POWER May 2012 - 72

WATER MANAGEMENT
treatment systems if makeup pretreatment
or any form of ZLD is used.
Membrane filtration (microfiltration,
ultrafiltration, or nanofiltration) should be
used any time plant makeup water is provided
by either surface or recycled water.
These waters typically contain large concentrations
of extremely small particles
that do not coagulate well. Traditional
multimedia filtration cannot consistently
remove these particles. Membrane filtration
can remove these particles and protect
downstream equipment. Potable and well
waters typically do not require membrane
filtration.
Influent demineralization, which is cost
prohibitive for most fossil plants, may
make economic sense for CSP plants with
an evaporation pond. Demineralizers are
generally reliable and relatively simple to
operate. They remove all cations and anions,
unlike simple softening with ion exchange
or lime. The waste stream produced
by demineralizers is concentrated and can
be sent directly to the evaporation pond.
Demineralizing half of the incoming water
would lower the concentration of all ions
by half, thus doubling allowable cooling
tower cycles of concentration and increasing
allowable RO recovery. A detailed economic
analysis should be performed, but
up-front demineralization usually provides
an excellent option for air-cooled plants.
Most of the water used in air-cooled plants
is for steam cycle makeup and mirror
washing, so the majority of water required
has to be demineralized in any case.
Always try to avoid complicated serial
water treatment processes. Lime softening
followed by RO followed by demineralization
may make sense in a fossil plant,
but it may be more trouble than it's worth
in a CSP plant. CSP plant water usage is
generally lower than in a similarly sized
fossil plant (because dispatch is lower), so
operating cost doesn't increase as much as
might be expected when simpler but lessefficient
processes are used. Simpler designs
normally have a lower capital cost.
New and emerging technologies often
show great promise with significant potential
savings in water and dollars, but the
process risk can be extremely high. Though
new approaches may provide benefit, any
design that incorporates new technology
should include redundancy or contingency
to mitigate the higher process risk.
Balance Water Quality
and Availability
Solar thermal plant locations exhibit remarkable
commonality. They're typically located
in extremely hot and arid environments far
72
from urban areas and other infrastructure.
Water is universally scarce. Groundwater
and/or surface water may be available, but
neither may available in sufficient quantity to
meet plant water consumption needs. Water
quality data is often nonexistent, particularly
its potential for scale and corrosion. Often,
the only option for water is drilling new water
wells. Water sources and supplies may
have to be mixed and matched.
New groundwater supplies must be
sampled early and often. At least one and
preferably multiple test wells should be
drilled and sampled monthly for at least a
year. Too many water systems have been
designed based on a single unrepresentative
sample only to result in deficiencies in
the water treatment plant operation. Further,
the aquifer must be modeled to determine
the potential for change over the life
of the plant.
For example, at one CSP plant location
groundwater quality was relatively good,
but other businesses held significant water
rights that had not been exercised in recent
years. Aquifer modeling determined that
groundwater quality would remain stable,
assuming that the current extraction rates
were maintained and the new CSP plant
demands were added, but water quality
would deteriorate significantly if other users
began to withdraw at their maximum
allowable rates. If such situations occur, a
plant can be designed for the current water
quality, but provisions must be made
to add additional treatment equipment in
future years if other users increase extraction.
That includes additional building
space, underground piping, evaporation
pond space, and so on.
Recycled water or degraded surface
water (usually agricultural run-off or
something similar) may also be available.
Though data may exist for the original
source water, data rarely exists for the
actual recycled water or degraded surface
water. Sampling becomes even more critical
in such cases because these waters often
exhibit significant seasonal variability.
The design water quality for each water
source should be determined by calculating
the average and standard deviation of
all sample results. The design water quality
becomes the average (mean) plus two
standard deviations for each constituent.
This approach provides the additional design
margin necessary to account for historical
variability.
In addition, water chemistry must be
modeled for each potential water source
and for mixtures if the plant intends to
use more than one water supply. Mineral
solubility and chemistry modeling must
be performed. As stated earlier, accurate
2. Calcite saturation at seven cycles of concentration in a cooling tower.
The blue columns indicate that no scale should form; red indicates a significant risk of scale
formation. When the same water is treated with a calcium carbonate dispersant, all of the bars
are blue. Source: WorleyParsons
100
90
80
70
60
50
40
30
20
10
8.25
8.5
7.75
8
130
7.5
120
110
F°
www.powermag.com
7.25
100
90
80
7
70
POWER | May 2012
pH
Operating range profile at 7.00 cycles

http://www.powermag.com

POWER May 2012

Table of Contents for the Digital Edition of POWER May 2012