Counteract defects with good forecasts
How can artificial intelligence help to identify quality deviations in the production process at an early stage? Machine learning methods and the linking of machine and quality assurance data make this possible.
Where do quality deviations occur in the production process?
For this project, we looked at the production process of a company that generates large amounts of machine data, which is stored in a software system. However, the results of quality assurance are recorded separately. There is currently no direct link between the production process and subsequent quality checks.
As a result, quality defects are only noticed during quality assurance – often up to an hour after the actual error occurred in the process. During this time, defective production continues, resulting in scrap and rework.
To improve this, the key challenge is to identify which process settings are responsible for quality deviations. AI models are used to identify these correlations so that the company can react early and avoid errors as far as possible.
AI model can provide solution
The aim of the project is to systematically evaluate production data and develop a prototype AI model that can predict quality defects at an early stage.
Specifically, the following questions, among others, need to be answered:
- Which process settings have a particularly strong influence on certain quality defects?
- At what data thresholds can poor quality be expected?
- Can an AI model be used to make reliable quality predictions?
The prototype AI model can serve as the basis for an early warning system. This allows process variables to be proactively adjusted before errors occur, reducing scrap and rework.
Implementation with company data
In the project, we combine existing machine data with the results of quality control and prepare it into a uniform database. On this basis, statistical evaluations and machine learning methods are used to identify relevant correlations between process settings and quality characteristics.
Building on this, we then develop an AI model that can predict quality deviations during ongoing production. The project clearly demonstrates how existing digital production systems can be further developed with the help of AI and used for predictive quality assurance.
Your DAISEC contact
If you have any questions or would like to find out more about how an AI strategy can help your company or organisation, please contact us.
Expert in AI technologies in manufacturing
