Bottleneck Mining: A data-driven bottleneck identification method via process mining in manufacturing systems

In the era of Industry 4.0, with the availability of production data, there is an increasing demand to detect the throughput bottleneck in manufacturing systems using a data-driven approach. Traditional data-driven approaches identify the bottleneck indirectly by analyzing machine metrics collected from the shop floor, such as machine starvation and blockage. While efficient in most cases, these approaches often fail to identify the bottleneck according to its sensitivity-based definition, potentially resulting in incorrect results. To address this gap, we propose a novel data-driven approach named “Bottleneck Mining”. This approach utilizes event logs as inputs to derive executable simulation models through process mining. Subsequently, it performs sensitivity analysis on the system throughput to accurately locate the bottleneck. The usefulness of the proposed approach has been demonstrated on several production lines.