- Email: [email protected]
Oilfield produced water treatment by ceramic membranes: Preliminary process cost estimation
Desalination 360 (2015) 81–86
Contents lists available at ScienceDirect
Desalination journal homepage: www.elsevier.com/locate/desal
Oilfield produced water treatment by ceramic membranes: Preliminary process cost estimation S.E. Weschenfelder a,b,⁎, A.C.C. Mello a,b, C.P. Borges c, J.C. Campos b a b c
Technology for Water Treatment and Reuse Department, Petrobras Research Center, Rio de Janeiro, Brazil School of Chemistry, Inorganic Processes Department, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil COPPE/Chemical Engineer Program, Federal University of Rio de Janeiro Rio de Janeiro, Brazil
H I G H L I G H T S • • • • •
The microfiltration process enables the treatment and reuse of produced water. The process costs have been related to cross flow velocity and water recovery. OPEX and CAPEX have been estimated for a real-scale plant under optimal conditions. The OPEX was found to be equal to US$0.23/m3. The CAPEX for a full scale plant was estimated at MUS$7.33.
a r t i c l e
i n f o
Article history: Received 17 November 2014 Received in revised form 11 January 2015 Accepted 12 January 2015 Available online xxxx Keywords: Produced water Treatment Ceramic membranes Cost estimation Process
a b s t r a c t The application of ceramic membranes for oilfield produced water treatment has been considered a very promising technology, mainly due to the oil and grease separation efficiency and process robustness. The purpose of this study was to obtain a preliminary estimate of operating expenses (OPEX) and capital expenditures (CAPEX) for a full scale ceramic membrane plant, based on data obtained in lab scale tests and from literature information. Different crossflow velocities (CFVs) and water recovery rates were simulated and the results were correlated to the OPEX, CAPEX and total cost (TC) per cubic meter of treated effluent. An increase of US $0.10/m3 in the OPEX and a 55% boost in the value referring to the CAPEX, by increasing the water recovery rate from 80% to 95% were observed. It was found that, under an optimal CFV (2.0 m·s−1) and considering the water recovery rate equal to 95%, the cost related to OPEX and TC were, respectively, US$0.23/m3 and US$3.21/ m3. The CAPEX for a full scale plant, capable of treating 1000 m3·h−1 of produced water, was estimated at MUS$7.33. In all of the experimental conditions assessed, it was possible to generate a permeate stream with oil and grease content (CO) lower than 5 mg·L−1. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Oilfield produced water is a byproduct of the oil extraction process from subsurface geological formations. Its composition typically includes the presence of oil, dissolved organic compounds and inorganic particles [1]. Depending on the oil producing field, the produced water may present oil and grease content higher than 500 mg·L−1 and salt concentrations (CS) between 80 and 200,000 mg·L− 1. In many cases this effluent is reused for the purposes of irrigation, reinjection into reservoir aiming to enhance oil recovery, or also for steam generation through the application of subsequent desalination processes. However,
⁎ Corresponding author at: Technology for Water Treatment and Reuse Department, Petrobras Research Center, Rio de Janeiro, Brazil. E-mail address: [email protected] (S.E. Weschenfelder).
http://dx.doi.org/10.1016/j.desal.2015.01.015 0011-9164/© 2015 Elsevier B.V. All rights reserved.
CO higher than 5 mg·L−1 may compromise the injection of water in the reservoir, as well as the efficiency of the salt removal processes [2–6]. Conventional produced water treatment processes may include flotation systems, hydrocyclones, and nutshell or mixed media filters. These types of equipment, however, have a reduced efficiency in removing solids and oil and grease particles whose dimensions are smaller than 5.0 μm, making it difficult to generate an effluent that is appropriate for reuse [2,7–9]. The membrane separation processes can be presented as an alternative technology to the conventional processes used in treating oilfield produced water. Ceramic membranes have been taken into consideration for presenting advantages connected to its greater mechanical, chemical and thermal resistance, in addition to its efficiency in removing oil and grease from streams with a high load of solids and oily contaminants, without the addition of chemical products [2]. The membranes used for that purpose may be produced from different materials,
82
S.E. Weschenfelder et al. / Desalination 360 (2015) 81–86
among which zirconium oxide, aluminum oxide and titanium oxide are highlighted. Many studies have been made comparing the performance of different inorganic materials for treating oily effluents. The reported results show that zirconium oxide membranes provide slightly higher fluxes than aluminum oxide and titanium oxide membranes [10–12]. Also different pore sizes have also been compared to oily water treatment. Hua et al. [13] reported that the use of membranes with pore sizes of 0.2 μm was not appropriate for high quality effluent. Similar results were also obtained by Srijaroonrat et al. [14] and Ebrahimi et al. [6]. They also concluded that the use of membranes with pore sizes of 0.1 μm gives the best results in terms of flux compared to the pore sizes of 0.05 μm and 0.5 μm. The purpose of this work was to assess the operating expenses and capital expenditures involved in a full scale ceramic membrane plant for the treatment of oilfield produced water. The performance of the microfiltration process was assessed in laboratory scale experiments using a synthetically-prepared effluent and simulating real operating situations. 2. Materials and methods 2.1. Membrane characteristics To perform the experiments, commercial zirconium oxide (ZrO2) membranes were evaluated. The membranes have 19 channels of 3 mm, microfiltration area equivalent to 0.0381 m2 and mean pore sizes of 0.1 μm. According to the manufacturer, Likuid Nanotek, these membranes can withstand pressures up to 8 bar, pH values between 0 and 14 and temperatures up to 100 °C. 2.1.1. Experiment system The experimental system consisted of a membrane module, a recirculation pump, pressure gauges and flow meters, flow control valves in the feed, permeate and concentrate streams, one tank to
collect the permeate and a heated and mechanically agitated feed tank. The permeate mass was continuously monitored by data acquisition. A schematic representation of the experiment set-up is shown in Fig. 1. During testing, the valves V-1 and V-2 remained opened and the feed flow rate was controlled by the frequency inverter connected to the pump B-1. The ΔPTM was adjusted through valve V-3 and calculated as the difference between the mean pressure given by PI-1 and PI-2 and the pressure given by PI-3. 2.1.2. Synthetic effluents The effluent used in the tests was synthetically prepared with distilled water, salt (sodium chloride) and oil. The oil was added to the saline mixture and immediately emulsified with a Turrax mixer (Ultra-Turrax T-50). The emulsion was deemed stable when the oil droplet mean size, as measured by a particle size analyzer (Malvern Mastersizer Micro) ranged from 10 to 30 μm. The oil had a density equal to 28°API and was obtained directly from an offshore oil production unit. Its concentration was determined by an infrared spectrophotometer (Horiba OCMA-350). 2.1.3. Experiments Before starting each experiment and after the chemical cleaning procedure, the membrane hydraulic permeability was determined. For such measurement, distilled water was applied and the permeate flux was recorded, under a turbulent flow (CFV = 3.0 m·s− 1) at 25 °C in different ΔPTM: 1.0, 2.0, and 3.0 bar. The hydraulic permeability was considered as the slope of a linear correlation between the permeate flux and ΔPTM. To assess the influence of salinity on the permeate flux, experiments were carried out using a synthetic solution containing CO = 180 mg·L−1 and various CS (0 mg·L−1, 25,000 mg·L−1, 50,000 mg·L−1, 75,000 mg·L−1 and 100,000 mg·L−1) under a CFV equal to 3.0 m·s−1 and ΔPTM equal to 0.5 bar.
Fig. 1. Schematic drawing of the experiment set-up.
S.E. Weschenfelder et al. / Desalination 360 (2015) 81–86
83
Table 1 Data base for calculating the CAPEX [15–21]. Item
Cost (US$)
Membranes Housings Pumps Peripherals
3609.00/module 1203.00/module 12,671.00/stage 50,000.00/stage
distilled water in order to eliminate the cleaning solution traces. The membrane was considered adequate for subsequent tests when at least 90% of its original permeability was recovered. 2.2. Methodology of the economic analysis
Fig. 2. Permeated flux as a function of CS (CO = 180 mg·L−1, ΔPTM = 0.5 bar, CFV = 3.0 m·s−1).
In subsequent tests, the permeate flux was correlated to ΔPTM for different CO and CFVs, keeping CS constant and equal to zero. After obtaining a steady state permeate flux, the ΔPTM was increased until reaching the limiting permeate flux for each experimental condition. The permeate and concentrate streams were realigned to the feed tank in order to maintain a constant CO. The following parameters were assessed at different levels: CO (180, 500 and 1800 mg·L−1 ± 5%), ΔPTM (0.5, 1.0, 2.0 and 3.0 bar) and CFVs (0.3, 1.0, 2.0, 3.0 m·s−1). The crossflow velocities correspond to a Reynolds number (Re) equal to 1000, 3000, 6000, and 9000, respectively. All experiments were carried out at room temperature (25 °C ±2) for a duration of 60 min. 2.1.4. Membrane regeneration At the end of each experiment, the membrane was regenerated to recover its initial permeability. The system was water washed prior to the chemical cleaning procedure, which involved the use of an aqueous solution composed of 1000 mg·L−1 of sodium hydroxide and 1000 mg·L−1 of sodium hypochlorite. This solution was heated to 70 °C and recirculated during 5 min with the permeate line closed and during 5 more minutes changing the flux direction (backwashing). Once this process was ended, the system was again washed with
The cost estimate was made by taking into consideration a microfiltration system containing several parallel stages, producing a permeate flux of 1000 m3·h−1, operating 24/7, 365 days a year. The limiting permeate flux values have been used to assess the capital expenditures (CAPEX), operating expenses (OPEX) and total cost (TC) per volume of water treated. The influence of the CFV in the operating expenses and capital expenditures were verified based on data obtained from the lab experiments and information found in literature [15–18]. The number of modules was determined as the ratio between the total membrane area required (AT) and the area of a single membrane module (AM). AM was considered as equal to 8.02 m2 and AT as being the ratio between the input flow and the limiting permeate flux (Jlim). Usually, ceramic membranes are used with high flow speeds to reduce the cake layer effect. In industrial systems, in order to reduce the membrane area required, part of the concentrate flow is recycled. Thus, the water recovery rate is kept relatively low between the input and output of the microfiltration modules. The global water recovery rate is obtained from the ratio between the system feed flow and the permeate flow, and may reach high values according to the unit's project. In the methodology adopted for the process economic assessment, oil initial concentration of 90 mg·L−1 and different water recovery rates (50, 80 and 95%) were considered. The recovery rate with a single pass through the modules was kept below 5%, ensuring uniform conditions for the microfiltration process throughout the module. It should be highlighted that this entails high recycling flows, increasing the system's energy demand. 2.2.1. CAPEX The CAPEX, or capital expenditure, was determined by adding up the costs of equipment, valves, pipes and instruments that are part of the process [16,17]. 2.2.1.1. Membranes and housings. The cost related to membranes and housings corresponds approximately to 35% of the total CAPEX. The membrane cost was considered as the average cost given in the literature [17–20]. Most membrane modules are provided with its housing. In this assessment, each membrane module was considered as having one housing, thus, the cost was estimated by multiplying the housing cost by the number of modules to be used in the plant. The housing cost was estimated as 30% of the membrane cost [15,18,21].
Table 2 Data base for calculating the OPEX [15–18].
Fig. 3. Permeated flux evolution overtime for different CFVs (CO = 180 mg·L−1, CS = 0 mg·L−1, ΔPTM = 0.5 bar).
Energy
US$0.05 kW/h
Depreciation Membrane service life Maintenance and manpower Cleaning
7-year linear method 5 years 10% of the fixed capital US$32.00/m2/year
84
S.E. Weschenfelder et al. / Desalination 360 (2015) 81–86
Table 3 Permeate flow obtained experimentally for different CO, CFVs and ΔPTM. CO (mg·L−1)
CFV (m·s−1)
180 180 180 180 500 500 500 500 1800 1800 1800 1800
0.3 1.0 2.0 3.0 0.3 1.0 2.0 3.0 0.3 1.0 2.0 3.0
J (L·h−1·m−2)
J (L·h−1·m−2)
J (L·h−1·m−2)
J (L·h−1·m−2)
ΔPTM = 0.5 bar
ΔPTM = 1.0 bar
ΔPTM = 2.0 bar
ΔPTM = 3.0 bar
128 170 209 214 102 136 167 171 45 75 119 128
173 240 293 291 145 180 229 238 60 100 158 170
200 320 385 405 160 215 310 305 80 111 222 227
210 320 420 438 150 216 328 320 62 135 215 233
2.2.1.2. Peripherals and pumps. For a cost estimate regarding pumps, each stage was considered as having 8 membrane modules and a recycling flow corresponding to each CFV. The total cost was estimated by multiplying the pump cost by the number of stages in the plant [21]. The peripheral cost includes costs with valves, instruments, equipment and pipes [15,17,20]. 2.2.2. OPEX Values related to energy consumption, depreciation, membrane replacement, maintenance, manpower and membrane regeneration were taken into account for operating expenses (OPEX) [15,16]. 2.2.2.1. Energy consumption. The energy required to operate the system (EQ) was defined by Eq. (1). In this equation, EQ is related to the recycling flow (Q), the pressure variation between the module's input and output (ΔP), the total membrane surface area (A) and pump efficiency (η) [15,16]. EQ ¼
ΔP Q : Aη
ð1Þ
The energy cost was determined by multiplying EQ by the unit value for electric energy and by the time of operation, considering the pump efficiency as being at 80%.
is reduced for plants with a low holdup and which only need one brief cleaning cycle per day, using NaOH and/or chlorine [15]. If an independent cleaning system is required, containing pumps, heat exchangers and controls, an amount between US$ 20,000 and US$ 50,000 should be added to the capital expenditure [16,18]. This analysis opted for more conservative values, that is, a complete cleaning system and an annual cost of US$32.00 to regenerate 1 m2 of membrane. 2.2.3. Total cost (TC) The total cost (TC) per volume of treated water produced was obtained from Eq. (2), which takes into account the capital expenditure (CAPEX) and the operating expense (OPEX). RCAPEX, expressed in Eq. (3), is the remuneration over the CAPEX. The interest sum in fraction terms (Σi) is determined by Eq. (4). T is the annual interest rate expressed in %, n is related to the period in years and V is the total volume of treated water produced over one year, expressed in m3. A 20-year period was analyzed. TC ¼ RCAPEX þ OPEX
RCAPEX ¼ X
2.2.2.2. Depreciation. Investment depreciation is normally considered for a period of 7 to 14 years [15,17]. The membrane costs are not included in this estimate and are considered an independent operating expense. Using a 7-year linear depreciation method, the depreciation was calculated by subtracting the membrane cost from the capital expenditure, divided by the period considered in the calculation.
i¼
CAPEX
X
ð2Þ i
Vn
X T n 1þ N: 100
ð3Þ
ð4Þ
2.2.2.3. Membrane replacement. The cost referent to membrane replacement depends on the type of membrane used and its service life. Inorganic membranes are more resistant than polymeric ones, and may resist for a period of 5 to 10 years [15,18]. A 5-year membrane service life was considered for this analysis. 2.2.2.4. Maintenance and manpower. Plants with a treatment capacity under 500 m3·day−1 require only 1 operator per shift. Plants with capacities over 2000 m3·day−1, however, require 1 operator per shift for every 500 m3·day−1. Manpower and maintenance related costs correspond to 2 to 14% of the fixed capital expenditure [15,16]. This assessment considered a nominal rate equivalent to 10% of the fixed capital. 2.2.2.5. Membrane regeneration. An annual expenditure between US$ 1.00 and 32.00 per m2 is estimated for membrane regeneration, considering the use of conventional detergents and cleaning products. The cost
Fig. 4. Limiting permeate flux as a function of CFVs for different CO under ΔPTM = 2.0 bar.
S.E. Weschenfelder et al. / Desalination 360 (2015) 81–86
85
(25,000 mg·L−1, 50,000 mg·L−1 and 100,000 mg·L− 1) under a CFV equal to 3.0 m·s−1. This velocity corresponds to a turbulent flow regime. Fig. 2 presents the results of the permeate flux after 60 min from the start of the test as a function of the salt concentration for CO = 180 mg·L−1, CFV = 3.0 m·s−1. After each experiment the membrane was submitted to a chemical cleaning process and the mean permeability with distilled water was 900 (±60) L·h−1·m−2·bar−1. As can be seen in Fig. 2, the increase in salt concentration caused a decrease of approximately 20% in the permeate flux. This phenomenon may be attributed to an increase in the viscosity, being possible to infer that no changes in the dispersed phase occurred. Zhang et al. [22] found that the permeate flux decreases practically at the same rate as the viscosity increases. However, other authors describe flow changes related to the coalescence in the dispersed phase with an increase in salinity [13,23,24]. The CFV applied in the experiments may have contributed to prevent an intensification of the coalescence in the dispersed phase. In order to simplify the economic assessment, experiments were performed to investigate the influence of CFVs, CO and ΔPTM, without the presence of salts. Fig. 3 presents the permeate flux evolution over time, for different CFVs, maintaining constant the CO = 180 mg·L−1, CS = 0 mg·L−1, ΔPTM = 0.5 bar and T = 25 °C. A sharp decrease can be observed in the permeate flux during the first few minutes of the experiments. After approximately 30 min, under all of the experimental conditions, a steady state permeate flux was almost obtained, which ensures the formation of the cake layer. As expected, the flux in these conditions depends on the crossflow velocity. An increase in the value of the CFV enhance the turbulence and, consequently, reduces the concentration in the dispersed phase near the membrane surface [1,25]. The estimates for the process cost only considered the limiting permeate fluxes. This parameter can be related to the maximum flux observed, near the steady state condition, when ΔPT is increased. Table 3 presents the permeate flux obtained for each CFV, ΔPTM and CO. Fig. 4 presents the behavior of the limiting permeate flux for different CO and CFVs. (See Tables 1 and 2.) The results, when compared under the same operational conditions and CO, presented some similarity with the results reported in the literature [5,26–28]. The permeate flux tends to increase with ΔPTM, but it simultaneously leads to the increased speed of the oil droplets accumulating onto the membrane surface, expanding the effect of the cake layer. Thus, the enhancement of the driving force for the solvent transport is counterbalanced by an increased mass transfer resistance, reaching the limiting permeate flux [29]. According to Fig. 4, it can be observed that the limiting permeate flux is strongly dependent on the CFVs and CO, since a higher concentration of the solute accelerates and increases the cake layer formation, reducing the permeate flux. For CFVs ≥ 2.0 m·s− 1 (Re ≥ 6000), no significant gain to the permeate flux was observed. For the cost analyses at CFV = 4.0 m·s−1, the limiting permeate flux was determined by extrapolating the graphic curves in Fig. 3. It is important to stress that, regardless of effluent characteristics and of the experiment conditions applied, the oil and grease concentration in the permeate stream was always lower than 5 mg·L−1. 3.2. Capital expenditures and operating expenses Fig. 5. CAPEX and OPEX variations as a function of CFVs for ΔPTM = 2.0 bar, CO = 90 mg·L−1 and different water recovery rates: 50% (a), 80% (b) and 95% (c).
3. Results and discussion 3.1. Experimental results The membranes were subjected to experiments with the purpose of checking the salt concentration effect on the permeate flux. The tests were performed with a synthetic effluent containing a fixed concentration of oil and greases (180 mg·L−1) and varying salt concentrations
As previously mentioned, the limiting permeate flux for each CFV and CO was considered for the process economic assessment. The variation in the capital expenditures (CAPEX) and operating expenses (OPEX) for a plant with a permeate generation capacity of 1000 m3·h−1 was related to the CFVs for different oil and grease concentrations, as shown in Fig. 5. As discussed in Section 2.2, the assessment of different CO allows for assessing the effects of different water recovery rates that may be adopted in the operation of a real plant. In the graphics presented in Fig. 5 it can be observed that, for all water recovery rates, the CFV equal to 2.0 m·s−1 promotes a reduction
86
S.E. Weschenfelder et al. / Desalination 360 (2015) 81–86
Table 4 Composition of the capital expenditure and operating expenses. 20-year capital return period and a water recovery rate equal to 95%. Item Annual volume (.106 m3) CAPEX (M US$) RCAPEX (US$/m3) OPEX (US$/m3) TC (US$/m3)
CFV = 0.3 m·s−1
CFV = 1.0 m·s−1
CFV = 2.0 m·s−1
CFV = 3.0 m·s−1
CFV = 4.0 m·s−1
8.76
8.76
8.76
8.76
8.76
23.88 9.71 0.74 10.45
11.35 4.62 0.35 4.97
7.33 2.98 0.23 3.21
7.47 3.04 0.25 3.29
7.58 3.09 0.27 3.36
in both the CAPEX and the OPEX. Using the CFV higher than 3.0 m·s−1 entails a slightly significant gain in process performance, but costs related to the investment in pumps and energy consumption increase considerably. On the other hand, reduced CFV values cause a higher membrane area demand for microfiltration, increasing CAPEX and OPEX. The increased recovery rate also influences greatly the process costs, being most significant in the 80 to 95% range. 3.3. Total costs Table 4 summarizes OPEX- and CAPEX-related costs for CFVs, considering a 20-year investment return period and a water recovery rate of 95%. The total cost was determined as a function of the unit volume of treated water in m3. As previously discussed, it can be noted in Table 4 that a CFV equal to 2.0 m·s− 1 leads to a reduction in the total costs of the processes. It should be highlighted that different scale mitigation strategies such as, for example, the use of back pulses and backwashing, have not been taken into consideration in this assessment. These strategies may reduce the process global cost. 4. Conclusions The ceramic membranes used in the experiments were capable of treating the oilfield produced water synthetically prepared. It was also able to generate a permeate stream with an oil and grease content lower than 5 mg·L−1, which enables the reuse of water. The performance assessment of the microfiltration process showed that the crossflow velocity and the oil and grease concentration were determining factors in the behavior of the permeate flux. The concentration of salts, however, did not significantly influence the performance observed. Limiting permeate flux values obtained in conditions close to the established regime have been used for the economic assessment of the microfiltration process in different CFVs and water recovery rates. The results of this assessment show that the operating expenses and capital expenditures can be minimized by using a CFV equal to 2.0 m·s−1. An increase of US$0.10/m3 was observed in the OPEX and a 55% boost in the value referring to the CAPEX, by increasing the water recovery rate from 80% to 95%. It was found that under an optimal crossflow velocity (i.e. CFV = 2.0 m·s−1) and considering the water recovery rate equal to 95%, the cost related to OPEX and the total cost were, respectively, US$0.23/m3 and US$3.21/m3. The total CAPEX for a full scale plant, capable of treating 1000 m3·h− 1 of produced water, was estimated at MUS$7.33. References [1] S.H.D. Silahi, T. Leiknes, Cleaning strategies in ceramic microfiltration membranes fouled by oil and particulate matter in produced water, Desalination 236 (2009) 160–169. [2] E.T. Igunnu, G.Z. Chen, Produced water treatment technologies, Int. J. Low Carbon Technol. (2012) 1–21. [3] N.I. Brasil, M.A.A. Araújo, E.C.M. De Souza, Processamento primário de petróleo e gás, LTC, Rio de Janeiro, 2011. 299.
[4] J. Zhong, X. Sun, C. Wang, Treatment of oily wastewater produced from refinery processes using flocculation and ceramic membrane filtration, Sep. Purif. Technol. 32 (2003). [5] M. Çakmakçe, N. Kayaalp, I. Koyuncu, Desalination of produced water from oil production fields by membrane processes, Desalination 222 (2008). [6] M. Ebrahimi, D. Willershausen, K.S. Ashaghi, L. Engel, L. Placido, P. Mund, P. Bolduan, P. Czermak, Investigations on the use of different ceramic membranes for efficient oil-field produced water treatment, Desalination 250 (2010). [7] A. Fakhru'l-Razi, A.R. Pendashteh, A.L. Chuah, A.B.D. Radiah, S.S. Madaeni, Z.A. Zurina, Review of technologies for oil and gas produced water treatment, J. Hazard. Mater. 170 (2009) 530–551. [8] A.L.S. Souza, C.J.A. Furtado, Reinjeção de água produzida na Petrobras: perspectivas e desafios, Bol. Téc. Prod. Petról. Rio J. 1 (n.2) (2006). [9] L. Li, R. Lee, Purification of produced water by ceramic membranes: material screening, process design and economics, Sep. Sci. Technol. 44 (15) (2009) 3455–3484. [10] C. Yang, G. Zhang, N. Xu, J. Shi, Preparation and application in oil–water separation of ZrO2/Al2O3 MF membrane, J. Membr. Sci. 142 (1998) 235–243. [11] P. Wang, N. Xu, J. Shi, A pilot study of the treatment of waste rolling emulsion using zirconia microfiltration membranes, J. Membr. Sci. 173 (2000) 159–166. [12] S.M. Kwan, K.L. Yeung, Zeolit micro-fuel cell, Chem. Commun. 31 (2008) 3631–3633. [13] F.L. Hua, Y.F. Tsang, Y.J. Wang, S.Y. Chan, H. Chua, S.N. Sin, Performance study of ceramic microfiltration membrane for oil wastewater treatment, Chem. Eng. J. 128 (2007) 169–175. [14] P. Srijaroonrat, E. Julien, Y. Aurelle, Unstable secondary oil/water emulsion treatment using ultrafiltration: fouling control by backflushing, J. Membr. Sci. 159 (1999) 11–20. [15] M. Cheryan, Ultrafiltration and Microfiltration Handbook, Technomic, Pennsylvania, 1998. [16] N. Singh, M. Cheryan, Process design and economic analysis of a ceramic membrane system for microfiltration of corn starch hydrolysate, J. Food Eng. 38 (1998) 57–67. [17] E. Salehi, S.S. Madaeni, S. Laki, Aplicabilility of ceramic membrane filters in pretreatment of coke-contaminated petrochemical wastewater: economic feasibility study, Ceram. Int. 40 (2014) 4805–4810. [18] H. Krawczyk, A.S. Jonsson, The influence of feed flow channel diameter on frictional pressure drop, membrane performance and process cost in full-scale tubular ceramic membranes, Chem. Eng. Res. Des. 92 (2014). [19] P. Sanciolo, S. Gray, Outcomes of Membrane Research: Impacts on the Australian and International Water Industry, Water Research Australia, 2014. [20] K. Guerra, J. Pellegrino, Investigation of low-pressure membrane performance, cleaning, and economics using a techno-economic modeling approach, Science and Technology Program Report, 174, 2012. [21] H.P. Loh, J. Lyons, C.W. White III, Process Equipment Cost Estimation Final Report, NETL, US Department of Energy, USA, 2002. [22] H. Zhang, Z. Shong, W. Xing, Application of ceramic membranes in the treatment of oilfield-produced water: effects of polyacrylamide and inorganic salts, Desalination 309 (2013). [23] M.C. Sterling, T. Ojo, R.L. Autenrieth, J.S. Bonner, C.A. Page, A.N.S. Ernest, Coalescence kinetics of dispersed crude oil in a laboratory reactor, Arctic and Marine Oilspill Program Technical Seminar, vol. 2, Environment Canada, 1999, pp. 741–754. [24] C.K. Schulz, Tratamento de efluentes oleosos utilizando processos de separação por membrana(Doctoral dissertation) Universidade Federal Do Rio De Janeiro, 2005. [25] T. Mohammadi, M. Kazemimoghadam, M. Saadabadi, Modeling of membrane fouling and decline in reverse osmosis during separation of oil in water emulsions, Desalination 157 (2003) 369–375. [26] T. Mohammadi, A. Pak, M. Karbassian, M. Golshan, Effect of operating conditions on microfiltration of an oil–water emulsion by a kaolin membrane, Desalination 168 (2004) 201–205. [27] S.R.H. Abadi, M.R. Sebzari, M. Hemati, F. Rekabdar, T. Mohammadi, Ceramic membrane performance in microfiltration of oily wastewater, Desalination 265 (2011). [28] M. Ebrahimi, G. Lavi, T. Schmidts, F. Runkel, P. Czermak, Development and production of oil-in-water vehicles sub-micron emulsion using tubular ceramic membranes, Desalination 224 (2008) 40–45. [29] M. Mohammadi, M. Kazemi, S.S. Madaeni, Hydrodynamic factors affecting flux and fouling during reverse osmosis of seawater, Desalination 151 (2002) 239–245.
Contents lists available at ScienceDirect
Desalination journal homepage: www.elsevier.com/locate/desal
Oilfield produced water treatment by ceramic membranes: Preliminary process cost estimation S.E. Weschenfelder a,b,⁎, A.C.C. Mello a,b, C.P. Borges c, J.C. Campos b a b c
Technology for Water Treatment and Reuse Department, Petrobras Research Center, Rio de Janeiro, Brazil School of Chemistry, Inorganic Processes Department, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil COPPE/Chemical Engineer Program, Federal University of Rio de Janeiro Rio de Janeiro, Brazil
H I G H L I G H T S • • • • •
The microfiltration process enables the treatment and reuse of produced water. The process costs have been related to cross flow velocity and water recovery. OPEX and CAPEX have been estimated for a real-scale plant under optimal conditions. The OPEX was found to be equal to US$0.23/m3. The CAPEX for a full scale plant was estimated at MUS$7.33.
a r t i c l e
i n f o
Article history: Received 17 November 2014 Received in revised form 11 January 2015 Accepted 12 January 2015 Available online xxxx Keywords: Produced water Treatment Ceramic membranes Cost estimation Process
a b s t r a c t The application of ceramic membranes for oilfield produced water treatment has been considered a very promising technology, mainly due to the oil and grease separation efficiency and process robustness. The purpose of this study was to obtain a preliminary estimate of operating expenses (OPEX) and capital expenditures (CAPEX) for a full scale ceramic membrane plant, based on data obtained in lab scale tests and from literature information. Different crossflow velocities (CFVs) and water recovery rates were simulated and the results were correlated to the OPEX, CAPEX and total cost (TC) per cubic meter of treated effluent. An increase of US $0.10/m3 in the OPEX and a 55% boost in the value referring to the CAPEX, by increasing the water recovery rate from 80% to 95% were observed. It was found that, under an optimal CFV (2.0 m·s−1) and considering the water recovery rate equal to 95%, the cost related to OPEX and TC were, respectively, US$0.23/m3 and US$3.21/ m3. The CAPEX for a full scale plant, capable of treating 1000 m3·h−1 of produced water, was estimated at MUS$7.33. In all of the experimental conditions assessed, it was possible to generate a permeate stream with oil and grease content (CO) lower than 5 mg·L−1. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Oilfield produced water is a byproduct of the oil extraction process from subsurface geological formations. Its composition typically includes the presence of oil, dissolved organic compounds and inorganic particles [1]. Depending on the oil producing field, the produced water may present oil and grease content higher than 500 mg·L−1 and salt concentrations (CS) between 80 and 200,000 mg·L− 1. In many cases this effluent is reused for the purposes of irrigation, reinjection into reservoir aiming to enhance oil recovery, or also for steam generation through the application of subsequent desalination processes. However,
⁎ Corresponding author at: Technology for Water Treatment and Reuse Department, Petrobras Research Center, Rio de Janeiro, Brazil. E-mail address: [email protected] (S.E. Weschenfelder).
http://dx.doi.org/10.1016/j.desal.2015.01.015 0011-9164/© 2015 Elsevier B.V. All rights reserved.
CO higher than 5 mg·L−1 may compromise the injection of water in the reservoir, as well as the efficiency of the salt removal processes [2–6]. Conventional produced water treatment processes may include flotation systems, hydrocyclones, and nutshell or mixed media filters. These types of equipment, however, have a reduced efficiency in removing solids and oil and grease particles whose dimensions are smaller than 5.0 μm, making it difficult to generate an effluent that is appropriate for reuse [2,7–9]. The membrane separation processes can be presented as an alternative technology to the conventional processes used in treating oilfield produced water. Ceramic membranes have been taken into consideration for presenting advantages connected to its greater mechanical, chemical and thermal resistance, in addition to its efficiency in removing oil and grease from streams with a high load of solids and oily contaminants, without the addition of chemical products [2]. The membranes used for that purpose may be produced from different materials,
82
S.E. Weschenfelder et al. / Desalination 360 (2015) 81–86
among which zirconium oxide, aluminum oxide and titanium oxide are highlighted. Many studies have been made comparing the performance of different inorganic materials for treating oily effluents. The reported results show that zirconium oxide membranes provide slightly higher fluxes than aluminum oxide and titanium oxide membranes [10–12]. Also different pore sizes have also been compared to oily water treatment. Hua et al. [13] reported that the use of membranes with pore sizes of 0.2 μm was not appropriate for high quality effluent. Similar results were also obtained by Srijaroonrat et al. [14] and Ebrahimi et al. [6]. They also concluded that the use of membranes with pore sizes of 0.1 μm gives the best results in terms of flux compared to the pore sizes of 0.05 μm and 0.5 μm. The purpose of this work was to assess the operating expenses and capital expenditures involved in a full scale ceramic membrane plant for the treatment of oilfield produced water. The performance of the microfiltration process was assessed in laboratory scale experiments using a synthetically-prepared effluent and simulating real operating situations. 2. Materials and methods 2.1. Membrane characteristics To perform the experiments, commercial zirconium oxide (ZrO2) membranes were evaluated. The membranes have 19 channels of 3 mm, microfiltration area equivalent to 0.0381 m2 and mean pore sizes of 0.1 μm. According to the manufacturer, Likuid Nanotek, these membranes can withstand pressures up to 8 bar, pH values between 0 and 14 and temperatures up to 100 °C. 2.1.1. Experiment system The experimental system consisted of a membrane module, a recirculation pump, pressure gauges and flow meters, flow control valves in the feed, permeate and concentrate streams, one tank to
collect the permeate and a heated and mechanically agitated feed tank. The permeate mass was continuously monitored by data acquisition. A schematic representation of the experiment set-up is shown in Fig. 1. During testing, the valves V-1 and V-2 remained opened and the feed flow rate was controlled by the frequency inverter connected to the pump B-1. The ΔPTM was adjusted through valve V-3 and calculated as the difference between the mean pressure given by PI-1 and PI-2 and the pressure given by PI-3. 2.1.2. Synthetic effluents The effluent used in the tests was synthetically prepared with distilled water, salt (sodium chloride) and oil. The oil was added to the saline mixture and immediately emulsified with a Turrax mixer (Ultra-Turrax T-50). The emulsion was deemed stable when the oil droplet mean size, as measured by a particle size analyzer (Malvern Mastersizer Micro) ranged from 10 to 30 μm. The oil had a density equal to 28°API and was obtained directly from an offshore oil production unit. Its concentration was determined by an infrared spectrophotometer (Horiba OCMA-350). 2.1.3. Experiments Before starting each experiment and after the chemical cleaning procedure, the membrane hydraulic permeability was determined. For such measurement, distilled water was applied and the permeate flux was recorded, under a turbulent flow (CFV = 3.0 m·s− 1) at 25 °C in different ΔPTM: 1.0, 2.0, and 3.0 bar. The hydraulic permeability was considered as the slope of a linear correlation between the permeate flux and ΔPTM. To assess the influence of salinity on the permeate flux, experiments were carried out using a synthetic solution containing CO = 180 mg·L−1 and various CS (0 mg·L−1, 25,000 mg·L−1, 50,000 mg·L−1, 75,000 mg·L−1 and 100,000 mg·L−1) under a CFV equal to 3.0 m·s−1 and ΔPTM equal to 0.5 bar.
Fig. 1. Schematic drawing of the experiment set-up.
S.E. Weschenfelder et al. / Desalination 360 (2015) 81–86
83
Table 1 Data base for calculating the CAPEX [15–21]. Item
Cost (US$)
Membranes Housings Pumps Peripherals
3609.00/module 1203.00/module 12,671.00/stage 50,000.00/stage
distilled water in order to eliminate the cleaning solution traces. The membrane was considered adequate for subsequent tests when at least 90% of its original permeability was recovered. 2.2. Methodology of the economic analysis
Fig. 2. Permeated flux as a function of CS (CO = 180 mg·L−1, ΔPTM = 0.5 bar, CFV = 3.0 m·s−1).
In subsequent tests, the permeate flux was correlated to ΔPTM for different CO and CFVs, keeping CS constant and equal to zero. After obtaining a steady state permeate flux, the ΔPTM was increased until reaching the limiting permeate flux for each experimental condition. The permeate and concentrate streams were realigned to the feed tank in order to maintain a constant CO. The following parameters were assessed at different levels: CO (180, 500 and 1800 mg·L−1 ± 5%), ΔPTM (0.5, 1.0, 2.0 and 3.0 bar) and CFVs (0.3, 1.0, 2.0, 3.0 m·s−1). The crossflow velocities correspond to a Reynolds number (Re) equal to 1000, 3000, 6000, and 9000, respectively. All experiments were carried out at room temperature (25 °C ±2) for a duration of 60 min. 2.1.4. Membrane regeneration At the end of each experiment, the membrane was regenerated to recover its initial permeability. The system was water washed prior to the chemical cleaning procedure, which involved the use of an aqueous solution composed of 1000 mg·L−1 of sodium hydroxide and 1000 mg·L−1 of sodium hypochlorite. This solution was heated to 70 °C and recirculated during 5 min with the permeate line closed and during 5 more minutes changing the flux direction (backwashing). Once this process was ended, the system was again washed with
The cost estimate was made by taking into consideration a microfiltration system containing several parallel stages, producing a permeate flux of 1000 m3·h−1, operating 24/7, 365 days a year. The limiting permeate flux values have been used to assess the capital expenditures (CAPEX), operating expenses (OPEX) and total cost (TC) per volume of water treated. The influence of the CFV in the operating expenses and capital expenditures were verified based on data obtained from the lab experiments and information found in literature [15–18]. The number of modules was determined as the ratio between the total membrane area required (AT) and the area of a single membrane module (AM). AM was considered as equal to 8.02 m2 and AT as being the ratio between the input flow and the limiting permeate flux (Jlim). Usually, ceramic membranes are used with high flow speeds to reduce the cake layer effect. In industrial systems, in order to reduce the membrane area required, part of the concentrate flow is recycled. Thus, the water recovery rate is kept relatively low between the input and output of the microfiltration modules. The global water recovery rate is obtained from the ratio between the system feed flow and the permeate flow, and may reach high values according to the unit's project. In the methodology adopted for the process economic assessment, oil initial concentration of 90 mg·L−1 and different water recovery rates (50, 80 and 95%) were considered. The recovery rate with a single pass through the modules was kept below 5%, ensuring uniform conditions for the microfiltration process throughout the module. It should be highlighted that this entails high recycling flows, increasing the system's energy demand. 2.2.1. CAPEX The CAPEX, or capital expenditure, was determined by adding up the costs of equipment, valves, pipes and instruments that are part of the process [16,17]. 2.2.1.1. Membranes and housings. The cost related to membranes and housings corresponds approximately to 35% of the total CAPEX. The membrane cost was considered as the average cost given in the literature [17–20]. Most membrane modules are provided with its housing. In this assessment, each membrane module was considered as having one housing, thus, the cost was estimated by multiplying the housing cost by the number of modules to be used in the plant. The housing cost was estimated as 30% of the membrane cost [15,18,21].
Table 2 Data base for calculating the OPEX [15–18].
Fig. 3. Permeated flux evolution overtime for different CFVs (CO = 180 mg·L−1, CS = 0 mg·L−1, ΔPTM = 0.5 bar).
Energy
US$0.05 kW/h
Depreciation Membrane service life Maintenance and manpower Cleaning
7-year linear method 5 years 10% of the fixed capital US$32.00/m2/year
84
S.E. Weschenfelder et al. / Desalination 360 (2015) 81–86
Table 3 Permeate flow obtained experimentally for different CO, CFVs and ΔPTM. CO (mg·L−1)
CFV (m·s−1)
180 180 180 180 500 500 500 500 1800 1800 1800 1800
0.3 1.0 2.0 3.0 0.3 1.0 2.0 3.0 0.3 1.0 2.0 3.0
J (L·h−1·m−2)
J (L·h−1·m−2)
J (L·h−1·m−2)
J (L·h−1·m−2)
ΔPTM = 0.5 bar
ΔPTM = 1.0 bar
ΔPTM = 2.0 bar
ΔPTM = 3.0 bar
128 170 209 214 102 136 167 171 45 75 119 128
173 240 293 291 145 180 229 238 60 100 158 170
200 320 385 405 160 215 310 305 80 111 222 227
210 320 420 438 150 216 328 320 62 135 215 233
2.2.1.2. Peripherals and pumps. For a cost estimate regarding pumps, each stage was considered as having 8 membrane modules and a recycling flow corresponding to each CFV. The total cost was estimated by multiplying the pump cost by the number of stages in the plant [21]. The peripheral cost includes costs with valves, instruments, equipment and pipes [15,17,20]. 2.2.2. OPEX Values related to energy consumption, depreciation, membrane replacement, maintenance, manpower and membrane regeneration were taken into account for operating expenses (OPEX) [15,16]. 2.2.2.1. Energy consumption. The energy required to operate the system (EQ) was defined by Eq. (1). In this equation, EQ is related to the recycling flow (Q), the pressure variation between the module's input and output (ΔP), the total membrane surface area (A) and pump efficiency (η) [15,16]. EQ ¼
ΔP Q : Aη
ð1Þ
The energy cost was determined by multiplying EQ by the unit value for electric energy and by the time of operation, considering the pump efficiency as being at 80%.
is reduced for plants with a low holdup and which only need one brief cleaning cycle per day, using NaOH and/or chlorine [15]. If an independent cleaning system is required, containing pumps, heat exchangers and controls, an amount between US$ 20,000 and US$ 50,000 should be added to the capital expenditure [16,18]. This analysis opted for more conservative values, that is, a complete cleaning system and an annual cost of US$32.00 to regenerate 1 m2 of membrane. 2.2.3. Total cost (TC) The total cost (TC) per volume of treated water produced was obtained from Eq. (2), which takes into account the capital expenditure (CAPEX) and the operating expense (OPEX). RCAPEX, expressed in Eq. (3), is the remuneration over the CAPEX. The interest sum in fraction terms (Σi) is determined by Eq. (4). T is the annual interest rate expressed in %, n is related to the period in years and V is the total volume of treated water produced over one year, expressed in m3. A 20-year period was analyzed. TC ¼ RCAPEX þ OPEX
RCAPEX ¼ X
2.2.2.2. Depreciation. Investment depreciation is normally considered for a period of 7 to 14 years [15,17]. The membrane costs are not included in this estimate and are considered an independent operating expense. Using a 7-year linear depreciation method, the depreciation was calculated by subtracting the membrane cost from the capital expenditure, divided by the period considered in the calculation.
i¼
CAPEX
X
ð2Þ i
Vn
X T n 1þ N: 100
ð3Þ
ð4Þ
2.2.2.3. Membrane replacement. The cost referent to membrane replacement depends on the type of membrane used and its service life. Inorganic membranes are more resistant than polymeric ones, and may resist for a period of 5 to 10 years [15,18]. A 5-year membrane service life was considered for this analysis. 2.2.2.4. Maintenance and manpower. Plants with a treatment capacity under 500 m3·day−1 require only 1 operator per shift. Plants with capacities over 2000 m3·day−1, however, require 1 operator per shift for every 500 m3·day−1. Manpower and maintenance related costs correspond to 2 to 14% of the fixed capital expenditure [15,16]. This assessment considered a nominal rate equivalent to 10% of the fixed capital. 2.2.2.5. Membrane regeneration. An annual expenditure between US$ 1.00 and 32.00 per m2 is estimated for membrane regeneration, considering the use of conventional detergents and cleaning products. The cost
Fig. 4. Limiting permeate flux as a function of CFVs for different CO under ΔPTM = 2.0 bar.
S.E. Weschenfelder et al. / Desalination 360 (2015) 81–86
85
(25,000 mg·L−1, 50,000 mg·L−1 and 100,000 mg·L− 1) under a CFV equal to 3.0 m·s−1. This velocity corresponds to a turbulent flow regime. Fig. 2 presents the results of the permeate flux after 60 min from the start of the test as a function of the salt concentration for CO = 180 mg·L−1, CFV = 3.0 m·s−1. After each experiment the membrane was submitted to a chemical cleaning process and the mean permeability with distilled water was 900 (±60) L·h−1·m−2·bar−1. As can be seen in Fig. 2, the increase in salt concentration caused a decrease of approximately 20% in the permeate flux. This phenomenon may be attributed to an increase in the viscosity, being possible to infer that no changes in the dispersed phase occurred. Zhang et al. [22] found that the permeate flux decreases practically at the same rate as the viscosity increases. However, other authors describe flow changes related to the coalescence in the dispersed phase with an increase in salinity [13,23,24]. The CFV applied in the experiments may have contributed to prevent an intensification of the coalescence in the dispersed phase. In order to simplify the economic assessment, experiments were performed to investigate the influence of CFVs, CO and ΔPTM, without the presence of salts. Fig. 3 presents the permeate flux evolution over time, for different CFVs, maintaining constant the CO = 180 mg·L−1, CS = 0 mg·L−1, ΔPTM = 0.5 bar and T = 25 °C. A sharp decrease can be observed in the permeate flux during the first few minutes of the experiments. After approximately 30 min, under all of the experimental conditions, a steady state permeate flux was almost obtained, which ensures the formation of the cake layer. As expected, the flux in these conditions depends on the crossflow velocity. An increase in the value of the CFV enhance the turbulence and, consequently, reduces the concentration in the dispersed phase near the membrane surface [1,25]. The estimates for the process cost only considered the limiting permeate fluxes. This parameter can be related to the maximum flux observed, near the steady state condition, when ΔPT is increased. Table 3 presents the permeate flux obtained for each CFV, ΔPTM and CO. Fig. 4 presents the behavior of the limiting permeate flux for different CO and CFVs. (See Tables 1 and 2.) The results, when compared under the same operational conditions and CO, presented some similarity with the results reported in the literature [5,26–28]. The permeate flux tends to increase with ΔPTM, but it simultaneously leads to the increased speed of the oil droplets accumulating onto the membrane surface, expanding the effect of the cake layer. Thus, the enhancement of the driving force for the solvent transport is counterbalanced by an increased mass transfer resistance, reaching the limiting permeate flux [29]. According to Fig. 4, it can be observed that the limiting permeate flux is strongly dependent on the CFVs and CO, since a higher concentration of the solute accelerates and increases the cake layer formation, reducing the permeate flux. For CFVs ≥ 2.0 m·s− 1 (Re ≥ 6000), no significant gain to the permeate flux was observed. For the cost analyses at CFV = 4.0 m·s−1, the limiting permeate flux was determined by extrapolating the graphic curves in Fig. 3. It is important to stress that, regardless of effluent characteristics and of the experiment conditions applied, the oil and grease concentration in the permeate stream was always lower than 5 mg·L−1. 3.2. Capital expenditures and operating expenses Fig. 5. CAPEX and OPEX variations as a function of CFVs for ΔPTM = 2.0 bar, CO = 90 mg·L−1 and different water recovery rates: 50% (a), 80% (b) and 95% (c).
3. Results and discussion 3.1. Experimental results The membranes were subjected to experiments with the purpose of checking the salt concentration effect on the permeate flux. The tests were performed with a synthetic effluent containing a fixed concentration of oil and greases (180 mg·L−1) and varying salt concentrations
As previously mentioned, the limiting permeate flux for each CFV and CO was considered for the process economic assessment. The variation in the capital expenditures (CAPEX) and operating expenses (OPEX) for a plant with a permeate generation capacity of 1000 m3·h−1 was related to the CFVs for different oil and grease concentrations, as shown in Fig. 5. As discussed in Section 2.2, the assessment of different CO allows for assessing the effects of different water recovery rates that may be adopted in the operation of a real plant. In the graphics presented in Fig. 5 it can be observed that, for all water recovery rates, the CFV equal to 2.0 m·s−1 promotes a reduction
86
S.E. Weschenfelder et al. / Desalination 360 (2015) 81–86
Table 4 Composition of the capital expenditure and operating expenses. 20-year capital return period and a water recovery rate equal to 95%. Item Annual volume (.106 m3) CAPEX (M US$) RCAPEX (US$/m3) OPEX (US$/m3) TC (US$/m3)
CFV = 0.3 m·s−1
CFV = 1.0 m·s−1
CFV = 2.0 m·s−1
CFV = 3.0 m·s−1
CFV = 4.0 m·s−1
8.76
8.76
8.76
8.76
8.76
23.88 9.71 0.74 10.45
11.35 4.62 0.35 4.97
7.33 2.98 0.23 3.21
7.47 3.04 0.25 3.29
7.58 3.09 0.27 3.36
in both the CAPEX and the OPEX. Using the CFV higher than 3.0 m·s−1 entails a slightly significant gain in process performance, but costs related to the investment in pumps and energy consumption increase considerably. On the other hand, reduced CFV values cause a higher membrane area demand for microfiltration, increasing CAPEX and OPEX. The increased recovery rate also influences greatly the process costs, being most significant in the 80 to 95% range. 3.3. Total costs Table 4 summarizes OPEX- and CAPEX-related costs for CFVs, considering a 20-year investment return period and a water recovery rate of 95%. The total cost was determined as a function of the unit volume of treated water in m3. As previously discussed, it can be noted in Table 4 that a CFV equal to 2.0 m·s− 1 leads to a reduction in the total costs of the processes. It should be highlighted that different scale mitigation strategies such as, for example, the use of back pulses and backwashing, have not been taken into consideration in this assessment. These strategies may reduce the process global cost. 4. Conclusions The ceramic membranes used in the experiments were capable of treating the oilfield produced water synthetically prepared. It was also able to generate a permeate stream with an oil and grease content lower than 5 mg·L−1, which enables the reuse of water. The performance assessment of the microfiltration process showed that the crossflow velocity and the oil and grease concentration were determining factors in the behavior of the permeate flux. The concentration of salts, however, did not significantly influence the performance observed. Limiting permeate flux values obtained in conditions close to the established regime have been used for the economic assessment of the microfiltration process in different CFVs and water recovery rates. The results of this assessment show that the operating expenses and capital expenditures can be minimized by using a CFV equal to 2.0 m·s−1. An increase of US$0.10/m3 was observed in the OPEX and a 55% boost in the value referring to the CAPEX, by increasing the water recovery rate from 80% to 95%. It was found that under an optimal crossflow velocity (i.e. CFV = 2.0 m·s−1) and considering the water recovery rate equal to 95%, the cost related to OPEX and the total cost were, respectively, US$0.23/m3 and US$3.21/m3. The total CAPEX for a full scale plant, capable of treating 1000 m3·h− 1 of produced water, was estimated at MUS$7.33. References [1] S.H.D. Silahi, T. Leiknes, Cleaning strategies in ceramic microfiltration membranes fouled by oil and particulate matter in produced water, Desalination 236 (2009) 160–169. [2] E.T. Igunnu, G.Z. Chen, Produced water treatment technologies, Int. J. Low Carbon Technol. (2012) 1–21. [3] N.I. Brasil, M.A.A. Araújo, E.C.M. De Souza, Processamento primário de petróleo e gás, LTC, Rio de Janeiro, 2011. 299.
[4] J. Zhong, X. Sun, C. Wang, Treatment of oily wastewater produced from refinery processes using flocculation and ceramic membrane filtration, Sep. Purif. Technol. 32 (2003). [5] M. Çakmakçe, N. Kayaalp, I. Koyuncu, Desalination of produced water from oil production fields by membrane processes, Desalination 222 (2008). [6] M. Ebrahimi, D. Willershausen, K.S. Ashaghi, L. Engel, L. Placido, P. Mund, P. Bolduan, P. Czermak, Investigations on the use of different ceramic membranes for efficient oil-field produced water treatment, Desalination 250 (2010). [7] A. Fakhru'l-Razi, A.R. Pendashteh, A.L. Chuah, A.B.D. Radiah, S.S. Madaeni, Z.A. Zurina, Review of technologies for oil and gas produced water treatment, J. Hazard. Mater. 170 (2009) 530–551. [8] A.L.S. Souza, C.J.A. Furtado, Reinjeção de água produzida na Petrobras: perspectivas e desafios, Bol. Téc. Prod. Petról. Rio J. 1 (n.2) (2006). [9] L. Li, R. Lee, Purification of produced water by ceramic membranes: material screening, process design and economics, Sep. Sci. Technol. 44 (15) (2009) 3455–3484. [10] C. Yang, G. Zhang, N. Xu, J. Shi, Preparation and application in oil–water separation of ZrO2/Al2O3 MF membrane, J. Membr. Sci. 142 (1998) 235–243. [11] P. Wang, N. Xu, J. Shi, A pilot study of the treatment of waste rolling emulsion using zirconia microfiltration membranes, J. Membr. Sci. 173 (2000) 159–166. [12] S.M. Kwan, K.L. Yeung, Zeolit micro-fuel cell, Chem. Commun. 31 (2008) 3631–3633. [13] F.L. Hua, Y.F. Tsang, Y.J. Wang, S.Y. Chan, H. Chua, S.N. Sin, Performance study of ceramic microfiltration membrane for oil wastewater treatment, Chem. Eng. J. 128 (2007) 169–175. [14] P. Srijaroonrat, E. Julien, Y. Aurelle, Unstable secondary oil/water emulsion treatment using ultrafiltration: fouling control by backflushing, J. Membr. Sci. 159 (1999) 11–20. [15] M. Cheryan, Ultrafiltration and Microfiltration Handbook, Technomic, Pennsylvania, 1998. [16] N. Singh, M. Cheryan, Process design and economic analysis of a ceramic membrane system for microfiltration of corn starch hydrolysate, J. Food Eng. 38 (1998) 57–67. [17] E. Salehi, S.S. Madaeni, S. Laki, Aplicabilility of ceramic membrane filters in pretreatment of coke-contaminated petrochemical wastewater: economic feasibility study, Ceram. Int. 40 (2014) 4805–4810. [18] H. Krawczyk, A.S. Jonsson, The influence of feed flow channel diameter on frictional pressure drop, membrane performance and process cost in full-scale tubular ceramic membranes, Chem. Eng. Res. Des. 92 (2014). [19] P. Sanciolo, S. Gray, Outcomes of Membrane Research: Impacts on the Australian and International Water Industry, Water Research Australia, 2014. [20] K. Guerra, J. Pellegrino, Investigation of low-pressure membrane performance, cleaning, and economics using a techno-economic modeling approach, Science and Technology Program Report, 174, 2012. [21] H.P. Loh, J. Lyons, C.W. White III, Process Equipment Cost Estimation Final Report, NETL, US Department of Energy, USA, 2002. [22] H. Zhang, Z. Shong, W. Xing, Application of ceramic membranes in the treatment of oilfield-produced water: effects of polyacrylamide and inorganic salts, Desalination 309 (2013). [23] M.C. Sterling, T. Ojo, R.L. Autenrieth, J.S. Bonner, C.A. Page, A.N.S. Ernest, Coalescence kinetics of dispersed crude oil in a laboratory reactor, Arctic and Marine Oilspill Program Technical Seminar, vol. 2, Environment Canada, 1999, pp. 741–754. [24] C.K. Schulz, Tratamento de efluentes oleosos utilizando processos de separação por membrana(Doctoral dissertation) Universidade Federal Do Rio De Janeiro, 2005. [25] T. Mohammadi, M. Kazemimoghadam, M. Saadabadi, Modeling of membrane fouling and decline in reverse osmosis during separation of oil in water emulsions, Desalination 157 (2003) 369–375. [26] T. Mohammadi, A. Pak, M. Karbassian, M. Golshan, Effect of operating conditions on microfiltration of an oil–water emulsion by a kaolin membrane, Desalination 168 (2004) 201–205. [27] S.R.H. Abadi, M.R. Sebzari, M. Hemati, F. Rekabdar, T. Mohammadi, Ceramic membrane performance in microfiltration of oily wastewater, Desalination 265 (2011). [28] M. Ebrahimi, G. Lavi, T. Schmidts, F. Runkel, P. Czermak, Development and production of oil-in-water vehicles sub-micron emulsion using tubular ceramic membranes, Desalination 224 (2008) 40–45. [29] M. Mohammadi, M. Kazemi, S.S. Madaeni, Hydrodynamic factors affecting flux and fouling during reverse osmosis of seawater, Desalination 151 (2002) 239–245.