应用神经网络实现中间产品加工工时定额的快速计算

Rapid Interim Product Man-hour Ration Estimation with Artificial Neural Network Method

准确地计算中间产品加工工时定额是制定合理生产计划的基础。以前的工时定额数据主要来源于已建造的同类型船舶的数据积累,这种工时估算过多地依赖于人的经验,而且也没有针对中间产品族的加工单独给出工时定额。为适应中间产品的专业化生产,需要提供不同中间产品族生产的工时定额,作为制定中间产品族生产单元作业计划的依据。本文通过对部件特征的描述,划分出了部件加工族,并以一部件族为例,通过分析该部件族的典型部件生产过程,得到了部件生产工时定额的影响因素,并且应用BP神经网络建立工时定额和影响因素之间的映射关系,从而可以通过神经网络方便、准确地计算出该部件族中每个部件生产的工时定额。

A rational production plan is based on accurate estimation on manhour of interim product. Previously, a new ship's manhours are estimated from similar ship built before. But accuracy of the data depends on work's experience. And they aren't suitable for interim product cellular manufacture. So, we need manhours of all interim product cluster to make out a rational cellular production plan. After description of subassembly's characteristics, subassembly families are obtained. As an example, work procedures of some typical subassemblies are analyzed, and effect factors to manhour ration of subassembly are got. And then the relationship between manhour rations and effect factors is established with BP artificial neural network. So manhour ration of each subassembly will be calculated efficiently and accurately.