Desalination: reconcile recovery, brine and energy use
Desalination recovery measures how much incoming water becomes product water. It does not by itself measure energy efficiency, product quality or economic return. A higher recovery percentage changes the relationship between intake and brine, but a research model still needs separate water, salt and energy balances before treating that change as an operating improvement.
The deliverable is a daily balance sheet with explicit metering boundaries. It should reproduce product volume, remaining brine, approximate salt concentration and energy per unit of product. Keep the example at the level of commercial evidence and accounting boundaries; it is not a design instruction or a claim that a particular recovery rate is technically achievable at a real facility.
Start with the process boundary
The US Department of Energy’s Desalination Basics distinguishes membrane-based desalination from thermal processes. In the Membrane-based Desalination discussion, reverse osmosis uses pressure and a membrane to separate water from dissolved salts. This supports identifying the separation and its energy input. The page is not a universal energy benchmark or an operating specification for all seawater plants.
Draw a box around the assets represented by each number. Does electricity cover only the membrane equipment, or also intake, pretreatment, post-treatment and product pumping? Does product volume mean water leaving the separation step or water delivered to the buyer? A numerator from the smaller equipment box divided by output from the larger delivery box produces an intensity that cannot be compared fairly with a full-site meter.
Use one period throughout the water balance. A daily average intake, month-end storage stock and annual shipment number do not close merely because all use cubic metres. Identify tanks and changes in inventory if the reporting period is not a steady-state approximation. Record whether downtime days enter the average.
Rebuild a fictional daily water balance
Assume a plant receives 100,000 cubic metres of seawater per operating day. It produces 40,000 cubic metres of product and discharges 60,000 cubic metres as brine. Ignore storage changes, evaporation and other streams for this first teaching case. Recovery is product divided by feed: 40,000 ÷ 100,000 = 40%.
Water volume closes because 100,000 = 40,000 + 60,000. That is a volume approximation for the stated case, not a statement that density differences and every real side stream can always be ignored. A more precise engineering mass balance would require suitable density and stream measurements. The purpose here is to detect missing commercial volumes before interpreting a ratio.
Suppose feed salinity is thirty-five kilograms of dissolved salt per cubic metre. Salt entering is 100,000 × 35 = 3,500,000 kilograms per day. Initially assume no salt remains in product water. All 3,500,000 kilograms then enter the 60,000-cubic-metre brine stream, producing a calculated concentration of approximately 58.33 kilograms per cubic metre.
| Stream | Water volume, cubic metres/day | Salt, kilograms/day |
|---|---|---|
| Feed | 100,000 | 3,500,000 |
| Product, initial simplification | 40,000 | 0 |
| Brine, initial simplification | 60,000 | 3,500,000 |
The zero-product-salt assumption simplifies the arithmetic; it does not certify product quality. Replace it when measured product salinity is available. If the product contains 0.5 kilograms of salt per cubic metre, product salt becomes 40,000 × 0.5 = 20,000 kilograms. Brine salt is then 3,480,000 kilograms, or fifty-eight kilograms per cubic metre. Keeping the assumption visible makes this correction reproducible.
Put energy on the correct denominator
Assume the documented site boundary consumes 160,000 kilowatt-hours per operating day and covers all the equipment included in the reported product volume. Energy intensity per product cubic metre is 160,000 ÷ 40,000 = four kilowatt-hours. Energy per feed cubic metre is 160,000 ÷ 100,000 = 1.6 kilowatt-hours. Both are mathematically valid, but answer different questions.
Calling the 1.6 figure the product-water intensity would understate this example’s purchased electricity per unit of product by 60%. Do not compare a feed-based figure with a competitor’s product-based figure and interpret the difference as efficiency. Write the denominator in the column heading, not just in a distant footnote.
Purchased electricity, gross equipment demand and net demand after energy recovery also need distinct labels. If the 160,000 figure already comes from the net site meter, subtracting an energy-recovery estimate again double counts the benefit. Ask whether auxiliary equipment and exported energy are included before calculating a revised intensity. Where the source does not say, present the ambiguity instead of inventing a net figure.
Test higher recovery without assuming free production
For a deliberately constrained sensitivity, keep feed at 100,000 cubic metres and assume product increases to 50,000, leaving 50,000 of brine. Ignoring product salt again, calculated brine concentration becomes seventy kilograms per cubic metre. If electricity were unchanged, the arithmetic energy intensity would fall to 3.2 kilowatt-hours per product cubic metre.
The phrase if electricity were unchanged is the central assumption. The balance alone does not show that the plant can achieve this recovery with the same pressure, pretreatment, membrane condition or electricity use. Higher concentration and altered operating conditions need facility evidence. Label 3.2 as an illustrative sensitivity, never a proven improvement obtained merely by changing the recovery cell.
A different boundary case concerns delivery. Suppose the original plant still produces 40,000 cubic metres, but 2,000 are consumed or lost between the product meter and buyer delivery. Delivered water is 38,000. With unchanged site energy, intensity per delivered cubic metre is approximately 4.21 kilowatt-hours. Recovery at the separation step remains 40%; delivery efficiency has changed a separate commercial denominator.
Build the evidence worksheet
| Input | Obtain | Check |
|---|---|---|
| Feed and product | Meter records for the same interval | Volumes, operating days and storage changes |
| Brine and side streams | Discharge and process records | Water balance closes within stated precision |
| Salinity | Stream-specific sampling and units | Kilograms/cubic metre versus other units |
| Energy | Meter scope and electricity ledger | Gross, net, auxiliaries and recovery treatment |
| Delivery | Product and customer meters | Separation output versus saleable delivered water |
| Change claim | Comparable before/after records | Same feed conditions and plant boundary |
Do not treat salinity units as interchangeable without conversion. A reported grams-per-litre concentration has the same numerical value as kilograms per cubic metre, but an unspecified parts-per-million figure may require a clearly stated mass basis and density assumption. Preserve the original source unit so the conversion can be inspected.
When water and salt balances disagree, first inspect period alignment, tank inventory, unreported cleaning streams and measurement precision. A difference is a question for source reconciliation, not automatic evidence of misconduct. Keep a residual column and show which missing input could explain it. Do not force the brine volume to close and then call it an independently observed measurement.
If the source supplies product and recovery but no feed meter, implied feed equals product divided by recovery. Label it derived. If energy covers an entire integrated utility and desalination’s allocation is absent, a desalination-specific energy intensity remains unknown. Allocation assumptions can be scenarios but cannot become actual facility measurements through repeated citation.
Update the worksheet when feed salinity, metering scope, recovery, delivery losses, production availability or the energy accounting method changes. Environmental discharge permissions and drinking-water compliance require their own jurisdiction-specific evidence; neither follows from this balance. The research conclusion should identify what closes, which denominator is comparable and which performance claim still lacks measured support.
Reconcile availability before annualising the day
The example uses an operating day. To build an annual volume scenario, obtain operating days or actual monthly product totals; do not multiply by every calendar day while also applying an availability rate already reflected in the daily average. A calendar-day average may already include outages, so a second availability deduction would count downtime twice.
Similarly, distinguish process recovery from plant availability. A facility can retain a 40% recovery ratio while producing less annual water because it runs fewer hours. Decompose an annual change into operating opportunity, feed rate and recovery using consistently defined inputs. Where these are unavailable, the daily balance cannot establish the annual cause.
Quality sampling also needs a time basis. A single high-salinity intake sample multiplied by full-day flow is a stress scenario unless the source establishes it represents the day. Preserve sampling frequency and averaging method. If the source’s precision cannot support a small reported improvement, show the unrounded inputs or a sensitivity range rather than adding decimal places to the final efficiency ratio. The residual should remain visible until a better measurement resolves it.
FAQ
Does a higher recovery rate prove lower electricity cost?
No. The new energy requirement, electricity price and delivery boundary need separate evidence.
Is calculated brine salinity a discharge-compliance result?
No. It is a simplified balance estimate and does not establish local environmental compliance.