Dieter Fichtner

Use of agents and neural networks for acquisition and preparation of distributed NC information to support NC planning

SPEP-NC information offers new possibilities for object-oriented data management. A strong need for high information content and the increasing object-oriented CAM software requires a new information level for preparation and execution tasks in part manufacturing. This paper describes applications of different intelligent methods especially to support distributed technological data management. The problem of current technological databases will be solved with the help of agent technology and machine learning for distributed information on the shop floor and in planning departments. The authors introduce a combination of agent-based organization and self-learning of feature-based technological information. The prototype includes three components for coordination, data acquisition and preprocessing of technological information to assist human engineers in planning and manufacturing.

A Process for Making Cutaneous Radiation Applicators Based on Digital Data

An optical modeling process for facial regions and other body surfaces has been developed. The body part in question is digitized using optical 3-D metrology to obtain a comprehensive dataset. The data is then prepared for further use by smoothing the point clusters. The radiophysically significant position of the radioactive material, e. g., seeds, aluminum tubes, etc., can be accurately determined using CAD modeling. Subsequently, a rapid prototyping process (fused deposition modeling [FDM]) is used to implement the CAD model directly in order to create a radiation applicator that can be used in practice. Biologically compatible polycarbonate can be used for this purpose.Es wurde ein optisches Abformverfahren für Gesichtsbereiche und weitere Körperoberflächen entwickelt. Der gewünschte Körperteil wird mittels optischer Dreikoordinatenmesstechnik digitalisiert, um einen umfangreichen Datensatz zu gewinnen. Dieser wird durch Glättung der Punktwolken für die weitere Nutzung aufbereitet. Die strahlenphysikalisch bedeutsame Lage des strahlenden Guts (Seeds, Aluminiumschläuche etc.) kann mittels CAD-Modellierung eindeutig festgelegt werden. Danach wird mittels eines Rapid-Prototyping-Verfahrens („fused deposition modeling“ [FDM]) das CAD-Modell direkt in den praktisch anwendbaren Strahlenapplikator umgesetzt. Dafür ist beispielsweise biologisch verträgliches Polycarbonat anwendbar.

Neural network based application and data management for cutting values

The manufacturing processes have to guarantee a flexible, high quality production at low cost. In order to fulfil the demands, intelligent information systems are required with special respect to the technology data type. A central point of technological data is the cutting values, i.e. feed and speed. The high speed milling process is one of the newly established processes with the need to set the optimum cutting parameters. The rapid development of advanced materials, tools etc. the results of a recent project* to solve cutting parameter generation problems by means of neural network, mathematical model and technology database system.

User Support and Experiences for CAD/CAM Process Chains for Five Axis Milling

In this paper some technological and economical aspects of five axis machining, especially by new CNC functions, and characteristics of the developed pilot solution are described.