A finer apparent cut is not automatically the best economic result. Capturing more fine solids can increase cake mass, entrained liquid and valuable barite loss. The correct operating point balances product quality, liquid recovery, weighting-material recovery, solids rejection and machine stability.
Cut point describes a probability, not a knife edge
For a defined particle population and a properly mass-weighted two-product partition curve, grade efficiency gives the fraction of each size class reporting to the nominated solids stream. The d50 is the size at which 50% of that defined size class reports to the solids stream and 50% to the liquid stream under steady tested conditions. Density, shape and agglomeration broaden the split.
Overall removal efficiency can hide the wrong separation
A centrifuge may show high total-solids recovery because it captures dense barite or coarse solids while allowing problematic colloidal LGS to remain. Conversely, a fine separation may remove LGS but discard expensive barite and liquid. Total percentage alone cannot distinguish these outcomes.
Steady-state solids recovery, Rₛ = ṁₛ,cake / ṁₛ,feed
Steady-state product-liquid recovery, Rₗ = ṁₗ,centrate / ṁₗ,feed
For weighted mud, separately estimate HGS, LGS, oil/water or brine and total liquid. Close the balance before interpreting efficiency.
Why liquid follows the cake
Cake contains interstitial liquid between particles, surface-wetting liquid and sometimes a less-drained bed created by high differential speed or shallow dry-beach exposure. Finer-particle beds often retain more liquid per unit solids mass because of greater surface area and smaller pores, although packing, permeability, chemistry and conditioning can change that direction. Therefore a setting that captures more fines may produce more wet cake even when the centrifuge is functioning correctly.
Worked economic logic
Suppose Setting A removes 800 kg/h solids with 1.2 m³/h valuable liquid in cake. Setting B removes 950 kg/h but loses 2.0 m³/h liquid and more barite. The additional 150 kg/h removal must be valued against 0.8 m³/h extra liquid loss, barite replacement, waste transport and downstream benefit. Without that calculation, “B removes more” is incomplete.
| Metric | Technical question | Commercial question |
|---|---|---|
| d50/grade curve | Which sizes are splitting? | Are they undesirable or valuable? |
| Solids recovery | How much solids reaches cake? | What disposal/processing cost follows? |
| Liquid recovery | How much usable liquid is retained? | What replacement value is saved? |
| HGS/LGS split | Which density fractions move? | What barite is lost or recovered? |
Finding the correct operating point
- Define unacceptable solids and valuable material.
- Measure matched flow and composition for feed and outlets.
- Build grade efficiency where PSD data permit.
- Calculate solids, liquid and HGS/LGS recovery.
- Attach replacement, treatment and disposal cost.
- Select the point that delivers the best whole-system outcome, not the finest claimed d50.
Constructing a grade-efficiency curve
Collect representative, time-aligned feed and outlet samples and determine stream flows or mass rates. Obtain PSD for feed and outlets using a method appropriate to the fluid and particle range. For each size bin, calculate the fraction reporting to the solids stream after correcting for stream mass. A curve based only on normalized PSD percentages without stream flows can be misleading because it ignores how much material is in each outlet.
Partition errors and closure
Sampling error, non-steady operation, agglomerate breakup, density differences and analytical limits can produce impossible partition values above 100% or poor mass-balance closure. Do not force a smooth d50 through bad data. Report closure, repeatability and the size range where the method is reliable.
Use a multi-objective operating window
Plot solids rejection against liquid recovery and, for weighted mud, barite recovery. Add torque margin and throughput. The preferred point is often a knee in the curve where further fine-solids capture demands a disproportionately large liquid or barite loss. That knee is a measured business and engineering decision, not a universal cut point.
Definitions and system boundary
The reported d50 is conditional on the tested material, density and shape distribution, sample preparation or dispersion, operating condition and a properly mass-weighted two-product partition curve. Mixed-density drilling mud does not have one intrinsic size-only d50.
Rs = ṁs,cake / ṁs,feed
Rl = ṁl,centrate / ṁl,feed
These require steady state, all streams captured and negligible accumulation. Check feed = centrate + cake ± accumulation and sampling error. Here recovered liquid means the nominated centrate/product stream, with dilution and recycle inside the stated boundary.
Common questions
What is a centrifuge d50 cut point?
For a defined test population and mass-weighted two-product partition curve, it is the size class for which 50% reports to the nominated solids stream and 50% to the liquid stream under steady tested conditions.
Why can a finer cut reduce liquid recovery?
Capturing more fine solids can send more associated interstitial and surface-wetting liquid to cake; the actual result also depends on packing, permeability, chemistry, conditioning and transport.
What should be optimized besides cut point?
Solids and liquid recovery, HGS/LGS split, centrate quality, cake liquid, torque, stability and total operating cost.
