|
 From
Electrical World Magazine, "Engineer’s Notebook" by Walter
Sass
March/April 2001 — While there
are few conceptually simpler ways to get more useful capacity from
transformers than to strategically parallel them, in reality, transformer
paralleling is anything but simple. For example, two 40–MVA
transformers could not individually service two buses with 10– and
50–MVA loads. They could if they were tied together, but virtually
any difference between them (from winding variations to MVA ratings)
could lead to equipment overloads, wasted energy, and operational
instability.
Parallel configurations where the primaries are connected to different
buses, pose even graver difficulties. And combining voltage–regulating
controls that were designed to work on independent transformers may
not work at all.
Problem Explained
Power supply sharing is one of the oldest and most persistant problems
in electrical engineering. Whether the power sources are transistors
or hydro–electric generators, problems arise when their outputs
get connected together.
For example, it might be tempting to combine two independent circuits,
each consisting of one battery powering one lightbulb, to best distribute
the energy from both batteries. The combined circuit of parallel
batteries and bulbs might work, but there may be unintended consequences.
The stronger battery may charge (dump energy into) the weaker one.
How much source–to–source energy transfer takes place
is a function of the source impedances and voltage characteristics.
In specific cases, such as identical batteries with identical states
of discharge, there might be no source–to–source currents.
Otherwise, the undesired energy transfer would occur unchecked, unless
some form of regulating circuitry – like blocking diodes for
example – stopped it.
In the case of output–interconnected distribution transformers,
undesired energy transfer can similarly occur. Unlike the battery
example, there is no possibility of adding diodes or other regulating
circuits. In fact, the only regulating means available are usually
the load tap changers (LTCs) that regulate the line voltages going
to the loads.
LTCs can’t change transformer impedances or block currents,
and their primary line regulation functions cannot be ignored, so
their effectiveness in facilitating transformer paralleling is inherently
limited. In fact, transformers with substantial impedance differences
(10% or greater expressed on the same base) may not be paralleled
safely.
But let’s assume it’s possible to use sophisticated controls
to operate paralleled LTCs to somehow optimally minimize energy transfer
(circulating current) and maintain line regulation. In the example
above, this would be equivalent to paralleling two batteries by adjusting
small series potentiometers to match voltage outputs of each. The
pots could be adjusted jointly to regulate the voltage going to the
lightbulb loads, and adjusted differentially to reduce circulating
current between the batteries.
Unfortunately, AC power circuits are more complicated. Using LTCs
to regulate the output voltages of parallel transformers primarily
acts to share reactive power (VAr) between them, not real power.
The optimal paralleled LTC control scheme may be required to restrict
VAr burdens on each transformer, further compromising its ability
to distribute the real power loads as intended.
Despite these problems and limitations, transformer paralleling is
becoming an increasingly common utility practice. Even if inevitable
circulating currents waste energy and heat up the transformers, or
if shifted VAr burdens run the transformers nearer their VA limits,
the possibility of picking up more load with existing equipment is
very compelling.
The key to operating transformers in parallel safely will be more
LTC control strategies that truly optimize line regulation, minimize
circulating currents, and limit VAr burden. Such control strategies
will require distributed observation of bus interconnections at the
substation and coordinated operation of LTCs that bypass or eliminate
their local controllers. Further, transformer manufacturers can expect
their gear to run hotter and experience more overstress as paralleling
becomes more common.
|
|
