Why a logbook of spot analyses answers what was rather than what is happening: the three purposes of measurement, why trend beats value, and where control points belong.
Water quality control on a plant usually looks like this: once a shift an operator takes a sample, carries it to the laboratory, and the result goes into a logbook. Formally, control exists. In practice it answers "what was" rather than "what is happening" — and that is a fundamental difference.
A laboratory analysis shows the state of the water at the moment of sampling. Everything that happened between two samples stays invisible. If a deviation occurred and ended inside that interval, the logbook will not show it — but the equipment has already lived through it.
Why we measure: three different tasks
Process control
Measurement is needed to make a decision here and now: how much reagent to dose, whether it is time to clean, whether blowdown should increase. This requires continuity rather than precision to the last digit. A sensor linked to a dosing pump does not merely record a deviation — it prevents it happening.
Demonstrating compliance
Here it is the opposite: accuracy, a recognised method and traceability matter. That is a laboratory task, and an online sensor does not replace it.
Diagnosis
Understanding why the system behaves differently from expectation. Here the measurement history is the most valuable thing: a single value says almost nothing, while a month of trend explains a great deal.
Conflating these tasks is a common mistake. A laboratory logbook is unsuitable for process control, and an online sensor is unsuitable for reporting. Both are needed.
What is usually measured continuously
- pH. The base parameter: coagulant performance, metal precipitation and corrosivity all depend on it. It changes fast and measures fast — an ideal candidate for automation.
- Conductivity. An indirect measure of dissolved solids. In recirculating circuits it shows cycles of concentration; on membrane plants it shows permeate quality and membrane integrity.
- Turbidity. Shows how clarification and filtration stages are performing. Rising turbidity is an early signal that an earlier stage has stopped coping.
- Oxidation-reduction potential. Reflects the presence of an oxidiser or reducer. A key parameter wherever residual chlorine must not reach the membranes.
- Flow. Not a quality parameter in itself, but without it dosing cannot follow load.
Why the trend matters more than the value
A single measurement is almost always misinterpreted, because there is nothing to compare it against. The real information is in the movement.
Steadily rising dissolved solids at unchanged blowdown means make-up or evaporation has changed. Turbidity climbing faster than usual before a clean says the preceding stage is degrading. Residual reagent falling faster than normal signals rising load long before it becomes visible.
That is why it matters not only to measure but to store measurements so they can be compared. A logbook you cannot build a trend from is an archive, not a tool.
Normalisation: without it comparison is meaningless
Some parameters depend on measurement conditions. Membrane plant output varies with temperature and pressure; summer and winter readings cannot be compared directly — the difference is physics, not equipment condition.
Without correcting to comparable conditions, degradation is noticed late: everything looks fine in summer, the winter drop is blamed on the cold, and real capacity loss accumulates unnoticed.
Where to put control points
The logic is simple: measure where a decision is made, and where responsibility passes from one stage to the next.
At the inlet, to know what arrived. After each key stage, to know whether it coped. At the outlet, to confirm the result. At the discharge point, for reporting. Monitoring only at the outlet tells you something is wrong but not where.
Common mistakes
- Sensors installed but never calibrated. An uncalibrated sensor is worse than none: it creates false confidence.
- Measuring without linking to dosing. The data exists, but the response is manual and late. The gain comes from the sensor–controller–pump chain.
- Sampling at a convenient time. Sampling outside peak load shows water that does not actually occur on site.
- Streams mixed before the measurement point. An average across the common sewer hides which department creates the problem.
- Data not retained. Without history you can neither diagnose nor demonstrate compliance over a period.
Frequently asked questions
Where do we start if we only have laboratory control now?
With pH and conductivity at critical points. Those are the cheapest and most informative measurements, and they show immediately whether the system is worth extending.
Does online monitoring replace the laboratory?
No. The tasks differ: an online sensor controls the process, the laboratory demonstrates compliance. Online data alone is usually not accepted as evidence.
How often should sensors be calibrated?
To the manufacturer's schedule and to operating conditions: aggressive media, fouling and high temperature shorten the interval. Regularity matters more than frequency.
How many parameters are enough?
As many as decisions require. An excessive set of sensors nobody watches is worse than a small set built into the control loop.
Getting started
Identify which decisions on site are currently made blind — that is where measurement is needed. More on the approach in chemical consulting; questions through the contact form.
