Engineering explained / SALT

The sea is full of water.
The salt pushes back.

Reverse osmosis is a pressure contest. Follow the water through a membrane, turn up the recovery, and watch the far end become the hard part.

The agent layer behind the engineering

Try Reflux on your own Aspen Plus.

Describe the work in plain English. Reflux makes changes in Aspen, runs the case, and reads back the result. Windows, your own Aspen Plus V14 licence, three free runs to start.

Try Reflux Student free →Join the trial waitlist
Conceptual cutaway of a spiral-wound reverse osmosis pressure vessel with layered membranes around a central permeate collection tube.
Inside the pressure vessel: feed travels along the membrane leaves. Water crosses the membrane into the permeate path; concentrated brine leaves separately. Generated cutaway, simplified for explanation.

01 / The far end

More fresh water. Less pressure margin.

Ordinary osmosis moves water toward the salty side. Reverse osmosis applies pressure to that side to reverse the net flow. As fresh water leaves, the remaining brine gets saltier. Its osmotic pressure rises while friction eats a little of the feed pressure.

Feed pressureOsmotic pressure
Pressure along the membrane tubeAt 50% recovery the osmotic pressure rises from 27 to 54 bar.80400barINLETOUTLET
Outlet salt’s pull54.0 bar
Outlet salinity70.0 g/L
Brine per 1 L fresh1.00 L
Outlet driving pressure9.0 bar

Pressure still exceeds the salt’s pull at the outlet.

Equations and assumptions
Cᵦ = C𝒇 / (1 − R)
πᵦ ≈ π𝒇 / (1 − R)
Vᵦ / Vₚ = (1 − R) / R
Driving pressure ≈ P𝒇 − ΔP − πᵦ

R is recovery. This illustrative model assumes complete salt rejection, negligible permeate salinity and pressure, π𝒇 = 27 bar at 35 g/L, proportionality between concentration and osmotic pressure, and 2 bar of feed-channel pressure loss. Cumulative recovery is distributed uniformly along the plotted tube. Real designs require membrane transport, concentration polarization, temperature, scaling and fouling limits, and actual hydraulics.

At 50% recovery, the brine-to-fresh-water ratio is 1:1. The reel’s 1.5:1 number is a different measure: a global average across desalination technologies in a 2019 study. It is not the mass balance for a 50%-recovery RO plant. That distinction is why the slider shows one litre of brine at its starting point.

02 / The energy ladder

There is a floor. It is not the whole bill.

Salt and water can only be separated by doing work. For seawater near 35 g/L at 50% recovery, a commonly cited reversible minimum is about 1.06 kWh per cubic metre of product water. Actual electricity use depends on the plant and on what its meter includes.

  1. 1.06Physics floor
  2. ~2.3Efficient plants
  3. 2.9–5.5Reported plant range
  4. >6Early SWRO / historical

All rungs: kWh/m³ of fresh water. These are reference benchmarks, not four directly comparable contemporary designs. Intake, pretreatment, pumping, membrane separation and post-treatment boundaries differ. The historical reduction reflects better membranes, pumps, plant design and energy recovery together.

What the simple model predicts

The ideal proportional-pressure model gives 1.04 kWh/m³ for the selected salinity and recovery. The 1.06 reference uses seawater thermodynamics; the small difference is expected.

Wₘᵢₙ ≈ (π𝒇 / 36) × [−ln(1 − R) / R]

1 bar·m³ = 1/36 kWh. This is reversible separation work; it excludes real process losses and auxiliary loads.

03 / The pressure comes back

The brine leaves with energy still in it.

Generated conceptual cutaway of a ceramic pressure exchanger rotor, exposing straight channels with blue brine meeting turquoise feedwater.
A rotating ceramic channel alternately accepts and discharges the two streams. Pressure transfers through brief direct liquid contact. This conceptual illustration is not a manufacturer drawing.

A rotary isobaric pressure exchanger uses the pressurized reject stream to pressurize incoming seawater. Its quoted efficiency can be around 98% for hydraulic energy transfer. That number belongs to the device, not to the entire desalination plant.

Hydraulic work without recovery3.61kWh/m³ product
Hydraulic work with 98% recovery1.90kWh/m³ product

For the selected conditions above, recovering brine pressure reduces this model’s hydraulic duty by 47%. The pump still supplies net work.

What is included in this comparison
W₀ = P𝒇 / (36 R)
Wᵣ = ηₑ (P𝒇 − ΔP) (1 − R) / (36 R)
Wₙₑₜ = W₀ − Wᵣ

ηₑ = 0.98 and ΔP = 2 bar. This is an idealized hydraulic energy balance per cubic metre of permeate. It is not metered plant electricity; motor and pump losses, booster configuration, mixing and auxiliary loads are excluded. If the selected recovery has no outlet pressure margin, the energy numbers describe that hypothetical balance, not a feasible operating point.

Read the underlying work

Sources and model boundaries

  1. Elimelech & Phillip, Science (2011): thermodynamic limits and the energy, technology and environmental context for seawater desalination.
  2. Water (2016), Predicting the Specific Energy Consumption of Reverse Osmosis Desalination: energy accounting, reference ranges and model dependence.
  3. Energy Recovery, PX Q400 white paper (2025): direct pressure exchange, 40–50% typical recovery, and device efficiency measurements.
  4. Jones et al., The state of desalination and brine production (2019): approximately 142 million m³/day brine and 95 million m³/day product in the historical global dataset.

The interactives are educational balances. They do not size membranes or predict potable-water quality, safe brine discharge, or a plant’s operating cost. Salt rejection, remineralization and outfall design are separate engineering problems.

Build the model. Keep the judgment.

Try Reflux on Aspen Plus, or tell us which simulator should come next.

Try Reflux Student free →Join the trial waitlist

Nathan Ruberto · Reflux

Explore the engineering library →