Clustering and Operation Analysis for Assembly Blocks Using Process Mining in Shipbuilding Industry

Abstract

A block assembly process in the shipbuilding industry consists of many work stages. Block assembly involves many workers in many shops. Each assembly block, which is a part of a ship, has a different structure requiring specific work processes. Therefore, in order to better understand such real processes, an information system for monitoring of block position has been developed. Recently, the necessity of using data accumulated in information systems has become greater. This paper proposes a new, clustering and operation analysis method for assembly blocks based on process mining techniques suitable for the shipbuilding industry. The approach consists of four steps: 1) trace clustering from the task perspective, 2) trace clustering from the work shop perspective, 3) definition of new clusters considering task and work shop simultaneously, and 4) comparison of new clusters with other clusters from the process perspective. The output of clustering and operation analysis can be used for production planning purposes such as resource allocation and operation scheduling for assembly blocks. The effectiveness of the proposed method was verified in a case study using real event logs generated from the Block Assembly Monitoring System (BAMS), an information system.