Analysis Framework using Process Mining for Block Movement Process in Shipyards

Dongha Lee, Hyerim BaeJournal of Korean Institute of Industrial Engineers 2013 · DOI

Abstract

In a shipyard, it is hard to predict block movement due to the uncertainty caused during the long period of shipbuilding operations. For this reason, block movement is rarely scheduled, while main operations such as assembly, outfitting and painting are scheduled properly. Nonetheless, the high operating costs of block movement compel task managers to attempt its management. To resolve this dilemma, this paper proposes a new block movement analysis framework consisting of the following operations: understanding the entire process, log clustering to obtain manageable processes, discovering the process model and detecting exceptional processes. The proposed framework applies fuzzy mining and trace clustering among the process mining technologies to find main process and define process models easily. We also propose additional methodologies including adjustment of the semantic expression level for process instances to obtain an interpretable process model, definition of each cluster's process model, detection of exceptional processes, and others. The effectiveness of the proposed framework was verified in a case study using real-world event logs generated from the Block Process Monitoring System (BPMS).

Introduction

Shipbuilding is a complex logistical operation involving the assembly of massive steel blocks. While primary production stages like assembly, outfitting, and painting are meticulously planned, the movement of these blocks between workshops and storage areas remains a significant "black box." Due to physical constraints and long production cycles, blocks are frequently moved to temporary storage or shifted to accommodate other tasks. These unscheduled movements are expensive and often result in "spaghetti models"—process maps too intricate to provide actionable insights. This research aims to transform raw data from Block Process Monitoring Systems (BPMS) into structured models to identify and eliminate logistical waste.

Methodology: The Four-Step Framework

The authors introduce a structured approach to analyze block movements using process mining techniques. The framework is designed to move from high-level process understanding to the detection of specific inefficiencies.

The proposed four-step block movement analysis framework.
  1. Understanding the Entire Process (Step 1): Using Fuzzy Mining, the framework simplifies complex event logs by filtering noise and aggregating infrequent activities. A key innovation is semantic expression adjustment, where specific workshop codes are grouped into broader categories (e.g., "Assembly Shop") to enhance readability.
  2. Log Clustering (Step 2): To address the diversity of block paths, Trace Clustering (k-means and hierarchical) is used to group blocks with similar movement patterns into distinct sub-logs.
  3. Discovering Process Models (Step 3): Standard process models are generated for each cluster. At this stage, the semantic level is refined back to detailed codes to ensure the models are useful for operational management.
  4. Detecting Exceptional Processes (Step 4): Actual logs are compared against standard models to identify three types of waste: Re-handling (multiple moves in the same area), Reverse Movement (moving back to a previous workshop), and Unnecessary Movement (moves between storage yards without production).

The study utilized 3,167 event logs from 212 unique blocks at a major shipyard, capturing data fields such as project ID, location, timestamp, and transporter ID.

Results and Insights

The framework successfully categorized the 212 blocks into three distinct clusters based on their movement complexity. Cluster 0 (84 blocks) followed a standard flow, while Cluster 2 (57 blocks) involved highly complex movements with frequent storage yard stays.

Process model highlighting the main process flow and identified abstracted clusters.

Quantitative analysis revealed significant inefficiencies:

  • Re-handling was the most prevalent issue, affecting 24.9% of events in the standard cluster and 27.3% in the complex cluster.
  • Unnecessary Movement accounted for 12.3% of moves in the complex cluster, highlighting a clear need for storage optimization.

The "dual-layer" approach—using Fuzzy Mining for a high-level view and Trace Clustering for detailed analysis—proved effective in resolving the "spaghetti model" problem. By formalizing definitions for re-handling and reverse movement, the framework provides a quantitative baseline for logistical performance.

Conclusion and Future Work

While the framework provides unprecedented visibility into block logistics, it is currently limited by a lack of "intent" data; the system records where a block moved, but not why (e.g., failed inspection vs. full workshop). Future research will focus on integrating real-time monitoring for immediate alerts, incorporating scheduling data for "planned vs. actual" comparisons, and developing expert systems to automatically categorize the reasons behind movement exceptions.

Summary updated

Abstract

In a shipyard, it is hard to predict block movement due to the uncertainty caused during the long period of shipbuilding operations. For this reason, block movement is rarely scheduled, while main operations such as assembly, outfitting and painting are scheduled properly. Nonetheless, the high operating costs of block movement compel task managers to attempt its management. To resolve this dilemma, this paper proposes a new block movement analysis framework consisting of the following operations: understanding the entire process, log clustering to obtain manageable processes, discovering the process model and detecting exceptional processes. The proposed framework applies fuzzy mining and trace clustering among the process mining technologies to find main process and define process models easily. We also propose additional methodologies including adjustment of the semantic expression level for process instances to obtain an interpretable process model, definition of each cluster's process model, detection of exceptional processes, and others. The effectiveness of the proposed framework was verified in a case study using real-world event logs generated from the Block Process Monitoring System (BPMS).
BibTeX
@article{lee2013analysis,
  author = {Dongha Lee and Hyerim Bae},
  title = {{Analysis Framework using Process Mining for Block Movement Process in Shipyards}},
  journal = {Journal of Korean Institute of Industrial Engineers},
  year = {2013},
  doi = {10.7232/jkiie.2013.39.6.577},
}
Figure