Abstract: A system and method for learning and asserting what portions of a utility GIS network model are incorrect or flawed as they relate to real world conditions, and what the correct real world relationships are in the field is described. The system and method leverage available smart grid data to assess the quality of a primary (GIS) source data set; quality data renders derived analyzes across the utility valid, sound, and action worthy. The system and method utilize existing partially correct electrical network distribution model data and various non-specialized source data including smart meter, spatial, and customer information data collected from the network to test, validate and suggest corrections to the connectivity model. By forming putative ground truth assignments between utility components, the system tests the assumptions by examining the geospatial proximity and correlating voltage and event data over time to form refined hypothesis.

Abstract: A system and method for learning and asserting what portions of a utility GIS network model are incorrect or flawed as they relate to real world conditions, and what the correct real world relationships are in the field is described. The system and method leverage available smart grid data to assess the quality of a primary (GIS) source data set; quality data renders derived analyses across the utility valid, sound, and action worthy. The system and method utilize existing partially correct electrical network distribution model data and various non-specialized source data including smart meter, spatial, and customer information data collected from the network to test, validate and suggest corrections to the connectivity model. By forming putative ground truth assignments between utility components, the system tests the assumptions by examining the geospatial proximity and correlating voltage and event data over time to form refined hypothesis.