realtime volume scanning & measuring
About
Company
AgristoLocation
Harelbeke, Nazareth, Wielsbeke (B) & Tilburg (NL)Competences
computer vision
digital twin
capture
dual
product selection
volume scanning & measurement
production metrics
food processing industry
shifting from 1 daily manual check to continuous realtime measurement
relevance
Since 1986 Agristo has developed into a world player for the development and production of frozen potato products. As a Belgian company with our headquarters in Harelbeke and more production units in Nazareth, Wielsbeke, Tilburg (The Netherlands) they specialize in offering an extremely wide range of fries and derived potato products, made to measure for each client.
When you produce more than 700.000 tons of potatoes per year, you don't want the supply of potatoes to your production lines to come to a standstill.
Being continuously informed of what is happening on the work floor and having a guarantee that all machines have the correct supply is crucial for the continuity of the factory.
"Realtime volume measurement is a real game changer in the daily operations of Agristo. At each moment we know exactly how much volume of potato's there're available and can now avoid stand stills."
Ruben Devos
Process Innovation Manager
Challenge
Agristo wanted to switch from periodic and manual checks by an operator to realtime measurements of all potato bunkers on the production site. Immediately vintecc proposed computer vision as the solution but that wasn't really the challenge.
The challenge was rather focussed on the sensors: how many were needed, where did they have to be positioned for optimal volume scanning and how durable do they have to be, specifically for this kind of harsh environment.
solution
1 LiDAR sensors for volume scanning
- Robust LiDAR sensors, normally used for ADAS-systems, where selected that could withstand the shocks and vibrations that come with these environments without the need of recalibration.
- These sensors provide the desired high resolution data that fit the next generation of smart factory.
2 digital twin for simulation
- A complete digital twin of the bunker environment was created and quickly set-up.
- Adding the LiDAR sensors in this environment - including their scanning area and 'reach' - an ideal coverage path and set-up could be determined by virtual simulation.
3 realtime volume measuring
- The high resolution scanning data was centralized in capture - the industrial IoT-platform of vintecc - and was transformed into volume data. This volume data was then transferred tot the internal production systems of Agristo for further integration.
results
1 no budget overspend in sensors
- Being able to virtual simulate the set-up of the LiDAR-sensors in advance allowed Agristo to select the correct type and the ideal number of LiDAR sensors.
2 continuos stock & volume analytics
- Due to the integration of computer vision technology the operator can know at all times how much potato's there are available in the bunkers.
3 avoid production downtime
- Knowing the volumes in each bunker in realtime, the change to come to a standstill in the production line is minimized.
4 well informed bunker selection for production
- To avoid production downtime and to optimize the production outcome, operators can see and objectively decide which potato bunker they will select for supplying the production lines.