Documente Academic
Documente Profesional
Documente Cultură
Aim
The aim of this lecture is to introduce Thermoeconomics as an analytical and
powerful tool for the cost accounting, diagnosis, improvement, optimization and
design of energy systems using the combination of Second Law of
Thermodynamics and Economics.
OBJECTIVES
The lectures will comprise the following objectives:
§ A vision about the importance of Thermoeconomics as a tool for improving
energy systems and prevent damage on envioronment.
§ Introduce the basic concepts of exergy, cost, exergetic and monetary costs,
fuel, product, unit exergetic consumption and efficiency.
§ Describe the process of cost formation and distinguish between the physical
and the thermoeconomic plant models. The productive structure and its
mathematical representation.
§ Analyze the basic concepts to diagnose and optimize energy systems using
thermoeconomics.
*
Director of CIRCE and Chair on Thermal Systems
2
CONTENTS
1. Introduction to Thermoeconomics
2. Basic Concepts.
2.1. The Concept of Cost.
2.2. Fuel, Product and Unit Exergetic Consumption.
2.3. Physical and Thermoeconomic Plant Models.
3. Calculating Thermoeconomic Costs
3.1. Economic Resources and Thermoeconomic Costs
4. Thermoeconomic Applications to Complex Energy Systems .
4.1. Operation Thermoeconomic Diagnosis
4.1.1. Technical Exergy Saving
4.1.2. Impact on Resources Consumption
4.1.3. Malfunction and Dysfunction Analysis
4.1.4. Intrinsic and Induced Malfunctions
4.2. Thermoeconomic Optimization
KEYWORDS
Thermoeconomics, diagnosis, optimization, costs, exergy, exergetic cost,
irreversibility, malfunction, fuel, product.
GLOSSARY
Thermoeconomics: Science that combines Thermodynamics and Economics in
order to avoid the natural resources consumption in processes.
Malfunction: Effect of an inefficiency in a/several process units of a system.
Exergy: Amount of available energy in a physical process.
Cost: Amount of resources to obtain a product.
1. Introduction to Thermoeconomics
As the human population grows, our finite world is becoming smaller and natural
resources are more and more scarce. We must conserve them in order to
3
survive and Thermoeconomics plays a key role in this endeavor. We should find
out how energy and resources degrade, which systems work better, how to
improve designs to reduce consumption and prevent residues from damaging the
environment. Thermoeconomics and its application to engineering energy
systems can help to answer these questions.
The production process of a complex energy system (e.g. a complex power
plant) can be analyzed in terms of its economic profitability and efficiency with
respect to resource consumption.
An economic analysis can calculate the cost of fuel, investment, operation and
maintenance for the whole plant but provides no means to evaluate the single
processes taking place in the subsystems nor how to distribute the costs among
them.
On the other hand, a thermodynamic analysis calculates the efficiencies of the
subsystems and locates and quantifies the irreversibilities but cannot evaluate
their significance in terms of the overall production process.
Thermoeconomic analysis combines economic and thermodynamic analysis by
applying the concept of cost (originally an economic property) to exergy (an
energetic property), (see e.g. Valero et al. (1986)). Most analysts agree that
exergy is the most adequate thermodynamic property to associate with cost
since it contains information from the second law of thermodynamics and
accounts for energy quality (Tsatsaronis (1987, 1998), Gaggioli and El-Sayed
(1987), Moran (1990)). Exergetic efficiency compares a real process to a
reversible one, (i.e. an ideal process of the same type). An exergy analysis
locates and quantifies irreversibilities in a process. Exergy based
thermoeconomic methods are also referred to as “exergoeconomics”
(Tsatsaronis and Winhold (1985)).
The physical magnitude connecting physics (thermodynamics) and economics is
entropy generation or, more specifically, irreversibility. This represents the
“useful” or available energy lost or destroyed (exergy destruction) in all physical
processes. All real processes in a plant are non-reversible and, as a
consequence, some exergy is destroyed and some natural resources are
consumed and lost forever, which creates cost. All natural resources have an
economic cost: the more irreversible a process, the more natural resources are
consumed (higher energetic cost) and the higher the required investment (higher
thermoeconomic cost). If we can measure this thermodynamic cost by
identifying, locating and quantifying the causes of inefficiencies in real
processes, we can provide an objective economic basis using the cost concept.
Thus, thermoeconomics assesses the cost of consumed resources, money and
system irreversibilities in terms of the overall production process. Consumed
4
2. Basic Concepts
The cost of a flow in a plant represents the external resources that have to be
supplied to the overall system to produce this flow. Thermoeconomic analysis
distinguishes between exergetic costs and monetary costs.
The exergetic cost (Valero et al (1986)) of a mass and/or energy flow is the
units of exergy used to produce it, e.g. the exergetic cost of the net power is the
exergy provided by the natural gas to generate the electrical power delivered to
the net by the co-generation plant (see Figure 1). These costs are a measure of
the thermodynamic efficiency of the production process generating these flows.
The unit exergetic cost of a mass and/or energy flow represents the amount of
resources required to obtain one unit of exergy. Thus, if the unit exergetic cost
of the electricity is three, three units of plant exergy resources are consumed to
obtain one exergy unit of electrical power.
6
1
Com bus tor
1
2 3
5 6
Compres sor Turbine
2 3
0 4
HRSG
Ai r 7
Gases 4
Natural Gas
Work 8
Water/Steam
The average costs are only known after production, when we know how many
resources were used and the production obtained. The average cost is not
predictive. Knowing the average unit cost of a product does not provide the cost
7
7 Branching 2 P3 = B5 + B6 = F2 = B5 = Wcp
Wcp + Wnet B6 = Wnet
Table 2. Fuels and Products of the process units of the co-generation plant.
A plant is analyzed using a physical model with a set of equations to describe the
physical behavior of the process units. It calculates parameters such as
temperatures, pressures, efficiencies, power generated etc. to describe the
physical state of the plant. Depending on the analysis, a decision has to be taken
on the detail required i.e., which flows and process units are to be considered.
The process units for the analysis do not necessarily correspond to physical
units. Various parts of the installation can be combined into one process unit and
physical units can be further disaggregated. It is important to choose an
appropriate aggregation level that properly defines the behavior of each process
unit and its purpose in the overall production process. The physical structure
(see Figure 1) depicts the process units, mass and connecting energy flows
considered in the physical model.
10
Comp F2 = B5
F3 = B3 – B4 Wnet
P3
Turb b2
P2 = B2– B0 Pj1 = B3
j1 b1
P1 = B3 – B2 P4 = B7 = Bheat
HRSG
F4 = B4
F1 =B1
Comb
where the index i refers to the input flows of the process unit l, the index j
refers to the output flows of the process unit l, and m is the number of flows
considered in the productive structure. Every flow is an input flow of a process
unit and an output flow of another process unit or the environment. For the flows
interacting with the environment, we define:
Bm-s+1 = ωi i = 1,…,s (8)
where s is the number of system outputs, and ωi is the total system product, i.e.
an external variable which determines the total product. The characteristic
equations for the system in Figure 2, are shown in Table 3:
1 Combustor F1 P1 F1 = g (x 1 , P1 ) = k P1
F1 cb
2 Compressor F2 = Wcp P2 F2 = g (x 2 , P2 ) = k P2
F2 cp
3 Turbine F3 P3 = Wgt F3 = g (x 3 , P3 ) = k P3
F3 gt
4 H.R.S.G. F4 P4 =B HEAT=ω 4 F4 = g (x 4 , P4 ) = k P =k ω4 = k B
F4 HRSG 4 HRSG HRSG
HEAT
Euler´s Theorem states that the homogeneous function of first order verify:
∂g
Bi = i B + ∂g i B + ... + ∂g i B l1 ,…,ls in Sl (10)
∂Bl l1
∂Bl l2
∂ Bl ls
1 2 s
∂g i
Bi = ∑ κ ij B j κ ij = i=1,...,m l=1,...,n. (11)
j∈Sl ∂B j
This property means that the input of a process unit varies at the same rate as
its outputs. Note that this property does not imply that the function must be
linear.
14
κij are the technical production coefficients and represent the portion of the i-th
process unit production:
∂g i
κ ij = (12a)
∂B j
The sum of κij coefficients of a unit is the unit exergy consumption of that unit:
n
n ∑ Fi Fj
κj= ∑κ ij = i =0 = (12b)
i =0 Pj Pj
This equation shows how the production of a process unit is used as fuel by
another unit or as a part of the total plant production. In the above expression,
Bij is the production portion of the i-th process unit that fuels the j-th process
unit, and Bi0 represents the production portion of the process unit i leading to the
final plant product (the subscript 0 refers to the environment, which is
considered another process unit interacting with the plant).
Equation (14) can be expressed in terms of the unit exergetic consumptions as:
n
Pi = Bi 0 + ∑ κ ij Pj i = 0,1,…,n (15)
j =1
P = Ps + KP P (16)
where Ps is a (n×1) vector whose elements contain the contribution to the final
production of the system Pi0 obtained in each process unit, and 〈KP〉 is a (n×n)
matrix, whose elements are the unit exergy consumption κij. This expression
helps to relate the production of each process unit as a function of the final
production and the unit consumption of each process unit:
In the same way, we can express the irreversibility of each process unit as:
I = I Ps where I ≡ (K D − U D ) P (18)
FT = tκ e P Ps (19)
Once the thermoeconomic model has been defined and the characteristic
equations corresponding to the productive structure of the system are known,
the costs of all flows in the productive structure can be easily calculated.
There are two different types of thermoeconomic costs: average costs and
marginal costs (equations 1 and 2). It is important to note that (as discussed
below) the average and marginal costs coincide when the characteristic
equations of the thermoeconomic model are first grade homogeneous functions
(Serra (1994), Reini (1994), Uche (2000)).
This result is very important since both costs can be calculated using the same
procedure. Marginal costs are a derivative (see equation 2) and can be
calculated by applying the chain rule of the mathematical derivation. Similarly,
average costs can also be obtained from the rules of the mathematical derivation
applied to the thermoeconomic model when the characteristic equations are first
grade homogeneous functions.
According to the previous premises, the cost of the plant resources can be
defined as:
e
Bo = ∑ k*o ,i Bi (20)
i =1
where e, is the number of system inputs, and k *o,i is the unit cost of the –i–
external resource.
Each flow, as a process unit input, is a function (defined by the characteristic
equation) of a set of internal variables, x, external variables ω and the output
flows of the process unit. The cost of the plant resources is then a function of
each flow, the set of internal variables of each process unit and the final product
of the plant B0 = B0 (Bi, x, ω ), according the relations (7) and (8).
18
∂B0 m ∂B0 ∂g j
= ∑ i = e+1,… ,m (21b)
∂Bi j =1 ∂B j ∂Bi
j ≠i
∂ B0
The expression represents the marginal costs which evaluate the additional
∂ Bi
consumption of the resources, when an additional unit of the flow –i– is
produced, under the conditions that the internal variables, x, do not vary
throughout this process.
∂g i
We can denote these marginal costs as k *i, and κ ij ≡ the marginal
∂B j
consumption of flow –i– to produce the flow –j–, then we can rewrite the
previous expressions, as:
Note that the unit exergetic cost of each fuel entering the plant is unity because
there is no energy quality degradation nor exergy destruction at the very
beginning of the productive process. Hence, the amount of exergy consumed to
obtain each plant’s fuel is its own exergy content and therefore its unit exergetic
cost equals one.
It can easily be proved that the cost of each flow P* ij of the productive
structure using the Fuel/Product notation is:
And the exergetic cost of the product of each process unit is the same as the
cost of the resources needed to obtain it, hence:
19
n
Pi* = Fi* = ∑ k*P , j B ji i = 1,…,n (24)
j =0
This cost equation can also be expressed in terms of the unit exergetic
consumptions:
n
k *P ,i = κ 0 i + ∑ κ ji k *P , j i =1,…,n (25)
j =1
which can be used to obtain the unit exergetic cost of the flows appearing in the
productive structure diagram as a function of the unit exergetic consumption of
each process unit.
Then, if the characteristic equations and the marginal consumptions for each
*
process unit are known, the marginal cost k for each flow can be obtained by
solving the system of linear equations (25).
Example 1
For the example of a co-generation plant (Figure 2), equations (21a), (21b) can
be written as:
∂B1
k*F1 =
∂F1
∂ B1 ∂B1 ∂ Pj 1
k*F3 = = = k*Pj 1
∂ F3 ∂Pj1 ∂ F3
∂B1 ∂ B1 ∂ Pj1
k*F4 = = = k *Pj 1
∂F4 ∂ Pj1 ∂ F4
∂ B1 ∂B1 ∂F1
k*P1 = = = k*F1 kcb
∂ P1 ∂ F1 ∂ P1
∂ B1 ∂B1 ∂F2
k*P2 = = = k*F2 kcp
∂P2 ∂F2 ∂P2
∂B1 ∂ B1 ∂ F4
k*P4 = = = k*F4 k HRSG
∂P4 ∂F4 ∂P4
∂B1 ∂ B ∂P3
*
kW = = 1 = k*P3
net
∂ Pnet ∂P3 ∂ Wnet
Economic Zl = Z l ( Bl , B j , B h)
Resources
B
j
B
xl B
0 B h
i
Flow (kW) 1 2 3 4 5 6 7 8
Design 11 781 2704 9614 3831 2977 2500 2355 388
Operation 11 914 2758 9753 3887 3056 2500 2355 424
Table 4. Design and operation exergy flow values of the co-generation plant
(Figure 1)
23
j =1
(31)
However, even though the methods based on Second Law Analysis (Kotas
(1985)) and Technical Exergy Saving are useful to quantify the additional fuel
consumption, they fail when trying to identify the real causes of the additional
resources consumption.
P = Ps + KP P
For the sake of simplicity we did not consider thermal and mechanical exergies
as separate entities. Two auxiliary variables also appear r1 = (B3-B2)/B3 and r2
= B3/B2, which correspond to the part of the fuel of the turbine and the heat
24
recovery steam generator (HRSG) coming from the combustor and the
compressor respectively. Flow #8, produced in part in the combustor and in the
compressor, also leaves the system as a residue. Only a part of the entering
gases to the turbine: B3-B8 are used as a fuel of other process units of the
system. Therefore, only a part of the combustor’s and compressor’s product is
used as a fuel for other process units (useful product). Accordingly, Figure 4
shows the chosen disaggregation scheme of the system and the Fuel/Product
values for the design conditions are shown in Table 5. The F-P definition is
shown in Table 6.
(5)
2
(6)
(3)-(4) 3
(2)
(3)-(2)
(7)
(4)-(8) 4
(1) (8)
1
F0 F1 F2 F3 F4 Total
P0 0 11 781 0 0 0 11 781
P4 2355 0 0 0 0 2355
Table 5. Fuel and energy flows (kW) in design conditions for the co-generation
plant shown in Figure 1
1 Combustor B1 B3-B2
2 Compressor B5 B2
3 Turbine B3-B4 B6
4 HRSG B4-B8 B7 B8
∆FT =∆ t κ e P 0 + t κ e ∆P (32)
The increase of the process unit production from equation (16) may be
expressed in terms of the unit exergy consumption as:
∆P = ∆Ps + ∆ KP P 0 + KP ∆P (33)
26
(
∆FT = ∆t κ e +t k*P ∆ KP P 0 ) (36)
or in scalar format:
n n
∆FT = ∑ ∑ k*P, j ∆κ ji Pi 0 (37)
i =1 j = 0
The ∆〈KP〉 matrix is the key to predict the impact on fuel of a physical variation
of a parameter in the system.
Using the above equation, the additional resource consumption ∆FT (also called
Fuel Impact; Reini (1994)) can be expressed as the sum of the contributions of
each process unit.
The variation of the exergetic unit consumption of each process unit increases
its resources consumption and its irreversibilities in a quantity ∆κ ji Pi0 , which we
call, malfunction. Consequently, this implies an additional consumption of
external resources given by k *P, j ∆κ ji Pi 0 , which is also named the malfunction
cost. Therefore, the total fuel impact can be written as the sum of the fuel
impact or malfunction cost of each process unit, as shown in equation (37).
The proposed method provides the exact values of the additional resource
consumption of each process unit malfunction for any operational state. Other
methods, such as the Theory of Perturbations (Lozano et al. (1996)), only
provide an approximate predictive value, based on marginal costs (Lagrange
multipliers) which is valid for an operating state close to the reference
conditions.
Figure 5 compares the fuel impact and the increase of irreversibilities or the
technical exergy saving of each process unit and also compares (first column)
the malfunction and the fuel impact for each process unit. Three malfunctions in
the plant are shown in the combustor, the compressor and the HRSG. The
largest irreversibilities increase is in the combustor, but the largest fuel impact is
27
in the compressor. The question that arises is, What causes the irreversibilities
increase and the fuel impact, and how are they related?
80
Fuel Impact
Malfunction
60 Technical Saving
40
20
0
Combustor Compressor Turbine HRSG
∆F1 ∆ I1
I1
∆P 1 ∆F 2 ∆I2
F1 I2
P F2
1 P2
1 2
∆I = ∆K D P 0 + (K D − U D ) ∆P (38)
28
n
MFi ≅ ∑ ∑ ∆κ rji Pi 0 (43)
r j =1
Many thermal systems are very complex due to the number of process units
and/or its strong interdependence. This complexity makes difficult the
optimization of the system design and operation. The knowledge of the costs of
a system, which in the final instance give an economic meaning to the structural
interactions between subsystems, allows us to formulate problems related to
their optimization and also to solve these problems, under certain conditions, in a
31
very symple way. This possibility allows us in turn to assume that with the
appropiate calculation strategy the problem of the global optimization of the plant
could be reduced to a sequence of subsystem to subsystem optimization. Here
we describe strategies for optimizing complex systems as proposed by Lozano et
al. (1996). They are based on sequential optimization from process unit to
process unit using the Thermoeconomic Isolation Principle (Evans (1980)). In
this section, we see how the cost of the resources consumed by the system
varies when the unit of the cost of the resources consumed, the technical
production coefficients of the productive structure and/or external demand of
products vary. Once we know the relationship between the technical production
coefficients and the design free variables the chain rule of derivation can be
applied to distinguish the effect of a design free variable on the internal economy
of subsystems.
A process unit of a thermal system is thermoeconomically isolated from the rest
of the system if the product of the unit and the unit cost of its resources (internal
product and/or external resources) are constant and known quantities. If a unit
of a thermal system is thermoeconomically isolated, the unit may be optimized by
itself (without considering the modifications of other variables of the rest of the
system) and the optimun solution obtained for the unit coincides with the
optimum solution for the whole system.
Of course, T.I. (Thermoeconomic Isolation) is an ideal condition which cannot
be achieved in most of the real systems: Pj and k *P,i change when design
variables of other process units change ,due to feedback. But the more constant
Pj and k *P,i are, the closer to T.I. conditions and the fewer iteration loops
needed to achieve the optimal solution for the whole system. Thus, the goal is
not to achieve T.I. but to approach it as much as possible in order to obtain
maximum advantages, which include:
(1) Improvements and optimal design of individual units in highly interdependent
complex systems are greatly facilitated, as well as of whole systems.
(2) The designers can be specialized and their efforts concentrated on
designing the variables of single units, while resting assured that these
efforts yield optimum design and/or improve the overall system
(3) The convergence of the solution is faster.
To optimize individual units, the objective function of the cost of product of the
process unit –j– could be defined as:
n
Min ∑ k ij k *P ,i P j (46)
k i=0
k
32
where the unit cost of the input resources k *P,i and the production Pj are known
and constant.
In real world optimization problems, the design free variables do not necessarily
coincide with the technical production coefficients. In practice there will be a
function of the actual design free variables which can be named –x–
We say that a free variable x is a local variable of a subsystem –j– when the
production coefficients κij of this subsystem only depend on x. When a design
variable is attached to several subsystems, the previous expression must be
extended to all concerned subsystems.
To determine whether a design free variable is local or not and which process
units are involved, the cost resource impact of the design variables to each
process unit can be computed:
n ∂ κ ij ∂z P , j
AC0x, j = ∑ k*P ,i + P ∆x
(47)
i=0 ∂x ∂x
and the ratio calculated:
∆C0x, j
ε =
x
j n (48)
∑ ∆C
i =1
x
0, i
If this ratio is equal (or close) to 1, the design variable is local for process unit –
j–, if it is equal (or close) to zero, the design variable is independent of the
referred j process unit. In other cases the design variable involves several
process units.
These ideas could be used to design a strategy for global optimization problems:
(0) Determine which variables are local and which are regional (involve several
process units)
(1) Determine a sequence for local optimization of each process unit
(2) Take an initial value of the design variables
(3) Calculate technical production coefficients and unit product cost
(4) Find optimum values for local variables
(5) Find optimum values for global variables
33
Iterate from (3) to convergence when design variables or unit product cost do
not vary in the next iteration. In each iteration the unit cost of total product must
decrease.
NOMENCLATURE
LATIN SYMBOLS
k*: Exergy unit cost.
B: Exergy flow (kW).
F: Fuel (kW).
P: Product (kW).
I: Irreversibility (kW).
k: Unit exergy consumption.
g: Characteristic function.
Z: Capital cost of a process unit ($).
T: Temperature (º C).
f: Function.
r: Exergy ratio.
MF: Malfunction generated in a component (kW).
DF: Dysfunction generated in a component (kW).
DI: Dysfunction generated by a component (kW).
C: Total economic cost ($/s).
X: Variable.
GREEK SYMBOLS
η: Exergetic efficiency.
ω: Total system product.
λ: Lagrange multiplier.
κ: Technical production coefficients.
α: Coefficient in the Cobb-Douglas function.
34
∆: Difference.
ε: Ratio (thermoeconomic optimization).
SUBSCRIPTS
o: outlet
i,j: index.
e: external (inlet).
t: total.
n: stage.
t: tubes.
SUPERSCRIPTS
-1: Inverse.
t: Transpose.
0: Design conditions.
L: Local.
G: Induced.
35
r: Operating parameter.
36
REFERENCES
1. Bejan A., G. Tsatsaronis and M. Moran (1997).
Thermal Design and Optimization. John Wiley and Sons Inc. New York,
USA.
2. Brodyansky V. and A. Bandura (1993).
The prognosis for macroeconomical development and exergy. Proceedings
of the International Symposium on energy systems and ecology ENSEC’93,
153-161. Cracow, Poland..
3. El-Sayed Y. M. and M. Tribus (1983).
Strategic use of thermoeconomics for systems improvement. ACS
Symposium Series 235, 215-238. Washington D.C.
4. El-Sayed Y. M. (1988).
A Decomposition Strategy for Thermoeconomics Optimization of a
Given New Configuration, Approaches to the Design and Optimization
of Thermal Systems. Wepfer and Moran eds., 41-47. ASME. New York,
USA.
5. El-Sayed Y. M. (1996).
Second-Law-Based Analysis and Optimization of Seawater Desalting
Systems. Private Communication.
6. Erlach B. (1998).
Comparison of Thermoeconomic Methodologies: Structural Theory,
AVCO and LIFO. Application to a Combined Cycle. Department of
Mechanical Engineering. University of Zaragoza.
7. Erlach B., L. Serra and A. Valero (1999).
Structural Theory as Standard for Thermoeconomics, Energy Conversion
and Management 40, 1627-1649.
8. Evans R. B. (1980).
Thermoeconomic isolation and essergy analysis, Energy 5, Nº 8-9, 805-822.
9. Frangopoulos C. A. (1983).
Thermoeconomic Functional Analysis: A method for the optimal design
or improvement of complex thermal systems. Ph. D. Thesis. Georgia
Institute of Technology.
10. Frangopoulos C. A. (1987).
Thermoeconomic Functional Analysis and optimization. Energy 12, Nº 7,
563-571.
37