AQUACHAIN GUIDE
Back to illustrated guides

Illustrated guides · Physicochemical treatment

How does a sand filter capture particles and backwash itself?

A rapid granular filter passes coagulated and clarified water downward through sand, anthracite, or mixed media, captures particles by transport and attachment within pore paths, then reverses clean water and optional air through the bed to detach and remove the accumulated solids.

Direct answer

Direct answer

A rapid sand, dual-media, or multimedia filter is not a fixed-pore sieve. Microfloc from upstream treatment follows tortuous paths through the bed and is intercepted, settles, collides with, and attaches to media grains. Some solids load the upper zone while a well-designed bed also uses depth. As deposits constrict pores, head loss and local velocity rise and flow seeks lower-resistance paths; continued operation can produce breakthrough, cracks, short-circuiting, or the terminal allowable head loss. Backwash should start from defined individual-filter criteria such as effluent turbidity or particle count, bed head loss, run time, or production. The filter is isolated, surface wash or air scour is applied when designed, and clean water moves upward through the underdrain to expand the media uniformly, rub grains, detach solids, and carry them to wash-water troughs. The bed then resettles and may be rested, slow-started, and filtered to waste until its own effluent is acceptable. Underwashing leaves mudballs, binding, and dead zones; overwashing can lose media, disturb support gravel, or damage underdrains. Operators therefore need individual-filter flow, head loss, turbidity/particles, bed expansion, distribution, wash-water quality, and recycle evidence rather than a timer alone.

Four links that must work in both filtration and backwash

Clear effluent now or a visibly turbulent wash does not prove that the bed is uniform or that retained solids actually left.

Pretreatment creates filterable microfloc

Unstable fines pass through when coagulation is weak; oversized fragile floc blinds the surface or breaks. pH, chemicals, mixing, clarified-water quality, and flow swings set the incoming burden.

Media and flow use bed depth

Grain size, density, grading, depth, and filtration rate set pore paths, shear, and solids distribution. A thick surface mat builds head loss early; cracks or media loss create shortcuts.

Backwash distribution produces the right expansion

The underdrain and air grid must loosen the whole bed. Water temperature and media properties change the required rate; too little leaves mudballs, while too much can lose media or upset support layers.

Return to service protects the clear-water system

Media must resettle and ripen while valves change gradually. Filter-to-waste, delayed start, or slow start should end on individual-filter turbidity or particle evidence under the plant SOP.

1

A filter bank can keep cells online while one cell backwashes

Headers and valves control each bay; a calm bay represents filtration, the turbulent bay represents wash, and the front overflow removes dirty wash water.

A filter bank can keep cells online while one cell backwashes：Headers and isolation valves、Online filtering cell、Turbulent backwash cell、Wash trough and waste overflow1234

What to identify

  1. 1Headers and isolation valves
  2. 2Online filtering cell
  3. 3Turbulent backwash cell
  4. 4Wash trough and waste overflow

Figure takeaway

Filters operate as a bank. When one cell washes, its flow shifts to the remaining cells; without a plant-flow reduction, they can be bumped into overload and pass particles.

How to verify it in the field

Trace valve sequence and actual flow. Log each cell's rate, levels/head loss, and effluent turbidity; during wash, verify the load on remaining cells, wash flow, and even trough overflow.

2

A loaded column shows capture and head loss developing through depth

Darker attached solids occupy the upper bed while side pressure taps compare head at several elevations; cleaner lower media shows why total differential pressure alone cannot describe distribution.

A loaded column shows capture and head loss developing through depth：Segment pressure taps、High-solids upper zone、Continued depth capture、Local head loss and flow shift1234

What to identify

  1. 1Segment pressure taps
  2. 2High-solids upper zone
  3. 3Continued depth capture
  4. 4Local head loss and flow shift

Figure takeaway

Filtration is not only surface screening. A sound bed uses depth; rapid surface blinding raises head loss early, while poor effluent at modest head loss points toward weak microfloc, cracks, or short-circuiting.

How to verify it in the field

Plot individual-filter head loss, rate, and effluent turbidity/particles against run time. Use segment pressures when available and depth samples to confirm where solids accumulated.

3

Filtration flows down and backwash flows up through the same bed

The left column holds a stationary layered bed during downflow; the right sends wash water and air from below, expands the media, and carries detached solids out at the top.

Filtration flows down and backwash flows up through the same bed：Downflow through a fixed bed、Media and support layers、Upflow wash water/air scour、Expanded bed and dirty wash water1234

What to identify

  1. 1Downflow through a fixed bed
  2. 2Media and support layers
  3. 3Upflow wash water/air scour
  4. 4Expanded bed and dirty wash water

Figure takeaway

The aim is not maximum agitation. Backwash must detach and lift solids to the trough without losing media, rearranging support layers, or damaging the underdrain.

How to verify it in the field

Record surface-wash, air-scour, and low/high water-wash sequence, time, and rate. Measure temperature, resting depth, maximum expansion, whole-bed motion, and media in the trough.

4

Three columns separate underwash, an effective window, and over-wash

The left has clumps and uneven channels, the center is uniform with clearer supernatant, and the right is violently expanded toward media loss—three outcomes that a shared wash time can hide.

Three columns separate underwash, an effective window, and over-wash：Mudballs, binding, and channels、Uniformly cleaned bed、Over-expansion and media-loss risk、Post-wash samples and media loss1234

What to identify

  1. 1Mudballs, binding, and channels
  2. 2Uniformly cleaned bed
  3. 3Over-expansion and media-loss risk
  4. 4Post-wash samples and media loss

Figure takeaway

Identical wash duration can produce different cleaning. The operating window depends on temperature, grain density and size, depth, water/air rate, and distribution.

How to verify it in the field

Probe before and after, compare turbidity/particles, and measure rise rate and expansion. Track mudballs, media elevation, backwash solids, and first filtrate rather than judging surface turbulence alone.

5

A drained inspection finds mudballs, cracks, and dead zones hidden online

Operators inspect the exposed surface and localized clumps, then keep media cores from different positions separate; elevations and exposed hardware help trace uneven wash or media loss.

A drained inspection finds mudballs, cracks, and dead zones hidden online：Grid-based operator inspection、Surface elevation and cracks、Mudball or localized binding、Multi-location media cores1234

What to identify

  1. 1Grid-based operator inspection
  2. 2Surface elevation and cracks
  3. 3Mudball or localized binding
  4. 4Multi-location media cores

Figure takeaway

Total head loss and combined effluent can hide local defects. Mudballs, depressions, cracks, uneven depth, or support problems create repeatable shortcuts and require an offline grid inspection.

How to verify it in the field

Map media elevation, depth, cracks, depressions, mudballs, and sand in the clearwell. Link samples to the wash-distribution map, underdrain/nozzle inspection, and media addition history.

Six steps from incoming particles to removed backwash solids

Separate filtration, run termination, cleaning, and restart to locate the failed link.

  1. 1 Pretreat

    Raw water → filterable microfloc

    Create particles that attach without immediately blinding the surface.

  2. 2 Filter down

    Water ↓ media → underdrain

    Capture particles through bed depth and produce low-turbidity water.

  3. 3 End the run

    Head loss/turbidity/particles/time → isolate

    Stop before breakthrough or terminal head loss.

  4. 4 Loosen and detach

    Air/surface wash + wash water ↑

    Expand uniformly and break retained deposits free.

  5. 5 Remove solids

    Dirty wash water → trough/residuals

    Carry solids out and manage any recycle.

  6. 6 Resettle and ripen

    Rest/slow start/filter-to-waste → service

    Keep the initial turbidity spike out of finished water.

Responsibilities of four subsystems

Media does not act alone; pretreatment, bed hydraulics, washing, and restart constrain one another.

Coagulation/flocculation/clarification

Main job
Control the amount, size, strength, and surface of particles reaching filters
Typical failure
Unstable fines pass, weak large floc blinds the surface, or a solids surge overloads
Evidence
Raw/settled turbidity and particles, pH/chemicals, floc, filter loading

Media and operating flow

Main job
Provide attachment surface and bed depth while distributing filtration
Typical failure
Wrong grading/depth, media loss, cracks, excessive rate, or maldistribution
Evidence
Cell flow, head-loss curve, effluent, media elevation/grading, depth cores

Underdrain, air, and backwash

Main job
Distribute water/air, loosen the bed, and transport solids out
Typical failure
Dead zones, mudballs, underwash, over-wash, support or nozzle damage
Evidence
Flow/pressure, temperature, expansion, surface pattern, waste solids, media loss

Valves, filter-to-waste, residuals

Main job
Switch safely, control initial spikes, and manage dirty wash water
Typical failure
Surge/air binding, fast loading, poor first filtrate, or recycle shock
Evidence
Valve sequence, slow-start curve, cell turbidity/particles, waste quality and recycle rate/location

Backwash rate and expansion must be verified for the installed media, equipment, and water temperature. Typical expansion ranges in guidance are review starting points, not replacements for design criteria, manufacturer limits, local rules, and measured rise-rate/expansion tests.

Put three evidence groups on one filter-run timeline

Influent and loading

Flow, clarified-water turbidity/particles, temperature, pH/chemicals, floc condition, and actual rate of every online filter.

Bed response

Cell inlet/outlet level, total and segment head loss, effluent turbidity/particles, run time, valve position, and rate changes.

Wash and restart

Air/water flow and pressure, sequence, temperature, expansion/distribution, waste quantity/quality, media loss, filter-to-waste, and first filtrate.

Diagnose filter trouble from combined signals

Combined signal
Head loss rises quickly while individual-filter effluent stays clear
Suspect first
High solids or weak large floc, surface blinding, high rate, or incomplete prior wash
Next step
Align clarified-water and run curves, inspect the surface and wash pattern, then adjust pretreatment/load or cleaning—not only the timer
Combined signal
Turbidity/particles break through at modest head loss or in a sudden spike
Suspect first
Poor destabilization, media crack/short circuit, media loss, or valve/flow shock
Next step
Check cell rate and chemistry, compare effluent locations, then inspect media/underdrain and event logs
Combined signal
Parts of the bed stay still while others erupt and mudballs recur
Suspect first
Blocked/damaged distribution, uneven support, inadequate flow, or mismatched air-water sequence
Next step
Map bed motion, rise rate, and expansion; drain, grid-probe, and repair distribution components
Combined signal
An initial turbidity spike follows immediate return to service and then decays
Suspect first
No resettling/ripening, incomplete wash, rapid loading, or valve disturbance
Next step
Use evidence-ended filter-to-waste, rest and slow start, while confirming wash endpoint and valve sequence

Four misconceptions

Sand filters only sieve through grain gaps

Particles also collide, attach, and settle along tortuous pores; pretreatment and surface interactions matter as much as openings.

A fixed timer is the safest backwash trigger

Time cannot represent changing solids load or bed condition; combine individual-filter head loss, turbidity/particles, and operating limits.

A harder backwash is always cleaner

Over-expansion can lose media, disturb supports, or damage distribution; the goal is uniform detachment and removal.

Clear wash water means immediate service

Media still must resettle and ripen, and first filtrate can spike; restart should be controlled by individual-filter evidence.

Continue learning

Next illustrated guideHow does chemical precipitation remove heavy metals?Physicochemical treatmentHow does ion-exchange resin soften water and regenerate?Physicochemical treatmentWhy is a high-rate clarifier faster than a conventional clarifier?