Experiment name: RANCH
Manufacturing end-user: Ophardt Belgien
The Business Sector
The use case focuses on the manufacturing of dispensers for industrial and medical skin protection for the common sanitary use with a special range for soap and disinfectant dispensers.
| Ophardt Belgien bvba is part of Ophardt Hygiene, an international company that specializes in all kinds of products that deal with hygiene. Their assortment includes hygiene-technical dispensers for hospitals, industry, households and public sanitary facilities. These include soap and disinfectant dispensers, paper dispensers, waste bins and waste baskets, combination dispensers for soap and paper, toilet paper dispensers and many other products that complete their product portfolio. Naturally, high-quality materials such as stainless steel, aluminium and plastic are used for this. Ophardt also produces plastic refill bottles and dispenser bottles made of polyethylene or polypropylene. Ophardt Belgien is engaged in the production of many of these products (e.g. soap dispensers, trash cans, tissue dispensers, etc.). Anodizing also takes place in Belgium in the Maaseik site. |
 |
The Challenge
The most important part of a soap dispenser is the shroud. The shroud is produced starting from a long aluminium profile that is sawn, milled and folded. After the shroud is finished, a surface treatment is done in the anodizing line.
There are two operators working at the end of the anodizing line. One operator racks down the aluminium shrouds and checks the quality of some shrouds. The other operator checks the rest and packs all the shrouds in a box.
The end of the anodizing line is also the most stressing for the operators. First, deracking the shrouds entails grasping the objects above head height and below hip height. Second, there is a quite high variance in judgement between operators when inspecting the shrouds, which in addition puts mental stress on the operators, as this might lead to discussions.
The solution:
To automate the workflow, we replaced the table with a conveyor setup. For deracking, the operator has to correctly position the rack in front of a Kuka iiwa cobot. Then, this cobot deracks shroud by shroud and put those onto the conveyor belt. The conveyor then transports the shrouds to a UR3 cobot. The UR3 then picks up each shroud and manipulates them in front of a machine vision camera, in order to shoot pictures of the shroud side and top surfaces. These pictures are then analyzed by a neural network to detect any defects. Meanwhile, the UR3 puts the shroud on a second conveyor, where the shrouds are put in a queue to be checked by an operator. The operator picks a shroud from the conveyor, and on a screen he can see the pictures of the shroud together with the neural network decision, checks the shroud for defects and records his definitive decision into the system. Finally the shroud is packed for delivery (or thrown away in case of serious defects).
Partners:
| Logo | Website link |
 |
https://www.imec-int.com |
 |
https://www.sirris.be/ |
 |
https://www.ophardt.com |
