This presentation was made at the NAFEMS UK Seminar "Welding Simlation - How Much Validation is Enough?"

Resource Abstract

Welding is a complex manufacturing process that connects structural elements under the influence of heat, mostly metals. Undesirable side effects develop during the weld process that might negatively affect the future in-service performance of the connected elements. These side effects are mostly studied using a combination of experience and expert judgement, which forms the basis of most modern design and fabrication standards. Poor (historic) welding procedures can lead to costly repairs on new build structures and structural integrity issues on existing structures.



The application of a numerical framework to study welding processes has been initiated from the second half of the previous century. The increasing computer power could lead to an accelerated usage of welding simulations which would then results in a better understanding of welding processes and a more optimum usage of the material. Although the computer power increases, the wider application of welding processes is not observed yet. The complexity of the simulations, excessive need for accurate material data and the need for model validation holds back a wider application of welding simulations in practice in our view.



Arup has gone through the drill of model validation using the numerical software LS-DYNA, experimental test data and the material package JMatPro. A relatively simple welding specimen has been studied extensively to gain trust and confidence in the numerical modelling approach. This has resulted in a conference paper that demonstrated the capabilities of welding simulations in welding simulations by experiments. Arup is currently examining additional capabilities that can potentially enhance the understanding of welding process such as the prediction of microstructures and phase transformations.



Welding simulations can be used on ongoing civil engineering projects, both new build and retrofitting projects on existing structures for various industries. The complexity of the welding simulations holds this back for both clients, but also less experienced users within our firm. The suspicion can be overcome by having a capability statement proving the accuracy and validity of the analyses. Many companies are struggling with this, but they are all struggling with this independently, not sharing available test data or knowledge. Why can’t the industry develop a method statement that illustrates a few arbitrary examples with known inputs and outputs that can be used to demonstrate the accuracy of the software and user? After this first round of validation, the user can comfortable use the analyses in practice, comforting himself/herself and the client.



A weld can contain many hidden problems that could be avoided and/or mitigated earlier if a wider application of welding simulations was adopted within the industry. We can detect issues earlier and use the material in a more optimal fashion such that we reduce our carbon footprint. We see a good future for this, but the gap between research and civil engineering industry needs to be closed by having a method statement that has been agreed by the entire industry, not just the isolated institutions or firms.