Ensuring data readiness for Infrabel Asset Management
Since 2012, Kapernikov has been ensuring the readiness of all master data used in Infrabel’s SAP E ...
Challenge: 20.000 references with different shapes
The Bilzen distribution center is responsible for processing and shipping windshields in a way that ensures they arrive intact at various service locations. This is no simple feat when dealing with a multitude of windshield models, each with distinct dimensions and curvatures.
The company’s own custom carriers, or “supplainers” as they call it, are designed to hold up to 20 windshields. However, a lot of expertise is needed as operators need to know precisely how to position the glass to prevent scratches or breakage during transport.
Despite their best efforts, placement choices made by the warehouse operators sometimes lead to errors, repeated adjustments and thus inefficient use of labor. The variation in product shape and characteristics yield a learning process for new employees taking months to years.
Carglass needed to find a way to bundle its best practices into a data-driven, automated solution: a digital agent for operator assistance.
Our role: a digital agent using 3D scans of every windshield
Kapernikov’s approach began with understanding the intricate process operators followed when handling windshields. The solution was clear: create a system capable of identifying and guiding the placement of windshields using 3D data. This required a robust machine vision system that could accurately measure the shape of glass, an inherently challenging task due to its transparency.
Key components of the solution:
- 3D scanning system: Kapernikov partnered with Keyence to implement a scanning system capable of capturing detailed 3D images of windshields using laser triangulation technology.
- Data optimization: A numerical mesh was applied to fill gaps in the 3D data, ensuring a complete representation of each windshield’s shape. Thanks to this capability, scanning could start before the optical setup was tuned to perfection. And future shapes don’t require optical adaptations.
- Database of 3D models: A centralized database is created to store the 3D shapes and associated parameters, forming the basis for automated decision-making on windshield placement.
- Ranking system: By digitally comparing windshields, Kapernikov developed a ranking algorithm that categorizes the different windshield types based on their physical properties. This system provides order pickers with optimal packing instructions.
The combination of advanced scanning, data processing, and intelligent ranking is transforming how Carglass handles their logistics, eliminating the guesswork involved in packing.
The result: A solution validated in the field
Carglass masters continuous improvement and takes validation of this solution seriously.
- Field tests in the logistics hub in Bilzen show that the digital instructions lead to the desired reduction in number of actions to fill a carrier.
- The data driven digital agent has the knowledge of senior order pickers and thus reduces the time to train younger operators. The experience of Carglass with previous improvement projects learns that a reduction in the numbers of breakage and customer complaints will follow.
The solution’s success in Bilzen sets the stage for a broader rollout across Belron’s global network. This innovative approach positions Carglass for sustained operational excellence.
Kapernikov integrates data & ai technology in a project for an industry leader, while coordinating hardware vendors and machine builders to reach the goal.