Research article
Gene expression programming for process parameter optimization during ultrafiltration of surfactant wastewater using hydrophilic polyethersulfone membrane

https://doi.org/10.1016/j.jenvman.2020.110444Get rights and content

Highlights

  • Treatment of surfactant wastewater was studied using polyvinylpyrollidone modified polyethersulfone membrane.

  • Hydrophilicity property improved with the modification of polyvinylpyrollidone.

  • The optimum condition of transmembrane pressure, feed concentration, and temperature are found to be 3 bar, 100 ppm, and 25°C, respectively.

  • The maximum surfactant rejection of 72% was achieved in modified membranes.

Abstract

Surfactants are the emerging contaminant and cause a detrimental effect on the ecosystem. In this study, an attempt is made to removal anionic surfactant Sodium dodecyl sulfate (SDS) containing wastewater using hydrophilic polyvinylpyrollidone (PVP) (5–15 wt%) modified polyethersulfone (PES) ultrafiltration membrane. The influence of operating variables on membrane performance was also sequentially analyzed using tests and three numerical modeling methods such as multiple linear regression (MLR), multiple Ln-equation regression (MLnER), and gene expression programming (GEP). Contact angle value of 10 wt% PVP modified PES membrane decreased up to 23.8°, whereas the neat PES membrane is 70.7°. This study indicates that the required hydrophilic property was improved in the modified membrane. The water flux and porosity also enhanced in PVP modified PES membranes. In performance evaluation, the optimum operating variable condition of transmembrane pressure (TMP), feed concentration, and the temperature is found to be 3 bar, 100 ppm, and 25 °C, respectively. Among the models, GEP has a good correlation with experimental anionic surfactant SDS filtration data. GEP performs better than other model with respect to statistical parameter and error terms. This study provides an insight into an adaptation of novel numerical modeling methods for the prediction of membrane performance to the treatment of surfactant wastewater.

Introduction

Water resources are depleting in recent years due to industrialization, urbanization and rapid growth of population. Industrial graywater has gained more interest in recent days for the production of potable water (ciabatti et al., 2009). The graywater from textile and dyeing industries constitutes of various pollutants (Mai, 2013). Among the pollutant, surfactants have toxic effect to biotic systems. Therefore, it is necessary to remove the surfactant from the wastewater. The conventional method of treatment methods are physical filtration, adsorption, flotation, and biological methods (Kim et al., 2008; Nicolaidis and Vyrides, 2014). Membrane technology has been implemented in various industries to meet the stringent regulation by environmental agency to discharge of wastewater and water reclamation (Janpoor et al., 2011; Braeken et al., 2004). Nanofiltration and ultrafiltration are prevalently used in the treatment of graywater (Bhattacharyya et al., 1987; Ciabatti et al., 2009; Afkham et al., 2016). Fouling is an inevitable drawback in membrane separation process. Membrane modification and optimization of process conditions are the two major methods to improve the membrane performance. Hence, in this research, both methods were adopted to improve the treatment of surfactant containing wastewater. In addition, there is sodium dodecyl sulfate (SDS) in gray laundry water because there is this substance in scourer and detergent that Sumisha et al. (2015) used polyethersulfate (PES) and polyinylpyrollidone (PVP) to treat the gray water of laundry. They showed that the efficiency of PES membranes is great to remove oil, detergent, and surfactant from gray laundry water.

As above mentioned, there are two methods to improve the performance of PES membrane, including changing the operation parameters and using sufficient solvent or nanoparticle. On the other hand, there are two categories of additive material, including polymers material and inorganic nanoparticles (NPs), which makes the PES membrane more hydrophilic (Otitoju et al., 2018).

Several studies focused on improving the performance of the ultrafiltration membrane by using an additive (Krishnamurthy et al., 2016; Jiang et al., 2017; Nasrollahi et al., 2019). Jiang et al. (2018) employed deep eutectic solvent (DES) to improve the performance of PES ultrafiltration membrane. They demonstrated that the rejection ratio is increased 91.5–99% approximately when DES was used as additive in PES membrane (Jiang et al., 2018). Some studies used combination several additive materials, such as nanoparticles, to modify hydrophilic performance of ultrafiltration membrane (Vatanpour et al., 2018; Farahani and Vatanpour, 2018; Nasrollahi et al., 2018). Rabiee et al. (2014) proposed to employ poly vinyl chloride and TiO2 nano composite to improve the performance of ultrafiltration membrane.

Although there are some important advantages in using a membrane to treat wastewater, the fouling membrane at the short time is the major disadvantage of this method (Koh et al., 2005; Zhao et al., 2013). There are some methods, such as adding hydrophilic substance to the membrane, backwashing membrane, and optimization effective operating parameters to overcome this issue. The operating parameters such as transmembrane pressure (TMP), cross flow and stirring speed were optimized to improve performance of membrane and increase the flux (Mohammadi et al., 2003; Sondhi et al., 2000). Although these parameters were optimized to improve the performance of membrane in treating wastewater in the previous studies, other operating parameters such as the temperature of wastewater, amount of PVP and kind of solvent should be evaluated and optimized to improve the performance and efficiency of the membrane.

There are several useful studies, which focused on the artificial neural network (ANN) and radial basis function to predict the performance of the membrane for implementing the treatment of the wastewater (Chen et al., 2019; Zhao et al., 2019). In addition, some mathematical models were used to predict the performance of membrane bioadsorber reactor process in the water treatment process that non-dimensionalized method was utilized to find out the relationship between inputs variables and output; for example, Tsai et al. (2005) used a pore diffusion model for predicting the performance of the membrane. Cinar et al. (2006) used the artificial neural network (ANN) to predict the performance e of membrane bioreactor for treating. Some studies focused on reliable parameters to predict critical flux for the membrane bioreactor process (Fan et al., 2006; Khan et al., 2009). Moreover, a hydraulic model was used to predict the performance of the ultrafiltration membrane (Mielczarek and Bohdziewicz, 2011). Some regression models, such as multiple linear regression (MLR) and multiple Ln equation regression (MLnER), and also optimization models, like Gene expression programming (GEP), have been used in complex problems to predict outputs of the problem in the recent years that are useful and strong tools in prediction problems (Kumar and Goyal, 2011; Shishegaran et al., 2020a; Shishegaran et al., 2020b). The nonlinear model is stronger than the linear model in short datasets (Singh et al., 2013); thus, the results of linear and nonlinear models in predicting outputs should be compared. Although there are numerous publications regarding the water filtration (Rahmawati et al., 2019) and wastewater treatment (Damtie et al., 2019; Pang et al., 2019) using a membrane, there is lack of studies on developing numerical models for predicting the performance of the membrane.

Studies of optimization of ultrafiltration of surfactant containing wastewater are limited. Hence, in this research, hydrophilic PVP modified PES membrane is intended to treat the surfactant containing wastewater. The effect of the input variables, including feed concentration, temperature, and transmembrane pressure, on surfactant ultrafiltration, were evaluated. The efficiency and performance of the membrane were compared by three various numerical models, including MLR, MLnER, and GEP. The first objective of this study is related to the use of the hydrophilic polyethersulfone membrane to treat the surfactant containing wastewater. The second objective of this study is related to the use of numerical models, including MLR, MLnER, and GEP, for predicting the performance of this membrane in this process.

Section snippets

Methodology

This study is divided into two parts. The first part includes carrying out experimental tests, and there is a step in this part. The experimental tests are designed based on the changes in the variables in 3–5 levels separately. On the other hand, the value of PVP, pressure, temperature, and feed concentration are changed in samples 1–3, 4–8, 9–11, and 12–14, respectively. The concentration of pollutant, temperature, pressure, kind of the solvent, and the amount of PVP are considered as the

Pure water flux of the membrane

Hydrophilicity and pore size of the membrane are two main factors that are the most effective parameters on the performance of pure water flux. Table 2 shows the ultrafiltration membrane parameters and its performance. In this case, three membranes with various percentage of PVP, including PES(95%)/PVP(5%), PES(90%)/PVP(10%), and PES(85%)/PVP(15%), were produced. Also, the effect of kind of solvent on pure water flux was measured. Based on these results, the best kind of solvent and the best

Conclusion

The performance of PES membrane with four kinds of solvent, different amounts of PVP, and the effect of various operating parameters such as pressure, temperature, and feed concentration were measured for the treatment of wastewater containing SDS.

The experimental result shows that DMF is the best solvent for these membranes. 10 percent of adding PVP is the best amount in order to produce the membrane. In addition, the results of ATR, CA, and morphology tests show adding 10% PVP can improve the

CRediT authorship contribution statement

Aydin Shishegaran: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Software, Validation, Writing - original draft, Writing - review & editing. Arash Nazem Boushehri: Conceptualization, Formal analysis, Investigation, Methodology, Resources, Software, Validation, Writing - original draft, Writing - review & editing. Ahmad Fauzi Ismail: Conceptualization, Methodology, Supervision, Validation, Writing - review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgement

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (46)

  • T. Mohammadi et al.

    Hydrodynamic parameters affecting flux and fouling during reverse osmosis of sea water

    Desalination

    (2003)
  • N. Nasrollahi et al.

    Preparation and characterization of a novel polyethersulfone (PES) ultrafiltration membrane modified with a CuO/ZnO nanocomposite to improve permeability and antifouling properties

    Separ. Purif. Technol.

    (2018)
  • C. Nicolaidis et al.

    Closing the water cycle for industrial laundries: an operational performance and techno-economic evaluation of a full-scale membrane bioreactor system Resources

    Conserv. Recycl.

    (2014)
  • T.A. Otitoju et al.

    Recent advances in hydrophilic modification and performance of polyethersulfone (PES) membrane via additive blending

    RSC Adv.

    (2018)
  • Mohammad Peydayesh et al.

    Fabrication optimization of polyethersulfone (PES)/polyvinylpyrrolidone (PVP) nanofiltration membranes using Box–Behnken response surface method

    RSC advances

    (2017)
  • R. Rahmawati et al.

    Finned spacer for efficient membrane fouling control in produced water filtration

    J. Environ. Manag.

    (2019)
  • A. Shishegaran et al.

    Computational predictions for estimating the maximum deflection of reinforced concrete panels subjected to the blast load

    Int. J. Impact Eng.

    (2020)
  • A. Shishegaran et al.

    Prediction of air quality in Tehran by developing the nonlinear ensemble model

    J. Clean. Prod.

    (2020)
  • K.P. Singh et al.

    Predicting adsorptive removal of chlorophenol from aqueous solution using artificial intelligence based modeling approaches

    Environ. Sci. Pollut. Control Ser.

    (2013)
  • S.I.V. Sousa et al.

    Multiple linear regression and artificial neural networks based on principal components to predict ozone concentrations

    Environ. Model. Software

    (2007)
  • H.P. Srivastava et al.

    Performance of modified poly (vinylidene fluoride) membrane for textile wastewater ultrafiltration

    Desalination

    (2011)
  • A. Sumisha et al.

    Treatment of laundry wastewater using polyethersulfone/polyvinylpyrollidone ultrafiltration membranes

    Ecotoxicol. Environ. Saf.

    (2015)
  • H. Susanto et al.

    Characteristics, performance and stability of polyethersulfone ultrafiltration membranes prepared by phase separation method using different macromolecular additives

    J. Membr. Sci.

    (2009)
  • Cited by (26)

    • New prediction models for the compressive strength and dry-thermal conductivity of bio-composites using novel machine learning algorithms

      2022, Journal of Cleaner Production
      Citation Excerpt :

      Finally, in 1985, Cramer introduced Genetic Programming (GP), which was later enhanced using a variety of sizes and shapes (Cheng et al., 2020; Koza and Koza, 1992), and in 1999, Candida Ferreira, 2001, 2006 invented the EA technique known as GEP. The GEP consists of linear chromosomes with fixed lengths for encryption programs and can predict highly complex nonlinear problems (Shishegaran et al., 2020) for solving modeling functions, regressions, forecasting, detecting, and data analysis (Li et al., 2019; Mousavi et al., 2015). In the GP technique, the Roulette wheel is utilized for picking the data that is simultaneously regenerated with the aid of genetic operators (Armaghani et al., 2018).

    • Application of data-driven machine learning to predict propranolol and trimethoprim removal using a managed aquifer recharge system

      2022, Journal of Environmental Chemical Engineering
      Citation Excerpt :

      Multilayer perceptron (MLP) neural networks have been effectively employed to predict the performance of polymeric inclusion membranes and micellar enhanced ultrafiltration for heavy metal removal from wastewater [20,23,32]. In addition, gene expression programming (GEP) can develop a relationship between the input and output, which is considered a robust prediction tool aiding in predicting the output for complex problems [33]. Although GEP modeling has been used in various engineering applications [27,34], there have been no studies on GEP-based predictive modeling for PPs removal using MARS.

    View all citing articles on Scopus
    View full text