The development of powerful commercial computer programs made the concept of J-integral as computational parameter of fracture mechanics to be a very attractive one. It is equivalent to SIF in linear case, it converges in numerical calculation and the same results are obtained by different codes (programs). Besides, it is widely thought that elastic-plastic analysis gives bigger values than elastic SIF ones what is good from regulatory point of view. Such stand was reflected in the recommended by IAEA TECDOC 1627 (February 2010) devoted to pressurized thermal shock analysis of reactor pressure vessels, where the embedded crack in FEM mesh, elastic-plastic analysis with simultaneous determination of J-integral was stated as the best option of analysis.
But at that time all the most widely used software were not able to treat the residual stresses, the thermal stresses in case of two different materials. Such a contradiction between requirements and the possibilities made a lot of problems for honest contractors especially in countries where the regulator had no own experience in calculation and completely relied on the authority of international documents. This means that at that time the said recommendations were harmful. The main reason of such a situation was the absence of the carefully elaborated examples.
Now the capabilities and accuracy of such software are increasing. Nevertheless, some principal ambiguities and divergences of computations results in various J-integral contours around the crack tip still exist. They are exhibited when the large plastic zone emerges at the crack tip. Other problem is influence of the history of loading and the specification of the time of crack insertion in the mesh including the time of emergence of residual stress.
This paper is invitation for discussion of the accuracy and restriction of computational J-integral. With this aim the detailed analysis of some simplified 2D examples of calculation of elastic -plastic J-integral for surface crack with accounting for residual stress, thermal stress and inner pressure are performed and commented.
The attainment of consensus among the engineering society for treating the outcome results is the prerequisite for practical application of computational elastic plastic J-integral.