A Hydraulic Sand Ejector is a jet pump that uses a high-pressure motive water stream passing through a venturi nozzle to entrain and lift sand-water slurry out of a sump, pit or borehole. The throat — a hardened, replaceable convergent-divergent section — is the single most important component, because its diameter sets the suction draw and the slurry concentration the unit can move. The ejector exists to clear sand where conventional centrifugal pumps clog or wear out fast. On caisson sinking, dewatering wells and tunnel invert cleaning, a 100 mm ejector run at 7 bar motive pressure routinely lifts 30–50 m³/h of slurry with no moving parts in the slurry path.
Hydraulic Sand Ejector Interactive Calculator
Vary ejector diameter and motive pressure to estimate Bernoulli jet speed and slurry capacity range.
Equation Used
The calculator uses Bernoulli pressure-to-velocity conversion for the clean motive jet, then scales the article capacity range by ejector area and the square root of motive pressure. It is an estimating tool for the stated hydraulic sand ejector example, not a substitute for manufacturer performance curves.
- Capacity range is anchored to the article example: 100 mm at 7 bar gives 30-50 m3/h.
- Clean motive water density is taken as 1000 kg/m3 for Bernoulli jet speed.
- Flow scaling uses area times square-root pressure; actual output depends on nozzle spacing, lift, solids and back-pressure.
Operating Principle of the Hydraulic Sand Ejector
The mechanism is plain Bernoulli at work. High-pressure motive water — usually 5 to 10 bar from a surface duty pump — accelerates through a converging nozzle, dropping its static pressure as velocity climbs past 25 m/s. That low-pressure region sits inside the suction chamber, which is open to the sand-laden water at the bottom of the pit. Atmospheric pressure plus the static head above the sand pushes the slurry into that low-pressure zone, where it mixes with the motive jet and accelerates together through the throat. The combined stream then decelerates in the diffuser, converting velocity head back to pressure head so it can climb the discharge line to surface.
Why build it this way? Because there are no impellers, no seals, no shafts down the hole. Sand is the most aggressive abrasive a pump will ever see — silica hardness sits at Mohs 7, harder than most pump bronze. A centrifugal pump impeller passing 30% sand slurry at 30 m/s wears through in days. The ejector takes the wear on a single replaceable hardened throat liner, typically tungsten carbide or alumina ceramic, and that's the only abrasion-exposed part you ever change. The motive pump up top runs clean water and lasts indefinitely.
The failure modes are tight and predictable. If the throat-to-nozzle spacing drifts more than ±0.5 mm out of factory setting, suction collapses — you'll see motive pressure rise on the gauge while discharge flow drops, classic dry-running symptom. If motive pressure falls below about 4 bar, the velocity at the throat won't generate enough vacuum to lift slurry against the static suction head, and the unit just recirculates its own motive water. Cavitation in eductors happens when motive pressure climbs too high relative to back-pressure on the discharge — the throat erodes in a pitted ring pattern and you'll hear a gravelly rattle through the discharge hose.
Key Components
- Motive Nozzle: Converging nozzle that accelerates motive water from supply velocity to roughly 25–40 m/s at the exit. Bore tolerance is typically ±0.05 mm on the throat exit diameter — a 0.1 mm wear groove drops jet coherence and suction draw by 15%.
- Throat (Mixing Tube): Hardened convergent-divergent section, usually tungsten carbide or alumina, where motive jet and entrained slurry mix and reach uniform velocity. Length-to-diameter ratio sits between 6:1 and 8:1 for sand service. This is the wear part — expect 200–600 hours service life in heavy sand before bore wears 5%.
- Suction Chamber: Open volume around the gap between nozzle exit and throat entry where slurry enters the unit. The nozzle-to-throat spacing is the most critical assembly dimension — factory-set, normally 0.8 to 1.2 nozzle diameters, and it must not drift.
- Diffuser: Diverging cone downstream of the throat that recovers velocity head as static pressure to drive the slurry up the discharge line. Half-angle of 5–7° gives best recovery without flow separation.
- Suction Strainer: Coarse screen, usually 25–40 mm openings, that excludes cobbles and debris too large to pass the throat. Throat will pass particles up to about 50% of throat diameter, so the strainer sets the upper particle size.
- Motive Supply Hose and Pump: Surface-mounted clean-water pump delivering 5–10 bar at the design motive flow. A 100 mm ejector typically needs 30–40 m³/h of motive water at 7 bar from something like a Grundfos CR or a diesel-driven Godwin set.
Where the Hydraulic Sand Ejector Is Used
Sand ejectors show up wherever you need to lift abrasive slurry from a place a real pump can't survive or can't fit. The common thread is depth, abrasion, or both. They handle sand concentrations up to about 30% by volume, particle sizes up to roughly 50% of throat diameter, and lift heights up to 40 m on a properly sized motive supply.
- Caisson and Shaft Sinking: Soletanche Bachy and Bauer routinely run 150 mm hydraulic ejectors inside open-bottom pneumatic caissons to remove sand and silt cuttings during sinking, where a centrifugal pump would not survive 30% sand slurry.
- Tunnel Construction: TBM invert cleaning on Crossrail and the Thames Tideway used hydraulic eductors to clear muck-pit fines from sumps behind the cutterhead trailing gear, lifting 20–40 m to surface skips.
- Water Well Development: Boart Longyear and Atlas Copco well-development rigs use small 50–80 mm ejectors lowered down newly drilled water wells to remove drilling sand and develop the screen — the airlift alternative needs more lift head than is often available.
- Marine Dredging: DOP submersible dredge pumps from Damen and the simpler Toyo-style hydraulic eductors handle harbour and marina silt removal where a vessel-mounted centrifugal dredge can't get into shallow corners.
- Power Station Intake Cleaning: Coastal power stations like Sizewell B use hydraulic ejectors to clear sand accumulation from cooling-water intake forebays and trash-rack pits during shutdowns.
- Mine Dewatering: Sump cleaning at quarry and open-pit mining operations — Anglo American and Rio Tinto sites run trailer-mounted ejector packages to clear sand-laden sumps that would chew through a submersible Flygt within a shift.
The Formula Behind the Hydraulic Sand Ejector
The single number that decides whether the ejector works is the entrainment ratio — the mass of suction slurry lifted per unit mass of motive water. Below the low end of the typical range, around 0.3, the unit moves so little slurry the operation is economically pointless. At the nominal sweet spot of 0.7 to 1.0 you get the rated lift and concentration the manufacturer publishes. Push beyond 1.5 and the throat starves — the motive jet can no longer accelerate the entrained mass, suction collapses, and you're back to recirculating motive water. The formula below gives you suction flow as a function of motive flow, area ratio and head ratio so you can sit the design squarely in the sweet spot before ordering hardware.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Qs | Suction (slurry) volumetric flow drawn into the ejector | m³/h | GPM |
| Qm | Motive water volumetric flow supplied to the nozzle | m³/h | GPM |
| At | Throat cross-sectional area | mm² | in² |
| An | Motive nozzle exit cross-sectional area | mm² | in² |
| Hm | Motive supply head at the nozzle inlet | m | ft |
| Hd | Discharge head at the diffuser exit | m | ft |
| Hs | Suction head at the strainer (negative if lifting) | m | ft |
Worked Example: Hydraulic Sand Ejector in a deep diaphragm-wall trench desanding pass on a Hong Kong MTR station box
You are sizing a 100 mm hydraulic sand ejector to clean fine sand and bentonite cuttings out of a 38 m deep diaphragm-wall panel trench on an MTR station box excavation in Kowloon. The motive pump up top is a skid-mounted multistage delivering 35 m³/h of clean water at 70 m head. The throat-to-nozzle area ratio on the chosen unit is 4.0. Discharge head into the desanding plant on the slab above is 8 m, and suction head at the trench bottom is −30 m (i.e. you are lifting 30 m).
Given
- Qm = 35 m³/h
- At / An = 4.0 —
- Hm = 70 m
- Hd = 8 m
- Hs = −30 m
Solution
Step 1 — compute the head ratio at the nominal operating point. Numerator is motive minus discharge, denominator is discharge minus suction:
Step 2 — compute the area term:
Step 3 — combine into the suction flow at nominal motive supply:
That's the design point — 77 m³/h of slurry up 30 m, comfortably inside the manufacturer's published curve for a 100 mm Toyo-style unit at a 4:1 area ratio. The entrainment ratio Qs/Qm is 2.2, which sounds high but is volumetric — on a mass basis with sand-laden slurry at SG 1.3 the mass ratio is closer to 2.9, still above the practical efficiency knee.
Step 4 — at the low end of motive supply, say the surface pump drops to 50 m head because two ejectors share the manifold:
You lose 17% of your slurry throughput for a 29% drop in motive head — eductors are not linear, and the head ratio penalty hits hard. At the high end, push motive head to 90 m:
In practice you don't see 89 m³/h for long — the motive nozzle starts cavitating once jet velocity exceeds about 45 m/s, and the throat liner pits within 50–100 hours instead of the normal 400+. The sweet spot is 65–75 m motive head.
Result
Nominal suction flow is 77 m³/h of slurry lifted 30 m to the desanding plant. That's enough to clean a typical 6 m × 1.2 m panel trench in roughly 90 minutes, which matches the MTR station-box programme. At the low end (50 m motive) you fall to 64 m³/h and the panel-cleaning slot stretches past 2 hours, eating into the concrete-pour window; at the high end (90 m motive) you nominally hit 89 m³/h but throat life collapses, so the practical ceiling is 75 m motive head. If your measured flow is 30% below predicted, the three things to check first are: (1) the motive nozzle bore — a 0.3 mm wear groove from sand ingestion through a damaged supply strainer will halve jet coherence; (2) air ingress at the suction-hose flange, which shows up as a frothy discharge and a noticeable drop in vacuum at the suction-side gauge; and (3) discharge line kink or partial blockage raising Hd — every extra metre of back-pressure costs you roughly 1 m³/h of suction flow on this unit.
Hydraulic Sand Ejector vs Alternatives
Hydraulic sand ejectors compete with submersible slurry pumps and airlift pumps. Each has a window where it wins, and the decision usually comes down to depth, abrasion intensity, and whether you have power or compressed air available at the surface.
| Property | Hydraulic Sand Ejector | Submersible Slurry Pump (Flygt-class) | Airlift Pump |
|---|---|---|---|
| Max sand concentration (% by volume) | 30–35% | 10–15% before rapid impeller wear | 20–25% |
| Practical lift height | 40 m | 60–80 m | Limited to ~60% of submergence depth |
| Wear-part service life in heavy sand | 200–600 h on throat liner | 50–200 h on impeller and volute | Effectively unlimited (no wear parts) |
| Capital cost (100 mm-class unit) | $3,000–$8,000 | $8,000–$25,000 | $1,500–$4,000 plus compressor |
| Surface support required | Clean-water pump, 5–10 bar | Power cable, control gear, lifting frame | Air compressor, 4–8 bar |
| Moving parts in the slurry path | None | Impeller, mechanical seal, bearings | None |
| Best application fit | Deep sumps, caissons, diaphragm walls, abrasive slurry | Construction dewatering with low sand content | Well development, shallow sand removal where compressor is available |
| Hydraulic efficiency | 20–30% | 60–75% | 15–25% |
Frequently Asked Questions About Hydraulic Sand Ejector
That pattern almost always points to fines packing in the diffuser or a partial blockage at the discharge fitting — the motive side reads normal because the nozzle is fine, but back-pressure has crept up and shifted you off the design head ratio. Check the discharge head with a gauge at the diffuser flange; if it has risen 3–5 m above commissioning value, pull the unit and inspect the diffuser cone for a sand wedge.
The other cause is the suction strainer plugging with clay-bound material. Bentonite-bearing trench fluid is the worst offender because it forms a gel layer over the screen openings. A quick reverse flush — close the discharge valve momentarily and let motive pressure backwash the strainer — clears it without pulling the unit.
Area ratio trades suction flow against lift capability. A 6:1 ratio gives you more entrainment (higher Qs/Qm) but only works when motive head is comfortably above 3× the lift head — push it past that and the throat starves. A 4:1 ratio handles deeper lifts more reliably because it can convert more motive head into suction work.
Rule of thumb: if your lift exceeds 25 m, pick 4:1 or 3:1. For shallow desilting at 10 m or less where you want maximum slurry throughput, go 6:1. For the MTR-style 30 m lift in the worked example, 4:1 is the right call — a 6:1 unit on the same motive supply would deliver maybe 40 m³/h instead of 77.
If motive flow and pressure both read nominal, the gap is usually not lost flow — it's mass-balance confusion. The formula gives total volumetric flow leaving the diffuser, which equals motive plus suction. If the weir is calibrated for clean water but the slurry is 25% sand by volume, your weir reading underreads by roughly 10–15% because the head-discharge curve shifts with density.
The other real cause is short-circuiting: if the suction inlet sits above the slurry layer, the unit pulls clean water from above the sand bed instead of slurry. Lower the ejector until the strainer is buried 100–200 mm into the sand and the discharge will jump immediately and visibly turn dark.
Yes, but the head losses do not split evenly and you will not get 2× throughput. Each branch sees roughly half the motive flow but the same friction loss in the shared header, so motive head at each nozzle drops more than you would expect. From the worked example, dropping motive head from 70 m to 50 m cut suction flow from 77 to 64 m³/h — so two ejectors might give you a combined 128 m³/h instead of 154 m³/h.
If you must manifold, oversize the motive pump by 20–25% and use equal-length supply hoses to balance the branches. Unequal hose lengths bias all the flow to the shorter branch and the longer-branch ejector starves.
Two distinct causes produce a similar sound. The benign one is large particles passing the throat — anything above about 30% of throat diameter rings the discharge line as it transits. That's normal and the unit is doing its job, though you should check your strainer for a hole if the noise is new.
The damaging one is throat cavitation, which happens when motive pressure is too high relative to discharge back-pressure — vapor bubbles collapse on the throat wall in a tight ring just downstream of the nozzle. This produces a higher-pitched, more continuous gravelly noise rather than discrete bangs. Pull the unit after a shift and look for a polished pitted band 10–20 mm into the throat. If you see it, drop motive pressure by 10–15% and the cavitation stops.
That gap controls how the motive jet enters the throat. Too tight (under 0.6 nozzle diameters) and the jet hasn't fully expanded — you get a high-velocity core surrounded by dead zones, and the suction chamber doesn't develop full vacuum. Too loose (over 1.5 nozzle diameters) and the jet starts decelerating and dispersing before it engages the throat, so mixing efficiency drops sharply.
In the field, a 1 mm drift on a 25 mm nozzle is enough to drop suction flow 20%. After any maintenance that opens the suction chamber, set the gap with a feeler-style spacer, not by eye — and check it dry before you lower the unit, because there's no easy way to diagnose a misset gap once it's 30 m down a panel trench.
References & Further Reading
- Wikipedia contributors. Injector. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
