Self learning radio frequency monitoring system for identifying and locating faults in electrical distribution systems

摘要

Electrical faults are detected in electrical distribution systems (EDS) by detection and location of radio frequency (RF) emissions generated by the fault with multiple time-synchronized radio frequency monitors (RFM) distributed about the EDS. The RFMs are coupled to a self-learning, electrical fault monitor (EFM) that characterizes and/or locates electrical faults based on operating state (OS) patterns learned from transmission of test signals generated within the EDS. RF emissions data samples are characterized as safe operation (SO) states or potential electrical faults by accessing a base of stored knowledge concerning fault emission characteristics and/or synchronized time of arrival at each RFM. Information in the base of stored knowledge is updated to include new EDS operating states (OS). Confidence level associations, location of new radio frequency emission patterns and whether those patterns are indicative of safe operating (SO) conditions or electrical faults are stored in the base of stored knowledge.

主权项

1. A method for monitoring electrical faults in a power plant electrical distribution system, comprising:
monitoring and sampling in real time with an EDS electrical fault monitor (EFM) operation state (OS) radio frequency (RF) emissions data from electrical distribution system (EDS) equipment that are obtained with a plurality of time synchronized radio frequency monitors (RFM) that are distributed about the EDS and coupled to the EFM, the OS RF emissions potentially indicative of electrical faults in the EDS; storing in an automated data storage device that is coupled to the EFM OS data samples for each RFM, sampled data including the RFM identification, data sample time and data sample waveform characteristics; determining with the EFM in real time a likelihood of whether one or a combination of the respective OS data samples is indicative of an electrical fault by:
referencing in an automated data storage device coupled to the EFM previously stored information associating OS data with any one or more of EDS safe operation (SO), electrical faults or RF emission location within the EDS;comparing at least one stored OS sample reading from each RFM with respective stored association information relevant thereto and making respective EDS safe operation (SO) first confidence level determinations;comparing a combination of the at least one stored OS sample reading from each respective RFM with respective stored association information relevant thereto, if such combination information is available, and making respective EDS safe operation (SO) second or more confidence level determinations; andcombining all prior sequentially determined confidence levels information to derive an EDS safe operation (SO) overall confidence level; and causing the EFM to generate an alarm when any of the sequentially determined first through overall SO confidence levels is indicative of an electrical fault within the EDS.