Detect Energy Theft: The primary use of interval meter
data is automation in billing and settlement. Usually, both
customer-based electronic meter and concentrator meters are
installed in the distribution system. The concentrator meters
aggregate and track data from multiple customer-based
meters. By analyzing an avalanche of paired data from
concentrator meters and customer-based meters, irregular
energy loss patterns could be easily identified [4]. If we add in
historical energy theft information, it is possible to further
filter out reasonable changes in energy consumption trends
and detect potential energy theft.
Detect EV and Rooftop Solar Integration: In the past,
residential customer load profiles have been dependent upon
larger loads, which have been heating & cooling systems.
With changes to these characteristic load shapes from new
technologies such as electric vehicles and rooftop solar PV
there is potential to detect when a customer adopts such a
system as well as system performance. Power flow returned
to the grid is an easy identifier of a photovoltaic system
installation and the magnitude (although the signal is mixed
with load) can be identified by recognizing the season and
expected load of the customer. Dependent on customer
behavior, the EV detection would use fuzzy logic to identify a
characteristic change in baseload. Tracking these two
adoption trends is critical for structuring power purchase
agreements, planning infrastructure upgrades, and informing
state policies.
Develop More Granular Load Forecast: The availability of
interval meter data also creates the opportunity to develop
more accurate and granular load forecast in terms of both
location and time. More accurate forecasts will benefit both
transmission and distribution system operations. A more
accurate spatial forecast in the transmission system has huge
impact on the unit commitment and dispatch process. A
reduction of 1% in mean absolute percentage error (MAPE)
could decrease annual variable generation cost in the United
States by approximately $160 million [5]. Currently, the
utilities typically submit aggregate load forecast in their
service territories to the market/system operator. The market
operator would then disaggregate the forecast to the
substation level by using weather forecast information and
historical load distribution factors. With more granular
customer consumption and behavior information, the utilities
are in a great position to improve the spatial load forecast
accuracy through mining both electricity consumption data
and weather information.
Develop More Granular Renewable Generation Forecast:
With rapid penetration of distributed renewable generation in
distribution system, the need for accurate distributed
renewable forecast becomes critical. For example, as
distributed renewable (mostly solar photovoltaic) penetration
levels in distribution circuits reach 15% and beyond in
Hawaii and Southern California, the distributed generation
starts to have significant impacts on distribution systems
planning and operation. An accurate spatial joint load and