Cleaning Schedule Optimization of Heat Exchangers with Fouling on Tube and Shell Sides: A Metaheuristic Approach

Abstract

Pre-heat trains (PHTs) significantly reduce refinery fuel consumption and carbon emissions. However, these benefits are diminished by fouling in heat exchangers (HEXs). Current methods for optimizing cleaning schedules often report high computation times due to the transient nature of the fouling process and do not consider shell-side fouling, which can be significant for some oil fractions. This paper addresses these issues by adding shell-side fouling to the model and by transforming cleaning time variables into integers, reducing the problem of optimizing cleaning schedules to an integer nonlinear programming (INLP) problem. The reformulated problem is solved using integer particle swarm optimization (PSO) coupled with a simple search strategy, where the number of cleaning actions is preset and their timing is optimized. The adopted approach achieved up to 84% lower computation times compared to previous ones. Additionally, the relationship between cleaning actions and PHT performance is nonlinear, with diminishing returns from additional cleaning, and optimal cleaning schedules are often asymmetric for different HEXs within the same PHT. The proposed approach effectively reduces operating costs and provides a framework for future optimization enhancements.