Research on the two-dimensional PSO based method for sub-assembly identification in ship block building
点击次数：
论文类型：会议论文
发表时间：2021-01-09
卷号：3
摘要：Sub-assembly identification is an essential process design in ship block building, as it determines the efficiency and quality of assembly work. Now in most shipyards, experienced method is widely applied to identify sub-assembly division, which costs much design time and therefore leaves room for a more automated and efficient stage. Subassembly design can be regarded as a parts clustering process characterized by fuzziness and quantitative uncertainty, so it is difficult to describe the problem as a one-dimensional vector. To better describe the clustering situation, and realize the automatic identification of sub-assembly for hull structure, a new discrete particle swarm optimization (DPSO) algorithm based on two-dimensional coding is proposed in this study. Firstly, by analyzing the characteristics of sub-assembly partition, a two-dimensional binary matrix is defined to represent clustering results. The updating operation of particle position and velocity is improved with the standard PSO method. Secondly, an assembly information model is established with the definition of link intensity matrix and interference relation matrix, by detecting the factors of welding form and space collision between parts. Moreover, considering the impact of internal tightness, external feasibility and granularity of assembly units, a mathematical optimization model for particle evaluation is designed. Finally, one typical block is selected to verify the applicability and feasibility of the proposed method. ? 2019: The Royal Institution of Naval Architects.
发表时间：2021-01-09