Fleet control
FAIR 1 Archive Notice
This page describes research activities and results from the FAIR 2019-2023 program period.
The content is retained for reference purposes and is no longer actively maintained. Information on this page, including research outputs, team members, contact details, collaborations and links, may be incomplete or outdated.
No direct successor use case has been defined within the FAIR 2024-2028 program. Please consult the current FAIR use cases for related research activities.
Use Case Description
The use case demonstrates how can AI methods be used in control of (fleet of) machines to make them intelligent, leaning from each other and adaptive to the process’s changes.
Results & Achievements
In 2021, we developed an AI enabled hot-start for the commissioning of controllers for machines with slider-cranks performing new machine-task combinations. It outperforms the industrial practice of experienced operators performing smart copy-paste, obtaining initial guesses performing 50% better, and requiring 25% less fine-tuning time.
In 2022, we continued exploring reinforcement learning (RL) algorithm to do closed loop control with the mechanisms driving compressors. In simulations with varying consumption patterns, it learns control actions 10% better than industrial rule-based controllers, and on par with Model Predictive Controllers (MPC), but without their need for accurate models and online optimization.
In 2023, we developed a set-up using compressors as a representative use case for AI-based demand-driven control while dealing with dynamics user’s profiles. Real data were used to develop robust learning control algorithms but also improving simulations to make them more accurate.
Use Case Information
Period
2019-2023
Contact
Bey Abdel
Research Group
Flanders Make, VUB (Jan Helsen)