Air Quality Control Systems
Why the parameters inside your scrubber, not the annual stack test, decide whether your permit holds up.
A wet gas scrubber that passed its acceptance test is not a scrubber that will pass every hour for the next ten years. Between two periodic measurements sits a year of wear, fouling and shifting process conditions, none of which announce themselves at the stack until the margin is already gone. This article looks at what actually drifts inside a scrubber, which measurements see it coming, and how to turn that data into a record an inspector will accept.
01 The blind spot between two stack tests
Periodic measurement answers one question, once: was the plant within its limit on the day the measurement team was on site. It is a snapshot, and everyone involved knows it. The permit, however, applies to every operating hour in between.
That gap is not a paperwork problem. It is where the expensive surprises live. A scrubber degrades slowly and quietly, and the first visible sign is usually the one measurement that cannot be argued with: an exceedance at the stack, recorded, reportable, and impossible to explain because nothing upstream was being logged.
Working from the stack test alone
- Confirms compliance for a single day
- Shows the result, never the cause
- Degradation is found after the fact
- Maintenance is scheduled by calendar, not condition
- An exceedance leaves you without a defence
Working from continuous process data
- Every operating hour is accounted for
- Cause and effect are visible together
- Drift is caught while margin remains
- Maintenance follows measured condition
- Deviations come with a documented trail
Continuous monitoring does not replace the statutory measurement. It removes the uncertainty either side of it.
02 What actually drifts inside a wet scrubber
Removal efficiency in a wet gas scrubber rests on a small number of physical conditions: enough liquid, finely enough distributed, in contact with the gas for long enough, with the chemistry to absorb what is in it. Every failure mode is an attack on one of those four. Most of them are gradual.
| What changes | What it does to performance | How it looks from the control room |
|---|---|---|
| Nozzle wear | A larger orifice passes more liquid at lower pressure. Droplets coarsen, total surface area drops, and the spray pattern loses its edges. | Liquid flow looks healthy, even slightly generous. Only the pressure has moved. |
| Partial plugging | Solids or scale block individual nozzles. Part of the cross-section is no longer irrigated, and gas takes the dry path. | Header pressure rises while total flow falls. Easy to miss without a baseline. |
| Falling L/G ratio | More gas, or less liquid, than the design point. Absorption capacity per unit of gas drops in direct proportion. | Both flows individually within range. The ratio between them is what moved. |
| Chemistry drift | pH, reagent concentration or dissolved solids move away from the operating window and absorption slows or stops. | Visible immediately in pH and conductivity, if anyone is trending them. |
| Pump or strainer degradation | Impeller wear, or a slowly blocking suction strainer, reduces the flow and the pressure delivered to the spray banks. | Header pressure and liquid flow fall together. That is what separates it from nozzle wear. |
| Mist eliminator loading | Carryover increases, taking dissolved and suspended load out of the stack with it. | Differential pressure rises. Visible plume, sometimes deposits downstream. |
| Operation above design gas flow | Contact time shortens and entrainment rises. Often the result of a process change nobody flagged to the scrubber owner. | Fan load and pressure profile shift together across the whole system. |
None of these are exotic. What they have in common is that the stack sees them last, and sees only the sum of them.
03 The measurements that see it first
A useful monitoring set is smaller than most people expect. The point is not to instrument everything, but to measure the few variables that move before the emission does, and to compare them against the values recorded when the plant was known to be good.
- 1
- Spray header pressure. Wear drives it down, plugging drives it up. The earliest and cheapest indicator on a spray scrubber.
- 2
- Liquid flow per bank. Measured per header, so a failing section cannot be masked by a healthy one.
- 3
- Differential pressure. Across the mist eliminator, and across the vessel as a whole. The slope of the trend matters more than the figure.
- 4
- pH and conductivity. The chemistry of the recirculating liquid, logged continuously rather than sampled by hand.
- 5
- Gas temperature in and out. The approach to saturation indicates evaporation, liquid loss and whether the gas is really being contacted.
- 6
- Stack concentration. The outcome, not an early warning. Shown in the darker marker because it is the last parameter to move.
Nozzle header pressure
The single most informative measurement on a spray scrubber, and usually the cheapest. Wear drives it down, plugging drives it up. Neither shows in a flow reading with any clarity.
Liquid flow per bank
Measured per header rather than as a plant total, so a failing section cannot be masked by a healthy one elsewhere in the vessel.
Differential pressure
Across the mist eliminator, and across the vessel as a whole. The slope of the trend matters more than the absolute figure.
pH and conductivity
The chemistry of the recirculating liquid, logged continuously rather than sampled by hand on day shift.
Gas temperature in and out
The approach to saturation is a quiet indicator of evaporation, liquid loss and whether the gas is really being contacted.
Stack concentration
The outcome, not an early warning. Its value in this set is that it lets you correlate every parameter above with the number that actually appears in your report.
The gain is in the correlation. A rising differential across the mist eliminator on its own is a maintenance note. The same rise alongside a falling header pressure and a slow climb in stack concentration is a diagnosis, and it arrives weeks before the limit does.
04 Data is not yet evidence
There is a difference between numbers on a screen and a record that survives scrutiny. Where continuous emission monitoring is part of the permit, the framework is already written down in EN 14181, and it is worth knowing which part does what.
- QAL1
- Establishes that the measuring system is suitable for the application before it is installed.
- QAL2
- Calibrates the installed system against a standard reference method and fixes the valid calibration range.
- QAL3
- The ongoing drift and precision checks that keep the system inside that calibration between audits.
- AST
- The annual surveillance test that confirms the calibration is still valid.
In practice, the records that get questioned are almost never the emission values themselves. They are the things around them: whether the zero and span checks were performed and by whom, whether the calibration gases were traceable and in date, how much data was available over the reporting period, and what happened to the measurements during the hours when the analyser was in maintenance mode.
Process data strengthens that position considerably. If an exceedance is accompanied by a logged pump trip, a documented response and a return to normal within the hour, the conversation is about an incident that was managed. Without that context, it is about a plant that was out of control for an unknown length of time.
Keep the process log and the emission log on the same time base. Reconciling two systems with clocks that differ by minutes is a problem you only discover when you are already under pressure.
05 Setting it up without over-engineering it
Most of the value comes from the first two steps. The rest is refinement.
Record the good state
Log header pressures, flows, differential pressures, chemistry and temperatures while the plant is demonstrably performing, ideally during the acceptance test. Without that reference, every later reading is a number without a meaning.
Set bands from the design point
Derive alarm and warning limits from the design case and the recorded baseline, not from whatever the instrument shipped with. A warning band that sits well inside the point of non-compliance is what buys you time.
Trend, do not just alarm
Wear and fouling are slopes, not events. A parameter moving steadily in one direction for three weeks is more informative than any single threshold crossing.
Tie each signal to an action
Every monitored parameter should have an owner and a defined response. A trend nobody is responsible for is a trend that gets acknowledged and left.
Close the loop at the next inspection
Compare what the data predicted against what you found inside the vessel. Two or three cycles of that and the bands are tuned to your plant rather than to a generic assumption.
06 Start where the performance is made
Scrubber performance is created in the spray: in the droplet size, the distribution across the vessel and the liquid-to-gas ratio at the point of contact. That is also where degradation starts, long before it reaches the stack. Nozzle selection, spray pattern and wear behaviour are Lechler's core discipline, and they are the parameters most often left unmonitored on an otherwise well-instrumented plant.
We design, build and service wet gas scrubbers, and we work daily with the analyser technology that measures their result. If you are reviewing an existing installation, a comparison of your current operating data against the original design case is usually the fastest way to find out how much margin you still have.
Frequently asked questions
Does continuous monitoring replace periodic stack measurement?
No. The statutory measurement remains the reference, and where a continuous system is installed it is the periodic reference measurement that calibrates it. Process monitoring covers the hours in between, which is where the risk actually sits.
What is the minimum sensible instrument set?
Pressure at each spray header, liquid flow per bank, differential pressure across the mist eliminator, and pH of the recirculating liquid. On most existing scrubbers a large part of that is already fitted but not logged or not trended.
Can this be retrofitted to an older scrubber?
Usually, yes. The instrumentation is standard and the tapping points are generally accessible. The work that takes the time is not the hardware but establishing what the correct values for your vessel are, which requires the original design data or a performance assessment.
How do we know the analyser data itself is reliable?
Through the ongoing quality checks the framework already prescribes: regular zero and span checks against traceable gases, documented drift assessment, and a maintained record of availability. An analyser without a current calibration record produces data, not evidence.
Standards referenced. Quality assurance of automated measuring systems, including QAL1 to QAL3 and the annual surveillance test (AST): EN 14181; certification of such systems: EN 15267.