1. Introduction
Fuels are still the main source of global energy generation. With the advancement and growth of industry, commerce, and population, there is an increasing demand for energy. The need for energy requires an increase in its production. To achieve this, it is necessary to refine crude oil more and more, as well as to consequently demand the unit operations involved, including heat transfer in the preheating battery exchangers (BPA).
The heat exchangers of the BPA reuse the heat from the distilled fractions of the atmospheric and vacuum tower petroleum output to heat the incoming crude oil and maintain the thermal efficiency of the processes. According to [
1], in only the petroleum preheating battery, the amount of energy consumed can reach 35–45% of the total energy consumption of a refinery. Thus, this reuse can lead to considerable savings. An important phenomenon in heat exchanger operation is deposition (fouling), which is manifested by the deposition of material on equipment surfaces. This deposition can occur as a result of the phase change that arises from the temperature differences between the surface and the fluid (deposition by crystallization), by chemical reactions on the surfaces (deposition by chemical reaction), or even by the growth of organisms on the surface (biodeposition).
It is an important phenomenon because the deposited material can restrict the cross-sectional area for fluid flow, causing an increase in pressure loss. This material primarily acts as a resistance to heat transfer, thus limiting heat recovery and increasing energy and cleaning costs [
2].
There is a consensus that deposition by chemical reaction is the most frequent among the phenomena, as it is triggered by the high temperatures of the process. Naturally, the reactive constituents of petroleum can undergo thermal decomposition, polymerization, self-oxidation reactions, precipitation or gelation of paraffins and asphaltenes, and condensation reactions catalyzed by FeS and other salts present to produce incrustation precursors [
3]. In this context, asphaltenes and polar molecules are the main causes of these phenomena and are potentiated under conditions of high temperature, pressure changes, and the nature of the solvent [
4].
Asphaltenes, as per their definition, are large polar molecules with high molecular weights, constituting a distinct fraction of oil that is soluble in toluene while insoluble in low-molecular-weight alkanes. They belong to the four primary categories found in petroleum, which are categorized based on their polarity, known as saturates, aromatics, resins, and asphaltenes (commonly abbreviated as SARA). In heavy oils, these molecules exist in a colloidal state and can undergo partial precipitation when changes occur in the fluid’s solubility properties due to factors such as oil blending, alterations in temperature, or shifts in pressure. The process of asphaltene deposition onto heated surfaces may be influenced by several factors, including the higher colloid instability index of the system and an enhanced mass transfer of the fluid. Consequently, a crucial aspect in assessing the escalation of asphaltene deposition rates on hot surfaces lies in comprehending the aggregation state of the precipitated asphaltenes [
5].
Deposition during oil processing, from the well and through the refinery, is so common that units need to be periodically stopped for cleaning. There are several incentives to mitigate deposition in refineries, as it is closely linked to energy, maintenance, and cleaning costs, as well as production losses due to plant shutdowns for cleaning. According to [
6], some control actions are applied to reduce the economic impact of fouling on refinery processes. For example, many companies invest in periodic chemical treatment programs to help significantly reduce fouling-related costs. A complete understanding of the fouling mechanism in heat exchange systems leads to designing chemical solutions that can achieve maximum fouling inhibition efficiency. Among the classes of chemicals used, we can mention dispersants to limit the size of solid particles in the system, corrosion inhibitors to minimize contact between the metal surface and the corrosive fluid, coordination metals that reduce the catalytic activity of the heat exchanger metal in fluid polymerization reactions, and polymerization inhibitors that reduce the free radical polymerization of olefins and some sulfur compounds present in petroleum.
Other solutions focus on predicting the deposition phenomenon based on the determination of thermal resistance to fouling (R
f), which, by definition, is the ratio of the deposit thickness (δf) formed on the heat exchanger wall to the thermal conductivity (λf) as the deposit increases. However, thermal conductivities are often deduced from changes in overall heat transfer performance and not from estimates of deposit thickness, which are therefore subject to large uncertainties [
7].
Thus, it becomes important to gather a set of experimental data for model development and for understanding and interpreting the phenomenon. There are several mathematical models that can predict the deposition or fouling rate over time. Semi-empirical system models consider the effects of operating conditions (temperature and velocity) and use experimental data to estimate parameters in a model based on the threshold fouling rate proposed by [
8,
9,
10].
In these models, the balance between the crude oil flow rate and the temperature needed to reach the activation energy for reactions and fouling formation determines the operating limit conditions so that the deposit does not initiate. To achieve this, understanding the fouling dynamics in heat exchangers can be obtained through a bench study aimed at replicating at a reduced scale the process conditions of crude oil preheating batteries and then feeding a validated model from the point of view of fouling rate and price; tests with carbon steel showed corrosion, and the authors suggest repeating the tests using stainless steel.
Two years later, the authors of [
11] evaluated deposition in heat exchangers using eleven different crude oils. Tests were performed with pre-filtration of the oil, using air and nitrogen to pressurize the system, and evaluating the composition of the oil in the deposition. The authors report that for light crude oils with low sulfur content, deposition is largely due to particles and gums. For medium sulfur crude oils, the formation of iron sulfides plays an important role in deposition. In unstable heavy oil systems, suspended asphaltenes are the fouling species. Traces of impurities, such as dissolved oxygen or suspended corrosion products, significantly increase deposit formation. The conclusion points out that the effects of velocity and temperature are more significant for adhesion of the fouling to the surface, to the detriment of the other parameters studied.
Valle [
12] investigated the fouling mechanisms of a light crude oil (density of 0.83 g/cm
3) based on its characterization and fouling tests using a bench-scale. The SARA analysis showed a quantity of saturates, aromatics, resins, and C5-asphaltenes of 58.39 wt.%, 30.91 wt.%, 7.18 wt.%, and 3.51 wt.%, respectively, which provided a high colloid instability index of 1.63, indicating that asphaltenes are unstable in the crude oil sample. Tests were performed with oil volume between 400 mL and 600 mL. The system was pressurized at 41.4 bar, and the working flow rate was 1 mL/min. The inlet temperature was 65 °C. Once the test started, the inlet and outlet temperatures of the oil in the heat exchanger were automatically collected every 6 s; the surface temperature (Ts) was measured through a movable thermocouple that can be positioned at 13 points along the tube from z = 0 mm to z = 60 mm. The fouling deposits were characterized by elemental analysis, scanning electron microscopy (SEM), thermogravimetric analysis (TGA), and photoacoustic infrared spectroscopy (PAS-IR). A fouling mathematical model was developed under a laminar flow regime following Epstein’s methodology. It was found that the fouling mechanism of the crude oil follows a laminar flow regime, with the transport of unstable asphaltenes to the hot surface, becoming fixed on this surface, and by means of chemical reactions, forming fouling deposits. The mass transfer of suspended particles carried in the crude oil also contributes to fouling, although it is not the main cause of the problem. However, according to the authors, under conditions of turbulent flow, such as those prevailing in industrial operations, suspended particles are expected to play a larger role in fouling.
Ho [
13] developed a combined experimental and modeling study aimed at relating crude oil properties to their intrinsic fouling tendencies. The experiments were conducted in a bench-scale with five types of crude oils in laminar shear flow. A mathematical model was developed by dividing the system into two subsystems to analyze the temperature drop data with the rate of fouling thickness growth from time series. The results yielded a quantitative understanding of fouling dynamics, primarily influenced by two dimensionless parameters, denoted as “/” and “r”. The former represents the Stanton number, reflecting the system’s heat transfer characteristics, while the latter is the product of a fouling Biot number and a speed ratio. Key properties of crude oil, such as S
BN/I
N (solubility blending number/insolubility number
), n-heptane insoluble (NHI), T
BN (total base number), and metals, are identified as significant factors in fouling. In contrast to previous studies focusing on the consequences for process equipment, this analysis addresses the timing of fouling occurrences, allowing for preventive measures. The correlation between fouling and properties could aid in developing advanced chemical treatments to minimize fouling and optimize the use of different crude oils.
Xing et al. [
14] studied the fouling behavior of thermally processed bitumen products obtained by visbreaking. The raw material was prepared by mixing crude oil bitumen with 15% volume of hydro-treated light oil. The visbreaking products obtained at different pitch conversions (525 °C + fractions) were characterized and tested in the laboratory for their fouling tendency in a benchtop heat exchanger test instrument. In order to further investigate the isolated contributions of olefins and asphaltenes to the fouling tendency, a selected visbreaking product was distilled into an olefin-rich fraction (less than 280 °C) and heavier fractions (greater than 280 °C) to avoid the loss of lighter fractions during the deasphalting process. The test results showed that the fouling tendency increased linearly with the total olefin content in the visbreaking products. It was found that higher conversions resulted in products with higher fouling tendency. The fouling tendency of visbreaking products reduced with increasing
p-value (ratio of solubility number to insolubility number) and increased with increasing olefin content. The deasphalting process significantly reduced the fouling tendency of the visbreaking product. With the presence of olefins, asphaltenes had a significant effect on fouling tendency, while without them, the effect of asphaltenes on fouling tendency was less significant.
The objective of this work was to perform an experimental study on the deposition dynamics of crude oil blends in a benchtop heat exchanger test instrument (BHETI). By collecting experimental data, resistance fouling (Rf) was calculated. The benchtop heat exchanger test instrument is a dynamic laboratory test used to predict scaling phenomena on a reduced scale in an accelerated manner to simulate industrial process conditions and determine the scaling tendencies of a fluid in the process unit. Finally, a study of the oil fractions present in the blends was conducted using the SARA method to investigate their influence on the deposition phenomenon. Based on the percentage content of each oil fraction, it was possible to determine geochemical parameters of oil maturity/degradation and compare them using the Colloidal Instability Index (CII) or ternary phase diagram. This work aimed to understand the effect of oil composition on deposition and how this information can be used to prevent or mitigate its effects, thereby reducing the impact on the equipment of a plant.
2. Materials and Methods
Deposition within heat exchangers represents a multifaceted phenomenon, characterized by a multitude of influencing parameters and a considerable time frame for its manifestation. Accurately predicting deposition remains an arduous endeavor. Consequently, specialized equipment has been engineered to expedite the simulation of deposition within a condensed timeframe. These apparatuses exert substantial stress on samples, thereby enabling the assessment of an oil’s propensity to generate deposits. Termed as benchtop heat exchanger test instruments (BHETI), these devices serve as invaluable tools in this context.
2.1. Deposition Tests in the BHETI
In this study, deposition experiments were conducted using an Alcor Petrolab BHETI-1 apparatus on two offshore field crude oil blends referred to as blend A and blend B. These experiments were carried out at the Fine Chemistry, Pharmaceuticals, Biotechnology, and Beverage Laboratory located within the SENAI CIMATEC Park. A schematic diagram of the experimental setup is shown in
Figure 1. The BHETI unit and the programmable logic controller (PLC) are shown in
Figure 2.
The characteristics of blends A and B are presented in
Table 1.
Testing involved a 50% volume blend mixture of blend A and blend B, referred to as MB 50%A/50%B. The benchtop heat exchanger test instrument (BHETI) unit facilitated experimentation under a comprehensive range of monitored conditions, encompassing fluid temperature, system pressure, and volumetric flow rate control, as shown in
Table 2.
The unit boasted a 2 L sample reservoir equipped with a heating jacket and temperature control via a thermocouple. Additionally, it featured an automatic agitator for efficient mixing. Transport lines were equipped with heating blankets for temperature maintenance.
The heat transfer section, known as the “hot finger”, was a shell and tube heat exchanger with a 1.5 mm annular diameter and a length of 60 mm, constructed from 1018 steel. It was equipped with two thermocouples for precise fluid temperature control and one thermocouple for monitoring wall temperature. The blends were pumped from the reservoir through the pipeline to the hot finger, which was electrically heated to facilitate a single passage of the oil through the section, leading to the formation of fouling deposits.
The system’s operational range extended to pressures of up to 500 psi of nitrogen (34.5 bar) and flow rates of up to 20 mL/min. A programmable logic controller was employed to regulate key test variables, including fluid flow rate and wall temperature. Various test conditions were investigated, including wall temperatures of 250 °C and 400 °C, as well as flow rates of 1 and 7 mL/min. The test duration was fixed at 4 h, with the system pressurized to 10 bar using nitrogen gas. Long-term tests were also conducted to assess the influence of time on deposition rate under specific conditions of 300 °C, 10 bar, and 1 mL/min. Following each test, the hot finger underwent heptane cleaning and subsequent weighing to evaluate the presence of deposited materials.
2.2. Gravimetric Determination of Fractions of Blends A and B by Open-Column Liquid Chromatography—SARA Method
Open-column chromatography is a liquid chromatography technique used for the separation of compounds based on their physicochemical affinity. The chromatographic system consists of a glass column with a specific size and diameter. The column contains two main phases that allow for the chromatographic process, where various components of a sample can be separated through the differential interaction between these phases. The first stationary phase, consisting of silica, alumina, or a mixture of phases, is responsible for separating the components of the sample. The second mobile phase, consisting of organic solvents with different degrees of polarity, is responsible for transporting the sample constituents according to their physicochemical affinity (intermolecular forces) with the stationary phase.
In the fractionation by the SARA method (saturates, aromatics, resins, and asphaltenes), based on ASTM D2007 [
15] and D4124-09 standards [
16], open-column liquid chromatography is used to separate the main groups of compounds in a petroleum sample. Based on their level of molecular complexity and polarity, this type of physicochemical analysis enables the separation of petroleum compounds into three main fractions called saturated hydrocarbon compounds, aromatic hydrocarbon compounds, and NSO compounds (nitrogen, sulfur, and oxygen), which can be considered as a mixture of resins and asphaltenes.
In this work, the petroleum blends and their mixtures were characterized by their fractions using the SARA method. The analyzed samples were blend A; blend B; blend of blends A and B (AB 50%/50%); blend A after deposition test (AP2); blend B after deposition test (BP2); and blend of blends A and B after deposition test (MB 50%/50%). The analyzed samples are catalogued and showed in
Table 3.
Based on the percentage content of each petroleum fraction, it is possible to determine geochemical parameters of petroleum maturity/degradation and compare it through the Colloidal Instability Index (CII), presented in Equation (1), or a ternary phase diagram.
2.3. Calculation of Rf
The effect of deposition on heat exchanger equipment calculations involves the deposit resistance (fouling factor), which is added to the convective and conductive thermal resistances in determining the overall heat transfer coefficient in the equipment affected by the deposit. This effect of adding thermal resistance in heat exchangers occurs with deposit adhesion, which occurs on both the shell and tube sides, with the influence of deposit on the hot fluid side being relevant according to the temperatures and flow rates involved.
The calculation of operating efficiency in heat exchangers in industrial projects is performed from the energy balance, where the heat transfer rate (Q) per unit area of the exchanger (A) is written as the ratio of the average thermal variation (ΔT) and the overall heat transfer resistance (R
T), which, by definition, is the inverse of the overall heat transfer coefficient (U), as can be seen in Equation (2).
By definition, the amount of heat can be expressed by Equation (3).
where ṁ is the mass flow rate, C
p is the specific heat, and ΔT is the temperature difference of the fluid. Along the length of a heat exchanger, both the hot and cold fluids can experience non-linear temperature changes; therefore, the average temperature difference (ΔT) is better defined as the logarithmic mean temperature difference. Thus, from Equation (1), we can obtain Equation (4).
where ΔT
L is the logarithmic mean temperature difference.
For industrial shell and tube heat exchangers, a correction factor (F
c) is added to the product of Equation (3) to minimize the effects of concurrent and countercurrent flow temperature on this type of exchanger, resulting in Equation (5).
where F
c is the correction factor for the number of passes in the tubes of the heat exchanger. Equation (6) is obtained as a function of the overall heat transfer coefficient for the clean exchanger conditions (U
c), based on the results generated during the first 15 min of the test. The Equations (2)–(6) are developed by mass and energy transfer and flow laws, and these were obtained in [
17].
According to [
18], fouling deposits are commonly quantified by estimating the deposit over time, which is called the instantaneous fouling resistance R
f(t) and can be determined from the difference of the inverse of the overall heat transfer coefficients under fouling and clean conditions using Equation (7).
According to [
2], another parameter used to quantitatively evaluate the fouling propensity of crude oils is the percentage of fouling at time t (%F
t). %F
t is a value that can be directly provided by the HLPS equipment for an instantaneous R
f, as seen in Equation (8):
Therefore, based on Equations (7) and (8), we can obtain Equation (9).
3. Results and Discussion
Table 4 presents the input and output variables of BHETI for the oil deposition tests carried out with blend A and blend B. The pressure and test time were fixed at 10 bar and 4 h, respectively. The tests were conducted in duplicate, and it was possible to determine an average deposition rate for each test, since no significant variations in the deposition rate were observed during the 4 h operation of BHETI.
It was observed that the inlet and outlet temperatures of the exchanger (dependent variables) were influenced by the wall temperature. The bulk temperature represents the average equilibrium temperature of the moving oil and was calculated as the average be-tween the inlet and outlet temperatures of the oil in the exchanger.
The results indicate that, for all tests, blend A was more susceptible to deposition than blend B, as evidenced by higher R
f values, suggesting that in blend A, due to its higher concentration of light hydrocarbons, there was greater instability of paraffinic chains [
19].
The wall temperature significantly influenced the deposition rate of the oils, where higher temperatures were accompanied by higher deposition rates. This phenomenon may be associated with polymerization reactions of the oil on the wall of the exchanger due to high surface temperatures. However, the oil flow rate showed an inversely proportional behavior to the deposition rate, where lower flow rates were accompanied by higher Rf values. This effect can be explained by the shearing flows that can naturally remove much of the material deposited on the exchanger wall if the flow evolves into a turbulent system or promotes an increase in deposition if the flow on the exchanger surface tends towards a laminar regime.
Thus, increasing the wall temperature combined with a reduction in oil flow in the BHETI generated a synergistic effect, increasing the values of the instantaneous deposition rate (R
f).
Figure 3 and
Figure 4 show the effect of temperature over time on different flow rates (1 mL/min and 7 mL/min, respectively). As mentioned earlier, at higher temperatures, higher deposition rates were observed, suggesting that this phenomenon was caused by an increase in the volatility of light-saturated compounds with a greater propensity for oil polymerization due to high temperatures. We observed that the shear rate caused a significant reduction in the deposition rate, as the shear effects of the flow promoted the removal of deposits on the exchanger surface. In deposition rate prediction models (deposition threshold models), such as the [
8] models, this phenomenon was corrected with a removal term, where the Reynolds number or flow velocity was considered.
In addition, it is interesting to note in
Figure 3 that over 3 h of testing, no changes in R
f values were observed. The deposition rate remained constant throughout the analysis, and even in long-term tests, this behavior was also observed.
In
Figure 5, a long-term test is presented comparing blend A under flow conditions of 1 mL/min and 300 °C with a mixture of blends A and B in a 50% blend A and 50% blend B mass ratio under the same conditions. As discussed previously, the deposition rate of individual blends A and B did not change over time, remaining constant. What we can observe is that the sample of the blend mixture underwent a variation in the deposition rate over time, where Rf increased as time increased, indicating that fouling processes were occurring on the heat exchanger surface.
If the behavior of the blend mixture graph is observed in more detail, it is possible to see periods of deposition interspersed with periods of removal over the 7 h test period. Possibly, the pressure increase caused by the decrease in the tube’s cross-sectional area due to deposition caused a momentary removal of fouling on the heat exchanger surface, subsequently returning to the incrustation process.
Thus, it is suggested that the blend mixture may have induced a thermodynamic imbalance in the petroleum. This was likely due to suboptimal behaviors of hydrocarbons and complex chemical reactions, potentially enhancing the instability of asphaltene compounds in the medium, thereby promoting the continuous occurrence of this phenomenon throughout the test duration. In accordance with Fan et al. [
18], in the presence of laminar flow conditions (1 mL/min), unstable asphaltenes tend to migrate to the hot finger surface, where they adhere and subsequently undergo chemical reactions, leading to the formation of deposits. By the end of the tests, the rate of deposition nearly doubled when compared to the initial value.
Table 5 presents the SARA method results for different oil blends and their mixtures along with the calculated values of the Colloidal Instability Index (CII) for different scenarios of the NSO fraction. The NSO fraction represents a group of nitrogen-, sulfur-, and oxygen-containing organic components that include asphaltenes and resins. Separating this fraction into asphaltenes and resins would lead to considerable sample losses and excessive solvent expenses, making it impractical to fractionate.
Since Equation (6) depends on the separated fractions of asphaltenes and resins, three scenarios were considered for the CII to improve the interpretation of its values. In the first scenario, it was assumed that all NSO was composed of asphaltenes; in the second scenario, all NSO was composed of resins; and in the third scenario, the NSO was composed of 50% asphaltenes and 50% resins.
One way to interpret CII is that for values equal to or greater than 0.9, the asphaltenes are unstable in the mixture of oil fractions, while for values below 0.7, the oil is unstable for deposition. For values between 0.7 and 0.9, it is not possible to conclude the stability without specific tests.
According to
Table 4, we can see that most samples have a saturation level of 50% or more, which favors a greater increase in the numerator of the CII equation, raising the instability of asphaltenes in the medium with values above 0.9. Even in situations where NSO was considered as 100% resins, the CII results reported values of 0.98 (AP2) and 0.96 (MB 50%/50%), confirming the instability of asphaltenes in the fractions of the petroleum blends, which could promote greater depositions. This way, as blend A CII index is greater than blend B, blend A stability to deposition is lower than blend A.
Another observation was that after the samples went through the deposition tests in BHETI, the saturation levels reduced, and the aromatics increased considerably, with NSO fractions remaining practically constant. This could be associated with greater losses of volatile saturated compounds due to high thermal exchange temperatures and the formation of polyaromatic compounds by polymerization of the chains, characteristic of degraded petroleum.
In
Figure 6, ternary diagrams of saturates, aromatics, and NSO of the tested samples are presented along with the one presented by [
7], which shows the trend of petroleum towards deposition.
The diagram, along with the results from
Table 4, shows that blend A had a higher content of NSO compounds, which comprise the high molecular weight polycyclic fraction, than blend B. Asphaltenes, components of this fraction, are insoluble in light alkanes and, therefore, precipitate with n-hexane, while resins are more soluble but highly polar. These data support the results found in
Table 3 where blend A was more susceptible to deposition compared to blend B. Therefore, blend A was found to have characteristics of crude oils in general, including tar compounds, while blend B behaved more like normal production crude oils.