Artificial intelligence can save green houses a fifth of electricity bills

When Putin invaded Ukraine, he attacked not only his blue-yellow neighbour to the west, but also unintentionally the growing season in Danish green houses. Since 24 February, gas prices have risen by 700 % and electricity prices have doubled.  

The green house Knud Jepsen A/S, located in the Aarhus suburb of Hinnerup, is the world's largest producer of Kalanchoe, popularly known as Flaming Katy. The green house is 120,000 m². Grow lights are installed throughout the nursery. It is expensive. Very expensive. But there is money to be saved.  

In collaboration with Aarhus University and the Knud Jepsen green house, researchers from the Mærsk McKinney Møller Institute at the University of Southern Denmark have developed an algorithm that optimises lighting control in relation to electricity prices.  

- You typically see electricity prices fluctuate a lot. Sometimes there is an oversupply on the grid, so energy prices become negative. In the DynaGrow research project, we have developed an algorithm that implements electricity prices in horticultural lighting control. We have found the right balance with the lowest possible energy consumption while maintaining the right quality of the plant, says associate professor Jan Corfixen Sørensen from the University of Southern Denmark.  

The computer-based systems implement weather forecasts so they can predict how electricity prices will be in the future. The weather forecast is also used to calculate how much growing light needs to be supplied to meet the needs of the plants. Energy Denmark provides a price forecast for the electricity the green house needs to buy, and DynaGrow then suggests when it is cheapest to switch on the grow lights. 

- We make forecasts that save money based on data-driven decision-making, says Jan Corfixen Sørensen.  

Has saved 20 percent on electricity 
And there is money to be saved. Production manager Gunner Holm from Knud Jepsen A/S says that the green house used 31 percent less electricity in January compared to last year.  

- If we look at the whole year, we've probably saved 20 percent on consumption. Now that prices are so crazy, there is really money to save, says Gunner Holm.  

Flaming Katy is a special flower. At the beginning of its life, it needs plenty of light to grow. Up to 20 hours a day. Therefore, it is interesting in the context of electricity savings.  

- It needs maybe 15 hours of light a day on average. Based on the weather forecast and the algorithm, we can then give it 10 hours one day if power is extremely expensive, and 20 the next when it's less so. At this stage, there is a real opportunity for optimisations in relation to electricity prices, says Gunner Holm. 

Putin has accelerated the green transition 
Initially, the Dynagrow project was launched to address the green transition, but with rising electricity prices, the speed of the project has accelerated sharply. Forced by economic circumstances. 

- We have tightened the bow hard and adjusted the power consumption a lot to see how much we could save. Under normal circumstances, we might have turned the knob a quarter turn, but instead, we've given it a full spin. It has meant that we have got to the point where we are now much faster than we would be otherwise, says Gunner Holm.  

He also says that the green house has observed that some varieties have been shown to perform better than others at lower light levels, which will be a parameter going forward when choosing new varieties. The green house will look for those characteristics and cross them with other varieties. 

Potential for savings in other sectors
Not all green houses and plants can benefit from electricity savings through data-driven lighting control. Tomatoes, for example, need as much light as possible, so it is not possible to adjust the power according to electricity prices. But then other industries can do it, says Jan Corfixen Sørensen. 

- In the transition to renewable energy sources, we see fluctuating energy prices depending on sun, wind and weather. At the same time, prices rise in the morning when people do their laundry, and again in the evening when dinner needs to be cooked and the electric car needs charging. We can take this into account and avoid the times when electricity prices are highest, says Jan Corfixen Sørensen. 

He goes on to say that green houses have been a prominent industry to test the technology, but that it can be extended to other energy-intensive industries where there is an opportunity to optimise energy consumption. Knud Jepsen A/S is one of more than 200 green houses in Denmark, so the potential in the industry is enormous. 

- It could be feed producers. Drying feed is costly. The construction industry is also an obvious choice, as concrete drying uses an incredible amount of energy. But we can't just transfer the technology one-to-one. We need to optimise domain-specifically. Green houses have specific requirements, other industries have different ones, so it's important to put the end user first and create tailor-made solutions.