In the fruit sector, increasing attention is being paid to the presentation of products for example apples and pears are packed so that their stalks all point in the same direction. Apples are also placed in trays so that their best, blushing side is facing upwards. Stas' new machine fully automates the aesthetic filling of such packaging.
Stas is family business based in Glabbeek, Belgium, that has been building and selling machinery for the fruit sector since 1986. "We started with the construction of rotary sorting tables," says Managing Director Kristof Stas. "Then came sorting machines with receding belts that still sorted fruit mechanically. In the latest generation of machines, this is done electronically where each piece of fruit is weighed and an output on a sorter is controlled accordingly. In recent years, various aspects of quality control have been added, such as measuring colour and monitoring the external and internal quality of the fruit."
After checking and sorting the fruit, packaging follows, which today is still often done manually, and because producers wants to achieve the best possible presentation. With the new machine, Stas has succeeded in automating that process too. Central to this is a vision system with artificial intelligence that looks at each piece of fruit from all sides to assess how it should be packed. Moreover, Stas wanted to develop a universal machine that could be used for various types of fruit.
"It was fairly quickly clear that this was going to be a complex and extensive project," says Kristof Stas. "We therefore started looking for a partner who had the know-how to co-develop all aspects of the machine and who was willing to think along with us, with the ambition of obtaining the best possible end product. We found that partner in Festo."
Jürgen Seymortier, Sales Engineer at Festo: "Thanks to Stas' R&D efforts and mindset, we can make suggestions and optimisations to achieve a good result. Fruit is difficult to sort and not all tests go perfectly from the first run. It requires a lot of resources and patience on the part of the client to take a step back sometimes to make a better jump. Thanks to Stas' long-term vision and realism, this led to a great result."
At Festo, Jurgen Seymortier took the project to heart. For the artificial intelligence and control programming, Raf De Vos was called upon, who gained a lot of experience as a programmer in the financial sector with fast, big data applications - experience that came in handy when integrating image processing and machine control.
At the centre of the packing machine is a unit developed specifically for this application called a tumbler. It is a tub with two conveyor belts at the bottom that allow a piece of fruit to be tilted and rotated. Inside the machine are eight of these tumblers, with cameras above them whose images are sent to a neural network. That searches the images for the position of the stalk. For apples, it can also determine at that moment where the red colour is that is desirable to have visible at the top in the packaging.
On the other side of the machine is a second vision system with a 3D camera that scans the packing trays. These are pre-formed trays in which cavities are provided to deposit the fruit. The vision system determines the positions and orientations of these cavities and can also monitor where the empty places are.
Based on all that data, a Festo PLC controls the processing system. This consists of a bridge that moves between the tumblers and the packing tray, with four heads on top that can take one piece of fruit each. Each head sits on a linear axis that is transverse to the bridge and can fill four locations in a pack sheet if needed, simultaneously. The vertical movement of each head is achieved using a telescope with both a pneumatic and electric carriage with the former providing speed of movement and the latter precision. Each head can also rotate to place the fruit in the right direction.
"For such a machine, it is important to achieve a high capacity," explains Kristof Stas. "A lot of design choices are aimed at making that possible. The tumblers, for example, not only allow us to see the fruit from all sides but are also used to orientate each piece correctly so that fewer movements are needed in subsequent handling, which increases speed. The choice of a telescopic shaft was also made to achieve higher speed."
Many additions have also been made to make the machine absolutely reliable, which again is crucial for maximum capacity. For example, there are lifts between the infeed conveyor and the tumblers that guarantee that only one piece of fruit is taken each time. A specially written algorithm chooses the optimal positions for depositing the fruit with minimum movements. This takes into account that the heads on the bridge can move in relation to each other but cannot cross each other.
"What is unique about the machine is its flexibility to pack different types of fruit," says Kristof Stas. "For the operator, there is a touch panel on which they can indicate what type of fruit is involved and whether orientation and colour should be taken into account when packing. Everything else is fully automatic, including the detection of the pack dimensions and the shape of the pack sheets." This way, the machine is more flexible and can also look at more than just fruit, such as packing peppers, for example.
In Festo, Stas found a partner who was involved throughout the design process and in the construction of the first prototype to translate the many concepts in the machine into efficient drive technologies. Moreover, together with the hardware, the software was also developed so that optimum choices could be made in terms of mechatronics.
This was not the first collaboration with Festo for Stas as the two previously worked together to develop a gantry robot that can empty large fruit bins into a water tank. Water transport is often used in the sector to avoid damaging fruit. To make the robot also suitable for working with pears, some modifications were made . Pears - unlike apples - sink in water. Therefore, a moving bottom was developed for the water tank at the start of transport, which can be pushed upwards with cylinders to bring the pears on the surface into the water stream towards the conveyor belt. In the process, the movement of the cylinders is perfectly synchronised to keep the bottom flat.
With the new packing machine, a complete line can now be built that automates the whole process of handling, sorting and packing. "In the fruit sector, many companies face staff shortages," says Kristof Stas. "With our automation solutions, we can support this."