A hydrocyclone can show the expected spray pattern and still move the wrong amount of solids or drilling fluid. A performance test needs the feed, overflow and underflow measured over the same stable operating interval. Otherwise, the apparent separation result is a guess assembled from unrelated readings.
Scope. This is a three-stream, surface-process acceptance method for desanders, desilters and hydrocyclone banks. It is not a universal operating-pressure specification, an API RP 13C screen test, or a substitute for the installed cone manufacturer's instructions. In weighted mud, rejecting the underflow may also reject valuable weighting material; the objective must be defined before interpreting any result.
First, state what the cyclone is supposed to achieve
For an unweighted system, the objective is usually reducing unwanted drilled solids while controlling the amount of useful liquid discarded with the underflow. In a weighted system, the same discharge can contain recoverable barite. A hydrocyclone or mud cleaner must be judged within the actual fluid program, including whether the underflow is discarded, screened and recovered, or routed elsewhere.
Record cone make, diameter, installed count, number operating, apex/vortex condition, pressure measurement point, feed-pump duty, mud formulation, density, temperature, solids concentration, and any intentional bypass. One pressure reading cannot establish both separation quality and flow balance.
Define a defensible test boundary
Use one steady operating window and capture all three streams. For a bank, include every operating cone and any bypass or separate discharge. Do not use a pumped-feed estimate from a nameplate rating or assume equal loading among cones on a manifold. If the feed, overflow or underflow is recirculated into a common tank during sampling, isolate the test boundary conceptually and document inventory effects.
| Measurement | Feed (F) | Overflow (O) | Underflow (U) |
|---|---|---|---|
| Mass flow, kg/h | Measured or Q × density | Measured or Q × density | Measured or Q × density |
| Dry-solids mass fraction, kg/kg | Representative sample | Representative sample | Representative sample |
| PSD | Sample method recorded | Same method/basis | Same method/basis |
| Operating state | Feed pressure and conditions | Flow path confirmed | Apex condition, discharge stability |
Sampling rule. Samples should represent the same interval and steady duty, with repeated or time-composited grabs where appropriate. A brief underflow cup sample collected during a pressure swing must not be compared with an earlier feed-tank sample. Record sample IDs, density method, solids method, units and collection times. Use suitable containment and the approved rig sampling procedure; do not sample from hazardous unguarded discharge points.
The three checks that matter
- Total mass closure: MF ≈ MO + MU, using measured kg/h, with any known inventory change or separate streams accounted for.
- Dry-solids closure: MFxF ≈ MOxO + MUxU, where x is measured dry-solids mass fraction.
- Liquid split: liquid to underflow ÷ feed liquid. Report this alongside solids to underflow; solids removal is not free if large amounts of base liquid are thrown away.
Use signed residuals and show the denominator: closure residual (%) = 100 × [feed in − outlets out] ÷ feed in. A result of zero is not a guarantee of perfect measurements: independent errors can cancel. There is no single defensible acceptance tolerance for every field test; set the tolerance from instrument uncertainty, sample variability and stability of the chosen test boundary.
Watch the calculated stream. Overflow is often hard to measure directly and gets calculated as feed minus underflow. If any stream is calculated by difference, total mass closure is zero by construction and stops being a check. The dry-solids closure is then the only independent test, so the report must say which stream was calculated and how.
Worked example: dry solids removed versus liquid lost
Illustrative calculation only. The figures are constructed for arithmetic demonstration, not taken from a rig test and not proposed as expected cyclone performance. Mass flow is used directly to avoid assuming that different slurry densities are interchangeable.
| Stream | Total flow (kg/h) | Dry solids (wt%) | Dry solids (kg/h) | Liquid (kg/h) |
|---|---|---|---|---|
| Feed | 50,000 | 12.0 | 6,000 | 44,000 |
| Overflow | 42,000 | 3.81 (1,600/42,000) | 1,600 | 40,400 |
| Underflow | 8,000 | 55.0 | 4,400 | 3,600 |
Mass closure: 50,000 − 42,000 − 8,000 = 0 kg/h. Dry-solids closure: 6,000 − 1,600 − 4,400 = 0 kg/h. Feed solids reporting to underflow: 4,400 / 6,000 = 73.3%. Liquid reporting to underflow: 3,600 / 44,000 = 8.2%. Underflow liquid/solids ratio: 3,600 / 4,400 = 0.82 kg/kg.
The 73.3% result means total measured feed dry-solids mass reporting to the underflow during this test. It is not a universal removal efficiency, a size-specific classification efficiency, or evidence that 73.3% of unwanted low-gravity solids were removed. The sample may contain weighting material and different minerals. Without particle-size and density-class information, that distinction cannot be resolved.
When the cut point is the real question
For each consistent particle-size band i, the measured partition to underflow is PU,i = (MU,solids × fU,i) / (MF,solids × fF,i). Here f is the measured mass fraction of total dry solids in that size band, determined with a method suitable for the mud and the relevant particle type. Validate the size distributions and reconciled solids balance before fitting a curve or quoting d50.
Fine particles can accompany liquid into underflow, so an uncorrected partition curve differs from a model-based bypass-corrected curve. Correction assumptions are not uniquely measurable, and different models can give different corrected cut points. State whether you report an observed partition or a model-corrected metric. Do not silently mix the two. University of Queensland analysis describes water and fine-solids entrainment in classification; Nageswararao (2016) examines the limitations of bypass correction. These are mineral-processing methods requiring adaptation to drilling fluids.
Diagnose the cause before touching a setting
| Observed evidence | Possible explanation | Next verification |
|---|---|---|
| High liquid loss in underflow | Apex setting/wear, high split or feed-condition change | Repeat liquid split and check OEM geometry and stable feed duty |
| Unexpected coarse solids in overflow | Plugged or partly blocked apex, overload, feed-head variation, worn internal parts or wrong classification duty | Check flow/pressure trend, cone condition and size-band partition |
| One cone roping while others spray | Underflow capacity exceeded on that cone: apex too small or partly plugged, local solids overload or low feed head from unequal manifold distribution. (A worn, enlarged apex normally widens the spray and raises liquid loss rather than causing a rope.) | Inspect manifold distribution, apex size and condition of the individual cone |
| Solids balance will not close | Non-simultaneous samples, density conversion error or unaccounted stream | Re-sample before making a performance claim |
A spray pattern is a diagnostic clue, not a calibrated solids-removal test. A single pressure target is not transferable across cone designs, mud densities and operating limits; use the fitted OEM pressure-flow operating window and the measured duty. For a step-by-step roping diagnosis, use the Hydrocyclone Roping Decision Tree and the roping troubleshooting guide.
Acceptance record: what to hand over
The completed report should include a one-page boundary diagram; installed/operating cone count; pressure and flow trend; feed, overflow and underflow sampling log; lab method and dry-solids basis; signed total/solids closure; liquid split; PSD comparison if obtained; observed defects; and a clear pass, conditional retest or fail against pre-agreed test objectives. If sampling uncertainty is larger than the claimed improvement, report that limit, not a false percentage gain.
This is one equipment-level test inside a wider system. For generation-to-disposal accounting and working field sheets, see the free SC DrillTech field engineering report From Formation to Waste. For related field checks, read Hydrocyclone Feed Head and Solids-Control Mass Balance.
Common questions
How do you measure hydrocyclone efficiency in the field?
Sample feed, overflow and underflow over the same stable interval, measure mass flow and dry-solids fraction for each stream, and close both the total and the dry-solids balance. Then report solids recovered to underflow together with the liquid lost to underflow. One number on its own hides the trade-off.
Is a spray discharge proof that a cone is working properly?
No. A spray tells you the underflow is not overloaded at that moment. It says nothing about how much solids or liquid each stream is carrying. Only a closed three-stream balance answers that.
What does roping mean on a desander or desilter?
The underflow capacity of that cone has been exceeded, usually from high solids loading, an apex that is too small or partly plugged, or low feed head. A roping cone sends solids it should reject into the overflow. A worn, enlarged apex does the opposite: wider spray and more liquid loss.
Technical basis and boundaries
Mass conservation and grade-partition calculations are general engineering principles. The specific operating envelope, sampling access and acceptance criteria must come from the site procedure and fitted OEM. Further background: University of Queensland (classification and water split); Nageswararao, K. (2016), “Modelling of hydrocyclone classifiers: a critique of ‘bypass’ and corrected efficiency”, Powder Technology 297. These sources do not specify drilling-mud field acceptance thresholds.
