Feb-2020

Estimating delayed coker yields

Correlations were developed for estimating product yields from delayed coker units to achieve optimum performance and refinery configuration.

KARTHIK RAMESH
Indian Oil

Article Summary

The delayed coking unit (DCU) is one of the most important carbon rejection secondary processing units for upgrading heavier fractions of crude oil residues (atmospheric and vacuum residue). The DCU is employed to produce a wide range of more useful products such as lighter hydrocarbons, LPG, gasoline, light gasoil, and heavy gasoil through thermal cracking. Delayed coking has been selected by many refiners as their preferred choice for bottom of the barrel upgrading due to its inherent flexibility to handle even the heaviest of residues while producing clean liquid products. The process provides essentially complete rejection of metals and relatively high conversion to LPG, naphtha, and gasoil. Thus, uninterrupted operation of the DCU is of the utmost importance to ensure profitability and higher levels of crude processing in the refinery. The goal is to operate the coker to maximise the yield of clean distillates and minimise the yield of coke.

The feed to the DCU is the most important process variable. The feed has the greatest impact on operating conditions, yield, and product quality. Feed quality depends on several factors including the quality and type of crude processed by crude distillation units and other processing units, the complexity of the refinery, and so on.

DCU feeds are blends of atmospheric and vacuum residue, unconverted residues from ebullated bed hydrocracking, and FCC clarified oil among others. The ability to predict the impact of changing feedstock quality on product yield, operating conditions, and product quality will enhance the scope for increasing gross refinery margin.

The yield from a DCU can be varied to meet the refiner’s objectives through selection of operating parameters (pressure, temperature and recycle ratio). These operating variables control yield and product quality. Increasing coking temperature decreases coke production and increases liquid yield and gasoil end point. Increasing pressure for the recycle ratio increases gas and coke make and decreases liquid yield and gasoil end point.

The most important point is that these potential yields are calculated using properties typically analysed in refinery laboratories, for instance Conradson carbon residue.

In this article, we analyse and deduce correlations for estimating product yields for DCUs. In addition, linear programming was performed and the effect on economics was analysed with changes in crude processed in the refinery. The correlations developed have been used in selecting a configuration for the capacity expansion of a refinery in India.

Conradson carbon residue (CCR)
The carbon residue of crude oil, heavy oil, or residue is the wt% of coke that remains after evaporation and cracking of a sample in the absence of air under specified conditions. Depending on the testing procedure, the result is reported as Conradson or Ramsbottom carbon residue in wt% units. The carbon residue indicates the asphalt content of oil or the amount of lubricating oils that can be produced from processing of the oil. It is normally measured by the ASTM B189  method.

Figure 1 shows typical values for CCR values from vacuum residue (540°C to FBP) derived from Middle Eastern crudes.

Estimation of product yields can be carried out using the correlations shown in Table 1 based on wt% of CCR in vacuum residue feeds in the CCR range 10-29.

Light and heavy naphtha yields are calculated based on TBP cut points and typical percentage yields of total naphtha from an operating DCU.
A comparison of product yields from two crude mixes based on these correlations and a commercially operating DCU is shown in Figure 2 and Figure 3.

Estimation of product yields can be carried out using the correlations shown in Table 2 based on wt% CCR in vacuum residue or atmospheric residue for CCR ranges below 10.

Based on these correlations, a comparison of product yields from a commercially operating DCU is shown in Figure 4.
As Table 2 shows, the product yields calculated from correlations are similar to product yields obtained from commercially operating units, with little error.

Application of correlations for a refinery configuration
The economics of petroleum refining are extremely complex. Crude oil can be processed by many possible refinery configurations. Each configuration includes various technology options that depend on market demand for finished products. In a competitive market, refiners are driven to optimise all potential investment options and select those options that provide the greatest profit. The expected refining margin will thus govern the adaptation of refining facilities to meet market demand. The refiner must meet fuel quality constraints and supply growing product demands while maximising profit. Thus, possible combinations of technology options can be considered. A competitive market drives the refiner to select the optimum refinery configuration that satisfies multiple objectives.

Linear programming (LP) algorithms are great mathematical tools to handle this complex problem efficiently. These are amenable to case studies and comparative statements which help in selecting a configuration and in justifying investment decisions.

A refinery wide linear programming model was run for both the crude cases (crude mix 1 with CCR 28.3 and crude mix 2 with CCR 26.8). Table 3 shows the findings of output results from the model.

In Table 3, Case 1 considers yields from correlations for a DCU for crude mix 1 (CCR 28.3). Case 2 considers the same yields from a DCU for crude mix 1 as the yield pattern of crude mix 2 (CCR 26.8).

As per the LP model’s results, it can be seen that incorrect consideration of DCU yield values results in a GRM loss of $15 million and a mismatch of secondary processing units and overall hydrogen consumption of the refinery. The above correlations can be used accurately for refinery configuration studies.

The correlations can also be used for validating crude for refineries, and for estimating product yields for new crudes processed in the refinery. They can be utilised in a linear programming model of a refinery, and to estimate or compare economics and refinery margins. Further, process parameters (temperature, pressure, recycle ratio) can be fine-tuned to achieve the predicted yields using the correlations to achieve optimised operation of a DCU.

Conclusion
The Conradson carbon residue of DCU feed is an important property to define and study. The knowledge developed by using the correlations in this article can be used to make better decisions and to optimise the DCU. The correlations can be used accurately for refinery configuration studies, validating crude for the refineries, estimating product yields from new crudes processed in the refinery, and for adjusting process parameters to achieve the most economical product pattern.