Article pubs.acs.org/est
Application of Hollow Fiber Forward Osmosis Membranes for Produced and Process Water Volume Reduction: An Osmotic Concentration Process Joel Minier-Matar,† Ana Santos,† Altaf Hussain,† Arnold Janson,† Rong Wang,‡ Anthony G. Fane,‡ and Samer Adham*,† †
ConocoPhillips Global Water Sustainability Centre, Qatar Science & Technology Park, Doha, Qatar Singapore Membrane Technology Centre, Nanyang Technological University, Singapore
‡
ABSTRACT: Produced and process water (PPW) from oil and gas operations, specifically in Qatar, are disposed of by deep well injection in onshore facilities. Disposing large volumes of PPW may affect deep well formation sustainability highlighting the need for effective PPW management. Forward osmosis (FO) was applied as an “osmotic concentration” process to reduce PPW injection volumes by 50% using brines and seawater as draw solutions (DS). The energy intensive step of restoring the salinity of the DS was eliminated; the diluted DS would be simply discharged to the ocean. Both hollow fiber and flat sheet FO membranes were tested and the former exhibited better flux and rejection; they are the focus of this study. Optimization experiments, conducted using Box-Behnken statistical design, confirmed that temperature and DS concentration had a substantial effect on performance. To validate the concept, a long-term experiment, under optimized conditions, was conducted with PPW as feed and brine from thermal desalination plant as DS which yielded an average flux of 24 L/m2h. The results confirmed that low-energy osmotic concentration FO has the potential for full-scale implementation to reduce PPW injection volumes. Pilot testing opportunities are being evaluated to demonstrate the effectiveness of this technology under field conditions.
1. INTRODUCTION
The selection of the draw solution depends on the availability of brines or seawater. However, there are several advantages of using brine as draw solution compared to seawater. First, the higher salinity of the brine15,16 results in a greater osmotic pressure gradient, ultimately leading to higher membrane fluxes.17 Second, no additional pretreatment is required for the brine since the pretreatment is already incorporated in the desalination plants.15 Additionally, the temperature of the brine streams discharged from thermal desalination plants is slightly higher than that of the seawater,15,16 which enhances FO performance.18,19 Operating parameters play an important role in the process performance.17,19,20 Published data indicate that the temperature and the draw solution salinity have a major impact on the flux.18,19,21 It has been also reported that the crossflow velocity influences the thickness of the boundary layer at the membrane surface.19 Most studies have conducted two-level factorial designs to optimize the operating conditions of the process. While this approach is the simplest to implement, it is less comprehensive than second order experimental designs.22
Water management is one of the key global challenges faced by the oil and gas industry,1 involving both produced water from reservoirs and process water from operations.2 The most common method of managing those streams onshore is by deep well injection.1,3−5 However, disposing large volumes can affect the formation integrity3 and a specific target was set in Qatar to reduce the injection volumes by 50% to ensure longterm sustainability of the formation.6 In this study, forward osmosis (FO) has been evaluated as an “osmotic concentration” process for volume reduction (Figure 1). While conventional FO requires two steps, a low-energy FO separation followed by an energy-intensive step to restore draw solution salinity,7−9 osmotic concentration (also known as osmotic dilution) involves only the first FO separation step.10−13 The readily available draw solution (seawater or reject brine from desalination plants) is discharged without water recovery and the relative low salinity feed, produced/ process water (PPW), with its volume reduced by 50%, can be sent to injection wells. Advantages include: • Low capital/operational cost and low energy process for PPW volume reduction10,14 • Beneficial environmental impact due to reducing salinity of the brine discharged15 © XXXX American Chemical Society
Received: September 30, 2015 Revised: April 10, 2016 Accepted: May 10, 2016
A
DOI: 10.1021/acs.est.5b04801 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
Figure 1. Osmotic concentration as applied for produced/process water volume reduction using brine or seawater as draw solution.
Figure 2. Hollow Fiber Bench Scale System Schematic.
Alternatively, response surface methodologies (RSM) can be applied to fit second order models. RSM are a set of statistical tools often used to identify optimum conditions by considering the combined effects of various process parameters.23,24 An advantage to this approach is the significant reduction of the number of experiments required to find specific trends. In this study, a Box-Behnken statistical analysis,22,25 one of the most common RSM, was used to determine the optimum operating conditions for the FO process. Membrane configuration also plays a role in the implementation of the FO process.26 Flat sheet (FS) membranes are commonly used in FO applications,10 mainly in the spiral wound configuration since they are easy to install due their modular design. Also FS membrane are commercially available and installed in full scale applications.26,27 Recently, new FO hollow fiber membranes (HF)28 have been developed that offer potential advantages over FS membranes: • Higher productivity per unit volume due to higher packing density.26,28−31 • Self-supported membranes eliminate the need for spacers, lowering manufacturing cost and improving performance.30−33 • Easy to scale up since fibers can be easily potted inside a holding vessel.7 • Improved hydrodynamics enhancing the shear forces on the membrane surface.8 However, one of the key drawbacks of HF membranes is their limited commercial availability in the market, indicating that vendor engagement and further applied research is needed in this area before full scale installation.34,35
In the present work, the performance of FO to reduce the disposal volumes of PPW was evaluated. During this evaluation, two different FO configurations were considered: commercially available FS membranes36 and HF membranes synthesized and provided by the Singapore Membrane Technology Centre.29 The main process parameters (temperature, draw solution salinity and feed crossflow velocity) were optimized using a Box-Behnken statistical design and different readily available draw solutions (seawater and brines), collected from desalination plants in Qatar, were evaluated. To validate the concept, a long-term experiment was conducted over 80 h and detailed water quality analyses were performed.
2. MATERIALS AND METHODS 2.1. Bench Scale System. The FO system used in the experiments was custom-built (Figure 2). Two variable speed pumps (KNF, Switzerland) were used to generate crossflow velocity inside and outside the fibers, creating two separate closed loops for the feed and draw solutions, respectively. The flow rates were kept constant using a PID controller integrated with the control system (National Instruments). The feed solution tank was placed on a digital balance (Meter Toledo) and weight changes were monitored to calculate the water flux. Feed and draw solution temperatures were adjusted using a water bath (Julabo, Germany). Temperature, pressure and conductivity (for both feed and draw loops) were recorded throughout the tests using various sensors as shown in Figure 2. The concentration of the draw solution was monitored using online conductivity measurements. When the draw solution conductivity dropped below the set-point, a dosing pump (KNF, Switzerland) was used to automatically add concenB
DOI: 10.1021/acs.est.5b04801 Environ. Sci. Technol. XXXX, XXX, XXX−XXX
Article
Environmental Science & Technology
2.5. Feed and Draw Solutions. 2.5.1. Feed Solution: Produced/Process Water (PPW). The PPW was a combination of gas field produced water extracted from an offshore reservoir and process water from onshore operations. The blend ratio between produced and process water was approximately 1:5. The PPW underwent deoiling, H2S removal and cartridge filtration (2 μm). On average, the temperature of the PPW ranges from 20 to 35 °C. Since the PPW was collected from a gas field, it is not as contaminated as flowback or produced water from an oil field. The detailed composition of this stream is shown in Table 1. In every experiment with PPW, the feed volume was reduced by 50%.
trated draw solution until it reached the desired level. The initial volume for both feed and draw solutions was 1300 mL. A similar set up was used for flat sheet testing.9 2.2. Membranes. Two different thin film composite (TFC) membrane configurations were evaluated: (1) Flat sheet (FS) membranes, commercially available.36 The FS membranes had a hydrophilic TFC active layer embedded in a polyester screen support. Each membrane coupon measured 190 mm by 140 mm with an effective membrane area of 0.014 m2. The reported A and B values for this membrane, at 20 °C, are 1.8 L/(m2h-bar) and 0.9 L/m2h, respectively.36 Standard diamond shape spacers (similar to those used in reverse osmosis (RO) spiral wound elements) were used on both feed and DS channels; the thickness of the spacer was 65 mils. (2) Hollow fiber (HF) membranes developed by Singapore Membrane Technology Centre.28,29 These membranes have a hydrophilic TFC active layer on the lumen of a poly(ether sulfone) HF substrate. The active layer was cast on the lumen (inner layer) of the fibers because if cast on the outer layer, it could be damaged due to the friction/interaction between the fibers in the module. Each module had 15 fibers with an active length of 230 mm. The fiber ID and OD were 0.98 mm and 1.34 mm, respectively. The module had an effective membrane area of 0.0106 m2. The A and B values for this membrane, at 20 °C, are 2.34 L/(m2h-bar) and 0.137 L/m2h, respectively. 2.3. Theoretical Model for Flux Analysis. The flux of the FO membranes was modeled by the solution-diffusion model (eq 1)29 and used to predict the ideal flux profile over time: ⎞ ⎛ Aπ draw + B ⎟⎟ Jv = K m ln⎜⎜ ⎝ Aπfeed + Jv + B ⎠
Table 1. Chemical Composition of the PPW and Draw Solutions parameter (mg/L unless specified) conductivity, μS/ cm TDS pH chloride (Cl−) sodium (Na+) calcium (Ca2+) magnesium (Mg2+) bromide (Br −) sulfate (SO42−) potassium (K+) phosphate (PO43−) total nitrogen (TN) total organic carbon (TOC) inorganic carbon (IC) osmotic pressure (25 °C), bar
(1)
Where Jv is the volumetric flux of water, and πdraw and πfeed are the osmotic pressures of the draw solution and the feedwater, respectively. A and B are the water and salt permeability coefficients respectively, determined empirically by RO experiments at various temperatures.29 Km is the membrane mass transfer coefficient, given by Km =
ε·D D = τ·l S
(2)
qρgh 3.6x106η
(3)
where q = flow rate (m /h), ρ = density of fluid (kg/m ), g = gravity (9.81 m/s2), h = differential head (m), η = pump efficiency For the calculations presented in this paper, the pump efficiency was assumed to be 0.6 and the differential head was assumed to be 20 m. The energy consumption for the RO process was estimated using the Reverse Osmosis System Analysis (ROSA) design software.37 3
brine from thermal desalination plant
brine from RO plant
seawater
1725
100 250
85 600
66 543
1550 8 284 345 38 8 5 347 4.5
