Machine learning models for fouling factor prediction in heat exchanger networks

This work will involve development of time-series (recurrent) machine learning models for predicting fouling factors in heat exchangers in a heat-exchanger network. In particular, the aim will be to predict future evolution of fouling factor given its past trajectories and trajectories of flowrates and temperatures of both shell and tube side fluids in a heat exchanger network. The work will be part of a live industry sponsored project and will involve working with real-industrial (refinery) data along with a project staff.