IEEE Electrification - September 2020 - 15

Machine learning algorithms can rapidly identify the grid
status with reduced numbers of observations, and reinforcement learning can be used to explore optimal inverter control strategies.
Take the emerging "electric spring" concept as an
example (Figure 12). Analogous to a mechanical spring,
an electric spring is a power electronics device that
can 1) provide electric voltage support, 2) store electric
energy, and 3) damp low-frequency oscillations. It can
compensate for the voltage drop in the distribution
network and stabilize the frequency. To achieve these
goals, electric springs need to accurately infer the grid
status and respond rapidly to power flow dynamics:
both ultrafast grid impedance estimation and ultrafast
control are needed. Increasing the switching frequency
of the electric spring devices can improve the control
bandwidth, reduce the electromagnetic interference
filter size, accelerate the overall control loop, and
enable other grid response functions that have not
been demonstrated before. Many similar active compensation devices have been proposed. Their performance and functionality can all benefit from higher
switching frequency.

electronics at the grid edge, the voltage quality and
current quality are cross-coupled and cannot be separated. High-frequency power electronics produces both
low-frequency harmonics (below 1 kHz) and high--
frequency harmonics (above 1 kHz). The low-frequency
harmonics can be mitigated by advanced control, but
the high-frequency harmonics appear in the distribution grid and cannot be easily eliminated, especially in
the distributed grid without isolation transformers.
Fixed frequency operation usually leads to groups of
harmonics around the integer multiples of the switching frequencies. Hysteresis control, used in smaller
converters, leads to a highly distorted frequency spectrum. If the switching frequency is close to a system
resonant frequency, it causes a large high-frequency
ripple on the voltage. An increasing penetration of
DERs with power-electronic interfaces will lead to an
increasing level of high-frequency harmonics. The
physical principles and system consequences of highfrequency harmonics remain unclear. Standardized
methods for measuring, islanding, and characterizing
high-frequency current and voltage harmonics are not
yet available.

Power Quality, Low-Frequency Oscillation,
and High-Frequency Harmonics

Packetized Energy and Peer-to-Peer Energy Exchange

Power quality measures the electrical interaction
between the power grid and the sources and the loads.
Conventionally, power quality comprises two parts: 1)
voltage quality, the way in which the supply voltage
impacts equipment operation, and 2) current quality,
the way in which the equipment current impacts the
system operation. With a large amount of power

As the deployment of DERs expands, packetized power
microgrids are emerging as promising solutions to effectively
incorporate DERs (Figure 13). The full benefits of a microgrid
can only be realized when energy can be freely exchanged
and traded between sources and loads with economic incentives. Dynamic peer-to-peer energy trading requires rapid
and precise balance of the sources and loads within a distribution network (Figure 14).

Solid-State
Transformer

GPS

Grid-Forming Inverter With Electric Spring
Figure 12. Grid-forming inverters and energy routers at the edge (see Hui et al.).

IEEE Elec trific ation Magazine / S EP T EM BE R 2 0 2 0

IEEE Electrification - September 2020

Table of Contents for the Digital Edition of IEEE Electrification - September 2020