Abstract In-service welding of pipelines without interrupting the internal gas flow is nowadays state of the art for the maintenance and extension of the grid. However, due to access restrictions, it is quite challenging or even impossible to measure local temperatures during welding and subsequent cooling. To address this issue, in the present work welding experiments were performed that consider representative weld configurations and process parameters relevant for in-service welding of gas pipelines. A flat steel sheet was multi-pass welded onto the outer surface of a straight steel pipe segment by manual shielded metal arc welding (SMAW) or by manual gas tungsten arc welding (GTAW). Thermocouples were used for measuring local temperatures at defined positions next to the weld seam. The influences of joint type, wall thickness of the sheet and pipe and heat input on the temperature distribution and, in particular, on the local peak temperatures and cooling rates were studied. Moreover, a finite element (FE) model of a selected weld configuration was created and validated for calculating the temperatures at positions that were inaccessible for direct measurements. The results of the measurements showed that the peak temperature at the inner surface of the pipe decreases with increasing wall thickness. However, independently from the wall thickness, the results of the simulations indicated austenitization next to the weld seam. The low cooling rates did not promote any martensite transformation, as confirmed by the moderate increase of the hardness and by microstructure characterization. The results are of practical interest for pipelines used to transport hydrogen gas, as the temperature-dependent composition of the microstructure and the temperature itself influence the absorption, diffusion and solubility of hydrogen in steel.
Kaiser et al. (Tue,) studied this question.