This full-text paper was peer-reviewed at the direction of IEEE Instrumentation and Measurement Society prior to the acceptance

and publication.

Is Reactive Power a Zero-Quantity?

Charles Trevor GAUNT Pitambar JANKEE

Department of Electrical Engineering Department of Electrical Engineering
University of Cape Town University of Cape Town
Cape Town, South Africa Cape Town, South Africa
[email protected] [email protected]
2023 IEEE 13th International Workshop on Applied Measurements for Power Systems (AMPS) | 979-8-3503-1240-9/23/$31.00 ©2023 IEEE | DOI: 10.1109/AMPS59207.2023.10297215

Abstract—Reactive power has been defined in various ways The waveforms of p(t) in the whole circuit and each
and discussed in technical and measurement papers almost component pR and pX are illustrated in Fig. 2. The integrals of
continuously since it was proposed. For over 100 years, instantaneous powers give respectively the total power P, the
consensus has eluded physicists and electrical engineers. This power dissipated in the resistor PR, and the power stored and
paper explores the physical, mathematical, and practical aspects returned PX. The apparent power S=VI represents the power
of reactive power. Average (active) power and average apparent were there no phase displacement between V and I, in other
power can be defined without reactive power, for which the words, neglecting one of the influencing factors that affect a
average is always zero. Two definitions are proposed that apply measurement.
consistently to all systems operating in relatively steady state –
for an apparatus and within a delivery system – with balance or The relatively small variation of the voltages and currents
unbalance, sinusoidal waveforms or distorted by harmonics, is intentional, since the interpretation focuses on the power
and for ac and dc. The implications for measurement systems components.
and devices in smart grids and conventional power systems
(including control), power quality standards, revenue metering, The power in the resistor is a double frequency oscillation
and scientific and legal measurement are outlined. There are between 0 and 8050 W, with an average of 4025 W. The total
also implications for defining dynamic or transient parameters. power varies from -475 to 8625 W, with an average of
4025 W. The average value of the negative part of the curve
Keywords—analysis, apparent power, measurement, metering, (the power returning to the source) over the duration of the
power factor, reactive power cycle is -61 W.
I. INTRODUCTION The power oscillating between the inductor’s magnetic
field and the circuit varies from -2227 to 2227 W with an
There is little difference between definitions of power
average of 0.
(also termed active power) in an electrical system or circuit,
such as by the IEC, IEEE, and DIN standards organizations The value of conventional reactive power Q is 2227 var,
and in textbooks. It is a physical quantity representing the rate which is the peak value of the power oscillation in the
of energy transfer or doing work. The measurement of inductor. Q is neither the average value of the oscillation nor
electrical power, also referred to as active power, is based on the average value of the total power returning to the source.
a physics-consistent concept model. Power is calculated as the (Note: var is the unit for reactive power [1], [2].)
average of the integral of instantaneous power over one or
more cycles, or as the product of rms voltage and current |I |= 20 Arms
components V and I scaled by cosφ, where φ is the |Vx| X
displacement angle between the waveforms. 50 Hz V = 230 Vrms
Two other conventional power components, apparent |Vr| R
power S and reactive power Q, have diverse non-physical
operational definitions in different standards, especially for Fig. 1. Simple series RX circuit with fixed voltage and impedance
polyphase circuits under conditions of unbalance or harmonic factor R/Z of 0.875, P=4025W, S=4600VA, Q=2227var,
distortion. Some of the diversity is illustrated by the IEEE Std displacement angle φ =28.955°. So, Z=11.5, R=10.0625, and
X=5.5674 Ohms.
100-2000 [1].
The paper examines some of the anomalies of reactive 10000
power. It shows average reactive power is zero and the term is 8000
unnecessary. The paper then offers and justifies two
alternative approaches to defining all power parameters under 6000

steady state conditions according to whether the terms relate 4000

to an apparatus or a power system.
2000
II. SINGLE PHASE POWER PARAMETERS 0
0 5 10 15 t [ms] 20
A. Sinusoidal Waveforms in One Cycle -2000
v(t) [V] i(t) [Ax5] vi(t) [VA]
The instantaneous values v(t) and i(t) of sinusoidal current, -4000 p_r(t) [W] p_x(t) [W] p(t) [W]
I cos(ω(t)+φ), during one sinusoidal cycle of voltage, Pav [W] Q [var] Sav [VA]
V cos(ω(t)), for a simple circuit described in Fig. 1 give the Fig. 2. Instantaneous values of v(t), i(t), p_r(t), p_x(t), p(t), and vi(t)
instantaneous power p(t)=v(t) i(t). for the circuit in Fig. 1, during one 20ms-period of V. Also shown:
Pav, Q, and Sav.

Authorized licensed use limited to: University of Cape Town. Downloaded on November 16,2023 at 09:02:20 UTC from IEEE Xplore. Restrictions apply.
 The power derived by neglecting the phase displacement PoC
between V and I varies from 0 to 9200 with an average of Apparatus
4600. As a conditional “what-if” measure of power, its unit Thévenin Equivalent R, X: load,
Impedance rTh, xTh I
would be W but apparent power S is assigned the unit VA. generator
VTh VPoC
or another
B. Reactive Power as Power Returning to Source network
A widely held interpretation of reactive power is that it is
the power oscillating between the source and load, being the VTh
cyclic exchange of energy stored in the magnetic fields of VTh x
generators and motors, which is needed for them to work. φAPP VPoC
VTh r
Energy is also stored and released in transmission lines,
transformer inductances, and the electrostatic fields of VX
capacitors. I
VR
The origin of this concept of reactive power is probably Fig. 3. Single phase power system equivalent circuit and its vector
the description by Kennelly in an IEEE paper of 1910 [3]. He diagram with VPoC as the reference vector.
described power in a generator with internal resistance and
inductance as being the average power plus a double- and distribution can be represented by the Helmholtz-
frequency component of oscillating power. The vector power Thévenin equivalent impedance (TEI) and an ideal voltage
triangle of P, Q and S was proposed in the paper. source.
However, it is clear from Fig. 2 that the quantification of The inclusion of the delivery system impedance allows the
Q=2227 var cannot be the average of the power oscillating loss directly attributable to the apparatus to be included in
between the load and source, because the average is zero and analysis.
the amplitude is 2227 W.
The current through the TEI determines the voltage at the
Another interpretation is that the reactive power returning PoC. The main effect of higher Thévenin reactance is to
to the source causes extra loss in the conductors, although the increase the rotation of VPoC with respect to VTh and higher
resistance of the wires from source to load is not included in Thévenin resistance increases the delivery loss.
the various definitions of reactive power [1] [2]. Kennelly did
not refer to loss associated with reactive power. Early III. POLYPHASE POWER PARAMETERS
researchers referring to wattless, quadrature or reactive power A. Standard Definitions
[4] knew that this power was not a physical power and could
not be measured in physical meters [5]. Its interpretation as a The IEC recently revised its definition of reactive power.
power component oscillating between source and apparatus is It is defined only for a two-terminal circuit under sinusoidal
invalid. conditions (IEV ref 131-11-44) and associated with apparent
power and the displacement angle φ [2].
Reactive power lacks a physical interpretation because it
is not a physical quantity. It is defined only by equations that Most other standards also adopt the PQS power triangle as
inherently violate the laws of physics. It exists only as the the basis for defining apparent and reactive power. Apparent
imaginary part of apparent power expressed as a vector P+jQ. power and reactive power are inseparably linked. In various
standards, one is measured and the other is derived from it and
C. Power Factor the power.
Power factor is defined as the ratio of active power to Various standards define apparent power and reactive
apparent power, P/S, and its strict units would be W/VA. power differently [1] [2] [7] [8]. No standards define terms in
Unity power factor occurs where P=S, such that there is no the presence of unbalance and harmonic distortion that
displacement angle between them in the PQS power triangle. characterize practical power systems. As a result, meters
This is consistent with Kennelly’s earlier definition of the nominally complying with the same standard are known to
impedance triangle for the geometric addition of resistance give different results [9].
and reactance to give impedance [6]. His impedance factor The large number of variables or influencing factors that
IF=R/Z is a dimensionless ratio and has the same numerical affect a measurement, or the degrees of freedom, are possible
value as the power factor. reasons for not developing a complete, general, and consistent
If the influencing factor of the angle between V and I is set of definitions for all cases.
neglected and S=|V| |I| is interpreted as a “what-if” quantity of The rest of this section identifies the physical concept that
active power, then P/S is a dimensionless ratio equal to IF and could be used for analyzing systems with so many variables.
does not need the concept of the power triangle or a definition
of reactive power. B. Concept Model for System with Sinusoidal V and I
D. An Apparatus in the Context of a Power System Polyphase systems can have any number M wires, such as
3-phase 3- or 4-wire systems with three source voltages but
The circuit of Fig. 1 is that of an apparatus connected to an different delivery systems and load topologies. The TEIs of
ideal voltage source. An apparatus is an assembly of the wires may be balanced or unbalanced since they depend
components and wires that might be simple or complex, or a not only on the conductors, but also other apparatus (loads and
power network with many nodes. sources) connected to the power system. In conventional
In practical power systems, as depicted in Fig. 3, the power systems with rotating machine generators, the voltage
terminal voltage is not independent of the load. From the point sources are equal and displaced by 120°.
of connection (PoC) of the load, the generators, transmission,

Authorized licensed use limited to: University of Cape Town. Downloaded on November 16,2023 at 09:02:20 UTC from IEEE Xplore. Restrictions apply.
 The apparatus, source, and TEIs of an M-wire power
system with sinusoidal voltages and currents are shown in VTh m,h rTh m.h xTh m.h PoC IS m,h
Fig. 4.
C. Harmonic waveform distortion eref Um,h Apparatus
Practical power systems, whether conventional or with
increasing penetration of power electronic devices, do not
Vref
operate only with sinusoidal V and I, and linear impedances;
they exhibit harmonic distortion. Fig. 4. 3-phase m-wire power system equivalent circuit for sinusoidal
voltages and currents and linear TEIs in which m=no of wires and
Fourier analysis allows distorted periodic waveforms to be h=harmonic order. Vref is the reference point for measuring U and
resolved into H+1 harmonic components, including dc. For eref is the offset voltage from Vref across the delivery system.
each harmonic, there is a circuit like that in Fig. 4, with its
frequency components of voltages, currents, and impedances.
Power is an extensible quantity (allowing it to be added - The optimum power sent out from the Thévenin point to
arithmetically), so the principle of superposition can be used deliver the same power to the PoC with the minimum loss
to solve each harmonic system separately and add the power is carried by a current vector IA distributed in proportion
quantities associated with each harmonic. At the same time, to the source voltages and inversely in proportion to the
power cannot be assigned to any single wire or sub-set of resistances.
wires. Similarly, apparent power defined conditionally as a
“what-if” power is also a parameter of the system, not a wire. - By resistance-weighting the voltage and current
component vectors, the power sent out from the Thévenin
D. A Mathematical Model point PTh is given by ||V'Th(null)|| ||I'A||, the delivery loss by
Linear algebra is a compact mathematical technique for ||I'A||2, and ||I'A|| is found by solving a quadratic equation.
handling vectors, such as the Mx(H+1) sets of voltage and
current components, and the TEIs. It is suitable for the - The same operation is repeated for each frequency
analysis of a power system with large numbers of variables. component (including dc) to obtain the scalar power
parameters that, using the principle of superposition, can
The equations describing the relationships between the be added.
variables in the concept model must be consistent with the
laws of physics. These include Ohm’s Law, Kirchhoff’s The details of the derivation and proof of the GPT are
Voltage and Current Laws, Fourier decomposition, the available in [10]. The calculation is a simple linear algebraic
Helmholtz-Thévenin equivalent circuit, and the Conservation process easily compiled in a spreadsheet, processing language
of Energy. As a result of complying with the laws of physics, or control algorithm. This section describes the process.
derived measurements will be capable of physical A. Inputs
interpretation.
Derived from regularly sampled scalar instantaneous
IV. GENERAL POWER THEORY voltages and currents at the PoC, the inputs to the algorithm
are the complex rms values (the rms value and the angle from
A General Power Theory (GPT) based on the concept the reference) voltages and currents, Vm,h and Im,h. The
model developed in Section III was derived and proved reference is usually one of the measured voltages at the PoC.
mathematically using linear algebra. The reasons for By convention, currents and power into the PoC are positive
formulating it were to develop better converter control and to and out of it into the delivery system are negative.
enable more representative analysis of power system
disturbances by currents such as leakage dc and There are many published methods of determining the TEI
geomagnetically induced currents that cause asymmetric from a PoC. Passive techniques depend on changes in power,
harmonic distortion in delivery systems. such as caused by a tap changer operation. Active techniques
disturb the currents in selected wires, such as by a power
The GPT is basically an optimization algorithm that electronic converter. The TEI is based on small changes of
minimizes the loss uniquely attributable to a load or source voltages and currents and is not the fault level impedance.
such as distributed generation. The principle is that of all the Fault impedance is based on transient generator impedance
infinite possible current components in the delivery system, values, and usually neglects the intermediate loads and the low
only those aligned with the voltages at the Thévenin voltage voltage at the point of fault that decreases the effect of shunt
source can deliver the power with minimum loss. All other capacitance in cables and high voltage lines.
current components incur avoidable loss. Also, the apparent
power is the maximum power at the Thévenin source for the B. Calculation
same loss as incurred by the uncompensated load or The average ac power at the PoC, PPoC, is the sum of the
generation at a PoC, and it should be used in the ratio of power powers of each harmonic system, PPoC,h for m=1 to M and h=1
factor. to H, where:
Key aspects of the GPT include: , =∑ , , , and
, = ∑ , , for the dc power component.
- The vectors of voltages and currents U and IS and the
scalar Thévenin equivalent impedances and voltage The current through the impedance of each wire causes
sources can be measured at the PoC. (Note that voltages voltage drop (or rise) which is added to the voltage at the PoC
and currents are scalar quantities, but a set of scalars to give the Thévenin point voltage referred to the reference
forms a vector quantity [2].) wire from which the voltages were measured:
, = , + , , … and thereby VTh (1)

Authorized licensed use limited to: University of Cape Town. Downloaded on November 16,2023 at 09:02:20 UTC from IEEE Xplore. Restrictions apply.
 The resistance elements of the TEI form a square matrix R Redistribution of the current components changes the
with M(H+1) rows and M(H+1) columns, where the diagonal voltages at the terminals of the apparatus. All forms of
vector is compensation have this effect, whether using passive
R(diag) = {(r1,0, r1,1, … r1,H), (r2,0, r2,1, ... r2,H), impedances or power electronics. Thus, although the GPT
…, (rM,0, rM,1, … rM,H)}, optimizes the loss incurred for delivering the same power to
and all other elements are zero. the PoC, the sensitivity of the apparatus to the change in
voltage will determine the need for re-compensation
One benefit of the resistance weighting approach is that responsive to the effects of voltage change. In conventional
||I'||2 gives the delivery loss, and ||VTh'|| ||I'|| gives the power power definitions based on a concept model limited to an
where ' indicates the resistance weighted variable. apparatus with ideal voltages at the terminals, the response of
Using the R matrix, a resistance-weighted unit vector 1h' practical power systems to compensation is neglected.
identifies the offset from the voltage measurement reference
of a null point consistent with Kirchhoff’s voltage law: V. VARIANT OF GPT FOR AN APPARATUS
= ∑ , , /∑ R , 1h′ A. Calculation of Apparent Power and Power Factor
This offset is subtracted from each , to give the
Assuming an impedance-less delivery system (ZTh=0) has
Thévenin line voltage with respect to the weighted null point: the effect of reducing the full concept model to that of an
VTh m,h(null)= VTh m,h - eref,h and apparatus supplied by ideal voltages. Then the GPT’s
. .- equations of power loss and voltage drop in the delivery
"#$$ = ∑ ∑ , % &'()* + , (2)
1
impedance are no longer needed, but the referencing of
giving ||0 "#$$ || voltages to a null point in accordance with Kirchhoff’s voltage
law is still relevant. The null reference is given by:
Without compensation, the loss due to the resistance
weighted original current ||IS'|| is ||IS'||2. & = ∑ , , .
The apparent power APAPP = ||IS|| ||VPoCnull|| (5)
The minimum loss, due to the resistance weighted active
where || 3 ||1 = ∑. ∑ 1
, and
current ||IA'|| is ||IA'||2. The power at the Thévenin point is 1 . ∑ 1
|| || "#$$ || = ∑ , % &'()* .
"#$$ || ||IA'||, so:
||IA'||2 – || "#$$ || ||IA'|| + PPoC = 0. (3) Then PFAPP = PPoC/APAPP
Solving this quadratic equation gives ||IA'||.
The apparent power is designated APAPP to distinguish it
With optimal compensation, the system apparent power from the system APSYS, and from the S of conventional
(APSYS in Watts) at the Thévenin point with the same loss as definitions that do not reference voltages to a null point. APAPP
with the original currents is: 2 343 = ||0 5677 || ||89 || (4) can be interpreted as conditional “what-if” power but to avoid
confusion with real power would retain the unit VA.
The system power factor PFSYS = P/APSYS = ||IA'|| / ||IS'||
is an index of the relative efficiency of the power delivery. Apparatus APAPP is potentially useful as an indicator of
equipment or apparatus ratings using the rated voltages and
C. Outputs currents. The apparatus approach is also useful for
The significant power parameters derived from the comparisons against the results of other power definitions that
calculation are the system apparent power APSYS and the do not explicitly include the effects of unbalance or distortion.
system power factor PFSYS. Both have physical
B. Example
interpretations.
An example of single-phase resistance between two phase
There is no form of reactive power in this approach. wires of a 4-wire wye system is given as example 3.2.2.7 and
Conventionally defined reactive power Q is not defined in the the circuit in Figure 4(a) in IEEE 1459-2010 [9]. It has been
power triangle in the same way as P and S, since the average extended in Fig. 5 to other m-wire apparatus topologies.
of the instantaneous power depicted in Fig. 2 is the zero
average of the power oscillating in an inductor. Consistent
with P and S, the value of Q is zero, but it is defined in ways
that assign it a complex number quantity. It is not a physical
quantity, so it cannot be modelled in a measurement approach
defined strictly in terms of circuit physics.
D. Compensation
The details of GPT-calculated compensation are beyond
this paper’s scope, except for considering two of its effects.
Optimal compensation redistributes the original current in
proportion to the Thévenin source voltages and inversely to
the delivery system resistances so that with minimum loss the
same power is available at the PoC. The compensation current
IC at the PoC is the difference between the active current IA
and the original apparatus current IS = IA + IC. The reduction
in loss is ||8: ||1 + 2 < 8= , 8: >. The real value of the second
term shows that the non-active current usually associated with Fig. 5. PFAPP for various systems supplying a 1-phase resistor. The
reactive power is not orthogonal to the active current, else the delivery system impedance shown, equal in all wires, is neglected in
the original example of the 3-ph wye, line-line circuit.
term would be zero.

Authorized licensed use limited to: University of Cape Town. Downloaded on November 16,2023 at 09:02:20 UTC from IEEE Xplore. Restrictions apply.
 Standard 1459-2010 compares the various approaches to techniques of a converter disturbing only a few current
measuring apparent power and three alternatives of vector, components at a time.
arithmetic, and effective power factor. The effective power
factor is derived from an intermediate step of defining an Advantages of using only measurements at the PoC are
effective line voltage and current. For the example of an that they avoid uncertainty associated with approaches of
unbalanced load of a resistor between two phase wires of a 4- making measurements across different locations [14] and
wire wye system, the effective power factor is given as 0.7071. dependance on communications that could be subject to cyber
This is consistent with PFAPP. The values of PFAPP for other interference.
topologies in Fig. 5 are given below each circuit. After the measurement of the input quantities, each single
The example demonstrates how the single formulation of output quantity of the GPT approach is given implicitly by the
this variant of the GPT can be applied easily to systems with measurement model and its algebraic calculations of the
any number of wires. It also shows how the number of wires physical properties of the system performance. There is no
in the system supplying an apparatus affects the measurement further need for statistical algorithms, empirical quantities, or
of apparent power and power factor. reference frame shifts based on an approximation of q.

There are significant similarities between the Fryze- VII. IMPLICATIONS OF THE NOVEL MEASUREMENT
Buchholtz-Depenbrock and Emanuel’s approaches that APPROACH
underlie the effective apparent power approach, and the The GPT approach produces a consistent method of
Currents Physical Components approach following defining, measuring and controlling performance in practical
Czarnecki. Each derives a null point for the voltages, though power systems.
using different methods. The GPT’s justification for
determining the null point is derived from Kirchhoff’s voltage The broad objective of GPT measurement is to guide
law and has the same effect on measuring apparent power and technical decision-making and fair commercial trade. Applied
power factor. to control, such as of power electronic compensators or
converter interfaced generation, the objective is to minimize
Although APAPP is possibly useful as an equipment rating delivery losses (and the energy needing to be produced) while
in VA, it has no relationship to loss incurred in delivery. Also, delivering the same power to the load and improving voltage
PFAPP is not an index of the relative delivery efficiency of the profiles and stability.
load or generation in the context of a whole power system, as
given by PFSYS. The implications are as broad as the range of applications
of power measurement. Measurement standards define what
VI. MEASUREMENT UNCERTAINTY parameters should be measured and how the measurement
The GPT is based only on samples of voltage and current calculation should be made. Meter standards define how the
at the PoC. In this respect, the same uncertainty is derived measurement standards will be embodied in physical meters.
from the inputs as any other digital measurements of power The meters provide data for input to system analysis, for
and apparent power in the frequency domain. systems operators and electromechanical control, and for legal
regulation of power quality and financial transactions between
The power terms apparent power and power factor have suppliers and customers of electrical power and energy. This
been derived by the GPT for steady state conditions, including last application suggests there is no basis in physical
constant frequency. Small frequency variations occurring in performance for tariffs or pricing of reactive energy, reactive
practical power systems can managed by averaging. This is no power, or reactive power capability.
different from conventional measurements.
Simulations have shown that decisions and control based
Voltages and currents in all wires are sampled at fixed on concepts of reactive power, such as in the analysis of
intervals and the voltage measurements must be made from a system stability in the presence of geomagnetically induced
common reference point or wire. Depending on the currents [15], or the volt-var control of distributed generation
conditions, these measurements will define distorted, to limit voltage rise, can lead to inappropriate decisions.
unbalanced sinusoids, with a special case being sinusoidal
balanced waveforms. The most significant uncertainty is As an oscillating power, reactive power is a steady state
usually considered to derive from the voltage and current average power over one or more cycles and it is always zero.
transducers. However, having a high sampling rate and many Measurement by its amplitude makes it inconsistent with the
samples is sufficient to reduce the uncertainty of the derived definitions of average power and average apparent power,
components [11]. with which it is combined by an operational definition – an
arbitrary equation. Its measurement model is incompatible
Identifying the exact length of a periodic cycle reduces with one based on the physical performance of practical power
spectral leakage when calculating voltage and current vectors systems.
U and Is. Without an accurate frequency estimation, spurious
harmonic orders are introduced by a Fourier Transform. A By contrast, the measurement model and formulation of
phase-locked loop robust to distortion and unbalance must be the GPT lead to physical power parameters at a connection to
used to identify the frequency. Then, the complex rms values any apparatus.
described in Section IV.A are derived for all harmonic orders Reactive power cannot be defined in the GPT because it is
using a Discrete Fourier Transform. non-physical. This raises questions about the role of reactive
The uncertainty associated with the estimation of the power in analysis and the suitability of controllers and
Thévenin equivalent parameters depends on the method of software using Q-based models. Any value of Q or any other
measurement [12] [13]. Passive techniques that depend on a non-active power component in any equation related to line
load change are less controllable than the active disturbance loss, voltage drop, or voltage stability introduces unknowable

Authorized licensed use limited to: University of Cape Town. Downloaded on November 16,2023 at 09:02:20 UTC from IEEE Xplore. Restrictions apply.
uncertainty about performance prediction. Further questions ACKNOWLEDGMENT
arise about the validity of instantaneous reactive power q in Our colleague Dr. David Oyedokun has assisted by
transient conditions, where the parameter is widely used, and checking draft manuscripts to identify any ambiguities and
the possibilities of alternative approaches focused on voltages technical errors. Any mistakes in the final manuscript are the
and currents. responsibility of the authors.
The GPT’s consistent approach to measuring power
parameters and its application to control have significant REFERENCES
implications for international, national and industry standards.
GPT-control of inverters suggests significant energy savings [1] The Authoritative Dictionary of IEEE Standards Terms, IEEE
can be achieved by decentralized loss optimization in power Standard 100-2000 (Inactive - withdrawn).
https://ieeexplore.ieee.org/document/4116787.
system operation. Its alternative approach to power theory
requires changes in teaching and exposes many new [2] IEV, International Electrotechnical Vocabulary, International
Electrotechnical Commission, https://www.electropedia.org/.
opportunities for research in steady state optimization and
[3] Kennelly, A.E., Vector power in alternating-current circuits, Proc.
probably in transient stability analysis and control. AIEE, vol. 29, issue 7, pp.1023-1057, 1910.
https://doi.org/10.1109/PAIEE.1910.6659879.
VIII. CONCLUSIONS [4] Browne, W.H., Letter to the Editors, Electrical World and Engineer,
The concepts of apparent power, reactive power, and vol. 38, p.1031, December 21, 1901.
power factor quantities are deeply embedded in the analytical [5] Browne, W.H., Power factor indicators, Trans. AIEE, vol. XVIII,
methods applied in power system design and operation, as pp.287-312, 1901, https://doi.org/10.1109/T-AIEE.1901.4764183.
definitions and measurements of the quantities, and as used [6] Kennelly, A.E., Impedance, Trans. AIEE, vol X, p175-232, 1893.
for input data in modeling system performance. However, https://doi.org/10.1109/T-AIEE.1893.4768008.
there are very few ways to test physically the validity of the [7] Standard definitions for the measurement of electric power quantities
under sinusoidal, nonsinusoidal, balanced, or unbalanced conditions,
results of the conventional approaches relative to the physical IEEE Standard 1459-2010.
performance. https://doi.org/10.1109/IEEESTD.2010.5439063.
The analysis presented in this paper shows that [8] Quantities used in alternating current theory - Part 2: Multi-line
circuits, DIN 40110-2, DIN German Institute for Standardization,
conventionally defined apparent power, reactive power, and 2002. https://www.en-standard.eu/din-40110-2-
power factor are inconsistent with the laws of physics that wechselstromgrossen-teil-2-mehrleiter-stromkreise/
describe the performance of electrical circuits and power [9] Definitions for Calculations of VA, VAh, VAR, and VARh for Poly-
systems. Phase Electricity Meters, NEMA C12.24 TR-2011.
[10] Malengret, M., and Gaunt, C.T., Active currents, power factor, and
Interpretations commonly given to reactive power are apparent power for practical power delivery systems, IEEE Access, 8,
invalid. Defined by its average value over an integer number 133095–133113, 2020.
of cycles, reactive power is always zero. Unlike average https://doi.org/10.1109/ACCESS.2020.3010638.
power P and average apparent power S, Q has a special value [11] De Capua, C., Landi, C., and Malafronte, G. C., Measurement
uncertainty in evaluating total harmonic distortion factor, in Proc.
and is given a special unit. Reactive power is not evident (and 18th IEEE Instrum. and Meas. Technology Conf., pp. 675-679,
not used) in a power theory rigorously based on a physical Budapest, 2001. https://doi.org/10.1109/IMTC.2001.928903.
measurement model. Reactive power is defined by an [12] Borkowski, D., Wetula, A., and Bien, A., New method for
empirical equation that neglects influencing quantities of noninvasive measurement of utility harmonic impedance, at 2012
unbalance and distortion. Its inherent uncertainty makes it IEEE PESGM. https://doi.org/10.1109/PESGM.2012.6345080.
potentially misleading in any equation with Q and can lead to [13] Hui, J., Freitas, W., Vieira, J.C.M., Yang, H., and Liu, Y., Utility
inappropriate decisions. Like the concept of the aether, the Harmonic Impedance Measurement Based on Data Selection, IEEE
Trans. Power Delivery, vol. 27, no. 4, pp. 2193–2202, 2012.
concept of reactive power leads to uncertain outcomes in https://doi.org/10.1109/TPWRD.2012.2207969.
analysis and it may be unnecessary. [14] Haji, M.M., and Xu, W., Online determination of external network
The General Power Theory, without reactive power, models using synchronized phasor data, IEEE Transactions on Smart
Grid, vol.: 9, no. 2, 2018. https://doi.og/10.1109/TSG.2016.255948.
defines consistent, interpretable power terms, including
[15] Chisepo, H.K., and Gaunt, C.T., Towards asymmetrical modeling of
apparent power and power factor, for systems with any voltage stability in the presence of GICs, at 2022 IEEE PES/IAS
number of wires and in the presence of unbalance and PowerAfrica, Kigali, Rwanda.
harmonic distortion. This has broad implications for scientific, https://doi.org/10.1109/PowerAfrica53997.2022.9905396.
technical, and legal metrology, for control, and for many .
analytical methods applied in power systems planning and
operation.

Authorized licensed use limited to: University of Cape Town. Downloaded on November 16,2023 at 09:02:20 UTC from IEEE Xplore. Restrictions apply.