Condition Based Maintenance for Data Center Operations Management

Power distribution system (PDS) of data center has designed for high reliability and fault tolerance by deploying a redundant topology. Optimal data center reliability requires 4 principals which consist of topology design by standard; reliable components or mean time between failures (MTBF); real-time monitoring by data center infrastructure management (DCIM); and preventive maintenance through data center operations. However, data center needs more proper operations and maintenance approach to avoid planned and unplanned downtime. The best practice for data center operations management should be applied predictive maintenance or condition-based maintenance (CBM) to monitoring of critical components and systems. These processes will ensure that the system parameters will perform on their functions, and estimate remaining useful life of components for critical mission. Time-based maintenance (TBM) (preventive maintenance) has been applies for supplementary PDS components that define as not mission critical systems. This research proposes a model of preventive and predictive maintenance (PPM) for PDS of data center, which has been verified to improve maintenance efficiency and reduce operating costs of repair time and inventory. How to identify potential problems in early state before component and system failures are the key to prevent planed and unplanned downtime.