TG1: Single Bead-on-Plate Weldment (Stainless Steel)
Task Group 1 has studied the residual stress field around a short TIG weld bead deposited on the surface of a grade 316 austenitic stainless steel plate (180 mm long, 120 mm wide, and 17 mm thick) by conducting measurement and numerical round robins in parallel.
From 2002 to 2010, fourteen independent sets of residual stress measurements were made using five different techniques, and subjected to detailed statistical analysis to evaluate best estimate profiles for validation of finite element predictions. In parallel, over forty finite element simulations were performed, examining a wide range of solution variables and their impact on the predicted residual stresses.
This apparently simple geometry proved to be very challenging, both for experimentalists and analysts.
Fig. 1. TG1 – illustration of simulated plate distortions
The numerical round robin was run in two phases:
-
Phase 1 with unconstrained thermal efficiencies
-
Phase 2 using a fixed 75% efficiency and advanced material hardening models
This large body of work has allowed NeT Task Group 1 to develop a reliable thermal solution strategy based on a global weld heat input calibrated against far-field thermocouples followed by derivation of detailed weld heat source characteristics by matching the weld fusion boundary profile. The accurate thermal solutions then allow the most important mechanical solution variables to be isolated and optimised. An important variable is found to be the material hardening model, with mixed isotropic-kinematic hardening being the most accurate for the AISI 316L plate material. Other solution variables, such as the welding efficiency, the mesh design and the thermal boundary conditions, are found to be of much less importance.
The following figures illustrate how significantly the agreement between measurements and simulations improved by the selection of an appropriate thermal solution and the most suitable material hardening model.
Fig. 2. TG1 – through thickness distribution of transverse residual stresses, simulation results compared to Bayesian mean of measurements
Fig. 3. TG1 – macrograph of weld cross section vs. simulated temperature distribution during welding
Lessons from TG1
-
Measure the welding thermal efficiency
-
Thermocouple all specimens
-
Shield thermocouples against electrical interference from the welding torch
-
Record temperatures at high enough frequency (>5 Hz)
-
Make more use of strain gauges
-
Specify measurement locations using prior knowledge of the global stress field
-
Specify and control reference specimens
-
Record what you measure and quantify measurement uncertainties
-
The actual heat input must be modelled for this weld type
-
Measure and model cyclic hardening behaviour of both the weld and parent materials for accurate stress predictions
The lessons learned from TG1 have proven to be invaluable assets for the definition of the activities in the later Task Groups, such as TG4, TG5, and TG6.
Upon request, TG1 material could be made available to external users to serve them as a benchmark for their own finite element simulation and/or residual stress measurement procedures.
Network on Neutron Techniques Standardization for Structural Integrity
