Thermodynamic Steady States in Simple Electrical Circuits
Thermodynamic Steady States in Simple Electrical Circuits
Thermodynamic Steady States in Simple Electrical Circuits
K.M. Browne
School of Engineering and Technology, Deakin University, Victoria, Australia
Communicated by E.E. Michaelides, New Orleans, USA
Registration Number 957
Abstract
Electrical circuits provide good examples of systems which gravitate towards the three
possible thermodynamic conditions of equilibrium, metastability or dissipative steady
state. Circuits are driven towards any one of the three conditions simply by increasing
total entropy or decreasing available energy. The thermodynamic condition for steady
state conduction in a number of simple LCR and hot filament circuits is found to be
1/ðdstot =dcÞ ¼ 0 or 1/ðdeavail =dcÞ ¼ 0 where stot and eavail are the total entropy and
available energy of the system and its surroundings, and c is any single valued
variable which describes the condition or state of the total system. This is the inverse
of the equilibrium and metastable conditions. It holds for all systems examined,
whereas the Onsager-Prigogine principle, dstot =dt ¼ 0 at steady state, holds only in
one case. This shows that the principle is not the causative principle of the steady state.
As expected, the analysis also demonstrates that electric circuit theory is fully
consistent with the second law of thermodynamics as well as with the first.
1. Introduction
Because the laws of thermodynamics apply to every physical change that occurs,
thermodynamics may well be regarded as a macroscopic ‘‘theory of everything’’. Its
theory should be applicable, therefore, to the behaviour of all types of electrical
circuits. Circuit theory, for example Kirchhoff’s and Joule’s laws, is known to be
consistent with the first law of thermodynamics. Thermodynamic steady states are also
determinable from the first law on the basis that the output power is equal to the input
power since the system itself does not change under steady state conditions. However,
only a few attempts have been made to analyse the second law of thermodynamics
of electrical circuits and in particular to determine why electrical steady states exist
[1–4]. Most attempts have been based on Onsager’s ‘‘principle of the least dissipation
(rate) of energy’’ according to which ‘‘the rate of increase of entropy plays the role of
a potential’’ [5]. This principle was originally applied to coupled irreversible processes
including the thermocouple. Onsager’s ideas were later extended to other dissipative
steady states (or stationary states) through a principle recently restated by Kondepudi
and Prigogine [4] as: ‘‘In the linear regime, the total entropy production (rate) in a
system subject to flow of energy and matter dstotal /dt reaches a minimum value at the
stationary state.’’ These authors also state that ‘‘the general formulation and the
demonstration of this principle is due to Prigogine’’ (see also [6] pp.75–77). However,
it is better called the Onsager-Prigogine principle since it is clearly based on Onsager’s
ideas of entropy production and energy dissipation rates. If steady states minimise the
energy dissipation rate, this could well be a useful principle in the design of electric
circuits, and in particular in the design of integrated circuits.
In attempting to prove the Onsager-Prigogine principle for electrical circuits,
Kondepudi and Prigogine state that for circuit elements in series ‘‘the entropy
production (rate) in the linear regime is minimised when all flows are equal’’ and
suggest that this is supported by Feynman et al. [3]. Feynman actually stated without
proof or reference that ‘‘if currents are made to go through a piece of material
obeying Ohm’s law, the currents distribute themselves inside the piece so that the rate
at which heat is generated is as little as possible. Also we can say (if things are kept
isothermal) that the rate at which energy (sic) is generated is a minimum.’’
Kondepudi and Prigogine assumed that current and voltage in LCR circuit elements
are linearly related. However, while voltage and current in resistors are linearly
related by Ohm’s law, this does not apply to inductors or capacitors in DC circuits. In
AC circuits, steady state current and voltage are linearly related by I ¼ V/Z, where
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Z ¼ R2 þ ðoL 1=oCÞ2 is the impedance, but the phase angle f ¼ artan(R/Z)
must be taken into account in determining the power dissipated. Thus, Kondepudi and
Prigogine’s ‘‘proof’’ of the principle fails because they assume that for a number of
circuit elements connected in series the power dissipated in the ith element is ViIi,
which is only correct if Vi and Ii are the potential difference across and current
through the total resistive component only.
Nevertheless, simple electrical circuits are very useful systems for testing the
Onsager-Prigogine principle since the associated electrical circuit theory is well
established and steady states are common. In general, electrical circuits dissipate
energy as heat by the irreversible conversion of electrical energy to heat energy in
resistors, but not in ideal capacitors or inductors where energy is stored reversibly in
electric or magnetic fields. For ideal LCR circuits, the loss of available energy is
simply the energy dissipated in the resistive elements; the entropy generated is the
heat dissipated divided by the environmental temperature T0, while its production rate
is equal to the power dissipated divided by T0. In the case of a heated filament, the
system is not maintained at the environmental temperature and entropy is determined
in a different way.
Here it is shown thermodynamically why simple inductor-capacitor-resistor (LCR)
circuits and heated filaments reach dissipative steady states. It is also shown that these
states do not necessarily coincide with a minimum entropy production rate as
suggested by the Onsager-Prigogine principle. Furthermore, in all cases the steady
state is characterised by 1/ðdstot =dcÞ ¼ 0 or 1/ðdeavail =dcÞ ¼ 0 where stot is the total
entropy of the system and its surroundings, eavail is the available energy and c is any
single valued variable which denotes the state or condition of the system. These are
the inverse of the equilibrium conditions. Thus all of these systems are driven towards
a dissipative steady state by available energy decrease or by entropy increase only,
regardless of its production rate. This finding for electrical circuits supports and
compliments recent work on other steady state systems [7]. Thus it appears that the
Onsager-Prigogine principle is not the causative principle for the development of
dissipative steady states, even in systems close to equilibrium.
Fig. 1. Systems which tend towards (a) a thermodynamically stable equilibrium state (b) a
metastable state and (c) a steady state. In (c) the switch S2 allows the capacitor to be charged
to a finite initial potential difference Vi.
dstotal deavail
¼0 or ¼ 0; (1a)
dVc dVc
which is, in terms of a general parameter c which represents the state of the system
dstotal deavail
¼0 or ¼ 0; (1b)
dc dc
eavail ¼ 0: (2)
In practice, the finite internal resistance of the capacitor would allow both capacitor
and battery to be discharged slowly over a long period of time.
A system tending toward steady state is represented by the circuit shown in Figure 1c.
This allows a leakage current Ip through the parallel resistor Rp. Switch S2 allows
the capacitor to be charged initially to zero potential (S2 in position 1 and S1 off ) or
Vi (S2 in position 2 and S1 off). If S1 is switched on and S2 is switched to position 1,
regardless of its initial position, a steady state is reached when the current through
the capacitor C is zero, in which case the steady state voltage on the capacitor
Vc,ss is:
Rp
Vc;ss ¼ Vo : (4)
Rp þ Rs
Regardless of the initial voltage across the capacitor, the currents through Rs, Rp
and the capacitor C, in terms of Vo, and the instantaneous potential across C, Vc,
are:
Is ¼ ðVo Vc Þ=Rs ;
Ip ¼ Vc =Rp ;
Ic ¼ ðIs Ip Þ ¼ ðVo Vc Þ=Rs Vc =Rp :
It is assumed that the whole process is carried out at constant temperature T0, and that
the heat output of the resistors is fed into a large heatsink at that temperature. It is
immaterial whether the heatsink and battery are considered part of the system or the
surroundings provided that the sign of the entropy contribution is correct. The rate of
total entropy production is the total instantaneous power dissipated in the resistors
divided by the environmental temperature, T0. In terms of the capacitor potential
difference Vc, this is:
!
ds 1 ðV0 Vc Þ2 Vc2
¼ þ :
dt T0 Rs Rp
The entropy production rate is quadratic in Vc, with ds/dt having a minimum value
where Vc ¼ Vo Rp =ðRp þ Rs Þ, which coincides with the steady state. This is shown in
Figure 2 for the particular conditions selected. In this case, the system behaviour is in
agreement with the Onsager-Prigogine principle.
Fig. 2. Rate of entropy production, for the CR circuit shown, as a function of the voltage
across the capacitor at T ¼ 300 K. Arrows show the process directions.
The time dependent voltage on the capacitor, from elementary circuit theory is:
Vc;ss Vi
Vc ¼ Vc;ss 1 expðt=tÞ ; (5)
Vc;ss
where t is the time constant given by t ¼ CRs =ð1 þ Rs =Rp Þ. For Vi ¼ 0:
Vc ¼ Vc;ss ½1 expðt=tÞ :
Now, since ds/dt is a function of Vc and T, the latter being held constant at the
environmental temperature T0 in this case, then Vc can be regarded as the independent
variable in terms of which the total entropy produced may be found. Noting from
Eq. (5) that
dVc 1
¼ ðVc;ss Vc Þ
dt t
and since ds=dVc ¼ ðds=dtÞ=ðdVc =dtÞ then
!
ds t ðVo Vc Þ2 Vc2
¼ þ : (6)
dVc T0 Rs ðVc;ss Vc Þ Rp ðVc;ss Vc Þ
The total change in entropy of the system and its surroundings, calculated from an
initial condition where the initial voltage on the capacitor is Vc;i , is found by
integrating Eq. (6):
ð !
t Vc ðVo Vc Þ2 Vc2
Ds ¼ þ dVc
T0 Vc;i Rs ðVc;ss Vc Þ Rp ðVc;ss Vc Þ
0 1
Vo2 Vc;ss Vc 2Vo Vc;ss Vc
B Rs ln Vc;ss Vc;i þ Rs Vc Vc;i þVc;ss ln Vc;ss Vc;i C
B C
t B
B
0 2 2
1 C
C
Ds ¼ B ðVc;ss Vc Þ ðVc;ss Vc;i Þ C:
T0 B B C C
B 1 þ 1 B 2 2 C C
@ Rs Rp @ 2 Vc;ss Vc A A
2Vc;ss ðVc;i Vc Þ þ Vc;ss ln
Vc;ss Vc;i
(7)
Equation (7) indicates that the entropy change relative to the initial conditions (in this
case, Vc;i ¼ 0 or 2
Vc;ss ) approaches an infinite value as the system approaches the
steady state. Thus, as the steady state is approached, jdDstot =dVc j ! 1, so that
The change in available energy of the system can be determined from the Guoy-
Stodola theorem [8], Deavail ¼ T0 Ds, so that jdDeavail =dVc j ! 1 and
Fig. 3. The entropy changes relative to the initial conditions, for the CR circuit shown,
illustrating the transition from a metastable equilibrium state (Rp ¼ 1) where entropy is a
maximum, to dissipative steady states having decreasingly smaller (but finite) values of Rp.
Equations (8a) and (8b) are effectively the inverse of those characterising the
equilibrium and metastable conditions (Eq. 1). The entropy and available energy
changes, for selected values of the circuit components, are shown in Figures 3 and 4.
If the parallel resistance Rp is infinite, then the system will come to a metastable state
with the capacitor charged and, theoretically, should remain in that state indefinitely.
In practice, leakage currents invariably exist so that the capacitor will not remain
charged forever. In that case, the system would move towards a steady state because
if Rp 6¼ 1 then, following a transient stage, a steady current would flow through Rp
until such time as the source emf is removed. It might be expected that there would be
little difference between a system tending towards a metastable state (Rp ¼ 1) and
one tending towards a steady state system (Rp @ 0). However, these two systems are
quite different with respect to the dissipation of energy, although part of their
transient behaviour is similar.
The transition from a system which tends towards metastability and one which tends
towards a steady state as Rp changes from infinite to finite is shown in Figures 3 and
4. Also shown is the system behaviour as the value of Rp is decreased to low values.
In all cases, the two initial conditions are taken as (i) the capacitor is fully discharged,
and (ii) the capacitor is initially charged to twice its steady state voltage determined
from Eq. (4) with V0 ¼ 100 V, Rs ¼ 1000 O and with Rp as shown on each curve.
Fig. 4. The available energy changes relative to the initial conditions, for the CR circuit shown
in Figure 3.
These selected conditions make each entropy and available energy curve symmetrical
about the steady state voltage for the particular conditions. The entropy curves are
given by Eq. (7). The curves of available energy change can also be obtained from
Eq. (7) by multiplying the entropy by (T0). The curves appear like inverted entropy
change curves as seen in Figure 4.
At the metastable equilibrium (when Rp ¼ 1), the entropy of the system is a
maximum and the available energy is a minimum as for normal equilibrium, but the
system still contains available energy in the form of stored energy in the battery and
capacitor.
While these curves can be determined directly from Eq. (7), they can also be
determined from a more fundamental model simulated numerically on a spreadsheet
or computer program which allows the capacitor to charge incrementally in finite
steps, while checking at each step to see if the available energy decreases (or entropy
increases) and if so, proceeding to the next step. This would appear to be the way in
which the process actually progresses in nature. This demonstrates the relationship
between the physics of progression of a physical process and the thermodynamic laws
which govern it. That is to say, nothing can happen unless the particular change
(either infinitesimal or extended to finite) brings about a decrease in available energy
(or an increase in the total entropy). This type of incremental model, therefore,
provides a fundamental explanation of why any physical change actually occurs.
Since available energy function has the appearance of a potential energy well, it fits in
with the idea of mechanical and other physical systems tending towards ‘‘lower
energy states’’ as, for example, when a rock rolls into a valley. For that reason, as well
as the fact that it has a clear physical meaning, available energy has a greater
conceptual and pedagogical value than does entropy. Thus, it may be assumed that all
processes and all changes are driven by reducing available energy. The fact that
entropy also increases may be regarded as an outcome of reducing available energy,
by virtue of the Guoy-Stodola theorem.
This methodology can be applied to systems which tend, by means of irreversible
processes, towards the three possible states, namely equilibrium, metastability or steady
state. That is to say, the first and second laws of thermodynamics are capable of explaining
how transient irreversible processes lead to steady states as well as to equilibrium and
metastable states and there is no need to make any other assumptions for steady states.
Thus regardless of its veracity in certain cases (which appears to be only coincidental
in the present case), the Onsager-Prigogine principle appears to be unnecessary, for
there is no reason to believe that entropy production rates have any effect on the final
state to which a system progresses. This will be confirmed in the following.
Fig. 5. Entropy production rate as a function of current for the LR circuit shown. Arrows show
the direction of change towards the steady state. S2 allows an initial current greater than the
steady state current.
steady state value Iss. Steady state is reached when I ¼ Iss ¼ V0 /R, at which time the
voltages across the resistor and inductor are VR ¼ V0 and VL ¼ 0, respectively.
Taking the current I as the independent variable, the power dissipated is P ¼ I 2 R so
that the entropy production rate is
ds ðI 2 RÞ
¼ : (9)
dt T0
where Ii is the initial current through the inductor which can be greater or less than
the steady state current I ¼ Vo /R, and t ¼ L/R is the time constant of the circuit.
The total entropy change of the system, relative to the initial conditions, is calculated
from an initial condition where t ¼ 0 and the current is Ii. Taking I as the independent
variable, and noting that ds/dI ¼ (ds/dt)/(dI/dt) and from Eq. (10) that
dI 1 Vo
¼ I ; (11)
dt t R
ds tR I2
¼ : (12)
dI T0 ðVo =R IÞ
The change in entropy given by Eq. (13) is plotted as a function of I in Figure 6 for
the particular circuit values shown.
Fig. 6. Entropy change as a function of current for the LR circuit shown at 300 K with initial
current of either 0 A or 2 A. Arrows show direction of change towards the steady state.
or more simply as T0Ds from the Guoy-Stodola theorem, both of which give the
change from the initial current conditions (either Ii ¼ 0 or 2A) as a function of
current. The change in available energy is plotted in Figure 7 for the particular LR
circuit shown.
From Figures 6 and 7, the steady state condition is given by
which can be seen to be the inverse of the equilibrium condition when written as:
1 1
¼0 or ¼ 0: (14b)
dDstot =dI deavail =dI
5. AC Circuits
Fig. 7. Available energy change from the initial condition as a function of current for the LR
circuit. Arrows show direction of change towards the steady state.
and y is the phase angle at time zero. The current through the circuit is then given by
V0
i ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ½sinðot þ y fÞ expðRt=LÞsinðy fÞ ; (15)
R þ o 2 L2
2
where f ¼ artanðoL=RÞ [9]. The waveform of the current with time consists of a
sinusoidal component plus a transient exponentially decaying component iexp given
by the exponential component of (15), that is
V0
iexp ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ½expðRt=LÞsinðy fÞ : (16)
R þ o2 L2
2
ðt
1
Ds ¼ I 2 Rdt0 :
T0 0
Near the starting point, each cycle represents a ‘‘bump’’ on the curve which becomes
smaller and smoother as the system approaches the steady state. However, the entropy
of the system and surroundings increases rapidly as the exponential component of the
current approaches zero, that is, as the system approaches the steady state condition.
Fig. 8. Transient initial current for the circuit shown having series resistance and inductance
of 10 O and 1.0 henry, respectively with a 50 Hz AC applied voltage of 100 V peak with zero
phase angle at t ¼ 0.
Thus,
1=ðdstot =dIexp Þ ¼ 0; (17)
which has the form of the general dissipative steady state case where 1=ðdstot =dcÞ ¼ 0.
This is the same condition at the steady state that occurs in all other cases examined.
The rate of entropy change, found from Eq. (9), is not very meaningful in this case
since it is zero wherever the current is zero and a maximum at peak current. Since a
minimum occurs every cycle, the Onsager-Prigogine principle cannot be tested in
this case.
Steady state currents in any electrical circuit are not instantaneous because (a) all
electrical circuits contain inherent inductance and capacitance, (b) heating and
cooling effects are associated with currents and (c) electrons or other carriers have
some inertia. The time lag associated with (c) is usually much smaller than for (a) or
(b) and is not included in the present analysis. Thus, ordinary steady current flow is
the result of the system changing in such a way that entropy continuously increases or
available energy continuously decreases which is characterised by the conditions
given by Eq. (14). In general, it appears that all three types of states, including steady
states, are determined only by an increase in total entropy or a corresponding
decrease in available energy and are independent of the rate at which these occur.
This will now be tested in the case of the heated filament.
7. Electrically-Heated Filament
Steady states are defined by a first law power balance, but how the system moves to a
steady state is a second law problem. An electrically-heated filament in a light globe,
for example, attains its steady state of operation by balancing the electrical input
power against its radiation and heat output. In reaching this condition, changes in
temperature, radiative power, entropy and available energy occur. A simplified
spreadsheet model of a filament, briefly described below, has been used to
demonstrate the thermodynamic process by which the steady state is reached from
a specified initial condition. It shows that the system is driven into steady state by
increasing entropy or decreasing available energy, but steady state does not coincide
with minimum entropy production rate. In developing the model, a number of
assumptions were made. Firstly, the filament operates in a vacuum, and has radiative
output but no convective or conductive heat output components. Secondly, the
filament is assumed to be straight with no radiative power flow from one part of the
filament to another. Thirdly, both resistivity and specific heat are assumed to be linear
functions of temperature approximating the properties of tungsten up to 2500 K.
Fourthly, the thermal emissivity of the filament e is assumed to be 1.0 which is higher
than the actual value for tungsten, but has the advantage that the instantaneous
temperature of the total radiation emitted is equal to the corresponding instantaneous
temperature of the filament. Thus the entropy of the radiation is easily determined.
The overall arrangement which has been modelled is represented in Figure 10. The
system is taken as the filament itself, heated by the electrical energy supplied by the
electrical energy sink. Its black body radiation is used to drive a heat engine having
Carnot efficiency (Z ¼ 1). The entropy of the radiation is equal to the heat rejected by
the heat engine at T0 (298 K) divided by T0. For the filament, the equations for
specific heat and entropy [10], the resistivity [11], resistance and electrical and
radiative power are, respectively:
V2
Pelec ¼ W
RðTÞ
Pradn ¼ sT 4
A W
In these equations, l is the length and a is the area of the cross section of the wire, V is
the fixed operating voltage, s is Stefan’s constant and A is the total radiating area of
the filament. At the steady state condition the first law power conversion is:
Pelec ¼ Prad’n.
For a filament 450 mm long and 0.02 mm diameter operating at 230 V, the steady
state condition occurs at 2395 K. In modelling the approach of the system to the
steady state, two initial conditions have been used. The first has the filament at T0
Fig. 11. The entropy of the filament and surroundings relative to 298 K as the filament is
heated or cooled from its initial condition of 298 K or 3000 K.
(298 K), the second has the filament temperature at 3000 K. The model operates on a
time sequential basis, with time increments between 20 ms and 200 ms. It first
determines the energy input and radiation output of the filament from its (previous)
temperature and resistance. The new temperature of the filament is then determined
from its previous, and the nett energy input over the particular time-step is calculated.
The entropy of the filament is calculated from the above equation, while that of the
black body radiation emitted is determined from ds ¼ der =Tr , where er and Tr are the
energy and temperature of the emitted radiation. The total entropy produced and the
rate of its production in moving from the two initial conditions to the steady state
condition are shown in Figures 11 and 12, respectively.
The model shows that the approach to steady state at 2395 K is characterised by a
very rapid increase in the total entropy of the system and its surroundings relative to
the initial conditions, as occurs in all other steady states examined.
Again, as in most other cases, a minimum in the rate of entropy production does not
occur at the steady state condition. Furthermore, the rate of entropy production is a
continuous function with respect to filament temperature, and is quite unaffected by
the maximum in the total entropy of the isolated system at 2395 K. Thus the particular
system does not behave according to the Onsager-Prigogine principle. It may be noted
that the entropy production rate associated with the system, that is the filament, is zero
at the steady state. This is because, by definition, there is no change in the system itself
at the steady state, although there is a steady flow of heat through it.
The physical cause of the rapid increase in total entropy with temperature, as
temperature approaches its steady state value, is because the rate of change of the
Fig. 12. The rate of production of entropy of the filament and surroundings relative to the
initial conditions as the filament is heated or cooled from 298 K or 3000 K to its steady state
temperature at 2395 K.
filament temperature slows greatly as the radiated power approaches the input power,
but the rate of entropy production changes only slowly.
8. Conclusions
Simple electrical circuits have been shown to provide useful models for determining
the nature of transient thermodynamic changes which lead to the three possible final
states of equilibrium, metastability and (dissipative) steady states. The thermo-
dynamics of four electrical systems (CR DC; LR DC; LR AC; and a hot electrically-
heated filament) have been examined with respect to their approach to a steady state,
in which a steady electric current passes through the system. It has been found quite
possible to analyse these non-equilibrium conditions using conventional thermo-
dynamics, without any need for the contemporary theory known as the ‘‘thermo-
dynamics of irreversible processes’’.
The discharge of a capacitor in a circuit which does not include a battery or emf
source leads to an equilibrium state having maximum total entropy and zero available
energy. However, the charging, by an ideal battery, of a perfect capacitor with no
leakage leads to a metastable state which is characterised by a maximum in the total
entropy and a non-zero minimum available energy of the system and its surroundings.
Thus, both equilibrium and metastability are characterised by:
dstotal deavail
¼0 or ¼ 0;
dc dc
where c is a single valued variable (such as voltage) which indicates the state of the
system. However, the available energy is finite for the metastable state, and is zero for
equilibrium.
Where a system does not move towards equilibrium or metastability, it moves
towards establishing a steady state in which energy is dissipated as heat in the
resistive element of the circuit. It has been shown that the approach to steady state
is characterised by an increasing total entropy or a decreasing available energy of
the system and its surroundings as the steady state is approached. In keeping with the
behaviour of other steady states which have been examined, the condition at the
steady state is characterised by
The transition from a system which tends towards metastability to one which tends
towards a steady state has been illustrated by decreasing the parallel (leakage)
resistance, Rp, in a RC circuit from a theoretical infinite value to a finite value.
Changing Rp from an infinite value to any finite value, causes the system to change
from a metastable state to a steady dissipative state with a corresponding change from
a maximum in total entropy to a total entropy which theoretically approaches infinity
at the steady state. Corresponding changes in the available energy are also illustrated.
All of the transient changes which occurred in all the systems were apparently driven
by increasing total entropy or decreasing available energy. Thus, it appears that there
is no fundamental difference in the thermodynamic causes of change in the different
types of system, whether the system moves toward an equilibrium state, a metastable
state or a steady state.
In one of the four circuits (direct current RC circuit with non-infinite parallel
resistance), the steady state coincided with a maximum in the rate of entropy
production, but the approach to the steady state was still characterised by con-
tinuously increasing entropy or decreasing available energy.
No transient changes of any system appeared to be driven by the minimisation of
entropy production rate as suggested by the Onsager-Prigogine theory, although in
one case only, the minimum entropy production rate coincided with the steady state.
Thus, it appears that all changes in all circuits are driven only by increasing entropy
and are unaffected by the rate at which entropy is produced.
Acknowledgments
The author would like to thank Professors A. Brownrigg and P. Hodgson for their
encouragement, discussions and suggestions. The support of an Australian Postgraduate
Award is acknowledged.
References
[1] Onsager, L., Reciprocal relations in irreversible processes II, Phys. Rev., 37 (1931), 2265–
2279.
Kenneth M. Browne
School of Engineering and Technology
Deakin University
Pigdons Road, Geelong
Victoria, 3217
Australia
Email: kmbrowne@deakin.edu.au