Field and repair welding is a common practice in the power generation, oil and gas, and petrochemical industries. Such welds can lead to formation of martensite in the heat affected zone (HAZ) and the weld metal of steel welds and in the substrate HAZ of weld overlays. Temper bead welding is a technique, which can effectively temper the fresh martensite with subsequent passes, eliminating the need for post weld heat treatment (PWHT).
The objective of this work was to develop a method for prediction of the tempering efficiency in the HAZ of two-layer weld overlays (WOLs) using computational modeling and predictions validation through comparison with experimental WOLs. Three two-layer weld overlays of Alloy 625 on Grade 22 steel plates were produced using GTAW cold wire procedures. The heat input level was varied between low, medium, and high. The thermal histories from these welds were recorded using fourteen Type-K thermocouples, which were placed in two rows along the plate length. The acquired thermal histories at each heat input were then processed to determine the tempering efficiency and hardness at the thermocouple locations.
In parallel, FEA models were developed to replicate the experimental WOLs. The weld bead profiles were measured using a profilometer and implemented in the FEA model. Weld thermal histories at the thermocouple locations of the experimental WOLs were predicted using the developed FEA model with the SysWeldTM software.
In parallel with experimental work, a Design of Experiment (DoE) module was developed to simulate temper bead welding scenarios efficiently. The DoE module is capable of automatically meshing the weld beads and base plate, which allows for quick parameter adjustment and simulation. However, the DoE model assumes approximations such as trapezoidal weld beads and simplified mesh geometry, which could create inaccurate results. For this reason, both the FEA models that recreate the actual WOL geometry and the DoE module will be calibrated and validated through comparison with the extracted thermal histories from the welding experiments. Predicted weld thermal histories form the FEA models and DoE module will be used for hardness predictions and then compared with experimental hardness data to evaluate prediction accuracy. Once validated, the DoE module can be applied in extensive studies for evaluation of the welding parameters on tempering efficiency and development of optimized temperbead welding procedures.
The results of this study are relevant for multi-pass welding applications, where post-weld heat treatment is difficult or non-feasible. The computational DoE module can be applied for simulation of different welding scenarios and will allow users to perform multifactor process optimization and tempering efficiency evaluation of welding procedures.