Mapping Water Addition Through Coconut Milk Extraction

Locate every water addition point

Water enters a coconut milk line at more than one place, and each entry has a different meaning. Kernel washing carries some water forward with the prepared material. Extraction may add water to the grated kernel before or during pressing. Formulation adds water as a recipe ingredient. Cleaning and product flushing use water in circuits that should be isolated from production. If all of these are recorded as one number, the resulting yield figure cannot be interpreted.

Start the map at the receiving scale. Record the mass of prepared kernel and its measured moisture or free-water condition. Then identify every point where water is deliberately added, metered or introduced by an upstream operation. Give each point a name, an owner and a measurement method. A line operator may know that a hose is used; the brief needs to know how much water that hose delivers and why.

The map should also show where water can leave the product path. Drainage from a washer, condensate, spills and cleaning returns are not recipe additions. They should be captured as losses or separate streams, not folded into the extraction calculation.

Separate kernel input from press meal

A useful extraction balance has at least four weighed streams: prepared kernel in, water added, extracted liquid out and pressed meal out. If the press has more than one liquid outlet, name each one. If a second extraction is used, treat it as a separate step with its own inputs and outputs rather than averaging it into the first.

Weighing every stream during a stable run is more informative than sampling a single product tank after the fact. Record the run duration and any interruptions. A press that is stopped and restarted may discharge material that is not representative of steady operation. Note where samples were taken and whether they were taken from a flowing line, a tank or a collected container.

The balance does not need to close perfectly on the first attempt. It needs to show where the uncertainty sits. A large unexplained difference may indicate a measurement problem, an unrecorded water addition or product held inside the equipment. Those are different findings, and each points to a different correction.

Read fat and solids on consistent mass bases

Fat and total solids results are only comparable when they refer to the same material and the same basis. A percentage measured on wet kernel is not directly comparable with a percentage measured on extracted milk. A sample taken before screening contains coarse fibre that a sample taken after screening does not. Record the sampling point, the sample preparation and whether the result is expressed by mass, by volume or on a dry basis.

During an extraction trial, analyse the kernel feed, the extracted milk and the residual meal. That trio supports a recovery calculation and shows whether fat is leaving with the meal. A single milk-fat result can describe the product but cannot establish how much fat entered the process. Where the feed is variable, take more than one sample across the run and report the range rather than a flattering average.

The same discipline applies to added water. If the water contains solids or is reused, its contribution should be measured or at least bounded. Potable water is not automatically a zero-impact ingredient in a mass balance, particularly where reuse or recovery is practised.

Why dilution does not enrich cream

Adding water increases the total mass of liquid collected, but it lowers the fat concentration unless the extra water carries additional coconut solids. A larger volume at a lower fat content is not evidence of a richer extraction. This is simple conservation, yet it is easy to obscure when a proposal quotes “milk output” without specifying composition.

The mistake becomes more serious in a cream project. If the incoming milk does not contain enough fat to meet the cream target, adding water cannot close the gap. The options are a richer feed, a different extraction arrangement or a separately assessed concentration step. Each has its own equipment, energy and product-quality consequences.

A trial report should therefore state both mass and composition at each boundary. “We collected more litres” is not a recovery result. “We recovered this much fat in the milk stream, with this much remaining in the meal, from this measured kernel input” is a result that another engineer can examine.

Build a measurement sheet before a recipe

The deliverable from this exercise is not a universal extraction recipe. It is a measurement sheet tailored to the plant: stream names, instruments, sample points, frequency, units and acceptance rules. It should be reviewed by the people who will actually collect the data, because a method that disrupts production will not survive a busy shift.

Use the sheet in a short trial before fixing equipment guarantees or output claims. Compare the measured water additions with the design assumptions. If the press requires more water than expected to convey material, that is a process finding, not a reason to restate the yield. The next design iteration can then address the real constraint.

Finally, keep extraction water separate from formulation water in the finished-product record. The first belongs to recovery; the second belongs to recipe and product identity. Combining them may make a line look efficient while making the product specification harder to defend.

What the exercise does not prove

A clean mass balance does not establish shelf life, microbial safety, consumer acceptance or a guaranteed output. It does not prove that a particular press will perform in the same way on a different kernel supply. It only shows what happened in the measured run, within the stated boundaries.

That is still valuable. It turns a vague discussion about yield into a set of observations that suppliers and plant teams can review together. In a project where water, fat and fibre all move at once, that clarity is the foundation for every later decision.