Page 3 - Reverse_osmosis2016
P. 3
S. Casimiro et al. / Desalination and Water Treatment 61 (2017) 183–195 185
π = π (1 − R ) (9)
p f j
Permeate flow through the RO membrane can be
expressed more completely by rearranging Eq. (1) taking
into account the effect of the permeate osmotic pressure,
average pressure drop in the RO vessel, permeate pressure,
and fouling factor Eq. (1) can be rewritten as follows:
P
Q = ( )( )(TCF FF P − cd − P − π + π
A S
)( )
p f 2 p ave p (10)
The fouling factor is applied to simulate aging and loss
of membrane permeability due to compaction and scale
fouling. Typically a fouling factor of 1 is applied to new
membrane, and a fouling factor between 0.65–0.85 for three
year old membranes and onwards. Also, because the per-
formance of the RO membranes is typically tested at 25 C,
o
a temperature correction factor (TCF) is considered were
suitable in the equations above to adjust the temperature
differences when running simulations assuming other feed-
water temperatures. TCF is determined using the following
equations [5]:
1 1
TCF = EXP 2640 − ; T ≥ 25 C (11)
°
+
298 273 T
1 1 Fig. 2. Energy recovery device.
°
TCF = EXP 3020 − ; T ≤ 25 C (12)
+
298 273 T
to the feed, Fig. 2), and three main pumps: a low pressure
3. ROSA validation pump (1 bar) between the intake and pre-treatment filters,
a high pressure pump (60 bar) forcing the water through
The reverse osmosis plant data used in the validation the membranes, and the energy recovering pump (56 bar).
procedure were provided by the plant operators of a desali- It also has a post-treatment system and a reservoir for pro-
nation plant in the southern city of Alvor, Algarve, Portugal duced water of 1000 m . A high efficiency energy recovery
3
on October 10 2014. The plant has a water production of system is used, it recovers energy by transferring most of
th
around 800 m d , and the data referred to nominal opera- the remaining pressure contained in concentrate stream
3
–1
tion of the plant. to a portion of the total feed water mass flow. This system
allows the mixing of a small amount of brine water (5–10%)
with the supply water, which can compensate for minimum
3.1. Plant configurations
required salinity to run the membranes to produce the tar-
The plant is composed of a pre-treatment system, 54 geted permeate quality.
semipermeable membranes (9 pressure vessels with 6 As part of this study, the reverse osmosis system anal-
membranes each) as seen in Fig. 1, an energy recovery sys- ysis (ROSA) model, a product of DOW FILMTEC , is
TM
tem (based on pressure exchange from the brine directly validated against another manufacturer’s desalination
membranes (Toray). The objective was threefold: 1) learn
about the design and operation of a RO plant and related
software using real plant data as reference, 2) describe in
the literature one of the few seawater RO plants operating
in Portugal, 3) confirm that ROSA can be used to simulate
the operation of a plant using membranes from other man-
ufacturers with seawater instead of brackish water (similar
research can be found in the literature [6], but it compares
the performance of these software with real RO plant data
making use of brackish water).
The Alvor plant uses Toray TM820C-400 membranes.
These are high rejection seawater membranes, with an
area of 37 m per element that according to the manufac-
2
turer maximize productivity and enable predictive system
design. The current membranes being used at Alvor were
replaced in a successive manner, one by one, starting from
Fig. 1. Alvor plant membrane assembly. the first maintenance operation carried out, and by May
