In today’s modern world, with the abundance of digital data, data science is identified as a rigorous discipline and machine learning (ML) techniques are commonly used.  The Prognostics & Health Management (PHM) field can successfully be executed utilizing a foundational approach where a digital hierarchy of needs is established for successful implementation of PHM on a large-scale system.  This paper rationalizes the digital hierarchy of needs as it applies to PHM, explains how each foundational concept is essential, and builds upon the base-level concepts through analysis and implementation. First, this paper expounds how the integration of digital data from the lower-level components to the system level is critical for the success of PHM-enabled Systems.  Once established, arguments will be presented for mapping the appropriate fault data to the corresponding components for the purpose of correlating failures to fault data.  Subsequently, a case is presented for using real data to conduct Fault Detection, Fault Isolation, and simple prognostics analyses.  Advanced PHM analyses can then be conducted utilizing data science and machine learning techniques with the intent of predictive maintenance analysis.  Lastly, an argument is presented answering the need for an approach to implement real-time predictive activities once the complex analysis is validated and verified. This concept can be seen graphically in the figure below.  Data integration is the foundation of the pyramid, supporting the identification of fault and parametric data, and followed by Fault Detection, Fault Isolation, and Simple Prognostics.  The top tiers of the pyramid then can integrate complex analysis such as ML, and real-time implementation.

No abstract