Hyung-Min Rho

A Hybrid Feature Recognizer for Machining Process Planning Systems

We describe a hybrid feature recognition method for machining features that integrates three distinct feature recognition methods: graph matching, cell-based maximal volume decomposition, and negative feature decomposition using convex decomposition. Each of these methods has strengths and limitations, which are evaluated separately. We integrate these methods in a sequential workflow, such that each method recognizes features according to its strengths, and successively simplifies the part model for the following methods. We identify two anomalous cases in the application of maximal volume decomposition, and their cure by introducing limiting halfspaces. Feature volumes recognized by all three methods are then combined into a unified hierarchical feature representation, which captures feature interaction information, including geometry-based machining precedence relations.

Generation of Modified Cutting Condition Using Neural Network for an Operation Planning System

Abstract For the rational scheduling of production in descrete machining shop, standard time needs to be estimated correctly. The estimation depends on the accuracy of cutting condition. Machinists usually make modifications to recommended cutting conditions suggested by process planner in order to satisfy requirements for individual operation. The systematization of the modification procedure by using neural network methodology is proposed in this paper. Also, other key functions of the operation planning system for prismatic components which includes the module for cutting condition are briefly described. The results of operation planning are demonstrated with an example part.

Managing Design Knowledge: Active Document System

Abstract Sharing design knowledge within a company is critical to retain market competitiveness, especially when the company develops a family of products. The goal of this study is to develop a Web-based design knowledge management system, named Active Document System, that helps designers retrieve critical design information promptly. In order to manage unstructured knowledge such as design know-how systematically, Active Document System classifies design knowledge in accordance with product functions as well as product structures. This paper explains Active Document System and a method to deploy the product function. The concept is implemented and applied to the management of design knowledge for the refrigerator design process.

Incremental Supervised Learning of Cutting Conditons Using the Fuzzy ARTMAP Neural Network

Abstract As a part of an effort to systematize operation planning for cutting processes, the fuzzy ARTMAP neural network has been applied to model the process of selecting cutting conditions and subsequently to learn cutting conditions for training the model. The fuzzy ARTMAP neural network is capable of incremental supervised learning, which enables the model to be reinforced continually and efficiently. In addition, a new algorithm called the replacement algorithm is proposed. When new cutting conditions that are more effective for a certain circumstance are obtained, the proposed algorithm deletes the old information learned, and then makes the network learn the better ones. Examples of decisions of cutting conditions using the fuzzy ARTMAP neural network and the replacement algorithm are provided and discussed

A rapid reverse engineering system for reproducing 3D human busts

Abstract The most critical problems in the current reverse engineering (RE) are often related to data registration and hole filling. This is because no single scan image is sufficient to describe the entire object since most objects occlude themselves, and some scanned data is likely to be missing on the object. In this paper, we establish a registration and a hole-filling algorithm to solve this problem. The registration module is based upon the standard Iterative Closest Point (ICP) algorithm with some additional heuristics, which help to overcome the limitation of ICP requiring that one of the data scans must be a subset of the other. The aligned triangle mesh is still unsuitable for model reproduction since it often has missing data. The proposed hole filling algorithm reconstructs these missing data with the iterative manner and these data are then smoothed with their neighboring polygonal meshes. We have developed a dedicated system for reproducing a human head and shoulder by using the implemented polygon-based algorithms. The system quickly reconstructs the surface datails of human busts and produces the replica of 3D human model using 4-axis CNC milling machine. Some typical examples are presented to validate the system and further research work is also discussed.

An Evaluation System of Order Acceptability under Consideration of Machine Loading in Die Manufacturing

Abstract In die manufacturing factories, orders are often accepted without reasonable evaluation. This incurs difficulties to generate an appropriate production schedule. By exploring some characteristics of die manufacturing, this paper proposes a heuristic method for quick order acceptance evaluation which considers of both real and potential loads at the order evaluation phase. This method is implemented into a quick order evaluation system as a part of the integrated production management system, and can perform the order evaluation quickly by comparing all the standard operation times for main parts of a die set with available capacity of machines.

Design Knowledge Management with Reconstructible Structure

Abstract Critical Issue In knowledge management Is how to Increase the adaptability of knowledge as well as how to reconstruct similar knowledge. In order to construct the design knowledge for new products, designers need to reconstruct the design knowledge with new relationships. A new design knowledge expression method, Design DNA, Is proposed In this work far representing complicated relationships of design knowledge. Design DNA facilitates rearrangement of design knowledge and promotes the flexibility of knowledge structure. Design DNA Is based on Layout-oriented domain knowledge and Function-oriented domain knowledge, which enables to generate new design knowledge that will result in new part geometries for given constraints on the part functions.

Tolerance Representation Scheme for Integrated Cutting Process & Inspection Planning

In order to improve machining quality of parts, tolerance information should be easily exchanged among product designers, process planners, operators and inspectors. The tolerance information exchange is not yet accomplished since common formats for tolerance information management are being under development. So far the tolerance is recorded as text information in the drawings of commercial 3D CAD systems, which requires human interpretation for its use in cutting process planning and inspection planning processes.

A study on 3D measurement for the composition of the human bust by using digital projection moire

Moire topography method is a well-known non-contacting 3-D measurement method. Recently, the automatic 3-D measurement by moire topography has been required since the method was frequently applied to the engineering and medical fields. 3-D measurement using projection moire topography is very attractive because of its high measuring speed and high sensitivity. In this paper, using two-wavelength methods of projection moire topography was tested to a measuring object with 2π-ambiguity problems. Experimental results prove that the proposed scheme is capable of finding absolute fringe orders, so that the 2π-ambiguity problems can be effectively overcome so as to treat large step discontinuities in measured objects.

Representation of assembly and inspection processes for the tolerance analysis of automobile bodies

An automobile body assembly is a complex system consisting of hundreds of compliant sheet metal parts. A number of locating schemes are used throughout the assembly and inspection processes. This paper presents a methodology to represent the assembly and inspection processes of an automobile body for tolerance analysis. The proposed representation methodology consistently describes dimensional variations with respect to various locating schemes that change through the assembly process.