Robust automated determination of the hierarchical structure of utility monitoring systems

摘要

An automated hierarchy classification algorithm (300) that searches for a child monitoring device's parent in a utility monitoring system (102) by segmenting the device data measured by a given device pair and calculating a segment correlation coefficient for each data segment. Devices to be placed in the hierarchy are filtered by calculating (304) the variance of their device data and eliminating devices with a low variance (306, 308). Devices are ranked by computing the sum of squares of their device data and ordering the devices accordingly from highest to lowest. The device data is segmented and segment correlation coefficients are averaged to produce an overall correlation coefficient. Criteria are evaluated to determine whether a device pair is linked. A correlation coefficient is calculated using the complete data series of a device pair (316), and the solution produced by this approach is compared (322) with the solution produced by segmenting the device data (314). If the solutions disagree, a likely candidate is determined (324) from a fuzzy logic module (120).