The injection molding industry is large and diversified. However there is no universally accepted way to bid molds, despite the fact that the mold and related design comprise 50% of the total cost of an injection-molded part over its lifetime. This is due to both the structure of the industry and technical difficulties in developing an automated and practical cost estimation system. The technical challenges include lack of a common data format for both parts and molds; the comprehensive consideration of the data about a wide variety of mold types, designs, complexities, number of cavities and other factors that directly affect cost; and the robustness of estimation due to variations of build time and cost. In this research, we propose a new mold cost estimation approach based upon clustered features of parts. Geometry similarity is used to estimate the complexity of a mold from a 2D image with one orthographic view of the injection-molded part. Wavelet descriptors of boundaries as well as other inherent shape properties such as size, number of boundaries, etc. are used to describe the complexity of the part. Regression models are then built to predict costs. In addition to mean estimates, prediction intervals are calculated to support risk management.
Doctor of Philosophy (Ph.D.)
College of Engineering and Computer Science
Industrial Engineering and Management Systems
Length of Campus-only Access
Doctoral Dissertation (Open Access)
Hillsman, Cyrus Clinton, "An Analogy Based Costing System For Injection Molds Based Upon Geometry Similarity With Wavelets" (2009). Electronic Theses and Dissertations, 2004-2019. 6131.