Approach and summary of the Thesis
   In the computer aided process planning for machining processes, one of the first stages consists of identifying the zones of material to eliminate in the raw material to generate the final part. The result is a set of entities called Machining Features, which have a clear relationship with the machining operations.
   The procedure of obtaining automatically these entities is named automatic feature recognition (AFR). This procedure starts with the 3D raw material model and the part model, establishing the adequate working entities (Machining Features). These entities contain the necessary information to carry out an automatic Process Planning. At the same time, the information is completed and expanded as the stages of Planning are developed.
   In the Thesis an automatic feature recognition is developed, as one of the first stages of the Process Planning, allowing integration with the computer aided design. This recognition should have a dynamic approach, offering different options. The provided solution should not be a static input, pre-established, for the remainder stages of the Planning. The process of recognition is strongly influenced by technological concepts and decisions (tool types, typical processes movements, linked cut influence, etc.) that guide and allow it to obtain valid results in the final application: the machining.
   Attending to this approach, the Thesis offers a complete and general solution to the process of automatic feature recognition, keeping in mind so-called conventional processes (turning, milling, filing, grinding, etc.). The proposed solution is not restricted to prismatic parts. It works with any regular surface (cylindrical, conical, flat, toric and spherical surfaces), giving solution to one of the most important problems that exist in the automatic feature recognition, the intersection/interaction among them. In order to obtain it, it works with 3D models in B-Rep format (Boundary Representation) in accordance to the standard ISO 10303, taking into account concavity/convexity relationships among the surfaces of the initial models.
   Technological knowledge is applied in the proposed operative, emphasizing the extraction of valid volumes in the process of mechanized destiny, obtaining at least one tool that should allow the elimination of the volume in this process, checking the accessibility in the zone of work to generate the suitable geometry, as well as the accessibility to the zone of work over the rest of geometry of the part. Only the verification of these concepts can guarantee that the part can be manufactured.
   The proposal does not obtain a unique solution, since each one of the volumes can be eliminated with different machine processes (including the existence of a tool and accessibility). At the same time, the proposal offers alternative tool profiles and its orientation respect to the accessibility.
   An important issue of the proposed method in the Thesis refers to how to extract the volumes. Each obtained volume must be removed with a unique tool, in only one orientation, without changing the working plane and its size must be the major possible. The proposal avoids to give solutions corresponding to sculpture surfaces machining, 3D machining, because they do not have relationship with the use of machining features. Basically, to extract the volumes in the proposed way support interactions among possible machining features.
   Final results data are stored in structures especially thought for the management of the information, for example, for combining Machining Features. These structures allow a direct migration to the Part 224 of the standard STEP. This standard is dedicated to the product definition based on Machining Features to be used in an automatic Process Planning.
   To verify the validity of the theoretical exposition of the thesis and its feasible algorithmic execution, main algorithms for all the processes are developed, as well as those specifics to the selected process. These algorithms have allowed test the proposed theoretical methodology on any part, regardless of its geometric complexity.