Cold Stabilisation Chill Load and Loss Calculator

Pulling a tank from 60F down to 26F is a bigger job than it sounds, because wine has most of the heat capacity of water and a full tank is several tonnes of it. The chiller does not care about tartrates — it only sees mass times specific heat times a temperature difference — and then the tank sits at temperature for a week or two while the compressor keeps fighting whatever leaks in through the shell. This works both halves, plus the wine the lees keep.

Your own target for this wine, from your own practice. This page has no opinion about what temperature anything should be held at.
kg/L
A dry wine sits a little under water because of the alcohol. Your own hydrometer figure is better than any printed one.
kJ/kg per K
Wine is mostly water, so the figure sits below the 4.18 of pure water by roughly the alcohol content. Use the value your own reference or your own refrigeration contractor gives — this page supplies no property data.
During harvest the same plant is holding ferments down. Whatever is left over is what pulls this tank down, and the honest figure is well under 100.
kW
Through the shell, the manway and the legs, plus anything running in the cellar. From your own refrigeration contractor or from how hard the compressor works on a holding tank.
The coefficient of performance of your plant at this suction temperature, from the plant data or your contractor. It falls as the target temperature falls, which is why cold stabilisation costs more per degree than cooling a ferment.
$
What stays behind when you rack off the crystals. Your own measurement — it depends on the tank shape and on where the racking valve sits more than on the wine.
Across every pass you run afterwards, not one.
Cold Stabilization Calculator — Chill Load and LossBuildFigure

Mass times specific heat times the drop

There is only one equation in the first half of this page and it has no wine in it. 5,000 litres at 0.99 kg/L is 4,950 kg. A drop from 60F to 26F is 34 Fahrenheit degrees, which is 18.89 kelvin, because a Fahrenheit degree is five ninths of a kelvin — the single most common mistake in this arithmetic and worth checking on any figure that came out of a spreadsheet somebody else built.

4,950 kg times 3.8 kJ per kg per kelvin times 18.89 K is 355,300 kJ, or 98.7 kWh of cooling, or 336,759 BTU. That is 28.06 ton-hours of refrigeration for one tank, which is why cold stabilisation gets scheduled around the ferments rather than alongside them.

The share field is the honest one

A 10 ton plant is 35.2 kW on paper. During harvest most of it is holding ferments down, and the fraction left over is what actually pulls a stabilisation tank. At 60 percent that is 21.1 kW, and 0.8 kW of leak-back through the shell takes it to 20.3 kW net. 355,300 kJ at 20.3 kW is 4.86 hours.

Set the share to 100 percent and it is 2.87 hours. Set it to 15 percent and the plant delivers 5.28 kW against 0.8 kW of leak-back, and the tank takes 22.05 hours. Set it low enough and the leak-back wins outright and the tank never gets there at all, which the page says plainly rather than printing a large number. That is not a hypothetical: it is September in a warm cellar.

The hold is often the bigger job

Ten days at 0.8 kW of leak-back is 192 kWh of cooling, against 98.7 for the pull-down. Nearly two thirds of the total load is the tank simply sitting there, and none of it depends on how cold the wine was when it arrived. Insulation and a shorter hold are the two levers, and only one of them is a winemaking decision.

At a coefficient of performance of 2.5 the 290.7 kWh of cooling draws 116.3 kWh of electricity, which at 16 cents is 18.60 dollars. Per bottle that is 0.279 cents. Cold stabilisation is a slow job and a scheduling problem, and on this evidence it is not an expensive one — which is worth knowing before anybody optimises it.

What the lees keep

The second half of the page is volume, not energy. 1.5 percent left with the crystals and 1 percent at the filter afterwards is not 2.5 percent: it is 2.485 percent, because the filter only works on what came off the lees. On 5,000 litres that is 124.3 litres, or 166 bottles, or 13.8 cases.

The line underneath is the one that gets used most. To finish with 5,000 litres rather than start with them, you need 5,127 in the tank — 5,000 divided by the 0.98515 keep factor, not 5,000 plus 2.5 percent, which would leave you 2 litres short and is wrong in a way that gets worse the longer the chain.

Questions people ask

How much energy does it take to cold stabilise a tank of wine?

Mass times specific heat times the temperature drop, and nothing else. 5,000 litres at 0.99 kg/L dropping 34F is 4,950 kg times 3.8 kJ/kg/K times 18.89 K, which is 98.7 kWh of cooling or 336,759 BTU. Then the hold adds its own load: ten days of 0.8 kW leaking back is another 192 kWh, which is usually the larger number of the two.

Why convert Fahrenheit degrees to kelvin?

Because specific heat is quoted per kelvin and a Fahrenheit degree is five ninths of one. A 34F drop is an 18.89 K drop. Feeding 34 straight into the equation overstates the load by 80 percent, and it is the single most common error in cooling arithmetic — the more so because the answer still looks like a plausible number of kilowatt hours.

How long does a chiller take to pull a tank down?

The heat to remove divided by the net cooling reaching the tank, where net means the plant share less whatever leaks back in. 355,300 kJ at 20.3 kW net is 4.86 hours. The share is the field that decides it: the same plant at 15 percent instead of 60 takes 22.05 hours, and low enough that the leak-back matches it, the tank never arrives at all.

How much wine is lost to cold stabilisation?

Whatever stays with the tartrate lees when you rack, plus whatever the filtration afterwards keeps. At 1.5 and 1 percent that is 2.485 percent chained rather than 2.5 added, or 124 litres on a 5,000 litre tank — 166 bottles. Both figures are yours to measure; the lees loss depends on the tank shape and the racking valve height more than on the wine.

How much wine do I need in the tank to finish with 5,000 litres?

5,127 litres. Divide by the keep factor of 0.98515 rather than adding the loss percentage back on. Adding 2.5 percent to 5,000 gives 5,125 and leaves you two litres short, and the error grows with every stage you chain, which is why the page prints the division rather than the addition.

Related