60. The crack depth was estimated from the surface crack length, measuring the aspect ratios (the ratio of depth to surface length) of beach marks left on the fracture surface and crack propagation rates were determined as the gradient of crack depth (a) to the number of loading cycles (N) curve. Crack closure behavior of the specimens was detected by ultrasonic method.
    The results of this study are summarized as follows;
  (1) In the regions where configurations of the initial notches didn't influence on the aspect ratios, those were about 0.4 and the fatigue crack propagation was predominant in striation formulation mechanism.
  (2) By using stress intensity factor range, ΔK as represented in equation (2) or (6), the relation between crack propagation rate in the direction of depth and ΔK was roughly consistent with that for through-cracks.
  (3) To avoid the influence of pressure ratio and evaluate the crack propagation rates more exactly, it was desired to use, instead of ΔK, effective stress intensity range, ΔKKetf, based on the crack closure level.