Hierarchical modeling of a manufacturing work cell to promote contextualized PHM information across multiple levels

Abstract

With the ever-increasing demand for reconfigurability and modularity in manufacturing, industrial work cells are increasingly integrating newer and more diverse technologies to not only support the production of a wider range of parts, but also ease the repair or replacement of faulty systems and components. Complex relationships between different elements of a work cell originate from the integration of multiple layers of hardware and software needed to successfully execute the complicated manufacturing processes. Much work within the science of PHM (prognostics and health management) has been dedicated towards the management of some of this complexity via monitoring, diagnostic, and prognostic technologies. The strategic application of PHM technologies has been shown to effectively reduce equipment/process downtime and lower maintenance costs. Part of the challenge of PHM, particularly for manufacturers, is to know exactly where, and how to apply PHM within their work cell operations to gain the maximum actionable information. This problem is further compounded for small to medium-sized manufacturers (SMMs) who are typically limited in their resources and investment capital. Effectively designing and implementing PHM requires a fundamental understanding of the overall work cell and its constituent physical components and sub-components. Likewise, understanding the relationships between these physical elements and how these elements relate to one another is critical to determining how the degradation of one element will impact the degradation of another. The National Institute of Standards and Technology (NIST) is researching various PHM technologies that aim to enhance decision-making at the factory floor to promote smarter maintenance and control strategies. Part of NIST’s research focuses on the decomposition of a work cell into a hierarchical structure to understand the physical and functional relationships among the overall system’s critical elements. This physical and functional decomposition is a necessity to promote a meaningful rollup of diagnostic and prognostic information from the lower levels to the higher levels of the hierarchy. The hierarchy seeks to encapsulate how the overall system, and its subsequent components, will behave when an element within the system is compromised or begins to fail. Neighboring components and sub-components could be subject to the ‘domino effect’ or the ‘ripple effect’, making diagnosing the root cause of a cascade alarms difficult without some reflective model of the system. This paper presents NIST’s efforts to develop a hierarchical decomposition methodology that will support PHM design and implementation within a complex work cell.