Structures & Equipment
Artificial intelligence: A digital brain for your greenhouse
By Don Horne
How innovations in technology are making the complicated simple and profitable
By Don Horne
There is an ever-growing demand for experienced greenhouse growers – and a concurrent need to control an increasingly extensive area due to increased scale of operations.
As global populations continue to grow and the need to feed more mouths expands apace, greenhouses are the perfect solution. But who will operate these greenhouses? Enter artificial intelligence (AI), and the computing power that comes with it.
Data and algorithms are being used to improve growing operations, says the CEO of Blue Radix, Ronald Hoek. “You can actually translate the growth of a tomato plant into data points, and use that information to optimize its development.” Algorithm-based solutions offer growers a digital brain for their greenhouse. According to Hoek, he believes that algorithms can feed the world, allowing growers to optimize their greenhouse climate, reduce their water and energy usage and limit their operational risks and their ecological footprint. And by drawing on the expertise from a wide range of experts and growers, algorithms can be generated that will optimize growth and yield for greenhouse growers.
“Algorithms can also make use of experience, tools and solutions from different expert companies. A good example is lighting,” says Hoek. “You can optimize lighting to a very detailed level. Why should we reinvent the wheel by finding the requirements for lighting and optimization solutions, if we can use the knowledge from an existing company or program that specializes in them? The same applies to IPM (integrated pest management). Artificial intelligence pulls all this data together.”
Blue Radix’s Crop Controller has been developed with support from Microsoft on the best possible usage of their Azure Platform. Crop Controller optimizes and steers climate and irrigation in a greenhouse with no human interference. The grower sets the crop strategy for the coming period. The algorithms of Blue Radix Crop Controller work with this crop strategy, learning and adapting to realise the greenhouse climate and irrigation accordingly.
Blue Radix colleagues have built more than six years of experience with the application of artificial intelligence in greenhouse operations around energy services at AgroEnergy – a leading Dutch Energy Company that offers smart autonomous energy solutions to greenhouses in the Netherlands.
For many, the idea of introducing automated systems raises the spectre of lost jobs. More computers and automation mean fewer people who traditionally did the work. Right? Wrong.
“The number of experienced and skilled growers is decreasing,” says Hoek. “As the sector grows, the amount of people who can operate greenhouses is decreasing. This is the problem, and that sense of urgency is growing to solve it.”
“Everyone is aware of the risks and problems that are coming up with the lack of people who are able to operate a greenhouse. And at the same time, artificial intelligence, technology, algorithms, data, and data tooling have become much more mature in the last three years.”
By incorporating all these factors into an AI greenhouse system, growers can make their company future proof, says Hoek.
It started with autonomous energy management
Hoek starts by explaining daily energy management in Dutch greenhouses. Many greenhouses in the Netherlands have combined heat and power (CHP) installations and are able to trade on the energy markets, bidding on the daily rate, in the hopes of securing the best price for heat and power. And that can be a costly and daily burden.
“In 2014 we introduced within AgroEnergy a system that is a combination of algorithms that does the thinking for the grower,” he explains. “The system takes all the different elements into account when optimizing the energy costs. Important considerations are the lighting strategy, weather conditions for the coming four days, energy prices, levels of the heat buffer, etc. We combine this with approximately 100 greenhouse-specific setpoints. That information is combined with data from a climate computer and translated into the most optimal transaction on our country’s energy markets.”
Through these algorithms, the optimal price is found and the CHP installation is ramped up or down to optimize cost savings. And the positive results are there for all to see when the utility bill arrives.
“You do not have to work with this daily energy management anymore, because the algorithms are not only thinking for you, but acting on your behalf on the energy market and steering installations. And that has proven to be a success.”
To date, 140 to 150 top greenhouses throughout the Netherlands are using autonomous energy management without the necessity of human interference. And a big plus for Dutch growers is that they don’t have to crunch the numbers on a daily basis, “because all the algorithms are doing the job for them,” says Hoek.
Adding autonomous climate management
So, with all this experience in building autonomous energy systems for growers, the logical next step is to let robust algorithms work for the crop as well.
Developed in close cooperation with Dutch vegetable growers, Crop Controller optimizes climate and irrigation and steers the climate computer and all relevant installations in one run. The system analyzes, calculates and steers every 15 minutes. And works without human interference. Of course, the grower remains always in control while benefiting from autonomous climate management.
The next step is to combine the strengths of autonomous energy management with autonomous climate management. Then every automated action is a balanced decision between optimal yield and lowest costs. Lighting is a good example once again: “What is the optimal lighting strategy for my crop with the lowest energy usage and costs?”
Many companies provide data insight tools for growers, supplying them with dashboards, alerts and digital advice. However, growers still need to assess the advice for themselves and perform the correct actions.
“We take it to the next level and offer solutions that go beyond human capacity,” says Hoek. “Algorithms really understand the strategy of the grower and perform actions fully autonomously. Only then will the grower be able to increase the span of control and make the company less dependent on the availability of human expertise.”
The Canadian Market
As Blue Radix markets its technology worldwide, Hoek has a unique perspective on the Canadian greenhouse sector compared to their global counterparts.
“When I’m speaking with Canadians, especially Ontario greenhouse vegetable growers, they really understand the necessity of these developments,” he says, adding that Canadian greenhouse growers operate very much like their Dutch counterparts.
He does see a “prove to me that it works” characteristic among Canadian growers, who want the hard numbers before embracing the technology. But the move to automation is moving forward. “The introduction of AI into irrigation and fertilization, for example, that is already happening,” he says. “Growers are innovating in their operations.”
But where is it all headed?
“In the long run, we won’t only be about climate, energy and irrigation,” says Hoek. “We are also discussing algorithm connections with labour, logistics and marketing; when you connect them in the right and consistent way, it can help a lot in optimizing the operations.”
“I think three big companies in the field of AI will be there in about five years; and we will be one of them.”