自动卷烟分拣机的控制策略优化与仿真

【作者】 李凯；

【导师】 吴耀华；

【作者基本信息】 山东大学 ， 系统工程， 2006， 硕士

【摘要】 目前在全国各地的烟草配送中心中，自动卷烟分拣系统的使用越来越广泛。多数自动分拣系统都是由多个分拣单机组成，在分拣过程中各分拣单机分拣工作时刻的不同，会给整个分拣系统的工作性能带来很大的影响。因此，对自动卷烟分拣系统的控制策略进行研究就有迫切的实际要求。 本文提出了两种新的控制策略，即：考察当前皮带位置的控制优化模型和预先分配皮带位置的控制优化模型。考察当前皮带位置模型的原理是：分拣机控制程序周期扫描卷烟落下位置的状态，如果为不被占用状态，则可以分拣卷烟，分拣卷烟后将该位置改为占用状态；否则不分拣卷烟。预先分配皮带位置模型的原理是：将皮带上的位置预先分配给各分拣机，产生各分拣机工作时刻，各分拣机在其分拣时刻分拣卷烟。 由于实际分拣作业是连续进行的，本文对订单连续分拣的分拣策略进行了分析。在控制优化后，可以计算出卷烟的位置，分拣时不需要等到全部卷烟移出分拣机区域后再开始下一订单，只需在前后两订单间空余出一定的间距。因此缩短了两订单之间的间距，有助于提高分拣效率。 本文针对分拣系统设计、开发了仿真程序，用于模拟不同分拣策略的分拣情况。该仿真程序主要仿真了卷烟从分拣机上落下的动作和卷烟落到皮带上后，随皮带一起前移的动作。通过仿真可以直观地看到不同分拣策略的卷烟分拣效果。该仿真程序还对仿真的结果数据进行了分析，可以为分拣系统设计者提供分拣时间、分拣效率等信息。仿真订单数据的输入既可以是人工输入，也可以从数据库中读取。从数据库中读取订单数据进行仿真，能使仿真过程更好地模拟分拣系统的实际工作情况。在该仿真程序中分拣系统的属性，如皮带速度，分拣机单机的数量位置等都可以设置。通过更改分拣属性的设置，可分析分拣参数、分拣机布局和订单特点对分拣系统性能的影响。 本文运用EIQ分析方法对订单进行了ABC分类。并对不同类型的订单进行了仿真，得出控制优化模型的仿真结果。仿真结果表明：两种优化模型都不仅使卷烟在皮带上的分布更加均匀，还提高了分拣效率。在大订单时使卷烟在皮带上均匀分布的效果更明显，对于减轻后续工作的压力很有帮助；在小订单时对分拣效率的提高作用更大一些。更多还原

【Abstract】 Nowadays automated picking systems of cigarette, are more and more extensive in cigarette distribution center. Most of the automated picking systems are composed by some single picking machines. The difference of single picking machines’ picking time in the picking process, weakened the working performance of the picking system. So, there is a urgent demand to investigate the control policy of automated picking system.Two control policies are designed in the paper, the model of check the present belt position, and the model of allocate the belt positions in advance. The principle of check the present belt position model is, to check whether the belt position to pick the cigarette, is occupied by other cigarettes. Can pick cigarette if the position is not be occupied, and the state of this position changed to be occupied. If the position is be occupied, then this cigarette can not be picked. The principle of the model of allocate the belt positions in advance is, distribute the positions on the belt to single picking machines, and record the picking time when the single picking machine will pick a cigarette.As the picking process is in succession, order by order, the paper analyses the picking policy of sequential order picking. After optimizing, the belt’s position of cigarette can be calculated, so next order can be picked after a few distance from previous order, not need to wait until all of the cigarettes of the previous order move away from the picking machine area. So the interval between two orders is be shortened, the efficiency of picking cigarettes can be improved.The paper has designed a simulation program, for simulate the picking condition of different picking policies. The main actions of simulation are cigarettes fall on the belt from picking machine, and cigarettes move forwards in the belt. Picking condition of different picking models can be seen intuitively through simulation program. This program still can offer picking information of simulation, such as picking time, efficiency, etc. The data of order can be input artificially, or read from database. The program can simulate the real picking更多还原