Prediction of the fouling thermal resistance on the sulphydric acid coolers

Heat exchangers’ fouling causes increased resistance to thermal exchange, with subsequent efficiency loss. Although related analysis has been exposed in previous studies, the available mathematical models do not consider all forms and mechanisms of deposition of unwanted material. This investigation proposed two models for prediction of the fouling thermal resistance in a system of hydrogen sulphide gas coolers under operations. The values for independent and response variables inherent to the process were obtained by applying the passive experimentation method. Correlations of 98,07 % and 97,23 % were achieved from the multivariable regression model (for the tubeside-shellside heat exchange and the shellside-jacket interaction, respectively), as compared to 99,63 and 99,03 % for the artificial neural network. The results confirm the validity of both techniques as reliable forecasting tools, with the neural network being the best predictor