How the s-X-AIPI web app can improve the aluminum sector's operations through AI-driven services
Our s-X-AIPI experts at AIMEN have developed a web application tailored to the aluminum use case being implemented in our research project. This web app is designed to provide AI-driven services to the end-user IDALSA, a company that specializes in the manufacture of aluminum ingots in various formats and chemical compositions from recycled waste and scrap.
The app's web interface is focused on ease of use, enabling interaction with both the AI Data Pipeline and remote external platforms such as the Autonomic Manager and Influx Platform. The Autonomic Manager remotely monitors local AI models at the production plants, offering external AI services to supervise the function of the corresponding AI Data Pipeline. This is made possible by the fact that the plant’s AI Data Pipeline sends metadata to the Autonomic Manager, to maintain and improve the performance of the local AI system. The ongoing data exchange ensures that the Autonomic Manager can dynamically support the functionality and adaptability of the plant's systems, even under a variety of operational conditions.
Currently, all components of the AI Data Pipeline have been deployed in a preliminary version and are actively in use at IDALSA plant. These include two key AI models:
A model for predicting the chemical composition of aluminum mixtures
A model for generating aluminum recipes based on user inputs (such as product type and quantity)
The video highlights how the AI services function, showcasing the process of a user submitting a request, reviewing the visualized results, and offering feedback. This feedback becomes part of the system's metadata, which activates the self-X capabilities supporting the Autonomic Computing features of the pipeline. Additionally, this input is incorporated into a Reinforcement Learning framework that refines the AI Generative model’s recipe suggestions, allowing the system to adapt its responses to better meet user preferences.
Any alarms triggered by the Autonomic Manager or Influx Platform are shown on the "Notifications" page, as demonstrated in the final moments of the video.
