Temporal Data Mining Approaches and Green Design Implementation for Data Center Chillers Management System

摘要

Temporal data mining refers time series data points recorded along with the time which they were observed. Time series is a sequence of data points collected at equal time interval for each point. We apply temporal data mining approaches to data center. A data center chillers management system is studied, planned and designed. In this study, we explore the 2 modules of the system: 1) chillers fault detection module and 2) chillers lifetime prediction or time for maintenance module. We propose temporal data mining approaches to solve 1) change point detection and 2) motif mining problems in the said system. The temporal data mining approaches are important to predict the chillers fault, usage lifetime and time for maintenance. They are also the main objectives of this chillers management system. This paper also illustrates the architecture and the framework of the data center chillers management system. It aims to justify the data center chillers management system from the information technology perspective.