This preview shows page 1. Sign up to view the full content.
Unformatted text preview: ation of compliant parts often have significant variation with wavelengths longer than the part-length, which cannot be accurately modeled using spectral analysis, and [Stout, 2002] presents a polynomial-based method to model such variation. [Tonks and Chase, 2004] develop a method to model long wavelength variation using a series of orthogonal polynomials. In this work, the CSTA method of [Merkley and Chase, 1996] is summarized, the need to model the effects of surface variation is explained, and typical surface variation is investigated. A hybrid geometric covariance model that combines the work of [Bihlmaier, 1999] and [Tonks and Chase, 2004] is presented. The hybrid geometric covariance model is used to predict the geometric covariance of a set of simulated parts and the covariance calculated from measured data taken from a set of sheet-metal parts. 2. LINEAR CSTA METHOD The CSTA method, first developed by [Merkley and Chase, 1996], finds the residual stress and springback after assembly due to dimensional and surface variation. The solution can be divided into three sections, solving for the total misalignment in the assembly; finding the covariance of the misalignment of each part, and determin...
View Full Document
- Spring '10
- The Land