Quantifying the impact of digitalization on manufacturing supply chain management (SCM) in a power generation company

Abstract

Industrial digitalization concepts such as Industry 4.0 or Smart Manufacturing are currently of great interest in academia and among industrial players. These concepts are expected to boost companies' manufacturing supply chain performance factors such as availability and productivity. For instance, greater availability of assets on the shop floor makes the product flow more predictable and smooth, thus reducing the necessity for high inventory and increasing inventory turnover. Although current studies of industrial digital transformation offer a large variable theoretical construct, they lack quantitative proof of their assumptions. The main goal of this thesis is to introduce a method to quantify the expectation that digital initiatives in heavy industry impact certain manufacturing supply chain performance factors. In particular, the study examines the visualization effect on the unplanned machine downtime, planned maintenance, and machine utilization. The assumption of the decrease in unplanned machine downtime, increase in early-stage planned maintenance, and increase in machine utilization are tested using non-parametric hypotheses test - Wilcoxon Signed Rank test. Measurement of these factors is conducted using data collected from a power generation equipment manufacturer. The showcase factory participates in an overall digitalization Smart Manufacturing program and is in its early stage of implementation. The results indicate a significant increase in machine utilization and planned maintenance. However, unplanned machine downtime was not significantly reduced, although the result shows an approximation toward statistically significant change. The importance of frequent analysis becomes obvious. Future tests are necessary to study the development in later stages of implementation of Visualization. The reduction in downtime could become significant and the planned maintenance should stop increasing and start decreasing over time. The proposed method serves as a step toward academic quantitative analysis of industrial digitalization.