AI for increased process effectiveness
Reference number | |
Coordinator | RISE Research Institutes of Sweden AB - RISE RESEARCH INSTITUTES OF SWEDEN AB, VÄSTERÅS |
Funding from Vinnova | SEK 4 559 553 |
Project duration | February 2021 - February 2023 |
Status | Completed |
Venture | Strategic innovation programme for process industrial IT and automation – PiiA |
Call | PiiA: Data analysis in industrial value chains, autumn 2020 |
Important results from the project
The goal is to demonstrate AI-based quality inspection and process control in the production at Nilar and Sura, that share the need to automate inspection routines (lot of manual work required). AI is used on existing process data in real time from production flows to detect quality deficiencies in product/process. The solution enables rapid data analysis of many images to identify surface defects as well as the monitoring of the fraction of defects in the process.
Expected long term effects
Results include (i) object detection tool for assembling process of Nilar batteries, (ii) automated detection and defect marking tool for Sura´s electrical steel sheets for electric vehicles, and (iii) ML tools for detecting deviations in batteries. They enable surface defects to be identified more efficiently and earlier in the process aiming to reduce material waste by >30% and manual work by >50% in the inspection process step, as well as increasing exploitation of available information.
Approach and implementation
Project aimed to demonstrate AI-based process control for automatic quality control in Nilar and Sura production. The project has demonstrated to Surahammar the usefulness of AI-based process control throughout their production processes and product quality control. Nilar has integrated AI solutions to identify defects on photos during assembling battery modules. With the enhanced AI-based vison system, Nilar can now remove the worst defects from being included in the modules.