139

Peter Martin1, Nathan Switzner1, Joel Anderson1, Emily Brady2, Pooya Delshad3, Peter Veloo3, USA.2, USA.3

1RSI Pipeline Solutions, New Albany, USA. 2Exponent, Houston, USA. 3Pacific Gas and Electric Company, San Ramon, USA

This work builds on previous efforts to use nondestructive composition data to mitigate outlier yield strength results during Instrumented Indentation Testing (IIT).1 It is widely known that IIT yield strengths are generally within ±10 % of the associated tensile test (TT) values, but that some ‘outlier’ results can deviate by 25 % or more. The fundamental cause of these outlier results is unclear, making it difficult to anticipate the accuracy of a given measured result. However, an empirical association between these outliers and the steel composition was used to develop an algorithm that improves the accuracy of strength estimates based on nondestructive measurements of the Mn and C concentrations. The current work expands on the prior report by: i) increasing the number of pipes used to validate the algorithm, ii) using a statistical method to explore the robustness of the algorithm, and iii) evaluating the impact of the adjustments when the strength values are used as inputs to a probabilistic grade calculator.

For a set of 89 pipe joints, the adjustment decreases the number of IIT-measured yield strengths deviating by more than 10 % from the corresponding TT values from 32 to 13 joints (from 36 % to 15 %). The largest adjustments occur in pipes where the IIT-measured YS values exhibit the largest deviations relative to the TT values, and where the ratio of Mn/C exceeds 10. This is interpreted as an indication that the adjustment works most effectively on modern, high-grade pipe since this type of pipe is expected to have a high Mn/C ratio. Validation of the model shows the algorithm is robust and likely to be broadly applicable. Finally, the use of adjusted strength results as inputs to a probabilistic grade calculator yields improved accuracy of grade estimates relative to the use of un-adjusted (as-measured) IIT results. Potential implementation strategies will be discussed, along with the applicability to other strength measurement methods.

1. M Scales, et al., (2022) Materials, 15, 832.

Read/Download PaperGradeAvailable after the event