IEEE Power & Energy Magazine - May/June 2022 - 63

Tailor-made forecasting models can now be designed
and trained to characterize the different dynamic
behaviors of each type of customer.
peaks or to force them to follow the availability of renewable
resources to improve the system efficiency.
In the same vein, this information is useful to perform expost
analysis and customer diagnosis. An electrical utility, or a
consumer itself, can compare its historical use of energy with its
cluster peers. By applying different statistical techniques, we can
identify, with a prespecified confidence level, possible anomalies
derived from defective appliances or even detect fraud.
Clustering tools can also help to design more reliable electrical
systems. The progressive adoption of new technologies,
such as household solar panels, electric vehicles, home batteries,
and so on, has a direct impact on customers' consumption
patterns that can be unveiled and characterized by clustering
tools. This information can be extrapolated into the
future, under different integration level scenarios, to quantify
the technical requirements of future generation systems
and transmission and distribution networks.
Finally, these tools can be used in demand forecasting
to make a more precise purchase in the wholesale electricity
market. Tailor-made forecasting models can now be
designed and trained to characterize the different dynamic
behaviors of each type of customer.
Conclusions
Most energy systems worldwide, including the European
energy sector, are undergoing a major transformation toward
a new model where energy efficiency and customer empowerment
are increasingly important.
How customers consume electricity is constantly changing.
The electrification of the demand, distributed renewable
generation, electric cars, and telecommuting are having a
major impact on consumption profiles, which are growing
increasingly heterogeneous. Thus, the customer segmentation
needs to be updated frequently and adapted to the new data
received from smart meters. The frequency of the measurements
and the number of monitored customers are expected
to continuously increase, creating opportunities for extracting
more information to gain insights into consumer behavior.
This new paradigm entails major challenges and opportunities
for both consumers and companies participating in the
electricity sector. There is a clear need to develop new and
efficient analytical clustering techniques to cope with larger
data volumes that include both structured and unstructured
data types (consumption and weather time series combined
with digital images, sensor data, social media, voice recordings,
and so on). In this regard, emerging clustering techniques
may/june 2022
in this field are expected to combine traditional statistical techniques
with machine learning approaches for capturing both
static and dynamic electricity consumption patterns.
Acknowledgments
The authors gratefully acknowledge the support from the
Spanish government and the Madrid government (Comunidad
de Madrid).
For Further Reading
S. Aghabozorgi, A. S. Shirkhorshidi, and T. Wah, " Timeseries
clustering: A decade review, " Inf. Syst., vol. 53, pp.
16-38, Oct./Nov. 2015, doi: 10.1016/j.is.2015.04.007.
A. M. Alonso, F. J. Nogales, and C. Ruiz, " Hierarchical
clustering for smart meter electricity loads based on quantile
autocovariances, " IEEE Trans. Smart Grid, vol. 11, no. 5, pp.
4522-4530, 2020, doi: 10.1109/TSG.2020.2991316.
A. M. Alonso, P. Galeano, and D. Peña, " A robust procedure
to build dynamic factor models with cluster structure, "
J. Econometrics, vol. 216, no. 1, pp. 35-52, 2020, doi:
10.1016/j.jeconom.2020.01.004.
J. Caiado, E. Maharaj, and P. D'Urso, " Time-series clustering, "
in Handbook of Cluster Analysis, C. Hennig, M.
Meila, F. Murtagh, and R. Rocci, Eds. London, U.K.: Chapman
and Hall, 2015, pp. 262-285.
R. J. Hyndman, X. Liu, and P. Pinson, " Visualizing big
energy data: Solutions for this crucial component of data
analysis, " IEEE Power Energy Mag., vol. 16, no. 3, pp. 18-
25, 2018, doi: 10.1109/MPE.2018.2801441.
T. Liao, " Clustering of time series data: A survey, " Pattern
Recognit., vol. 38, no. 11, pp. 1857-1874, 2005, doi:
10.1016/j.patcog.2005.01.025.
Biographies
Andrés M. Alonso is with Universidad Carlos III de Madrid,
Getafe, 28903, Spain.
Eduardo Martín is with Endesa Energía, Madrid, 28042,
Spain.
Alicia Mateo is with Endesa Energía, Madrid, 28042, Spain.
Francisco J. Nogales is with Universidad Carlos III de
Madrid, Leganés, 28911, Spain.
Carlos Ruiz is with Universidad Carlos III de Madrid,
Leganés, 28911, Spain.
Andrea Veiga is with Endesa Energía, Madrid, 28042,
Spain.
p&e
ieee power & energy magazine
63
IEEE Power & Energy Magazine - May/June 2022

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - May/June 2022
