THERMAL CRACKING OF PETROLEUM FEEDSTOCKS: KINETICS AND PROCESS MODELLING STUDIES

Abstract

The residues obtained after distillation and solvent deasphalting process are the least valuable of the refinery streams. These residual feedstocks pose a lot of refining problems, due to the presence of a large amount of carbonaceous material as well as metals. These carbonaceous components and metals make the processing difficult due to coke formation as well as catalyst deactivation and poisoning. Due to this reason, the thermal cracking process becomes more attractive, as compared to catalytic one, for the upgradation of residual feedstocks. The refinery engineers are gaining interest in residue upgradation due to the continuing decline in fuel oil markets and growing demand for transportation fuels. In addition, refiners are exploring the production of transportation fuels from low value residuals rather than from increased crude processing. Residue upgradation technologies are being developed worldwide, with the objectives of improving the process and catalyst performance, lowering capital cost, and addressing safety and environmental concerns. An attempt has been made in the present thesis, to gain a deeper insight into the thermal cracking behavior of residual feedstocks of different origins and varying physical characteristics. The mild form of thermal cracking, also called Visbreaking, is very old process, yet the subject of its kinetic modelling did not get much attention initially. The reason was lack of understanding of reactions because of their complex nature. A deeper understanding is mandatory for the efficient design and operation of industrial visbreaker. An exhaustive literature review suggests that the models available for the thermal cracking of residual feedstocks, are related to severe cracking conditions, where the coke formation is an integral part of the process. Further, the detailed study of thermal cracking in terms of lumps of distillate fractions is not available, which is required for the detailed modelling of a mild thermal cracking process. In view of the above, the present study has been undertaken with the objectives to conduct experiments on the thermal cracking of residual feedstocks, and to develop a suitable kinetic model, which may be useful for the design of industrial units. Experiments were conducted in a 400 ml Stainless Steel batch reactor fabricated in-house. A common inlet / outlet was provided for the charging and evacuating the reactor. Provisions were made to measure liquid, and vapcr temperatures separately, and pressure as well. For pressure regulation and gas discharge during and after the reaction, a needle valve was provided. The heat required for the reaction was provided by a salt bath containing a eutectic mixture of NaN03, KN03, and NaN02 in the ratio of 7, 53 and 40%respectively. Four residual feedstocks of Indian and Middle East origin, which are processed in Indian refineries, namely North Gujarat short residue (NGSR), Bombay High short residue (BHSR), visbreaker feed from Mathura Refinery (MVBF), and asphalt from Haldia refinery (HRA) were taken for the kinetics studies. The feedstocks were taken in such a manner so that a wide range of certain key characteristics of practical interest is covered, in order to study their effect on the cracking behaviour. The kinetics experiments were carried out with 120 g of vacuum residue (feed), at a Nitrogen pressure of 12 kg/cm2(g),and four temperatures from 400 to 430 °C, at an interval of 10 °C. The gas formed during the cracking is measured by releasing through the gas meter. The exit gas was collected and analyzed by gas chromatography for some representative runs. It was observed that the outlet gas composition did not vary significantly during experiments. The liquid product from the reactor contains unconverted vacuum residue and cracked products. It is quantified, and the distillate fraction boiling up to 500 °C is separated out by atmospheric and vacuum distillation. The distillate fraction is analyzed by ASTM D-2887 (SIMDIST). Analysis of the cracked sample has yielded percent distilled at various temperatures with cut range of every 1% distilled, which provided the flexibility to make the industrially important lumps (fractions) for the analysis. Therefore, the results were analyzed with four product lumps, namely Gas (C5 -) collected at room temperature, Gasoline (IBP-150 °C), Light Gas Oil (150-350 °C), and Vacuum Gas Oil (350-500 °C). The results from the present study confirm the first order reaction kinetics for the overall cracking of vacuum residues, as reported earlier in the literature. The activation energies for the conversion of residues have been found of the order of 102-206 kJ/mol. A five lump kinetic model was hypothesized for the description of the thermal cracking. The lumps considered are vacuum residue feedstock, and four cracked products namely Gas, Gasoline, Light Gas Oil and Vacuum Gas Oil. Initially, a ten kinetic parameter model was hypothesized. The rate constants were estimated numerically with the experimental data, using differential evolution method, which is based on the genetic algorithm. Delplot analysis [Bhore et al. (1990)] was performed in order to establish the reaction pathways. The data were also analysed on the basis of selectivities of the products. The selectivities of all these product lumps were computed. These data reveal that the selectivities of Gas, Gasoline, and LGO lumps show an increase with the increase of residence time, whereas the selectivity of VGO decrease with increasing residence time as the reaction temperature is increased. This indicates the further conversion of VGO to LGO and GLN fractions at increased severity. The analysis of data resulted in the reduction of rate constants from ten to seven. Consequently, a five lump seven parameter kinetic model was yielded. The reaction rate constants as well as activation energies were again estimated for all the established pathways. The pathways evolved were further confirmed on the basis of instantaneous fractional yields of the lumps. It is our view that the detailed analysis of experimental results of present investigation and developed five lump seven parameter kinetic model shall in render a deeper insight into the thermal cracking of residual feedstocks for the design and optimal operation of industrial units.