Accessibility-driven spatial partitioning: A step towards automated design of multi-piece molds

Quite often complex objects are not manufacturable as a single piece. Due to manufacturing constraints, the desired complex object needs to be partitioned into a number of smaller and simpler manufacturable pieces. After manufacturing individual pieces, the complex object is realized by assembling various pieces together. Representative application domains where manufacturability-driven spatial partitioning is commonly used include mold and die machining, sheet metal products, layered manufacturing of ceramic and composite parts. In order to develop next-generation computer-aided design and manufacturing systems, we need new algorithms to automatically solve manufacturability-driven spatial partitioning problem.; This thesis focuses on the accessibility-driven spatial partitioning problem for material removal processes and provides a systematic approach to solving it. The following problems are studied in this thesis. First, this thesis provides a framework for representing accessibility constraints of various material removal processes as geometric constraints and provides algorithms for detecting violations of these geometric constraints in the object to be produced. Second, in order to eliminate these violations in the entire object, this thesis provides algorithms for partitioning the object into a set of components such that each of them is individually free of accessibility problems and therefore manufacturable. Third, this thesis provides algorithms for adding assembly features that constrain the unnecessary relative degrees of freedom among the manufactured components and facilitate easy assembly. Finally, for those classes of objects that can be modeled using the feature-based representation, this thesis also provides a feature-based partitioning approach for solving the accessibility-driven spatial partitioning problem for them.; The approach developed in this thesis for solving accessibility-driven spatial partitioning problem is adopted to design of sacrificial multi-piece molds. Given a part to be molded, we get the gross mold shape by subtracting the part shape from an enclosure. Then the algorithms developed in this thesis are used to guide the decomposition of gross mold shape into manufacturable components that form the mold assembly.; Besides direct application in automated sacrificial multi-piece mold design, the various algorithms developed in this thesis for solving accessibility-driven spatial partitioning problem will also be applicable to many other computer-aided design and manufacturing applications. Such applications include setup planning for machining, sheet metal product design, and inspection planning.