Towards Energy-Efficient Reverse Osmosis
Reverse Osmosis (RO) represents the state of the art in desalination technology today. Owing to its affordability and relatively low energy consumption, it has clearly overtaken thermal desalination technologies such as multistage flash (MSF) and multiple effect distillation (MED) (Weaver, Howells, & Brown,
2018). The energy usage of a desalination process is measured in terms of its specific energy consumption (SEC) [kWh/т3], which is defined as the energy consumed per volume of water produced. Though the SEC of RO desalination is much lower than that of thermal desalination, it is still several times higher than for other types of water treatment such as standard treatment of surface water by coagulation and filtration (Voutchkov, 2018a). The energy usage of RO desalination remains significant because it accounts for about half the cost of the product water. Moreover, since the energy to drive desalination is typically obtained from burning of fossil fuels, RO has a high carbon footprint compared to conventional water treatment. For these reasons, there is much interest in reducing the energy consumption of RO desalination to make it even more affordable and to reduce its environmental impact.
It is possible to define an ideal minimum SEC in the thermodynamic sense for any desalination process, including RO. The thermodynamic minimum SEC is independent of the technology used to carry out the process, applying in principle to thermal (e.g. MED. MSF), pressure-driven (e.g., RO), and electrical technologies (e.g. electrodialysis) alike. No matter how much effort is put into developing these technologies, they will never beat the thermodynamic ideal minimum SECjdeai - they can only approach it. Although any one of these technologies could in theory approach the minimum, in practice some are able to approach it much more closely than others. In particular, RO technology has been quite successful in gradually approaching the ideal minimum SEC.
The textbooks on chemical thermodynamics describe the concept of chemical potential as the work needed to remove a unit of a certain species from a solution (Smith, 2004). When this general concept is applied to the case of removing water from a solution of salt in water, chemical potential becomes equivalent to SECidea|. Moreover, there is a fundamental relation between chemical potential and osmotic pressure, such that the SECideal is also equivalent to the osmotic pressure of the salt solution to be desalinated. The only difference lies in the units conventionally used to measure these two quantities. Thus, for standard seawater containing 3.5% salt, the osmotic pressure is 26 bar=2.6 MPa resulting in a SECideai of 2.6/3.6=0.72 kWh/т3. This means that no future desalination technology will ever be able to extract a cubic meter of freshwater from the sea while consuming less than 0.72 kWh of energy. An exception arises if the seawater is taken from an area of the sea, such as an estuary, that is less salty than standard seawater. This is because the osmotic pressure (and thus SECidea|) varies more or less in proportion to the salt concentration. Therefore, the SECideal for estuarine water may be less than 0.72 kWh/nr3.
There are many reasons why SEC exceeds the ideal minimum in real desalination plants. A fundamental reason relates to the recovery ratio (/•), i.e., the fraction of freshwater that is recovered from the incoming seawater. The theoretical concept of chemical potential applies to a hypothetical situation in which a small volume of freshwater is recovered from an infinite reservoir of seawater, such that /■ is virtually zero. This is not the case in practice for several reasons. For example, the seawater has to be pumped onshore to reach the desalination plant. This incurs costs of intake pipework and power supplied to the pump. To avoid pumping an excessive amount of seawater, the recovery ratio of most desalination plants is kept below 50%. As water is extracted from a finite amount of seawater, the osmotic pressure must go up and the corresponding SECjdeai for a recovery ratio r>0 increases according to (Qiu & Davies, 2012a):
where SEQdea|i() is the value of SECideal at r = 0. (This equation is obtained by a process of mathematical integration and is based on the assumption that the osmotic pressure increases in proportion to the salt concentration, which is quite accurate, except at high concentrations). For a recovery ratio of 50%, Equation 11.1 gives SECjdeai = 1 kWh/т3, whereas (as shown subsequently) real RO desalination plants have an SEC of more than twice this value.
A second reason for this disparity between real and ideal SEC relates to the membrane area and capital cost. In the RO process, permeate water is driven through a selective RO membrane by means of a transmembrane pressure Др that must exceed the osmotic pressure Дл. The difference (Др - Дл), called the net driving pressure, determines the flux Jw of water through the RO membrane (i.e., the flow per unit membrane area) according to the equation:
Here, A is the permeability - a property of the membrane that depends on the technology used in its fabrication. The value of A is always finite, such that the net driving pressure has to be above zero to maintain a certain flux. The flux should not be too small, otherwise this results in a large area of membrane being needed for a desired flow of product water, with associated increase in the cost of the desalination plant. On the other hand, whenever (Др - An) exceeds zero substantially, there is an energy penalty in SEC because the thermodynamic ideal case is to apply a pressure across the membrane only jusl exceeding An. This shows that there is a trade-off between minimizing SEC and minimizing the cost of the RO plant.
To obtain a better trade-off in this respect, one approach would be to select a membrane with higher value of permeability A. Indeed, RO membrane manufacturers have gradually increased A, such that fluxes from RO membranes today are in practice nearly twice the values possible 20 years ago. Here, however, there is another trade-off to consider that relates to the ability of the RO plant to reject salt. The flux Js of salt through an RO membrane is, unlike the flux of water, virtually independent of the transmembrane pressure. It depends instead on the difference Ac in concentration across the membrane.
where В is the salt diffusion coefficient which, like A, is a property of the membrane. It is not desirable for Js to be too high, because this would result in too much salt reaching the product water, making it unsuitable for drinking. We would then say that the RO plant has not rejected enough salt. The difficulty in selecting membranes with higher A, however, is that such membrane also tend to have higher B. As such, they are not sufficiently selective in excluding salt and allowing water to pass through. This is referred to as the selectivity-permeability trade-off that is a general feature of membrane separation processes (Werber, Deshmukh. & Elimelech, 2016).
In summary, we have introduced the idea of an ideal minimum SEC for a RO desalination plant and highlighted some fundamental trade-offs to consider when aiming towards this ideal minimum. Not only do we wish to minimize SEC of the plant, we also wish to minimize its capital cost, increase freshwater recovery, and increase salt rejection. These four objectives tend to conflict. In the rest of this chapter, we explore in more detail the reasons for nonideal SEC and the advances that are being made to lower SEC in the various areas of technology that are used in desalination plants. In the next section, we take a closer look at the SEC of real RO plants.
ENERGY CONSUMPTION OF RO PLANTS
The SEC of RO plants is constantly decreasing. In the 1990s, it ranged from 5 to 10 kWh/т3 (Amy et ah, 2017). Today, the industry average SEC has been reduced to 3.1 kWh/т3, with modem plants achieving between 2.5 and 2.8 kWh/т3 (Voutchkov, 2018a). Although these values are much better than those of the 1990s, they are still more than twice the ideal value of 1 kWh/m3. It is therefore important to investigate where and how energy is consumed in a typical RO plant, as this will help in understanding what further progress can be made to reduce the SEC.
The energy consumption of a RO plant depends on its overall design and the design of its components. Other factors such as age, mode of operation, and feed water conditions also have an influence. An RO plant consumes energy in the form of electricity. The following breakdown of electrical energy consumption has been provided by Qasim et al. for a typical seawater RO desalination plant (Qasim, Badrelzaman, Darwish, Darwish, & Hilal,
2019). Figure 11.1 shows that the RO system uses 71% of the total energy, while the remaining 29% is used by the intake, pretreatment, delivery, and other auxiliary systems.
The intake system normally consists of a low-pressure pump drawing seawater through a submarine pipe and inlet grating. It has to collect seawater without ingesting wildlife and litter - and without endangering marine traffic or bathers. The inlet should be sufficiently distant from outlet pipes of other plants or the desalination plant itself, otherwise it would suck in contaminated seawater. Typically, the intake pump consumes about 5% of the total plant power. This depends on the pump selected and the pipe design, which in turn is influenced by the siting of the plant and local bathymetry.
Pretreatment is the second most energy intensive process in the RO plant, after the RO system itself. It can include conventional methods such as
FIGURE 11.1 Energy consumption distribution of a typical RO plant (Qasim et al., 2019).
coagulation, flocculation, sedimentation, and media filtration or advanced methods such as ultrafiltration. Pretreatment uses about 11% of the total plant energy consumption to get the water to the required standard prior to admission into the RO system. The pretreatment system is crucial to prevent excessive fouling, membrane damage, and to ensure membrane longevity.
Product Water Delivery
The permeate water must be post-treated to meet the drinking water standards and then pumped into the water distribution network. Post-treatments may include boron removal, remineralization, and chlorine removal. This process requires approximately 5% of the entire plant energy consumption.
Other Processes and Services
Commercial RO plants have a compound design that relies on multiple subsystems and processes that contribute to the water production cost and overall energy consumption. These processes, which include chemical dosing, membrane cleaning, concentrate discharge, waste disposal, as well as plant lighting, heating, instrumentation, and routine maintenance activities, could make up nearly 8% of the overall RO plant energy consumption.
The major energy load in the desalination system is the high-pressure pump (HPP) that drives the RO process. The HPP normally provides a pressure in the range 50-70 bar. Around 50% of this energy is required to overcome essential osmotic pressure difference; 20% is lost to inefficiencies in the HPP and the energy recovery device (ERD); 2.5% is lost through flow friction in the concentrate and permeate channels; and 2.5% is lost owing to concentration polarization. The remaining 25% is consumed by the HPP to overcome:
• Pressure losses corresponding to hydrodynamic losses in the membrane. This energy depends on the permeability A (see section 11.1).
A higher value of A would result in lower energy being required.
• Pressure losses corresponding to the increase of salt concentration, and therefore osmotic pressure, from the inlet to the outlet of the RO modules. Because the applied pressure must at all positions exceed the maximum osmotic pressure occurring at the outlet, there is an excess pressure at the inlet. The resulting wastage of energy is discussed further in section 11.4.1.
Because RO systems consume the most energy, the rest of this chapter examines more closely the different factors affecting their energy consumption, beginning with RO membranes and the prospects to improve plant performance through use of new and better membrane technology (Karabelas, Koutsou, Kostoglou, & Sioutopoulos, 2018).