Analysis of fouling rate of crude oils in the literature: experimental data, models fit and parameter estimation

Abstract

An important problem in petroleum refineries is the resistance to thermal exchange, due to the fouling in heat exchangers for heating of crude oil, typically described by threshold models. These models involve parameters that must be estimated from laboratory or refinery data. Based on this, some questions need to be asked: from typical experimental data, is it possible to determinate all the parameters simultaneously? On the contrary is necessary to maintain constant a set of parameters for which no error can be attributed; If the previous answer is positive, is it possible to state which of these parameters best fit the experimental data? In this context, the objective of this dissertation is: (1) To evaluate the identifiability of threshold models parameters making use of experimental data from the literature. For this step, a preliminary study was made in order to obtain experimental errors based on data provided by some authors in the literature; (2) Check among the selected models if it is possible to identify which ones canfit the experimental data. For the first step using the experimental errors obtained in the preliminary study, relative errors of the rate of deposition for a single run are below 5%. The compensation effect between pre-exponential factor and activation energy is confounded with statistical error in the most cases. The parameters estimation methodology lead to similar results. In the second part, the results have shown that, despite most of the models were able to correlate the trend of experimental and calculated values, the experimental error obatined for pseudo-experimental data is quite low. Therefore, few models were able to predict data within this precision.