- Email: [email protected]
Treatment of vegetable oil refinery wastewater using alumina ceramic membrane: optimization using response surface methodology
Journal of Cleaner Production 112 (2016) 3132e3137
Contents lists available at ScienceDirect
Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro
Treatment of vegetable oil refinery wastewater using alumina ceramic membrane: optimization using response surface methodology s a, Nikola Maravi Zita Sere c a, *, Aleksandar Taka ci a, Ivana Nikoli c a, c a, Cecilia Hodur b Dragana Soronja-Simovi c a, Aleksandar Joki a b
University of Novi Sad, Faculty of Technology, Bul. cara Lazara 1, 21000 Novi Sad, Serbia University of Szeged, Faculty of Engineering, Szeged, Mars t er 7, 6724 Szeged, Hungary
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 March 2015 Received in revised form 12 October 2015 Accepted 18 October 2015 Available online 26 October 2015
New regulations in environment protection and increasing market demands for “green” companies are forcing the industry to consider finding new and sustainable methods of wastewater treatment. The valuable information regarding treatment of the wastewater from the edible oil industry is presented in this research. The obtained results confirm a promising application of the third generation membranes, comprised of ceramic material (aluminium oxide), in treatment of wastewater from the edible oil industry. Positive results regarding chemical oxygen demand reduction and turbidity removal were noticed. Permeate flux values were used for process optimization in order to achieve a cost-effective process. Response surface methodology was used for the experimental design. The effects of wastewater temperature, transmembrane pressure, and feed-flow rate on the microfiltration model fit were studied. The experiments showed that microfiltration of this type of wastewater is a convenient technique for a possible large scale industrial application as a secondary step in treating wastewater from the oilseed processing facilities. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Microfiltration Chemical oxygen demand Wastewater Edible oil Response surface methodology
1. Introduction The main task of wastewater treatment is to eliminate pollution to the extent that the treated wastewater can be discharged into the recipient without consequences or that it can be re-used (Kralj, 2015). The technological process of wastewater treatment can consist of many processing stages, depending on the characteristics of the raw wastewater and the required quality of the treated wastewater, and for each of these stages there are several options. A sufficient knowledge of the amount and degree of wastewater contamination can help us to design a simple and efficient plant for wastewater treatment (Joss et al., 2005; Kastelan-Macan et al., 2007). Numerous reports regarding the occurrence of nonregulated water contaminants, such as the emerging organic contaminants, have expressed concern about their possible undesirable effects in the environment (Bolong et al., 2009; Daughton, 2004; Daughton and Ternes, 1999). Satyawali and Balakrishnan (2008) concluded that physicochemical methods, like membrane
* Corresponding author. Tel.: þ381 21 485 3685. E-mail address: [email protected] (N. Maravi c). http://dx.doi.org/10.1016/j.jclepro.2015.10.070 0959-6526/© 2015 Elsevier Ltd. All rights reserved.
filtration, are capable of organic load reduction and are therefore being widely field-tested. According to the OECD (The Organisation for Economic Cooperation and Development) report from 2013 vegetable oil production in Europe reaches 21,829,000.00 Mg per year. Therefore, the emission of poor quality effluents by the oilseed processing facilities is posing a dangerous threat to the fresh water resources (Saatci et al., 2001). The oilseed is usually processed in five stages: seed receiving and storage, seed preparation, solvent extraction unit, refining of crude oil and packaging of oil (Nucci et al., 2014). However, the harmful effluent is mostly discharged from degumming, deacidification and deodorisation stages, although the wastewater content and quality may significantly differ from one edible oil industry to another (Kale et al., 1999; Boyer, 1996). Flocculation, coagulation, air floation, oil-skimming and biological processes (anaerobic and aerobic digestion) are typical techniques normally used for edible oil wastewater treatment (Pathe et al., 2000). However, most of these techniques separate only a part of undesired components or require an appropriate pretreatment of the wastewater. Directive 91/271/EEC on urban wastewater treatment and Directive 96/61/EC on Integrated Pollution Prevention Control
Z. Seres et al. / Journal of Cleaner Production 112 (2016) 3132e3137
illustrate the current and future EU policy to encourage development of processes and standards to prevent negative effects on €der et al., 2007). water, using the best available techniques (Schro The methods of wastewater treatment are numerous. That is illustrated in the Best Available Techniques (BAT) for the protection of the environment in the food industry and in Best Available Techniques in the treatment of wastewater and waste gases in the chemical industry, issued by the Commission of the European Union. One of the options recommended by these documents is the application of membrane separation techniques. Membrane technology offers numerous advantages over traditional separation techniques providing higher efficiency, economical savings and faster processes. New membrane techniques are convenient for separation of suspended solids, colloids and high molecular weight materials that are present in treated wastewater (Cassini et al., 2010). The driving force is the pressure difference which enables a continuous separation process (Noble and Stern, 1995). The idea is to microfilter the wastewater, after the separation of oil by the skimmer, where the permeate passes through the membrane and can be recycled and re-used in the vegetable oil refining process. Recent investigations showed promising results and possible large scale application (Buecker, 2007). The main goal of this research was to investigate the possibility of microfiltration of wastewater from vegetable oil refinery through application of the new generation of ceramic membranes to reduce the chemical oxygen demand (COD) of edible oil refinery wastewater. Wastewater was introduced into the laboratory plant for microfiltration treatment under different conditions of flow, transmembrane pressure (TMP) and feed temperature values. Besides COD, permeate flux was measured in order to obtain optimal filtration conditions. Statistical analysis of the data was conducted in order to obtain an appropriate mathematical model of the process. Finally, the model was analyzed and the influence of different factors on the permeate flux and COD was discussed. 2. Materials and methods Treated wastewater was sampled from oilseed processing plants “Victoriaoil”, “Victoria group”, Serbia. Characteristics of this wastewater are: chemical oxygen demand in the range of 5000e18,000 mg O2/L, turbidity in the range of 200e2500 NTU. Our research included pre-treatment of wastewater by filtration process through the filter cloth with pore size of 1 mm in order to separate larger particles that may interfere the microfiltration process. The starting volume of feed mixtures for the cross-flow microfiltration was circa 3 L, which is optimal for such a laboratory plant (Fig. 1), given that this is the sufficient volume of feed mixture needed to fill the pipe system and apparatus to enable the undisturbed running of the pump. Batch cross-flow microfiltration runs were performed during 90 min, with permeate collected in vessel 5 and retentate in vessel 6. A ceramic tubular membrane with a pore size of 200 nm, produced by GEA Westfalia, Germany, was used for a cross-flow microfiltration, under the transmembrane pressure in the range of 1e3 bar, temperature range from 40 to 60 C and a flow rate of 100e300 L/h. During microfiltration, the permeate flux J was monitored, followed by the chemical oxygen demand and turbidity of wastewater (as a feed mixture), permeate and retentate. Chemical oxygen demand was determined by the titrimetric method JUS ISO 6060, Official Gazette of FRY, no. 45/94 (1994). Turbidity was determined by the device Turb 550 IR. The measurements were performed automatically.
3133
Fig. 1. Laboratory apparatus for cross-flow microfiltration (1 e feed tank, 2 e thermostat, 3 e rotary pump, 4 e module with membrane, 5 e vessel for permeate, 6 e vessel for retentate, 7, 8, 9 e valves, 10 e thermometer, 11, 12 e manometer, 13 e rotameter).
The RSM was applied to evaluate the effects of microfiltration parameters and optimize conditions for various responses. BoxeBehnken experimental design (BBD) with three numeric factors on three different levels was used (Myers and Montgomery, 1995). Design included 13 randomized runs with one replicate at the central point. In Table 1 the experimental design is given. Response surface methodology (RSM) is a statistical method of multifactorial analysis of experimental data which provides a better understanding of the process than the standard methods of experimentation, since it is able to predict how the inputs affect the outputs in a complex process where different factors can interact among themselves. All the coefficients of the different polynomial equations were tested for significance with an analysis of variance (ANOVA) (Martí-Calatayud et al., 2010). In such work, the utility of statistical multifactorial analysis of experimental data for understanding the effects of different factors and their interactions in the process were confirmed. For responses obtained after the experiments, a polynomial model of the second degree is established to evaluate and quantify the influence of the variables:
Y ¼ b0 þ b1 x1 þ b2 x2 þ b3 x3 þ b11 x21 þ b22 x22 þ b33 x23 þ b12 x1 x2 þ b13 x1 x3 þ b23 x2 x3 ; (1) where Y is the predicted response, b0 is an intercept; x1, x2 and x3 are the coded levels of input factors; b1 to b33 are regression coefficients; x1x2, x1x3 and x2x3 represent interactions of input factors, while x21 , x22 and x23 represent quadratic terms (Montgomery, 2001). The adequacy of the model was evaluated by the coefficient of determination (R2) and model p-value. Statistical analysis was
Table 1 Variables and levels in the BoxeBehnken experimental design. Factor levels
Input factors Transmembrane pressure (bar) Feed flow rate (L/h) Temperature ( C) Dependent responses Permeate flux (L/(m2 h)) Chemical oxygen demand reduction (%)
1
0
1
1 100 20
2 200 40
3 300 60
Z. Seres et al. / Journal of Cleaner Production 112 (2016) 3132e3137
3134
performed using Statistica software. Response surface plots are shown for two factors, where the third factor is set to its medium value. 3. Results and discussion Table 2 shows the experimentally obtained values according to the BoxeBehnken experimental design for each response under different operating conditions. Permeate flux values presented in Table 2 are obtained after 90 min of microfiltration. Although steady-state permeate flux occurred mostly after 10e20 min of microfiltration, there were numerous fluctuations in permeate flux values under various operating conditions. Therefore, in order to provide representative comparison of the obtained results, the permeate flux values after 90 min of microfiltration were used for further investigations in this research. Chemical oxygen demand rejection, RCOD (%), was calculated in the following way:
RCOD ð%Þ ¼
1
! CODp $100; CODf
(2)
where RCOD (%) e is the chemical oxygen demand rejection, CODf e chemical oxygen demand of feed wastewater (mg O2/L), CODp e chemical oxygen demand of permeate (mg O2/L). Also permeate flux was calculated using the well known formula:
J¼
V ; A$t
(3)
where J e is permeate flux (L/(m2 h)), V e volume of permeate (L), A e membrane surface (m2), t e time of microfiltration (h) 3.1. Permeate fluxes First, the change of permeate flux during the concentration of feed mixtures (without the use of the static mixer) will be presented and analyzed. In Fig. 2 permeate flux decline curve obtained under the experimental conditions: TMP ¼ 2 bar, t ¼ 60 C, feed flow rate ¼ 300 L/h and TMP ¼ 3 bar, t ¼ 60 C, feed flow rate ¼ 200 L/h are shown. The trend was similar for other measured values. As Akay and Wakeman (1994) previously concluded, the decline in the permeate flux values was caused by membrane fouling and gel (or concentration) polarization which lead to the resistance formation on the membrane surface during
Fig. 2. Typical permeate flux J vs time curve of ceramic membrane microfiltration.
microfiltration of the wastewater, from the compounds present in the wastewater. As expected, exponential model:
J ¼ a$t b ;
(where a, b are model parameters) was the best fit for each of the conditions with R2 > 0.95. The regression coefficients obtained by RSM model are given in Table 3. Significance of input factors and their interactions in the observed model were determined by the p-value. A factor is considered to affect the response if the p-value is less than the used probability level. The significance was judged at probability levels of 0.05, 0.10 and 0.15. For the response of permeate flux J coefficient of determination was found to be 0.92, which indicates that only 8% of the variations could not be explained by this model. The most significant effect on the permeate flux value was the interaction between feed flow rate and transmembrane pressure. Apart from this interaction, the only significant linear effect affecting the permeate flux was feed flow rate. Feed flow rate had the biggest impact on the permeate flux during the performed experiments. Pre-treated wastewater from oilseed processing plants had a low density and low viscosity that
Table 3 Estimated coefficients for responses J and RCOD. Regression coefficients
1 2 3 4 5 6 7 8 9 10 11 12 13
Input factorsa
Dependent responses J (L/(m2 h))
Table 2 BoxeBehnken experimental design and responses. Run
(4)
Coefficient Responsesb
TMP (bar)
Q (L/h)
Temp ( C)
J (L/(m2 h))
RCOD (%)
2 3 1 2 1 2 3 2 1 3 3 1 2
100 200 200 300 200 100 200 300 100 100 300 300 200
20 20 20 20 60 60 60 60 40 40 40 40 40
29.03 33.29 35.38 41.2 3.33 20.58 45.6 109.55 28.9 11.59 144.52 6.23 41.3
0.34 0.33 0.33 0.50 0.83 0.84 0.75 0.77 0.66 0.88 0.76 0.70 0.99
a TMP, transmembrane pressure; Q, feed flow rate; Temp, feed mixture temperature. b J, permeate flux; RCOD, chemical oxygen demand reduction.
Intercept b0
798.048
RCOD (%) p-Value
Coefficient
p-Value
0.235156
2.0475
0.036532a
Linear b1 b2 b3
185.310 311.418 6537.6
0.516396 0.132212c 0.640534
0.72938 0.34125 0.07509
0.071241b 0.1221c 0.011a
Quadratic b11 b22 b33
25.524 17625.6 43.2
0.675924 0.441332 0.774774
0.14688 0.03353 0.00071
0.086075b 0.208196 0.016434a
Interaction b12 b13 b23
84.024 1996.2 2073.6
0.044101a 0.410168 0.196935
0.024 0.00094 0.00173
0.43065 0.698955 0.282733
a b c
Effects are statistically significant, p < 0.05. Effects are statistically significant, p < 0.10. Effects are statistically significant, p < 0.15.
Z. Seres et al. / Journal of Cleaner Production 112 (2016) 3132e3137
provided good flowability through the membrane pores. Therefore, rising of TMP and temperature values with fixed value of feed flow rate did not affect the permeate flux significantly. That can also be explained as a result of a formed cake layer on the membrane surface with low compression rate, therefore TMP could not have contributed to the indicative increase of permeate flux values. On the other side, increase in the feed flow rate under the intermediate TMP values had a positive effect on the permeate flux. Increased flow rates led to slower formation of the cake layer due to the higher feed velocity. Moreover, higher feed flow rates provided the constant removal of the particles from the upper layers of the cake which resulted in less cake build-up. Response surface plot showing the combined effect of transmembrane pressure and feed flow rate is given in Fig. 3. Observing Fig. 3, it can be clearly concluded that the highest values of permeate flux were obtained under the intermediate TMP values and the highest feed flow rates. The temperature of feed wastewater varied in the range of 20e60 C and within that range did not affect the permeate flux significantly (highest p-values). Still, it could be expected that temperature of this kind of wastewater has a higher influence on the permeate flux, because it should contain a certain amount of edible oil from the extraction process. However, wastewater pretreatment using the filtration cloth or skimmer (industrial-scale) separates aggregated oil particles from the upper oil-phase of wastewater accumulated in the collector. Therefore, treated wastewater contained only a small percent of edible oil at the level insignificant for the temperature influence on the permeate flux during microfiltration. 3.2. Chemical oxygen demand rejection (RCOD) Now, the results of the influence of input factors on RCOD value will be given and analyzed. In Fig. 4 the positive effect of microfiltration with the membrane pore size of 200 nm on RCOD value decrease is evident. The values before (wastewater) and after microfiltration (permeate and retentate) are presented. Namely, RCOD in permeate had an average decrease of 67%, which is significantly higher reduction comparing to the previous investigation of wastewater microfiltration (Chai et al., 1999). Besides RCOD decrease, the application of 200 nm pore size microfiltration
Fig. 3. Combined effect of TMP and feed flow rate on permeate flux values.
3135
Fig. 4. Average chemical oxygen demand RCOD and turbidity before microfiltration in wastewater and after microfiltration in permeate-purified wastewater, and retentate.
resulted in the significant turbidity removal (99%), as can be seen in Fig. 4. Significantly high value of coefficient of determination (0.961) indicates goodness of the model fit for the response of RCOD, with only 3.9% of the variations which could not be explained by this model. If our model is compared to the models obtained by other researchers dealing with wastewater treatment using RSM (Guo et al., 2013; Ghafari et al., 2009; Yuliwati et al., 2012), our model had a highly significant coefficient of determination indicating that the model was sufficient for estimation of the response in the investigated range of independent variables. Temperature of the feed mixture (wastewater) proved to be the most significant effect. Both linear and quadratic effect of wastewater temperature had the most prominent influence on the chemical oxygen demand reduction. Nevertheless, the linear and quadratic effect of TMP had a moderately significant impact on RCOD. Likewise, a small positive influence on RCOD by the feed flow rate increase was detected, as can be seen in the response surface plot of combined effect of wastewater temperature and feed flow rate (Fig.5). The value of RCOD in wastewater is in direct correlation with the content of organic matters that are mainly comprised of
Fig. 5. Combined effect of wastewater temperature Temp and feed flow rate Q on RCOD.
Z. Seres et al. / Journal of Cleaner Production 112 (2016) 3132e3137
3136
polysaccharides, colloidal proteins or humic substances in the aggregate (Drewes and Fox, 1999). Thus, the amount of retained particles directly influences the permeate RCOD value. Microfiltration through the membrane with a pore size of 200 nm provided good removal efficiency, expressed through the high RCOD rates. Due to relative easy separation, it can be assumed that majority of organics in wastewater from vegetable oil refinery were macromolecules with high molecular weight. Moreover, enhanced removal efficiency for middle and small molecular organics was enabled by macromolecular organics that first blocked the membrane pores, effectively making the membrane pore size smaller and leading to the efficient interception of middle and small molecular organics (Huang et al., 2012). Tahri et al. (2012) concluded that dissolved organic and inorganic substances' retention increases with increasing of TMP. Therefore, moderately significant influence of rising TMP values on RCOD (Fig. 6) is expressed through the forming of the more compact cake layer, which provides prominent removal efficiency of small and middle sized organics. However, the input factor with the greatest positive influence on RCOD was the temperature. It is assumed that increased wastewater temperature influences the enhanced formation of aggregates of the organics and chemical reactions between the components which resulted in compounds of larger molecular masses, which are then easier to remove from the wastewater using the applied membrane pore size. This positive effect is highly pronounced when comparing the RCOD after the microfiltration at 20 C (under 40%) and RCOD after microfiltration on 60 C (over 75%) (Table 2).
4. Conclusions The microfiltration of wastewater from vegetable oil refinery using alumina ceramic membrane with pore size of 200 nm proved to be adequate as a secondary treatment preceding discharge of wastewater into the recipient or its re-use in the industrial process. Positive effects were noticed through the analysis of the most important objectives: RCOD and permeate flux. The best RCOD in permeate, over 75%, was obtained through the microfiltration of
Fig. 6. Combined effect of TMP and feed flow rate Q on RCOD.
wastewater at 60 C. The highest permeate flux was measured during the use of intermediate TMP and high feed flow rates. The permeate flux values during microfiltration were used in optimization of hydrodynamic conditions of the microfiltration process in order to ensure process sustainability. Moreover, the obtained retentate (concentrate) could be applied to the various energy resources, and cause improvements in their energy value. Obtained permeate with reduced RCOD and without turbidity could be re-circulated in the process. Further investigation should include determination of a residual hexane in permeate, as hexane is used in solvent extraction of oil from the oilseeds, and thus provides process sustainability. Acknowledgements for donations, The authors would like to thank “Victoriaoil”, Sid, as well as office GEA, Westfalia Separator GmbH in Belgrade for donated ceramic membrane. This study was part of national project of Ministry of Science and Technological Development of the Republic of Serbia. References Akay, G., Wakeman, R.J., 1994. Mechanisms of permeate flux decay, solute rejection and concentration polarisation in crossflow filtration of a double chain ionic surfactant dispersion. J. Membr. Sci. 88 (2), 177e195. Bolong, N., Ismail, A.F., Salim, M.R., Matsuura, T., 2009. A review of the effects of emerging contaminants in wastewater and options for their removal. Desalination 239 (1), 229e246. Boyer, M.J., 1996. Environmental Impact and Waste Management, Bailey's Industrial Oil and Fat Products, fifth ed. Wiley, New York. Buecker, B., 2007. Microfiltration for industrial water treatment in power generation. Ind. WaterWorld 20e23. Cassini, A.S., Tessaro, I.C., Marczak, L.D.F., Pertile, C., 2010. Ultrafiltration of wastewater from isolated soy protein production: a comparison of three UF membranes. J. Clean. Prod. 18 (3), 260e265. Chai, X., Mi, Y., Yue, P.L., Chen, G., 1999. Bean curd wastewater treatment by membrane separation. Sep. Purif. Technol. 15 (2), 175e180. Daughton, C.G., 2004. Non-regulated water contaminants: emerging research. Environ. Impact Assess. 24 (7), 711e732. Daughton, C.G., Ternes, T.A., 1999. Pharmaceuticals and personal care products in the environment: agents of subtle change? Environ. Health Perspect. 107 (Suppl. 6), 907. Drewes, J.E., Fox, P., 1999. Fate of natural organic matter (NOM) during groundwater recharge using reclaimed water. Water Sci. Technol. 40 (9), 241e248. Ghafari, S., Aziz, H.A., Isa, M.H., Zinatizadeh, A.A., 2009. Application of response surface methodology (RSM) to optimize coagulationeflocculation treatment of leachate using poly-aluminum chloride (PAC) and alum. J. Hazard. Mater. 163 (2), 650e656. Guo, J., Yang, C., Zeng, G., 2013. Treatment of swine wastewater using chemically modified zeolite and bioflocculant from activated sludge. Bioresour. Technol. 143, 289e297. Huang, W., Chu, H., Dong, B., 2012. Characteristics of algogenic organic matter generated under different nutrient conditions and subsequent impact on microfiltration membrane fouling. Desalination 293, 104e111. €bel, A., McArdell, C.S., Ternes, T., Siegrist, H., 2005. Joss, A., Keller, E., Alder, A.C., Go Removal of pharmaceuticals and fragrances in biological wastewater treatment. Water Res. 39 (14), 3139e3152. Kale, V., Katikaneni, S.P.R., Cheryan, M., 1999. Deacidifying rice bran oil by solvent extraction and membrane technology. J. Am. Oil Chem. Soc. 76 (6), 723e727. Kastelan-Macan, M., Ahel, M., Horvat, A.J., Jabucar, D., Jovancic, P., 2007. Water resources and waste water management in Bosnia and Herzegovina, Croatia and the State Union of Serbia and Montenegro. Water Policy 9 (3), 319e343. Kralj, A.K., 2015. The re-usages of wastewater within industry: the positive impact of contaminants. J. Clean. Prod. http://dx.doi.org/10.1016/j.jclepro.2015.02.054. Martí-Calatayud, M.C., Vincent-Vela, M.C., Alvarez-Blanco, S., Lora-García, J., ~ os-Rodríguez, E., 2010. Analysis and optimization of the influence of Bergantin operating conditions in the ultrafiltration of macromolecules using a response surface methodological approach. Chem. Eng. J. 156 (2), 337e346. Montgomery, C.D., 2001. Design and Analysis of Experiments. John Wiley and Sons, Pte. Ltd., Singapore. Myers, R.H., Montgomery, C.M., 1995. Response Surfaces Methodology: Process and Product Optimization Using Designed Experiments. Wiley, New York. Noble, R.D., Stern, S.A. (Eds.), 1995. Membrane Separations Technology: Principles and Applications, vol. 2. Elsevier. Nucci, B., Puccini, M., Pelagagge, L., Vitolo, S., Nicolella, C., 2014. Improving the environmental performance of vegetable oil processing through LCA. J. Clean. Prod. 64, 310e322.
Z. Seres et al. / Journal of Cleaner Production 112 (2016) 3132e3137 Pathe, P.P., Nandy, T., Kaul, S.N., 2000. Wastewater management in vegetable oil industries. Indian J. Environ. Prot. 20 (7), 481e492. Saatci, Y., Hasar, H., Cici, M., 2001. Treatability of vegetable oil industry effluents through physicalechemical methods. Fresen. Environ. Bull. 10 (12), 854e858. Satyawali, Y., Balakrishnan, M., 2008. Wastewater treatment in molasses-based alcohol distilleries for COD and color removal: a review. J. Environ. Manag. 86 (3), 481e497. €der, P., Navarro-Avin ~o , J., Azaizeh, H., Goldhirsh, A.G., DiGregorio, S., Schro Komives, T., et al., 2007. Using phytoremediation technologies to upgrade waste water treatment in Europe. Environ. Sci. Pollut. Res. Int. 14 (7), 490e497.
3137
Tahri, N., Masmoudi, G., Ellouze, E., Jrad, A., Drogui, P., Amar, R.B., 2012. Coupling microfiltration and nanofiltration processes for the treatment at source of dyeing-containing effluent. J. Clean. Prod. 33, 226e235. Yugoslavian Standard JUS ISO 6060, 1994. Water Quality, Chemical Oxygen Demand Determination. Official Gazette of FRYno. 45/94, Belgrade. Yuliwati, E., Ismail, A.F., Lau, W.J., Ng, B.C., Mataram, A., Kassim, M.A., 2012. Effects of process conditions in submerged ultrafiltration for refinery wastewater treatment: optimization of operating process by response surface methodology. Desalination 287, 350e361.
Contents lists available at ScienceDirect
Journal of Cleaner Production journal homepage: www.elsevier.com/locate/jclepro
Treatment of vegetable oil refinery wastewater using alumina ceramic membrane: optimization using response surface methodology s a, Nikola Maravi Zita Sere c a, *, Aleksandar Taka ci a, Ivana Nikoli c a, c a, Cecilia Hodur b Dragana Soronja-Simovi c a, Aleksandar Joki a b
University of Novi Sad, Faculty of Technology, Bul. cara Lazara 1, 21000 Novi Sad, Serbia University of Szeged, Faculty of Engineering, Szeged, Mars t er 7, 6724 Szeged, Hungary
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 March 2015 Received in revised form 12 October 2015 Accepted 18 October 2015 Available online 26 October 2015
New regulations in environment protection and increasing market demands for “green” companies are forcing the industry to consider finding new and sustainable methods of wastewater treatment. The valuable information regarding treatment of the wastewater from the edible oil industry is presented in this research. The obtained results confirm a promising application of the third generation membranes, comprised of ceramic material (aluminium oxide), in treatment of wastewater from the edible oil industry. Positive results regarding chemical oxygen demand reduction and turbidity removal were noticed. Permeate flux values were used for process optimization in order to achieve a cost-effective process. Response surface methodology was used for the experimental design. The effects of wastewater temperature, transmembrane pressure, and feed-flow rate on the microfiltration model fit were studied. The experiments showed that microfiltration of this type of wastewater is a convenient technique for a possible large scale industrial application as a secondary step in treating wastewater from the oilseed processing facilities. © 2015 Elsevier Ltd. All rights reserved.
Keywords: Microfiltration Chemical oxygen demand Wastewater Edible oil Response surface methodology
1. Introduction The main task of wastewater treatment is to eliminate pollution to the extent that the treated wastewater can be discharged into the recipient without consequences or that it can be re-used (Kralj, 2015). The technological process of wastewater treatment can consist of many processing stages, depending on the characteristics of the raw wastewater and the required quality of the treated wastewater, and for each of these stages there are several options. A sufficient knowledge of the amount and degree of wastewater contamination can help us to design a simple and efficient plant for wastewater treatment (Joss et al., 2005; Kastelan-Macan et al., 2007). Numerous reports regarding the occurrence of nonregulated water contaminants, such as the emerging organic contaminants, have expressed concern about their possible undesirable effects in the environment (Bolong et al., 2009; Daughton, 2004; Daughton and Ternes, 1999). Satyawali and Balakrishnan (2008) concluded that physicochemical methods, like membrane
* Corresponding author. Tel.: þ381 21 485 3685. E-mail address: [email protected] (N. Maravi c). http://dx.doi.org/10.1016/j.jclepro.2015.10.070 0959-6526/© 2015 Elsevier Ltd. All rights reserved.
filtration, are capable of organic load reduction and are therefore being widely field-tested. According to the OECD (The Organisation for Economic Cooperation and Development) report from 2013 vegetable oil production in Europe reaches 21,829,000.00 Mg per year. Therefore, the emission of poor quality effluents by the oilseed processing facilities is posing a dangerous threat to the fresh water resources (Saatci et al., 2001). The oilseed is usually processed in five stages: seed receiving and storage, seed preparation, solvent extraction unit, refining of crude oil and packaging of oil (Nucci et al., 2014). However, the harmful effluent is mostly discharged from degumming, deacidification and deodorisation stages, although the wastewater content and quality may significantly differ from one edible oil industry to another (Kale et al., 1999; Boyer, 1996). Flocculation, coagulation, air floation, oil-skimming and biological processes (anaerobic and aerobic digestion) are typical techniques normally used for edible oil wastewater treatment (Pathe et al., 2000). However, most of these techniques separate only a part of undesired components or require an appropriate pretreatment of the wastewater. Directive 91/271/EEC on urban wastewater treatment and Directive 96/61/EC on Integrated Pollution Prevention Control
Z. Seres et al. / Journal of Cleaner Production 112 (2016) 3132e3137
illustrate the current and future EU policy to encourage development of processes and standards to prevent negative effects on €der et al., 2007). water, using the best available techniques (Schro The methods of wastewater treatment are numerous. That is illustrated in the Best Available Techniques (BAT) for the protection of the environment in the food industry and in Best Available Techniques in the treatment of wastewater and waste gases in the chemical industry, issued by the Commission of the European Union. One of the options recommended by these documents is the application of membrane separation techniques. Membrane technology offers numerous advantages over traditional separation techniques providing higher efficiency, economical savings and faster processes. New membrane techniques are convenient for separation of suspended solids, colloids and high molecular weight materials that are present in treated wastewater (Cassini et al., 2010). The driving force is the pressure difference which enables a continuous separation process (Noble and Stern, 1995). The idea is to microfilter the wastewater, after the separation of oil by the skimmer, where the permeate passes through the membrane and can be recycled and re-used in the vegetable oil refining process. Recent investigations showed promising results and possible large scale application (Buecker, 2007). The main goal of this research was to investigate the possibility of microfiltration of wastewater from vegetable oil refinery through application of the new generation of ceramic membranes to reduce the chemical oxygen demand (COD) of edible oil refinery wastewater. Wastewater was introduced into the laboratory plant for microfiltration treatment under different conditions of flow, transmembrane pressure (TMP) and feed temperature values. Besides COD, permeate flux was measured in order to obtain optimal filtration conditions. Statistical analysis of the data was conducted in order to obtain an appropriate mathematical model of the process. Finally, the model was analyzed and the influence of different factors on the permeate flux and COD was discussed. 2. Materials and methods Treated wastewater was sampled from oilseed processing plants “Victoriaoil”, “Victoria group”, Serbia. Characteristics of this wastewater are: chemical oxygen demand in the range of 5000e18,000 mg O2/L, turbidity in the range of 200e2500 NTU. Our research included pre-treatment of wastewater by filtration process through the filter cloth with pore size of 1 mm in order to separate larger particles that may interfere the microfiltration process. The starting volume of feed mixtures for the cross-flow microfiltration was circa 3 L, which is optimal for such a laboratory plant (Fig. 1), given that this is the sufficient volume of feed mixture needed to fill the pipe system and apparatus to enable the undisturbed running of the pump. Batch cross-flow microfiltration runs were performed during 90 min, with permeate collected in vessel 5 and retentate in vessel 6. A ceramic tubular membrane with a pore size of 200 nm, produced by GEA Westfalia, Germany, was used for a cross-flow microfiltration, under the transmembrane pressure in the range of 1e3 bar, temperature range from 40 to 60 C and a flow rate of 100e300 L/h. During microfiltration, the permeate flux J was monitored, followed by the chemical oxygen demand and turbidity of wastewater (as a feed mixture), permeate and retentate. Chemical oxygen demand was determined by the titrimetric method JUS ISO 6060, Official Gazette of FRY, no. 45/94 (1994). Turbidity was determined by the device Turb 550 IR. The measurements were performed automatically.
3133
Fig. 1. Laboratory apparatus for cross-flow microfiltration (1 e feed tank, 2 e thermostat, 3 e rotary pump, 4 e module with membrane, 5 e vessel for permeate, 6 e vessel for retentate, 7, 8, 9 e valves, 10 e thermometer, 11, 12 e manometer, 13 e rotameter).
The RSM was applied to evaluate the effects of microfiltration parameters and optimize conditions for various responses. BoxeBehnken experimental design (BBD) with three numeric factors on three different levels was used (Myers and Montgomery, 1995). Design included 13 randomized runs with one replicate at the central point. In Table 1 the experimental design is given. Response surface methodology (RSM) is a statistical method of multifactorial analysis of experimental data which provides a better understanding of the process than the standard methods of experimentation, since it is able to predict how the inputs affect the outputs in a complex process where different factors can interact among themselves. All the coefficients of the different polynomial equations were tested for significance with an analysis of variance (ANOVA) (Martí-Calatayud et al., 2010). In such work, the utility of statistical multifactorial analysis of experimental data for understanding the effects of different factors and their interactions in the process were confirmed. For responses obtained after the experiments, a polynomial model of the second degree is established to evaluate and quantify the influence of the variables:
Y ¼ b0 þ b1 x1 þ b2 x2 þ b3 x3 þ b11 x21 þ b22 x22 þ b33 x23 þ b12 x1 x2 þ b13 x1 x3 þ b23 x2 x3 ; (1) where Y is the predicted response, b0 is an intercept; x1, x2 and x3 are the coded levels of input factors; b1 to b33 are regression coefficients; x1x2, x1x3 and x2x3 represent interactions of input factors, while x21 , x22 and x23 represent quadratic terms (Montgomery, 2001). The adequacy of the model was evaluated by the coefficient of determination (R2) and model p-value. Statistical analysis was
Table 1 Variables and levels in the BoxeBehnken experimental design. Factor levels
Input factors Transmembrane pressure (bar) Feed flow rate (L/h) Temperature ( C) Dependent responses Permeate flux (L/(m2 h)) Chemical oxygen demand reduction (%)
1
0
1
1 100 20
2 200 40
3 300 60
Z. Seres et al. / Journal of Cleaner Production 112 (2016) 3132e3137
3134
performed using Statistica software. Response surface plots are shown for two factors, where the third factor is set to its medium value. 3. Results and discussion Table 2 shows the experimentally obtained values according to the BoxeBehnken experimental design for each response under different operating conditions. Permeate flux values presented in Table 2 are obtained after 90 min of microfiltration. Although steady-state permeate flux occurred mostly after 10e20 min of microfiltration, there were numerous fluctuations in permeate flux values under various operating conditions. Therefore, in order to provide representative comparison of the obtained results, the permeate flux values after 90 min of microfiltration were used for further investigations in this research. Chemical oxygen demand rejection, RCOD (%), was calculated in the following way:
RCOD ð%Þ ¼
1
! CODp $100; CODf
(2)
where RCOD (%) e is the chemical oxygen demand rejection, CODf e chemical oxygen demand of feed wastewater (mg O2/L), CODp e chemical oxygen demand of permeate (mg O2/L). Also permeate flux was calculated using the well known formula:
J¼
V ; A$t
(3)
where J e is permeate flux (L/(m2 h)), V e volume of permeate (L), A e membrane surface (m2), t e time of microfiltration (h) 3.1. Permeate fluxes First, the change of permeate flux during the concentration of feed mixtures (without the use of the static mixer) will be presented and analyzed. In Fig. 2 permeate flux decline curve obtained under the experimental conditions: TMP ¼ 2 bar, t ¼ 60 C, feed flow rate ¼ 300 L/h and TMP ¼ 3 bar, t ¼ 60 C, feed flow rate ¼ 200 L/h are shown. The trend was similar for other measured values. As Akay and Wakeman (1994) previously concluded, the decline in the permeate flux values was caused by membrane fouling and gel (or concentration) polarization which lead to the resistance formation on the membrane surface during
Fig. 2. Typical permeate flux J vs time curve of ceramic membrane microfiltration.
microfiltration of the wastewater, from the compounds present in the wastewater. As expected, exponential model:
J ¼ a$t b ;
(where a, b are model parameters) was the best fit for each of the conditions with R2 > 0.95. The regression coefficients obtained by RSM model are given in Table 3. Significance of input factors and their interactions in the observed model were determined by the p-value. A factor is considered to affect the response if the p-value is less than the used probability level. The significance was judged at probability levels of 0.05, 0.10 and 0.15. For the response of permeate flux J coefficient of determination was found to be 0.92, which indicates that only 8% of the variations could not be explained by this model. The most significant effect on the permeate flux value was the interaction between feed flow rate and transmembrane pressure. Apart from this interaction, the only significant linear effect affecting the permeate flux was feed flow rate. Feed flow rate had the biggest impact on the permeate flux during the performed experiments. Pre-treated wastewater from oilseed processing plants had a low density and low viscosity that
Table 3 Estimated coefficients for responses J and RCOD. Regression coefficients
1 2 3 4 5 6 7 8 9 10 11 12 13
Input factorsa
Dependent responses J (L/(m2 h))
Table 2 BoxeBehnken experimental design and responses. Run
(4)
Coefficient Responsesb
TMP (bar)
Q (L/h)
Temp ( C)
J (L/(m2 h))
RCOD (%)
2 3 1 2 1 2 3 2 1 3 3 1 2
100 200 200 300 200 100 200 300 100 100 300 300 200
20 20 20 20 60 60 60 60 40 40 40 40 40
29.03 33.29 35.38 41.2 3.33 20.58 45.6 109.55 28.9 11.59 144.52 6.23 41.3
0.34 0.33 0.33 0.50 0.83 0.84 0.75 0.77 0.66 0.88 0.76 0.70 0.99
a TMP, transmembrane pressure; Q, feed flow rate; Temp, feed mixture temperature. b J, permeate flux; RCOD, chemical oxygen demand reduction.
Intercept b0
798.048
RCOD (%) p-Value
Coefficient
p-Value
0.235156
2.0475
0.036532a
Linear b1 b2 b3
185.310 311.418 6537.6
0.516396 0.132212c 0.640534
0.72938 0.34125 0.07509
0.071241b 0.1221c 0.011a
Quadratic b11 b22 b33
25.524 17625.6 43.2
0.675924 0.441332 0.774774
0.14688 0.03353 0.00071
0.086075b 0.208196 0.016434a
Interaction b12 b13 b23
84.024 1996.2 2073.6
0.044101a 0.410168 0.196935
0.024 0.00094 0.00173
0.43065 0.698955 0.282733
a b c
Effects are statistically significant, p < 0.05. Effects are statistically significant, p < 0.10. Effects are statistically significant, p < 0.15.
Z. Seres et al. / Journal of Cleaner Production 112 (2016) 3132e3137
provided good flowability through the membrane pores. Therefore, rising of TMP and temperature values with fixed value of feed flow rate did not affect the permeate flux significantly. That can also be explained as a result of a formed cake layer on the membrane surface with low compression rate, therefore TMP could not have contributed to the indicative increase of permeate flux values. On the other side, increase in the feed flow rate under the intermediate TMP values had a positive effect on the permeate flux. Increased flow rates led to slower formation of the cake layer due to the higher feed velocity. Moreover, higher feed flow rates provided the constant removal of the particles from the upper layers of the cake which resulted in less cake build-up. Response surface plot showing the combined effect of transmembrane pressure and feed flow rate is given in Fig. 3. Observing Fig. 3, it can be clearly concluded that the highest values of permeate flux were obtained under the intermediate TMP values and the highest feed flow rates. The temperature of feed wastewater varied in the range of 20e60 C and within that range did not affect the permeate flux significantly (highest p-values). Still, it could be expected that temperature of this kind of wastewater has a higher influence on the permeate flux, because it should contain a certain amount of edible oil from the extraction process. However, wastewater pretreatment using the filtration cloth or skimmer (industrial-scale) separates aggregated oil particles from the upper oil-phase of wastewater accumulated in the collector. Therefore, treated wastewater contained only a small percent of edible oil at the level insignificant for the temperature influence on the permeate flux during microfiltration. 3.2. Chemical oxygen demand rejection (RCOD) Now, the results of the influence of input factors on RCOD value will be given and analyzed. In Fig. 4 the positive effect of microfiltration with the membrane pore size of 200 nm on RCOD value decrease is evident. The values before (wastewater) and after microfiltration (permeate and retentate) are presented. Namely, RCOD in permeate had an average decrease of 67%, which is significantly higher reduction comparing to the previous investigation of wastewater microfiltration (Chai et al., 1999). Besides RCOD decrease, the application of 200 nm pore size microfiltration
Fig. 3. Combined effect of TMP and feed flow rate on permeate flux values.
3135
Fig. 4. Average chemical oxygen demand RCOD and turbidity before microfiltration in wastewater and after microfiltration in permeate-purified wastewater, and retentate.
resulted in the significant turbidity removal (99%), as can be seen in Fig. 4. Significantly high value of coefficient of determination (0.961) indicates goodness of the model fit for the response of RCOD, with only 3.9% of the variations which could not be explained by this model. If our model is compared to the models obtained by other researchers dealing with wastewater treatment using RSM (Guo et al., 2013; Ghafari et al., 2009; Yuliwati et al., 2012), our model had a highly significant coefficient of determination indicating that the model was sufficient for estimation of the response in the investigated range of independent variables. Temperature of the feed mixture (wastewater) proved to be the most significant effect. Both linear and quadratic effect of wastewater temperature had the most prominent influence on the chemical oxygen demand reduction. Nevertheless, the linear and quadratic effect of TMP had a moderately significant impact on RCOD. Likewise, a small positive influence on RCOD by the feed flow rate increase was detected, as can be seen in the response surface plot of combined effect of wastewater temperature and feed flow rate (Fig.5). The value of RCOD in wastewater is in direct correlation with the content of organic matters that are mainly comprised of
Fig. 5. Combined effect of wastewater temperature Temp and feed flow rate Q on RCOD.
Z. Seres et al. / Journal of Cleaner Production 112 (2016) 3132e3137
3136
polysaccharides, colloidal proteins or humic substances in the aggregate (Drewes and Fox, 1999). Thus, the amount of retained particles directly influences the permeate RCOD value. Microfiltration through the membrane with a pore size of 200 nm provided good removal efficiency, expressed through the high RCOD rates. Due to relative easy separation, it can be assumed that majority of organics in wastewater from vegetable oil refinery were macromolecules with high molecular weight. Moreover, enhanced removal efficiency for middle and small molecular organics was enabled by macromolecular organics that first blocked the membrane pores, effectively making the membrane pore size smaller and leading to the efficient interception of middle and small molecular organics (Huang et al., 2012). Tahri et al. (2012) concluded that dissolved organic and inorganic substances' retention increases with increasing of TMP. Therefore, moderately significant influence of rising TMP values on RCOD (Fig. 6) is expressed through the forming of the more compact cake layer, which provides prominent removal efficiency of small and middle sized organics. However, the input factor with the greatest positive influence on RCOD was the temperature. It is assumed that increased wastewater temperature influences the enhanced formation of aggregates of the organics and chemical reactions between the components which resulted in compounds of larger molecular masses, which are then easier to remove from the wastewater using the applied membrane pore size. This positive effect is highly pronounced when comparing the RCOD after the microfiltration at 20 C (under 40%) and RCOD after microfiltration on 60 C (over 75%) (Table 2).
4. Conclusions The microfiltration of wastewater from vegetable oil refinery using alumina ceramic membrane with pore size of 200 nm proved to be adequate as a secondary treatment preceding discharge of wastewater into the recipient or its re-use in the industrial process. Positive effects were noticed through the analysis of the most important objectives: RCOD and permeate flux. The best RCOD in permeate, over 75%, was obtained through the microfiltration of
Fig. 6. Combined effect of TMP and feed flow rate Q on RCOD.
wastewater at 60 C. The highest permeate flux was measured during the use of intermediate TMP and high feed flow rates. The permeate flux values during microfiltration were used in optimization of hydrodynamic conditions of the microfiltration process in order to ensure process sustainability. Moreover, the obtained retentate (concentrate) could be applied to the various energy resources, and cause improvements in their energy value. Obtained permeate with reduced RCOD and without turbidity could be re-circulated in the process. Further investigation should include determination of a residual hexane in permeate, as hexane is used in solvent extraction of oil from the oilseeds, and thus provides process sustainability. Acknowledgements for donations, The authors would like to thank “Victoriaoil”, Sid, as well as office GEA, Westfalia Separator GmbH in Belgrade for donated ceramic membrane. This study was part of national project of Ministry of Science and Technological Development of the Republic of Serbia. References Akay, G., Wakeman, R.J., 1994. Mechanisms of permeate flux decay, solute rejection and concentration polarisation in crossflow filtration of a double chain ionic surfactant dispersion. J. Membr. Sci. 88 (2), 177e195. Bolong, N., Ismail, A.F., Salim, M.R., Matsuura, T., 2009. A review of the effects of emerging contaminants in wastewater and options for their removal. Desalination 239 (1), 229e246. Boyer, M.J., 1996. Environmental Impact and Waste Management, Bailey's Industrial Oil and Fat Products, fifth ed. Wiley, New York. Buecker, B., 2007. Microfiltration for industrial water treatment in power generation. Ind. WaterWorld 20e23. Cassini, A.S., Tessaro, I.C., Marczak, L.D.F., Pertile, C., 2010. Ultrafiltration of wastewater from isolated soy protein production: a comparison of three UF membranes. J. Clean. Prod. 18 (3), 260e265. Chai, X., Mi, Y., Yue, P.L., Chen, G., 1999. Bean curd wastewater treatment by membrane separation. Sep. Purif. Technol. 15 (2), 175e180. Daughton, C.G., 2004. Non-regulated water contaminants: emerging research. Environ. Impact Assess. 24 (7), 711e732. Daughton, C.G., Ternes, T.A., 1999. Pharmaceuticals and personal care products in the environment: agents of subtle change? Environ. Health Perspect. 107 (Suppl. 6), 907. Drewes, J.E., Fox, P., 1999. Fate of natural organic matter (NOM) during groundwater recharge using reclaimed water. Water Sci. Technol. 40 (9), 241e248. Ghafari, S., Aziz, H.A., Isa, M.H., Zinatizadeh, A.A., 2009. Application of response surface methodology (RSM) to optimize coagulationeflocculation treatment of leachate using poly-aluminum chloride (PAC) and alum. J. Hazard. Mater. 163 (2), 650e656. Guo, J., Yang, C., Zeng, G., 2013. Treatment of swine wastewater using chemically modified zeolite and bioflocculant from activated sludge. Bioresour. Technol. 143, 289e297. Huang, W., Chu, H., Dong, B., 2012. Characteristics of algogenic organic matter generated under different nutrient conditions and subsequent impact on microfiltration membrane fouling. Desalination 293, 104e111. €bel, A., McArdell, C.S., Ternes, T., Siegrist, H., 2005. Joss, A., Keller, E., Alder, A.C., Go Removal of pharmaceuticals and fragrances in biological wastewater treatment. Water Res. 39 (14), 3139e3152. Kale, V., Katikaneni, S.P.R., Cheryan, M., 1999. Deacidifying rice bran oil by solvent extraction and membrane technology. J. Am. Oil Chem. Soc. 76 (6), 723e727. Kastelan-Macan, M., Ahel, M., Horvat, A.J., Jabucar, D., Jovancic, P., 2007. Water resources and waste water management in Bosnia and Herzegovina, Croatia and the State Union of Serbia and Montenegro. Water Policy 9 (3), 319e343. Kralj, A.K., 2015. The re-usages of wastewater within industry: the positive impact of contaminants. J. Clean. Prod. http://dx.doi.org/10.1016/j.jclepro.2015.02.054. Martí-Calatayud, M.C., Vincent-Vela, M.C., Alvarez-Blanco, S., Lora-García, J., ~ os-Rodríguez, E., 2010. Analysis and optimization of the influence of Bergantin operating conditions in the ultrafiltration of macromolecules using a response surface methodological approach. Chem. Eng. J. 156 (2), 337e346. Montgomery, C.D., 2001. Design and Analysis of Experiments. John Wiley and Sons, Pte. Ltd., Singapore. Myers, R.H., Montgomery, C.M., 1995. Response Surfaces Methodology: Process and Product Optimization Using Designed Experiments. Wiley, New York. Noble, R.D., Stern, S.A. (Eds.), 1995. Membrane Separations Technology: Principles and Applications, vol. 2. Elsevier. Nucci, B., Puccini, M., Pelagagge, L., Vitolo, S., Nicolella, C., 2014. Improving the environmental performance of vegetable oil processing through LCA. J. Clean. Prod. 64, 310e322.
Z. Seres et al. / Journal of Cleaner Production 112 (2016) 3132e3137 Pathe, P.P., Nandy, T., Kaul, S.N., 2000. Wastewater management in vegetable oil industries. Indian J. Environ. Prot. 20 (7), 481e492. Saatci, Y., Hasar, H., Cici, M., 2001. Treatability of vegetable oil industry effluents through physicalechemical methods. Fresen. Environ. Bull. 10 (12), 854e858. Satyawali, Y., Balakrishnan, M., 2008. Wastewater treatment in molasses-based alcohol distilleries for COD and color removal: a review. J. Environ. Manag. 86 (3), 481e497. €der, P., Navarro-Avin ~o , J., Azaizeh, H., Goldhirsh, A.G., DiGregorio, S., Schro Komives, T., et al., 2007. Using phytoremediation technologies to upgrade waste water treatment in Europe. Environ. Sci. Pollut. Res. Int. 14 (7), 490e497.
3137
Tahri, N., Masmoudi, G., Ellouze, E., Jrad, A., Drogui, P., Amar, R.B., 2012. Coupling microfiltration and nanofiltration processes for the treatment at source of dyeing-containing effluent. J. Clean. Prod. 33, 226e235. Yugoslavian Standard JUS ISO 6060, 1994. Water Quality, Chemical Oxygen Demand Determination. Official Gazette of FRYno. 45/94, Belgrade. Yuliwati, E., Ismail, A.F., Lau, W.J., Ng, B.C., Mataram, A., Kassim, M.A., 2012. Effects of process conditions in submerged ultrafiltration for refinery wastewater treatment: optimization of operating process by response surface methodology. Desalination 287, 350e361.