UBC Theses and Dissertations
Efficient and accurate geometric simulation of multi-axis milling operations Joy, Jimin
Geometric modeling is an essential part of process planning and verification step in the modern manufacturing practice that employs complex operations such as multi-axis milling. Geometric modeling by itself is used for tool path generation and verification. It is also essential to create important input for mechanistic simulation. Due to this great relevance, many geometric modeling methods have been employed for machining simulation. However it is still a challenge to obtain acceptable combination of accuracy, efficiency and robustness from most of the existing methods. The best known modeling methods also appear to have reached a saturation point. Yet the industrial machining cases are ever increasing in complexity and it demands for a faster method maintaining the acceptable level of accuracy. This thesis presents an enhanced voxel representation format for modeling the machined workpiece geometry in general milling operations. The modeling format is named as Frame-Sliced Voxel representation (FSV-rep) as it uses a novel concept of frame-sliced voxels to represent the boundary of the workpiece volume in a multi-level surface voxel representation for memory-efficient implementation. Frame-sliced voxels enables approximation of the workpiece surface to achieve sub-voxel details. This thesis further identifies an efficient three-step update process that can be followed to compute machined part geometry from an initial FSV-rep workpiece model and set of tool paths. To be computationally feasible and yet robustly handling all tool path types, suitable swept volume representations are identified for various tool path categories. The three-step update process is then used in customized ways for the different categories to utilize the salient features of each. A robust and efficient approach to generate standard surface representation of the machined part geometry from the updated FSV-rep model is also developed. Results show that the FSV-rep model is able to provide acceptable accuracy levels while being significantly faster than popular modeling methods for machined part geometry computation in general multi-axis machining. The specialized swept volume representation identified for planar and 3-axis straight cut operations is further improving the FSV-rep update performance to be up to an order of magnitude faster than possible with general sampled swept volume representations.
Item Citations and Data
Attribution-NonCommercial-NoDerivatives 4.0 International