Ophardt seeks to enhance the production of soap dispensers by means of automation

Specialist in hygiene products Ophardt Belgien investigated whether it would be possible to optimise part of their soap dispenser production, while also simplifying the tasks of the operators. The idea was to do this by introducing a cobot and vision system.

The HORSE project (Smart integrated Robotics system for SMEs controlled by Internet of Things based on dynamic manufacturing processes) intends to bring progress to the production industry by proposing a new flexible model of a smart factory which incorporates the collaboration of people, robots, AGVs and machines to efficiently perform industrial tasks. Within this project several demonstrators must be built in cooperation with SMEs. Sirris helped build such a demonstrator under the name RANCH (Robotics And Neural networks Combined in HORSE), a collective project together with imec, UGent and Ophardt Belgien, a subdivision of the international Ophardt Hygiene, based in Maaseik.

Production of soap dispensers

Ophardt Hygiene is a global specialist in quality hand hygiene solutions. This manufacturing company wanted to investigate whether it could optimise part of its soap dispenser production for industrial and medical use. To this end they decided to introduce the use of cobots, while also relieving the work of operators. The quality control was also simplified by means of a vision system.

The most important part of the soap dispensers is the holder. It is made of an aluminium profile that is sawn, milled and folded, and finished by means of anodisation. For this purpose the holders hang on a rack and are immersed in an anodisation bath. After this treatment, they are removed from the rack and undergo quality control before they are packaged. Two operators stand at the end of the anodisation line: one removes the holders from the rack, the other inspects them for production errors and places them in a box. These tasks are repetitive, ergonomically stressful and require constant concentration, which is mentally stressful. The quality of the inspection also depends on the operator performing the task.


In the Sirris and imec labs the project partners developed a set-up with a cobot that can remove the holders from the rack. For this purpose, the cobot is equipped with proximity sensors to locate the holders. An additional challenge is that the holders are sometimes stuck to the rack due to the anodisation process. The cobot must therefore be able to feel how much force is needed to loosen them.

After removing them from the rack, the cobot puts the holders on a conveyor belt that takes them to the point of quality control, where a second cobot picks them up and holds them in front of a vision system (developed by UGent) that takes pictures of each side. The photos are analysed automatically. An operator examines the holders for a final inspection and can also see the photos and analysis on his screen.

Promising test results

The system to remove the holders from the rack, was 100 percent successful, the automatic quality control achieved a precision of 97 percent on a separate test batch. The ergonomic load on the operator was improved by adding a cobot to keep the holders out of the rack. In addition, the consistency of the quality inspections was increased by adding the vision system to support the operators.

For a cost-efficient implementation in Ophardt’s production, the process needs to be optimised to achieve better cycle times. For better automated quality control, there is a need for a larger dataset collection, so that the system is trained to recognize all types of errors. The quality of the photos is also very important.