Changing filter screens on a fixed schedule is simple and usually wrong. Some cycles foul fast, some barely foul at all, and a calendar treats both the same. Differential pressure across the screen is the honest signal – provided the threshold came from your process rather than from a round number.
What pressure drop actually tells you
As a screen fouls, the open area falls and the resistance to flow rises. Measure the pressure either side of the screen and the difference tracks the fouling. It is a direct, continuous, cheap measurement, and it works whether the deposit is a cake, a blockage or a film.
Why not flow rate
Flow rate is the thing you actually care about, but it depends on everything else in the system – pump performance, other restrictions, valve positions. Pressure differential isolates the screen, which makes it a cleaner signal. If your system is already flow-metered, use the flow, but normalise it against the pressure.
Setting the threshold
Do not start from a number. Start from your system’s tolerance:
- What differential can your pump deliver against before flow becomes unacceptable?
- At what differential does the risk of screen deformation or bypass become real?
- What does the process need in terms of minimum flow?
Then set the change point below the tightest of those with a margin, and instrument it. A gauge either side of the housing is enough – it does not need to be a transmitter.
The shape of the curve matters more than the number
Log the differential over a full run and look at the shape. Gradual rise means cake formation and a predictable interval. A sudden jump means something changed – a change in the feed, or a screen that has started to deform. Flat then steep means a threshold effect, and it will be hard to predict.
That curve also tells you whether cleaning is worth doing. If the differential climbs back to the threshold within a short time after cleaning, the screen is not the constraint and cleaning is wasted effort.
Common mistakes
- Measuring downstream of other restrictions, so the reading includes pipework and fittings
- No baseline, so nobody knows what a clean screen reads
- A threshold set too low, which wastes screen life
- A threshold set too high, which risks damage and bypass
- Ignoring temperature, because viscosity changes with it and so does the reading
What to record
Clean differential, differential at change-out, hours or volume between changes, and what the screen looked like. Four data points per cycle, and after a few cycles you will know your interval properly instead of guessing.
What to do when the reading is noisy
Differential readings are often erratic, which leads people to distrust them and go back to a calendar. Usually the noise has a cause worth finding rather than ignoring.
Common ones: tapping points that are partly blocked, a gauge that is too large for the range being measured, readings taken at different flow rates, and temperature variation changing the viscosity. All are fixable, and a stable reading is worth having because it makes everything else easier.
If the noise is genuine process variation, use a trend rather than an absolute threshold – the rate of rise is often a better signal than the level, and it is less sensitive to shifts in operating conditions.
What we can supply
We build filter screens to a stated open area, which makes your clean differential predictable and comparable between batches. If your clean reading drifts from one delivery to the next, that is worth investigating – it usually means the mesh specification has quietly changed.