Basler Electric BE1-CDS240 General Information - Manual (Page 17)

Problem 6: Transformer Energization Inrush and Overexcitation
General
When the zone of protection includes a transformer, excitation and energization inrush currents appear as
differential current because they flow into the zone of protection and not back out again. Normally,
transformer excitation losses are small. However, when the transformer is overexcited or upon initial
energization or sympathetic inrush conditions, they can cause sensitive differential protection to operate.
BE1-CDS240 Solution
Overexcitation and inrush are nonlinear and the currents have high harmonic content. Inrush is
characterized by high, even harmonics such as the second harmonic. Overexcitation is characterized by
high, fifth harmonics. The BE1-CDS240 calculates the magnitude of the second and fifth harmonic
components of the differential current. If the differential current is made up of a significant proportion of
either of these two components, the user can select to inhibit the sensitive percentage restrained
protection from operating.
Since CT saturation can also cause the currents seen by the relay to be rich in harmonics, an
unrestrained instantaneous differential element is included to provide dependability for an internal fault
with CT saturation.
During inrush, the second harmonic component may not be equally shared on all three phases. Because
of this, misoperation can occur on a phase with low second harmonic content. Our unique method of
second harmonic sharing improves security by allowing the harmonic restraint elements to respond to the
ratio of operate current to the sum of harmonic current measured on all three phases. This is superior to
other methods of cross blocking since each phase element operates independently in its comparison of
operating current to harmonic current. Thus, security is enhanced without sacrificing dependability
because a faulted phase will not be restrained by inrush on unfaulted phases.
To further enhance security from false tripping on inrush, the operating characteristic responds only to the
fundamental component of this highly distorted current thus reducing sensitivity to inrush current yet
allowing improved sensitivity to power system faults.
Problem 7: Digital Measurement Errors
General
Numerical relays sample the current and use digital signal processing algorithms to extract magnitude
and angle information to develop their operating quantities. Most methods used are accurate only when
the sampled power system quantity is operating at nominal frequency. Unless steps are taken to
compensate, accuracy falls off very quickly with deviations in the power system frequency.
Another problem that must be overcome with digital technology is the need to use anti-aliasing filters prior
to the sampling process to prevent harmonic components from affecting the measurement. Analog filters
introduce phase shift errors and are subject to attenuation drift that can introduce magnitude and angle
errors to the measurement.
These problems are more acute for differential protection since the operating quantity (differential current)
is derived from the difference between the measured quantities. Any error in magnitude or angle
measurement can result in large inaccuracies in the differential current measurement.
BE1-CDS240 Solution
Digital relay designers must decide on which of three solutions to use for this problem. They can allow
misoperation at off-nominal frequencies. Alternatively, they can disable the protection at off-nominal
frequency. Finally, they can compensate to maintain accuracy at off-nominal frequency. The BE1-
CDS240 uses frequency tracking to adjust the sampling interval to maintain full accuracy across a wide
frequency range so that it is both secure and dependable in all applications. For example, tripping of
important transformers during a disturbance that causes the system to go unstable can have a
catastrophic effect on an already over stressed power system. Generator and motor differential protection
applications are another situation where accuracy across a wide frequency range is important.
To eliminate the errors introduced by analog low-pass filters, the BE1-CDS240 uses digital signal
processing technology and 144 samples per cycle over-sampling to provide digital low-pass filtering.
9365200990 Rev M BE1-CDS240 General Information 1-15