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.
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:
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.
Tell the testing laboratory what decision the test must support. Common purposes include:
| Purpose | Typical material requirement |
|---|---|
| Sizing a new vacuum disc filter | Large composite |
| Comparing conventional vs. ceramic media | Medium to large, split sample |
| Estimating cake moisture and filtrate clarity | Standard leaf-test quantity |
| Evaluating an ore-body change | Depends on variability range |
| Investigating a capacity bottleneck | Large, time-matched to problem periods |
| Comparing operating conditions or reagents | Multiple sub-samples |
| Confirming performance for an expansion | Large, representative of future feed |
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.
Collect from a flowing process stream that represents the proposed filter feed. Avoid:
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.
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:
Record any process disturbance during collection. A composite taken through a thickener upset must not be labeled as normal steady-state feed.
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:
Keep these samples separate and clearly labeled. Do not combine unlike materials unless the intended plant feed is genuinely that blend.
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:
Filtration behavior can change between the plant and the laboratory. The following degradation mechanisms are common:
| Mechanism | Effect on filtration |
|---|---|
| Cooling | Increases viscosity, slows dewatering |
| Oxidation | Changes mineral surface wettability |
| Bacterial growth | Generates gas, alters chemistry |
| Salt precipitation | Blocks pores, changes permeability |
| Reagent degradation | Flocculants break down over hours to days |
| Agglomeration / dispersion | Alters effective particle size |
Ask the laboratory about:
Do not add water, dispersant, flocculant, acid, alkali, or preservative unless the testing plan specifies it. Even seemingly harmless dilution changes the feed condition.
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:
Photograph the containers and labels before shipment. This simple step has saved countless investigations when questions arise later.
The physical sample is only half the test package. Provide a sampling report containing:
| Data category | Specific information |
|---|---|
| Material identity | Mineral or material description |
| Throughput | Dry-solids tonnage to be treated |
| Feed properties | Solids concentration, slurry density, solids density |
| Particle size | Full PSD, at minimum percent passing key sizes |
| Chemistry | Slurry temperature, pH, reagents and dosages |
| Upstream context | Thickener or conditioning information |
| Performance targets | Target cake moisture, required filtrate quality |
| Operating context | Expected operating hours, availability, site elevation |
| Variability | Known ore variability, sampling and compositing method |
Laboratory analysis can confirm some properties, but it cannot reconstruct missing plant context.
Before testing begins, define the variables to be measured and reported. For a vacuum filtration program, the standard set includes:
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.
The laboratory should inspect and record:
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.
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.
| Info@tzrfiltration.com | |
| +86 139 1246 6955 |