The Advancement of Adaptive Relaying in Power Systems
Protection
2.1 Abstract
The electrical distribution system in the United States is considered one of the most
complicated machines in existence. Electrical phenomena in such a complex system can
inflict serious self-harm. This requires damage prevention from protection schemes. Until
recently, there was a safe gap between capacity to deliver power and the demand. Therefore,
these protection schemes focused on dependability allowing the disconnection of lines,
transformers, or other devices with the purpose of isolating the faulted element. On some
occasions, the disconnections made were not necessary. The other extreme of reliability calls
for security. This aspect of reliability calls for the operation of the protective devices only for
faults within the intended area of protection. There is a tradeoff here; where a dependable
protection scheme will assuredly prevent damage, it is prone to unnecessary operation which
can lead to cascading outages.
Where a secure scheme will not operate unnecessarily, it is prone to pieces of the system
becoming damaged when relays fail to operate properly. With microprocessor based relaying
schemes, a hybrid reliability focus is attainable through adaptive relaying. Adaptive relaying
describes protection schemes that adjust settings and/or logic of operations based on the
prevailing conditions of the system. These adjustments can help to avoid relay
miss-operation.
Adjustments could include, but are not limited to, the logging of data for post-mortem
analysis, communication throughout the system, as well changing relay parameters. Several
concepts will be discussed, one of which will be implemented to prove the value of the new
tools available.
2.2 Power Systems History
The distribution of electricity in the United States of America can be traced back to
September 4th, 1882, when Thomas Edison opened the Pearl Street Station in lower
13
Manhattan.
Serving about a quarter of a square mile, this direct current (DC) system was primarily
used for lighting in the financial district of New York. Edison showed that it was possible to
efficiently provide electricity from a central generating station. The issue with DC systems
was that the end consumer had to be located within a few miles of the generating station. The
problem was that the low voltage used with this type of distribution led to higher currents and
higher losses on the lines used to distribute the electricity. This forced the generating plants to
be small which reduced efficiency, and it meant that only small distribution systems in
densely populated areas would be effective [1].
In distribution systems the voltage is held constant and the current flowing through the
lines depends on the load being served. The losses associated with the lines used to distribute
the electricity vary with the square of the current running through the lines. So if the current
through the lines doubles, the losses associated with the lines actually quadruples. At the time
Edison started implementing his systems, there was no way to easily change the voltage in a
DC system.
The ability to vary the voltage of the distribution lines would allow for the reduction of
current during transmission, which was being developed within alternating current (AC)
systems. AC allowed for transformers to increase the voltage required by transmission and a
reduction to a voltage level that is safe for end consumers to use; this significantly reduces
the losses of sending electricity over longer distances. Nikola Tesla was the pioneer of this
AC technology as well as the concept of polyphase distribution [1].
These competing strategies of electrical distribution, AC and DC systems, led to what is
commonly known as the Battle of the Currents. Thomas Edison, owning the patents for DC
systems, argued that AC and the higher voltages associated with it was unsafe. At the same
time, however, George Westinghouse was building AC transmission lines that stretched for
miles. This, along with Nikola Tesla’s development of an AC motor among other
developments, led to the ultimate victory of AC systems. This victory of alternating current
led to the electrical distribution system we have today in which large generating stations
delivering power over long distances at high voltages, which is both economical and efficient
14
in comparison to the original DC systems [4]. This did, however, lead to several engineering
issues to which solutions are still being developed today.
The AC system pioneered by Westinghouse and Tesla has developed into one of the
most complex machines in the world. The growth started with many small independent
systems. For reliability purposes, these systems were interconnected. This interconnection of
many small systems meant that the number of machines necessary for reserve operation
during peak loads was lowered. The interconnection also enabled utility companies to get the
cheapest possible power from their neighbors. These interconnections grew into the massive
system which we have today. There are issues that arose with the creation of this massive
system; these issues include higher fault currents, cascading failures in which multiple
smaller systems are affected when the problem only occurred in one of them, and a very
delicate balancing act that occurs between systems. The planning that goes into this system,
especially the protection of the system itself, is very complicated [5]. This system is generally
broken down into generation, transmission, and loads. The transmission portion is divided
into transmission, subtransmission, and distribution; each having different voltage levels
controlled using transformers.
2.3 Generator
Generators are used to convert different forms of energy into electrical energy. Most
generators in use today convert mechanical energy into electrical energy using magnetic field
interactions. This mechanical energy is generally provided in the form of a spinning prime
mover. The prime mover usually has a magnetic field associated with it, and its spins within
the stator coils; the stator is the stationary portion of a generator, and the field on the rotor
induces currents within those stator coils. The spinning action can be provided using a steam
turbine where some source of heat boils water to drive that turbine, or in the case of a
hydroelectric dam, water could spin a turbine directly. Sometimes internal combustion
engines can also be directly coupled to a prime mover. Steam power plants generate their heat
by burning coal, natural gas, or oil as well as using nuclear reactions to generate heat. In the
case of using a prime mover type generator, the speed at which that generator spins is
extremely important because it determines the electrical output frequency. The great thing
15
about all the types of generation discussed so far is that their output levels can be controlled
by varying the amount of energy put into the prime movers.
Other, less controllable, forms of generation include renewables like solar and wind
power. Solar power can be in the form of photovoltaic energy which needs to be converted
from DC to AC to contribute to the system, or solar thermal which can be incorporated like
any other thermal based generation. Wind power generates electricity with a prime mover, but
because wind speeds are not constant the electricity must be conditioned using power
electronics to ensure the output has the correct voltage and frequency. The main issue with
these types of generation is that there is no way to control their output, so there isn’t any way
to predict accurately how these sources will contribute. Another issue is that when small scale
projects are implemented and feed energy back into the grid, current flow can change
direction which may affect the operation of certain types of protective relays. So while it is
good to have a contribution from renewable resources, there is a tradeoff in the predictability
of operation.
2.4 Transformers
Transformers are an essential part of the electrical distribution system, as discussed
earlier. Generation is generally done at voltage levels between 13.8 kV and 24 kV.
Consumption of this electricity is generally done at voltage levels between 110 V in homes
and up to 4160 V in large industrial plants. Transmission of electricity can occur at levels of
115 kV to 765 kV in the United States, and go as high as 1 megavolt in other parts of the
world. Transformers are what make this wide range of voltage level capabilities possible.
Without transformers and the ability to vary voltage levels, it would be much less efficient to
transmit power over great distances.
Transformers operate based on Faraday’s law of induction. Faraday’s law states that if
magnetic flux passes through a coiled conductor it will induce a voltage in that conductor that
is directly proportional to the derivative of that flux and the number of turns in the conductor
coil. In a transformer, a flux is induced by a primary coil that is wrapped around a
ferromagnetic core. The ferromagnetic core is used to give a path to the flux that has a high
permeability. There is then a secondary coil which is wrapped around the same ferromagnetic
16
core which has a voltage induced on it by the flux traveling through the core. The amount of
flux is dependent upon the voltage and number of turns on the primary coil, and the voltage
on the secondary coil is determined by the flux and the number of turns in the coil. Because
the number of turns directly determines the ratio of the primary voltage to the secondary
voltage, this ratio is commonly referred to as the turns ratio.
In an ideal world, a transformer would take a voltage from one level to another without
any type of losses, but this isn’t the case. Transformer losses include copper losses, eddy
current losses, hysteresis losses, and leakage flux. Copper losses are due to the resistance
associated with the coil of wire itself and are proportional to the square of the current flowing
through the coils of the transformer. Eddy currents are losses from unwanted currents induced
on the core of the transformer and are proportional to the square of the voltage across the
terminals of the transformer. Hysteresis losses are due to the rearrangement of magnetic
domains in the core and are a function of the voltage applied to the transformer. Copper, eddy
current, and hysteresis losses are all consumers of real power and are modeled as resistances.
Leakage flux is simply flux that is not captured by the core and is passed to the other coil in
the transformer. It is a function of the current flowing through the coils. Leakage fluxes are
consumers of reactive power and are modeled as inductive impedances. These losses,
however, are small in comparison to the losses that would occur in transmission if
transformers were not available.
2.5 Adaptive Protection Schemes
Adaptive protection schemes are the result of the application of microprocessors in the
area of protective relays and are growing in importance in the electrical power systems in the
United States and worldwide. These schemes may have complicated implementations as far
as programming, but their concepts can be explained fairly easily. Many of these concepts are
simply expansions on previous protection applications. Several of these concepts will be
explored, including previous system events that could have been mediated with the help of
these new concepts.
2.6 Differential Protection
Differential protection schemes are set up simply to check for any difference between
17
two quantities at a given instance. Limitations on time synchronization made this
implementation only reasonable for equipment protection and difficult for other applications
until the recent advent of GPS signals. On the other hand, for signals collected from distant
points into a system, the burden of communication made the implementation of differential
protection difficult or unattainable. While this type of protection could be useful in detecting
a difference in current from one substation to the next, historically its application required the
two measurements to be taken very close to one another because of the constraints on
communication. So the scheme was limited generally to transformer and generator protection.
Before microprocessor-based systems and IEDs the nature of these two types of protection
were limited by several issues, specifically mismatches in current transducers.
Percentage differential protection of transformers finds the difference between two
current levels that should be close to equal. This is done by putting the output of two current
transducers in parallel with a relay that detects current flow. With the proper connection of
polarities of the CTs, if both the secondary currents are equal, no current will flow through
the relay. Issues with this include the previously stated mismatch due to CT limitations, as
well as CT error mismatches, transformer’s magnetizing currents, and tap changing elements
which will change the effective ratio of the transformer itself. These problems are alleviated
by establishing a restraint current. The restraint current is simply the average of the secondary
currents. The relay operates when the current that it sees exceeds a certain percentage of the
restraint current. The smaller the percentage required, the higher the sensitivity of the relay.
Additionally, magnetizing current can cause errors during energization and fault removal,
and its harmonic content can cause issues as well. The magnetizing inrush current is caused
when an unloaded transformer is brought online and needs to gain the flux necessary for
steady state operation to occur. This can also happen when a fault is cleared and the current
changes significantly. On top of magnetizing current, transformer over-excitation can also
become a problem. The saturation during these times can cause the differential relay to react
unnecessarily. Lastly, if there is a fault outside of the transformer it is possible that the CTs
will saturate at different current levels. If the difference between these saturation levels is
large, the differential relay will operate unnecessarily for a fault that is not within the
18
transformer.
Computer based relays can offer solutions to all of these issues to significantly increase
the accuracy of operation for a percentage differential protection scheme. The main error
caused by mismatched ratios is quickly mediated by the fact that a computer can take the
output from any CT and scale it according to the turns ratio necessary for the secondary
currents to match up. In fact, the CTs do not even need to have secondary currents that are
close to each other, but simply take the current low enough for an analog to digital conversion
to be given to the computer. CTs can then be chosen based on their accuracy as well as their
saturation limits in order to prevent some of the other issues discussed.
The computer itself can be given inputs on different phenomena going on to prevent
unnecessary operation. For example, if the transformer is being brought online the computer
can be set to recognize and ignore the issues brought about by the inrush currents. In the case
of a transformer with tap changing occurring, the relay could be set up to allow for any inrush
currents expected as well as change the necessary ratios on the CTs. With certain
communication parameters, it could even be possible for a computer based differential relay
to recognize when faults outside of the transformer will affect operation of the relay. In the
field of communication, there are many opportunities to improve protection schemes.
