WREC 1996

VALUE ANALYSIS OF ADVANCED HEAT REJECTION SYSTEMS

Carl Bliem Federica Zangrando Vahab Hassani

ABSTRACT

A computer model is developed to evaluate the performance of the binary geothermal power plants (Organic
Rankme Cycles) with various heat rejection systems and their impact on the levelized cost of electricity.

KEzwoRDs

Heat Rejection, Organic Rankine Cycles, LEC, Computer Model

INTRODUCTION

Researchers at the National Renewable Energy Laboratory (NREL) have developed a computer model to
evaluate the performance of the geothermal power plants with various heat rejection systems and their impact
on the levelized cost of electricity (LEC). The computer model developed in this work is capable of
simulating the operation of a geothermal power plant which consists mainly of an Organic Rankine Cycle
(binary plants) with different types of working fluids such as pure hydrocarbons and some binary mixtures
of the most promising combinations of hydrocarbons, The computer model performs the cycle analysis and
component sizing for binary systems with various heat rejection systems. A spread sheet is then used to
carry on the economic data calculation and analysis. The value analysis technique used in this work is the
method described by Demuth and whitbeck (1982). In this method, the impact of a new system on LEC is
calculated based on an incremental technique which determines the change in the cost of producing electricity
between a baseline configuration and another configuration expressed as a fractional change. The cycle
analysis computer program accepts five different heat rejection models. The following configurations are
now operational in the computer program: l)Air-cooled condenser, 2)Shell-and-tube (surface) condenser with
evaporative cooling tower, 3)Shell-and-tube (surface) condenser with the cooling water first cooled in an air-
cooled exchanger and then in an evaporative cooling tower, 4)Parallel configuration of 1 and 2, 5)Series
contiguration of 1 and 2 (air-cooled condenser at the hot end). The program operates in two modes: “design
mode” in which all components are sized for a particular set of design parameters such as pinch points, and
“operational mode” in which the component hardware is fixed and the cycle state points change to
accommodate a different set of ambient conditions.

TYPES OF SYSTEMS ANALYZED

Figures 1 shows the type of binary cycle system which can be- analyzed by the computer program. Figure
1 shows a simple Rankine cycle, which may be a boiling cycle or a supercritical cycle depending on the
heater pressure. This paper will emphasize the work with the unrecuperated cycle. Studies at a later time

1250
 WFtEC1996
time will investigate the recuperated cycle.

Five different heat rejection systems can presently be analyzed by the computer program. First, and
perhaps the simplest system, consists of an air-cooled condenser. This is, at the present time, the most
prevalent heat rejection system for binary geothermal plants. The advantages of this system are its
simplicity and lack of a requirement for water. The primary disadvantage of this system may be a high
condensing temperature produced when ambient temperatures are high (Summer and daytime). This
system will be referred to in the remainder of the report as Case 1. The second most prevalent type of
heat rejection, Case 2, is a totally evaporatively cooled system. The working fluid from the turbine or
recuperator is condensed in a shell-and-tube (surface) condenser with cooling water on the other side of
the exchanger. Case 3 is one of three hybrid systems using both evaporative and dry cooling. This may
be thought of as an extension of Case 2 in which the cooling water from the condenser first passes
through an air-cooler and then through the evaporative tower of Case 2. Case 4 is the second hybrid
arrangement. Here, the working fluid is split as it leaves the turbine (or the recuperator) and part is sent
through a standard air-cooled condenser (similar to the one in Case 1) and the remainder is sent through
a shell-and-tube (surface) condenser with cooling water cooled in an evaporative tower (as in Case 2).
For this arrangement, the flow split may be varied in response to varying ambient conditions to optimize
the system performance. Case 5 is the other hybrid arrangement. Here, all of the turbine exhaust goes
through an air-cooled condenser. Then, the uncondensed portion goes through a shell and tube condenser
with the cooling water evaporatively cooled. If the working fluid is a pure substance, the liquid which
is condensed in the air-cooled condenser will bypass the second condenser. If the working fluid is a
mixture, this choice is more difficult. Bypassing the partially condensed liquid will result in worse
performance than if the phases can be intimately mixed in the second condenser. However, this may be
difficult to do. Further investigation of this problem will be considered at a later date.

COMPUTER PROGRAM FOR CYCLE ANALYSIS AND COMPONENT SIZING

This section discusses the computer program developed to predict the performance and size the major
components of the baseline and modified systems. The computer program provides two options: a
“design” option, which sizes the various components for a given cycle and an “off-design performance”
option, which takes the hardware from a design run and determines its performance under different
ambient conditions.

Method of Solutioa

The computer program determines the Rankine cycle for a prescribed set of initial conditions. The first
pass through the program is a “design” run. Given the design constraints, the program determines the
equipment necessary to meet the constraints. The output of this program consists of the plant
performance and an estimate of the sizing of the major components.

The “off-design performance” option takes the size parameters for the heater, turbine and heat rejection
components and with the new ambient conditions varies pinch points and flow ratios until the size
parameters are sufficiently close to the set values. This iteration assumes that the heater pressure remains
constant and that the working fluid to coolant flow rates remain the same. (This means that the working
fluid flow rate is fixed by the turbine while the coolant flow rates are the same as in the “design” case
and the condenser pinch point, cooling tower approach and geofluid flow rate can vary.)
. .
asehne Bm

Four plants were considered to span the low temperature region for small binary plants. Geofluid inlet
temperatures of 220 and 280 F were used where the geofluid might come directly from a low temperature
1251
 WREC 1996

resource or it might come from the exhaust of a high temperature plant. For the cases of a bottoming
cycle, geofluid from a high temperature plant, the geofluid reinjection temperature was restricted to 190
F to minimize silica deposition. For the low temperature resources, no outlet limit was applied. To
incorporate reliable data into the computer simulations, NREL used a database obtained from an industry
partner (Barber-Nichols) on design and cost information of these systems for both air cooled and water-
cooled evaporative heat rejection systems. Using Barber-Nichols data base, the performance and cost data
for these systems were evaluated and utilized as baseline for the computer program. Using “rules of thumb”
and case studies, optimum systems were determined. Isobutane was used in these binary cycles.

APPLICATION OF THE VALUE ANALYSIS TOOL AND COMPUTER MODEL

The use of lighter hydrocarbons have been recommended for geothermal binary plants with low resource
temperatures. To explore this recommendation and to illustrate the use of the computer model developed
here, the replacement of isobutane in the baseline power plant with propane for the 280 F resource with
no geofluid reinjection temperature limit was studied. This example will demonstrate that the value
analysis method and the computer model can be used to minim& the LEC by changing the basic design
parameters of the cycle, e.g., the pinch points in the heat exchangers and condensing pressure.

Figure 2 shows a composite of the optimization choices with the new working fluid. The fractional changes
in the LEC are with respect to the isobutane baseline described earlier. The highest curve is for a heater
pressure of 500 psia. This is higher than the 3 13 psia pressure in the baseline case. No account was taken
of the impact of this higher pressure on cost of piping and heater because the piping, fittings and valves will
be of the same class of service and that any increase in cost of the heat exchanger shell will be negligible
for this change in pressure. The 500 psia heater pressure did not produce a decrease in LEC over the
baseline value. Increasing the heater pressure to 600 psia did result in a decrease in LEC for condensing
pressures below 94 F. A brief study of the heater pinch point using 4 and 6 F indicated little difference with
the larger pinch point giving the lower LEC. The optimum condensing temperature was 85 F with a 600
psia heater pressure and a 6 F heater pinch point. The resultant decrease in LEC was almost 4% below that
of the baseline cycle.

Figure 3 indicates the differences in net geofluid effectiveness (Net W h/lb, of geofluid). The increase in
net geofluid effectiveness for all of the propane cycles is 20 to 30% above that of the baseline system. This
means that the power plant will use 20 to 30% less geofluid to produce the same amount of new electrical
energy and deplete the reservoir at a much lower rate. One might expect this large improvement in
performance to be reflected in a larger decreased energy cost. Field costs were only 15 to 25% of the LEC
for the baseline cases considered here. Classically, for larger plants this number is near 50% which would
result in a much greater impact on LEC. It is planned in future work to consider larger size plants. This
may result in a stronger preference for propane over isobutane in this type of application.

CONCLUSIONS

A computer model was developed to evaluate the performance of the binary geothermal power plants
(Organic Rankine Cycles) with various heat rejection systems and their impact on the levelized cost of
electricity. The computer model is capable of simulating the operation of an Organic Rankhre Cycle with
different types of working fluids such as pure hydrocarbons and some binary mixtures of the most
promising combinations of hydrocarbons. The computer model performs the cycle analysis and
component sizing for binary systems with various heat rejection systems. A spread sheet is then used
to carry on the economic data calculation and analysis. The method of computing the LEC is an
incremental method to determine the percentage change from a base case. The cycle analysis computer
program accepts five different heat rejection models. The program operates in two modes: “design”
mode, and “off-design performance” mode in which the component hardware is fiied and the cycle state
1252
 WREC 1996

points change to accommodate a different set of ambient conditions.

REFERENCES

Demuth, O.J. and J.F. Whitbeck (1982). Advanced Concept Value Analysis for Geothermal Power
Plants. EGG-GTH-5821.

; Condensing Temperature,F

Fig. 2 LEC for Propane cycle compared to the Baseline lsobutane cycle

D
s
.$ 1.5
0
g 1
t
'--
x
$ 0.5
'i;
5 0
'ii
Y 80 85 90 95 100 105 110
t
Condnesing Temperature,F
Fig 3. Net Effectiveness for a 280 F resource using Propane

/
1253