Measurement Algorithms for Digital Protection - Digital Signal Processing in Power System Protection and Control - page 165

(A) Classification

• Basic classification of the protection criteria measurement algorithms related
to the principle of measurement allows distinguishing methods applying:
signal averaging, orthogonal components and delayed orthogonal compo-
nents. Other special approaches met include counting of impulses and solving
of the fault circuit equation.
• Very similar methods and principles as well as adopted requirements apply to
algorithms of current and voltage magnitude and power measurement.
• The algorithms for estimation of impedance, phase shift and some of the
frequency measurement equations differ from the ones for magnitude and
power in the sense that they are quotient type, however, they are also similar
in the sense that the variables in the numerator and denominator are deter-
mined with the algorithms from previous point, i.e., they are usually powers
and magnitudes of certain signals.
• Slightly different and more diverse are some of the algorithms for frequency
measurement and symmetrical components filtration.

(B) Dynamics

• Dynamic features of measurement are dependent on applied filters and other
methods of introductory digital signal processing.
• A bit part in resulting dynamics play signals delaying being applied in some
algorithms (most often few times lower that the filter window length); the
total duration of measurement transients is a sum of filter window length and
delay value.
• Dynamics of the averaging methods depends clearly on the period of signal
averaging.

(C) Filtering efficiency of the algorithms

• Filtering, elimination or suppression of signal distortions is determined
mainly by the frequency characteristics of applied filters or other signal
processing methods. The algorithm itself has generally little influence in this
context. Sparse examples are exceptions to that rule; e.g., the algorithm of
magnitude estimation employing averaging of squared sample values (being a
kind of filter, incidentally). If one uses orthogonal components obtained from
signal delaying (without filtration), the resulting algorithms perform in fact
only spectral transformation of the processed AC components.

(D) Computational complexity

• Decisive for ease or difficulty of numerical implementation of the measure-
ment algorithms are again applied filters. Complexity of the algorithms is
rather small, since they usually base on simple mathematical operations like
additions, multiplications, divisions or squaring of numbers.