![Sample our Computer Science journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5928/TOC-Subject-Sample-2021-Computer-Science.jpg"})
Abstract
Compared to single-cutter machining, using multiple cutters in 5-axis finish machining of freeform surfaces can produce shorter tool-paths; hence the increased machining efficiency. In our previous work, a method to evaluate a cutter’s accessibility at any point on a machining surface has been developed. In this paper, this method is used to identify feasible cutters and construct their machining regions. These cutters can make up many cutter combinations that can finish the entire machining surface, among which there will be an optimal set that produces the shortest tool-path. To find this optimal combination, we propose to use the tool of neural network to predict the tool-path length for a machining region without actually generating the tool-path. The neural network is trained extensively with a large set of carefully designed training data extracted from actual machining jobs. Finally the validity of our method is proved with testing data sets that have never been exposed to the neural network before.