Historical alarm and near-realtime facility data analysis

This paper describes the current effort to analyze historical alarm data and discusses applications and methods for secure, near-realtime communications of data. The objective of analyzing historical alarm data is to determine if trends in system or subsystem performance can be detected that may affect system maintenance activities and operational costs. A wealth of data is potentially available, and this paper discusses initial efforts to extract useful information from that data. As an example, some sensors may report excessive nuisance alarms that correlate well with certain weather conditions, and sensor re-installation may prove remedial. In the future, fielded sensors that report ID codes could also report serial numbers that map into a vendor's database with manufacturing dates and recall information. In a more timely context, system maintainers often require the assistance of others-either component manufacturers, system integrators, or simply "recognized experts"-to diagnose problems and effect appropriate repairs. The ability to "reach back" to assistance personnel from a remote location is highly desirable.