The statistical properties of fracture toughness test results from weld metal and heat affected zone of structural steel developed for arctic applications are investigated. Methodology and requirements to testing and statistical treatment of the results are proposed to achieve robust estimates of the characteristic values that can be used in design or engineering critical assessment. The results indicate that the SINTAP procedure provides robust, conservative, but sometimes overly conservative estimates of the lower tail of the statistical distribution. As the lower bound fracture toughness appears to be independent of constraint, it is proposed to use high constraint specimens for qualification testing of smaller data sets as the likelihood of revealing lower bound values is higher. Weld thermal simulation confirms the temperature dependence in accordance with the master curve concept in the heat affected zone.
Qualification of structural steel and welding procedures for low temperature applications require fracture mechanics testing. Requirements are implemented in several standards including ISO 19902. The draft ISO standard for arctic structures refers to ISO 19902 for fracture toughness requirements. The acceptance criteria for fracture toughness are normally not given in design standards. Reference is normally made to an engineering critical assessment (ECA) such as BS 7910 or API RP 579. In this paper, methods for determination of characteristic toughness values for use in ECA are investigated. Requirements to testing and reporting test data are proposed in order to provide consistent and statistically significant results. On this basis, reliability based calibration can be performed to establish safety factors to be used in design and ECA The main data source for the quantitative analysis is the R&D project “Arctic Materials”, funded by The Research Council of Norway, oil companies, material suppliers and offshore industry. The project is executed by SINTEF.