(566c) Evaluation of the Impact of Uncertainties of Measurement of Process Parameters in the Calculation of Incrustation Resistance in Crude Oil Heating Heat Exchangers

An important phenomenon in the operation of heat exchangers is fouling, which is characterized by the sedimentation of solid material on the surface in the walls or tubes of such equipment. The deposition can have three origins: crystallization, chemical reaction or biodeposition. It is an important phenomenon since the deposited solid material can restrict the cross-sectional area for fluid flow, causing increased pressure drop and resistance to heat transfer, limiting the performance of the equipment in energy terms and increasing energy costs and cleaning.

Fouling is a particularly complex phenomenon when it occurs in crude oil pre-heating (BPA) batteries, which have the function of supplying the oil stream, which feeds the refineries at a temperature level which is sufficient to maintain the efficiency of the atmospheric distillation processes and under vacuum. The complexity is due to the multiplicity of phenomena involved that influence and are influenced by deposition, such as activation energy of chemical reactions, solubility of asphaltenes, presence of particulates, flow velocity, among others. These parameters affect fouling, making any mathematical model applicable only to very restricted conditions, which in many cases do not represent the operational conditions of a refinery.

Nevertheless, many efforts have been made to construct mathematical models capable of being widely applied in predicting the dynamic behavior of deposition in GAPs. The great difficulty factor of the construction of these models consists of the multiplicity of phenomena that act in a synergistic way in the generation of the deposition, so that the complexity of the phenomenon is not usually fully represented by the contour conditions of the models. Thus, many of these models show deviations from the operational data obtained in industrial units. However, the analysis of the pertinence of the models is generally carried out without considering the uncertainty of the data obtained in the industrial plants, which can make difficult a more accurate assessment of the degree of approximation of reality.

This work presents a methodology to estimate the thermal resistance (R_f) measurement errors in refinery oil preheating batteries. For this, mathematical correlations were developed to estimate the error in the R_f calculation based on the temperature and flow measurement uncertainty reported by the instrument manufacturers. The methodology was implemented by calculating the error range of R_f calculations performed from available data from oil refineries. As a result, relative errors of R_f of the order of 50% were observed, indicating that some mathematical models may be considered valid if uncertainties in the measurement of operational parameters are taken into account.