Noise Control in Novel Power Electronics for SMART Grid

As society and technology are developing, the amount of electrically powered
devices is ever increasing. The traditional electrical grid, structured in a hierarchical way, is not capable of sustaining the rapid development and implementation
of a more dynamical consumer. By incorporating solar panels and
wind turbines at farm, businesses and even households the consumer becomes a
producer as well. This transforms the conventional grid into a more dynamical
and also distributed one. The project of which this thesis was a part of, deals
with the integration of renewable energy by applying a new architecture that
enables point-to-point power transmission and thus reduces instabilities and
improves dynamical behavior. The key objective for the University of Twente
(UT) is to find the best options for reducing interference, associated with fast
switching semiconductors as applied in the novel converter type.
This thesis starts with the development of a theoretical model based on the
control signals applied to switching power devices. The model can be used to
predict and estimate the conducted Electroagnetic Interference (EMI) generated
in switching devices using sinusoidal Pulse Width Modulation (sPWM)
waveforms. The model was verified in a conducted emission measurement.
Combining the model with a radiated emission estimation model, gives a description
of a noise source that is applied in the newly developed multi-level
converter.
For application in multi-level converters a sPWM generator was developed,
which with its flexibility is used to perform measurements on Galium-Nitride
(GaN) and Silicon-Carbide (SiC) based DC/AC converters, and multilevel converters.
In case of conducted EMI measurements, the results were used to verify
the theoretical model. In case of the magnetic radiated EMI measurements,
a time domain measurement technique was developed that is comparable to
using a traditional EMI receiver. The technique reduces measurement times
from minutes to several seconds per orientation and placement. In case of large
stacked multi-level converters it was deemed necessary to asses the magnetic
radiation produced in such a structure. The electric field measurements are
eventually used together with the developed mathematical model to determine
the effective radiation efficiency of the system under test.
The developed technique of determining the effective radiation efficiency,
together with the concept for optimal placement of EMI in fully integrated systems are considered to be part of the main contributions of this thesis and
can be seen as Mitigation through EMI placement. Another mitigation technique,
which is a fairly classical Electromagnetic Compatibility (EMC) one,
consist of implementing a filter along the propagation path of the disturbance.
Part of the research objectives was developing behavioral circuit models based
on full wave model and thus optimized filters implementing nano-crystalline
materials.
This thesis contributes with the development of an automated method for
back annotating field effects into equivalent circuit simulations. Part of this
work was the development of Gauss-Newton optimization algorithm that can fit
impedance curves to equivalent circuit elements. This was applied to measurements
and full-wave 3D simulations of relatively simple components like capacitors,
which showed optimal capacitor placement can be investigated through
circuit simulations rather then Electro-Magnetic (EM) field simulations. Extending
the research to more complex structures and components, required the
development of a 3D full wave high frequency models. This has been done for
a two phased sectionally winded Common Mode Choke (CMC), incorporating
complex permeabilities through a dispersion model.
Overall it can be concluded that this thesis has contributed to the development
of the future electrical grid, by investigating components that are
attributed to the Multifunctional Multilevel Modular Converter (M3C), which
is considered to enable the development of the ’smart grid ’. Much work still
needs to be done, from refining and applying the developed measurement techniques
to larger and fully integrated systems to developing more accurate and
faster fitting algorithms for determining equivalent circuit component values.
Also the effect of proposed mitigation techniques on various functionalities of
the M3C have to be investigated.