Basic Principles of Reverse Osmosis: Sergio G. Salinas-Rodriguez Jan Cornelis Schippers
Basic Principles of Reverse Osmosis: Sergio G. Salinas-Rodriguez Jan Cornelis Schippers
net/publication/351606373
CITATION READS
1 664
3 authors:
Maria D Kennedy
IHE Delft Institute for Water Education
212 PUBLICATIONS 5,026 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
A study of Pb, Cd & As removal from a drinking water source using pristine and modified Cocoa shell and Miscanthus biochars View project
Microbial desalination for low energy drinking water (MIDES) - H2020/685793 project View project
All content following this page was uploaded by Sergio G. Salinas-Rodriguez on 08 September 2021.
Basic principles of
reverse osmosis
Sergio G. Salinas-Rodríguez, Jan C. Schippers, Maria D. Kennedy
• Understand and be able to apply the basic principles of reverse osmosis, such
as: recovery, salt passage, salt rejection, concentration polarization, effect of
temperature, energy consumption.
2.1 INTRODUCTION
Reverse osmosis (RO) systems are capable of separating dissolved ions from a feed stream
based on salt diffusion mechanism. In RO systems, feed water is split into two streams: one
with a (very) low salinity and one with a high salinity. The low salinity stream is known as
permeate or product water while the high salinity stream is known as concentrate, brine, or
reject.
Figure 1 Basic schematic of a reverse osmosis system (Adapted from DuPont, 2020)
The quantity of water (Qw) flowing through a membrane is proportional to the differential
pressure feed-permeate (∆P), membrane surface area (A) and permeability of the membrane
(Kw). This relationship is expressed with the following equation:
dV
Qw = = (∆ P − ∆ π )·K w· A Eq. 2.1
dt
In membrane technology, flux is defined as the ratio between the permeate flow and surface
area of the membrane. It is expressed as:
Q 1 dV
J= w = Eq. 2.2
A A dt
Flux (J) is the permeate flow through a membrane surface area (Qw/A) (m3/m2.h or L/m2.h).
Then,
1 dV
J= =( P ) Kw Eq. 2.3
A dt
Answer:
a) 10 bar · 1 L/[Link] = 10 L/m2.h then 10 L/m2.h · 35 m2 = 350 L/h
b) 10 bar · 5 L/[Link] = 50 L/m2.h then 50 L/m2.h · 35 m2 = 1,750 L/h
In reverse osmosis systems the osmotic pressure is governed by the salinity of the feed water
and the recovery at which the RO system operates.
Osmotic pressure is the pressure which needs to be applied to a solution to prevent the
inward flow of water across a semipermeable membrane (Voet, et al., 2001). Osmotic
pressure is also defined as the minimum pressure needed to cancel out osmosis. Figure 2.2
illustrates the process of reverse osmosis in which the applied pressure needs to overcome
the osmotic pressure head to force water to pass through the semipermeable membrane.
Osmotic
pressure
Water
head
Salt water
Figure 2 Illustration of reverse osmosis principle. (Adopted from Salinas Rodriguez et al., 2015)
The osmotic pressure reduces the effect of hydraulic pressure; as a consequence, the effective
pressure or net driving pressure (NDP) is equal to the hydraulic pressure minus the osmotic
pressure.
where:
ΔP differential hydraulic pressure (pressure feed – pressure permeate) (bar)
Δπ differential osmotic pressure (osmotic pressure feed – osmotic pressure permeate) (bar)
In membrane filtration, the osmotic pressure hinders the water flow as illustrated in Figure 3.
Flux
Osmotic pressure
Feed pressure
Chapter 2 31
Table 1 Osmotic pressure for seawater and brackish water
Water type Salinity, mg/L Osmotic pressure, bar
Brackish water 1,000 ~0.7
Seawater 35,000 ~25
In practice a salinity of about 1,300 to 1,400 mg/L equals to 1 bar of osmotic pressure. From
this number, the osmotic pressure (in bar) can be estimated with a rule of thumb, as follows:
Where the salt concentration (C) is in milligrams per liter (mg/L). Hydranautics suggests that
an approximation for osmotic pressure may be made by assuming that 1,000 mg/L of total
dissolved solids (TDS) equals about 11 psi (0.76 bar) of osmotic pressure (Hydranautics,
2001).
On the other hand, the feed osmotic pressure can be calculated more accurately by using the
equation provided by ASTM 2000 based on the van’t Hoff equation:
Where:
πf osmotic pressure (kPa)
φ Osmotic coefficient
Estimates of osmotic coefficients for brackish and seawater are 0.93 and 0.90, respectively.
Tf temperature of feed stream (°C)
Σmi summation of molalities of all ionic and non-ionic constituents in the water
Membrane manufacturers make use of similar formulas, for instance DOW in its Filmtec
technical manual (Dupont, 2020) recommends the following equation for the feed water
osmotic pressure:
Where:
πf osmotic pressure (psi). 1 psi = 6.8948 kPa = 0.068948 bar
T temperature of water (°C)
Σmi summation of molalities of all ionic and non-ionic constituents in the water
Dupont uses the following simplified equations when calculating the average osmotic
pressure in the feed-concentrate stream as a function of salinity. For Cfc < 20,000 mg/L, the
osmotic pressure in bars:
Cfc (T + 320)
fc
= Eq. 2.8
491,000
Where:
T temperature (°C)
Cfc salt feed-concentrate concentration (mg/L)
Hydranautics recommends the following equation for the feedwater osmotic pressure
(Hydranautics, 2001):
fc
=1.19 (T +273) mi Eq. 2.10
Where:
πfc osmotic pressure (psi). 1 psi = 6.8948 kPa
T temperature of water (°C)
Σmi summation of molalities of all ionic and non-ionic constituents in the water
Theory suggests that the chemical nature of the membrane is such that it will absorb and
pass water preferentially to dissolved salts at the solid/liquid interface. This may occur by
weak chemical bonding of the water to the membrane surface or by dissolution of the water
within the membrane structure (Solution Diffusion Theory). The chemical and physical
nature of the membrane (e.g., surface charge and pore size) determines its ability to allow
the preferential transport of water over salt ions.
Under specific reference conditions, flux and rejection are intrinsic properties of membrane
performance (Dupont, 2020).
With:
∆P = Pf – Pp Eq. 2.12
And
∆π = πf – πp Eq. 2.13
Chapter 2 33
Where:
Jw water flux (L/m2.h)
ΔP hydraulic differential pressure (pressure feed - pressure permeate) (bar)
Δπ osmotic pressure difference between the feed water and product water (permeate) (bar)
Kw permeability constant for water (m3/[Link]) (L/[Link])
Answer:
Feed osmotic pressure = 35 × 0.7 = 24.5 bar
NDP = 25 - 24.5 = 0.5 bar ≈ 0.
Concentrate osmotic pressure is much higher than feed osmotic pressure.
Therefore, no production is expected.
Answers:
Flux: (60 - 25) bar · 0.8 L/[Link] = 28 L/m2.h
Permeate production: 28 L/m2/h · 35 m2 = 980 L/h
Answer:
The concentrate concentration is 44,000 × 1/(1-R) = 50,000 mg/L.
Now, we can calculate the average feed-concentrate concentration = 47,000 mg/L.
πavg = 36 bar (using the Dupont formula, Eq. 2.8)
The feed pressure will be: Pf = J / Kw + πavg
Pf = 25 / 1.0 + 36 = 25 bar + 36 bar = 61 bar
Where Cf is the salt concentration in the feed water and Cp is the salt concentration in the
product water.
Questions: Calculate the salt rRejection of the plant and calculate the salt passage of the
plant.
Answers:
SP = (Cp / Cf) × 100 %= (500 / 45,000) × 100 % = 1.1 %
SR = 100 % - SP = 98.9 %
Chapter 2 35
2.4 SALT FLOW
Water can pass a reverse osmosis membrane; salts as well, however, at a much lower rate.
The transport of salts through RO membranes is due to diffusion. Diffusion is a result of the
motion of ions in water and the tendency of salts to move from high concentration to low
concentration. Diffusion is a slow process, but cannot be neglected.
The salinity of the product water (Cp) depends on the relative rates of water and salt transport
through a membrane. This relationship is expressed by the following equation:
Qs
Cp = Eq. 2.18
Qw
Qs = ∆C · Ks · A Eq. 2.19
With
∆C = Cf – Cp Eq. 2.20
Where:
Qs Flow rate of salt through membrane (kg/s)
ΔC Salt concentration differential across membrane (kg/m3) = Cf - Cp
Ks Membrane permeability coefficient for salt (m3/m2.s)
A Membrane area (m2)
Cf Feed concentration (mg/L)
Cp Permeate concentration (mg/L)
Replacing terms,
Cp =
(C f )
– C p ·K s
Eq. 2.22
( P – P – ∆ π )·K
f p w
Dividing the whole equation by Cf, and rearranging the equation, we have the salt passage:
⎛ C ⎞
⎜C f – p ⎟·K s Eq. 2.23
Cp ⎜ C f ⎟⎠
⎝
=
Cf (
Pf – Pp – ∆ π ·K w )
Looking at the right side of the equation 2.24, since Cp is small compared to Cf, the ratio Cp
/ Cf is much smaller than 1, therefore the salt transport (Qs) is constant at a certain Cp and is
independent of the pressure. As a consequence, the salt passage (SP = Cp / Cf) is lower at high
pressure (Pf) and vice versa. This is because the same quantity of salt (Qs) will be diluted by a
larger volume of (product) water and vice versa.
Membrane
Cf = Cm
Cp
The salt flux (Js) is defined as the salt transport per membrane area per hour (mg/m2.h) (mg
per m2 per hour). The salt flux is proportional with the concentration difference between
membrane surface at: feed water side (Cf) and product water side (Cp), and the permeability
coefficient of membrane for salts (ions) Ks.
Remark: The larger the pores, the larger the permeability for salt and water. In general, Cp is
low in comparison with Cf , so Cp can be neglected in this formula.
Js ≈ Cf · Ks Eq. 2.26
Where:
Js salt flux (mg/m2.h)
Cf concentration at feed side membrane (mg/L)
Cp concentration at product side membrane (mg/L)
Ks Permeability for salt [(mg/m2.h) / (mg/L)]
Chapter 2 37
Example 6 – Salt passage and salt concentration permeate
The flux in an RO membrane is 10 L/m2/h. Salt concentration in feed water is 41,000
mg/L and Ks value is 0.08 [(mg/m2.h)/(mg/L)] = 0.08 L/m2/h.
Questions: What is the salt concentration in the permeate? What is the salt passage?
Answer:
In 1 hour, 10 L water is produced by 1 m2 membrane surface area. In 1 hour, the following
amount of salt will pass 1 m2 membrane surface area.
C perm = =
( )
J s C f – C p ·K s C f ·K s
≈
Jw Jw Jw
Cp 328
SP= ·100%= ·100%=0.8%
Cf 41,000
or
SP = 100 % – SP = 100 % – 0.8 % = 92 %
Answer:
Salt transport is the same (since the concentration of the feed water does not change) and
water transport (flux) is double, so Cp will be two times lower. Consequently, SP will be
two times lower.
J s C f ·K s
Cp = =
Jw Jw
Answers:
The flux will be 50% lower, since the capacity is reduced by 50 %.
Cp = (Cfc × Ks) / J
When we assume that Cfc will not change (because the recovery is constant), then Cp will
become two times higher
Since salt rejection (SR) depends on flux, Ks is not directly available from manufacturers
information. Thus, Ks has to be calculated from test results under standard conditions.
The permeate salinity can be chosen to a certain extend by the choice of:
• Type of membrane and manufacturer
• Recovery: lower recovery results in lower salinity
• Flux: higher flux gives lower salinity
For drinking water, 500 mg/L is usually the guideline. For industrial waters much lower
guidelines are often adopted, e.g., 10 to 50 mg/L. Usually a second pass is installed when
lower salinity is required. For instance, the product water seawater RO plants use to be
rather high.
Chapter 2 39
2.5 RECOVERY AND CONCENTRATION FACTOR
Recovery (R) is also known as conversion. Recovery is by definition the part of feed water
that is converted in product water and is expressed as percentage.
Qp
R= 100% Eq. 2.27
Qf
Where:
Qp Product water flow rate (m3/h)
Qf Feed water flow rate (m3/h)
Recovery affects salt passage and product flow. As recovery increases, the salt concentration
on the feed-concentrate side of the membrane increases, which increases the salt transport
and the permeate salinity (Cp).
High salt concentration in the feed-concentrate solution increases the osmotic pressure,
which consequently reduces the net driving pressure (NDP). As a result, the product water
flow rate is reduced and the permeate salinity (Cp) increased.
Qf Qc
Cf Cc
Qp
Cp
Qf = Qp + Qc Eq. 2.28
Qf · Cf = Qc · Cc + Qp · Cp Eq. 2.29
Cc Q f – Q p Q f Q p
= = – =1– R Eq. 2.31
Cf Qf Qf Qf
Cf = R · Cp + (1 – R) · Cc Eq. 2.32
And
C f R C f (1– SR) (1– R) Cc
= + Eq. 2.37
Cc Cf Cf
Then
Cc
(1– R)· C =1– R·(1– SR) Eq. 2.38
f
Then
C f ·⎡⎣1– R ·(1– SR)⎤⎦
Cc = Eq. 2.39
1– R
Or
⎡1– R ·(1– SR)⎤
⎣ ⎦
CF = Eq. 2.40
1– R
Since the salt passage is usually low, we may assume that the total amount of salt entering
the plant (Qfeed × Cfeed) will increase in concentration in the concentrate (brine) stream. As
a result:
1
CF = Eq. 2.41
1– R
Chapter 2 41
and
Cf
Cc = Eq. 2.42
1– R
In the concentrate, concentrations of salts are increased, including sparingly soluble
compounds. In seawater mainly calcium carbonate, while in brackish/fresh e.g., calcium
carbonate, calcium sulfate, strontium sulfate, barium sulfate and silica (SiO2).
As soon as the solubility is exceeded, precipitation / scaling might occur (scaling potential
and kinetics play a role). Results in lower membrane permeability (Kw). Precipitation of
calcium carbonate can be avoided by acid dosing and/or adding antiscalants. Supersaturation
to a certain extend is allowable by antiscalant dosing.
Answers:
Recovery: Qp / Qf = (200 / 500) × 100 % = 40 %
Concentrate flow: Qf - Qp = 500 – 200 = 300 m3/h
Salinity concentrate: CF × Cf = 1 / (1 - R) × Cf = 1 / (1 - 0.4) × 40,000 = 66,667 mg/L
Assuming salt rejection 100 %, the concentration factor for various recoveries is presented
in table 2.
Answer:
R = Qp / Qf or Qf = Qp / R
Qf = (1,000 / 3) / 0.45 = 740 m3/h per unit.
Total feed flow is 3 × 740 = 2,220 m3/h/L
Answers:
In seawater RO, recovery is usually 40 %.
Concentration factor: CF = 1/ (1-0.4) = 1.667
Salinity concentrate: Cc = 47,000 × CF = 78,300 mg/L
Feed osmotic pressure: πf = 47 × 0.7 = 32.9 bar
Concentrate osmotic pressure: πc = 78.3 × 0.7 = 54.8 bar
Pressure higher than 32.9 bar is required to exceed the osmotic pressure.
One stage.
Pressure drop in feed – concentrate channel can be calculated with the formula of Schock
and Miquel, (1987).
0.5·λ ·ρ ·v 2 ·L
∆ P= Eq. 2.43
dh
R ( v dh ) Eq. 2.45c
e=
Chapter 2 43
Or
0.5·ρ 0.7 ·v1.7 ·η 0.3 Eq. 2.46
∆ P=
d h0.7
Where:
ΔP = pressure drop across spacer λ = friction coefficient
ρ = density water v = velocity
L = length membrane dh = hydraulic diameter
The normalized pressure drop (NPD) in a feed – concentrate channel can be calculated with
the following formula.
Where:
∆Pact is actual pressure drop
CF(Q) is correction factor for flow
CF(T) is correction factor for temperature.
Where:
Qfc,ref average reference feed/concentrate flow
Qfc,act average actual feed/concentrate flow
ηT,ref viscosity at reference temperature
ηT,act viscosity at actual temperature
m = 1.4 (Hydranautics); or m = 1.7 (Schock and Miquel)
n = 0.34 (Hydranautics); or n = 0.3 (Schock and Miquel)
The procedure to determine the reference pressure drop, which equals normalized pressure
drop at startup, is the following:
1. Measure ∆Pstart at startup, preferably for each stage
2. Measure Qfc,start at startup, preferably per stage. Use this value as reference Qfc,ref
3. Choose a reference temperature e.g., 20 °C
4. Calculate ∆Pref with formula and ηT,ref and ηT,start
5. Use ∆Pref as reference pressure drop.
Under steady-state conditions, the concentration at the membrane surface (Cm) is constant.
The crossflow along the membrane surface enhances back diffusion of salts to the bulk. This
increase in salt concentration at the membrane surface is called concentration polarization.
As a result, the concentration at the membrane surface (Cm) is higher than in the bulk (or
feed water) (Cb).
This phenomenon results from: water flows through a membrane; salts (ions) are rejected;
retained salts (ions) accumulate at the membrane.
Cm
J·C J·Cp
Feed Flow
Cb
Figure 6 Concentration polarization - Concentration profile under steady state conditions (Mulder,
2003)
At steady-state the convective transport of salt to the membrane is equal to the sum of the
permeate salt flow plus the diffusive back transport of salt:
dC
J ·C p = D· + J ·C Eq. 2.49
dx
With the following boundary conditions:
When x = 0 then C = Cm
When x = δ then C = Cb
After integration:
⎛ C – C ⎞ J ·δ
ln= ⎜⎜ m p
⎟=
⎟ D Eq. 2.50
C
⎝ b – C p ⎠
Chapter 2 45
The ratio of the diffusion coefficient (D) and the thickness of the boundary layer (δ) is called
the mass transfer coefficient (k).
D
k= Eq. 2.52
δ
The intrinsic retention (Rint) of the membrane is the salt concentration at the membrane
minus the concentration of salt in permeate over the salt concentration at the membrane.
Cp
Rint =1− Eq. 2.53
Cm
Then, Cm / Cb becomes:
⎛J⎞
⎜ ⎟
Cm e ⎝k⎠
Eq. 2.54
=
Cb ⎛J⎞
⎜ ⎟
⎝k⎠
Rint +(1− Rint )·e
The ratio Cm / Cb is called the concentration polarization factor. This ratio increases (i.e.,
the concentration at the membrane (Cm) surfaces increases) with increasing flux (J), with
increasing retention (Rint) and with decreasing mass transfer coefficient (k).
When the salt is completely retained by the membrane (Rint = 1 and Cp = 0), then
⎛J⎞
Cm ⎜ ⎟
⎝k⎠
=e Eq. 2.55
Cb
This is the basic equation for concentration polarization which demonstrates the two
factors (the flux “J” and the mass transfer coefficient “k”) and their origin (membrane part
“J”, hydrodynamics “k”) responsible for concentration polarization.
The pure water flux (specific permeability) is determined by the membrane used and this
parameter is not subject to further change once the membrane has been selected. On the
other hand, the mass transfer coefficient depends strongly on the hydrodynamics of the
system and can therefore be varied and optimized.
Using the arithmetic average of feed and concentrate flow as average feed flow, the
concentration polarization factor can be expressed as a function of the permeate recovery
rate of a membrane element Ri.
⎛ 2·R ⎞
⎜⎜ i ⎟
⎟
⎝ 2−Ri ⎠
β = Kp· e Eq. 2.58
Chapter 2 47
Where:
β = CPF = concentration polarization factor
Kp constant depending on type (manufacturer) membrane (usually 0.99)
Qp permeate flow of an element (m3/h)
Qc concentrate flow in an element (m3/h)
Ri recovery of a membrane element
The value of the concentration polarization factor of 1.2, which is the recommended
Hydranautics limit, corresponds to 18 % permeate recovery for a 40 inches long membrane
element (Hydranautics, 2001). Equations 2.57 and 2.58 and frequently applied by Koch
and Hydranautics membranes manufacturers (Hydranautics, 2001).
Where:
β Concentration polarization factor
R Recovery
The recommended recovery (by Dupont) per RO element varies with the quality of the feed
water e.g.:
• Seawater 10-12 %
• Filtered treated domestic wastewater 10-12 %
• Pre-treated surface water 15-18 %
• Softened well water 19-25 %
In sea water RO systems, the concentration polarization will decrease with increasing
recovery. Consequently, the CPF is the last element is lower than in the first element. The
reason for that is that the flux is reducing dramatically with increasing recovery.
The mass transfer coefficient (k), is related to the Sherwood number (Sh).
k·d n
Sh= = a·Re b ·Sc c Eq. 2.60
D
Where: a, b and c are constants. With
ρ ·v·d n
Re= Eq. 2.61
η
And
η
Sc= Eq. 2.62
ρ ·D
Where:
k Mass transfer coefficient (m/s)
Re Reynolds number, dimensionless
Sc Schmidt number, dimensionless
dh Hydraulic diameter (m)
v Flow velocity (m/s)
ρ Water density (kg/m3)
η Dynamic viscosity (kg/m.s)
D Diffusion coefficient (m2/s)
Chapter 2 49
Correlations for mass transfer coefficients depend on physical characteristics of the system
and the flow conditions (e.g., laminar or turbulent). The mass transfer coefficient (k) is
mainly a function of the feed flow velocity (v), the diffusion coefficient of the solute (D),
the density and the module shape and dimensions. Of these parameters, flow velocity and
diffusion coefficient are the most important.
0.33
Re Sc d h
Channel Sh= 1.85 Sh = 0.04 · Re0.75 · Sc0.33
L
In the channel of the feed spacer of a spiral-wound RO element, Schock and Miquel (1987)
found that the mass transfer coefficient could be predicted by the following equation, when
calculations for the velocity in the channel and the hydraulic diameter took the presence of
the spacer into account:
D
k = 0.023· ·Re0.875 ·Sc0.25 Eq. 2.63
dn
Temperature has an effect on Kw. The higher is the water temperature the higher the
permeability will be. The change in permeability is about 3 % per °C. Kw is linked with the
viscosity of water.
When dealing with the membrane permeability, the correction will be as follows:
Where:
TCF temperature correction factor
t temperature in °C
Kwt membrane permeability at temperature “t”
Kw25 membrane permeability at 25 °C
The membrane permeability and the salt passage increase with temperature. Salt
permeability is connected with the diffusion of salt ions through the membrane. The
diffusion coefficient is defined with the following formula:
k B ·T
D= Eq. 2.66
6·π ·η ·r
Where:
kB Boltzmann constant
T absolute temperature 273 + T°C
η viscosity of water
r radius of ion
In the diffusion coefficient equation, the viscosity is a dominant factor. Viscosity will
decrease with temperature.
Chapter 2 51
A frequently applied formula for normalizing the salt permeability is:
Where:
Kst Salt permeability at temperature “t”
Ks25˚C Salt permeability at 25 °C
t Temperature in °C
The consequence of the previous equation is that the higher the temperature the higher the
salt passage will be. Similar to the effect of temperature on required pressure, the effect on
salt passage can be derived:
C pt1 C pt1
C pt 2 = = Eq. 2.68
TCF 1.03(t1−t 2)
Answer:
The salt permeability will increase with approximately 3 % per °C.
Since the flux (J) and Cf are constant, Cp will increase with approximately (25-5) = 20 °C
times 3 % per °C or about 60 %.
So, Cp will increase approximately with 60 %, from 5 mg/L to 8 mg/L.
Using the TFC we get the more accurate answer:
Cp t2 = Cp t1 / TCF = Cp t1 / 1.03(t1-t2) ; {Cp t1 = 5 mg/L and t = 5 °C and t = 25 °C}
Cp t2 = 5 mg/L / TCF = Cp t1 / 1.03(5 - 25)
Cp t2 = 5 mg/L / 0.55 = 9 mg/L
The permeate flux and the RO membrane salt rejection are important operational
performance parameters of a reverse osmosis system. Dupont (2020) summarizes the
parameters influencing the flux and salt rejection as follows:
• With increasing effective feed pressure, the permeate salinity will decrease (increased
salt rejection) while the permeate flux will increase.
• With increasing temperature (and all other parameters are kept constant), the permeate
flux and the salt passage (less salt rejection) will increase.
• With increasing recovery, the permeate flux will decrease and stop if the salt concentration
reaches a value where the osmotic pressure of the concentrate is as high as the applied
feed pressure. The salt rejection will reduce (more salt passage) with increasing recovery.
(a) (b)
Permeate flux
Permeate flux
Salt rejection
Salt rejection
Pressure Temperature
(c) (d)
Permeate flux
Permeate flux
Salt rejection
Salt rejection
Recovery Feed concentration
Figure 7 Impact of pressure (a), temperature (b), recovery (c), and feed concentration (d) on
reverse osmosis performance (Dupont, 2020)
To pressurize water cost energy. The theoretical minimum energy can be calculated with
the formula:
Taking into account the efficiency of the pump the formula changes into
0.0275 P
E=
Eq. 2.70
Npump
In RO and NF the recovery is less than 100%. As a consequence, the energy consumption
per m3 water produced will be higher according the formula:
0.0275 P
E=
Eq. 2.71
Npump R
Where:
E = energy consumption in kWh/m3
Chapter 2 53
P = pressure in bar
Npump = efficiency pump + motor
R = recovery
Answer:
E = (0.0275 · P) / (R · Npump)
Assuming: Npump = 70 %
We have: E = (0.0275 · 15) / (0.75 · 0.7) = 0.8 kWh/m3
Usually no energy recovery is applied for brackish water RO.
As a reference, in Table 4, the energy to just overcome the osmotic pressure seawater
is presented. At 50 % recovery, the theoretical minimum energy is about 1 kWh/m3. In
brackish water this energy is much lower.
In practice, energy recovery devices are applied to optimize the energy consumption in RO
plants.
Feedwater Product
Pump
Turbine
Discharge concentrate
The energy consumption can be reduced by e.g., recovering energy from the brine with a
turbine:
In practice the energy consumption equals: 0.5 – 1 kWh/m3 for brackish (without energy
recovery device); 3.0 – 4.0 kWh/m3 for sea (with energy recovery device).
T Brine flow
Chapter 2 55
Example 19 – Energy consumption in SWRO with pressure exchanger energy recovery
Same information as in the previous example.
Recirculation
M pump
M
Booster
Brine flow
Membranes
High pressure
pump
Figure 9 Schematic of a seawater RO system with a pressure exchanger (PX) and image of a
pressure exchanger (Adopted from Energy Recovery, 2021)
Energy consumption represents together with the investment, the largest part of the cost of
seawater RO. Seawater RO has improved tremendously over the past 20-25 years. Energy
consumption has been reduced: from 8 kWh/m3 in 1980 to 4 kWh/m3 in 2000.
Since 2006 further reduction down to about 3 kWh/m3 was achieved, due to lowering the
feed pressure down to 60 bar and applying pressure exchangers.
According a commonly applied rule of thumb for brackish water: “When conversion has
to be higher than about 50 % a second stage (array) is used as well”. The number of vessels
in the next stage is about 50 % of the previous one. Because the ratio feed flow to permeate
flow at the entrance of the next stage is the same. In the second stage about 50 % is converted
in product. This brings the total conversion at about 75 %.
Modern membranes have a very high permeability. Consequently, the required pressure is
much lower than in the past. Head loss due to the spacer and increasing osmotic pressure
reduce the net driving pressure. As a result, the permeate flow in the last elements in a vessel
are substantially lower, which allows a higher recovery before the CPF arrives at 1.21.
Chapter 2 57
2.13 REFERENCES
ASTM D4516 - 00 (2000) Standard Practice for standardizing reverse osmosis performance data
ASTM, West Conshohocken.
DUPONT (2020) FILMTEC™ Reverse Osmosis Membranes Technical Manual. In: Water solutions
(ed), pp. 207.
Energy Recovery. (2021). PX Pressure exchanger [Online]. Available: [Link]
water/px-pressure-exchanger/ [Accessed 20 April 2021].
Hydranautics (2001) Terms and Equations of Reverse Osmosis.
Mulder M (2003) Basic Principles of Membrane Technology, Second edn Kluwer Academic, Dordrecht
/ Boston / London
Salinas Rodríguez, S. G., Schippers, J. C. & Kennedy, M. D. (2015). The process of reverse osmosis. In:
Burn, S. & Gray, S. (eds.) Efficient Desalination by Reverse Osmosis: A best practice guide to RO.
IWA Publishing.
Schock G, Miquel A (1987) Mass transfer and pressure loss in spiral-wound modules. Desalination
64: 339
Voet D, Voet JG, Pratt CW (2001) Fundamentals of Biochemistry (Rev. ed.), New York