The deformations of polycrystalline materials are modelled by linking a constitutive equation for the crystallographic slip of a single crystal to the macroscopic behavior of the aggregate. In this study, anisotropic elasticity (lattice stretching) of a cubic crystal is incoporated into the anisotropic plasticity from crystallographic slip. The constitutive description for the aggregate, derived from a crystal plasticity theory, is used to formulate a Consistent Penalty Finite Element Method for the anisotropic elasto-viscoplastic deformation of polycrystalline materials. As an application, a plane-strain forging process is simulated and the effects of the initial textures on the deformation behavior of the workpiece are examined.

This paper is concerning the development of multiple neural networks system of problem domains where the complete input space can be decomposed into several different regions, and these are known prior to training neural networks. We will adopt oblique decision tree to represent the divided input space and sel ect an appropriate subnetworks, each of which is trained over a different region of input space. The overall architecture of multiple neural networks system, called the federated architecture, consists of a facilitator, normal subnetworks, and tile networks. The role of a facilitator is to choose the subnetwork that is suitable for the given input data using information obtained from decision tree. However, if input data is close enough to the boundaries of regions, there is a large possibility of selecting the invalid subnetwork due to the incorrect prediction of decision tree. When such a situation is encountered, the facilitator selects a tile network that is trained closely to the boundaries of partitioned input space, instead of a normal subnetwork. In this way, it is possible to reduce the large error of neural networks at zones close to borders of regions. The validation of our approach is examined and verified by applying the federated neural networks system to the configuration design of a midship structure.

An axial flow fan design system has been made by integrating the self-developed programs and I-DEAS. By using the system, an axial flow fan was designed, manufactured and verified through the wind tunnel experiments in coorperation with a refrigerator appliance division. It has been shown that the optimal design without the ambiguity of the design parameters can be possible by the three-dimensional flow simulations using a self-developed CID code, FANS-3D. (Flow Analysis code using Navier Stokes aguations in Three-Dimensional curvilinear coordinates). By virtue of the fluency of the data flow, an optimally designed fan which satisfies design conditions can be selected in a short time and less cost. The manufacturing processes of a Mock-up and an injection molding die have been automated through the self-made interface programs which connnect from the start to the end. It has been shown that the newly developed fan by this system has a superior performance characteristics to an existing fan.

Most design changes are due to interferences and fit-up as parts are assembled if such a large product as an automobile or an aircraft is developed by many concurrent engineers. In this part (1) of the thesis, the assembly-centric modeling methodology with CAD systems is proposed in order to reduce the design changes. Unlike part-centric modeling method, a part is modeled with its own coordinate system which is used in downstream process as machining and measuring. The part coordinates initially have the same orientation as its assembly which is predefined in BOM (Bill of Material). Then, the corrdinates origin of the part is moved to its location to be assembled from that of its assembly coordinate system. To implement this methodology, the position data of the part w.r.t. its assembly are stored in a database to build the same hierarchical assembling structure as BOM structure. This modeling approach has the advantage of reflecting asembling sequence, because the process of positioning parts is similar to that of real assembling. And with the method, a designer can easily adjust all of the part positions of an assembly to resolve interferences if he modifies just the coordinates origin of the assembly, which results in moving included parts and assemblies together.

In the development of large complex assembly, most design changes are from the following reasons.: 1) insufficient consideration of assembling sequence of downstream process. 2) interference and fit-up of related parts. 3) initial design changes which are not completely resolved. In order to reduce the first reason of above, the assembly-centric modeling methodology was proposed in the part (1) of the thesis. In this part (2), a Digital Mock-Up system, which is a tool to build a virtual mock-up in design stage, has been developed in order to prevent the other two reasons. The system can build a virtual assembly in various ways as followings.: 1) assembling parts which are located in user defined envelope. 2) assembling parts with the specified disciplines. 3) assembling parts that are selected in the part list. It can also make an assembly by the combination of above methods. In addition to that, it has the functions to read/write the virtual assembly and to explode parts of the assembly in desired direction. With the virtual assembly, engineers can design interference free parts without making physical mock-up. The system has been implemented with Oracle database management system in CATIA environment.

This study intends to develop a fixture planning module as a part of the planning system for cutting. The fixture module uses machine learning method to reuse previous failure results so that the system can reduce the repeated failures. Machine learning is one of efforts to incorporate human reasoning ability into a computerized system. A human expert designs better than a novice does because he has a wide experience in a specific area. This study implements the machine learning algorithm to have a wide experience in the fixture planning area as a human expert does. When the fixture planner finds a setup failure for the suggested operations by a process planner, it makes the process planner store its attributes and other information for the failed setup. Then the process planner applies the learned knowledge when it meets a similar case so that the planner can reduce possibility of setup failure. Also the system can teach a novice user by showing a failed setup with a modified setup.

Ball rolling blending is a popular technique for blending between parametric surfaces. The ball rolling blend surface is conceptually a trajectory of a ball rolling between two base sufaces. It is constructed by sweeping a circular arc along a ball contact curve pair. Since a ball rolling blend surfaces does not have a polynomial form like a Bezier surface patch, it is impossible to apply this method directly to a commercial CAD/CAM system. In this paper an algorithm is developed to approximate a ball rolling blend surface into Bezier surface patches. Least square method is applied to obtain proper Bezier surface patches under a given tolerance. The Bezier surface patches have degree three or more and guarantee VC1-continuity.