47 
IJCPE Vol.11 No.4 (September 2010) 

 
Iraqi Journal of Chemical and Petroleum Engineering 

 Vol.11 No.3 (September 2010) 47- 54 
ISSN: 12010-4884 

 

 

 

POWER GENERATION FROM “PRO” PROCESS USING FLAT SHEET TFC–

ULP KOCH MEMBRANES 
 

Adil Al–Hemiri, Adel O. Sharif
*
 and Saif T. Manji 

Department of Chemical Engineering, College of Engineering, Baghdad University 

                                           *Department of Chemical Engineering, Surry University 

 

 

 

ABSTRACT 

The production of power using the process of pressure–retarded osmosis (PRO) has been studied both 

experimentally and theoretically for simulated sea water vs. river water and deionized water under two 

cases: the first is for simulated real conditions of sea water and river water and second under low brine 

solution concentration to examine the full profile of the power- pressure.   The influence of 

concentration polarization (CP) on water flux has been examined as well.  

Keywords: Pressure–retarded osmosis; renewable energy; TFC-ULP membrane; concentration 

polarization 
 

 

INTRODUCTION 

Osmosis hydro power is one of the recently 

attractive types of renewable energy. Suggested 

only in the 1970s by Sidney Loeb (the co-

inventor of Reverse Osmosis) through utilizing 

the osmotic pressure in power generation using 

membranes. It is a natural process occurred 

when salt solution and fresh water are portioned 

in two chambers separated by a semi-permeable 

membrane, made for example of a biological 

membrane, e.g. of hog bladder, fresh water will 

permeate through the membrane.  

The driving force is capable of elevating the 

salt water level above the fresh water level, 

whereas the potential energy is obtained from 

the static water height. The process stops when 

the hydraulic pressure is equal to the potential 

osmotic pressure of the salt water. In theory, a 

stream flowing at 1 m3/s could produce 1 MW 

of electricity [1]. The global potential for 

Pressure-Retarded Osmosis is calculated to be 

about 1600 TWh/y and for Europe 200 TWh/y 

[2]. Significant research efforts took place 

between mid 1970s to the early 1990s, but due 

to ineffective membranes, the key part of 

process, the  

construction of a PRO power plant was not 

considered. In 1997, Statkraft, one of the 

leading energy providers in Norway, started 

their research on PRO, they have constructed a 

PRO prototype and claimed to construct 

commercial power plants in 2015[2]. 

Concentration  

 

polarization a phenomenon of increasing the 

solute concentration at the low concentration 

side  

 

(concentrative External Concentration 

Polarization) and d Concentration Polarization) 

Iraqi Journal of Chemical 

and Petroleum Engineering 

 

University of Baghdad 

College of Engineering 



 

Power Generation from PRO Process Using Flat Sheet TFC–ULP Koch Membranes  

 

                                             IJCPE Vol.11 No.4 (December 2010) 48        

 

 

and the build up or dilution of solute at the 

interface between the membrane active and 

support layer (Internal Concentration 

Polarization).  

 

The first two types can be reduced greatly by 

PRO Membrane Cell and Experimental 

Procedure  

 

The PRO membrane cell is comprised of a two 

3 in diameter QVF vessels capable of 

withstanding pressure up to 1.5 bar as shown in 

Fig. 1. The membrane area is 45.6 cm2. 

providing agitation or circulation while the last 

type is strongly dependant on the resistance to 

solute transfer. This variable governs the rate of 

solute transfer within the support layer of the 

membrane and it is recommended to be as low 

as possible so that the solute will transport 

easily and hence prevent excessive 

concentration or dilution at the membrane inner 

interface [3]. 

 

The objective of the present study is describe 

the performance of PRO process under 

simulated real conditions using flat sheet TFC-

ULP Koch membrane and to derive a simple 

relation to predict theoretically that  

EXPERIMENTAL  

Membrane 

The membrane used for all experiments was 

provided by Koch Membrane systems. The 

membrane chemistry is uncertain but is 

believed to be composed of polyamide. 

 

Brine and Fresh Water Solutions  

Deionized water was used to prepare the 

sodium chloride solutions. These solutes were 

used to represent sea water (brine solution) and 

river water (fresh water). The brine solution 

concentration taken are as 26.5 and 35 g/l NaCl 

to resemble the actual concentration of NaCl in 

sea water for the first and if the whole salts in 

sea water are represented by NaCl for the 

second. The osmotic pressures of the various  

 

solutions were calculated using HYSYS® 3.1 

software. 

per

 
 

 

 Fig.1 Schematic diagram of the PRO 

membrane cell 

formance. 

 

 

This cell accepts flat sheet membranes. A 

plastic mesh spacer is provided on the fresh 

water side of to provide mechanical support to 

the membrane. The brine solution is flowing 

cocurrently to the fresh water and both 

solutions flow in closed circuits via plastic 

tubing. Two centrifugal pumps from STUART 

TURNER LTD. (England) were used for 

circulating the solutions through the system. 

Flowmeters measured the volumetric flow 

rates, which were fixed at 7 l/min for brine 

solution side and 5 for fresh water side with 

flow area around 11.4 cm2. Both solutions flow 

at the same constant temperature 25  1.5oC. 

The detailed diagram of the PRO system is 

given in Fig. 2. 

 

 



Adil A. A. Al–Hemiri, Adel O. Sharif, Saif T. Manji 

 

49 IJCPE Vol.11 No.4 (September 2010)                                                    

 

 
 

 

Fig. 2 Schematic diagram of the PRO system 

 

 

A pressure gauge followed by control valve is 

installed at the brine solution outlet on the 

membrane cell in order to pressurize the solution 

to any desired pressure. Water and salt fluxes are 

determined by mass balance by measuring salt 

concentrations using calibrated conductivity 

meter provided by WTW (Germany) in both 

solutions every 1 hour of experiment for a total of 

five hours. The power generated is calculated 

from the product of the pressure difference across 

the cell and water flux according to Eq. 1  
PJW W       (1)  

Where W is the specific power generated in 

Wm2, JW is the water flux in m3/ m2 (Lee et al, 

1981) 

 

 

RESULTS AND DISCUSSION  

 

In PRO operation, the brine solution side is 

pressurized as mentioned earlier; the positive 

difference between the brine solution and fresh 

water sides causes the water flux to decrease and 

the FO process to deviate towards Reverse 

Osmosis mode if the pressure difference does not 

surpass the osmotic pressure difference. Figs. 3 

and 4 show this behavior. 

 

 

 

 

Fig. 3 Measured water flux vs. pressure for 

35 g/l brine solution concentration using 

DI and river water at 7 and 5 l/min flow 

rates for brine solution side and fresh water 

respectively. All solutions are at 25 ± 

1.5oC 

 

 
Fig. 4 Measured water flux vs. pressure for 

26.5 g/l brine solution concentration using 

DI and river water at 7 and 5 l/min flow 

rates for brine solution side and fresh water 

respectively. All solutions are at 25 ± 

1.5oC 

 

The increase of the pressure difference is 

associated with drop in the water flux and 

is described by Eq. 2 
 PAJW        (2) 

 

Where A is the water permeability 

coefficient, Δπ is the osmotic pressure 

difference, and ΔP is the pressure 

difference [4]. 



 

Power Generation from PRO Process Using Flat Sheet TFC–ULP Koch Membranes  

 

                                             IJCPE Vol.11 No.4 (December 2010) 50        

 

 

 

The decrease in the water flux with increasing the 

pressure difference seems to lower the power 

produced but the increasing pressure compensates 

that drop due to increased momentum of the 

solution. 

 

 
Fig. 5 Measured specific power in Watts/m2 vs. 

pressure for brine solutions of 35 and 26.5 and 

concentration and DI water at 25 ± 1.5oC 

 

 
 

Fig. 6 Measured specific power in Watts/m2 vs. 

pressure for brine solutions of 35 and 26.5, 

concentration and river water at 25 ± 1.5oC 

 

Figs. 5 and 6 justify the increasing power 

generation resulted from increasing pressure 

difference regardless of the decrease in the water 

flux, however above figures show that the power 

produced is susceptible to further increase at 

higher pressures. The experimental observation 

of Thorsen and Holt (2009) and simple 

mathematical analysis provided elsewhere[3] 

reveal that the optimum operating pressure is 

about one-half the osmotic pressure 

difference show that the power pressure 

difference curve has a parabolic behavior 

were there is maximum point for the 

produced power and since such high 

pressure demanding operating condition is 

unattainable an alternative approach must 

be followed to predict experimentally the 

entire power – pressure curve. Such 

approach can be reached by using brine 

solution of low concentration that 

possesses an osmotic pressure within the 

operating pressure range of the current 

membrane cell (up to 1 bar). A relevant 

concentration of brine solution is 1.5 g 

NaCl/l which has an osmotic pressure of 

1.17 bar and optimum pressure difference 

of 0.6 bar. This concentration is also 

studied for the cases of DI water and 

simulated river water, see Figs. 7 and 8. 

 

 
 

 

Fig. 7 Measured specific power and water 

flux against pressure difference for 1.5 g 

NaCl/l brine solution and DI water at 25C 

± 1.5oC 

 

Fig. 7 has matched the behavior of the 

earlier study5] in its parabolic form. The 

continuous increase of the pressure 

difference from FO conditions (where W 

has zero value) is accompanied by a 

decline in water flux and simultaneous 

increase in the specific power until a 

certain point in which Wmax (peak of the 



Adil A. A. Al–Hemiri, Adel O. Sharif, Saif T. Manji 

 

51 IJCPE Vol.11 No.4 (September 2010)                                                    

 

curve) is reached. Wmax from Fig. 7 is about 

0.00798 W/m2 at ΔP = 0.6 bar, this value is 

subjected to comparison with the theoretical 

value from [3] 

4AW
2

m ax       (A.4) 

 

The water permeability coefficient for this system 

is found to be 2.335×10-12 m/s.Pa and with Δπ 

calculated from HYSYS 3.1 software Wmax is 

0.0078 W/m2 which appreciably agrees with the 

experimental value. 

 

 

Fig. 8 Measured specific power and water flux 

against pressure difference for 1.5 g NaCl/l brine 

solution and river water at 25C ± 1.5
o
C 

 

From Fig. 8 Wmax 0.0032 W/m
2
 at ΔP = 0.4 bar 

with Aapp = 2.096×10
-12

 m/s.Pa and Δπ = 

78,567.142 Pa, theoretical value of W is found to 

be 0.003235 W/m2. The agreement between the 

experimental results and results of Eq. A.4 is 

ascribed to the trivial polarization effects. 

 

The complete power – pressure difference curve 

using brine concentrations 35 and 26.5 g/l will be 

studied by the aid of Eq. 2 which neglects any 

polarization effects and modified form of Eq. 1 

that allows for the concentration polarization to 

take place, see Appendix A for details. 

 

 

 

 

 

 

 

Table 1 Values of permeability coefficient 

A, osmotic pressure difference and 

reflection coefficient for PRO system 

 

System 

A 

(m/s.Pa) 

×1012 

∆ 

(bar) 
 

35 g/l vs. 0.01 

g/l 
1.351 27.8 0.7494 

26.5 g/l vs. 

0.01 g/l 
1.558 21.0 0.7913 

35    g/l vs. 

0.5  g/l 
1.363 27.4 0.7032 

26.5 g/l vs. 

0.5  g/l 
1.533 20.6 0.7275 

 

The values of A are determined from the 

slopes in Figs. 3 and 4, ∆ is calculated 

using HYSYS 3.1, and   (reflection 

coefficient) is found from fitting data in 

Figs 5 and 6 to Eq. A.5 

Figures 9 – 12 show a predicted profile of 

the specific power against pressure 

difference under pressure values up to 20 

bars. The consideration of polarization 

effects in Eq. A.5 has greatly reduced the 

power generation to about 1 W/m2 as in 

Fig. 9 and 1.6 W/m2 as in Figs 10 – 12. 

 
Fig. 9 Specific power against pressure 

difference for 35 g/l NaCl brine solution 

and DI water. Specific power is predicted 

from Eqs. A.2 and A.5 based evaluated 

parameters in Table 1 



 

Power Generation from PRO Process Using Flat Sheet TFC–ULP Koch Membranes  

 

                                             IJCPE Vol.11 No.4 (December 2010) 52        

 

 

 
Fig. 10 Specific power against pressure 

difference for 35 g/l NaCl brine solution and 

simulated river water. Specific  

power is predicted from Eqs. A.2 and A.5 based 

evaluated parameters in Table 1 

 

 

 

 

 
Fig. 11 Specific power against pressure 

difference for 26.5 g/l NaCl brine solution and DI 

water. Specific power is predicted from Eqs. A.2 

and A.5 based on evaluated parameters in Table 1 

 

 

 

Fig. 12 Specific power against pressure 

difference for 26.5 g/l NaCl brine solution 

and simulated river water. Specific power 

is predicted from Eqs. A.2 and A.5 based 

on evaluated parameters in Table 1 

 

In Fig. 13 a comparison between the 

experimental results of Thorsen and 

Holt[5] and Eqs. A.2 and A.5 is made. The 

comparison shows the validity Eq. A.5) 

and also shows the error encountered when 

concentration polarization effects are 

omitted. Eq. A.5 has been in agreement 

with the experimental data to about 99.9 

using Chi square test. 

 

 
Fig. 13 Comparison between experimental 

results of Thorsen and Holt (2009) and the 

results of Eqs. A.2 and A.5. () symbols 

denote for experimental results, solid line 

for Eq. A.5 and dashed line for Eq. A.2 

 



Adil A. A. Al–Hemiri, Adel O. Sharif, Saif T. Manji 

 

53 IJCPE Vol.11 No.4 (September 2010)                                                    

 

CONCLUSION 

Osmosis hydro power seems to be very 

promising since it posses a capacity of 27 bar for 

sea water vs. river water but with the available 

membranes it is obvious that commercial 

production is beyond achieving yet. The reason is 

that the currently available membranes still have 

low water permeability coefficients.  

The concentration polarization CP have reduced 

the water flux greatly and hence power 

generation by a value about 1.6 W/m2, with value 

of  evaluated between 0.7 and 0.8 which 

represent the fraction that the total driving force 

∆ has reduced to. 

The comparison of the modified Eq. A.5 with the 

experimental work of Thorsen and Holt (2009) 

yielded an excellent agreement of 99%. 

 

Nomenclature  

A    water permeability coefficient (m.s-1.pa-1) 

JW   water flux (l.m-2.h) 

P     pressure of solutions (bar) 

W    specific power (Wm2) 

Greek letters 

π     osmotic pressure (pa) 

σ     reflection coefficient 

Superscripts 

*   denote for either P or W with  

     considering concentration polarization 

  

 REFERENCES  

1-Wick, G.L., and Schmitt, W.R. (1977), 

"Prospects for renewable energy from sea", 

Marine Technology Society Journal, 11. 

2-Statkraft (2006), "Osmotic Power — A huge 

renewable energy source, Information material", 

www. Statkraft.com. 

3-Lee, K.L., Baker, R.W., and Lonsdale, H.K. 

(1981), "Membranes for power generation by 

pressure-retarded osmosis", Journal of Membrane 

Science, 8. 

4-McCutcheon, J.R. and Elimelech, M. (2006), 

"Influence of concentrative and dilutive internal 

concentration polarization on the flux behavior in 

forward osmosis", Journal of Membrane 

Science, 284. 

5-Thorsen, T., and Holt, T. (2009), "The 

potential for power production from 

salinity gradients by pressure retarded 

osmosis", Journal of Membrane Science, 

335. 

 

Appendix A  

Evaluation of Reflection Coefficient 

Starting from Eq. 1 
PJW W         (1) 

Suppose that water flux is calculated from 
 PAJW        (A.1) 

Substituting Eq. A.1 into 1 to get 

 2PPAW 
     (A.2) 

 

The maximum attainable theoretical power 

(for negligible concentration polarization 

effects) is evaluated at ΔPoptimum 

obtained from differentiating Eq. A.2 and 

equating to zero [3]. By donning so 

ΔPoptimum will be half the osmotic 

pressure difference 
2Poptimum        (A.3) 

Therefore substituting (A.3) into (A.2) 

4AW
2

m ax        (A.4) 

Eq. A.4 presents an idealized approach for 

the calculation of the specific power since 

it is based on Eq. A.1 which does not allow 

for effects of ECP and ICP. Therefore the 

introduction of the reflection coefficient σ 

into Eq. A.1 is essential. By doing so result 

will be 

 2* PPAW      (A.5) 
 

Recalculating the optimum pressure and 

the maximum specific power in similar 

manner 

2P
*
optim um       (A.6) 

4AW
2*

m ax      (A.7)