High speed and low speed are not quality labels. They are operating modes selected to move different portions of a mixed solids population. The correct mode depends on which solids should report to cake, which should remain in liquid, and what valuable fluid or weighting material must be retained.
Start with the separation objective
In an unweighted low-solids system, the objective may be to reject the finest practical drilled solids while recovering liquid. In a weighted mud, the objective may be to recover coarse/dense barite into the underflow while allowing finer LGS to follow the liquid to a second treatment step, or to remove LGS while accepting a calculated barite loss. Dewatering uses chemistry and separation to produce a reusable liquid and disposable cake. These are different flowsheets.
What “high” and “low” actually mean
The labels must be converted into bowl diameter, actual RPM and RCF. A “high-speed” setting on a large bowl may produce a different acceleration from another machine at the same RPM. Speed also acts with feed rate, viscosity, pond depth and geometry; mode names alone cannot predict cut point.
| Mode objective | Desired split | Primary checks |
|---|---|---|
| LGS removal | Increase capture of the targeted fine-solids population; quantify simultaneous barite and liquid reporting to cake | LGS trend, liquid/barite loss, PSD |
| Barite recovery | Recover HGS, route finer LGS onward | HGS/LGS balance and recovered mud quality |
| Dewatering | Clean reusable liquid and concentrated cake | Chemistry, centrate, cake liquid, polymer |
Density and size act together
Centrifugal settling responds strongly to particle size and to density difference between particle and liquid. Barite is denser than drilled solids, so a smaller barite particle may settle similarly to a larger LGS particle. Their distributions overlap; there is no perfect speed that separates every barite particle from every drilled-solid particle.
This overlap is why retort data alone cannot validate the split. PSD shows how size classes partition; retort/density-based HGS–LGS estimates and, where needed, mineral- or density-specific analysis are required to infer barite versus drilled solids. Neither PSD nor retort alone proves mineral-by-size recovery. In weighted mud, include barite and base-fluid recovery as well as LGS reduction.
A selection workflow
- Define the target streams and acceptable valuable-material loss.
- Measure mud density, retort fractions, rheology and representative PSD where available.
- Confirm actual bowl radius and calculate RCF.
- Begin at a conservative flow and manufacturer-approved speed.
- Collect matched feed, centrate and cake data.
- Calculate HGS/LGS and fluid balance.
- Adjust speed and flow one variable at a time.
Dual-centrifuge arrangements
A first machine may recover a denser/coarser fraction while a second machine treats the first centrate to reject finer LGS. Alternatively, units may run in parallel for capacity. The correct arrangement depends on mass balance; “two centrifuges” does not automatically mean better control if both are fed incorrectly or produce an uncontrolled recycle.
Classification versus clarification
A decanter may be operated as a clarifier, maximizing solids removal from the liquid, or as a classifier, deliberately creating a split between particle populations. Barite recovery is a classification problem: the desired dense fraction and undesired LGS overlap. Operating reports should name the mode and desired destinations rather than label one outlet universally good or bad.
Mode transition control
Changing between modes requires more than changing RPM. Feed source, flow, pond, differential strategy and outlet routing may also change. Material remaining in the bowl and lines contaminates early samples. Document transition time, flush or displacement procedure where applicable, and the point at which samples represent the new mode.
Weighted-mud economics
Calculate the mass and replacement value of barite in each stream, the liquid carried with the recovered/discarded solids, the LGS removed and the volume of conditioned mud returned. A mode that reduces retort LGS slightly while discarding large barite and base-fluid value may not be optimization.
Limits of HGS/LGS reporting
HGS/LGS values are method-dependent estimates, not direct mineral identification. State the API field-test method, assumed densities and corrections, and report closure and uncertainty. Salts, brine, internal phase, chemical solids, drilled-solids density and retort uncertainty can affect the estimate.
Common questions
Is high-speed mode always for LGS removal?
It is commonly used to target finer solids, but the actual split depends on bowl geometry, RCF, flow, viscosity, PSD and density contrast.
Can a centrifuge perfectly separate barite from drilled solids?
No. Their size and settling-response distributions overlap, so the split always involves a recovery/rejection trade-off.
How is the correct mode verified?
By matched feed and outlet sampling, HGS/LGS and liquid mass balance, PSD where available, and economic evaluation of recovered and lost material.
