Excerpts from the California Energy Commission Report 2016 regarding

The Use of Wind Barriers to Mitigate the

Effect of Wind on Air-Cooled Condensers

Prepared for: California Energy Commission

Prepared by: Maulbetsch Consulting

July 2016
CEC-500-2016-047
a
EXECUTIVE SUMMARY

Introduction
A major challenge facing deployment of thermal power plants in California is the
significant water demands these facilities may impose on local limited freshwater
supplies. To significantly reduce power plant water demand, air-cooled condensers are
increasingly being used. Use of this technology is commonly referred to as dry cooling.
An air cooled condenser rejects heat to the atmosphere and consists of an elevated A-
frame arrangement of parallel finned tube bundles. Steam is piped from the turbine or
boiler to the top of the condenser and then as the steam flows down the tubes and
condenses, heat is transferred to air passing outside the tubes. The condensate is then
reused in the power plant. To increase air flow pass the tube bundles, air-cooled
condensers are generally elevated many meters above the ground and utilize a number
of large, low speed axial fans located beneath the condenser.
While air-cooled condensers conserve significant amounts of water, they require higher
capital and operating costs, and impose some penalties on the efficiency and output of
the generating unit. A further challenge to the use of this cooling technology is that high
ambient winds, often in conjunction with high ambient temperatures, can reduce power
plant performance and can cause potential physical stress and damage on air-cooled
condenser fans.

Project Purpose
The adverse effect of wind on the performance of large air-cooled condensers has been
recognized for many years. Wind has been shown to cause significant degradation of the
thermal performance of air-cooled condensers, and impose stress on some mechanical
elements, particularly the fans, the fan blades, the fan motors, and the gearboxes.
Deterioration in thermal performance due to wind is attributed to two mechanisms:
recirculation and fan performance degradation. Recirculation occurs when wind
patterns around the air-cooled condenser cause the hot air exiting from the top of the
condenser to be blown down, and redirected into the air entering the condenser,
resulting in air temperatures higher than ambient conditions and lower condenser
thermal efficiency.
Degraded fan performance occurs when wind passing beneath the condenser causes a
high pressure zone which reduces the amount of air being entrained by the affected
fans. This reduces the amount of air passing by the condenser tubes, resulting in lower
condenser thermal efficiency. These winds can also cause fan fatigue and failure.
It is thought that fan airflow degradation is the more important of the two wind effects.
However, this can vary with the details of the site topography, the presence of nearby
obstructions, air-cooled condenser orientation relative to the prevailing winds and wind
conditions.
A number of approaches have been taken in attempt to mitigate the effects of wind on
air-cooled condensers, including:

2
 • The installation of additional cells to add compensating capacity
• Modification of original fans
• Thermal performance enhancements with inlet spray cooling
• Use of physical wind barriers to modify the wind flow patterns around the air-
cooled condenser. This approach has been used either as a retrofit or as part of
the original design at several sites.
While a substantial amount of research has been conducted on the more general effects
of wind on air-cooled condensers, there is a lack of detailed information on the
effectiveness of this mitigation measure. The purpose of this study is to address this lack
of information.

Project Process
To understand the ways wind barriers diminish the effect of wind on air-cooled
condenser performance, the study took a three-pronged approach: field testing at an
operating power plant site, physical modeling in a wind tunnel, and mathematical
simulation using computational fluid dynamics methods. The field tests were used to
guide, calibrate, and validate the physical and computational models.
Field testing was conducted to determine the effects of windscreens on air-cooled
condenser airflow and performance, and to obtain direct measurements of fan blade
vibrations, stresses, and how they are mitigated by windscreens. The field tests were
conducted at the only air-cooled condenser in the United States that is equipped with
retractable windscreens. This allowed comparisons of the effect of wind on air-cooled
condenser performance and physical stress with and without screens under very similar
conditions.
Instrumentation was installed on and around the air-cooled condenser to monitor the air
flow and temperature patterns in addition to the performance of selected fans and
stresses on fan blades.
A physical model of the plant site and air-cooled condenser was constructed for testing
in an atmospheric boundary layer wind tunnel. Wind speed and direction under and
around the air-cooled condenser was measured to produce detailed airflow patterns
around the condenser. Mesh material of appropriate porosity was attached to the side of
the air-cooled condenser model to simulate the windscreens. Comparison with field data
was used to validate the physical model results. The model was then used to explore
limited variations on the windscreen arrangement in the field to determine the most
effective configuration.
Using computational fluid dynamics, the researchers used analytical modeling of air
flow patterns around the air cooled condenser, with and without wind barriers.
Comparison with field and wind tunnel data was used to calibrate and validate the
numerical simulation. Ultimately, the model was used to explore the behavior of
alternative wind protection approaches, for which no physical data was available.

3
Results
The dynamic loading on the fan blades was significantly reduced by the deployment of
the windscreens. On the basis of field measurements averaged over all wind directions
from northwest to southwest, the results suggest that a 50 percent deployment of the
screens was the most favorable in that there was less reduction in airflow at the lowest
wind speeds compared to full deployment. Overall, the indicators of thermal
performance showed substantial variability with wind speed and direction but relatively
little variation with screen position.
Wind tunnel tests were run for a range of wind speeds coming from the West (normal to
the long side of the air-cooled condenser) with and without windscreens in place. Model
results show that the presence of the wind screen is evidently beneficial; with the wind
screen in deployed position there is less horizontal flow interfering with the vertical
flow into the fans and allows for more through the ACC fan.
The computational fluid dynamic models achieved good representation of field
conditions, but only at very low wind speeds, suggesting that the flow of the incoming
wind is not adequately represented in the model. Although quantitative results were not
obtained and the goal of producing a computational tool capable of generalizing test
results from the field or the wind tunnel was not achieved, some increased
understanding of the important physics was obtained to serve as a starting point for
additional modeling efforts.
This study was undertaken with the goal of increasing the understanding the
mechanisms by which wind screens helped, to quantify the beneficial effects and to
develop guidelines for the selection and design of wind screens.
Increased understanding of the physical mechanisms which determine the effect of
windscreens was obtained through a year of continuous field testing and extensive wind
tunnel modeling. However, the complete development of an analytical model which
was to have been used to generalize the test results was not achieved. As a result, the
goal of producing guidelines for windscreen selection and design could not be met.

1.2 Wind Effects

The adverse effect of wind on the performance of large ACCs has been recognized for
many years. Kröger (1998), in a chapter entitled “Effect of Wind on Air-Cooled Heat
Exchangers”, states the situation succinctly:
“In general, winds have a negative effect on the performance of mechanical-draft heat exchangers.
Plume air recirculation tends to increase while fan performance is usually reduced during windy
periods.”
Wind has been known to create significant problems for the thermal performance of
ACCs and to impose stresses on some mechanical elements, particularly the fans, the fan
blades, the fan motors and the gearboxes. Air-Cooled Condenser Design, Specification,
and Operation Guidelines (Wilbur & Maulbetsch 2005) notes that:
“The impact of ambient wind on ACC performance is not well understood by owner/operators or
their representatives in the specification and bid/evaluation process,” and that “this area of wind

4
effects in total represents the major challenge associated with ACC specification, design, and
performance.”
The effect of wind on ACC thermal performance is illustrated in Figure 1 which shows
the variation in turbine exhaust pressure vs. ambient temperature for a range of wind
speeds. Conditions such as these are typical of hot, arid desert locations in the
southwestern part of the US.

Figure 1: Effect of Wind on Turbine Exhaust Pressure

At ambient temperatures above 100F, the difference between no wind or low wind and
high winds (above 20 mph) can be 1.5 to 2. inches Hga with a correspondingly
significant effect on turbine output.

Physical damage resulting from wind is usually evidenced by incipient surface cracking
or occasional failure of the fan blades, by failures of gearboxes or by motor trips from
excessive vibration or current variations. Examples of fan blade damage are illustrated
in Figures 2 and 3.

Deterioration in thermal performance is attributed to two mechanisms: recirculation and

fan performance degradation.
Recirculation occurs when wind patterns around the ACC causes the hot exhaust air
from the top of the ACC to be blown down and re-entrained into the inlet air stream

5
resulting in an inlet air temperature higher than the far-field ambient air temperature.
This is illustrated in Figure 4.
Fan performance degradation, resulting in significantly reduced air flow into the ACC,
is illustrated in Figure 5, where a smoke plume is shown passing directly under a fan
without being entrained.
While both mechanisms contribute to degraded ACC thermal performance, the effect on
fan performance is the more significant in most cases. This is illustrated in Figure 6.

Figure 2: Surface Cracking on ACC Fan Blade

6
Figure 3

7
Figure 4: Hot Air Recirculation on an ACC

8
Figure 5: Degradation of Fan Performance and Inlet Air Flow Reduction

9
Partial Listing of ACCs with Wind Screens

Chapter 5: Field Test Results

Testing was conducted at the Caithness Power Plant that has an ACC that is arranged as
3 x 6 so it has 18 fan modules. The ACC was retrofitted with motorized windscreens in
an effort to reduce fan blade and U-bolt damage that had been occurring as well as
excessive fan motor trips. The ACC startup date was in 2009 and the retrofit occurred in
2012. It is believed that this is the only ACC in the U.S. that has motorized windscreens.
The wind screens were supplied by Galebreaker Industries. Field testing was conducted
in the 2014 and 2015-time period. Test instrumentation, setup and data acquisition
system was supplied by the Howden Fan Company. The air inlet of the ACC is 15

10
meters tall and the wind screen installation has an upper 2 meter fixed section and is
combined with a 5-meter motorized section. For the purposes of the testing program, the
fixed section was shortened to approximately 1 meter in height. Thus for the purposes of
the test program, a fully deployed screen has a height of 6 meters and a 50% deployed
screen has a height of 3.5 meters. Since the motorized screen rolls up from the bottom, a
fully retracted screen has a screen section greater than 1 meter in height with a rounded
edge.
Wind Screens Fully Retracted

11
 Wind Screens Fully Deployed

Figure 75 presents the average velocities in Cell 3.4 for the 5 screen positions. Figure 76
presents the same plot for Cell 2.4 and Figure 77, for the average velocities in both cells.
The most dramatic effects are apparent in Figure 75 for the upwind, edge cell (Cell 3.4).
For lower wind speeds of less than about 5.5 to 6.0 m/s (~ 12 to 13.5 mph), the
deployment of the wind screen at all levels reduces the average inlet velocity below the
value for the fully retracted case. For all wind speeds above 2.5 m/s, the average inlet
velocity for the two least deployed cases (0 percent and 25 percent) decreases
monotonically with increasing wind speed while the average inlet velocity for the more
deployed cases (50 percent, 75 percent and 100 percent) increases from wind speeds of
2.5 m/s up to wind speeds of about 5.5 to 6. m/s and then decreases slightly or levels off
for further increases in wind speed. At the highest wind speed of 8 m/s (~ 18 mph),
screens at 50 percent, 75 percent and 100 percent deployment show essentially identical
results of an approximately 9 percent increase in average inlet velocity over the retracted
case.

12
Figure 75: Cell 3.4 Average Inlet Velocity for Varying Screen Position

Figure 76: Cell 2.4 Average Inlet Velocity for Varying Screen Position

13
Figure 77: Cells 3.4 and 2.4 Average Inlet Velocity for Varying Screen Position

Figure 77 indicates that the combined effect on the two cells essentially balances out. At wind
speeds below 5 m/s (~ 11 mph), the screens reduce the average inlet velocity to the combined
cells slightly; above 5 m/s (~ 11 mph), the higher deployments show a significant increase in
comparison to the other two and they level out with wind speeds up to 8 m/s (~ 18 mph) at
approximately a 5.5 percent increase in average inlet velocity compared to the fully retracted
case. Again the 50 percent deployed condition gives the best performance on balance across the
entire speed range.
Two items are noteworthy. First, over the full range of wind speeds, the 50 percent case
exhibits the best overall performance. It produces the least reduction in average inlet
velocity in comparison to the fully retracted setting at the lower speeds and better
performance than the higher deployment settings over the entire range. Second, the
curve for the 25 percent deployment case exhibits significantly different characteristics
than the other cases. Unlike the 50 percent, 75 percent and 100 percent cases, the average
inlet velocity decreases monotonically over the entire speed range. In comparison to the
fully retracted case, the average inlet velocity is significantly lower at all speeds and the
shape of the curve vs. wind speed is essentially linear unlike the 0 percent case. No
reason for this behavior is known.

Some general observations can be made from comparisons of the previous three figures.

For the screens fully retracted:

• At the lowest wind speeds, the screen position has a much greater influence on the
airflow into the windward cell (Cell 3.4) than it does on the downwind cell (Cell 2.4). For
wind speeds up to about 4.5 m/s, the windward fan captures more incoming air than the

14
downwind fan. At the lowest wind speed (2.5 m/s), the average inlet velocity into the
windward cell is about 6.9 m/s; into the downwind cell, about 6.7 m/s.
• As the wind speed increases above 5 m/s, the incoming air increasingly bypasses the
windward cell and is captured by the downwind cell. When the wind speed reaches 7.5
to 8 m/s. the average velocity in the windward cell has fallen to about 5.8 m/s, while the
average downwind cell inlet velocity has remained essentially constant at about 6.6 to
6.7 m/s.

For the screens fully deployed:

• At the lower (2.5 to 5. m/s) wind speeds with the screens fully deployed, the airflow
into the windward cell is partially impeded and the average inlet velocity is as low as 5.9
m/s, well below that with the screen retracted. In the downstream cell, it is about 6.9 m/s,
well above that with the screen retracted, as incoming air is impeded from entering the
windward cell is captured by the downwind cell.
• As wind speed increases above 5 m/s, the blockage effect on the windward cell is
offset by reduced bypassing of the cell as the velocity of air passing through the screen is
reduced and the average inlet velocity into Cell 3.4 increases up to a wind speed of
about 6 m/s and then levels off. The average inlet velocity into the downwind cell is only
modestly affected by the deployment of the screen at the lowest wind speeds, but then,
as the wind speed increases, the airflow into the downwind cell increases up to the
highest wind speed of about 8 m/s.
For the two cells combined, the deployment of the screens to 50, 75 or 100 percent
provides an increasingly significant improvement in total airflow into the two cells
combined at wind speeds above about 6 m/s with an approximately 7 percent
improvement at the highest wind speed of 8 m/s. The trend of the curves at that point
suggests that the percentage improvement in total airflow would continue to increase at
still higher wind speeds.

5.7 Wind-induced Dynamic Blade Loading

As noted in the introductory discussion in Section 1, an important effect of wind on
ACCs is the imposition of wind-induced stress on the fan blades leading, in some
instances, to blade damage or failure. To study this effect, dynamic loading
measurements on 7 of the 9 blades on Fan 3.4 were recorded throughout the test period.
Two representative examples of these measurements are shown in Figures 100 and 101.
The following material is based on the Howden report (Holders 2015).
Figure 100 displays long-term, average values of the dynamic blade loading on Blade 3
of Fan 3.4 (Sensor 54L in Figure 49) vs. wind speed measured at the tank for the five
screen positons.

15
Figure 100: Long-Term Average Blade Loading vs. Wind Speed for 5 Screen Positions

The dynamic blade loading is taken as 3 times the standard deviation of the output
signal from load cell 54L. The high resolution (250 Hz) readings were averaged over 10
second (2,500 points) intervals and binned into 1 m/s ranges of wind speed as measured
on the water tank (Sensor 21A in Figure 49).
The results show a significant increase in blade loading with increasing wind speed
during periods when the screens are fully retracted or only 25 percent deployed. A 50
percent to 100 percent deployments, both the loading itself and the increase in loading
with increasing wind speed are significantly reduced. During periods when the screens
were fully deployed, the loading was reduced by a factor of between 2 and 3 at the
highest (7.5 to >8 m/s) wind speeds and the effect of wind speed over the range from 2.5
to > 8 m/s was essentially eliminated. At the lowest wind speeds, the lower screen
deployments (25 percent and 50 percent) appeared to increase the dynamic loading
slightly. The reason for this is not fully understood but minimal screen deployment may

16
introduce some flow disturbances at levels close to the fan inlet which are reduced or
eliminated as the screen is further deployed.
To understand the mechanism by which an increase in wind speed results in increased
dynamic blade loading and the effect of screen deployment on the relationship between
wind speed and blade loading, the high resolution data were examined in more detail. A
few short (180 second) periods during which significant changes in wind speed occurred
were selected. The following plots and discussion are excerpted from the Howden
topical report.

5.7.1 Blade Excitation Frequency

The 250 Hz load cell output from several of blades (Blades 1, 3, 5 and 7) were filtered to
extract the signal that matched the fan operating frequency (93 rpm = 1.55 Hz). Figure
101 shows the normalized signals from the 4 blades and the tachometer pulse which
occurs once per revolution. The pulses are separated by approximately 160 (≈ 250/1.55)
samples.

Figure 101: Filtered Sample Load Cell Outputs and Tachometer Pulse

The filtered data sets were shifted to align them all with the Blade 1 signal. The aligned peaks
were separated by the sample time corresponding to one revolution as determined by the “one per
revolution” tachometer pulse (Sensor 25 on Figure 49). The phase shift required to bring the
different blade data sets into alignment coincides with the angle between the sampled blades
confirming that the blades were excited once per revolution as shown in Figure 102.

17
Figure 102: Aligned Pulses and Tachometer Pulse

18
 Figure 103: Smoke Trace Under Fan

Figure 105: Inlet Air Velocity Variations

Detailed high resolution measurements of the variation in air inlet velocity over the Cell 3.4 fan
inlet area are shown in the upper panel in Figure 105 for the following locations for a three-
minute period on July 30, 2014 (3:18 pm to 3:22 pm).

19
Front of the fan inlet area: Sensors (28F + 31F)/2
• Back of the fan inlet area: Sensors (29F + 30F)/2
• Average over the fan inlet area: Sensors (28F through 35F)/8

During this period, the inlet air velocity is reasonably uniform (Vfront ≈ Vback ≈ Vaverage)
across the fan inlet area. The wind speed at the edge of Cell 3.4 is consistently below
about 7.5 m/s corresponding to a wind speed at the tank of less than 4 m/s. The dynamic
blade loading is low in the range of +/- .01.

5.7.3 Effect of Screen Deployment

The results discussed above have been for periods during which the wind screen was
fully retracted. The results in Figure 100 indicated that the increase in dynamic loading
with wind speed was reduced or eliminated as the wind screen was deployed. This is
attributed to the effect of the wind screen in reducing the non-uniformity in the air inlet
velocity. This effect is seen in Figure 107 which plots inlet velocity variation from front
to back of the fan inlet area for both fully retracted and fully deployed screens and also
in Figures 108 and 109 which was taken from a separate analysis analyzing the same
trend but plotting it against off-site wind speed. A reduction in non-uniformity is seen
with screen deployment.

20
Figure 105: Wind Speed, Inlet Velocity, and Blade Loading at Cell 3.4

Figure 106 illustrates, in the same format, a three-minute period (also on July 30, 2014
from 3:45 pm to 3:48 pm) during which (at approximately 3:47 pm) the wind speed
jumps to about 10 m/s at the edge of Cell 3.4 (corresponding to approximately 8.5 m/s at
the tank). The increase in wind speed corresponds precisely to an increase in non-
uniformity of the inlet air velocity and a significant (about threefold) increase in the
dynamic blade loading.

21
 Figure 106: Effect of Wind Speed Increase

The test results, as illustrated by the selected examples discussed above, demonstrate conclusively
that the wind screens essentially eliminated any increase in wind-induced blade loading when
deployed at Caithness up to wind speeds at the edge of the monitored cell up to over 10 m/s.

22
Figure 107: Effect of Screen Deployment on Inlet Velocity Variability

23
Figures 108 & 109 Effects of Screen Deployment Using Offsite Wind Measuremnts

24
6.1.3.2 Plant/ACC operating conditions
• All fans at full speed
• Constant steam unit output/constant ACC heat load
• Screens fully deployed or fully retracted

Extensive data searching throughout the entire 18-month test period yielded a number
of periods satisfying these criteria. Plots were constructed for these periods of inlet air
velocity for Cells 3.4 and 2.4 and static pressure in Cells 3.4, 2.4 and 1.4. A few example
plots are shown in Figures 122, 123, and 124. The plots are segregated into separate time
periods, designated by the small numbers above the abscissa (1 through 8), which are of
different duration and may be months apart but have similar characteristics.

Figure 122: Inlet Air Velocities in Cell 3.4 During Selected “Steady” Periods

25
Figure 123: Inlet Air Velocities in Cell 2.4 During Selected “Steady” Periods

26
Figure 124: Static Pressure in Cells 3.4, 2.4, and 1.4 During Selected “Steady” Periods

An important feature of these plots is that measurements of inlet velocities and static
pressures under comparable, steady conditions are very consistent in the different time
periods even though the periods may be months apart. This essentially confirms that the
site and wind characteristics which control ACC behavior have been correctly identified
and that the selection, placement and methodology of field data acquisition are apparently
correct and reliable.

27
6.1.4 Modeling of Field Test Conditions
Extensive wind tunnel testing was conducted on a model of the full-scale ACC at
Caithness. The wind tunnel, model, instrumentation and testing protocol were described
briefly in Section 4.3 and are presented in detail in Larson (2015) and Larson (et.al. 2015).
The following section presents the results and conclusions of the tests.

The overall conclusion from this set of 12 observations is that the screens reduce the air
speed flowing across the fan inlets which would be expected to improve fan
performance and result in higher airflow into the cells. At the wind speeds modeled this
is consistent with the averaged field results shown in Figures 75 through 77.
Measurements at the lowest wind speeds at which field tests indicated a reduction in
flow to Cell 3.4 with screen deployment could not be modeled in the wind tunnel.
Therefore, there is no physical model confirmation of that field observation. The primary
effects are the expansion of the region of reduced velocity behind the screen in the upper
region near the fan inlets and an extension of the accelerated high speed region near
ground level further downwind under Cell 2.4. The downwind region under Cell 1.4 is
affected very little as a function of increasing wind speed.

Figure 132: Velocity Measurement Under Row 4 Cells at 4. m/s; z = .67 (Retracted
vs. Deployed)

28
29
CHAPTER 7: Summary and Conclusions

7.1 Field tests

Measurements of fan inlet velocity, fan static pressure, fan motor current, cell inlet
temperature and blade loading were made on an 18-cell ACC equipped with retractable
wind screens. The screens could be fully retracted, fully deployed or set at any
intermediate position. Ambient wind speed and direction were monitored along with
plant operating conditions including unit output, steam flow, turbine exhaust pressure
and fan status (off, half speed, full speed for all 18 fans). Test measurements were made
at one minute intervals over the course of 18 months and presented as a function of
wind speed and direction.
The results, which are presented in detail in Section 5, can be summarized as follows:

7.1.1 Effect on Air Flow

The following discussion is based on a nominal “base case” of fully retracted screens
and very low (essentially still air) wind speeds.
For the windward cell (Cell 3.4):
• With the screens fully retracted, wind speeds above 3 to 4 m/s caused a drop-off in the
average inlet velocity from just less than 7 m/s to approximately 5.8 m/s at a wind speed
of over 8 m/s.
• With the windscreens fully deployed at the highest wind speed of over 8 m/s, the inlet
velocity was approximately 6.3 m/s, significantly increased above the fully retracted
case.
• At the lowest wind speeds, deployment of the screens reduced the inlet velocity from
just less than 7 m/s to less than 6 m/s.

7.1.1.2 For Cells 3.4 and 2.4 combined:

• The inlet velocities averaged for the two cells, showed a slight decrease below the base
case conditions at the lowest wind speeds and an increase at the highest speeds.

On the basis of measurements taken over two months of summer testing in 2014 and
averaged over all wind directions from northwest to southwest, the results suggest that
a 50 percent deployment of the screens was the most favorable in that:
• There was less reduction in airflow at the lowest wind speeds compared to full
deployment.
• The increase in airflow at the highest wind speeds experienced during the test period
was essentially the same as for the fully deployed screens and an improvement over the
fully retracted screen conditions.

30
Additional tests are being conducted at the site following the conclusion of this study.
Measurements at higher wind speeds than were ever experienced at Caithness should be
investigated to determine site characteristics for which a reduced screen size selection
might be recommended.
7.1.4 Blade loading
• Deployment of the screens resulted in a significant decrease in dynamic blade loading
under higher wind speeds.
• The reduction is related to the presence of the screens creating a much more uniform
inlet velocity profile for the air entering the fans.
• It is expected that the reduction in dynamic loading will have the effect of protecting
the fan blades against wind-induced damage and of prolonging their life.
• Although no quantitative measurement of damage or blade lifetime could be made as
part of this study, it is noted that no fan blade problems or failures have occurred since
the installation of the screens, which was the purpose for which they were installed.

7.4 Additional observations

7.4.1 Agreement with Observations Elsewhere
While the measurements made in the current study demonstrated no discernible
improvement in ACC thermal performance from the deployment of the screens, such
improvements have been reported elsewhere.
• Figure 11 and 14 of this report displays data showing clear performance enhancements
after the installation of solid wind barriers at WyGen (Figure 11) and a porous,
cruciform screen at El Dorado (Figure 14)
• Figures 23 and 25 show reported reductions in turbine exhaust pressure at Coryton
(Figure 23) and Kings Lynn (Figure 25), both plants in the UK.
• Results from the Mystic Plant in Massachusetts (Avyazain 2015) presented at the
recent Air Cooled Condenser Users Group meeting indicated a significant increase in
plant output as a result of the installation of wind screens.

Prior to the start of testing, plant staff stated that they had never observed wind effects
on plant performance of any consequence. It is believed that the absence of an observed
thermal performance improvement at Caithness may be the result of several factors.
1. Neither the wind speeds nor the ambient temperatures at Caithness are at levels
which would result in high turbine backpressures for the design plant heat load and the
size of the ACC.
2. At the highest turbine backpressures encountered during the test period, predictions
based on the properties of condensing steam and the turbine characteristics suggest that
a modest improvement in ACC thermal performance would not result in a significant
improvement in plant performance.

31
3. These conditions differ from those at other plants where high winds and high
summertime temperatures results in operating at much higher turbine backpressures
where effects of small changes on plant performance are to be expected. This was
certainly the case at WyGen, El Dorado and Mystic and perhaps so at Coryton and Kings
Lynn.

7.4.2 Fully Retracted Screens Vs. No Screens Installed

It should be recognized that the configuration with the screens “fully retracted” is not
the same as the configuration that would exist if no screens were installed at all.
Significant testing was performed prior to the decision to install the wind screens;
however, these tests were not part of this program. When the screens are fully retracted
the fabric rolls up and the length of screen still exposed when “fully” retracted extends
28 inches (0.7 m) below the bottom of the wind wall and support beam and blocks an
area that would be open if no screens existed. The computational modeling results
suggest that the dynamics of the flow approaching and flowing around the bottom of
the wind wall can have an important effect on the calculated results. However, the
importance of this difference could not be explored in the field tests.

7.4.3 Cleanliness of Screens

It was noted that airborne dirt and debris led to some degree of blockage on the screens
during the test period. This may have had some effect on the results, but this effect could
not be quantified.

32
Galebreaker Industrial Ltd Gary Mirsky
Galebreaker House O.E.M. & North and South American
New Mills Ind., Estate, Representative for Galebreaker
Herefordshire HR8 2SS, UK. cell: 908 -797-4515
tel: +44 (0)1531 637900 land line Florida office: 772-800-3844
fax: +44 (0)1531 637901 web: www.galebreaker.com