基于BL-NFP神经网络算法的二维不规则钣金零件排样  

Layout for two-dimensional irregular sheet metal parts based on BL-NFP neural network algorithm

作　　者：刘玲玲 赵罘[1] 龚堰珏 Liu Lingling;Zhao Fu;Gong Yanjue(Institute of Artificial Intelligence,Beijing University of Industry and Commerce,Beijing 100048,China)

机构地区：[1]北京工商大学人工智能学院,北京100048

出　　处：《锻压技术》2021年第12期54-60,共7页Forging & Stamping Technology

基　　金：国家自然科学基金资助项目(51975006)。

摘　　要：针对传统工业生产中二维不规则钣金零件的利用率不高、计算时间较长的问题,提出了以能量函数为载体的BL-NFP(Bottom Left-No Fit Polygon)神经网络算法。该算法是将BL(Bottom-Left)定位算法和临界多边形(No-Fit Polygon,NFP)几何特性相结合,同时模拟了钣金零件的排样过程。并采取对待排入零件优先进行面积大小核算、再排入待排物体的方式,并利用Matlab对算法的输出数据和前人所列出的数据结果进行了测试对比。结果表明:BL定位算法能够合理地计算出零件的排入位置,NFP能够有效地解决不规则零件排样利用率小的问题,神经网络算法则能够有效地提高求解速度。针对二维不规则钣金零件的排样问题,与传统神经网络算法相比较,采用该算法缩短了钣金零件下料机器计算最优解40%的时间,并提高了约8%的钣金材料利用率。For the problems of low utilization rate and long calculation time of two-dimensional irregular sheet metal parts in traditional industrial production,a BL-NFP(Bottom Left-No Fit Polygon)neural network algorithm with energy function as the carrier was proposed,which combined BL(Bottom-Left)positioning algorithm with the geometric characteristics of critical polygon(No-Fit Polygon,NFP),and the layout process of sheet metal parts was simulated.Then,the method of calculating the area sizes of the parts to be placed first,and then placing the objects to be arranged was adopted.Furthermore,the output data of the algorithm and the data results listed by the predecessors were tested and compared by Matlab.The results show that BL positioning algorithm can reasonably calculate the placing position of parts,NFP can effectively solve the problem of low layout utilization rate for irregular parts,and the neural network algorithm can effectively improve the solution speed.Aiming at the layout problem of two-dimensional irregular sheet metal parts,compared with the traditional neural network algorithm,the time of the optimal solution calculated by blanking machine of sheet metal parts is shortened by 40%using this algorithm,and the utilization rate of sheet metal material is improved about 8%.

关 键 词：二维不规则钣金零件 能量函数 BL-NFP神经网络算法 排样 材料利用率 

分 类 号：TH164[机械工程—机械制造及自动化]