Clean water from clean energy: removal of dissolved contaminants from brackish groundwater using wind energy powered electrodialysis
Around 770 million people lack access to improved drinking water sources (WHO 2013), urgently necessitating implementation of contaminant removal by e.g. desalination systems on a large scale. To improve water quality and enable use of brackish water sources for human consumption in remote arid areas, a directly coupled wind – electrodialysis system (Wind-ED) was developed. Modularity, sustainability and above all suitability for the practical use in off-grid locations were the main motivations and design objectives. The direct coupling of wind energy with membranes reduces the system costs as well as technical drawbacks associated with using intermediate energy storage systems. During this research, systematic experiments were performed using the Wind-ED system in order to determine desalination performance and clean water production, specific energy consumption (SEC) and current efficiency (ηc) under relevant conditions, such as varying: i) wind speed, ii) wind turbulence intensity, iii) oscillation periods, iv) varying NaCl concentrations and v) flow rates. Moreover, the competitive removal of four commonly available inorganic contaminants in brackish groundwater sources, nitrate (NO3-), fluoride (F-), sulphate (SO42-) and chloride (Cl-), were investigated. Firstly, to establish a systematic understanding of how and to what extent energy fluctuations influence the transport of the salt (i.e. NaCl) ions across the membranes, experiments were conducted using pulsed electric field assisted electrodialysis (pulsed-ED) over a wide range of frequencies (0.001 – 10 Hz) and duty cycles (20 – 80). The results showed that pulsation applied in the sub-limiting regime resulted in reduced water production, explained by the delays caused by the off-periods during the pulsed desalination process. At higher current densities, pulsation led to considerable improvements in current (e.g. up to 95%, for a feed solution of 500 mg/L and a pulse regime of 1 Hz at 50 V peak voltage) and significant reduction in water dissociation, explained by a reduction of concentration polarisation. Importantly, the pulsation had no significant effect on energy consumption or current efficiency suggesting that ED could be suitable for direct coupling to fluctuating energy sources such as wind energy. ED was consequently coupled to a wind turbine system and a series of desalination tests were performed over a wide range of wind speeds (2-10 m/s), turbulence intensities (TI of 0-0.6) and oscillation periods (0-180 s). Results showed that water production and SEC increased with wind speed. However, both the water production and SEC stopped increasing as the power output from the turbine levelled off at wind speeds above the rated value (vrated: 7.9 – 8.4 m/s). The impact of wind speed fluctuations on the system performance were insignificant up to a TI of 0.4. The desalination performance declined under high turbulence intensity fluctuations (TIs ≥ 0.5) and long periods of oscillation (> 40 s), as the wind-ED system periodically cycled off in response to operation below the cut-in wind speed of the wind turbine (vcut-in: ~ 2 m/s). The off-cycling of the system caused significant delays in the desalination process, and thus resulted in reduced water production. Further reduction in the water production resulted as the wind-ED system operated under intermittent wind speed conditions with off-wind periods longer than 10 s. It was concluded that the main challenge in direct coupling of ED to a wind resource was not the magnitude of the fluctuations but the impact of the power cycling off during long periods of oscillation and lengthy periods of no wind. Interestingly, the SEC of the process remained relatively unaffected by the fluctuations and intermittencies in the wind resource. The effect of energy fluctuations on the competitive transport of F-, Cl-, NO3- and SO42- from artificial brackish water (TDS ~4350 mg/L) was investigated using different sets of real wind data. The ion removal, independent of the wind regime tested, followed the order: NO3- ≥ Cl- > F- > SO42-. The competitive removal of the ions was linked to differences in physicochemical properties (i.e. hydration energy, ionic mobility and valence). The specific selectivity (e.g. preferential transport of NO3- over SO42- ions) was found to increase with concentration polarisation being either minimised (by lowering the mean wind speed) or disrupted (by fluctuations in the wind resource). The results from flow rate and feed concentration experiments, showed that power production of the wind turbine depended on not only the available wind energy but also the resistance of the load (i.e. the ED stack). Thus, increasing the feed concentration and the flow rate resulted in reduced resistance in the ED stack (Rstack), which inversely influenced the current induction counter torque force applied on the shaft of the wind turbine and caused the rotor to spin at a lower angular velocity. This led to increased sensitivity of the wind-ED system to wind speed fluctuations (e.g. system cycled off due to extreme fluctuations and intermittencies with low TDS feed concentration of 2400 mg/L) and hence a reduction of desalination performance. Impact of flow rate on the SEC was found to be negligible; this was attributed to the automatic voltage to current adjustments done by the wind turbine, in order to minimise the impacts of Rstack on the power production by the turbine at a given wind speed. Increased flow rate and resulting shrinkage of the boundary layer’s thickness, caused the concentration profiles at the solution-membrane interface to become steeper. This favoured the transport of ions with the highest diffusion coefficients in the mixture (i.e. Cl- and NO3-). Decreased flow rate favoured the transport of ions with larger valence numbers and higher electric mobility inside the electrolyte (i.e. SO42-); as the former property governed the faster migration of SO42- ions through the thick boundary layer and the latter property assisted with the improved affinity of the ion-exchange membrane to SO42- ions compared to the monovalent anions in the mixture. Increasing the feed concentration of Cl- from 500 to 2,550 mg/L led to reduced transport numbers for the other anions in the mixture and significantly reducing their removal rate. The results obtained from both the pulsed-ED and wind-ED experiments showed that, despite direct coupling to the fluctuating energy source the SEC of the process remained relatively unaffected by the energy fluctuations. Although the desalination process might require more time to be completed when operating under extreme wind speed fluctuations and intermittencies, the quality of the drinking water produced was always within the WHO standards. In conclusion, the findings from this research prove the wind-ED system to be an energetically robust and a reliable off-grid desalination technique suitable for the treatment of brackish groundwater in water stressed remote regions.