Buying light when it is cheap: DLI targets and half-hourly tariffs

In a lit growing operation the lights are usually the biggest variable cost, and the schedule running them was often set at commissioning and never looked at again. On a time-of-use or agile tariff, the price of the electricity behind that schedule changes every half hour.

Start from light, not hours

A crop does not need twelve hours of light as such. It needs a daily light integral (DLI): an amount of photosynthetically active light per square metre per day. With a steady canopy PPFD, the hours follow from arithmetic:

hours = DLI × 1,000,000 / (PPFD × 3600)

DLI is in mol/m²/day and PPFD in µmol/m²/s, and Apogee explains the measurement in more detail. The formula assumes a steady light level; if output changes through the day, or daylight contributes in a glasshouse, the light has to be added up over time instead.

Use a PPFD that represents the crop, not one bright reading directly under a fixture. And DLI is not the whole recipe: intensity and the light and dark periods matter too, and some crops will not tolerate their hours being moved at all.

Then price every start time

Once the hours are known, what they cost depends on when they start. BeeGrow takes the lighting load you enter and a half-hourly rate profile for the day, prices every possible start time, including windows that run past midnight, and shows the cheapest one that still delivers the light. Overnight is often cheapest, but not always: on agile tariffs the shape of the day moves.

The sums are done in integer micro-pence and converted once at the end, so adding up 48 half-hourly costs does not build rounding errors into a comparison that may turn on fractions of a penny. The standing charge is kept separate, because moving the lights changes what you pay per unit, not the fixed daily charge. The Octopus API documentation describes the tariff and consumption data behind this.

What it does not tell you

A cheaper window still has to suit the crop and the site. The comparison does not model how moving the lights changes heating or cooling load, which on some sites matters a great deal. It is priced from the rate data you select, so a historical profile describes the past rather than tomorrow’s prices. And it cannot tell you whether the DLI target in your recipe is right: if the target is wrong, the scheduler just makes you efficiently wrong.

BeeGrow proposes the schedule; it does not move your lights on the strength of a tariff. Measure actual consumption and crop results during a trial before counting a projected saving as a real one. More detail is on the energy and tariff optimisation page.