Situational awareness for distribution grids

Until now, distribution grids have been operated with only limited information about real-time load conditions. The reason is that, traditionally, those systems were overdimensioned and intended to be operated only in a passive fashion. With the emergence of distributed energy sources, electric vehicles and demand response programs, electricity distribution has been undergoing fundamental changes posing substantial challenges to grid operators.

A key to manage those challenges is to improve the visibility on the actual state of the network. IBM Research - Ireland has developed a set of technologies for enabling such services. The fundamental idea is to exploit redundancy, e.g., from sensoring through telemetry and smart meters and from prior knowledge of the expected loads, and compute an estimate of the unavailable measurements with associated confidence bounds quantifying the uncertainty.

This allows us to:

  1. Infer real-time load conditions even in network areas where sensoring is sparse and
  2. Detect, localize and diagnose measurement inconsistencies, e.g., related to outages, unaccounted generation or customers assigned to wrong phases.

Illustration of the approach

outage localisation

IMAGE 1: Example of the voltage estimated through a distribution feeder based on SCADA measurements available only at the head of the feeder and at 3 reclosers and on prior models of the load at the buses.


IMAGE 2: Limited sensor information (head of feeder + 3 reclosers) can be used to localise an outage to an area of the feeder, as well as to estimate the pattern of the anomaly, which in this case is negatively correlated with a residential load pattern (the data show less load than expected ). The spatial resolution of the anomaly localisation, that is the size of the areas where any anomaly can be localised, is strictly related to the number and location of available sensor measurements: more sensors allow a better resolution.


  1. Integrated state estimation and load modelling for distribution grids with ampere measurements
    Francesco Fusco, Mathieu Sinn,
    4th European IEEE Conference on Innovative Smart Grid Technologies (ISGT), pp. 1-5, 2013.
  2. General bad data identification and estimation in the presence of critical measurement sets
    Francesco Fusco,
    IEEE PES General Meeting, pp. 1-5, 2014.
  3. Sensor Placement for Optimal Estimation in Power Distribution Grids
    Francesco Fusco, Jonas C. Villumsen,
    IEEE PES Conference on Innovative Smart Grid Technologies (ISGT) 2015.