Fast Estimation Method of Machinable Area of Workpiece Surface for 3+2-Axis Control Machining Using Graphics Device – Visualization Algorithm of Machinable Area and Minimum Shank Length with Texture Projection Technique –

Abstract

This study proposes a new method of estimating tool posture in 3+2-axis control machining process. The proposed method focuses on two different properties of the workpiece surface, the machinable area and then minimum shank length. The distribution of these properties on the workpiece surface is determined by the tool posture, workpiece shape, and the shape of the cutting tool. In the planning process of 3+2-axis control machining, CAM and CAPP operators often determine the combination of tool posture and tooling conditions through trial and error. Considering these processes, it would be extremely useful to have a fast method of visualizing these properties on the workpiece surface to realize CAM and CAPP systems with an interactive interface. Therefore, this study proposes a fast estimation method that visualizes both the machinable area and the distribution of the minimum shank length as a color image for each tool posture candidate. In order to reduce the calculation time of the proposedmethods, a graphics device known as a Graphics Processing Unit (GPU) is introduced. In the proposed algorithm to adapt several features to GPU hardware, the offset shape of the workpiece surface is generated from depth information in rendering 3DCG. Furthermore, the unmachinable area is estimated by the inverse-offset operation and shadow mapping function in 3D-CG techniques. In the visualization phase of the required shank length on the workpiece surface, a color image is generated from the depth information. Then, the color image is projected on the workpiece shape using the texture projection technique. Because most calculation processes can be executed inside the GPU hardware, the developed prototype system can visualize both the unmachinable area and the distribution of minimum shank length within several dozen milliseconds for each tool posture candidate.