PHM framework design based on data warehouse

Data warehouse (DW) has experienced tremendous growth in recent decades. It has been applied so widely that gradually become the core of business intelligence (BI), and at the same time, in a relatively short period of time, DW starts to play an important role in the prognostics and health management (PHM) system owing to the advantages of improved performance, better data quality, and the ability to consolidate and summarize data from heterogeneous legacy systems. This paper presents a framework for the development of PHM system based on data warehouse different from traditional data warehouse. The historical data is used by data mining technique to discover regularities and improve current operating health assessment of equipments, so the further decision support can be generated by the advisory system in real time. Self-adaptive means system can achieve feedback of data feature, according to the frequency of fault, the result of equipments maintenance and the requirement of decision-makers, then rationally adjust the extraction of historical data in the DW to improve efficiency and accuracy. This architecture covers all the phases of PHM offboard system design (conceptual, logical and physical) and combines with data management technology and Machinery Information Management Open Systems Alliance (MIMOSA) standards in all of them. Moreover, DW is clearly and formally established by using operational data store, on-line analytical processing and data mining to obtain more effective maintenance identification and scheduling from the early stages of development to the final implementation. The final goal of the work is to suggest an improved methodology for potential users of the data warehouse in their Off-board PHM system design process.