Multi-node power systems profiling with modified state estimation.

摘要

Electrical equipment poses a danger to maintenance personnel when the equipment must be serviced energized. While de-energized maintenance is preferred, for some tests, doing so will not provide the needed performance data to validate important performance metrics of the equipment. In particular, sites with a high cost of downtime (health care centers, data centers, continuous process manufacturing, etc.) identifying a problem that could affect series integrity of conductors (continuity) of service is critical. The infrared thermal imaging test is a common method of inferring the resistance of a connection by measuring the temperature of a component, and when normalized to current flow, can be used as a proxy for the effective resistance of the current path. Deviations from historical ratios of temperature to current may indicate a change in the junction resistance. The problem is that IR cameras require line-of-sight visibility with the energized electrical component to be tested. Many times due to available fault currents, fault clearing times and/or voltage levels, this cannot be done from a far enough distance to afford the operator safety even when wearing the maximum available PPE (personal protective equipment). The challenge has been how to perform this measurement safely and accurately, but still recognize that capital (for specialized test equipment) and expenses (for labor to perform testing) are frequently limited. Therefore, having a method of performing these tests throughout a facility, ideally performed from a central location while not requiring specialized equipment (leveraging the already installed IEDs [intelligent electronic devices]), would be a valuable solution. This research discovered a solution that reads low-resolution data from legacy IEDs but processes it to improve resolution to a point where ultimately it can be used to measure junction-to-junction impedances to sufficient resolution to detect failing connections. The research outlines the statistical analysis, filter design and conductor temperature normalization algorithms necessary to achieve this goal.