Research On Path Optimization Of Two Dimension Laser-Cutting

Laser cutting is one of the most widely applied cutting technologies in the manufacturingindustry because of its wide range, high speed, narrow kerfs and small heat-affected zone. Inorder to improve the efficiency, quality, and cost.the laser cutting process is further studied inthis dissertation.First， the significance of the topic and the domestic, foreign research status,frequently-used cutting technologies and mechanisms in laser cutting are introduced. And thetechnological treatments of laser cutting for sheets are discussed, including setting up lead-inand lead-out lines，arranging auxiliary cutting path, the treatment in cutting sharp corners andso on. In this paper we consider the problem of path optimization for laser cutting with twodifferent constraint cutting conditionsFor the idle stroke problem in laser cutting, an algorithm to optimize the cutting path ispresented. At first, centroid coordinates are used to express part positions, which transformsthe problem into a Travelling Salesman Problem, so that the best cutting sequence can be gotby genetic algorithm. Then traditional heuristics methods are adopted to determine thepiercing point of each contour. Computational experiments shows that the idle stroke isreduced by as much as11.77%.To avoid the torch falling down into the finished cut cavity during its vacancy course, acollision-avoiding algorithm is presented. Firstly，two local search heuristics algorithm areused to determining the cutting sequence. Secondly，make sure that the torch has travelledthrough the cut area as few as possible. Finally，the former path is adjusted to totally preventcollisions. Experimental results demonstrate that the algorithm can completely avoid thecutting collision. The largest path optimization is as high as38.32%Two key problems in laser cutting path planning are studied in this dissertation, forwhich a shortest path optimization algorithm and a collision-avoiding algorithm are presented,whose effectiveness is tested in computational trials. The experimental results show thatproposed algorithms have high efficiency and strong practicability.