PSLE Science Reality Lab Vol No.256 | “Pilot-Scale Test Removed 90%” — Will the Full-Scale System Always Remove 90%?

Stable internal ID: PSLE-SCI-REALITY-0256

Wait, what? A small pilot water-treatment unit is built beside a large plant. Engineers run water through it and report: “90% removal achieved.” A poster then announces: “The full-size plant will remove 90% every day.”

The pilot result may be excellent evidence. The poster still makes one extra move.

Pilot-scale testing exists because moving directly from a bench experiment to a huge real system can be risky and expensive. A pilot gives scientists and engineers a controlled way to test whether the important process relationships survive at a larger, more realistic scale. But scale-up is itself a scientific question. The full system can have different flows, mixing patterns, loading changes, maintenance conditions, start-up behaviour and day-to-day variation. The pilot result does not disappear when we remember those differences; it simply has a boundary.

This Reality Lab owns one narrow evidence-transfer job: how a Primary 5/6 learner should read a pilot-scale success claim without turning it into a promise that a full-scale system will give the identical percentage under every operating condition.

Quick Answer

No. “90% removal at pilot scale” means the tested pilot configuration achieved that result under its stated conditions. It can be important evidence that the process is promising and can support scale-up when the pilot represents the full system well. It does not automatically guarantee 90% removal in every full-scale run.

Pilot scale is between a small laboratory test and a full operating system.
A good pilot tries to preserve the important physical and operating relationships of the future full-scale process.
Some things do not scale by simply multiplying every length, volume or amount.
Real full-scale operation may face wider variation, longer durations, changing feed conditions, maintenance and unexpected disturbances.
The right question is not “Did the pilot work?” alone, but “What exactly did the pilot test, and which conditions must remain comparable for the result to travel?”

The Exact Learner Job This Page Owns

This page is not a standalone owner of engineering scale-up, wastewater treatment, fluid dynamics, chemical reactors or industrial design. Those specialist concepts belong elsewhere. It also does not replace the existing PSLE Science owners for fair tests, variables, method limits, evidence selection or conclusions.

Its job is narrower: evaluate a real-world claim that uses a pilot-scale result as evidence for a full-scale promise. The communication object could be a news article, technology brochure, school infographic, project report or product-comparison graphic.

Three Scales, Three Evidence Jobs

Scale	Typical scientific job	What it cannot automatically prove
Laboratory or bench scale	Explore mechanisms, ranges and controllable conditions using small quantities.	That the same performance will survive real operational scale.
Pilot scale	Test a more realistic process arrangement and gather evidence for scale-up.	That every full-scale operating condition has already been tested.
Full scale	Operate the system at the size, throughput and environment required in real use.	That all future conditions will be identical to the validation period.

The U.S. Environmental Protection Agency explains that laboratory-scale work is easier to control, while pilot-scale units are used to model or predict full-scale behaviour and must accurately represent the full-scale process when scale-up is the goal. EPA quality-assurance guidance also warns that important departures in process parameters between pilot and full scale can invalidate a scale-up claim.

Rebuild the Claim Object

Original composite case: A treatment process is designed to remove coloured particles from water. The pilot unit receives 100 litres per hour. Water enters a mixing tank, receives a treatment step and passes through a filter. During a five-day pilot test, the incoming water contains about 50 units of the measured material and the outgoing water contains about 5 units. The report summarizes the result as roughly 90% removal.

A future full-scale unit is planned to treat 10,000 litres per hour.

A weak reading says: “100 times more water means build everything 100 times bigger and the result stays 90%.”

A stronger reading asks what made the pilot successful: contact time, mixing, temperature, loading, filter area, flow pattern, chemical dose, start-up state, maintenance and the variability of the water entering the system.

Observed, Claimed and Inferred

Layer	Example
Observed	The pilot unit’s inlet and outlet were measured during defined tests.
Calculated	Removal was approximately 90% under those test conditions.
Supported claim	The process showed strong pilot-scale performance for the tested configuration.
Extra inference	The future full-scale system will always remove exactly 90% under any condition.

The extra inference may later become well supported. But it needs scale-up evidence. It is not contained automatically in the pilot percentage.

Why “Just Make It Bigger” Can Fail

Many processes depend on relationships, not just amounts. Imagine stirring sugar into a cup of water. Now imagine mixing a swimming-pool-sized tank. You cannot simply use a spoon one hundred times larger and assume the mixing pattern is identical. Distance, shape, flow and energy matter.

Similarly, a small filter may receive water evenly across its surface while a large filter develops channels. A small tank may mix almost completely while a large tank leaves slow zones. A short pilot may use new equipment, while a full-scale system must survive weeks and months of fouling, cleaning and maintenance.

For Primary Science, the transferable lesson is not advanced engineering mathematics. It is this: when size changes, check whether the important conditions and relationships stayed equivalent.

Worked Case 1: Same Percentage, Different Flow Pattern

Pilot Unit A is small enough that water passes evenly through the filter. Full-Scale Unit B has a much wider filter bed. Some water follows easier paths and moves through faster than other water.

Even if the same filter material is used, the actual contact between water and filter may differ. The pilot’s 90% removal is still evidence that the material and process can work. But the full-scale design now needs evidence that flow distribution remains suitable.

Worked Case 2: The Pilot Saw One Kind of Input

A pilot is tested during dry weather when incoming water is fairly stable. The full plant later receives much muddier water after storms. A brochure repeats the 90% pilot result as if input conditions never change.

The issue is not scale alone. The feed condition changed too. A strong scale-up study should test or otherwise account for the range of conditions the full system is expected to face.

Worked Case 3: Five Excellent Days

The pilot runs for five days using freshly cleaned equipment. The full plant must run for months. Over time, surfaces become coated and filters begin to clog.

Can five good pilot days prove six months of stable performance? No. They can show short-term performance. A longer-duration or repeated test is needed for long-term reliability claims.

Time is another dimension of scale.

Worked Case 4: Same Dose per Litre, Different Mixing

A chemical treatment is added at the same concentration in pilot and full scale. A student concludes that performance must therefore be identical.

But “same dose per litre” controls only one variable. The full-scale result can also depend on how quickly the chemical mixes, how long it contacts the water, temperature, particle size and the incoming material. A fair comparison needs the variables that actually influence the process, not merely the easiest one to match.

Worked Case 5: The Full-Scale System Performs Better

Scale-up is not automatically bad news. Suppose the full system has better sensors, automatic flow control and more stable operating conditions than the pilot. Its performance may equal or exceed the pilot.

This is important because scepticism must not become one-directional. The correct statement is not “full scale will always be worse.” It is “full scale is a new condition that must be checked.”

Worked Case 6: A Demonstration Is Not Yet a Scale-Up Test

A viral video shows a small transparent column turning coloured water clear. It is visually impressive. The presenter says, “Imagine this multiplied to clean a whole lake.”

The demonstration shows that something happened in that column. It does not by itself establish throughput, energy use, maintenance, material life, waste handling, performance under changing water conditions or lake-scale feasibility. The communication has moved from phenomenon to system claim without the intermediate pilot evidence.

The Scale-Up Bridge

A useful way to read a pilot report is to look for a bridge between the small system and the large one. The bridge is made of preserved relationships.

Was the same basic process sequence used?
Were flow and contact-time relationships representative?
Were vessel shapes and mixing behaviour scaled appropriately?
Was the pilot challenged with realistic input conditions?
Were important operating limits explored?
Were failures, shutdowns and maintenance needs recorded?
Were measurements collected consistently at inlet and outlet?

The more complete that bridge is, the more confidently the pilot result can inform full-scale design. But it remains evidence, not a magical guarantee.

Representation Check: What Does “90%” Refer To?

A single percentage can hide several scientific choices.

90% of what measured quantity?
Measured at which inlet and outlet locations?
Average over how many runs?
Was it 90% at each run, or an overall average?
Was the incoming concentration stable?
Was the output below a detection or reporting limit?
Did the percentage refer to mass, concentration, count or another quantity?

Before debating scale-up, make sure the original pilot claim itself is understood correctly.

Baseline and Comparison Check

A pilot result can look impressive because the percentage is large. But what is the comparison?

If untreated water falls naturally from 50 units to 20 units during the test while the pilot outlet is 5 units, the process probably contributed additional removal, but the fair comparison needs an untreated control or another appropriate baseline. If the pilot receives different input water from the comparison system, the headline percentage alone may not isolate the process effect.

Method and Variable Check

Use the PSLE investigation habit: identify what changed, what should remain comparable and what was measured.

Potentially important condition	Why it matters to scale-up evidence
Flow rate and flow pattern	They affect how long material spends in each part of the process.
Mixing	A larger vessel can develop uneven regions.
Input composition	A process tested on one type of water or material may respond differently to another.
Temperature	Some physical, chemical and biological process rates change with temperature.
Loading	Higher mass or particle loads can overwhelm equipment.
Duration	Fouling, wear and accumulation may appear only after longer operation.
Maintenance	A process may depend on cleaning, replacement or calibration intervals.

Alternative Explanations for a Successful Pilot

“The process works” may be a reasonable explanation. But a careful learner also asks whether the pilot result was helped by conditions that may not survive scale-up:

operators gave the pilot unusually close attention;
the input was selected to be easy to treat;
equipment was new and freshly cleaned;
the trial was too short for deterioration to appear;
the pilot used a different geometry or flow pattern;
only the best run was highlighted;
the result came from a narrow operating range.

Those possibilities are not accusations. They are hypotheses that can be tested by better design and reporting.

What Evidence Would Strengthen the Full-Scale Claim?

The pilot was designed to represent the full-scale process rather than merely resemble it.
Important operating relationships were documented and preserved.
The pilot was challenged across realistic ranges of input and operating conditions.
Testing lasted long enough to observe fouling, wear or other time effects relevant to the claim.
Multiple runs and independent measurements showed consistent performance.
Unsuccessful or lower-performing conditions were reported, not hidden.
A staged scale-up or demonstration at larger scale reproduced the performance range.
The final full-scale system was commissioned and then independently monitored.

What Would Weaken It?

Only one short pilot run is shown.
The pilot and planned full system have major differences in flow, mixing, geometry or operating conditions.
The pilot used unusually clean or easy input material.
The full-scale claim extends far beyond the tested loading or temperature range.
The headline says “always” or “guaranteed” while the report shows variation.
Only a successful average is shown, with no individual runs.
The pilot’s maintenance burden is omitted from the full-scale promise.
No post-scale-up verification is planned.

How Far Can the Conclusion Travel?

A pilot result belongs first to the tested process, configuration, operating range, input material and duration. With strong scale-up design, it can support predictions about a full-scale system. But every widened claim needs a bridge.

Do not automatically travel from:

one pilot size to every larger size;
one feed condition to every feed condition;
five days to a whole year;
fresh equipment to aged equipment;
one operator team to all operating contexts;
one measured outcome to every environmental benefit.

The route can be travelled. It just needs evidence at each important transfer.

Tempting but Invalid Reasoning

“The pilot removed 90%, so full scale must remove 90%.” Pilot evidence supports a prediction, not a guarantee.
“The full plant is 100 times larger, so multiply every dimension by 100.” Scale depends on geometry and process relationships, not one multiplier.
“The pilot was small, so its evidence is worthless.” A well-designed pilot can be powerful evidence precisely because it tests scale-up conditions.
“Different full-scale performance proves the pilot was dishonest.” Scale-up differences can arise even when both studies are careful; investigate the mechanism.
“One successful demonstration proves commercial operation.” Demonstration and sustained operation answer different questions.
“If the average was 90%, every run was 90%.” Always inspect variation.

PSLE-Style Transfer Case: The School Rainwater Cleaner

Original classroom case: Students build a pilot rainwater-cleaning column. It treats 2 litres per minute and reduces a measured cloudiness value by an average of 85% across six trials. The school considers a much larger system for a storage tank.

A proposal says: “Because the pilot achieved 85%, the school tank system will definitely achieve 85%.”

A strong response would say:

The pilot shows the treatment arrangement can work under the tested conditions.
The larger system must preserve important conditions such as flow through the filter, contact time and filter loading.
Rainwater can vary between storms, so the input range should be tested.
Longer operation may reveal clogging that the short pilot did not show.
The full-scale result should be measured after installation rather than assumed from the pilot.

Transfer Case: Same Process, Different Geometry

A small rectangular tank mixes evenly. The large design is much taller and has the same mixer shape placed at one corner. A learner says, “Same mixer means fair comparison.”

Not necessarily. The variable “mixer type” is the same, but the geometry and mixing coverage may not be. A real fair comparison preserves the functional condition—effective mixing—not merely the label on one component.

Explained Practice

Practice 1 — What Was Actually Proven?

A pilot system removed 92% during three controlled trials. What is the strongest safe statement?

Answer: The pilot achieved about 92% removal during those controlled trials. Full-scale performance still needs scale-up evidence.

Practice 2 — New Condition

The full system will operate at twice the temperature range tested in the pilot. Why does that matter?

Answer: The claim is travelling beyond a tested condition. Temperature may affect process performance, so the wider range should be evaluated.

Practice 3 — Time Scale

A two-day pilot is used to claim one year of maintenance-free operation. What evidence is missing?

Answer: Longer-duration evidence about fouling, wear, cleaning and performance over time.

Practice 4 — Percentage

Five pilot runs remove 98%, 95%, 92%, 80% and 75%. Is “90% removal” enough information?

Answer: The average may be around that value, but the variation matters. The full report should show the range and conditions associated with lower runs.

Practice 5 — Correct Scientific Attitude

Which is better: “Pilot tests cannot be trusted” or “Pilot tests are valuable when we know what they represent and verify scale-up assumptions”?

Answer: The second. Healthy scepticism evaluates evidence; it does not reject evidence automatically.

Delayed Independent Return

After several days, give the learner three cards labelled Lab, Pilot and Full Scale. Add evidence statements such as “easy to control variables,” “tests realistic flow arrangement,” “faces everyday operational variation,” “uses small quantity,” and “must be monitored after commissioning.” Ask the learner to place each statement and explain what evidence transfer is still needed between cards.

The delayed task checks whether the learner has learned the architecture of scale-up rather than memorised the phrase “pilot is smaller.”

Parent and Tutor Teaching Guide

Build the lesson with familiar scale changes. A paper bridge that holds one eraser does not prove a geometrically larger copy will hold one hundred erasers. A recipe doubled from two servings to four is easier to scale than a recipe moved from a home bowl to a factory mixing tank. Ask the child what relationships must be preserved.

A good teaching sequence is:

Claim: What is the pilot result?
Bridge: Which conditions make the pilot representative of full scale?
Stress: Which full-scale conditions were not yet tested?
Verification: What should be measured after scale-up?

Avoid teaching “small test bad, large test good.” Scientists often use small and pilot tests because they allow careful learning before committing resources. The lesson is about matching certainty to the evidence stage.

Routes to Existing Canonical PSLE Science Owners

Authoritative Sources

Quiet Return

A pilot result is not a promise printed by a smaller machine. It is a carefully bounded piece of evidence.

Respect it enough to ask what it really demonstrated. Respect the full-scale question enough not to pretend it has already been answered. Then build the bridge: preserve the important relationships, test the wider conditions, measure the scaled system and update the conclusion.

Scale-up is not “make it bigger.” Scale-up is “prove the evidence still travels.”

Small Group Tutorials