Work Roll Cooling System Design Optimisation in Presence of Uncertainty

The paper presents a framework to optimise the design of work roll based on the cooling performance. The framework develops Meta models from a set of Finite Element Analysis (FEA) of the roll cooling. A design of experiment technique is used to identify the FEA runs. The research also identifies sources of uncertainties in the design process. A robust evolutionary multi-objective algorithm is applied to the design optimisation I order to identify a set of good solutions in the presence of uncertainties both in the decision and objective spaces.