A Case Study to Monitor Parallel‐Connected Ni–Cd Batteries on EMU Trains Through Compositional Data Control Charts

ABSTRACT Electric multiple unit (EMU) trains, widely used in modern railway transportation, use Nickel–Cadmium (Ni–Cd) batteries to power auxiliary systems on board, including lighting, ventilation, control electronics, and emergency systems. These batteries are often connected in parallel to ensure a wider power supply range for electrical loads that may vary in response to external demands. When the condition of a parallel‐connected battery system is adequate, each battery contributes approximately the same portion of the total electrical load, as the current flowing through each battery is ideally equal. Monitoring such battery systems is crucial for the early detection of performance degradation and potential failures. To ensure that monitoring focuses on the relative contributions of each battery regardless of electrical load, we propose a statistical process monitoring approach based on compositional data analysis. In this case study, we explore the use of Hotelling's and multivariate exponentially weighted moving average control charts, applied after mapping the compositions into isometric log‐ratio coordinates, for the Phase II monitoring of parallel‐connected Ni–Cd batteries installed on a modern high‐speed train to power auxiliary onboard systems. Travel direction is treated as a known categorical covariate whose systematic effect is removed before monitoring, and the two charts are also considered jointly in a combined scheme. We demonstrate that the proposed approach guarantees a pre‐specified in‐control average run length, regardless of the electrical load. The R code and the real data to reproduce the main figures and tables are made openly available as .