Physics-driven shape variation modelling at early design stage

Modern markets are becoming increasingly competitive emphasizing the importance of achieving Right First Time (RFT) during the early design stage as a key enabler facilitating cost and time-to-launch (or time-to-market) reduction. One of the leading challenges to deliver RFT is the lack of effective methods to model product errors at early design stage. Usually, the assembly process is designed under the assumption of ideal (nominal) products. On the contrary, it has been demonstrated that product errors (both geometrical and dimensional) affect the performance of the final assembly. To facilitate easy decision making at early design stage, new methods and models are required to support design engineers. In this study, a framework has been proposed for early design support to generate product variation. International standard provides guidelines for product control and inspection (ISO-GPS or ASME-GD & however, the integration of tolerance standard into nominal sized CAD models is not yet achieved. Current, Computer Aided Tolerancing (CAT) tools mainly capable to model orientation and position tolerance specifications, whereas part shape errors are omitted. This paper presents an innovative physics-driven simulation framework to model shape errors of compliant sheet metal parts at early design stage. The modelling framework consists of three important stages: (i) initial shape error prediction using physic-based simulation, such as, stamping process simulation; (ii) individual orthogonal shape error modes/patterns identification based on decomposition techniques, such as, Geometric Modal Analysis (GMA); and, (iii) simulation of shape error variation classes by assigning distribution to each orthogonal shape error modes. The proposed approach enables to generate shape errors at early design stage of assembly process which can be utilized to optimize the assembly process, including fixture design and joining process parameters. An industrial automotive component illustrates the proposed methodology. (C) 2015 The Authors. Published by Elsevier B.V.