Why Coagulant Trial Results Change Between Production Batches

Three things change at once

A jar test result is the product of the water, the product and the method. When only the result is written down, the cause cannot be recovered afterwards, and the next batch will be compared against a number whose conditions nobody remembers. The practical response is not a better test; it is a test designed so that the causes can be separated, plus a record that makes the separation possible.

Rule out the method before blaming the product

Most of the drift between two runs happens before the coagulant is poured. Check the following, and check them against the conditions of the run that worked:

· Sampling point, depth and the time of sampling relative to the process

· Holding time and temperature between sampling and testing

· Whether the sample was sealed or left open to the air

· Dilution water: the same source, the same temperature, the same volume

· Dosing solution concentration and the interval since it was prepared

· Pipette or pump calibration, and the volume actually delivered

· Mixing device, speed setting and duration, recorded as a number

· Settling time, sampling depth and the moment the reading is taken

· The instruments used and their calibration — turbidity meter, pH meter

If two of these moved between the good run and the bad one, the result has already changed for reasons that have nothing to do with the product, and the comparison is not yet a comparison.

Characterise the water, including what the plant was doing

Water is where most of the real variation sits, and it is the part least often recorded. Measure it on every run, not once at the start of the campaign.

Table 1 — Water record for each trial run

Two or three lines of production context save a week of guessing: which line or batch was discharging, whether there was a cleaning, shutdown or start-up in the hours before sampling, how long the flow spent in the equalisation tank, and whether rain or a return stream entered the works.

Check the product and how it was prepared

Product variation is real, but it is usually smaller than the variation in the water, and it is easier to document. Take the following from the label and the COA rather than from memory:

· Grade and batch number, and whether this drum is from the same batch as the earlier trial

· Active content and basicity as stated for that batch

· Storage history — heat, frost, direct sun, opened and resealed, long storage

· Dissolution ratio used, and the water and temperature used to make the solution

· Appearance of the product and of the diluted solution, recorded rather than remembered

A four-way test that gives an answer

The decisive move is to keep a sample of the water from the run that worked. Seal it, keep it cool, write the date and time on the container, and use it within the period your laboratory considers valid. Then run four combinations on the same day, on the same bench, with the same dose ladder of at least four points and duplicate beakers at the control and the expected optimum.

Table 2 — Four-way comparison and what each result establishes

The reference cell is the control of the whole exercise. A session in which the reference does not reproduce has produced four numbers and no conclusion.

The batch comparison record

One sheet per production batch keeps the history readable when the next change arrives.

Table 3 — Batch comparison record, one sheet per production batch

Conclusions to avoid

· A single beaker is not a result; without duplicates it is an anecdote.

· Two changes made at the same time cannot be separated after the fact.

· Floc that looks large is not the same as water that settles clear.

· One dose cannot compare two products; the shape of the dose ladder carries the information.

· A product that is worse at low dose and better at high dose has not failed; it has a different working window.

· Do not change the dilution or the mixing at the same time as the product.

What to send with your inquiry

If you want us to look at a batch that behaves differently, send:

· The water analysis for both runs — pH, alkalinity, turbidity, suspended solids, COD

· The production schedule and product mix discharged before sampling

· The coagulant grade, batch and COA you are using now

· The target you must meet — turbidity, suspended solids or colour

· Monthly quantity and destination port, if a price comparison is also needed

With the two analyses and the two COAs we can usually tell you which of the three variables moved, and which grade of the same family covers the changed water. We will not promise a fixed dose or a fixed removal rate from a jar test result; those come from your own operating record.

A batch that behaves differently is information, not a fault. Recorded properly, it becomes the specification for the next one.