Process Controllers


A University of Melbourne breakthrough has developed a new approach that allows plant operators to put more trust in a wide range of unit controllers (such as for thickeners). It combines expertise in physical and chemical processing models with control theory and machine learning to great effect.

Statistical machine learning is used to estimate the parameters of controllers, based on self-learning that optimises both unit operations and operations across the system. Artificial intelligence techniques are used for failure detection and improving the set points across different operating regimes and shifts. This improves the trust of the operators in the virtual plant and controller, resulting in a more strategic, proactive and consistent operating environment to better meet production objectives.