IEEE Power & Energy Magazine - November/December 2017 - 27

Increasingly, forecasting is being used not only to schedule
thermal and hydropower generation but also to
optimize energy storage assets.
time. This places a significant value on improving wind and
solar power forecasting in the day-ahead unit commitment
process-although the industry is also pursuing advances in
short-term intraday and intrahour forecasting.
Solar power forecasting follows one of two approaches:
1) PV performance models aim to use irradiance forecasts
to determine the performance of solar generation facilities, and 2) statistical models use artificial neural networks,
regressions, and other statistical methods to predict future
behavior based on historical data. Hybrid models are being
developed that incorporate irradiance forecasts into neural
networks. In all cases, forecasts can be created for individual
power plants or for an ensemble of plants based on the purpose of the forecast.
A forecast tuned for an individual plant might be designed
to maximize market value, whereas a regional forecast used
by a system operator might be designed to minimize risk. Many
early forecasting programs focused on creating deterministic predictions for power output, providing a single value
for a moment in time; however, more recent advances use
probabilistic methods to incorporate information such as
the upper and lower bounds of possible forecasts and/or confidence intervals around a particular forecast value. Increasingly, forecasting is being used not only to schedule thermal
and hydropower generation but also to optimize energy storage assets.
MISO has been using weather-based load forecasting since
the launch of its energy market in 2005. In 2009, with wind
penetration increasing, MISO began implementing weatherbased wind forecasts into all aspects of markets and operations, which is one of many factors that enabled penetration
levels to exceed 25% of annual energy in 2016 for MISO's
northern region. Building on the success of wind forecasting,
MISO started forecasting solar in 2016.
MISO's wind and solar forecasts are a blend of five
forecasting methods designed to accurately predict generation throughout various time scales and levels of
fidelity. Maximum economic potential is obtained by looking at the feasible weather-based output of wind and solar
power plants along with transmission limitations to gain
a realistic understanding of future wind and solar output
throughout MISO. In 2010, MISO designed and implemented a market mechanism to take advantage of advances
in wind technology that make the concept of nondispatchability less applicable. The introduction of dispatchable
intermittent resources (DIRs) allows such resources to
fully participate in the energy markets and has resulted
november/december 2017

in more economic and reliable grid operations. Currently, 85% of the wind resources in MISO are registered as DIRs.

Fact Four: While Our Electric Power
Markets Were Not Originally
Designed for Variable Renewables,
They Can Be Adapted
In the 1990s, as capacity margins and natural gas prices
increased, electricity regulators, economists, and engineers
began to implement new ways to reduce electric power costs.
Large portions of electricity systems in the United States and
Europe, along with those in several other countries, were subjected to regulatory reforms that created competitive wholesale markets. Transmission owners gave operational control
of their assets to ISOs, which established centralized markets for energy, ancillary services, and capacity that aimed to
minimize the cost of wholesale power in serving load while
meeting the complicated operating and reliability requirements of power systems within different time frames (from
seconds to years). In some regions, retail competition was
also introduced and sustained.
The ISO markets-which vary dramatically in geographic
size (among many different countries), from one state or
province to very large multistate regions and entire nations-
have generally been effective at reflecting changes in the
fixed and variable costs of power supply and the resulting
market prices while minimizing market power. As a result,
market prices fluctuate, sometimes significantly, but usually
within regulated limits. Although the transparency of market rules and price results have been an ongoing issue, there
is no question that market participants, including renewable
energy suppliers, have much more information about how
they are valued within a market than before these institutional changes were implemented. In principle, the markets
are technology agnostic, but they need to adapt to public
policies as well.
These ISO markets have been very effective at integrating a variety of resources and have also been flexible enough
to handle dramatic changes in the composition of the generation fleet as well as extreme events such as frigid winter
conditions, hurricanes, and drought. They have an excellent
reliability record. However, the addition of wind and solar
to the system, combined with the shale gas revolution in the
United States and overcapacity in some regions of the United
States and Europe, has-with a few exceptions-suppressed
electricity prices for several years.
ieee power & energy magazine

Table of Contents for the Digital Edition of IEEE Power & Energy Magazine - November/December 2017