Membrane-driven processes account for over half the existing renewable energy powered desalination in existence. Some of the reasons for this include that they are a modular technology, easy to install, compact in size, and simple to operate. Many of these advantages are also mirrored by renewable energy microgenerators, such as PV modules. These are also modular, contain no moving parts, have a long life (> 20-year warranty) and involve low maintenance. The modularity of both of these technologies has also assisted in cost reduction being achieved via economies-of-scale. Wind turbines are also available in a wide variety of sizes (from 100 W up to MW scale) and multiple turbines can be included in a system design. Therefore, it is possible to scale a renewable energy powered membrane system to almost any size. These factors, combined with RO and NF exhibiting a very low SEC for seawater and brackish water, respectively, makes membranes an obvious choice when powering such systems with renewable energy.
Further advantages for small-scale systems can be realized coupling the DC output of PV modules and small wind turbines to power the necessary DC pump(s) and electronics, as well as possibly storing some energy in batteries. A DC only system increases system efficiency by 5-10% due to the avoidance of losses in power conversion (DC-AC) and rectification (AC-DC). In addition, the majority of renewable energy powered membrane filtration systems tend to use batteries to avoid energy fluctuations to enable continuous operation and avoid variations in pressure and flows. While energy storage enables a membrane system to produce a known amount of water at the desired quality, the use of batteries results in several problems:
1. The charge-in/charge-out efficiency of a typical deep-cycle lead acid battery is 75-80% , which results in a loss in system efficiency on the order of 20-25%. In order to overcome this loss, a 20-25% larger PV array is needed, substantially increasing the system cost.
2. Batteries both perform worse and degrade faster at higher temperatures, which is likely to coincide with arid regions where PV technology will be implemented. Specifically, with increasing operating temperature, the battery capacity decreases, followed by the charge efficiency decreasing and the self-discharge rate increasing . This has resulted in battery banks requiring replacement in as little as 2 years after installation, thus adding considerably to maintenance costs .
3. Even for a "long" battery-life of 5 years — representing over 1500 charge—discharge cycles — the battery bank will require replacement on average four times during the life of the system, since PV systems are designed to have at least a 20-year life, thus further adding to the lifecycle cost of the system.
4. A follow-on problem is that lack adequate disposal/recycling facilities rarely exist in remote regions, and improper disposal can create further environmental hazards .
For these reasons, renewable energy powered membrane systems are being investigated where the energy is stored in the form of the product water. This means that the system may have to be slightly oversized to account for variations in the energy resource availability, for example, to store enough water to account for a very cloudy day with minimal clean water production. However, as long as the water stored in the permeate tank remains free from biological contamination, this approach can lead to a lower lifecycle cost — and hence cost of water — as well as a much more robust system design that facilitates autonomous operation.
Therefore, it is interesting to investigate the performance of batteryless RE membrane systems. While PV-powered water pumping systems, which are directly DC-coupled between the PV panel and pump motor, operate very successfully without any form of energy storage , relatively little is known about the consequences of variable operation (flow, pressure) on NF and RO membrane systems [17,34]. This research is being pursued for both PV- and wind-powered membrane filtration systems [17,35].
Field trials performed in outback Australia have demonstrated that while relatively large variations in solar irradiance occur, due to large clouds passing overhead, the system still produces good quality water. This is demonstrated in the graphs in Fig. 8, which detail the performance of a 300 W PV-powered RO filtration system when treating brackish feedwater with an electrical conductivity (EC) of 8.2mS/cm during October 2005 (spring) . The two gray curves in Fig. 8a show the incident solar irradiance measured on the horizontal (dashed line) as well as that falling on the PV panels attached to a single-axis (east-west) solar tracker (solid line). This clearly indicates the advantage of having the PV modules track the path of the sun throughout the day, producing 36% more electricity throughout the day [9.5 kWh/(m2 day) instead of 7.0kWh/(m2day)]. Fig. 8a also plots the power output from the PV panels, which closely matches the solar resource availability. The maximum occurs at slightly less than the 300 W rating of the PV module due to temperature effects.
300 250 200 150 100 50 0
30 25 20 15 10
—i—i—i—i—i—i—i—i—i—i—i—r (a) . • . •
06:00 09:00 12:00 15:00 18:00 Time of Day (07.10.2005)
1200 ^ 1000 800 600 400 200 0
1000 800 600 400 200 0
1000 800 600 400 200 0
1000 800 600 400 200 0
1200 1000 800 600 1 400 200
12:30 12:32 12:34 12:36 12:38 12:40 Time of Day (09.10.2005)
Figure 8 (a) Pump power (black) as fluctuation of tracked solar irradiance (solid gray) throughout the solar day on 07.10.2005. The solar irradiance measured on a horizontal surface is also plotted (dotted gray). This results in (b) a varying feed flow, (c) transmembrane pressure (TMP), and (d) flux when using the BW30 membrane. Higher resolution permeate EC data (e) is plotted over short period of high solar irradiance fluctuation for NF90 membrane between 12:30 and 12:40 on 09.10.2005.
The DC power produced by the modules closely is electronically optimized to power the positive displacement pump. Feedwater is sucked through an ultrafiltration (UF) prefilter at a pressure of about —0.6 bar. The resulting feedflow reaches a maximum of about 400 L/h between 10:00 and 16:00 h (Fig. 8b) while the transmembrane pressure (TMP) is typically around 10.5 bar during this period, as shown in Fig. 8c. Under full sunlight, the flux is around 16L/(m2h) (Fig. 8d), which corresponds to a daily permeate production of 1.1 m3 with the Dow Filmtec BW30 membrane at an average permeate EC of 0.28 mS/cm. The retention was over 96% on average, while recovery was 28%. The average SEC for this experiment over the whole solar day was 2.3 kWh/m.
Similar experiments performed with other RO membranes including Dow Filmtec NF90, Hydranautics ESPA4, and Koch TFC-S yielded interesting results. Under similar solar conditions, the system produced 1.4 m with the NF90 membrane, albeit at a slightly higher permeate EC (0.52mS/cm). The performance with the ESPA4 membrane looked very promising, as even on a rainy and overcast day, the system still produced 0.85 m of permeate that exhibited a permeate EC of 0.81 mS/cm, which is only fractionally over the Australian Drinking Water Guideline value of 0.78 mS/cm, which is equivalent to 500mg/L TDS (ref). When using the TFC-S membrane, the system was not able to produce was of good quality (permeate EC — 2.1 mS/cm).
It was noted, however, that during periods of low power availability, the permeate EC value occasionally exceeds the guidelines. This is a result of stagnated water being flushing out of the system during periods of cloud cover. Fig. 8e examines this effect in more detail, showing 10 min period that was recorded with higher resolution conductivity data. This shows that despite a sudden drop in solar irradiance by 50%, the fluctuation in permeate quality is minimal. These encouraging results indicate show the ability of the system to perform well under partial cloud coverage. Further research is currently underway to characterize the system further under a wide range of fluctuating power conditions.
Was this article helpful?