Smart Manufacturing

Physics-enhanced Artificial Intelligence for Industrial Process Control

The current manufacturing environment places a growing demand on autonomous control and optimization of manufacturing processes, especially for unattended machines. With the development of modern information technology — and particularly of the new generation of artificial intelligence (AI) technology — new opportunities are available for the development of intelligent machines and related process controllers. Nevertheless, due to the high complexity of the production context, pure AI applictions are usually not enough, and the extrapolation capability of human operator based-on domain knowledge are still necessary.

Chen et al. (2019)

In this context, a new generation of process controllers that integrate physics and domain knowledge with Machine Learning in optimization techniques is investigated. They are based on advanced grey-box models that are continuosly tuned on field data, while preserving the generality and prediction capability of first principle approaches and technologist’s experience. The related research activities include:

  • Grey-box models indentification;
  • Development of Digital Twins of industrial processes;
  • Model-based and data-driven optimization of process parameters.