Optimizing heat exchanger efficiency: Predictive modeling to minimize fouling in crude oil refining

Fouling in heat exchangers significantly compromises energy efficiency in crude oil refining, leading to increased operational costs and environmental impacts. This study presents a predictive model aimed at enhancing heat exchanger performance by minimizing fouling resistance. Model fitting was conducted using approaches of varying complexity, with measures taken to avoid overfitting. The models were subsequently refined to incorporate key variables, such as inlet and outlet temperatures and mass flow rates, ensuring robustness and generalizability. The final simplified model comprises only 19 terms, yet achieved high predictive performance ( R 2 = 0.961; predicted R 2 = 0.956) and effectively addressed multicollinearity. The selected model identified significant linear, quadratic, and interaction effects among thermal and flow parameters, with the mass flow rates of the hot fluid (MFH) and cold fluid (MFC) emerging as particularly influential. Notably, the model demonstrated that fouling resistance decreases substantially with increasing hot fluid flow rate. Optimization using a desirability function identified 37 parameter combinations that achieved a fouling resistance (RFC) of 0.001 m 2 °C/W with a maximum desirability score of 1.00, consistently favoring high MFH values (~93.00 kg/s) and hot fluid outlet temperatures (THO) near 43.00°C. These findings confirm the model’s robustness and practical applicability, providing actionable insights for operational strategies aimed at minimizing fouling while maintaining thermal efficiency.