IEEE Power & Energy Magazine - May/June 2018 - 12
big data analytics
making the smart grid smarter
According to google trends
(Figure 1), web search interest in the
term "analytics" had a jump in 2005, a
peak in late 2011, and then a slow decline. After the wave in the early 2010s,
the trend has flattened in recent years.
today, analytics attracts three to five
times more search interest than many
other buzzwords, such as artificial intelligence and machine learning.
the cover of the september/october
2012 issue of IEEE Power & Energy
Magazine reads "the emerging Field of
data Analytics: How data grows into
Knowledge." the issue featured five applications of data analytics, including
situational awareness, security assessment, state estimation, fault analysis, and
wind power forecasting. Five years later,
as we are revisiting this subject, data analytics is no longer a stranger to the power
and energy sector. the power industry is
gradually moving up the analytics ladder,
from descriptive analytics, to predictive
Digital Object Identifier 10.1109/MPE.2018.2801440
Date of publication: 18 April 2018
Search Interest (%)
analytics, and all the way up to prescriptive analytics. Meanwhile, many power
companies have accumulated large volumes of data from various sources for
advanced analytics. now big data analytics, or applying analytics to big data, has
become an emerging field.
As an interdisciplinary field, big data
analytics has a broad range of applications in the power industry. When preparing this issue, we understood that
it couldn't possibly be comprehensive.
therefore, we have handpicked a collection of novel ideas, powerful tools,
and new implementations to present.
the topics include data visualization,
system operations and planning, energy
markets, renewable energy integration,
and education. this issue also covers a
variety of data sources, such as household-level smart-meter data, synchrophasor data, distribution outage data, and
high-resolution weather data.
data visualization is a crucial step in
the analytics workflow. since a picture
is worth a thousand words, carefully designed image displays can help analysts
identify the proper predictive models,
offer comprehensive information to decision makers, and ease communication
with stakeholders. As energy data is getting bigger and bigger, many visualization
tasks can no longer be performed through
simple point and clicks on a spreadsheet.
in the first article, "Visualizing Big energy data," Hyndman et al. demonstrate
several modern visualization solutions
via three different data sets: householdlevel smart meter demand data, phasor
measurement unit (PMU) data, and probabilistic wind power forecasts.
As PMUs are transforming the way
we look at power transmission networks,
some power distribution companies
have started investigating applications
of distribution-level PMUs, a.k.a. microPMUs (μPMUs). A typical μPMU reads
120 times/s, which is 108,000 times
faster than a typical smart meter that
reads once every 15 min. How to translate these ultrahigh frequency data into
improved distribution operations and
planning becomes an emerging challenge to both industry and academia. in
Google Trends on Analytics, Artificial Intelligence, and Machine Learning
figure 1. Web search interest for the terms analytics, artificial intelligence, and machine learning. A value of 100 is the
peak popularity for the term. A value of 50 means that the term is half as popular.
ieee power & energy magazine