PERFORMANCE EVALUATION OF PV DRIVEN NF-RO HYBRID WATER TREATMENT SYSTEM

Abstract

Inland salinity of groundwater, having a total dissolved solids (TDS) of 1500–3000 mg/L, has been found in substantial volumes throughout the majority of India. However, these parts of the India also receive a 5.5 – 6 kWh/m2/day of annual average global horizontal irradiance (GHI) making the photovoltaic (PV) the apparent selection as a renewable energy source. Nanofiltration (NF) and Reverse Osmosis (RO) membrane are processes used for the removal of divalent ions (Calcium, Magnesium, Sulphate etc.) and monovalent ions (Sodium, Chloride etc.) respectively [131, 215, 220]. Integration of renewable energy (PV) with desalination technologies (NF and RO) may provide the solution of drinking water problem in areas suffering from brackish ground water sources as well as having non-existent or limited electricity accessibility. Several solar powered membrane filtration systems have been investigated earlier. The performance and overall cost of such systems has been based on water productivity, rejection capability and power consumption. As a higher feed water recovery would result in smaller installation size of the membrane unit as well as have less capital and operating costs, many attempts have been made to increase the water recovery of membrane processes. In addition to this, increase in recovery by using concentrate staging has also been investigated earlier resulting in a significant reduction in Specific Energy Consumption (SEC), henceforth the overall system cost. Although some researchers had shown an influence of NF-RO hybridization on plant performance in terms of water quality in permeate staging configuration but, similar studies with concentrate staging configuration were not reported to minimize the SEC enabling its integration with solar PV system. In view of the above, NF-RO hybrid systems with both the configurations apparently need to be studied with an aim to increase water recovery, minimize the SEC and assessment of its compatibility with the PV system.The main objectives of this research work are: 1. Process optimization for NF and RO membrane systems in isolation and hybrid modes for both the staging configurations. 2. To check the operational feasibility of PV system driven NF-RO hybrid system and evaluate seasonal fluctuation of solar radiation on plant performance. 3. Economic assessment and recommendation for up-scaling to pilot studies. iv This thesis is composed of seven chapters. Chapter 1 deals with the introduction which includes problem statement, research gaps, objectives, research design and methodology, significance and organisation of thesis. Chapter 2 presents the literature review, discussing the studies carried out in the field of solar desalination with special interest of PV-RO technology and related issues along with the need of research in the present scenario. Chapter 3 presents methodology of the research work performed in this thesis. Overview of the system configuration and experimental procedure has been discussed in detail. A laboratory scale hybrid membrane unit was designed and procured to perform various membrane experiments in isolation and hybrid modes. A 1.5 kWp grid connected PV system consisting of multi-crystalline silicon PV modules was employed in combination with grid connected solar inverter capable to supply loads of up to 3 kWp, metering devices and distribution system. Optimization experiments were carried out to maximize water recovery and TDS rejection and to minimize SEC of small scale brackish water RO process [116] using pH, feed temperature, feed pressure and concentration of feed solution as input parameters. Six TFC RO and NF (molecular weight cut off-MWCO 100, 250 and 400 Da) membranes from four leading manufacturing companies (CSM, Dow, Vontron and Permionics) in spiral wound configuration were used to perform laboratory scale experiments. In present study, synthetic water was formulated in the laboratory on the basis of major ionic elements of actual groundwater and further used for performing different membrane experiments. Furthermore, to validate the accuracy of this formulation, concentrations of elements of the synthetic groundwater were analysed experimentally and compared with that of actual ground water. The two results displayed a net correlation of 0.9982. Chapter 4 deals with membrane characterization and optimization of input process parameters. This chapter presents the details of characterization of six commercially available small scale RO and NF membranes and also the optimization experiments using central composite design (CCD) of response surface methodology (RSM). Physical aspects of characterizing NF and RO membranes from various manufacturing brand were investigated. Physical surface characteristics including surface roughness, occurrence of functional groups and hydrophobicity/hydrophilicity properties were determined by atomic force microscopy (AFM), fourier transform infrared spectroscopy (FTIR) and contact angle measurement, respectively. A correlation between surface properties and membrane filtration results were obtained. From AFM analysis, it was revealed that the CSM RO membrane was the smoothest, with an RMS (Rq) v value of 33.99 μm. From FTIR analysis graphs it could be predicted that all RO and NF membranes contained thin polyamide layer with polysulfone support [152]. The smaller contact angle (higher hydrophilicity) and smoother surface of CSM membrane among RO membranes and NF250 among NF membranes could be the reason of its better performance. In the present study, optimization was performed employing RSM using CCD. Furthermore, experiments to validate these RO results were conducted employing the optimized process variables values derived from RSM prediction. Since RSM proposes the surface plots offering a better approach to envisage relations between independent and dependent variables, both the RSM and artificial neural network (ANN) methods were applied for modelling using the same experimental data. Finally, the validation of optimized process conditions suggested by the RSM was also carried out by the generated ANN model in MATLAB. After employing RSM models for process parameters optimisation on all NF and RO membranes, CSM membrane (among RO membranes) showed the best performance at 31.92oC temperature, 0.79 MPa pressure, 1500 mg/l feed salt concentration and 6.53 pH (very near to the actual i.e. 6.7) with 19.25% water recovery, 89.2% salt rejection and 17.6 kWh/m3 of SEC. Also, NF250 showed the best performance (among NF membranes) at 30oC temperature, 1.08 MPa pressure, 1500 mg/l feed salt concentration and pH 7.15 with 18.98% water recovery, 70.64% salt rejection and 9.35 kWh/m3 of SEC. Removal efficiency of major ions of validation experiment was observed. The removal efficiency of divalent ions (Ca2+<Mg2+<SO42+) was found to be higher as compared to monovalent ions (NO3-<Na+<Cl-) through RO and NF membranes. However, the overall removal efficiency of ions was generally higher in RO membranes than the NF membranes. Furthermore, ANN model was used to validate the RSM predicted optimized process conditions. Feed water temperature (31.94 oC), pressure (0.78 MPa), salt concentration (1500 mg/L) and pH (6.53) were used as input parameters for the ANN model. ANN predicted 19.51% and 18.59% of water recovery, 88.92% and 71.4% of TDS rejection and 16.60 kWh/m3 9.43 kWh/m3 of SEC for CSM and NF250 membranes respectively, at optimal process conditions. A comparison of the predicted values between ANN and RSM revealed that the values predicted by both RSM and ANN model were much closer to experimental values. Chapter 5 deals with the RO - NF hybrid experiments with PV system. After RSM optimization, best RO and NF membranes were selected to perform validation runs followed by NF-RO hybrid experiments. According to the operating conditions of feed, concentrate and permeate flow streams, the membrane performance data in isolation and hybrid configurations vi were analysed. The membrane filtration unit was designed to operate for about 6 to 8 hrs./day, depending on peak sun shine hours (PSSH). Monthly variation in current generation mainly depends on the sunny or cloudy nature of sky. Generated current was compared with the current required by hybrid membrane system. It was apparent that the amount of current generated during the PSSH was enough to operate the NF-C-RO hybrid membrane unit. Excess energy, which was not utilized by membrane filtration system,could be utilized in maintaing the temperature of feed water and/or for pumping the water. The chemical analysis by scanning electron microscope-energy dispersive using X-Ray (SEM‐EDX) demonstrated that Ca, Mg, Cl, O, C and S were the major elements of the inorganic deposits. The next highest cation was Mg involved in the inorganic fouling process. The SEM-EDX analysis of inorganically fouled membranes signified that substantial quantity of deposits contained inorganic material and minerals. For every possible compound initial search was performed using X-Ray diffraction analysis (XRD) data base. The XRD patterns of the all inorganically fouled membranes were almost similar. Calcite (CaCO3) was observed as a common element of the crystalline phase deposits observed on the all membrane surface The aim of the chapter 6 is the economic assessment to estimate the water production cost. For estimation of water production cost, common technical assumptions, specifications and design parameters were considered for PV assisted RO and NF membrane systems in isolation and hybrid mode. The water production cost of NF-RO hybrid system (Rs.99.81/m3) was about 1.6 times lesser than NF (Rs.158.46/m3) and 4 times lesser than RO (Rs.400.49/m3) system in isolation. Moreover, this cost could be further reduced on increasing the capacity of the membrane system and by providing more subsidy on solar system as an incentive to its users. The water production cost of Rs.146.5/m3, Rs.370.26/m3 and Rs.92.87/m3 for the PV-NF, PV-RO and PV-NF/RO hybrid membrane systems respectively could be reached when the plant life increases to 30 years. Finally chapter 7 presents the conclusions and contributions made by the research work as well as the recommendation for future work.