A few years ago I joined a team of experts writing a revised standard for phasor measurement units (PMUs), devices that seemed to perform amazing feats of measurement in electric power systems. Not long after I joined the group, I began to be troubled by what we were writing about the measurement of frequency. In particular, it seemed to me that the term frequency had not been defined if it was changing, and since the PMU was expected to return a value for the rate of change of frequency (ROCOF), it was obviously expected to be changing. If the term is not defined precisely, I thought, how would you know if the measurement was being made accurately?

%B IEEE Instrumentation & Measurement Magazine %V 19 %P 19 - 24 %8 12/2016 %N 6 %! IEEE Instrum. Meas. Mag. %R 10.1109/MIM.2016.7777647 %0 Report %D 2014 %T PMU Measurement Technology %A Harold Kirkham %A Jeffery E. Dagle %A Sun, Y. %K AA14-001 %K AARD %K CERTS %K phasor measurement units (PMUs) %X Phasor measurements are without doubt extraordinarily valuable and informative. The number of phasor measurement units in service continues to increase, to the great benefit of the power system. And yet the measurement method used is deeply flawed. This report uses an epistemological approach to explain why the measurement of synchrophasors, as it is presently implemented, is unjustified. Fundamentally, the problem is that the definition for the quantities being measured is incorrect. We expect the PMU to furnish values for the three parameters that define a phasor (its amplitude, frequency and phase), when in fact the signals we furnish to the PMU are not phasors. It follows that the best we can expect is an approximation. It also follows that we cannot know the quality of that approximation. As an example of what this means, consider that the PMU is required to furnish a value for the rate of change of frequency, and yet by the definition of a phasor, that rate of change is zero. The report goes on to present an alternative method of describing the input signals to the measuring system. Instead of imposing the requirement that the signal be described by a phasor, the assumption of stationarity is relaxed, and the value of each of the three parameters is permitted to change during the measurement. The result of changing the definition is that a new measurement technique is possible. As a byproduct of the new definition for the input signal, new (and concrete) definitions emerge for the terms “apparent frequency” and “instantaneous frequency.” %I Pacific Northwest National Laboratory (PNNL) %8 04/2014