Even the most careful filtration test produces unreliable results if the slurry sample does not truly represent the plant. Sampling error is especially serious in mineral processing: coarse particles settle in seconds, fines stay suspended, and process chemistry can shift during transport. This guide covers every step, from selecting the sampling point and building a time-based composite to preserving chemistry, documenting process data, and verifying sample condition on arrival, so your filter sizing rests on credible data.

1. Why representative sampling matters

Filter performance depends on far more than the mineral name. Two iron concentrates or two lithium tailings streams can behave very differently due to variations in:

• Particle-size distribution: the single most influential parameter for cake formation and moisture

• Mineralogy and particle shape: affect packing, permeability, and cake release

• Feed-solids concentration: directly impacts cycle time and throughput

• Clay and ultrafine content: even a few percent can blind filter media

• Slurry temperature: changes liquid viscosity and dewatering rate

• pH and dissolved salts: alter particle surface charge and agglomeration

• Flotation reagents, flocculants, oils, or other additives: can either help or hurt filtration

• Degree of oxidation or ageing: mineral surfaces change with time and exposure

The risk in both directions: A sample that loses its coarse fraction may appear harder to filter than the real plant stream, leading to oversized equipment. A sample that misses intermittent slugs of fines may appear easier, leading to an undersized filter that cannot meet production targets.

2. Define the purpose before collecting material

Tell the testing laboratory what decision the test must support. Common purposes include:

Purpose

Typical material requirement

The laboratory can then specify the sample mass or volume, number of containers, preservation method, and process data needed. Comparative or pilot work always needs more material than a basic leaf test.

3. Choose the correct sampling point

Collect from a flowing process stream that represents the proposed filter feed. Avoid:

• Quiet corners of tanks

• Surface-only scoops

• Drain residue

• Settled material from container bottoms

unless those locations are the deliberate subject of investigation.

The sampling point should capture either the full stream or a representative cross-section. For important design programs, automatic or correctly designed process samplers are strongly preferred. They take increments consistently across time and flow.

No existing plant? Work with the process engineer to select a comparable pilot stream, locked-cycle test product, or prepared sample that reflects the expected commercial material.

4. Use multiple increments instead of one grab

A single grab captures only one moment. Plant feed changes with ore blend, grinding conditions, cyclone performance, thickener operation, and reagent dosage. A time-based composite made from multiple increments is far more representative of average operation.

Before sampling, define all six parameters:

1. Production period to represent

2. Increment frequency

3. Increment size

4. Total sample quantity

5. Operating conditions that would invalidate the sample

6. How increments will be combined and mixed

Record any process disturbance during collection. A composite taken through a thickener upset must not be labeled as normal steady-state feed.

5. Sample the difficult conditions too

Designing only for the average leaves a filter unable to handle predictable feed variations. If the mine plan or plant history shows meaningful swings, collect separate samples for:

• Average ore: the baseline case

• Finest expected feed: worst-case blinding risk

• Highest-clay or highest-mud condition: maximum moisture challenge

• Lowest feed-solids concentration: highest volumetric load

• Different ore zones or blends: if geology varies across the deposit

• Startup or seasonal conditions: if relevant to the project

Keep these samples separate and clearly labeled. Do not combine unlike materials unless the intended plant feed is genuinely that blend.

6. Prevent segregation during collection

Slurry begins to segregate the moment agitation stops. Use containers and transfer methods that minimize settling and material loss. Before dividing a bulk sample, mix it thoroughly using a method that does not grind particles, introduce foreign material, or alter chemistry.

Critical rules:

• Never pour only the liquid from the top of one container and assume the rest has the same solids distribution.

• If the sample must be split, use appropriate slurry-splitting equipment or repeated increments from a well-mixed bulk.

• Inspect transfer hoses, buckets, and funnels afterward, coarse solids or sticky fines left behind are part of the sample, and their loss changes the test material.

7. Preserve the original process chemistry

Filtration behavior can change between the plant and the laboratory. The following degradation mechanisms are common:

Mechanism

Effect on filtration

Ask the laboratory about:

• Maximum acceptable time before testing

• Temperature requirements during transport

• Whether airtight containers are needed

• Whether headspace should be minimized

• Permitted preservatives (if any)

• Hazardous-material classification and shipping rules

Do not add water, dispersant, flocculant, acid, alkali, or preservative unless the testing plan specifies it. Even seemingly harmless dilution changes the feed condition.

8. Select and fill containers correctly

Containers must be clean, leak-resistant, chemically compatible, and strong enough for transport. Leave only the headspace needed for safe mixing and thermal expansion. Several manageable containers are often better than one oversized drum, large containers can be nearly impossible to remix after settling.

Each container label and shipping document must include:

• Project and plant name

• Stream name and sampling point

• Sample identification number

• Date and time of collection

• Container number (e.g., "1 of 4")

• Sampler's name or team

• Known hazards and handling precautions

Photograph the containers and labels before shipment. This simple step has saved countless investigations when questions arise later.

9. Send process data with the sample

The physical sample is only half the test package. Provide a sampling report containing:

Data category

Specific information

Laboratory analysis can confirm some properties, but it cannot reconstruct missing plant context.

10. Agree on the filtration test conditions

Before testing begins, define the variables to be measured and reported. For a vacuum filtration program, the standard set includes:

• Absolute vacuum pressure (not just gauge)

• Cake-forming time

• Drying time

• Cake thickness

• Filtrate volume and clarity

• Wet and dry cake mass

• Cake moisture test method

• Filter-medium specification

• Cake release behavior

• Specific dry-solids throughput (kg/m虏/h)

High-altitude installations: Discuss test pressure and scale-up basis explicitly. Reporting only a relative gauge vacuum can be insufficient when the atmospheric pressure at site differs significantly from laboratory conditions.

11. Check sample condition on arrival

The laboratory should inspect and record:

• Leakage or damaged seals

• Unusual odor or gas generation

• Temperature on arrival

• Evidence of excessive ageing

• Hard settling that resists remixing

• Visible differences in solids concentration between containers

If the sample is no longer credible, collecting a new one is almost always cheaper than designing equipment from misleading data.

After homogenization, compare measured solids concentration and particle size with the plant records. Significant disagreement must be investigated before the sizing result is accepted.

12. Questions to ask when reviewing the test report

1. Was the entire submitted sample homogenized properly?

2. Do measured feed properties match the plant sampling data?

3. What pressure, medium, area, and cycle times were used?

4. How were cake moisture and dry-solids throughput calculated?

5. Were repeat tests consistent? (If not, why?)

6. Which specific result is being used for scale-up?

7. What design margin is recommended for variability?

8. Are additional tests needed for the difficult ore cases?

Conclusion: Good equipment sizing starts with a credible sample

Representative sampling reduces uncertainty across every dimension of filter design: filter area, cake moisture, filtrate quality, consumables selection, and auxiliary-equipment sizing. It also gives manufacturer and customer a common, defensible technical basis for performance discussions and guarantees.

Tongzhiren Filtration evaluates mineral concentrate, tailings, and industrial slurry duties for vacuum disc filter applications. Contact our engineering team before collecting material. We will confirm the sample quantity, test objectives, and process information required for your specific project.

Contact us

Technical references

• Metso, Slurry Sampling Solutions (PDF)

• Metso Research & Testing, Dewatering Laboratory and Representative-Sample Considerations

• ANDRITZ, Leaf Testing and Filtration Process Simulation (PDF)