1887

Abstract

In order to ensure long-term sustainability of the reservoir, the gas industry in Qatar is faced with the challenge of reducing the volume of produced and process water (PPW) sent to disposal wells by 50% [1-3]. Recently, Qatargas initiated a project to recycle process water and thus, reduce disposal volumes using commercial advanced water treatment technologies [4]. One emerging technology, “osmotic concentration” (OC) has been identified that offers a low-energy alternative to conventional thermal or membrane volume reduction methods. Osmotic concentration is a membrane filtration process that mimics first step in a forward osmosis (FO) system. It requires a high salinity draw solution (DS) which passes on one side of a semi-permeable FO membrane while the feed passes on the other side. Water from the feed is drawn through the membrane, via natural osmosis, reducing the feed volume and increasing the volume of the draw solution. This paper summarizes the results of bench-scale volume reduction tests with PPW collected from Qatar's North Field operations as the feed and either seawater or the concentrated brine from thermal desalination plants as the draw solution. While in conventional forward osmosis, the draw solution is regenerated, in OC, there is no regeneration of the draw solution. The diluted seawater or brine would be simply discharged to the Arabian Gulf. For future projects/developments, the authors have proposed OC for PPW volume reduction, which can be a cost-efficient alternative to achieve 50% reduction in disposal volumes (Fig. 1). This approach is particularly applicable in Qatar due to close proximity of desalination plants and gas processing facilities. In all membrane processes, the driving force for permeation is pressure. The mechanism by which the pressure is created differentiates various membrane processes: Reverse osmosis: static pressure generated by a pump Membrane distillation: vapor pressure differential due to a difference in temperature Osmotic concentration: osmotic pressure differential due to a difference in salinity. In these examples, the driving force or transmembrane pressure (TMP) can be measured in units of kPa or bar. In reverse osmosis, the TMP ranges from 15 to 60 bar depending on the salinity of the feed. In osmotic concentration, a comparable TMP of 15 to 60 bar is generated simply by high salinity the draw solution, i.e. without any static pressure being required and hence the low energy requirements for OC processes. In addition to significantly lower operating energy requirements [5], an OC process offers the following advantages over reverse osmosis (RO): Lower capital cost because pressure-rated vessels and high pressure pumps are not required [6]; There is strong evidence that FO membranes are less prone to irreversible fouling than RO membranes and foulants can be removed by simple flushing with clean water with no addition of chemicals [7,8]; The water discharged to the Arabian Gulf is of lower salinity and that provides an environmental benefit. The primary disadvantages of OC are that there is no water recovery after the separation process and there is limited experience from full-scale water treatment installations. The main objective of this study was to investigate the feasibility of OC to concentrate PPW from gas operations by 50% using brine from thermal desalination plants as draw solution. The PPW was a combination of gas field produced water extracted from an offshore reservoir and process water from onshore operations. The blending ratio between produced and process water was approximately 1:5. The PPW underwent deoiling, HS removal and cartridge filtration (2 mm). The detailed composition of PPW and brine from thermal desalination plant are shown in Table 1. During this study, experiments were conducted to evaluate OC performance in treating PPW in the following areas: membrane configuration membrane fouling effect of pretreatment process optimization long-term stability/performance. In all tests, the active layer faced the feed solution (AL-FS mode) since this configuration provides better control of feed-side fouling. Membrane configuration During this project, two membrane configurations were evaluated: Flat sheet (FS) membranes, commercially available [9]Hollow fiber (HF) membranes, Singapore Membrane Technology Centre [10, 11]. The membrane surface areas were 0.014 and 0.0106 m2 for the FS and HF modules, respectively. Experiments were conducted using two feed solutions: DI water and PPW. Results showed that the HF membranes had improved performance from both the water flux and reverse solute flux (RSF) perspectives. The HF membrane flux was ≈ 35 to 45% higher for both DI water and PPW (Fig. 2). With PPW as feed at 25?°C, the RSF was measured for both membranes and the results showed that HF membranes exhibited ≤ 3 mmol/m2 h RSF for Na+ and Cl−  while FS membranes showed a RSF of ≈ 20 mmol/m2h for both ions. RSF is highly sensitive to operating temperature. Because HF membranes showed superior performance and there are also commercial advantages (higher packing density, lower fabrication cost), experiments focused on evaluating HF performance in treating PPW. To assess if membrane fouling occurred, a benchmark test with DI water as feed solution and 1M NaCl as draw solution at 25?°C was conducted before and after each fouling test. A decline in the benchmark flux after treating PPW would indicate that membrane fouling had occurred. The fouling tests were conducted on two feed streams: synthetic PPW (mimicking only the inorganic content of PPW) and real PPW. During the experiments, the initial volume of PPW was reduced by 50% and the draw solution (DS) was 1M NaCl. The DS concentration, for both benchmark and fouling tests, was maintained constant throughout the experiments by adding concentrated NaCl solution based on conductivity measurements. While the results for synthetic PPW showed that no fouling had occurred, the results showed that PPW could cause fouling on the membrane surface since the benchmark flux decreased from 17.5 to 15 L/m2?h (Fig. 3). The fouling was attributed to the organics present in the PPW since no flux decline was observed on the when synthetic PPW was used as feed. These results highlighted the need for effective pretreatment to remove organics. To determine if pretreatment could remove the organics responsible for fouling, a number of methods were screened and ultimately powdered activated carbon (PAC) was selected for pretreatment of the PPW. PAC is widely used for organics removal and previously evaluated for similar applications [12]. Lab results showed that at a dosage of 500 mg/L PAC, the TOC from the PPW was reduced from 132 to 45 mg/L. The PAC dosage was considered very high for a full-scale application and further pretreatment optimization is needed before field implementation. OC performance experiments showed no decline in benchmark flux when the volume of pretreated PPW was reduced by 50% indicating that pretreatment is essential for the successful implementation OC to reduce PPW disposal volumes. Results also showed that the HF membranes have good rejection for organics. For both the treated and untreated PPW, the TOC in the draw solution after OC treatment was below the 1 mg/L detection limit indicating that the membrane rejection of the organics was >99%. The main operating parameters for this application were optimized using a Box-Behnken design (BBD) [13, 14]. BBD is a response surface methodology that explores the effect of different input variables (temperature, draw solution concentration and feed crossflow velocity) on the output response (flux). This statistical tool takes into consideration the combined effects and interdependence of the input variables and significantly reduces the number of tests required as compared to the conventional factorial experimental design. The following parameters and test values were used during the BBD experiments: Temperature: 25, 35 and 45?°C Draw solution concentrations: 40, 55 and 70 g/L NaCl Feed crossflow velocity: 40, 60, 80 cm/s. Results showed that the temperature has the greatest impact on the flux, since it influences the water viscosity. DS concentration directly affects the osmotic pressure, influencing the flux. The feed crossflow velocity did not affect the process performance over the range tested. These results are consistent with authors’ expectations and general published results. Based on the BBD analysis, the optimized process conditions were: 45°C temperature, 70 g/L draw solution concentration and 80 cm/s feed crossflow velocity. Figure 4 shows the comparison of the OC performance, to achieve 50% feed volume reduction, at benchmark (25?°C, 70 cm/s, 58.5 g/L NaCl) and optimized (45?°C, 80 cm/s, 70 g/L NaCl) conditions. The elimination of the water recovery step in osmotic concentration makes it an energy efficient process [5]. To confirm if the brine from thermal desalination plants is a suitable DS for the treatment of PPW, a long-term experiment was performed simulating full-scale operation. In earlier experiments, the concentration of the DS was maintained constant, by adding concentrated DS. Since this is not feasible in full-scale applications, a process stability experiment was carried out without controlling the DS's concentration, allowing it to dilute with time as water permeates through the FO membrane. The experiment was conducted using pretreated PPW as feed and brine from thermal desalination plant as DS. The solution temperature was 45?°C since this is the expected temperature of the brine discharged from the desalination plant. The feed and draw solutions crossflow velocities were 80 and 40 cm/s respectively. The test was conducted for 80 hours. The PPW initial volume was 42 L and it was reduced in volume by 50% while the volume of the draw solution, initially 21 L, was increased to 42 L, reducing its salinity by 50%. A relatively stable performance was observed throughout the experiment with a 30% decrease in flux (from 28 to 20 L/m2?h) due to the decrease of the effective osmotic pressure and to the influence of the internal concentration polarization (ICP). After a sharp initial decline in osmotic pressure differential, the decline in flux almost tracked the decline in the osmotic pressure differential (Fig. 5). DI water fluxes during benchmark tests conducted before and after the experiment remained constant at 19.4 L/m2?h indicating that negligible fouling occurred. To evaluate the ability of HF membranes to reject specific contaminants, various water quality analyses were performed. Results showed that HF membranes have high rejection capabilities. The ions with the highest solute fluxes values were sodium and chloride with a RSF of 120 and 91 mmol/(m2?h) respectively at 45?°C (Table 2). Although a small amount of nitrogen passed through the membrane from the PPW to the draw solution, at the levels found (5.8 mg/L), it was below the discharge limits set by the State of Qatar and the European commission (10 mg/L) [15]. The results also showed that the organic carbon present in the PPW was rejected by the membrane and retained in the PPW. A slightly increase in the TOC concentration in the DS was observed and it could be attributed to uncertainties in the analysis since the results were at the low end of the measurement accuracy. Finally, preliminary cost estimates and energy calculations showed that OC is economically feasible to reduce PPW injection volumes from gas fields in an environmentally sustainable manner. The research team is currently evaluating different pilot testing opportunities to further demonstrate the cost-effectiveness of this technology under relevant field conditions.

Loading

Article metrics loading...

/content/papers/10.5339/qfarc.2016.EEPP1948
2016-03-21
2020-07-04
Loading full text...

Full text loading...

http://instance.metastore.ingenta.com/content/papers/10.5339/qfarc.2016.EEPP1948
Loading
This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error