Mode-based Decomposition of Part Form Error by Discrete-Cosine-Transform with Implementation to Assembly and Stamping System with Compliant Parts

A discrete-cosine-transformation (DCT) based decomposition method is proposed for modeling part form error, which decomposes the error field into a series of independent error modes. Compression, which ensures a compact model, is achieved by correlation reduction or mode truncation based on good energy compaction property of DCT. The part error related to assembly process (rigid body modes) and part distortion during manufacturing (deformation modes) can be separated and identified. DCT has been proven to be equivalent to least square regression with 2D cosine-base for modeling of part variation pattern. Estimation of these parameters in the model has also been developed. The proposed method was applied to model and evaluate assembly and stamping errors at one of the US stamping plant.