Symbolic Time Series Analysis in Aladdin for On-line Condition Monitoring

The objective of this paper is to briefly introduce a symbolic-based approach for on-line condition monitoring (OCM) of physical processes and its ongoing evaluation within a large-scale decision support monitoring system such as Aladdin. The premise is that symbolic manipulation of process information is rich in analytical expressiveness, computational power, and on-line performance, hence significantly enhancing the analysis portfolio for the next-generation of decision support systems. By manipulating symbolic information of plant conditions, instead of unprocessed plant measurements, the computational complexity required to implement advanced monitoring functions is kept within acceptable limits, while significantly improving the applicability, responsiveness, and effectiveness of deployed systems. Under a recently-established strategic partnership for on-line condition monitoring, diagnostics, and prognostics between the the Idaho National Laboratory (INL) and the Halden Reactor Project (HRP), a novel analytical framework and supporting methods based on Symbolic Time Series Analysis (STSA) will be implemented within the Aladdin code that will enable conducting decision-making functions at a high level of abstraction/intelligence, serving as a supportive, vigilant agent to operators for assessing process conditions and suggesting condition-based maintenance (CBM) activities. Benefits from this international cooperation include the development of novel methods for the symbolic analysis of signals and for the symbolic integration within a supervisory scheme of existing data-driven analysis techniques.