Material failure and plastic instability are currently in ASME and EN pressure vessel and piping standards prevented by applying experience-based safety factors. Due to the few existing experimental test results and the inaccuracy of the incorporated design equations, these safety factors are quite high. In particular, modern high strength steels are penalized because of the small yield-to-tensile ratio. Apart from monotonic loading conditions, an economical design of pressurized components to prevent material failure due to ultra-low-cycle fatigue (ULCF) is a big topic.
Experimental tests are in general very expensive and not suitable for the common design. Therefore, a numerical based prediction of the actual burst pressure and the resistance against strong cyclic loading conditions would be favourable. During the last decades innovative damage mechanics concepts have been developed and successfully validated in the scope of structural and plant engineering. Depending on the size of the component and the applied loading condition, several different damage mechanics models are available.
During the studies presented in this paper, a phenomenological damage mechanics model proposed by Bai and Wierzbicki has been utilized. It has been developed to describe the material failure in the upper shelf of the toughness temperature transition curve in case of monotonic loading conditions. Considering an extension based on the effective strain concept proposed by Ohata and Toyoda, the combined model is also able to predict failure in the range of ULCF.
The presented work aims at validating the accurate numerical prediction of failure by using this damage mechanics model. Validation is performed by comparing numerical and experimental results of a burst test of a pressure vessel and cyclic deformation tests of almost constant pressurized bended pipes. The pressure vessel made of HSLA steel P690Q had a length of 3.0 m, a diameter of 1.2 m and a wall thickness of 50 mm. Tests on bended pipes can be separated in two series. A first series made of steel X60 with a diameter of 16 “ and a wall thickness of 9.5 mm as well as a second series made of steel X65 with a diameter of 8.625 “ and a wall thickness of 5.6 mm.
The comparison of experimental and numerical results shows an acutely satisfying prediction of failure. Both, time point and location of failure coincide well. Mesh size dependency must be considered but the approach seems to be promising for further applications.