Smart Farms: Getting To the Heart of The Matter

August 3, 2021

Take a look behind the scenes of vertical plant automation and find out what it takes to bring a smart farm to life with exciting exploded-view diagrams.

Labour costs account for up to 50% of the market price of vegetables, berries, and herbs. However, just like the tractor that once transformed agriculture, the new technological revolution greatly reduces the need for manual, skilled and specialist labour. Today, the work is done by a team of robots, sensors, artificial intelligence, and a core team of highly skilled professionals who manage hundreds of farms at once using special software. As a result, we witness the “battle for the harvest” moving into the IT domain.

A vertical farm functions as an organism, and its equipment can be compared to human organ systems. Its sensors and senders are the nervous system, its air conditioners and dehumidifiers act as the respiratory system, while the solution supply pipeline represents the circulatory system. We can compare the nutrient mixing unit to the endocrine system, tray conveyors and lifts to muscles. Naturally, this organism has a “brain” — the Growtune IT platform.

The degree to which the brain can accurately and confidently “take control” determines how ready various systems, nodes, and technologies are to operate without human intervention. Below, we describe all of the tech that lies inside a vertical farm. How, where, and at what production stages does the automation work? For this purpose, we’ll leverage the somewhat forgotten “exploded-view diagram” style. This way, we'll be able to show all the devices the farm uses to take care of the plants.

1. Breathing: ventilation and air conditioning

Let’s remember biology classes. Plants use photosynthesis to turn light, water and CO2 into oxygen and energy. By doing so, they can consume far more CO2 than is normally present in the air. Since the farm, like a space station, uses a closed-loop air exchange system, with no air inflow from the street, creating an optimal environment and climate for plants is purely an engineering challenge. This includes purifying, conditioning, and adding some CO2 to the room to allow plants breathe freely and grow intensively.

“People who walk into a vertical farm for the first time notice how nice and clean the air is. That’s no surprise because it doesn’t contain any dust — there’s nowhere dust can appear from in the closed-loop circuit of the farm.

Alexey Novokreshchenko, Head of Automation at iFarm

The farm has multiple air conditioning units. It’s not just a matter of maintaining a certain temperature and humidity. We need to make sure that the climate is as homogeneous as possible throughout the entire arm, including at different heights. So, instead of one super-powerful unit, we install several medium-sized devices. This provides the right air circulation and creates a failure-proof system.

Hydroponic technologies tend to create a lot of moisture, and the farm air typically needs to be dehumidified (iFarm has developed and patented a dehumidification system that allows you to reuse this water to feed your plants). There are times, though, when you need to do the opposite: for example, you may need to humidify the air, especially at the initial stage of starting a farm when you don’t have a lot of plants.

2. Energy: grow lights

Photosynthesis, that is the conversion of light energy into chemical energy, only occurs when there is light. Agronomists call artificial lighting of plants “supplemental lighting”. It may vary from one area of the farm to another, depending on the crop and its cultivation phase. The light control unit creates “day” and “night” conditions for plants. The most frequent setup resembles a nice summer day: 14–16 hours of light and 8–10 hours of darkness. Darkness is crucial: the night is when cellular respiration occurs and other important processes take place.

“The plant depends on the day-and-night schedule. Its structure and maturation phases vary depending on the light cycle. Keep the microgreen sprouts in the dark a little longer, and they'll stretch out but will become coarser. By changing the lighting, you can experiment with taste. For example, if you keep sunflower microgreens in the dark for 10 days and then expose them to the light for a full day, you’ll get long and very sweet sprouts.”

Nikolai Podvigin, Head of the Agrolaboratory at iFarm

Agronomists at pilot farms refine cultivation flow charts. For example, they are currently experimenting with variable brightness to grow pine strawberries: the lamps are equipped with a dimmer so you can gradually add brightness. It seems that the agronomists’ hypothesis holds true — the berry evidently "enjoys" long dawns. Through smooth, incremental changes in lighting, the osmotic pressure in the stem and leaves grows gradually, whereas its surge upon sudden light exposure can have a tissue-damaging effect.

Sometimes you have to focus not only on what the plants need but also on organisational or purely economic factors. For example, the cost of electricity. True, the lamps we manufacture and install in racks are highly energy-efficient: they convert every watt of electricity into 0.6 watts of photosynthetically active radiation (FAR). Still, energy costs make up a large portion of the costs of growing herbs indoors. Automation can (and should) be leveraged to ensure optimal load distribution. If the farm can use a night rate, we change “days” with “nights” and reschedule the supplemental lighting for the nighttime to minimise operating costs.

3. Nutrition: irrigation and nutrient solution supply

The nutrient solution is made from several components stored in individual containers which can easily be replaced. The solution is prepared in the mixing unit, a sophisticated automated system capable of mixing dozens of litres of stock solutions with high precision and in the right proportions.

The nutrient solution is delivered to the roots of the plants through a tube system. Since the specific recipe depends on the type of the plant and its cultivation stage, the system continuously monitors (and sometimes adjusts) the composition of the solution.

“Plants eat and plants drink. We provide them with both by mixing nutrition with water. This water continuously runs through the rack. To create an individual “menu” for each plant type, we divide the farm into different watering areas and install an additional storage tank in each of them. Each area is equipped with an additional set of sensors that dynamically monitor the changes in the solution. In fact, these sensors control whether the plant “ate well” today and whether they have everything they need.”

Alexey Novokreshchenko, Head of Automation at iFarm

The water used for preparing the solution is added to the system after it is properly treated with a reverse osmosis filter. This removes contaminants, such as minerals, heavy metals, bacteria and viruses, leaving the water virtually pure.

4. Well-being: IoT sensors

The purity of the source water, the ratio of nutrients in the solution, its acidity, the brightness of the light, the amount of carbon dioxide in the air, temperature, and humidity — you need to control all these things with precision. A large farm can have hundreds of sensors.

Based on sensor measurements, the farm's central processing unit controls the equipment and hardware in real-time, always keeping climatic and biochemical parameters within the limits set by the cultivation flow chart.

If these limits are exceeded, it triggers an alarm and remote technicians are alerted. The algorithm helps identify potential human errors, among other things. For example, it will immediately alert you if a technician incorrectly replaces the cans with the source materials for the solution.

“Plants can be extremely sensitive, some may “dislike” even a one-degree deviation from the usual temperature. We covered the racks with sensors and cameras so we could understand what a crop is sensitive to during cultivation. Now we’ve stabilised the process by defining the optimal number of devices to ensure accurate environment control.

Nikolai Podvigin, Head of the Agrolaboratory at iFarm

5. Analytics: computer processing of Big Data

The “head” of the system is a cloud server. It collates and computes data indirectly from the array of actual measurements collected. For example, it calculates the efficiency of air circulation across the farm based on the data from all of the climatic sensors. What’s even more important is the ability of Growtune, the farm management software platform, to process all of the Big Data collected on the farm.

“The algorithms we use to prepare nutrient solutions are based both on recipes and on the processing of collected data. The outcomes of plant nutrition in previous cycles affect future mixing. The controller monitors the plants’ reaction to “dinner” on the previous day and adjusts the solution accordingly. And, after reviewing the combined data from different farms, we gradually optimise the flow charts.”

Alexey Novokreshchenko, Head of Automation at iFarm

Farms are increasingly using machine vision in their “nervous system”. Images from the cameras installed on the racks are analysed by a neural network trained to determine if the plants are healthy. Technicians can also use this neural network by uploading a picture of plants through a chatbot to gain critical information about its health. Another neural network calculates the dynamics of herbage mass accumulation using the same stationary video cameras or drone cameras.

6. Movement: automation of tray movements

Problems usually dealt with by rare and expensive specialists, agronomists, biochemists, and phytopathologists can be solved using artificial intelligence, special equipment, or remote consultants. But, until recently, there hasn’t been a fully automatic farm where electronics also handle routine manual work, such as moving trays.

In February 2021, iFarm engineers implemented and tested a design prototype of a fully automated vertical farm. It features a conveyor belt that moves trays with pots through the planting, germination, growth, ripening, and cutting areas. The farm can also be equipped with individual robot modules for packaging products and preparing trays for the next cycle.

7. A human helping hand

As with any sophisticated modern equipment, an automated farm has special service message protocols. Naturally, these protocols are intended for a person, such as an employee of a certified service provider. Automatic warnings about errors, worn-out nodes, or the need for consumables to be replaced translate into assignments for replenishment or troubleshooting.

The reports of deviations in the growth of herbs are handled by an iFarm agronomist. The first thing this expert does is double-check that the actual conditions match the ones in the flow chart. They then go through a list of tasks almost identical to the sections of this article. So, what’s happening with the air, climate, energy, and nutrition of the plants on the farm? Do the sensors provide correct data? When suspecting a technical failure, the agronomist connects a service employee to the diagnostic interface with one click. They can also connect to the technician’s camera or AR glasses if the employee is already on site.

“We have several continuous operation circuits, millions of second-by-second measurements, sensors, robots, conveyors, drones, cameras, and neural networks. It seems we’ve automated almost everything we can. The only thing you can’t automate is the R&D process itself. This is why people in the company focus on this process. And this is exactly why we are fighting for strategic investments.”