Advances in desalination energy recovery technologies

Bennett looks at the current market for desalination equipment, the types of desalination technology available and associated pumping requirements and the types of energy recovery devices currently available. He then evaluates energy efficiencies and predicts how future developments are likely to increase efficiencies still further alongside a range of other developments that are driving the cost of desalination, especially plants containing membrane systems, on a downward trend.

Figure 1. Reverse Osmosis occurs when the external pressure applied exceeds the osmotic pressure. Image courtesy of Designua/Shutterstock.


The desalination of seawater (see Figure 1) to produce fresh water for human consumption can be achieved by using thermal or membrane processes, or a hybrid combination of these two process types.

We shall firstly briefly consider thermal technologies. In the late 19th Century, the first major technical advance in desalination technology was the development of the multiple effect distillation (MED) process. Here, preheated feed water flowing over tubes in the first distillation ‘effect’ (or tank) is heated by prime steam, resulting in evaporation of a fraction of the water content of the feed. A process of evaporation-plus-condensation is repeated many times to produce the product water from effect to effect, hence the term ‘multiple effect.’ In the mid-1960s multistage flash (MSF) distillation became popular. Here, a flashing-cooling process is repeated from one stage to the next stage to produce the product water.

The first reverse osmosis membrane was produced at about the same time as MSF was being developed. Reverse osmosis membranes are based around the principle of utilising osmosis, a natural process involving water flow across a semi-permeable membrane barrier. Osmosis is selective in the sense that the water passes through the membrane at a faster rate than the dissolved solids. The difference of passage rate results in the separation of water and solids. The direction of water flow is determined by its chemical potential, which is a function of pressure, temperature and concentration of dissolved solids.

Pure water in contact with both sides of an ideal semi-permeable membrane at equal pressure and temperature has no net flow across the membrane because the chemical potential is equal on both sides. If a soluble salt is added on one side, the chemical potential of this salt solution is reduced. Osmotic flow from the pure water side across the membrane to the salt solution side will occur until the equilibrium of chemical potential is restored. Equilibrium occurs when the hydrostatic pressure differential resulting from the volume changes on both sides of the membrane is equal to the osmotic pressure. This is a solution property that is independent of the membrane.

Application of an external pressure to the salt solution side equal to the osmotic pressure will also cause equilibrium. Additional pressure above the equilibrium point will raise the chemical potential of the water in the salt solution and cause a solvent flow to the pure water side, because it now has a lower chemical potential. This phenomenon is called reverse osmosis (RO) (Figure 1). Applying this RO principle to treating seawater is challenging as high pressures (60-70 bar) are required to overcome the osmotic pressure.

In the early years of RO desalination, positive displacement and centrifugal pumps provided 100% of the energy to power a seawater RO (SWRO) plant, but innovations in the field of energy recovery have improved energy efficiency significantly over the last couple of decades. Waste energy from RO systems can currently be recovered from the concentrate flow, and can account for 25-30% of the energy required to overcome the osmotic pressure of seawater (Figure 2). This lowers the total energy requirement of desalination plants dramatically. Energy is recovered by utilising isobaric energy recovery device (ERD) technology.

Figure 2. An unlimited but expensive source for production of freshwater. Image courtesy of Longjourneys/Shutterstock.



The amount of new desalination capacity that came on line during 2013 was 50% more than previous year’s total, according to data from the International Desalination Association (IDA) and Global Water Intelligence. Desalination plants with a total capacity of 6 Million m3/d became operational during 2013, compared with 4 Million m3/d in 2012. Data has yet to be confirmed for the last twelve months, but preliminary information suggests that the total capacity of plants that are online or under construction exceeds 82 Million m3/d.

Whilst the 2013 growth rate was somewhat lower than 2010, when 6.5 Million m3/d of new capacity was installed globally, clearly the demand for desalination continues to grow. From 2010 - 2013, 45% of new desalination plants were ordered by industrial users such as power stations and refineries, while in the previous four years, only 27% of new capacity was ordered by these industrial water users.

Industrial applications for desalination grew to 7.6 Million m3/d for 2010-2013 compared with 5.9 Million m3/d for 2006-2009. Seawater desalination continues to represent the largest percentage of online global capacity at 59%, followed by brackish water applications at 22%, river water projects at 9%, and wastewater recovery at 5% and pure water systems at 5%.

The world’s largest SWRO seawater desalination plant at the time of writing (January 2015), the Ras Al-Khair plant in Saudi Arabia, produces 1,036,000 m3/d, sufficient to meet the daily drinking water requirements of around 3.5 million people. The plant produced its first freshwater in early 2014, although the project is actually scheduled for completion in December 2015. As the world’s largest hybrid plant, the project uses both membrane technology (RO at 309,360 m3/d) and thermal technology (MSF with a capacity of 727,130 m3/d). This plant also features the largest single MSF trains composed of 8 units with capacity of over 91,000 m3/d each. The RO plant has 17 trains.

The Ras Al-Khair plant is dual purpose in that it produces fresh water for drinking but also power in the form of electricity. It has an export production capacity of 1.025 million m3/d desalinated water and an electricity production capacity of 2,400 MW, providing 1350 MW for the nearby Maaden Aluminium Complex, 1050 MW to the Saudi Electricity Company, and about 200MW for internal consumption on site.

The combined cycle power plant at Ras Al-Khair is one of the more efficient power plants in the world. The total length of the double transmission lines from Ras Al-Khair plant to Riyadh and Hafr Al-Batin region will be 1,290 km. The cost of Ras Al-Khair desalination and power plant project and the transmission lines from the plant has so far reached in excess of a massive US$ 6 Billion, according to the IDA.

The largest MED thermal desalination plant in the world is currently the Jubail Water and Power Plant in the United Arab Emirates, a Marafiq plant, built by SIDEM with an 800,000 m3/d production capacity from 27 MED units. The cost for this plant was US$ 1 Billion. This is also a dual purpose plant generating 2744 MW electricity in addition to desalinated water.

The largest hybrid MED-RO plant is the Fujairah II project, also in the United Arab Emirates, constructed by SIDEM and Veolia as a green field development producing 2000 MW of power and 591,000 m3/d of drinking water. The hybrid system includes five high-efficiency gas turbines operated in combined cycle mode. The Fujairah I project, owned by Emirates Sembcorp Water and Power Company and commissioned in 2004, comprises a hybrid MSF-RO system again combined with power production with an electricity generation capacity of 893 MW and a seawater desalination capacity of 455,000 m3/d.

The largest membrane-only SWRO plant so far has been built by IDE Technologies: the 624,000 m3/d Soreq SWRO plant near Tel Aviv, Israel. It came on line in October 2013. This plant has the unique feature of 16” membrane elements installed in vertical pressure vessels. This compares with the widely accepted, traditional design of 8” diameter membranes installed horizontally.

The Ras Al-Khair, Fujairah II and Soreq SWRO plants all incorporate isobaric ERD units. The first major contract to incorporate the isobaric ERD was awarded for the Ashkelon SWRO project in Israel in 2003. The Ashkelon plant was the largest in the world at the time and was developed as a BOT (Build-Operate-Transfer) project by a consortium of three international companies: Veolia water, IDE Technologies and Elran.

In March 2006, the Ashkelon project was voted "Desalination Plant of the Year" in the Global Water Awards. Since that time, many more ERD projects have been awarded, the technology becoming standard on SWRO projects.

However, Andrews (2010) comments that the developers of earlier SWRO plants were nervous about the reliability and performance of isobaric ERDs. They favoured a safer conservative approach with lower capital expenditure on ERD units in exchange for higher energy losses and higher maintenance costs.

Figure 3. Isobaric energy recovery devices recover energy that would previously have been lost in the concentrated seawater waste stream.


Energy recovery

In the SWRO process the high pressure pump is by far the main consumer of power because of the need to pressurise the feed water to 60-70 bar to overcome the osmotic pressure of seawater. The pumps are generally multistage ring section type. Variable frequency control, an important energy saving measure, is generally adopted as an international design standard for all the seawater booster and energy recovery booster pumps.

High pressure pumps are one of the main process utilities in SWRO applications and they offer a substantial potential for energy savings. For seawater RO, the use of pressure centres with larger pumps serving more than one RO train in parallel offers the possibility of having higher efficiencies and a great potential for efficiency improvement and energy reduction.

Depending on seawater salinity, according to the IDA, the energy footprint of a modern SWRO system is 3.5-6 kWh/m3. The more water we recover per unit of seawater the more energy we need, but we can theoretically reduce down to 0.7 kWh/m3 if the water recovered is only 20% of the seawater mass. This energy may be supplied as electricity only or as electricity and heat.

Ultra-filtration (UF) and micro-filtration (MF) for SWRO pre-treatment is taking over from traditional sand and media pre-filtration and this offers the possibility of in-line UF or MF system configurations where seawater is pressurised from upstream of the membrane system with the filtered water passing directly to the inlet of the high pressure pumps prior to RO treatment. In this case the membrane pre-treatment feed pumps become an important utility as these pumps need to be variable speed and both the size and duty necessary to feed the RO system.

Stover (2005) describes the history of isobaric ERD development. The first patent for a rotary isobaric device was filed by Cheng in 1965. Various developments through the 1960s to 1980s included a large chamber device for SWRO and an RO-isobaric device for use on lifeboats,

Reflecting the rapidly expanding market for membrane desalination technology highlighted above, a number of companies now manufacture ERD units including AquaLyng, Danfoss, Energy Recovery, Flowserve, GE Water, KSB and Spectra. Devices are either sold as separate process units or they are engineered within multiple process units alongside other components of the overall SWRO system.

The ‘DWEER’ from Flowserve is an independent ERD that can process concentrate flows of 350 m3/h at pressures up to 75 bar, those which we have shown are typically required for high pressure SWRO systems. The units are designed to pass objects up to 50 µm without incurring damage, and they have a low noise level of 83 dB. Flowserve say that their unit attracts extremely low maintenance costs, with a 25-year service life contributing to a low total cost of ownership.

Marketed as ‘A leap forward in the science of energy recovery’ the latest PX-Q Series from Energy Recovery includes the PX-Q300, another independent ERD with low noise levels. The units are designed to be engineered into any size of RO desalination plant, and they have a guaranteed efficiency of 97.2%. Energy Recovery claims that the unit has the lowest lifecycle cost of any ERD on the market.

The Recuperator, a further independent ERD manufactured by Aqualing, is currently undergoing third generation development. The device has only been installed on Aqualing projects to date but it is due to be released as a third party product, and Aqualing claim it will allow waste energy recycling efficiencies of up to 98.5%,

These independent devices use the concentrate flow from the RO membranes to pressurise pre-treated seawater in a sequential process regulated by the concentrate flow from the RO membranes (Figure 1).

The Aqualing device, for example, consists of vertically standing pairs of duplex stainless steel chambers that work alternatively in a compression-transfer and decompression-discharge sequence. Pre-treated seawater comes from a pressurised feeding tank that ensures a constant flow and pressure into the SWRO system. Alternatively, a regulated flow and pressure to the system can be achieved by including a low pressure feed pump.

Isobaric ERDs allow the pressurised concentrate to ‘recycle’ energy back to the RO membranes in conjunction with a high pressure pump so that the concentrate is replaced with feed water at an identical flowrate (Figure 3).

The Clark Pump is a two-in-one combined unit comprising an energy recovery device and pressure boosting unit developed and manufactured by Spectra. It is supplied by a flow of low-pressure water provided by a separate pump. The Clark Pump boosts the pressure of the feed water to the RO membrane.

GE’s IPER system is another example of a two-in-one integrated high pressure pump and ERD system specifically developed for their1000 m3/day SeaPRO SWRO system. IPER is a highly efficient desalination pumping solution, based on positive displacement (PD) technology. Unlike other PD pumps, the water displacement cylinders are not driven by a large, fast rotating crankshaft. Instead, they are driven by a highly efficient hydraulic pump and cylinders, allowing for smooth and slow operation, which minimises vibration and wear and tear. A separate energy recovery device is not required – the ERD component is built into the unit.

Danfoss claim that their patent-pending iSave unit was the first integrated isobaric ERD with its own positive displacement booster pump and electric motor. This three-in-one system brings the process management benefit of fail-safe flow control without the need for separate high-pressure flow meters and their associated control.

The KSB SalTec N unit is an example of a four-in-one integrated system, with an additional low pressure booster pump to provide a controlled suction flow to the high pressure pump. Instead of two electric motors for the high pressure pump and the low pressure booster pump, the system only requires one electric motor drive. The SalTec N was specially developed for use in decentralised, mobile container units. KSB claim that the system uses 30 % less energy compared to an equivalent RO plant without ERD technology.


The above discussion highlights that leading ERD manufacturers claim efficiencies of over 97% which could draw us to the conclusion that there is little scope for improvement of these devices and they are already optimised for high efficiency. However, Andrews points out that there is no industry-agreed formula for calculating efficiency, so comparing claims between companies can be complicated and misleading although producing an industry standard for ERD efficiency would be difficult to achieve. However, the losses of isobaric ERDs are known, so we can say that if such losses are zero, then the efficiency would be 100%. But how do we get closer to 100% efficiency?

The individual energy losses in an RO system with an isobaric ERD system comprise the following components:
• High Pressure Differential Pressure
• Low Pressure Differential Pressure
• Leakage (sometimes called lubrication flow)
• Mixing

Andrews carried out an analysis of two leading ERD devices and how they would perform in a 50,000 m3/d SWRO plant with theoretical improvements to these losses accomplished. He concluded that various efficiency improvements would add 50% (US$ 1.9 Million) to capital expenditure (CAPEX) but there would be a significant saving of US$ 8.8 Million on the net present value calculation. This equated to a saving representing 4.6 times the extra investment, corresponding to a breakeven period of less than three years along with a reduced cost of water produced per volume water treated. This analysis demonstrates what could be achieved by increasing efficiencies still further above current levels.

Since the payback period is in Andrews’ calculation is less than three years, it is clear that even customers with a short time frame and accounting period should consider such an Improved ERD – if it becomes available. Reduced operating cost (OPEX) on the Improved ERD would result in further savings but it still would be tolerable to consider increased CAPEX on the Improved ERD.

The future

Our brief evaluation of the competitive ERD market above has shown that it should be possible to make isobaric ERDs with reduced losses and increased efficiency. The SWRO market needs isobaric ERDs with reduced losses, even at higher CAPEX, and preferably lower OPEX.

Now that isobaric ERDs are a proven technology, Andrews maintains that developers should take a longer term view and invest in improving efficiencies.

Andrews lists a number of areas for ERD manufacturers to consider in research and development activities. Improving the integrity of the high-pressure ERD seal will be critical to eliminate leakage and ensure long term integrity. The brine-feed water barrier will also be an important focus along with the need for pulsations and vibration (and the associated noise) to be reduced. Operational control in multi-unit systems will be important along with addressing corrosion problems. Optimising unit capacity alongside SWRO configurations will also be important.

In large scale SWRO plants, defined as those producing over 250,000 m3/d of fresh water, it is clear that we will continue to see a further reduction in energy costs per unit desalted water, due to these more efficient ERDs being developed alongside the introduction of new membrane types (that are more resistant to fouling), larger membrane housings and faster flows in RO plants.

It is likely that we will see the efficiency of all desalination plants significantly increase, not just SWRO systems. For example, the IDA predicts that we will see large MED plants with gain-output ratios (GOR) of 15-20. This means that they will produce 15-20 tons of water per each ton of steam used, compared to a typical GOR of 10 today. MED capacity is also likely to grow.

In SWRO we will most likely see established use of larger 16” diameter modules, compared to the 8” diameter standard, their use bringing a further reduction in cost and space requirements. Despite improvements in thin-film composite RO membrane design and construction over recent years, there are still shortcomings which will be improved. For example, we can envisage high temperature membranes which can increase fluxes and recovery at 50 degrees Celsius and above. This will bring more challenges for pumping systems and ERD units in terms of material selection at those higher temperatures.

Larger RO trains and improved fouling resistance will be part of the effect of increasing energy efficiency, reliability, and the environmental impact of these much larger SWRO systems, alongside implementing more efficient ERD units.

Pumps are a crucial component for a desalination plant regardless of the technology used. In the past, desalination plants, both RO and thermal, have suffered from shortcomings in pump selection due to poor material specification and often poor design of the system without considering cavitation, vibration, noise and performance issues.

Often CAPEX reduction has been the driver for the selection of a poor pumping system without or with minimal energy efficiencies. The situation is how transformed and we can only envisage an ongoing increasing market for desalination technology and associated energy recovery devices.