Mar-2015

Light tight oil crude unit fouling root causes and troubleshooting

Fouling in preheaters, crude furnace and atmospheric tower bottoms exchangers directly impacts throughput and profitability, requiring rigorous evaluation strategies

Gregory Savage and Kailash Sawhney
Nalco Champion

Article Summary

Crude preheat and furnace fouling is an observed problem in processing price advantaged crudes like light tight oils (LTO). Foulant material has been found in the crude preheat train upstream of the crude furnace in the crude side of the exchangers, in exchangers processing the atmospheric tower bottoms stream, and in crude heaters. As the preheat exchanger train fouls, the temperature of the crude furnace inlet falls. Consequently, trending the crude furnace inlet temperature (FIT) loss can provide a measure of preheat exchanger train fouling. Figure 1 shows the effect of processing light tight crude slate on FIT loss at a US refinery.

Deposition in preheat exchangers lowers heat transfer rates, thus decreasing the temperature of the crude as it enters the furnace. This requires increased furnace firing (higher duty), which increases tube metal temperatures. Deposition in furnace tubes also results in increased tube metal temperatures. Preheat exchanger train and furnace fouling results in greater energy usage, reduced unit throughput and more frequent heater decoking.

Crude units processing high proportions (>75%) of LTO frequently experience high crude furnace fouling rates often measured by a tube metal temperature (TMT) increase. Figure 2 shows the crude furnace normalised TMT increase for three different US refineries processing greater than 75% LTO. Prior to processing LTO, crude heater run lengths for these refineries were one to four years. As shown in Figure 2, typical run lengths are less than 180 days, and in severe cases may be as short as 30 days.

Blending of paraffinic crudes, such as LTO, with heavier asphaltenic crudes can result in lower crude stability and potential asphaltene precipitation. Resins in crude oil are bound to large asphaltene structures forming micelles, which serves to keep the asphaltenes suspended and dispersed in the crude. Loading and blending crudes containing asphaltenes with paraffinic crudes can reduce the asphaltene stability by solvating the resins and weakening the resin-asphaltene interaction.² The crude blend stability change with increasing LTO is seen in Figure 3. As the proportion of LTO increases, the overall crude stability decreases.

The crude stability index (CSI) is a measure of the stability of fouling precursors in the crude (predominantly the asphaltenes) and is measured by titration with an aliphatic solvent. The ‘peak’ in the titration curve (the so-called ‘flocculation point’) is indicative of their stability. The amount of solvent added at the peak is noted and converted to a CSI value. This measurement allows a direct comparison of crudes and crude slates and indicates the fouling tendency due to asphaltene destabilisation of the crude prior to refinery processing.

The proportions of each crude type and the order in which they are mixed also strongly determines the potential for asphaltene destabilisation. Once the asphaltenes agglomerate and deposit, re-solvating the asphaltenes requires time, temperature and strong solvents, and is not always successful in field applications. Consequently, it is important to determine blend ratios and the order of loading and blending. Titration of an aliphatic hydrocarbon into crude and aromatic solvent blends produces flocculation points that can be used to determine the crude’s intrinsic stability (S), oil solvency power (So), and the (Sa) ability of the asphaltenes to remain in a colloidal dispersion (modified ASTM D7157-05).

The (Sa) is linked to the solubility of the asphaltenes due to its size and structure. The oil solvency power (So) is the aromatic equivalency of the oil. The intrinsic stability (S) is a measure of oil’s available solvency power with respect to its asphaltene solubility.2 Figure 4 shows the resulting change in intrinsic stability as a consequence of blending order. Although the overall blend is intrinsically stable, by loading the asphaltenic crude over the LTO, the resulting blend stability is measurably reduced.

The strength of the resin-asphaltene interaction decreases upon heating as the resins are disassociated from the asphaltenes. This allows the asphaltenes to agglomerate and form particles of foulant. The extent to which the asphaltenes are stabilised at higher temperatures depends upon the strength of the asphaltene-resin interaction, which can be reduced through blending operations as previously discussed. The more strongly the resins are bound to the asphaltenes, the less prone the asphaltenes are to agglomeration and fouling. Fouling from destabilised asphaltenes is often found in the hottest exchangers in a preheat train.

The stability parameters of LTO are difficult to measure experimentally due to the relatively low amount of asphaltenic material present. However, the parameters may be calculated using a modified ASTM D7157-05. An additional industry stability measure is the colloidal instability index (CII) and is defined by Equation 1 below:⁴

Colloidal instability
         index =    (saturates + asphaltenes)
                                       (aromatics + resins)

The saturates, aromatics, resin and asphaltene (SARA) analysis results in the loss of volatile species, which can include both saturates and aromatics. Because of the material loss, the CII is never used in isolation, but instead used in conjunction with other test methods to better understand relative stability risks. Generally, a CII greater than 1.2 is considered a potential stability risk. The CII of tight oil crude blends tends to vary between 3 and 12, which indicates potential stability risks with heptane insoluble material such as asphaltenes or coke precursor (polycyclic aromatic hydrocarbon) molecules.

LTO samples subjected to thermal stressing under an inert atmosphere (N2) for an extended period of time then treated with hexanes show increased precipitation relative to unstressed samples, as shown in Figure 5. When thermally stressed, the associations between polar components of the LTO with stabilising species such as resins are disrupted. As a result, these polar materials become unstable in the LTO paraffinic environment. This phenomenon is even more pronounced when the LTOs are exposed to the much higher temperatures that occur in crude heaters and hot preheat exchangers. Under these conditions the heavy components of the LTO can undergo dehydrogenation, resulting in the formation of cyclic aromatic structures that can undergo further fusion to form polycyclic aromatic hydrocarbons (PAH), which may eventually lead to coke formation.

The precipitated polar aromatic particles were also observed in hot liquid process simulator (HLPS) testing, as well as in field deposits taken from exchangers processing atmospheric tower bottoms streams (see Table 1). The HLPS flow loop evaluates fouling mechanisms and chemical performance at representative temperatures and under dynamic conditions. Deposit weights from the HLPS indicate that foulants are typically 0.2–50 mg/L in the crude. The organic portion of the deposit is typically the majority of the foulant and has a carbon to hydrogen ratio between 9–13, with less than 30% dichloromethane extractable, which indicates the formation of coke precursor (ie, PAH molecules). As seen in Table 1, inorganic solids not removed in the desalter can bind with organic foulants and contribute to deposits on heat transfer surfaces.