ORIGINAL ARTICLE Predicting tolerance on the welding distortion in a thin aluminum welded T-joint Cristina Renzi 1 & Davide Panari 2 & Francesco Leali 2 Received: 2 March 2017 /Accepted: 26 January 2018 /Published online: 23 February 2018 # Springer-Verlag London Ltd., part of Springer Nature 2018 Abstract Welding is a widely accepted process used in the assembly of aluminum chassis structures in the automotive industry. Finite element analysis (FEA) is usually adopted to predict distortions caused in the welding process. However, only nominal distor- tions result from FEA simulations. Welding distortions could be more accurately predicted by introducing the prediction of tolerances due to a modification of the input parameters. The aim of this work is therefore to introduce the tolerance evaluation in the FEA model, by varying the welding input parameters (geometrical and dimensional tolerances on the parts, heat input…). To find the most suitable FEA model to investigate welding process tolerance, three FEA models are compared: one is the thermo- elastic-plastic (TEP) model, and two are based on the inherent strain method. The case study uses a thin (2 mm) aluminum T-joint, which is commonly used in automotive chassis assembly. Results deriving from FEA simulations were compared with exper- imental data. Among the various input parameters affecting the welding process, the authors combined the dimensional tolerance on the plate thickness with the variability of the heat input. The results provided a tolerance range value for the angular distortion of the T-joint. Keywords Design methods . Thermo-Elastic-Plastic analysis . inherent strain . welding Tjoint . FEA . automotive design 1 Introduction In the automotive industry today, welding is a consolidated assembly process [1] for thin aluminum structures (plate thickness less than 3 mm) used typically in chassis assemblies. However, welded aluminum structures suffer from prob- lems of distortion [2], due to high heat conductivity, an expan- sion coefficient almost twice that of steel and a modulus of elasticity one third that of steel [3]. The problem of welding distortion has been widely tackled in literature, mainly on steel large structures, characterized by butt joints or thick T-joints (e.g., [4–11]). CAX methods have been extensively used in automotive literature to optimize engineering design parameters correlat- ed to performance, starting from geometrical specifications (e.g., [12, 13]), using MCDM techniques in the evaluation and selection of concept design alternatives (e.g., [14, 15]), also in comparison to design methods in SMEs [16]. In particular, thermal elastic plastic (TEP) finite element analyses (FEA) have been widely used to evaluate the welding distortion in a welded joint (e.g., [17–20]). TEP simulation of the welding process is required to replicate the behavior of the residual stresses and deformations generated in the cooling phase. Firstly, a thermal analysis is carried out, replicating both welding and cooling phases. Thereafter, an elastic anal- ysis is performed, taking the resulting thermal energy distri- bution as the initial load condition. Deformations are due to the residual stress after the cooling phase in the welding process. * Cristina Renzi cristina.renzi@unimore.it; cristina.renzi@gmail.com Davide Panari davide.panari@unimore.it Francesco Leali francesco.leali@unimore.it 1 InterMech-MO.RE., Interdepartmental Research Centre for Applied Research and Services in the Advanced Mechanics and Motor Sector, University of Modena and Reggio Emilia, via Vivarelli 10, Modena, Italy 2 Department of Engineering “Enzo Ferrari”, University of Modena and Reggio Emilia, via Vivarelli 10, Modena, Italy The International Journal of Advanced Manufacturing Technology (2018) 96:2479–2494 https://doi.org/10.1007/s00170-018-1677-7