You are on page 1of 28

REVERSE OSMOSIS DESALINATION PLANT DESIGN

TRAINING PROGRAMME

1 Introduction:

1.1 Osmosis:
Osmosis is a natural flow process. In this process, there is a spontaneous flow of pure water into an
aqueous salt solution or from a less to a more concentrated aqueous salt solution across a
membrane. This membrane is referred to as semipermeable membrane by convention. This is
because it allows the passage of water molecules through it, but not the salt molecules / ions, which
are larger in size.

Every flow process requires a driving force, for e.g., difference in pressure. The flow resulting from
the difference in pressure is called viscous flow. Chemical potential is the driving force for osmosis,
i.e., it is essentially a diffusion process. The salt concentration is a measure of the chemical
potential.

Since the salt concentration is higher on one side of the membrane and the membrane permits only
the water molecules to pass through it, the water molecules flow from the less to more concentrated
solution side and dilute the more concentrated solution. This results in increase in water level or
water pressure on the concentrated solution side. This difference in pressure causes the water to
flow back to the low concentration solution side, i.e., viscous flow occurs. Initially, the osmotic
flow is higher and pressure builds up on the concentrated side. At a particular pressure, equilibrium
is reached and both the osmotic & viscous flows balance each other. This pressure, in the specific
case of pure water alone being present on one side of the membrane and salt solution on the other
side, is called the osmotic pressure. This will vary with salt concentration and is dependent only on
the salt solution properties and not on the membrane properties. Osmotic pressure is determined as
follows:

Osmotic pressure in bar, Po = c.R.T (1)


Where,c = sum of ionic molar concentration of all ions in solution in mole/litre,
R = gas constant = 0.083 litre.bar/Kelvin.mole and
T = Solution temperature in Kelvin.

For typical seawater with 35,000-mg/L salt concentration, the ionic salt concentration is 1.1 mol/L,
the osmotic pressure at 27oC, i.e., 300 K, works out to 27.4 bar.

1
Osmosis is an important biological process. Drinking salty water kills a man and the reason is
osmosis. When salty water is drunk, the stomach wall acts as a semipermeable membrane and
draws in water from the surrounding tissues through osmosis. Eventually, one dehydrates & dies.

1.2 Reverse osmosis:


The term reverse osmosis is actually a misnomer. When a pressure greater than the osmotic
pressure is applied on the concentrated solution side of the membrane, the viscous flow due to the
pressure differential across the membrane predominates and water flows from the concentrated
solution side to dilute solution side across the membrane. That is why, some prefer to call reverse
osmosis as hyperfiltration. Some are of the opinion that diffusion also plays a part in this process
apart from the viscous flow. It is said that the chemical nature of the membrane is such that it
preferentially adsorbs and passes water molecules compared to salt molecules / ions.

Whatever be the mechanism, the pressure differential across the membrane is the driving force for
the transport of water across the membrane. The side where the salty water is present is called the
feed side and the side where water with lower salt content is collected is called the permeate side.
The equation for water transport across the membrane is:
Water flow rate in m3/h, Qp = kw.A.{(Pf Po,f) (Pp Po,p)} (2)
Where,kw = Membrane flow factor for water in m/bar.h,
A = Membrane area in m2,
Pf, Pp = Pressure in bar on the feed & permeate sides respectively and
Po,f, Po,p = Osmotic pressure in bar on the feed & permeate sides respectively.

It should also be noted that there is no ideal membrane, which does not pass salt. The salt transport
across the membrane is purely a diffusion process, i.e., it is dependent on the salt concentration
differential and independent of the operating pressure. The equation for salt transport across the
membrane is:
Salt passage rate in kg/h, Ws = ks.A.(Cf Cp) (3)
Where,ks = Membrane flow factor for salt in m/h,
A = Membrane area in m2 and
Cf, Cp = Salt concentration in kg/m3 on the feed & permeate sides respectively.

The membrane flow factor for water depends on the membrane permeability for water and
membrane thickness. Similarly, the membrane flow factor for salt depends on the membrane
diffusivity for salt and membrane thickness. The membrane permeability for water and the

2
membrane diffusivity for salt are dependent on temperature apart from the membrane type. They
increase by about 3% per oC increase in temperature.

The salt passage percentage and permissible specific water flux are the two parameters that define
the performance of a membrane. The salt passage percentage is determined as follows:
Salt passage %, Ps = (Cp/Cf).100 (4)
Though the term salt rejection is frequently used to specify the membrane performance, it is better
to use the salt passage %. This is because, there does not seem to be much difference between 99 &
98% salt rejection percentages, but the corresponding salt passage percentages are 1 & 2%, i.e., a
doubling of salt passage, which affects the permeate quality very much.

The permissible specific water flux, i.e., permeate flow rate per unit membrane area, depends on the
type and construction of the membrane element and feed water quality. It is discussed in more detail
in a latter section. The permeate flow rate can also be related to the feed flow rate through permeate
recovery percentage and this is a useful parameter in designing RO processes. It is calculated as
follows:
Permeate recovery %. Rp = (Qp/Qf).100 (5)
Where,Qf = Feed water rate in m3/h.
Though, it may seem that there should be no limitations on permeate recovery %, in practice, the
permeate recovery % has to be limited to certain optimum value to obtain optimum membrane
performance. This is also discussed in a latter section.

1.3 Applications:
The major application for reverse osmosis is the production of potable water from brackish water or
seawater. A major industrial application is production of boiler feed water and high purity water for
pharmaceutical and electronic industry applications. In these cases, reverse osmosis is followed by
ion exchange process and this 2-stage process is very economical compared to the use of ion
exchange process alone. Concentration of fruit juices is also an area where reverse osmosis has
found some applications. Effluent treatment is a major application area at present. Since reverse
osmosis produces pure water as a product, which can be recycled, it is a very attractive process for
effluent treatment. The effluent stream is dewatered to produce a smaller quantity of concentrated
effluent stream, which can be treated more economically. Some typical examples are tertiary
treatment of sewage, treatment of textile dyeing industry effluent and treatment of fermentation
process effluent.

3
1.4 Seawater desalination:
Of the water resources available on earth, seawater forms 97% and fresh water only 3%. Even out
of this 3%, only 0.8% from rivers and underground sources is usable. With water requirements
increasing day-by-day, the only alternative available is seawater desalination. The current world
desalination capacity is about 20 million m3/day.

The two major processes used for seawater desalination are the thermal and reverse osmosis
processes. The thermal processes include multiple stage flash evaporation and multiple effect
evaporation. These processes use thermal energy to evaporate water into steam, which is condensed
to obtain pure water. Higher thermal efficiencies are achieved by carrying out multi-stage
operations. Reverse osmosis has entered the field of sea water desalination only recently, but it has
gained ground rapidly and has become the preferred process on account of the lower cost of potable
water produced.

A conventional single stage evaporation process uses 620 kWh of thermal energy to produce 1 m3 of
potable water. Multiple stage flash evaporation requires only 100 kWh of thermal energy and 3.5
kWh of electrical energy for pumps per m3 of potable water. On the other hand, reverse osmosis
process requires only 30 40 kWh for producing 1 m3 of water containing 500 mg/L total dissolved
solids from seawater. However, with energy recovery systems being employed in the large systems,
the energy consumption goes down to 7 kWh per m3 of potable water.

In a recent Israeli tender (2002) requiring a target cost of USD 0.50 per m3 of product water, with no
restrictions put on technology to be used, all the bidders opted for the reverse osmosis process. The
tender was finalised for a product water cost of USD 0.527 per m 3 against the earlier lowest cost of
USD 0.65 per m3. The project cost was USD 250 million for a 100 million m 3 per year plant. This
plant is expected to be commissioned during 2005. Another RO seawater desalination plant of
110,000 m3/d of treated water capacity being constructed in Singapore is expected to cost S$ 200
million. The treated water cost is estimated to be $ 0.78/m3.

2 Treatment system process design:

A reverse osmosis (RO) treatment system comprises of:


Feed water intake system,
Pre-treatment system,
RO module,
Post-treatment system and
Product water storage.

4
The design of feed water intake system and product water storage will not be considered in this
programme. The input data required for the process design of a RO treatment system are the
product water (permeate) rate & quality, i.e., the output from the system, feed water analysis and
preferred product recovery ratio / percentage (permeate rate to feed rate %).

The first step in the process design of any RO process is obtaining the complete analysis of the feed
water to be used. The analysis should cover physical, chemical and microbiological parameters. The
compliance of the radioactive parameters with the applicable standards should also be checked in
cases where there is natural background radiation and/or the plant is to be located near a nuclear
plant. Based on this analysis, the pre-treatment of the feed water is decided upon to ensure a long
life for the RO membranes. The estimated feed water analysis after pre-treatment is also
determined.

Based on the analysis of the input water to the RO system and the preferred product recovery ratio /
percentage, the RO module is designed and the required flow configuration within the module is
determined. The recovery ratio / percentage is varied, if necessary, to obtain an optimum design.
Thus, the required feed water rate will be firmed up at this stage only. The quality of the output
water from the RO module, i.e., its estimated analysis, is also determined in this stage.

The pre-treatment system is designed now based on the required feed water rate. The post-treatment
system after the RO process is decided based on the output water analysis from the RO module and
the required product water quality. The process design of the post-treatment system is then carried
out, completing the process design of the complete RO system.

2.1 Feed water analysis:


A complete feed water analysis is carried out covering the following parameters.

Physical parameters:
1. pH
2. Temperature
3. Turbidity
4. Silt density index (SDI)
5. Total suspended solids (TSS)
6. Total dissolved solids (TDS)
7. Conductivity and
8. Langelier saturation index (LSI)

Chemical parameters:

5
A. Cations:
1. Calcium (Ca)
2. Magnesium (Mg)
3. Sodium (Na)
4. Potassium (K)
5. Ammonium (NH4)
6. Barium (Ba)
7. Strontium (Sr) and
8. Iron (Fe).

B. Anions:
1. Carbonate (CO3)
2. Bicarbonate (HCO3)
3. Sulphate (SO4)
4. Chloride (Cl)
5. Bromide (Br)
6. Fluoride (F)
7. Nitrate (NO3) and
8. Silica (SiO2).

C. Dissolved gases:
1. Carbon dioxide (CO2) and
2. Hydrogen sulphide (H2S).

Microbiological parameters:
Normally, saline waters like seawater are not conducive for microbiological growth. However, a
total plate count (aerobic microbial count) test is carried out to check for the microbiological
quality.

2.2 Pre-treatment:
The feed water to a RO module should conform to certain minimum requirements to ensure
minimum fouling of the membranes. Membrane fouling can occur as layers of deposition on the
membrane surface (cake fouling), a hardened layer on the membrane surface (scaling), particle
insertion into the membrane pores (pore blockage) and chemical attachment of particles to the
membrane (adsorption). Membrane fouling reduces the membrane capacity and thus the useful life
of the membranes.

6
The most essential requirement for feed water relates to suspended solids including colloidal
particles, whether inorganic, organic, biological or a combination. They can deposit on the
membrane surface, block the feed channels and thus foul the membrane. Colloidal silica fouling is
one of the most difficult-to-handle factors in a RO process. Colloidal particles formed by a
combination of iron, silica and organics also poses a major problem. TSS, turbidity and SDI all are
indicators of suspended solids present. TSS is usually a measure of particles greater than 0.45-
micron size only. Thus, it does not indicate the presence of colloidal particles, which are 0.01 to 0.1
micron in size. Turbidity gives a measure of colloidal particles present. But, SDI is the best
indicator, since it is a measure of decrease in flow rate across a 0.45 micron filter under 2-kg/cm 2
gauge pressure over a period of 15 or 5 minutes and thus directly indicates the fouling tendency of
the feed water.

The feed water turbidity should not exceed 1 NTU and its SDI should not exceed 4. For long-term
reliable operation of the RO module, the average feed water turbidity should be less than 0.5 NTU
and its SDI should be less than 2.5.

The other major quality parameter for the feed water is the saturation levels of sparingly soluble
salts. These salts, as their concentration increases in the concentrate stream of the RO module, can
precipitate on the membrane surface and cause fouling. The indicators for the saturation levels for
these salts under the prevailing conditions of temperature, pH and ionic strength are Langelier
Saturation Index (LSI) and saturation ratios or percentages.

The LSI indicates calcium carbonate saturation. It is calculated from the analysis of water. If it is
negative, it indicates that calcium carbonate will remain in solution and if it is positive it indicates
possible calcium carbonate precipitation. It is preferable that feed water LSI does not exceed 0.2.
If an inorganic descalant like sodium hexametaphosphate (SHMP) is added, LSI can go up to 0.5
and if an organic descalant is added, LSI can go up to 1.8.

Saturation ratio is the ratio of the actual concentration of a sparingly soluble salt in the water to its
theoretical solubility at the given temperature, pH and ionic strength. This is also expressed as a
percentage. Saturation ratio or % is used to measure the saturation of other sparingly soluble salts
like sulphates of calcium, barium & strontium and silica. This silica is reactive silica, i.e., it reacts
with the standard molybdate reagent, and is different from colloidal silica mentioned earlier. Other
potentially scale forming salts like calcium fluoride and phosphate present in feed water seldom
pose any problem.

The solubility limits for these sparingly soluble salts at 30oC are:

7
Calcium sulphate : 2,090 mg/L
Barium sulphate : 2 mg/L
Strontium sulphate : 114 mg/L
Silica (@ pH of 8) : 120 mg/L.

Normally saturation % should not exceed 100%. If effective descalants are used, saturation limits
for sparingly soluble salts in the concentrate stream of a typical RO module can go upto:
Calcium sulphate : 230%
Barium sulphate : 6,000%
Strontium sulphate : 800%
Silica : 100%.

Further, the feed water should be free of residual chlorine, iron, dissolved hydrogen sulphide gas
and microorganisms including bacteria. The soluble ferrous iron salts can get oxidised and get
precipitated as ferric iron salts on the membrane surface, fouling as well as staining the membrane.
Similarly hydrogen sulphide can also get oxidised to sulphur, which can foul the membrane. The
microorganisms will get filtered by the RO membrane. But they can get attached to the membrane
surface and grow on the surface resulting in biofouling.

2.2.1 Coarse filtration: Coarse filtration normally forms a part of the water-intake system. Bars or
racks are used to remove very large sized objects including biological debris. Screens in the form of
partially submerged rotating drums or travelling screens are used to remove coarse particles,
typically 1 mm or greater in size. Grit chambers are used to separate high-density particles like
sand, typically 0.25 mm or greater in size, by settling. A suitable marine growth inhibitor can also
be added in the water intake system to prevent the growth of marine organisms in the intake channel
and filters.

2.2.2 Disinfection/chlorination: If any microorganisms are present in the feed water, it is


disinfected, generally using chlorine. The chlorine can be used directly or in the form of sodium
hypochlorite solution. The effectiveness of disinfection is proportional to the disinfectant
concentration and contact time. The same level of disinfection can be achieved by a higher
concentration with lower contact time or lower concentration with higher contact time. Chlorine
also reacts with any ammonia or organic nitrogen present in the water. 1 mg/L of ammonia present
in water requires 8 10 mg/L of chlorine. Typical chlorine concentrations required for disinfection
vary from 5 to 25 mg/L and typical contact times vary from 15 to 45 minutes. The required
optimum chlorine dosage and contact time are determined based on laboratory trials. When the pre-

8
treatment process includes a flocculation stage, contact time available will be generally higher and
lower dosage of chlorine will be sufficient. Otherwise, higher dosage of chlorine will be required.

Chlorine also plays a part in the acceleration of the oxidation of soluble ferrous salts to insoluble
ferric salts, which precipitate and are removed during the subsequent filtration operation. A chlorine
dosage of about 1 mg/L is required to remove 1 mg/L of iron. Soluble ferrous salts are normally
present in subterranean waters only and not in surface waters like seawater. This is because
dissolved oxygen present in surface waters oxidises any ferrous salts present. Chlorine also keeps
the manganese green sand, used in the subsequent multi-media filtration stage, in an oxidised
environment. Manganese green sand is used as a backup to ensure that no iron passes onto the
membranes in the RO module. Otherwise, green sand will require more frequent regeneration with
potassium permanganate.

Chlorine also reacts with hydrogen sulphide gas, if present in feed water, forming sulphuric acid
and / or sulphur particles, depending on its pH. These sulphur particles can foul the membranes in
the RO module and hence it is to be ensured that they are filtered before the feed water enters the
RO module. A chlorine dosage of around 7 mg/L is required to remove 1 mg/L of hydrogen
sulphide when the water pH is around 7. However, hydrogen sulphide is normally not present in
seawater.

Liquid chlorine from 1-tonne cylinders is evaporated in chlorine evaporators, the chlorine gas is
injected into a stream of water in the injector and the hypochlorous acid solution formed is injected
into the feed water. Chlorinators of various capacities are available as standard units and hence their
design is not covered in this programme.

Sodium hypochlorite solution with 3 to 12% available chlorine is used, when the chlorine
requirements are less and its higher cost is less than the cost of handling 1-tonne chlorine cylinders.
Sodium hypochlorite solution can be added to the feed water through a metering pump at the
required dosage and well mixed with it using a static mixer or flash mixer.

The addition of chlorine gas or sodium hypochlorite changes the chemical composition, TDS, pH
and dissolved carbon dioxide of the feed water and the feed water composition has to be changed
accordingly. In case of chlorine gas, pH decreases and dissolved carbon dioxide & TDS increase. In
case of sodium hypochlorite, pH & TDS increase and dissolved carbon dioxide decreases. When
chlorine is used only for disinfection and other reacting compounds are not present in the feed
water, the change in pH will be only slight and increase in TDS will also be insignificant in the case
of seawater.

9
Other chemicals that may be required to be added at this stage like flocculants are also added in the
same static mixer or flash mixer. Standard static mixing units are available and hence their design is
not covered in this programme. The process design of the flash mixer is presented below.

A turbine agitator is used for flash mixing of various chemicals with the feed water. The agitator is
mounted on a vessel, where the mixing occurs. The capacity of the vessel depends on feed water
flow rate and the time required for mixing. Normally a residence time of 1 to 3 minutes is sufficient
for the flash mixing of chemicals with the feed water. When flocculants are also being added, the
residence time should be about 1 minute only so that the formed flocs do not break due to the
vigorous agitation. Thus, the volume of the mixing vessel is calculated as:
Mixing vessel volume in m3, Vm = Qf.tr/60 (6)
Where,Qf = Feed water flow rate in m3/h and
tr = Residence time in minutes.

The vessel dimensions are selected such that the diameter of the vessel is approximately equal to the
liquid height in the vessel. For a vertical cylindrical vessel with a dished bottom having 6% knuckle
radius, the vessel diameter, such that it is equal to liquid height, can be determined as follows:
Vessel diameter in m, Dv = (1.35 Vm)1/3 (7)
The nearest vessel diameter, which is a multiple of 0.1 m, is selected. The height of the cylindrical
portion of the vessel is taken to be 1.25 times the vessel diameter. In this case also the nearest
multiple of 0.1 m is selected.

A turbine type agitator with 4 nos. flat blades pitched at an angle of 45o is used. A turbine rotational
speed of 100 rpm is normally used, since it is a convenient rpm for the agitator drive assembly
considering the cost & maintenance factors. For this rpm, the turbine diameter is selected as a ratio
of the vessel diameter as given below, based on the residence time required, to give the optimum
mixing performance.
Turbine diameter in m, Dt = 0.33 Dv (for tr = 3 min)
= 0.40 Dv (for tr = 1 min) (8)
The turbine blade width is selected based on the turbine diameter as follows:
Blade width in m, wb = 0.18 Dt (9)

The turbine agitator is centrally located within the mixing vessel, with a clearance of (D v/3) m
between the blade bottom and vessel dish bottom. The motor power required to drive the agitator
assembly is calculated for the case of pitched-blade turbine in turbulent flow regime as follows:
Power in kW, P = 7.5 d.Dt5 (10)

10
Where,d = feed water density in kg/m3.
An electric motor with the next higher standard kW output rating is selected. Its rpm and the
reduction ratio of the reduction gear drive are selected such that the final agitator rpm is 100.
Geared motor of the required kW and 100 rpm can also be used. The direction of turbine rotation is
selected such that the liquid is pushed down by the turbine towards the vessel bottom.

The vessel is provided with 4 nos. vertical baffles along the full height of the vessel wall, spaced 90o
apart to provide proper mixing. These baffles are of Dv/12 m width and are placed with a clearance
of Dv/72 m away from the wall. The feed of water and chemicals to the mixing vessel is given in the
middle of the vessel above the turbine agitator. The output from the vessel is taken from the side of
the vessel just below the top liquid level.

2.2.3 Flocculation: Flocculation process is used to enable agglomeration of suspended colloidal or


very fine particles, which otherwise would not settle or be removed by sand or multi-media
filtration. This process also aids the separation of ferric salts and sulphur precipitated on treatment
with chlorine as described above. This process is also used along with chemical precipitation
processes like lime or lime-soda water softening processes, which are used to decrease the water
hardness. However, softening processes are not required for seawater feed.

Colloidal or very fine particles have large surface areas compared to their mass and hence do not
settle and remain as a suspension. In addition, these particles acquire negative electrostatic surface
charges generally (sometimes positive charges also) due to their large surface area, which cause
them to repel each other and resist adhering together. Coagulants or flocculants like alum, ferric
sulphate, ferric chloride and sodium aluminate neutralise the negative charges and provide a nucleus
for the suspended particles to adhere to. The particles now agglomerate and form flocs, which can
be settled. These flocs are still fairly small and light in density. Their settling can be accelerated by
adding certain flocculant / coagulant aids like polyelectrolytes. These are normally high molecular
weight, water-soluble polymers. These are available as cationic, anionic and non-ionic types. These
enable the formation of larger-sized flocs, which settle faster.

The pH of feed water is an important factor that affects flocculation. The flocculants and the
flocculant aids to be used, their dosages, optimum pH and the time required for flocculation are
determined by conducting jar tests. The pH of feed water is adjusted, if necessary, by adding an acid
or alkali. The optimum pH in the case of alum can vary from 5.5 8.0, with a pH of 6 7 being
preferred. The optimum pH in the case of iron salts is 4.0 11.0, with a pH of 6 7 again being
preferred. The addition of alum or iron salts decreases the water pH and increases TDS & dissolved

11
carbon dioxide and the addition of sodium aluminate increases the pH & TDS and decreases the
dissolved carbon dioxide. Chemical composition is also changed in all the cases. However, these
changes are insignificant in the case of seawater.

The various chemicals required for the flocculation process are also added along with sodium
hypochlorite solution, if used, to the feed water and mixed either using a static mixer or flash mixer
as described in the earlier section. All the chemicals are added in form of solutions in water,
normally 1% solutions. Metering pumps are used to add the chemical solutions at the required
dosages.

A flocculation tank is provided, if necessary, as determined by jar tests, to provide the necessary
residence time for the formation of flocs, which can be easily filtered. The normal residence time
required is 15 30 minutes. The tank is constructed as a series of chambers with downward and
upward flow in the alternate chambers. The tank can also be provided with a slow speed agitator to
promote flocculation.

Nowadays, an alternative to flocculation process, viz., dissolved air flotation process, to remove oil
& grease and suspended solids, is also being used as a better option.

2.2.4 Sand / multi-media filtration: Pressure sand filtration is the most common filtration process
used in water treatment to remove the suspended solids. A bed of granular solids loaded in a
pressure vessel is used to filter the suspended particles from water. The most common flow pattern
used is downflow, i.e., the water to be treated is fed to the top of the bed. When the suspended
particles get trapped in the bed, the pressure drop across the bed increases and when it reaches a
specified level, the bed is given a backwash, i.e., filtered water is fed to the bed from the bottom at a
higher flow rate such that the bed expands and all the trapped particles are dislodged and removed
along with the water being fed.

The bed can be made up of single medium or upto four media. In the latter case, it is called multi-
media filtration. Use of single medium does not give optimum performance as explained below. For
efficient filtration, graded medium, i.e., medium with different particle size ranges, should be used
with larger particles at the top of the bed and smaller particles at the bottom. This will prevent the
top portion of the bed from getting clogged faster. However, during backwashing, this arrangement
of particles in the bed gets disturbed with the larger particles of the medium tending to settle at the
bottom of the bed. Hence it is preferable to use multi-media filtration with media of different
densities, the denser medium with smaller particle size range being loaded at the bottom of the bed
and progressively lighter medium with progressively larger particle size range being loaded at the

12
top of the bed. This ensures that the size gradation of the bed at different levels does not get
disturbed during backwashing.

The most common medium used is sand and it is used in the single medium beds. The other media
used in the multi-media filtration are anthracite and ilmenite or garnet. The density of sand falls
between these two media, with anthracite being the most light. Polymer beads, which are lighter
than anthracite are also being used, but mostly in proprietary filters using the upflow principle.
Green sand, which is based on manganese dioxide, is also used, especially when it is required to
remove the remaining iron present in the feed water. Sand and anthracite are used in the dual media
bed and ilmenite or garnet, sand and anthracite are used in tri-media filter. In the case of quad-
media filter, 2 layers of anthracite of different particle size ranges are used at the top of the bed. In
all the cases a bed of graded gravel is used at the bottom to support the filter media. It is made up of
2 to 4 particle size ranges, with the larger particles at the bottom, is used at the bottom, covering the
under-drain system, which takes away the filtered water and distributes the backwash water.

The inner diameter of the pressure sand filter is determined based on the flow rate of feed water per
m2 cross section of the bed. The normal rate lies between 4.8 24 m 3/m2.h, the typical value being
12 m3/m2.h. The gravel support medium is usually made up of particles in the size range 1 25 mm
with a height of about 0.3 m. In the case of a dual bed, sand of 0.4 0.8 mm particle size range is
filled upto a bed height of 0.3 m at the bottom above the gravel bed and then anthracite of 0.8 2.0
mm particle size range is filled upto a bed height of 0.45 m, thus providing a total filter media
height of 0.75 m. In the case of tri-media filter, the three layers above the gravel layer from the
bottom are ilmenite or garnet of 0.2 0.6 mm particle size range to a height of 0.1 m, sand of 0.4
0.8 mm particle size range to a height of 0.25 m and anthracite of 1 2 mm particle size range to a
height of 0.4 m, the total filter bed height being 0.75 m again.

When the pressure drop across the bed builds up to the specified level, the bed is given a backwash
for 2 minutes using filtered water. The backwash rate used is 48 72 m 3/m2.h. Normally two or
more pressure sand filters are used in parallel so that the production of treated water does not
completely stop during backwashing.

2.2.5 Activated carbon adsorption: Activated carbon adsorption, also called as activated carbon
filtration, is used to remove organics and chlorine from the feed water. Activated carbon reacts with
chlorine forming hydrochloric acid and carbon dioxide, about 1 g of carbon reacting with 10 g of
chlorine. Whereas, it only physically adsorbs the organics and hence can be regenerated in a furnace
by oxidising the organic matter and thus removing it from the carbon surface. Activated carbon

13
adsorption is also carried out in pressure vessels used for pressure filtration, with a bed of activated
carbon granules being used over the gravel support bed instead of the beds of filter media. Provision
for backwash is also given to prevent undue pressure build-up due to the accumulation of carbon
fines and other particles.

The residence time required in the case of organics removal is much higher than in the case of
chlorine removal. Flow rate and bed height similar to those used in the pressure sand filtration, i.e.,
a flow rate of about 12 m3/m2.h and a bed height of 0.75 m, can be used for activated carbon
dechlorination also. However, larger beds are required for organic removal, for a given flow rate,
since the residence time required will be around 15 minutes. Hence, pressure vessels of larger
diameter and activated carbon bed heights upto 3 m are used for this purpose. The residence time
required for organics removal is best estimated from pilot plant studies.

Activated carbon granules used should be sufficiently hard so that they do not generate too much
fines during normal operation. They are available in about four particle size ranges from 0.5 to 4.75
mm. Their bulk density is around 500 kg/m3 and they have a surface area of 950 1100 m2/g and an
iodine absorption value of 900 1050 mg/g. Normally two activated carbon adsorbers are used in
series, with the first one being removed from service when it has got exhausted and then used as the
second adsorber in series after regeneration of spent carbon and/or make-up with fresh activated
carbon.

In case dechlorination is the only operation required and not removal of organics, it may be
economical to dechlorinate using chemicals like sulphur dioxide, sodium sulphite or sodium
metabisulphite. Where chlorine is used directly for chlorination, it may be preferable to use sulphur
dioxide for dechlorination since same type of equipment as chlorinators can be used for this
purpose. A sulphur dioxide dosage of 2 mg/L will be required to remove 1 mg/L of chlorine. Where
sodium hypochlorite solution is used for chlorination, it may be preferable to use sodium sulphite or
metabisulphite solution for dechlorination. These solutions are added through a metering pump to a
static or flash mixer, where other chemicals like acid for pH adjustment and descalants are also
added. A dosage of about 4 mg/L of sodium sulphite / metabisulphite is required to remove 1 mg/L
of chlorine.

2.2.6 pH adjustment: If LSI of the feed water after the above pre-treatment process is positive
indicating its tendency for calcium carbonate precipitation, sulphuric or hydrochloric acid can be
added to decrease the pH and thus, LSI. The acid is added through a metering pump to the static or
flash mixer along with other chemicals to be added at this stage.

14
2.2.7 Antiscalant addition: If LSI and saturation percentages indicate the possibility of scaling, the
addition of antiscalants will prevent scaling of membranes in the RO module. The use of
antiscalants enable the use of a feed water with slightly higher LSI and saturation percentages as
indicated in the earlier section. The antiscalant solution is also added to the static or flash mixer
through a metering pump as indicated above.

2.2.8 Cartridge filtration: The feed water is filtered through a 5-micron cartridge filter to ensure
the removal of any remaining suspended or extraneous particles (from the vessels & pipelines) that
may be present in the feed water and thus protect the membranes in the RO module. The cartridge
filters are sized based on their flow rate pressure drop characteristics provided by their vendors.

2.2.9 Microfiltration & Ultrafiltration: The recent developments in the pre-treatment equipment
include backwashable capillary tube microfiltration & ultrafiltration membrane modules. These
systems can replace practically the entire pre-treatment system described above. They can operate
reliably at very high recovery rates and low feed pressure and provide better quality feed water.
Their cost is still high, though they are the only solution in case of certain feed waters.

2.3 Reverse osmosis:


The first step in the process design of the reverse osmosis module is the selection of the membrane
element, the pressure vessel to be used and number of elements per pressure vessel and the product
recovery ratio / percentage, which is nothing but the permeate recovery % for the RO module as a
whole, consisting of several membrane elements. Then a preliminary selection of the number of
membrane elements required, number of pressure vessels required, the flow configuration, etc. are
selected. Standard software programs are available for the process design of the RO system. A
preliminary design is carried out based on the preliminary data used. If the value of any parameter
falls outside the design limits in the preliminary design, the design process is repeated by changing
the necessary parameters. This process is continued till a valid design is obtained.

2.3.1 Membrane element: Commonly used RO membranes are cellulose acetate and thin film
composite (TFC) polyamide membranes. Cellulose acetate membranes have smooth surface and
little surface charge and hence are less prone to fouling and are more tolerant to chlorine. But, the
TFC polyamide membrane can operate at higher specific water flux and higher salt rejection at
lower operating pressures and is stable over a wider pH, pressure and temperature ranges and hence
is the first choice for most of the applications including seawater desalination.

The two most common membrane element constructions are the spiral wound & tubular. The
former scores on cost, packing density and operating pressure capability. The latter scores on

15
fouling resistance and cleanabilty. It has been found that a good pre-treatment plus spiral wound RO
membrane element is the optimum cost effective combination for the RO process. The special
advantages of the tubular construction make it ideally suitable for the pre-treatment of feed water,
where operating pressure is less and product recovery is high, like microfiltration & ultrafiltration.
Both the types of membrane construction employ cross flow pattern, i.e., feed / concentrate flows
parallel to the membrane surface and the permeate flows perpendicular to the membrane surface.
The main advantage for this flow pattern is that any potential scale forming or fouling substances
are swept away by the feed / concentrate flow and thus prevented from settling on the membrane
surface.

TFC polyamide spiral wound membrane elements, specially designed for seawater desalination, are
available. Their typical salt rejection percentages are above 99%, i.e., less than 1% salt passage
percentage. Depending on the required product (treated water) quality, the membrane element with
the required salt passage percentage is selected. These elements are normally available in 63.5 mm
(2.5), 101.6 mm (4) and 203.2 mm (8) diameters and 1016 mm length. For large capacity
applications like seawater desalination, the 203.2 mm diameter membrane element is commonly
used. Its design permeate flow capacity, Q p,d, is around 0.8 0.9 m3/h. This is determined based on
the permissible specific water flux and membrane area. The permissible specific water flux will
vary with the feed water quality. It ranges from 0.015 m 3/m2.h for a feed water SDI of 4 to 0.05
m3/m2.h for a feed water SDI < 1.

A preliminary selection of the number of membrane elements required for a given product flow rate
is made based on the design permeate flow capacity.
Number of membrane elements, Nm = Qp/Qp,d (11)

Normally six membrane elements can be accommodated in the standard pressure vessels used by
the RO industry and thus, a preliminary estimate for the number of pressure vessels required can be
estimated.
Number of pressure vessels, Npv = Nm/6 (12)
The number of pressure vessels is taken to be the next higher integer to the value obtained above.
The pressure vessels are normally of FRP construction.

2.3.2 Design parameters: As discussed at the beginning, the process of reverse osmosis is
governed by the three equations for osmotic pressure, water flow rate and salt transport rate
respectively. They are:

Osmotic pressure in bar, Po = c.R.T (1)

16
Water flow rate in m3/h, Qp = kw.A.{(Pf Po,f) (Pp Po,p)} (2)
Salt passage rate in kg/h, Ws = ks.A.(Cf Cp) (3)
It can be observed from these three equations that, once the membrane element is selected, the
parameters affecting the reverse osmosis process are feed water salinity, i.e., salt concentration, feed
water pressure and feed water temperature. Temperature affects the membrane flow factors, k w and
ks, also and hence it has an effect on water flow rate & salt passage rate also.

Feed water salinity: The osmotic pressure is directly proportional to feed water salinity (equation
(1)). Hence, higher the salinity, higher the feed water pressure required. As the feed water flows on
the feed-side of the membrane element, its salinity progressively increases since the rate of water
flow across the membrane is much higher than the rate of salt passage across the membrane. Thus,
the osmotic pressure at the end of the membrane element, where the feed water comes out as the
concentrate, is higher than that at the beginning.

When six membrane elements are connected in series in a pressure vessel, the osmotic pressure
corresponding to the salinity of the concentrate coming out of the sixth element will be still higher.
When the concentrate from the first pressure vessel is fed to another pressure vessel in series, the
osmotic pressure corresponding to the concentrate coming out of the second pressure vessel will be
even higher. Thus, the feed water pressure should be higher than the maximum osmotic pressure
corresponding to the salinity of the final concentrate and should also provide for the pressure drop
in the system. Or, a booster pump should be located between the pressure vessels to increase the
feed water pressure to the required level.

As the osmotic pressure of the feed water increases from the start to the end of the membrane
element, the feed water pressure decreases, and thus the pressure difference across the membrane
decreases, the water flow rate across the membrane decreases (equation (2)). Same will be the case,
when the concentrate from one pressure vessel flows as feed into the next pressure vessel.

As the feed water salinity increases, the permeate salinity will also be higher (equation (3)). Hence
the permeate salinity at the end of the membrane element, and similarly at the end of the last
pressure vessel, will be higher than that at the beginning. The salinity of the permeate obtained from
the RO module will be a flow rate-weighted average of the average salinity from each membrane
element. The salt passage rate will increase by about 3 to 17% per year as the membrane element
ages. Thus, the salinity of the permeate can be expected to increase by an average of 10% every
operating year.

17
Feed water pressure: The feed water pressure should be higher than the osmotic pressure
corresponding to the highest feed water / concentrate salinity in the system as explained above. As
the feed water pressure is increased further, the permeate flow increases (equation (2)), but the salt
passage remains the same (equation (3)). Hence the salinity of the permeate comes down as the feed
water pressure is increased and vice versa.

As the membrane ages, the permeate flux is expected to decline by about 2% to 10% per year, the
lower value corresponding to a feed water SDI of < 1 and the higher value corresponding to a feed
water SDI of 4. Since the RO systems are designed for a constant permeate flow rate, the high-
pressure pump should be capable of providing higher feed water pressures to provide for this
decrease in permeate flux. Similarly, this pump should be selected such that it can provide for the
change in feed water pressure required, corresponding to any expected change in feed water salinity
and temperature also.

Feed water temperature: The osmotic pressure is directly proportional to the feed water
temperature in K, i.e., about 0.33%/K at the normal room temperature. As mentioned earlier, both
the water flux and salt passage increase by about 3% per oC (or K) increase in feed water
temperature, due to the increase in membrane flow factors. Thus, the effect of temperature on
osmotic pressure is much less compared to that on water flow rate and salt passage rate across the
membrane. Hence, the permeate salinity should not be normally affected. However, in practice, the
permeate flux will be maintained constant by decreasing the operating pressure as the feed water
temperature rises. But, the salt passage will increase, thus increasing the permeate salinity.

2.3.3 Design limits: There are certain design limits, which are specific to the membrane element
chosen. Of these, salt passage percentage and permeate flow capacity have already been discussed
in the previous section. Others are maximum feed flow rate, minimum concentrate flow rate and
permeate recovery percentage, the last one being limited by the phenomenon of concentration
polarisation.

Maximum feed flow rate: This maximum feed flow rate is specified so that the integrity of the
spiral wound element is not affected by employing higher flow rates and correspondingly higher
operating pressures. The pressure drop on the feed side of the membrane element should also be
reasonable. Typical maximum feed flow rate for a 203.2 mm diameter and 1016 mm long TFC
polyamide membrane element is 17 m3/h.

Minimum concentrate flow: A minimum concentrate flow is required to maintain a minimum


velocity on the feed side of the membrane element to minimise the chances of membrane fouling or

18
scaling. Typical minimum concentrate flow rate for the above-mentioned membrane element is 2.7
m3/h.

Concentration polarisation & Permeate recovery percentage: As the salt is rejected at the
membrane surface, its concentration near the membrane surface (boundary layer) increases and
becomes higher than the salt concentration in the bulk of the liquid. This phenomenon is called
concentration polarisation. This concentration polarisation adversely affects the performance of the
membrane element as explained below.
The osmotic pressure at the membrane surface is higher (equation (1)) than that in the bulk
of the liquid and this reduces the permeate flow rate (equation (2)).
The salt passage rate increases (equation (3)).
The salt saturation percentage at the membrane surface may exceed the saturation
percentage limit and thus cause membrane scaling.
The concentration polarisation factor is defined as follows:
Concentration polarisation factor, Fcp = Cs/Cf (13)
Where,Cs = Salt concentration in kg/m3 at the membrane surface.

The concentration polarisation increases as the permeate flow increases, since the salt transport to
membrane surface increases. An increase in the feed water flow rate increases the turbulence and
decreases the thickness of the boundary layer having higher salt concentration and thus decreases
concentration polarisation. Hence the concentration polarisation factor can be determined as
follows:
Fcp = kcp.exp(Qp/Qf) (14)
Where,kcp = Membrane constant based on its geometry.
This equation shows that the concentration polarisation factor is dependent on the permeate
recovery ratio.

The maximum limit for the concentration polarisation factor is normally 1.2 for the optimum
performance of the membrane element. This corresponds to a permeate recovery percentage of
about 18% for an individual membrane element of the above-mentioned specification.

Membrane compaction: As the membranes are exposed to high water pressures, they get
compacted, i.e., their density increases. This compaction is maximum during the initial operating
period. Higher the feed water temperature, higher is the level of membrane compaction. This
compaction results in decreased water flow and salt passage rates across the membrane, resulting in
lower permeate flow and decreased permeate salinity. As the operating pressure is increased to

19
maintain the specified permeate flow, there is a further decrease in permeate salinity as explained
earlier. However, in practice, there is a net increase in salt passage percentage or permeate salinity
due to membrane fouling discussed in the next section. Membrane compaction is more significant
in the case of cellulose acetate membranes than in TFC polyamide membranes and also in seawater
RO, since the operating pressures are higher.

Membrane fouling: Fouling occurs as a result of the inorganic and organic substances present in
the feed water getting deposited on the membrane surface. The initial effect of fouling on the RO
process is similar to membrane compaction, i.e., decreased permeate flow, requiring increased feed
water pressure, and decreased permeate salinity. As fouling increases, it degrades the membrane,
thus increasing the salt passage percentage and hence permeate salinity. Membrane degradation also
occurs due to the harsh effects of the cleaning chemicals used to clean the fouled membrane. When
proper feed water pre-treatment is employed based on the feed water analysis, fouling will be
minimum. The normal life of a membrane element in a well-maintained RO system is a minimum
of three years.

2.3.4 Design procedure: Almost all the membrane element manufacturers have developed their
own software for the design of RO modules based on the membrane elements manufactured by
them. Since the design of an RO module requires a trial & error procedure, it is best done by such
software. Such software are user friendly and based on the Windows operating system and are
offered free of cost to the actual users by the manufacturers of the membrane elements. A general
description of the design procedure using such software is presented below.

The first step is to obtain the feed water analysis and enter the same in the analysis screen of the
software. Normally, the software itself balances the anions and cations and matches them with TDS,
calculates electrical conductivity, dissolved carbon dioxide based on the pH and bicarbonate
concentration, LSI and saturation percentages of sparingly soluble salts. If acid is to be added to
reduce the scaling tendency of the feed water, the software calculates the effect of such addition
also on the feed water analysis. It also calculates the osmotic pressure for the feed water.

The next step is to enter the product water flow rate required and the assumed product recovery
ratio / percentage in the design screen of the software. A product recovery percentage of 45 50%
is normally used for seawater RO desalination. Then a membrane element suitable for the required
product water flow rate and quality is selected from the design screen itself based on the nature of
feed water and salt rejection percentage and allowable permeate flow rate of the membrane element.

20
The number of membrane elements required and number of pressure vessels required are estimated
as per the equations 11 & 12 discussed earlier.

The next step is to select a flow configuration for the RO module. There can be more than one
concentrate or permeate stages. Feeding the concentrate from one stage to the next stage is called
concentrate staging. Similarly, feeding the permeate from one stage to the next stage is called
permeate staging. Each stage can contain several pressure vessels connected in parallel, with each
pressure vessel containing upto six membrane elements in series. The concentrate stages are called
arrays and permeate stages are called passes. A typical RO module will consist of a single pass and
two arrays. The same flow configuration can be used as the first choice for seawater RO
desalination. Since the concentrate flow rate from the first array, which is fed to the second array as
feed, is less than the feed flow rate to the first array, the number of pressure vessels in the second
array is assumed to be half the number in the first array in the first trial of design. Thus, two-third of
the estimated number of pressure vessels required will be in the first array and balance one-third in
the second array. These figures are entered in the design screen and the first trial of the RO module
design is run.

The design program performs an iterative calculation in which the program first estimates a feed
pressure to satisfy the desired permeate recovery and then calculates the performance of the first
element of the system. The concentrate from the first element becomes the feed to the second
element, and a second calculation of membrane element performance is made, and so on from
element to element through the complete array of the proposed design. The program then sums the
permeate flow from all elements, and compares this value to the target value. The program adjusts
the feed pressure based on this comparison, causing the program to converge on the required feed
pressure to achieve the required permeate recovery given the user defined system parameters. The
program calculates the permeate and concentrate salinity also based on the salt passage rates for the
selected membrane element.

The design output from the software is displayed on the screen and can be printed also. The output
indicates any design limit violation with respect to maximum feed flow rate, minimum concentrate
flow rate and maximum concentration polarisation factor apart from potential scaling problems. The
scaling potential is indicated by the LSI and saturation percentages of sparingly soluble salts with
respect to the concentrate. The product recovery percentage or number of pressure vessels in each
array is changed, as required, in case of design limit violations and the design is run again. This is
repeated till the RO module design is satisfactory with respect to all parameters. The design output
also gives the required feed water rate (which is used to size the pre-treatment system), the feed

21
water pressure required (used to size the high pressure pump) and the flow rates and pressures for
each pressure vessel and membrane element.

The other flow configurations that can be used are concentrate recirculation, permeate throttling,
interstage concentrate pumping, etc. In the case of permeate staging described earlier also, an
interstage permeate booster pump is required.

Concentrate recirculation is used only in small and compact single stage RO units to increase the
product recovery percentage. Its disadvantages are higher unit power consumption, higher feed
pressure requirement and increased permeate salinity.

Permeate throttling and interstage concentrate pumping are used to equalise the permeate flow rates
from the first and second concentrate stages, which is the requirement in certain cases. The
permeate throttling is simpler and requires less capital cost, but its unit power consumption is
higher. The interstage concentrate pumping requires increased capital cost but its unit power
consumption is less.

Permeate staging is used when seawater RO modules are to be operated at higher product recovery
percentages and/or at higher water temperatures and when very low salinity water is required from
brackish water RO modules.

2.3.5 Energy recovery system: The discharge of high-pressure concentrate from the RO module by
throttling is a waste of energy available in the concentrate stream. The energy loss is especially high
in the seawater RO systems since they employ high feed water pressures. Hence energy recovery
systems have been designed to recover the energy from the concentrate stream. The normal practice
is to use this energy for pumping a portion of the feed water with another pump being used to pump
the balance portion of feed water. The concentrate stream is used to run a turbine (for e.g., a Pelton
wheel), which runs a pump. There are other proprietary energy saving devices also. It is estimated
that atleast 30% of the power required for pumping feed water can be obtained from the concentrate
stream.

2.3.6 Membrane cleaning: The membranes can get fouled over time with inorganic, organic or
biological matter including colloidal particles. This results in increased feed water pressure
requirement and increased salt passage rate, thus affecting the performance of the RO module.
Hence it is essential to monitor the performance of the RO module regularly and establish and run a
regular cleaning schedule.

The operational data is normalised to account for the effect of changes in variables like temperature,
which also affect the performance. In general, membrane elements in a RO module are cleaned

22
when the normalised permeate flow rate has decreased by 10 15% (from the stat-up value);
pressure drop over a stage or the whole module has increased by 10 15% and the permeate salinity
has significantly increased.

A cleaning-in-place (CIP) system is used to clean the membrane elements. This consists of a storage
tank of sufficient capacity to hold the total volume of the RO module, i.e., the volume of all
pressure vessels and the volume of interconnecting piping, plus the volume of the CIP system
pipelines, a circulating pump and a cartridge filter and piping connection from the filter to the RO
module and back from RO module (both the concentrate and permeate sides) to the storage tank.
The storage tank is provided with a heating system, if the membrane element supplier recommends
cleaning at higher temperature. The pump and the pipelines in the CIP system are sized based on the
recommendations of the membrane element manufacturer for the cleaning solution flow rate and
pressure.

The cleaning chemicals and the cleaning procedure suggested by the manufacturer of the membrane
elements should be strictly followed.

2.4 Post-treatment:
The post-treatment of the permeate water obtained from the RO module is carried out to adjust its
pH to the required level and improve its taste. The pH of the permeate is low and it is necessary to
increase it to a value between 6.5 and 7.5. Other chemicals are added as required to improve the
taste of product water.

2.4.1 Degasification: The low pH of permeate is essentially caused by dissolved carbon dioxide.
Hence the pH of the permeate can be increased by degasifying the permeate with a stream of
filtered air to remove the dissolved carbon dioxide.

2.4.2 Chemical dosing: If the pH is still to be increased, alkalis like sodium bicarbonate, sodium
carbonate, sodium hydroxide and calcium hydroxide are added in the form of solutions through a
metering pump to the permeate stream in a static mixer or flash mixer. A combination of acids and
alkalis or salts can also be added to improve the taste of the product water.

3 Plant layout:

The primary consideration in plant layout is to provide an economical plant, safe and easy to
operate and maintain. The layout should be compact. The three major steps involved in the plant
layout are: 1. concept design, 2. equipment layout and 3. piping layout. Though these three steps
generally follow one after the other, there is considerable overlap between these activities. These
steps are discussed in more detail below.

23
3.1 Concept design:
At this stage only essential data like process or block flow diagrams, preliminary equipment sizes
and site information are available. The choice of indoor or outdoor plant or a suitable combination
of them is made, wherever applicable / permissible. The need for stand-by equipment like pumps or
stand-by power supply is identified and provided for. Horizontal and vertical equipment
relationships are planned and a preliminary equipment layout is prepared. Space allocation is made
for the different sections of the plant like process plant(s) (can be one or more), control room(s),
motor control centre(s), utilities, storage, effluent treatment plant, other auxiliary services like
workshop, laboratory, office, parking, garden & landscaping and area for future expansion. A master
plot plan is prepared arranging the above sections / areas in the site plan, considering the
constructional, erectional, process, operational, maintenance, safety and environmental aspects and
providing for proper access to the site and individual sections of the plant, emergency escape routes,
interconnections between the different sections of the plant and material-handling requirements, as
required. The internal roads are generally 6 7.5 m wide with a minimum headroom of 4.5 5m. A
plant model can also be made.

A typical arrangement of an outdoor plant is as follows: Process equipment is arranged in process-


flow sequence on both sides of a central pipe rack. Pumps are lined up in two rows under the pipe
rack. A road is provided between the two pump rows for maintenance and operation access. Two
roads along the periphery of the plant, parallel to the central road, allow access for construction and
maintenance equipment and operations.

Indoor plants are used when frequent attention by operators, material handling, cleanliness, weather
protection, environmental control and safety are the major considerations. The arrangement of
several indoor plants is also similar to the outdoor plant layout. In the case of multilevel plants,
vertical relationship of process equipment like provision for gravity flow is an important
consideration.

As the process flowsheet, equipment sizing & layout, etc. are finalised, the concept design is fine-
tuned, resulting in the final master plot plan or site plan or overall plant layout. The plant model, if
made, is also updated. Changes in process, operating philosophy or equipment type & size can
result in major changes in the concept design.

3.2 Equipment layout:


This is the detailed arrangement of equipment within a process plant or section considering the
horizontal and vertical relationships and interconnections among them. The information required at

24
this stage include: 1. preliminary piping & instrumentation (P & I) drawings with the specifications
& insulation data, 2. equipment drawings with elevation, plan, pipe & instrument connection &
insulation details, 3. preliminary routing of interconnecting piping, 4. instrumentation details
including panels, instrument-line trays, transmitters, etc. and their locations or routing, 5. special
requirements for the equipment like pump NPSH, 6. weight & size of equipment that require
handling during erection, operation or maintenance, 7. electrical requirements including cable trays,
cable runs, junction boxes, starters, etc. and their sizes and routing or locations and 8. safety &
environmental requirements.

The equipment should be arranged in process flow sequence in plan and elevation. Space should be
provided around equipment for convenient operation and maintenance access. Access arrangements
should be straight and simple. The width of walkways, catwalks or platforms in a plant should be a
minimum of 0.75 m with a minimum headroom of 2.25 m. A typical outdoor plant layout has been
described earlier. The central access road below the pipe rack in this case should be 4.5 m wide with
a minimum headroom of 3.5 m. In general the pump rows located outdoors should have an access
space of 3 m on either side. The maximum vertical run of ladders between ground / platforms
should be 6 m.

In case of indoor plants, each area of operation should have a minimum of two entry / exit points.
The operating and inspection points should be accessible and visible from operating aisles,
preferably without the help of auxiliary platforms and ladders. Equipment parts, instruments, valve
handwheels and piping should not protrude into access aisles. The main access aisles should be 2.4
m wide with a minimum headroom of 2.7 m. Access aisles over 7.5 m long should not be dead-
ended. The distance between a pump row and wall should be 1.2 m. The distance between two
pump rows should be 1.8 m. The height of the manhole on the top head of a vessel and the
platform / floor should be 0.75 1.2 m. Similarly, a horizontal manhole should be located 0.6 1.5
m above floor / platform. The minimum width of the staircase should be 0.75 m and the vertical rise
of a stairway between landings should be a maximum of 4.5 m with the landing length in the
direction of stairs being a minimum of 0.9 m.

Similarly all the access space and clearances should be provided based on the operational,
maintenance and safety needs.

3.3 Piping layout:


Piping should be arranged in an orderly manner and routed as directly as practical, preferably in
pipe racks. Generally, process lines, utility lines other than water, electric cables, HVAC ducts and

25
instrument lines will be carried on overhead pipe racks at specified elevations. As far as practically
possible, the piping in the north-south and east-west banks should be run at different elevations,
with its elevation changing with changing direction. The water lines can be run below ground or
over pipe racks, as convenient. Where future extension is anticipated, a valve should be provided at
the end of the line so that additions to the pipeline can be made without costly shutdown. The
valves should be preferably located at a height of 0.9 1.5 m height. The vent valves should be
located at easily accessible points. Drain valves should be provided at the low points of the pipe
system. Relief valves should be accessible for inspection and testing, where possible. Control valves
are generally located 0.6 m above ground / floor. All instruments should be located at accessible
points.

4 Hydraulic calculation:

Hydraulic calculations are performed to size the various pipelines in the plant and to determine the
pressure drops across the individual equipment and pipelines. Based on the pipeline pressure drops,
the equipment levels and required operating pressure (for e.g., feed pressure required for RO
module), the operating pressures at the inlet and outlet of the individual equipment, including
pumps, are determined. The pumps are sized based on its suction and discharge pressures, liquid
flow rate and the liquid properties.

The pipelines are sized based on a liquid velocity of about 2 m/s. The nearest or next higher
standard pipe size is chosen. The pressure drop in a pipeline is calculated based on liquid velocity in
the pipe, pipe size and length, friction factor and pipe roughness as follows:
Pressure drop in Pa, dP = 2.f.v2.d.L/D (15) (1 bar = 100,000 Pa)
Where,f = Fanning friction factor,
v = Liquid velocity in pipe in m/s,
d = Liquid density in kg/m3,
L = Equivalent pipe length in m and
D = Pipe inside diameter in m.

Fanning friction factor is obtained from pipe flow chart or the formula given below.
f = 0.0625/(log10(0.27.e/D + (7/Re)0.9))2 (16)
Where,e = Pipe inside surface roughness in m,
Re = Reynolds no. = D.v.d/vis and
vis = Liquid viscosity in Pa.s.

26
Equivalent pipe length is obtained by adding equivalent lengths of all pipefittings and valves to the
total pipe length. The equivalent length of a pipefitting or valve is obtained from the equivalent
length/pipe diameter ratio (L/D) and pipe diameter as given below.
Equivalent length of a fitting in m, Le,fitting = (L/D)fitting.D (17)

The (L/D) ratios for common pipefittings & valves are listed below.
Globe valve (open) 340
Gate valve (open) 13
Check valve (swing type) 135
Butterfly valve (open) 20
90o standard elbow 30
Std. tee (flow through run) 20
(flow through branch) 60

The pressure drop due to pipe enlargement / contraction and entrance / exit losses is calculated in
terms of number of velocity heads, K. The velocity head is nothing but v2/2g.
Pressure drop in Pa, dP = K.v2.d/2 (18)

K values are calculated as given below.


K for sudden enlargement = (1 (D1/D2)2)2 (19)
Where,D1 = Smaller pipe diameter in m and
D2 = Larger pipe diameter in m.
K for sudden contraction = 0.5.(1 (D1/D2)2) (20)
These K values are based on the velocity in smaller diameter pipe. Exit is a special case of
enlargement with D1/D2 being 0, i.e., K = 1. Similarly, entrance is a special case of contraction and
K value for sharp edged entrance is 0.5.

The pressure drop for liquid flowing through a bed of solids, like in multi-media & activated carbon
filters, is determined as follows.
Pressure drop in Pa, dPbed = fbed.vbed2.d.(H/Dp).((1-e)/e3) (21)
Where,fbed = Bed friction factor,
vbed = Liquid velocity in bed in m/s,
H = Bed height in m,
Dp = Particle size of solids in m and
e = Bed porosity or void fraction.

The bed friction factor is calculated as follows.


fbed = 1.75 + 150.(1-e)/Rebed (22)
Where,Rebed = Reynolds no. = Dp.vbed.d/vis.

27
5 Instrumentation & Process control:

The main instruments required are pressure gauges to measure the system pressure at various
locations. Flow meters are used to measure the feed, permeate and concentrate flow rates and also
the flow rates of chemical additives. Permeate, feed and concentrate salinities are also measured on-
line based on electrical conductivity. The pH of the feed to the RO module and permeate before and
after post-treatment are also monitored on-line. The feed temperature is also monitored since it can
affect the performance.

Automatic process controllers are used mainly to maintain the required feed flow rate using a
control valve in the discharge pipe of the high pressure pump feeding the RO module and the
required feed pressure or permeate flow rate using a control valve in the concentrate discharge pipe
of the RO module. Controllers also can be used to dose the chemical additives in proportion to the
feed / permeate flow rate or based on the water pH in case of acid / alkali. The backwash operation
of the multi-media and activated carbon filters can also be automated based on a set pressure drop
across the filter.

The liquid levels in the tanks like flash mixers in the system are also maintained by using control
valves in the inlet / outlet lines.

28

You might also like