A treatability study puts your real wastewater through candidate treatment steps in a laboratory, so you learn what removes what, at what dose, and what residue it leaves, before anyone sizes equipment. Pilot testing then runs the chosen steps continuously, usually on site, to see how they cope with the variability, fouling and upsets that a beaker never shows.
Both cost money and time, and both can be done badly: the wrong test, an unrepresentative sample, or a pilot run until the budget ends rather than until the question is answered. Brine Consulting is an Amsterdam-based independent advisory firm for industrial water. Because we sell no treatment equipment, no test we plan has a process of ours to favour. We write the programme around the design decisions it has to support, specify what the laboratory or pilot provider must run and report, and turn the results into a basis of design. To have a programme priced, send us the stream and the question you need answered.
Where testing fits: after the audit, before the design
An industrial wastewater audit tells you what water you have: flows, sources, loads and how they vary.
A feasibility study judges, mostly on paper, which treatment routes are worth pursuing and roughly what they cost.
Treatability and pilot testing check whether the shortlisted steps actually work on that water, and set the numbers the design will rest on.
Design then fixes the basis, the balances and the specification that suppliers bid against.
Skip testing and the design inherits assumptions. Test too early, before the audit, and you may test the wrong stream.
Three tiers, each roughly five times the cost of the one before
The clearest public framework for tiered testing comes from the US Environmental Protection Agency. Its Guide for Conducting Treatability Studies under CERCLA (EPA/540/R-92/071a, 1992 final edition, which updated the 1989 interim guide; we found no later edition) was written for contaminated-site clean-up, but its logic carries straight over to industrial wastewater.
| Tier | What it involves | Data you get | Testing time | Cost range given by EPA in 1992 (US dollars) |
|---|---|---|---|---|
| Screening | Small batch tests on the bench, with off-the-shelf equipment, single or duplicate | Qualitative: does it work at all? | A few days | 10,000 to 50,000 |
| Selection | Bench or pilot tests, batch or continuous, in duplicate or triplicate | Quantitative performance and cost at the conditions tested | Days to weeks at bench scale | 50,000 to 250,000 |
| Design | Mostly pilot or full-scale equipment in the field | Detailed design, cost and performance data | Weeks to months | 250,000 to 1,000,000 |
Ignore the dollar amounts, which are more than 30 years old and from a different sector. The ratio is what matters: each tier costs roughly five times the one before. In our view, most sites should therefore answer every question they can at the lowest tier, and keep the pilot for what only continuous operation can show.
Which test answers which question
A common way to waste a test budget is to choose the test before writing down the question. The table pairs common design questions with the standard test that answers each, how long it takes, and what it cannot tell you.
| The design question | Test that answers it | How long | What it cannot tell you |
|---|---|---|---|
| Which coagulant, flocculant, dose, pH and order of addition removes solids, colour or precipitated metals? | Jar test, ASTM D2035-19 (current edition) | Hours to a few days | How a continuous clarifier performs; the practice covers coagulation followed by gravity settling, so flotation needs its own trial |
| Will this effluent suppress the bacteria in an activated sludge plant? | Respiration inhibition, OECD Test No. 209 (2010 edition) | Oxygen uptake measured after 3 hours of exposure | Whether the biomass adapts over weeks of continuous feed |
| Will the organic load break down at all, given time? | Inherent biodegradability, OECD Test No. 302B (Zahn-Wellens/EVPA, 1992) | Up to 28 days | A design loading rate: it is a batch test, with removal tracked as dissolved organic carbon or COD |
| Can a biological process meet the limit continuously, and at what loading? | Continuous bench reactor or on-site pilot | For aerobic systems, HDR advises at least three sludge ages after acclimation, typically 2 to 3 months | Seasonal swings or production campaigns that fall outside the run |
| Will particles in the feed foul a reverse osmosis membrane? | Silt density index, ASTM D4189-23 (current edition) | A short filtration test | An absolute particle count; the method applies only to low-turbidity water below 1.0 NTU, and results change with temperature and filter make |
| Will the membranes hold their output over time? | Membrane pilot, with data normalised following ASTM D4516-19a (current edition) | Weeks to months | The standard is written for brackish water and seawater; its use on wastewater is not assured, so the normalisation method should be agreed in advance |
| How far can a brine be concentrated before scale, foam or solids appear? | Bench evaporation of the real brine in steps of concentration, then a pilot if the limit is close to the target | Days to weeks at bench scale | Effects that build up over time, such as recycle accumulation and cleaning frequency |
The first three tests are cheap and fast. Run them before committing to anything continuous, and use their results to decide what the longer tests need to cover.
When is a pilot worth paying for?
A pilot earns its cost when the answer depends on time or on variability:
the feed changes with season, product campaign or cleaning cycle, and the bench sample cannot represent all of it;
the risk is cumulative: biological acclimation, membrane fouling, scaling in an evaporator, or impurities building up in a recycle loop;
the process is unfamiliar for this kind of water, with no reference plant on comparable feed;
a regulator or lender will only accept demonstrated performance.
The last point can pay for itself. In a Burns & McDonnell white paper on a drinking water project in Wichita, Kansas, the state’s standard limited dual-media filters to 4 gallons per minute per square foot, while the design used 5. The state regulator said the higher rate could be approved if a pilot showed the design consistently met turbidity limits, and required it to run continuously for 30 days with round-the-clock monitoring. The paper reports that value engineering reduced the number of filters from 20 to 16. The same paper notes that pilots range from about a month to more than a year.
A bench programme is usually enough when the process is well proven on similar water, the feed is stable, and the question is a dose or a chemical choice rather than long-term behaviour.
An example test plan for a saline stream
This is a hypothetical plan for a common case with saline wastewater: reverse osmosis reject that a site wants to concentrate further before evaporation or disposal. It shows how each step decides whether the next one is needed.
Check the data before testing anything. Validate the existing analyses first: the charge balance error should be small and the summed ions close to the reported TDS. Our Water Analysis Validator runs both checks. Then work out the concentration factor the project needs with the Brine Recovery and Concentration Factor Calculator, so every later test has a target.
Screen pretreatment on the bench. Jar tests compare softening or precipitation chemistry for the ions that limit concentration. Stop here if no chemistry brings the limiting ions down far enough at a reasonable dose, and rethink the target.
Concentrate real brine on the bench in steps. Record at which concentration solids form, whether the brine foams, and the composition of what remains. If the limit is comfortably beyond the target and repeat samples agree, take it into the basis of design with a stated margin.
Pilot only what is still uncertain. If the bench limit sits close to the target, or samples disagree, a continuous pilot tests the effects that need time: recycle build-up, scaling between cleanings and cleaning recovery.
Our ZLD guide chapter on laboratory and pilot testing covers the evidence each test scale can and cannot provide for high-salinity streams in more depth.
What makes test results usable for design
The question and the pass mark are written before testing. State outlet concentrations to reach, not only a percentage removal. Removing 80% of a high-strength load can still miss a discharge limit.
Samples are representative and documented: where and when each was taken, what the plant was doing, and how the sample was stored and transported.
Residues are measured, not assumed: sludge volume and dewaterability, concentrate flow and composition, and chemical consumption per cubic metre treated.
The test closes its own balance. Water and the key constituents in should equal what leaves, within a stated error.
The report separates measured results from design estimates, and states the range of conditions tested so nobody applies the result outside it.
Who should run the pilot?
Equipment suppliers often offer pilot units, and their pilots are valuable: they know their equipment and can run it well. The risk is that the supplier also writes the protocol and the success criteria, so the test shows what their process does best. Whoever runs the unit, agree in advance the feed, the duration, the measurements, the normalisation method and the pass mark, and ask for the raw data as well as the summary. Our part is that agreement: we draft the protocol and pass marks before the unit arrives, review the data as it comes in, and say when the question has been answered so the pilot can stop.
Two questions to settle before testing
How long does testing take, and what drives the cost? A programme is as long as its slowest essential test, from hours for a jar test series to months for a continuous biological run (see the table above). Cost follows the same logic: the number of questions, the samples and analyses each one needs, and above all whether any question needs a continuous pilot, which sits a cost tier above bench work.
Do we need testing if the supplier guarantees performance? A guarantee is only as good as the feed it is written against. If the feed in the contract is not the feed you have, the guarantee may not apply when you need it. Testing tells you which feed to write into it.
Request a treatability study quote
Tell us the stream, the treatment question and any data you already have. We will reply with the tests that answer it and the scale each one needs. For sites outside the Netherlands and the EU we plan and review the programme remotely or under a defined engagement, with the testing done by a laboratory or pilot provider you appoint. Send us your project details for a quote, book a scoping call, or write to christos@brineconsulting.com.
Written by the Brine Consulting advisory team.
The Brine Consulting advisory practice includes Osis G. Kalache, P.Eng., Senior Consultant, Wastewater Process Engineering and Optimization, whose expertise combines operational know-how, data analytics, AI-enabled optimization, piloting and innovation management across demanding industrial and oil and gas applications, and Dimitra Aravani, External Advisor, Industrial Wastewater and Water Reuse, who specializes in industrial wastewater treatment, water reuse, resource recovery and water efficiency strategies, with more than a decade of experience supporting industrial and municipal clients. Meet our experts.