The project was done during the second block of SMR in Summer 2023.
During the Minor Smart Manufacturing and Robotics we were asked to build a prototype version of an automated scooping system for the Startup Foodmills in Rotterdam. The startup plans to produce localised pasta in so called micro factories which are inside of a 20 feet cargo shipping container. Their vision is to increase the biodiversity in the agriculture by just combining the different vegetables and ingredients into unique pasta. So with our prototype we not only built a proof of concept, we also planned to improve their current production area. Since it is still a startup, they produce only at small scale.
Our task was to scoop and weigh different ingredients which mostly behave like flour but can be different regarding their granularity and moisture. Therefore, we needed to find a solution which can dose automatically while getting direct feedback from the scale to stop the process if the needed amount is reached. We also got a space limitation which is an area of roughly 1 cubic meter to make the implementation into the future micro factories easily. There was also the need for an HMI where the owner can setup different recipes depending on the available resources. Also, this HMI shows error messages etc. The whole setup also needs to be cleaned very quickly and easily so we had to keep this in mind for our design choices. Also, the setup needs to scoop a minimum of 6,5kg of dry ingredients per hour because then it will be mixed with around 3,5 Liters of water. Foodmills then ends up producing 10kg pasta per hour 24/7
The benefits of our solution are that we did not use vibration at all in our system which is usually very loud and noisy as well as using air pressure to transport our ingredients. To be able to do this one would need a fully closed system and we were not able to do this at our workspace. The same reason applies to our used materials. We know that the system is not food safe now because we did not use stainless steel, but it is only a prototype proof of concept, so the functionality is the focus. We are sure that a switch from 3D printed parts with their rough surfaces to stainless steel with smooth surfaces will only increase the reliability and functionality of this system. The system can fit 10 hoppers overall however, because of the small timeframe we had (7 weeks) we only build 5 hoppers to show how the system works.
How it works:
At the HMI one can select and configure up to 10 different recipes and specify the amount of each individual ingredient for each recipe. These values are percentage values because then the system only needs the final weight of the recipe batch one wants to have and calculates the rest. Afterwards the PLC rotates the needed hoppers one by one on top of the collector bin and extrudes the needed amount. The collector bin is placed on top of a scale which communicates to the PLC through a RS232 cable to stop the dosing at the right time. We use only one motor which is pushed forward by a pneumatic piston and with a coupling connects to the hopper in front of it. Basically, we only need 2 stepper motors overall for the whole system. One motor is for the dosing and the second one is for the rotating table all the hoppers are placed on.
Major decisions during the project:
Nobody of us worked with food ingredients before from an industrial perspective so first we needed to do some research about what mechanisms the food industry uses if they work with flour or other materials which are more powder like. Based on our goal to not use air pressure or vibration we stumbled upon the principle of an Archimedes screw used in the industry to dose powder. Immediately we tested it with flour and it worked but the flour got stuck on the surfaces of the 3D-printed hopper and it built up an arch above the screw so after a few seconds no ingredients were falling into the screw to be then extruded. We fixed this major problem by adding another wheel on top of the Archimedes screw to scoop and push the ingredients into the Archimedes screw. Therefore, we finally had a hopper design which worked reliable and is independent of air or vibration which is commonly used in the industry. The next major design decision was our idea to build a rotating table to place the hoppers in. The benefits are that we are still using less than 1 m2 ground space for all the 10 hoppers and we can have 1 central point to collect all the ingredients needed for the recipes. In addition to that we now only need 2 motors instead of 1 motor for each hopper. This decreases the amount of wiring and controlling with the PLC as well as the complexity of the whole system. And this was something we wanted to achieve because of the limited amount of time we had.