Data Driven Prognosis and Health Maintenance for Complex Industrial Processes with Production Optimization and Energy Savings

Abstract

Prognosis and Health Management (PHM) is an increasingly important area of research in the past few years. This is due to the requirement on economical health management for complex industrial processes that consists of a large number of units and equipment which operate together in order to fulfill their production requirement in terms of guaranteed product quality, improved production efficiency and reduced energy cost. This paper reviews some aspects on PHM and discusses possible future directions. In particular, three new areas of research are pointed out, namely the stochastic distribution control based active PHM, the passivity analysis based PHM and the multiple data driven based models PHM. The key feature of these novel PHM strategies is that they can link the remaining useful life (RUL) estimation with decision variables such as control loop set points in the complex industrial processes. Such a link forms a group of models by which pro-active PHM strategies can be implemented in practice.