Cycle Time Reduction of 5-axis Laser Drilling via Time-optimal Trajectory Generation and Sequence Optimization

trajectory planning;traveling salesman problem;laser drilling;optimization

Woo, Kyongjaeadvisor:Erkorkmaz, Kaan ; affiliation1:Faculty of Engineering ; Erkorkmaz, Kaan ;

University of Waterloo

关键词: laser drilling;    traveling salesman problem;    Master Thesis;    optimization;    trajectory planning;   

Others  :  https://uwspace.uwaterloo.ca/bitstream/10012/14585/1/Woo_Kyongjae.pdf

瑞士|英语

来源: UWSPACE Waterloo Institutional Repository

【 摘 要 】

Cycle time reduction is one of the crucial tasks in manufacturing that needs to be achieved tomaximize productivity and profits. Laser drilling processes, depending on the size and complexityof the parts, require few hundreds to few thousands of holes to be drilled. Therefore, cycle time isdirectly related to in what order and manner the holes are visited. In this thesis, a method of cycletime reduction for 5-axis percussion laser drilling process is presented via generation of time-optimaltrajectory and optimization of hole visiting sequence.In percussion laser drilling, a series of laser pulses are fired to each hole while the workpiece isstationary. Once a hole is completely opened up, then drilling of the next hole continues byrepositioning the workpiece with respect to the beam. This stop-and-go nature of the drillingprocess enables one to describe the sequence optimization problem as a well-known TravelingSalesman Problem (TSP) in combinatorial optimization. The objective of TSP is to find aminimum cost sequence of points when the point-to-point cost information for every possible pairis known. In the case of the minimum cycle time problem, the point-to-point cost is the travel time,and the objective of TSP is to find a sequence with the minimum overall travel time.In planning of time-optimal trajectory for point-to-point motion under a specified path, industryuses CNC controller’s G00 (rapid traverse) + TRAORI (5-axis transformation and tool orientationretaining tactic) commands. To be practically beneficial, time-optimal trajectory generationstrategies discussed in this thesis is focused on closely estimating these CNC controller’sbehaviors. A total of four strategies are studied, and the most accurate strategy is chosen bycomparing the results with the experimentally measured CNC trajectories. The most accurate onespecifies the tool paths in Workpiece Coordinates followed by iterative velocity profiling of thetool path parameter to achieve minimum time trajectory under the machine’s velocity,acceleration, and jerk limits.With every hole-to-hole travel time calculated from the above strategy, sequence optimization canbe conducted. In this thesis, two methods from the industry partner, the proposed method, and theoptimal solver method are discussed. Due to licensing limitations, the proposed method isdeveloped in-house instead of using existing non-commercial TSP algorithms. The proposedmethod uses local search heuristics approach inspired by famous Lin-Kernighan heuristics. Theivresults are compared to the optimal solutions generated from the non-commercial state-of-the-artTSP solver called Concorde for benchmarking purposes.To understand the impact of the research in a real environment, one sample part and its originaldrilling process information have been made available by the industry partner. Although the fullexperimental results are not yet acquired at the moment of writing this thesis, the simulation resultsshow that the proposed sequencing optimization in conjunction with the proposed hole-to-holetrajectory generation strategy for correct estimation of travel time improves the overall cycle timeby 26.0 %.