Fehrer optimises manufacturing with Visual Inspection AI
Workforce productivity increased
Experts focus on other tasks that are more rewarding
Manufacturing processes streamlined
Data is a factor of production in the manufacturing process
Product quality optimised
Easier and more accurate quality control
Challenge
Fehrer is one of the world’s leading producers of components for vehicle interiors and a key partner of many top automotive manufacturers. With a proud heritage dating from 1875, the company has committed to remain at the forefront of innovation and manufacturing excellence.
Delivering high-quality vehicle components is critical to the company’s success. A key element of this is to spot potential defects and anomalies quickly and accurately. Although this is achievable manually, Fehrer could see the potential benefits of automation through Visual Inspection AI.
The strategic aim of the project is to partially automate the quality inspection process of the produced foam parts to improve defect detection rates, raise overall quality and generate a cycle of continual improvement through quality metrics. As well as reducing the “escape rate” of defective components, expert inspectors might be redeployed to value adding processes. With the right approach, the project promised to generate qualitative and quantitative benefits.
Fehrer needed practical help to develop a solution that could deliver all the benefits on offer and become part of its applied AI strategy. Having delivered several successful projects in collaboration with Google, GFT had the right track record. As an experienced Google Cloud data and AI partner, GFT was also very well versed with Google Cloud Platform (GCP) methods, tools and technologies and could deliver a solution quickly.
GFT was chosen to design, deliver and implement a proof-of-concept (PoC) solution. In addition to its practical expertise, GFT was perceived to have the right service culture and delivery capabilities using its unique mix of local and nearshore locations.
“From the outset we felt that GFT was a company we could work with. The project had clear objectives and we all understood the transformational potential of AI within manufacturing.”
Engagement
Being a world leader in the manufacturing of car components, Fehrer strives continually to drive process improvements. The company wished to prove that an AI-powered quality inspection process could deliver on its promise to improve both manufacturing quality and efficiency. A successful solution would also harness the power of data to reduce/eliminate false positives as well as false negatives, that is defective products being delivered to the customers. This would improve overall customer quality ratings.
Although Fehrer regarded this project as part of its strategic AI initiative, the partners decided to begin with foam pads for car seats, which are simple in their design/ construction, with inserted clips and the subsequent inspection of clips in the foam. The key inspection criteria were to ascertain the presence of the clips and whether they were defective, that is completely missing or misaligned, or not in a quick and precise manner. An escape rate of 0% was the target to be achieved.
While the criteria were easily defined, practical inspection is always very difficult, as defects can be hard to detect or distinguish from perfect products. Moreover, many fault types and a diverse range of misalignments can occur over dozens of production lines, foam part variations and insertion parts.
To accelerate progress, full use was made of Google Cloud Platform tooling, including VIAI and Vertex AI. This also meant that the project adopted best practice from the outset and could be easily scaled and redeployed across the company as a strategic AI solution.
The “eyes” of the solution comprise cameras on the production line as part of an Edge deployment at the Fehrer data centre. Edge is essential to deliver results in just a few seconds. In technical terms it is a truly hybrid solution that embraces modern methods and the best of cloud and on-premise processing.
When moved into a production environment, the solution will be highly intuitive to use with key metrics available on dashboards and a “traffic light” system showing inspection status. It has been rigorously tested, to prove its viability beyond doubt and to show that it offers a genuine improvement over manual inspection.
Benefit
The solution was delivered within three months. Throughout the project, the partners worked shoulder to shoulder so issues could be resolved quickly and decisively. This culture of total transparency enabled GFT to work as an extension of Fehrer’s own operation and ensured the solution met all stakeholders’ expectations.
Visual Inspection AI marks a new way of working for Fehrer. Like many manufacturing companies it relies on a dedicated, skilled workforce to deliver quality products to satisfied customers. With its new solution, some of these experts can focus on other tasks so Fehrer can attract and retain the best people with rewarding work.
Building on the success of the PoC, the partners are scaling the solution horizontally and evaluating new deployments. As well as offering a step-change improvement in visual inspection, the new solution will empower Fehrer to harness the data as a factor of production in its manufacturing process. Potential applications include:
- Root cause identification
- Anomaly detection
- Predictive maintenance
- Digital twins
- In-line quality control
Fehrer always viewed this assignment as part of a bigger story. The solution puts data at the heart of its manufacturing process and the company can tighten control and monitor quality more precisely. Manufacturing output can be aligned exactly with market requirements and waste reduced.
This assignment is a “lighthouse” project for Fehrer and Visual Inspection AI could be deployed across many more production lines. The success of the project shows how digital technologies enable companies to collaborate and deliver innovative transformational products with minimal speculative investment and at no risk.
“The PoC delivered robust evidence that AI-powered inspection is the way ahead in manufacturing. With this solution we can improve overall quality and lift our efficiency potential.”
Download - just one click away
-
Success story
Fehrer optimises manufacturing with Visual Inspection AI
